MADRONO A WEST AMERICAN JOURNAL OF BOTANY VOLUME XLVII 2000 BOARD OF EDITORS Class of: 2001 — Robert Patterson, San Francisco State University, San Francisco, CA Paula M. Schiffman, California State University, Northridge, CA 2002 — Norman Ellstrand, University of California, Riverside, CA Carla M. D'Antonio, University of California, Berkeley, CA 2003 — Frederick Zechman, California State University, Fresno, CA Jon E. Keeley, U.S. Geological Service, Biological Resources Division, Three Rivers, CA 2004 — David Wood, California State University, Chico, CA Ingrid Parker, University of California, Santa Cruz, CA Editor — Kristina A. Schierenbeck California State University, Chico Department of Biology Chico, CA 95427-05 1 5 kschierenbeck@csuchico.edu Published quarterly by the California Botanical Society, Inc. Life Sciences Building, University of California, Berkeley 94720 Printed by Allen Press, Inc., Lawrence, KS 66044 Madrono, Vol. 47, No. 4, pp. ii-iii, 2000 TABLE OF CONTENTS Alverson, Edward R. (see Zika, Peter E, Edward R. Alverson and Loverna Wilson) Baldwin. Bruce G., President's report for Volume 47 287 Baldwin, Bruce G., Roles for modern plant systematics in discovery and conservation of fine-scale biodiversity 219 Baldwin, Bruce G. (see also Mishler, Brent D., et al.) Barkley. Theodore M., Floristic studies in contemporary botany 253 Barron, Robin (see Corbin, Beth Lowe) Beidleman. Richard. Willis Linn Jepson — "The botany man" 273 Beidleman, Richard (see also Jepson, Willis Linn) Bermejo- Velazquez, Basilio (see Ledig, F. Thomas, et al.) Boyd, Robert S., Michael A. Wall, and James E. Watkins, Jr., Correspondence between Ni tolerance and hyperaccumulation in Streptanthiis (Brassicaceae) 97 Boyd, Steve (see Provance, Mitchell C., et al.) Boyd, Steve (see also Soza, Valerie, et al.) Bramlet, David (see Provance, Mitchell C, et al.) Brooks, Matthew L., Review of A Natural History of the Sonoran Desert ed. by S. J. Phillips and P. W. Comus 68 Bruederle, Leo P. (see Kuchel, Shannon D.) Buckmann, Allan (see Hrusa, G. Frederic, and Allan Buckmann) Ceska, Adolf (see Christy, John A.) Ceska, Oldriska (see Christy, John A.) Charlet, David A., Coupling species-level inventories with vegetation mapping 259 Christy, John A., Oldriska Ceska and Adolf Ceska, Noteworthy collections from Oregon and Washington 212 Clark, Dina, and Tim Hogan, Notewothy collections from Colorado 142 Capo-Arteaga, Miguel A. (see Ledig, F. Thomas, et al.) Coleman, Ronald A., Noteworthy collection from Arizona 138 Constantine-ShuU, Helen, and John O. Sawyer, Noteworthy collections from California 209 Corbin, Beth Lowe, James L. Reveal and Robin Barron, Eriogomim spectabile (Polygonaceae): A new species from northeastern California 134 del Moral, Roger (see Wood, David M.) Dodd. Shana C, and Kaius Helenurm, Floral variation in Delphinium variegatum (Ranunculaceae) 116 Ertter, Barbara, Our undiscovered heritage: Past and future prospects for species-level botanical inventory 237 Ertter, Barbara (see also Jepson, Willis Linn) Ertter, Barbara (see also Mishler, Brent D., et al.) Ewing, Kern, Environmental gradients and vegetation structure on south Texas coastal clay dunes 10 Felger, Richard, Noteworthy collections from Arizona and Sonora, Mexico 21 1 Felger, Richard Stephen, Noteworthy collections from Sonora, Mexico 21 1 Flores-Lopez, Celestino (see Ledig, F. Thomas, et al.) Hart, Jefferey A. (see Hrusa, G. Frederic, and Jefferey A. Hart) Hartman, Ronald L., Review of Synthesis of the North American Flora. Version 1.0 by John T. Kartesz and Christopher A. Meacham 207 Helenurm, Kaius (see Dodd, Shana C.) Hipkins, Valerie D. (see Safiya, Samman) Hobbs, Richard J., Review of 2nd Interface Between Ecology and Land Development in California edited by J. E. Keeley, M. Baer-Keeley, and C. J. Fotheringham 206 Holiday, Susan, A fioristic study of Tsegi Canyon, Arizona 29 Hogan, Tim (see Clark, Dina) Hrusa, G. Frederic, Noteworthy collection from California 139 Hrusa, G. Frederic, and Allan Buckmann, Noteworthy collection from California 138 Hrusa, G. Frederic, and Jeffrey A. Hart, Noteworthy collection from California 138 Jepson, Willis Linn, Richard Beidleman and Barbara Ertter, Willis Linn Jepson's "Mapping in forest botany" 269 Knops, Johannes M. H., and Walter D. Koenig, Annual variation in xylem water potential in California oaks .. 106 Koenig, Walter D. (see Knops, Johannes M. H.) Kuchel, Shannon D., and Leo P. Bruederle, AUozyme data support a Eurasian origin for Carex viridula subsp. viridula (Cyperaceae) 147 Kuykendall, Keli (see Zika, Peter E, Keli Kuykendll and Barbara Wilson) Ledig. F. Thomas, et al.. Locations of endangered spruce populations in Mexico and the demography of Picea chihuahuana 71 Mahony. Thomas M., and John D. Stuart, Old-growth forest associations in the northern range of coastal redwood 53 Mapula-Larreta, Manuel (see Ledig, F. Thomas, et al.) Markos, Staci (see Mishler, Brent D., et al.) 2000] TABLE OF CONTENTS iii Mayer, Michael S., Laura M. Williams, and Jon P. Rebman, Molecular evidence for the hybrid origin of Opuntia prolifera (Cactaceae) 109 Meyers-Rice, Barry A., Ramona Robison and John M. Randall, Noteworthy collections from California 209 Mishler, Brent D., The need for integrated studies of the California flora 230 Mishler, Brent D., et al.. Introduction to The Jepson Herbarium 50th anniversary celebration and scientific symposium: Discovery, communication, and conservation of plant biodiversity in California 217 Moe, Richard, Electronic activities of the University and Jepson Herbaria 265 Odion, Dennis, Seed banks of long-unburned stands of maritime chaparral: Composition, germination behavior, and survival with fire 195 Parnell, Dennis R. (see Saroyan, J. Phillip) Parsons, Lorraine S., and Adam W. Whelchel, The effect of climatic variability on growth, reproduction, and population viability of a sensitive salt marsh plant species, Lasthenia glabrata subsp. coulteri ( Asteraceae) 1 74 Preston, Robert E., Noteworthy collections from California 138 Provance, Mitchell C, et al.. Noteworthy collections from California 139 Provance, Mitchell C. (see also Soza, Valerie, et al.) Randall, John M. (see Meyers-Rice, Barry A.) Rebman, Jon P. (see Mayer, Michael S.) Reveal, James L. (see Corbin, Beth Lowe) Reyes-Hernandez, Valentin (see Ledig, F. Thomas, et al.) Robison, Ramona (see Meyers-Rice, Barry A.) Safiya, Samman, Barbara L. Wilson and Valerie D. Hipkins, Genetic variation in Pinus ponderosa, Purshia tridentata, and Festuca idahoensis, commuity-dominant plants of California's yellow pine forest 164 Sanders, Andrew C. (see Provance, Mitchell C, et al.) Sanders, Andrew C. (see also Soza, Valerie, et al.) Saroyan, J. Phillip, Dennis R. Parnell, and John L. Strother, Revision of Corethrogyne (Compositae: Astereae) ... 89 Sawyer, John O. (see Constantine-Shull, Helen) Schierenbeck, Kristina A., Editor's report for Volume 47 288 Soza, Valerie, et al. Noteworthy collections from California 141 Soza, Valerie (see also Provance, Mitchell C, et al.) Spickler, James C. (see Stillett, Stephen C.) Stephens, Scott L., Mixed conifer and red fir forest structure and uses in 1899 from the central and northern Sierra Nevada, California 43 Stephenson, Nathan L., Estimated ages of some large giant sequoias: General Sherman keeps getting younger .... 61 Stillett, Stephen C, James C. Spickler and Robert Van Pelt, Crown structure of the world's second largest tree .. 127 Strother, John L., Hedosyne (Compositae, Ambrosiinae), a new genus for Iva ambrosiifolia 204 Strother, John L. (see Saroyan, J. Phillip) Stuart, John D. (see Mahony, Thomas M.) Tiehm, Arnold, The taxonomic history, identity, and distribution of the Nevada endemic, Plagiobothrys glome ratus (Boraginaceae) 159 Van Pelt, Robert (see Stillett, Stephen C.) Wall, Michael A. (see Boyd, Robert S.) Watkins, James E. (see Boyd, Robert S.) Wells, Philip V, Pleistocene macrofossil records of four-needled pinyon or juniper encinal in the northern Viscaino Desert, Baja California del Norte 189 Whelchel, Adam W. (see Parsons, Lorraine S.) Williams, Laura M. (see Mayer, Michael S.) Wilson, Barbara (see Zika, Peter E, Keli Kuykendll and Barbara Wilson) Wilson, Barbara L. (see Safiya, Samman) Wilson, Loverna (see Zika, Peter E, Edward R. Alverson and Loverna Wilson) Windham, Michael D., Chromosome counts and taxonomic notes on Draba (Brassicaceae) of the Intermountain West. 1: Utah and vicinity 21 Wolf, Adrian L. (see Provance, Mitchell C, et al.) Wood, David M., and Roger del Moral, Seed rain during early primary succession on Mount St. Helens, Washington 1 Zika, Peter E, Edward R. Alverson and Loverna Wilson, Noteworthy collections from Oregon and Washington .. 213 Zika, Peter E, Keli Kuykendll and Barbara Wilson, Noteworthy collections from Oregon 144 Dates of Publication of Madrono, Volume 47 Number 1, pages 1-70, published 8 December 2000 Number 2, pages 71-146, published 5 March 2001 Number 3, pages 147-216, published 25 June 2001 Number 4, pages 217-300, published 31 August 2001 t UNITED STATES Statement of Ownership, Management, and Circulation 1 POSTAL SERVICEr. (Required by 39 USC 3685) Madrono 3. Filing Date 4 Issue Frequency Quarterly 6. Number of ssues Published Annually 6. Annual Subscription Price $27 00 California Botanical Society. Inc.: Herbana. Life Sciences Building University of California; Berkeley. CA 94720 Contact Person Roy Buck Telephone 510-848-4169 . Complete Mailing Address of Headquarters or General Business Office of Publisher (Not p California Botanical Society. Inc.; Herbaria, Life Sciences Building University of California. Berkeley, CA 94720 9. Full names and Complete Mailing Addresses of Publisher, Editor and Managing EdHor (Do not leave blank) Publisher {Name and complete mailing address) California Botanical Society, Inc ; Herbaria, Life Sciences Building Untversfty of California, Berkeley, CA 94720 Editor [Name and complete mailing address) Knstina A. Schierenbeck; California State University. Ch CO, Dept of Biology; Chico, CA 95429-051 5 Managing Editor (Name and complete mailing address} Full Name Complete Mailing Address California Botanical Society, Inc Herbana, Life Sciences Building University of California Berkeley. CA 94720 1 1 Known Bondholders, Morlgages. and Other Securrty Holders Owning at Holding 1 Percent or Moie of Tolal Amount of Bonds, Morlgages. or Complete Mailing Address 0 Has Not Changed Dunng Preceding l2 Months □ Has Changed Dunng Preceding 12 Months (Pub 3 Publicabon Title Extent and Nature of Circulation b Paid and/or Requested Circulation Paid/Requested Outside- County Mail Subscriptions stated on Form 3541 (Include adverHser's proof and exchange copies) Paid In-County Subscriptions (I Total Paid and/or Requested Circulation (Sum of 15b(1} . (2).(3). and (4)) | Outstde-County as Stated on Form 3541 In-County as Stated on Form 3541 B Distribution Outside the Mail (Cerrlorz or ottier rr Total Free Distribution (Sum of 15d and 15e) g. Total Distribution (Sum of 15c and 15f) 16 Publication of Statem [21 Publication required Will be printed in the 47/4 issijfe of this/^ublica IT^^i^^^^dTitte of Edft^Publi^^u^^ i Date for Circulation Data Below Average No. Copies Each Issue Nearest to Filing Date Date i3 I Instructions to Publishers 1 Complete and Tile one copy of this fomi with y your records 2 In Cases whe 3 Be sure to furnish all circulation information called tor Ir 4 Item 15h , Copies not Distnbuted, must include (1) new estimated returns from news agents, and (3), copies fo I. include In items 1 0 and n the name of the per; esses of individuals who are stockholders who own or hokJ 1 percent or [ties of the publishing corporation In item 11 . if none, check the box Use 15 FreedrcutationmustbeshowninitemslSd. e, andf d copies originally stated on Fomn 3541 . and returned to the publisher. (2) 3 use, leftovers, spoiled, and all other copies not distributed. Circulation must be published; it must be pnnled in any i< issue pnnted after October 6 In Item 1 6. indicate the date of the issue in whteh this Sta PS Form 3526, September 1998(Reverse) 1 October or, if the publication is not published during C it of Ownership will t>e published Computerized Facsimile VOLUME 47, NUMBER 1 JANUARY-MARCH 2000 MADRONO A WEST AMERICAN JOURNAL OF BOTANY Seed Rain During Early Primary Succession on Mount St. Helens, Washington David M. Wood and Roger del Moral 1 Environmental Gradients and Vegetation Structure on South Texas Coast- al Clay Dunes Kern Ewing 10 Chromosome Counts and Taxonomic Notes on Draba (Brassicaceae) of the Intermountain West. 1 : Utah and Arizona Michael D. Windham 21 A Floristic Study of Tsegi Canyon, Arizona Susan Holiday 29 Mixed Conifer and Red Fir Forest Structure and Uses in 1 899 from the Central and Northern Sierra Nevada, California Scott L. Stephens 43 Old-growth Forest Associations in the Northern Range of Coastal Redwood Thomas M. Mahony and John D. Stuart 53 Estimated Ages of Some Large Giant Sequoias: General Sherman Keeps Getting Younger Nathan L. Stephenson 61 A Natural History of the Sonoran Desert. Edited by S. J. Phillips and P. W. COMUS Matthew L. Brooks 68 ANNOUNCEMENT 70 \ CONTENTS WOUNCEMENT i PUBLISHED QUARTERLY BY THE CALIFORNIA BOTANICAL SOCIETY ■ "-* * • Madrono (ISSN002^^T75 is published quarterly by the California Botanical Society, Inc., and is issued from the office of the Society, Herbaria, Life Sciences Building, University of California, Berkeley, CA 94720. Subscription information on inside back cover. Estabhshed 1916. Periodicals postage paid at Berkeley, CA, and additional mailing offices. Return requested. Postmaster: Send address changes to Madrono, Roy Buck, % University Herbarium, University of California, Berkeley, CA 94720. Editor — Kristina A. Schierenbeck California State University, Chico Department of Biology Chico, C A 95429-0515 kschierenbeck @ c suchico.edu Editorial Assistant — David T. Parks Book Editor — Jon E. Keeley Noteworthy Collections Editors — Dieter Wilken, Margriet Wetherwax Board of Editors Class of: 2000 — Pamela S. Soltis, Washington State University, Pullman, WA John Callaway, University of San Francisco, San Francisco, CA 2001 — Robert Patterson, San Francisco State University, San Francisco, CA Paula M. Schiffman, California State University, Northridge, CA 2002 — Norman Ellstrand, University of California, Riverside, CA Carla M. D' Antonio, University of California, Berkeley, CA 2003 — Frederick Zechman, California State University, Fresno, CA JoN E. Keeley, U.S. Geological Service, Biological Resources Division, Three Rivers, CA 2004 — David M. Wood, California State University, Chico, CA Ingrid Parker, University of California, Santa Cruz, CA CALIFORNIA BOTANICAL SOCIETY, INC. Officers for 2000-2001 President: Bruce Baldwin, Jepson Herbarium and Dept. of Integrative Biology, 1001 Valley Life Sciences Bldg. #2465, University of Cahfomia, Berkeley, CA 94720. First Vice President: Rod Myatt, San Jose State University, Dept. of Biol. Sciences, One Washington Square, San Jose, CA 95192. rmyatt@email.sjsu.edu Second Vice President: Rob Schlising, California State University, Chico, Dept. of Biol. Sciences, Chico, CA 95424. rschhsing@csuchico.edu Recording Secretaij: Dean Kelch, Jepson and University Herbarium, University of Cahfomia, Berkeley, CA 94720. dkelch@sscl.berkeley.edu Corresponding Secretary: Susan Bainbridge, Jepson Herbarium, University of California, Berkeley, CA 94720. suebain @ SSCL.berkeley.edu Treasurer: Roy Buck, % University Herbarium, University of California, Berkeley, CA 94720. The Council of the California Botanical Society comprises the officers listed above plus the immediate Past President, R. John Little, Sycamore Environmental Consultants, 6355 Riverside Blvd., Suite C, Sacramento, CA 95831; the Editor of Madrono; three elected Council Members: Bl\n Tan, Strybing Arboretum, Golden Gate Park, San Francisco, CA 94122; James Shevock, USDI National Park Service, Pacific West Region, 600 Harrison Street, Suite 600, San Francisco, CA 94107; Diane Elam, U.S. Fish and Wildlife Service, 3310 El Camino Avenue, Sacramento, CA 95825; Graduate Student Representative: Kirsten Johanus, Jepson Herbarium, University of California, Berkeley, CA 94720. @ This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Madrono, Vol. 47, No. 1, pp. 1-9, 2000 SEED RAIN DURING EARLY PRIMARY SUCCE HELENS, WASHINGTON David M. Wood Department of Biological Sciences 515, California State University, Chico, CA 95929 Roger del Moral Department of Botany Box 355325, University of Washington, Seattle, WA 98195 Abstract Seed rain into sites undergoing primary succession on Mount St. Helens was measured from 1982 to 1986 and again from 1989 to 1990. Study sites were devastated in 1980 by pyroclastic flows of pumice, searing blasts, and lahars. Most sites were several km or more from seed sources. Seed rain density averaged 34 seeds 0.1 m"^ yr"' in mid-elevation barren sites, 1083 seeds 0.1 m ^ yr ' in mid-elevation vegetated sites and 2 seeds 0.1 m^ yr ' at subalpine barren sites. A total of 33 species was collected in traps. The relative abundance distributions of species were generally similar across years and sites. A few wind-dispersed species accounted for most of the seed rain: Anaphalis margaritacea (L.) Benth. & Hook., Epilobium angustifolium L., E. watsonii Barbey {E. ciliatum Raf.), Hieracium albiflorum Hook., and Hypochaeris radicata L. Seeds of trees and shrubs were virtually absent. The common species in the seed rain were also the most common species in the vegetation, although their absolute abundance is determined by environmental factors. Many uncommon species occurred in the vegetation that were not recorded in the seed rain. Two taxa common in the vegetation, Lupinus lepidus Douglas and Salix spp., were rare in the seed rain. For Salix, this is because seed dispersal occurred before traps were in place for the season. Lupinus lepidus is not wind dispersed and seeds are not likely to enter traps. We conclude that the seed rain on Mount St. Helens is apparently sufficient to initiate colonization but is depauperate in species. At present the vegetation generally reflects the incoming seed rain. One cause of succession is differential species availability at a site after a disturbance (Pickett et al. 1987). For vascular plants this differential avail- ability occurs mostly by vegetative regrowth, seed banks, or seed dispersal. In primary succession col- onization results mainly from seed dispersal. On large scale primary successional landscapes the in- put or "rain" of seeds from long-distance dispersal is the main source for the establishment of most species, because seed banks and regrowth are ab- sent. A complete interpretation of primary succes- sion in particular, and community assembly in gen- eral, must therefore include measurements of the density and species composition of the seed rain along with an assessment of environmental and substrate conditions (Wood and del Moral 1987; del Moral 1993; Chapin et al. 1994; Booth and Larson 1998; Dlugosch and del Moral 1999). For example, the absence of a species from a particular site or serai stage could be due as much to its absence from the seed rain as to its inability to establish. Conversely, high abundance of a colonist may be explained as much by its abundant seed rain as by its environmental tolerance, growth rate, or com- petitive ability. Hypothesized mechanisms of suc- cession such as facilitation (Connell and Slatyer 1977; Morris and Wood 1989) must also consider differential species availability through the seed rain. Studies of seed dispersal generally are of two types: studies of individual species with the parent plant and its seed shadow as the focus (reviews in Harper 1977; Willson 1993), or studies of the long- distance seed rain into sites where specific seed sources cannot be identified precisely. Seed rain measurements are most appropriate for studies of community assembly in primary succession, but published studies are few (Ryvarden 1971; Stocklin and Baumler 1996; Archibold 1980; Jefferson and Usher 1989; Chapin et al. 1994). This paper de- scribes the density and species composition of the seed rain in several contrasting regions and habitats undergoing primary succession on Mount St. Hel- ens, WA. The rate of vegetation recovery and plant species composition at various sites have been described for Mount St. Helens following the catastrophic eruption in 1980 (del Moral 1983; Wood and del Moral 1987; del Moral and Wood 1988; del Moral 1993; del Moral and BHss 1993; del Moral and Wood 1993a, b; del Moral et al. 1995; del Moral 1998, del Moral 1999), but detailed species seed rain data have not been reported with the exception of Dale (1989) who sampled lower elevation lahar (mudfiow) sites not included in this study. We pose these questions: For a given site, what is the density and species composition of the seed rain? How do islands of established, reproducing vegetation affect 2 . \ MADRONO [Vol. 47 Fig. 1. Location of study areas. PP = Pumice Plain sites, PA = Plains of Abraham sites, BC = Butte Camp sites. See text for details of seared, blowdown, and blast zones. the local seed rain? Does the species composition of the seed rain reflect the species composition of the colonizing flora? Are there species present in the seed rain but absent as colonists? Are there spe- cies present as colonists but absent from the seed rain? Study Area The Mount St. Helens volcano is in the Cascade Range of southwestern Washington at 46°12'N, 122°irw. The catastrophic north-directed eruption of May 18, 1980, produced a variety of impacts including: a debris avalanche; pyroclastic flows, or incandescent flows of gas and pumice; and lahars, flows of water-saturated debris ("mudflows") trig- gered by rapidly melting snow and ice (Lipman and Mulhneaux 1981; Decker and Decker 1981). Im- pacts on the vegetation are categorized into four regions: the blast zone, in which most life was de- stroyed; the blowdown zone, in which adult trees were knocked over but some saplings and under- story vegetation survived; the seared zone, in which trees remained standing but had their foliage singed by the hot gases of the blast, and the mud- flow (lahar) zone in which most vegetation was de- stroyed (Fig. 1). In addition to these four regions, a large region south of the crater received 5-20 cm of tephra (airfall ash and pumice) and most vege- tation survived. We studied three main areas separated by several km and at different elevations: Pumice Plain, Plains of Abraham (both in the blast zone), and Butte Camp (Fig. 1). The Pumice Plain is a 20 km^ region between the crater and Spirit Lake that received the full force of the north-directed lateral eruption, re- ceiving a debris avalanche and pyroclastic flows. These new deposits now overlie what was formerly a montane forest of Tsuga heterophylla (Raf.) Sarg., Pseudotsuga menziesii (Mirbel) Franco, and Abies amabilis (Douglas) James Forbes (Krucke- berg 1987). Elevations range from 1000 to 1200 m. Although much of the Pumice Plain was initially flat to moderately hilly, numerous erosion gullies continue to form and deepen. The Plains of Abra- ham, also a pumice landscape, is located approxi- mately 3 km NE of the crater at 1350 m elevation. It received searing blasts and deposits of tephra, and is predominantly flat with numerous small gul- lies (del Moral and Wood 1993b). The pre-emption vegetation of this high montane region was de- scribed by Kruckeberg (1987) as a region high in species richness of forbs and grasses but low in cover with scattered conifers. The Butte Camp re- gion is located on the southwest side of the volcano I 2000] WOOD AND MORAL: SEED RAIN ON MOUNT ST. HELENS 3 Table L Mean ± Standard Error of Seed Rain Density (seeds 0.1 m ^ yr"') for 1983 to 1986 Estimated from I Fallout Traps at Mid-Elevation Barren Sites. Number in parentheses is number of traps. In 1983 and 1984, mean values for the Pumice Pond and Spirit Lake sites are significantly different. Pumice Plain Sites 1983 1984 1985 1986 Pumice Pond Spirit Lake 75.5 34.9 ± 9.5 (10) ± 5.0 (10) 38.4 ± 9.6 (5) 13.0 ± .9 (9) no data 25.02 ± 3.0 (10) no data 26.9 ± 4.9 (10) at 1500 to 1600 m. This area of subalpine vegeta- tion was disturbed by several lahars and also re- ceived tephra deposits (del Moral 1983; del Moral and Wood 1986). However, seed rain was only measured at the primary succession lahar sites. Methods Seed collections. Seeds of vascular plants were collected from both wet pitfall traps and dry fallout traps. Pitfall traps, whose primary purpose was to collect ground-dwelling insects (Edwards 1986), consisted of 10 cm diameter plastic cups filled with ethylene glycol, set flush with the ground surface, and covered by a plywood square elevated 1 cm above the cup with comer nails. Pitfall trap data were collected in 1982, 1983 and 1985. The fallout traps consisted of 33 X 33 cm (0.1 m^) wooden frames 3 cm high with fine nylon mesh bottoms. Frames were filled with a single layer of used golf balls and set flush with the surface (Edwards 1986; Edwards and Sugg 1993). The golf balls (approxi- mately 3 cm in diameter) were used because they were an easily obtainable uniform sphere that mim- icked the size and surface texture of the pumice. We also wanted baseline density estimates of the seed rain for relatively flat, open ground. These baseline values may then be adjusted upwards if desired for sites of seed accumulation, e.g., against boulders or in gullies and other depressions (Dale 1989). Sticky traps (Werner 1975) or wet pitfall traps could yield over-estimates of density for flat ground in this open, windy environment due to seed accumulation (Johnson and West 1988). Thus, only fallout traps were used to estimate seed rain den- sity. Fallout traps were used in all years of the study except 1982, and were the only type of trap used in 1989 and 1990. Data from both pitfall traps and fallout traps were used to estimate relative abun- dance. Traps were set out in June of each year after snowmelt when the roads to the sites became ac- cessible and contents were collected approximately twice a month until November. Fallout traps were collected only once a year in late October, after the fall dispersal period and before sites became inac- cessible due to snow. Seeds were stored in alcohol or formalin-acetic acid-alcohol (FAA) and were identified using a reference collection obtained from field specimens and herbarium sheets. Al- though seed germination was not measured in this study, only those seeds with morphology and col- oration similar to viable seeds were counted. Ex- tensive seed germination experience with many species from Mount St. Helens suggests that the appearance of viability under a dissecting micro- scope is a good predictor of germination — most species had germination rates from 60 to 90% when abnormal-appearing seeds were excluded (Wood and del Moral 1987; Wood unpublished data). No- menclature follows Hitchcock and Cronquist (1973) with parenthetical updates from Hickman (1993) to correspond to Titus et al. (1998). Study sites. From 1982 to 1986 two sites on the Pumice Plain were sampled. Pumice Pond and Spir- it Lake. The Pumice Pond site was near the head- waters of the North Fork of the Toutle River on the northwest side of the Pumice Plain about 5 km NNW of the crater. Unfortunately, severe erosion at this site forced it to be abandoned in 1985 (Ed- wards and Sugg 1993). The Spirit Lake site was near the eastern edge of the Pumice Plain about 2 km south of Spirit Lake and 3 km north of the crater. Traps of both types (pitfall and fallout) were placed at 10 m intervals along 100 m transects, al- though resultant sample sizes vary because traps occasionally were filled with erosional material or lost (Tables 1 and 3). In 1986 mean percent cover of vegetation on the Pumice Plain in the vicinity of these sites was estimated at 0.09% (Wood and del Moral 1988) and had increased to only 1.4% by 1990 although some small patches exceeded 50% (Wood unpublished data). In 1989 and 1990, the number and placement of fallout traps was increased by including a greater variety of habitats within the Pumice Plain. Pumice Plain sites I and II were established in barren areas (defined as having <3% cover) close to the old Spirit Lake site at 1100 m elevation. Each site con- tained 16 fallout traps arranged in a 4 X 4 grid with traps separated by 10 m (hereafter referred to as a "16-FT grid"). Due to occasional trap disturbance (e.g., by ravens and elk) resultant sample size again varied (Table 2). The Pumice Ridge site was on an exposed, barren ridge 50 m above the Pumice Plain and contained 5 traps along a 30 m transect. The 16-FT grid Lupine Patch site was in a patch of dense flowering Lupinus lepidus Douglas, (>50% cover) a few hundred m from Pumice Plain I and II. The Willow Spring 16-FT grid was in a rela- tively open, moderately vegetated site (<20% cov- er) but was surrounded by a stand of dense, repro- ductively mature vegetation adjacent to a spring (Wood and del Moral 1988). This vegetation in- cluded Salix spp. (primarily S. sitchensis Bong, and 4 MADRONO [Vol. 47 Table 2. Mean ± Standard Error of Seed Rain Density (seeds 0.1 m ^ yr~') in 1989 and 1990. Number in parentheses is number of traps. Means with the same letter within a year are not significantly different at P = 0.05 by Tukey's HSD. MEB = mid-elevation barren; MEV = mid-elevation vegetated; HEB = high-elevation barren. Habitat 1989 1990 Butte Camp Sites Lahar I Lahar II Plains of Abraham Sites Abraham I Abraham II Pumice Plain Sites Pumice Plain I Pumice Plain II Pumice Ridge Lupine Patch Willow Spring HEB HEB MEB MEB MEB MEB MEB MEV MEV 1.2^ 2.8'' 0.7 (15) 0.8 (16) 11.8'' ± 1.7 (16) 9.0'' ± 2.0 (5) 24 + 4 1 (15) 21.9'''^ ± 3.3 (15) 68.2-^^ ± 19.4 (5) 94.1' ± 10.8 (16) 1709.1 ± 489.7 (16) no data no data 5.3" ± 2.8 (16) 66.5"'' ± 12.3 (4) 50.6" ± 8.2 (16) 50.6" ± 8.2 (16) 9.6" ± 2.3 (5) 355.6'' ± 48.6 (16) 2174.4 ± 645.1 (16) S. commutata Bebb), Anaphalis margaritacea (L.) Benth. & Hook., Epilobium angustifolium L., E. watsonii Barbey {E. ciliatum Raf.), Hypochaeris radicata L., and Lupinus lepidus. The Plains of Abraham area contained two sites in barren areas, one 16-FT grid (Abraham I) and one five-trap transect (Abraham II) similar to the Pumice Ridge site described above. Both sites were on nearly level ground and were spaced 200 m apart. In 1989 and 1990, mean percent cover on the Plains of Abraham was estimated at 0.12% and 0.23%, respectively (del Moral and Wood 1993b). The Butte Camp area contained two sites on la- hars (Lahar I and II), both 16-FT grids spaced 200 m apart. Percent cover on the Butte Camp lahars was estimated at 2-3% in 1989 (del Moral 1993). Results Density. Seed rain density varied widely over both sites and years (Tables 1 and 2), from a low of 1.2 seeds 0.1 m"^ yr"^ at Lahar II in Butte Camp in 1989 to a high of 2174 seeds OA m ^ yr i at Willow Spring on the Pumice Plain in 1990 (Table 2). In 1983 and 1984 the Pumice Pond site received more than twice as many seeds as did the Spirit Lake site (Table 1; t-test, log transformation, P < 0.01 in each year). In both 1989 and 1990, ANOVA revealed a significant difference among the Pumice Plain, Plains of Abraham, and Butte Camp areas as well as significant differences among sites within the Pumice Plain (Table 2; log transformation, P < 0.001 in each year). In both 1989 and 1990, Willow Spring had a significantly greater seed rain density than all other sites (Table 2; Tukey's HSD multiple comparisons, P = 0.05). Lupine Patch had the sec- ond highest seed rain density in both years, al- though mean density at this site was not signifi- cantly different from Pumice Ridge in 1989 or Abraham II in 1990 (Table 2). Statistical tests were not performed on year-to-year differences within a site due to the lack of clear hypotheses, as variation could be due to unmeasured factors such as differ- ences in wind patterns or growing conditions and seed production in surrounding landscapes. When sites were classified by habitat, the varia- tion in density was reduced and a clearer pattern emerged. Mid-elevation barren sites (Pumice Pond, Spirit Lake, Pumice Plain I and II, Pumice Ridge, Table 3. Relative Abundance of Common Species in the Seed Rain for 1982 through 1986. The Pumice Pond and Spirit Lake sites are combined. See text for additional species. Distributions between years are not significantly different by a Wilcoxon Signed Ranks test. Relative Abundance (%) 1982 1983 1984 1985 1986 Anaphalis margaritacea 16 36 21 10 70 Epilobium angustifolium 26 8 48 74 13 Epilobium watsonii (E. ciliatum) <1 5 10 3 5 Hypochaeris radicata 3 2 3 5 3 Hieracium albiflorum 3 2 3 1 3 Senecio sylvaticus 36 39 7 1 1 Lupinus lepidus 0 0 0 0 0 Number of pitfall traps 38 32 0 35 0 Number of fallout traps 0 20 14 10 10 2000] WOOD AND MORAL: SEED RAIN ON MOUNT ST. HELENS 5 Abraham I and II) had an overall mean density of 33.6 seeds 0.1 m^^ yr ', ranging from 5.3 at Abra- ham I in 1990 (Table 2) to 75.5 at Pumice Pond in 1983 (Table 1). High-elevation barren sites (Lahar I and II) had a much lower overall mean density of 1.9 seeds 0.1 m"^ yr The highest densities were recorded at mid-elevation vegetated sites (Lupine Patch and Willow Spring) where densities ranged from 94.1 at Lupine Patch in 1989 to 2174 at Wil- low Spring in 1990 (Table 2) with an overall mean density of 1083. Relative abundance. The relative abundance of the most common species in the seed rain is pre- sented in Tables 3 and 4. Relative abundance dis- tributions were generally consistent from year to year and from site to site. The most distinctive sites were Lahar I and II, where subalpine species char- acteristic of that habitat appear. However, no com- parison of abundance distributions is significantly different, either among years from 1982 to 1986 (Table 3), between years within a site, or among sites in 1989 and 1990 (Table 4; Wilcoxon Signed Rank Test, all P > 0.5). Unfortunately, separate es- timates of relative abundance for the Pumice Pond and Spirit Lake sites are not available because sam- ple collections from these sites were combined after counting the total number of seeds in a given trap. Six species accounted for 85% of the measured seed rain at the two Pumice Plain sites in 1982 and >90% from 1983 to 1986: Anaphalis margaritacea, Epilobium angustifolium, E. watsonii (E. ciliatum), Hypochaeris radicata, Hieracium albiflorum Hook., and Senecio sylvaticus L. (Table 3). Six species also accounted for >90% of the measured seed rain at all sites in 1989 and 1990 except for the subalpine sites Lahar I and II (Table 4). These were the same six species listed above except that Lupinus lepidus replaced S. sylvaticus. The decline of S. sylvaticus and the increase of L. lepidus were the most note- worthy changes in relative abundance during this study. S. sylvaticus decreased from 39% relative abundance in 1983 to zero in 1989 and 1990 at all sites except Lahar I and II (Tables 3 and 4). Lupinus lepidus was not recorded from 1982 to 1986 but dur- ing the 1989-1990 sampling period it occurred at all sites except Abraham II at least once. A total of 33 species was collected, including two unidentified grasses (one seed each). Species not listed in Tables 3 or 4, all with three or fewer seeds trapped except as noted, are: Acer circinatum Pursh, Achillea millefolium L., Agoseris grandiflora (Nutt.) E. Greene, Agrostis sp., Antennaria sp., Carex mertensii Prescott, Carex rossii Boott, Carex sp., Cinna latifolid (Goeppert) griseb., Cirsium vul- gare (Savio) Ten., Epilobium luteum Pursh, Juncus parryi Engelm., Lactuca muralis (L.), Fresen. Pen- stemon cardwellii Howell, Salix spp. (12 seeds), Saxifraga ferruginea Graham, Senecio vulgaris L., Sitanion hystrix (Nutt.) J. G. Smith (Elymus ely- o ^ ON ^ Q M z S ON OC 73 On O ^ 8 si w u ^ O OX) u o 2 ^ < Oooomiorj-Hor^t^r^or- — ' (N (N r-moooooooooooo -HOfNOOrJOOOOOOOO IT) On in m ON o o o o o o o oooooooo ^^On-^-^-hOOOOOOO 00 V V ^-Hin-H^oooooooo (NNOOON'xI-OOO— '—hOOOO OO^^OOO— 'ONOOOOOOO r-r^ioooooooooooo ooorir^inmooooooo (NO>nr^(NOOOOOOOO ON^ON'^r^-^OOOOOOO s 5 ~~ s s ^ a ..s; .~ ^ ^ Cl, o Qj o =§1 ^ S ^ -2 ?3 ^. ^ S 6 MADRONO [Vol. 47 moides (Raf.) Swezey), Sonchus asper (L.) Hill, and Taraxacum officinale Wigg. As with density, a classification of sites by hab- itat resulted in a clearer pattern of relative abun- dance. In both the mid-elevation barren sites and Lupine Patch (which had vegetation mostly <15 cm in height), Anaphalis margaritacea and Epilo- bium angustifolium dominated the seed rain. At Willow Spring, which had taller surrounding veg- etation (up to 2 m) including a vigorous flowering population of E. watsonii (E. ciliatum), seeds of E. watsonii were dominant. Still, densities of A. mar- garitacea and E. angustifolium at Willow Spring were similar to the other mid-elevation sites. The species recorded in the seed rain at Lahar I and II were distinct from all other sites (Table 4), as expected given the higher elevation, different surrounding flora, and greater physical exposure of these subalpine sites. Although sampled densities were very low (Table 2), making interpretation speculative, A. margaritacea and E. watsonii were conspicuously absent from the seed rain although E. angustifolium was present. Species characteristic of the surrounding subalpine flora that were record- ed at Lahar I and II, but were rarely trapped else- where, included Spraguea umbellata Torr. {Calyp- tridium umbellatum) (Torrey) E. Greene; Polygo- num newberryi Small, Aster ledophyllus A. Gray, Lomatium martindalei (J. Coulter & Rose) J. Coul- ter & Rose, Juncus parryi, and Hieracium gracile Hook. (Table 4). Eriogonum pyrolifolium Hook., a dominant species in many subalpine sites on Mount St. Helens (del Moral and Wood 1986, Chapin and Bliss 1989, del Moral 1993), was not recorded in the seed rain. Discussion Seed rain is a critical factor in determining spe- cies composition and abundance in early primary succession on Mount St. Helens. All species in the seed rain with >1% relative abundance at any site are present in the vegetation, and the most conmion species in the seed rain were also the most conmion species in the vegetation during the study period (Wood and del Moral 1988; del Moral 1993; del Moral and Wood 1993a; see also Stocklin and Baumler 1996). No species with consistent, rela- tively abundant seed rain appeared to be excluded from establishing at least some individuals on Mount St. Helens due to a lack of ecological tol- erance. However, the absolute abundance in the vegetation on Mount St. Helens is determined by a host of other factors in addition to seed rain density including safe-sites for germination (Wood and Morris 1990; del Moral and Wood 1993b; Titus and del Moral 1998) and facilitation (Morris and Wood 1989; del Moral and Wood 1993a). Most of the common species in the seed rain have seeds adapt- ed for wind dispersal: a feathery coma in Epilobium angustifolium and E. watsonii, and pappuses in An- aphalis margaritacea, Hypochaeris radicata, Hier- acium albiflorum, and Senecio sylvaticus. The consistency in species composition of the seed rain among both years and sites suggests that the vegetation will also be similar from site to site, with the exception of the subalpine lahar sites. This prediction is upheld for sites of similar elevation except where patches of Lupinus lepidus have de- veloped (del Moral et al. 1995). The species com- position of the seed rain also gives some indication as to seed sources. Seeds of conmion montane spe- cies such as Anaphalis margaritacea, Epilobium angustifolium, E. watsonii (E. ciliatum), Hypo- chaeris radicata, Hieracium albiflorum, and Sene- cio sylvaticus probably originated in seared and blowdown forest 10-20 km to the west and north of the study areas (Fig. 1) where recovery of these species occurred relatively rapidly (Halpem et al. 1990). Westerly prevailing winds likely transported these species up the Toutle River valleys to the Pumice Plains sites (Fig. 1). Willson (1993) reports a wide range in maximum dispersal distances of herbaceous species with wind dispersal adaptations, from a few m to >4000 m. The dispersal ability of Epilobium in particular is extraordinary — Solbreck and Andersson (1987) estimated the maximum dis- persal distance of E. angustifolium to be hundreds of km under windy conditions. Seeds of the ruderal species in the seed rain such as Cirsium arvense (L.) Scop., C. vulgare. Taraxacum officinale, Son- chus asper, Lactuca muralis, and Senecio vulgaris probably had their origin in low -elevation clearcuts or agricultural fields tens of km to the west. Dale (1989) captured several of these same ruderal spe- cies at lower elevation on the debris avalanche along the Toutle River to the west. The only shrub or tree species trapped besides Salix was one seed of Acer circinatum, in a sample of >75,000 seeds. Since the montane sites on Mount St. Helens will, in the absence of another eruption, eventually succeed to a coniferous forest, the low abundance of late-successional woody spe- cies suggests a strong seed dispersal limitation. Similarly, Chapin et al. (1994) detected no spruce seeds and negligible alder seeds in the pioneer stage of primary succession at Glacier Bay at dispersal distances comparable to those of this study (ap- proximately 10 km from seed sources for spruce and 3 km from alder sources). Although not de- tectable in the seed rain, conifer and shrub seed- lings such as Pseudotsuga menziesii, Abies amabi- lis, Tsuga heterophylla, Pinus contorta Loudon, Al- nus sinuata (Regel) Rydb. (A. viridis (chain) DC), Rubus spp., and Vaccinium spp. do occur in low numbers at most of the sites sampled here (see also del Moral et al. 1995). These individuals are either establishing from extremely low seed source inputs and/or our seed trap design did not adequately sam- ple their mode of dispersal (see below). Seed traps were designed to estimate the seed rain onto relatively flat, open ground. True densities 2000] WOOD AND MORAL: SEED RAIN ON MOUNT ST. HELENS 7 may exceed our estimates in microsites where seeds accumulate, such as in depressions or wet sites, or about rocks (Dale 1989; Titus and del Moral 1998). Higher densities also may occur in vegetated sites where short-distance dispersal supplements the long-distance seed rain. For example, at Willow Spring and Lupine Patch, seeds produced on or near the site probably equaled or even exceeded the number of seeds arriving by long-distance dispers- al. Also, variation among traps was highest at Wil- low Spring, with standard errors of 29% and 30% of the density means for 1989 and 1990, respec- tively (Table 2). This suggests that established veg- etation islands augment the long-distance seed rain in a patchy manner, in contrast to barren sites which receive a more predictable, albeit low input. The sharp decline in the seed rain of Senecio sylvaticus (Tables 3 and 4) may be explained by its life history — a biennial, it exploits forest clearcuts for only one or two generations before being out- competed by more aggressive serai species (West and Chilcote 1968; Halpem et al. 1997). The 1980 eruption of Mount St. Helens apparently created brief but favorable growing conditions for S. syl- vaticus in surrounding forests that resulted in a pulse of seed rain in 1982 and 1983. The overall mean density of 33.6 seeds 0.1 m^ yr ' for mid-elevation barren sites on Mount St. Helens is similar to that found by Ryvarden (1971; calculations from Rabinowitz and Rapp 1980), who reported 34.2 to 65.3 seeds 0.1 m~^ yr ' for primary succession at the base of a retreating glacier in Nor- way, and to that of Stocklin and Baumler (1996) who found 12.5 seeds 0.1 m~^ yr ' for newly ex- posed terrain in glacial forelands in Switzerland. Archibold (1980) reported 240.0 to 380.0 seeds 0.1 m"2 yr ' in stripmine wastes in Saskatchewan, but this higher figure may be due to the closer prox- imity of seed sources. The very low mean density of 2 seeds 0.1 m"^ yr ' for the subalpine lahar sites was probably because seeds of well-dispersed spe- cies such as Anaphalis margaritacea and E. wat- sonii did not reach that elevation, and because seeds of species in the surrounding vegetation have poor adaptations for dispersal (Wood and del Moral 1987). Many species that occur in the vegetation on Mount St. Helens were not recorded in the seed rain. Most of these species are unconmion or rare. This suggests that either their seed rain is below our detection limits or that their mode or timing of dispersal is such that they eluded capture. Although we think that low species richness in the seed rain is more likely, our traps were designed to capture wind-dispersed seed and thus may have missed capturing seeds of species with other dispersal modes. One possible dispersal mode that may be important on Mount St. Helens is that of secondary wind dispersal across hard snow surfaces. Matlack (1989) showed that seeds of Betula lenta were dis- persed greater distances by secondary dispersal than by primary dispersal to the ground. Because Mount St. Helens receives abundant winter snow and freeze-thaw cycles are conmion, hard surfaces conducive to secondary dispersal by wind probably occur. Water dispersal (hydrochory) is another un- measured variable. In addition to permanent streams, numerous small temporary streams com- monly develop during spring snowmelt and fall rains, and sheet flow occurs during particularly heavy rains. Seeds can be transported along these watercourses (Stocklin and Baumler 1996). Either secondary dispersal or water dispersal may be re- sponsible for the spread of non-wind dispersed spe- cies such as Lupinus lepidus and the occurrence of the late successional woody species listed above. Animal dispersal (zoochory) is another unmeasured vector. We consider animals to be less important than either wind or water, but we cannot rule out their effect. Plant taxa with fleshy fruits are rare on Mount St. Helens (e.g., Vaccinium, Rubus; del Moral 1993) suggesting that frugivory as a means of seed dispersal is also rare. However, birds and large mammals such as elk and coyotes travel long distances to the study sites and may disperse seeds by defecation or transportation in their feathers or hair. The potential importance of a rare colonization event that results in local seed production and pop- ulation spread should not be underestimated. Whereas relative abundance of a species in the seed rain is a good indicator of its relative abun- dance in the vegetation, the reverse is not neces- sarily true. A few species are common in the veg- etation but unconmion in the seed rain. These in- clude Lupinus lepidus, Salix spp., and Eriogonum pyrolifolium. Lupine is the species with the greatest disparity between its estimated seed rain density and its abundance. Lupine survived the eruption in a variety of high elevation sites around the volcano (del Moral 1983, 1993; del Moral and Wood 1986) and was present on the Pumice Plain as early as 1981 (C. Crisafulli personal communication), pos- sibly establishing from seeds or root fragments washed down from high-elevation survivors. Now lupine occurs across the Pumice Plain and other sites on Mount St. Helens (Morris and Wood 1989; Bishop and Schemske 1998; Titus et al. 1998). In spite of this early record of population growth, seeds of L. lepidus were not captured in the seed rain until 1989, presumably because of its limited seed shadow. Lupine seeds have no obvious dis- persal adaptations except for ballistic dispersal when legumes dehisce, but this type of dispersal probably only achieves a few m (Willson 1993). Thus the rapid increase of L. lepidus on the Pumice Plain was due to vigorous seedling recruitment in close proximity to early colonists, not to long dis- tance dispersal (Wood and del Moral 1988; Morris and Wood 1989). The low abundance of Salix spp. in the samples is probably due to a flaw in the sam- pling design. Salix began reproducing as early as 1985 at Willow Spring (Wood personal observa- 8 MADRONO [Vol. 47 tion) but each year due to impassible roads our traps were put out too late to catch dispersing wil- low seeds. The high abundance of Salix around Willow Spring would undoubtedly have contributed greatly to the seed rain at this site and would have resulted in lower relative abundances of other spe- cies such as E. watsonii. Eriogonum pyrolifolium, a dominant subalpine species, has relatively heavy, round seeds with no obvious dispersal adaptations (Wood and del Moral 1987) and thus its seed shad- ow apparently did not extend to the seed traps. The vast majority of incoming seeds in the seed rain fail to establish. Vegetation remained generally sparse by 1990 in spite of a rain of hundreds of seeds mr^ yr ' onto most sites. Previous studies demonstrated that limits to abundance on Mount St. Helens are set by environmental factors. Morris and Wood (1989) and del Moral and Wood (1993a) showed that Lupinus lepidus may facilitate the es- tablishment of several species including Anaphalis margaritacea, Epilobium angustifolium, and Hy- pochaeris radicata. Wood and Morris (1990) showed that manipulation of substrate moisture and microtopographic heterogeneity positively affected the rate of establishment of A. margaritacea and E. angustifolium. Del Moral and Wood (1993b) showed that most species on the Plains of Abraham established in favorable microsites more often than expected by chance. There is also a tradeoff be- tween seed mass and probability of establishment — heavier seeds have a greater likelihood of estab- lishing on Mount St. Helens due to increased seed- ling vigor but are less likely to disperse a long dis- tance (Wood and del Moral 1987; Wood and Morris 1990). Titus and del Moral (1998) further demon- strated the importance of microsites in seedling es- tablishment. Thus, the vegetation of early primary succession on Mount St. Helens is composed pri- marily of well-dispersed species in low abundance. Stochastic events such as chance colonization of species with low long-distance seed rain result in heterogeneous communities with little structure (del Moral et al. 1995). As succession proceeds, com- munity composition will become increasingly un- correlated with the long-distance seed rain. Acknowledgments We thank L. Zemke and A. Ziegler for assistance in the herbarium in Seattle; R. Sugg and J. Edwards for design- ing the fallout traps and collecting early samples; V. Do- raiswamy and J. Mcirr for excellent work in seed identi- fication and counting; J. Hubbell for invaluable field as- sistance; and P. Frenzen and C. Crisafulli of the Mount St. Helens National Volcanic Monument for friendship and logistical support. Useful comments on the manuscript were provided by J. Hubbell, M. Potvin, J. Titus and S. James. This research was supported by a California State University, Chico, Research Award to David Wood, NSF grants BSR 82-07042 and BSR 84-07213 to Roger del Moral and BSR 89-06544 to Roger del Moral and David Wood. Literature Cited Archibold, O. W. 1980. Seed input as a factor in the regeneration of stripmine wastes in Saskatchewan. Canadian Journal of Botany 58:1490-1495. Bishop, J. G. and D. W. Schemske. 1998. Variation in flowering phenology and its consequences for lupines colonizing Mount St. Helens. Ecology 79:534-546. Booth, B. D. and D. W. Larson. 1998. The role of seed rain in determining the assembly of a cliff commu- nity. Journal of Vegetation Science 9:657-669. Chapin, D. M. and L. C. Bliss. 1989. Seedling growth, physiology and survivorship in a subalpine volcanic environment. Ecology 70:1325-1334. Chapin, E S. Ill, L. R. Walker, C. L. Fastie, and L. C. Sharman. 1994. Mechanisms of primary succession following deglaciation at Glacier Bay, Alaska. Eco- logical Monographs 64:149-175. Connell, J. H. AND R. O. Slatyer. 1977. Mechanisms of succession in natural communities and their role in community stability and organization. American Nat- uralist 111:1119-1144. Dale, V. A. 1989. Wind dispersed seeds and plant recov- ery on Mount St. Helens debris avalanche. Canadian Journal of Botany 67:1434-1441. Decker, R. and B. Decker. 1981. The eruptions of Mount St. Helens. Scientific American 244:68-80. DEL Moral, R. 1983. Initial recovery of subalpine vege- tation on Mount St. Helens. American Midland Nat- uralist 109:72-80. DEL Moral, R. 1993. Mechanisms of primary succession on volcanoes: a view from Mount St. Helens. Pp. 79- 100 in J. Miles and D. H. Walton (eds.). Primary Succession On Land. Blackwell Scientific Publica- tions, London. DEL Moral, R. 1998. Early succession on lahars spawned by Mount St. Helens. American Journal of Botany 85: 820-828. DEL Moral, R. 1999. Plant succession on pumice at Mount St. Helens, Washington. American Midland Naturalist 141:101-114. DEL Moral, R. and L. C. Bliss. 1993. Mechanisms of primary succession: insights resulting from the erup- tion of Mount St. Helens. Advances in Ecological Re- search 24:1-66. del Moral, R. and D. M. Wood. 1986. Subalpine veg- etation recovery five years after the Mount St. Helens eruptions. Pp. 215-221 in S.A.C. (ed.). Mount St. Helens: Five Years Later. Eastern Washington State University Press, Cheney, WA. DEL Moral, R. and D. M. Wood. 1988. Dynamics of herbaceous vegetation recovery on Mount St. Helens, Washington, USA, after a volcanic eruption. Vegeta- tio 74:11-27. DEL Moral, R. and D. M. Wood. 1993a. Early primary succession on the volcano Mount St. Helens. Journal of Vegetation Science 4:223-234. DEL Moral, R. and D. M. Wood. 1993b. Early primary succession on a barren volcanic plain at Mount St. Helens, Washington. American Journal of Botany 80: 981-991. DEL Moral, R., J. H. Titus, and A. M. Cook. 1995. Early primary succession on Mount St. Helens, Washing- ton, USA. Journal of Vegetation Science 6:107-120. Dlugosch, K. and R. del Moral. 1999. Vegetational het- erogeneity along elevational gradients. Northwest Science 73:12-18. Edwards, J. S. 1986. Derelicts of dispersal: arthropod fall- 2000] WOOD AND MORAL: SEED RAIN ON MOUNT ST. HELENS 9 out on Pacific Northwest volcanoes. Pp. 196-203 in W. Danthanaraya (ed.), Insect Flight: Dispersal and Migration. Springer Verlag, New York, NY. Edwards, J. S. and P. Sugg. 1993. Arthropod fallout as a resource in the recolonization of Mount St. Helens. Ecology 74:954-958. Halpern, C. B., p. M. Frenzen, J. E. Means, and J. F. Franklin. 1990. Plant succession in areas of scorched and blown-down forest after the 1980 eruption of Mount St. Helens, Washington. Journal of Vegetation Science 1:181-194. Halpern, C. B., J. A. Antos, M. A. Geyer, and A. M. Olson. 1997. Species replacement during early sec- ondary succession: the abrupt decline of a winter an- nual. Ecology 78:621-631. Harper, J. L. 1977. Population Biology of Plants. Aca- demic Press, London. Hickman, J. C. (ed.). 1993. The Jepson Manual: higher plants of California. University of California Press, Berkeley. Hitchcock, C. L., and A. Cronquist. 1973. Flora of the Pacific Northwest. University of Washington Press, Seattle. Jefferson, R. G. and Usher, M. B. 1989. Seed rain dy- namics in disused chalk quarries in the Yorkshire Wolds, England, with special reference to nature con- servation. Biological Conservation 47:123-136. Johnson, C. K. and West, N. E. 1988. Laboratory com- parisons of five seed-trap designs for dry, windy en- vironments. Canadian Journal of Botany 66:346-348. Kruckeberg, a. R. 1987. Plant life on Mount St. Helens before 1980. Pp. 3-23 in D. Bilderback (ed.). Mount St. Helens: One Year Later University of California Press, Berkeley, CA. LiPMAN, P W. AND D. R. Mullineaux. 1981. The 1980 eruptions of Mount St. Helens, Washington. U.S. Geological Survey Professional Paper 1250. Matlack, G. R. 1989. Secondary dispersal of seed across snow in Betula lenta, a gap-colonizing tree species. Journal of Ecology 77:853-869. Morris, W E and D. M. Wood. 1989. The role of Lu- pinus lepidus in succession on Mount St. Helens: fa- cilitation or inhibition? Ecology 70:697-703. Pickett, S. T. A., S. L. Collins, and J. J. Armesto. 1987. Models, mechanisms and pathways of succession. Botanical Review 53:335-371. Rabinowitz, D. and J. K. Rapp. 1980. Seed rain in a North American tall grass prairie. Journal of Applied Ecology 17:793-802. Ryvarden, L. 1971. Studies in seed dispersal. I. Trapping of diaspores in the alpine zone at Finse, Norway. Nor- wegian Journal of Botany 18:215-226. Solbreck, C. and D. Andersson. 1987. Vertical distri- bution of fireweed, Epilobium angustifolium, seeds in the air. Canadian Journal of Botany 65:2177-2178. Stocklin, J. and E. Baumler. 1996. Seed rain, seedling establishment and clonal growth strategies on a gla- cier foreland. Journal of Vegetation Science 7:45-56. Titus, J. H., and R. del Moral. 1998. Seedling estab- lishment in different microsites on Mount St. Helens, Washington, USA. Plant Ecology 134:13-26. Titus, J. H., S. Moore, M. Arnot, and P J. Titus. 1998. Inventory of the vascular flora of the blast zone. Mount St. Helens, Washington. Madrono 45:146- 161. Werner, P. A. 1975. A seed trap for determining patterns of seed deposition in terrestrial plants. Oecologia 59: 145-156. West, N. E. and W. W. Chilcote. 1968. Senecio sylva- ticus in relation to Douglas-fir clearcut succession in the Oregon Coast Range. Ecology 49:1101-1107. WiLLSON, M. F. 1993. Dispersal mode, seed shadows, and colonization patterns. Vegetatio 107/108:261-280. Wood, D. M. and R. del Moral. 1987. Mechanisms of early primary succession in subalpine habitats on Mount St. Helens. Ecology 68:780-790. Wood, D. M. and R. del Moral. 1988. Colonizing plants on the Pumice Plain, Mount St. Helens, Washington. American Journal of Botany. 75:1228-1237. Wood, D. M. and W. F. Morris. 1990. Ecological con- straints to seedling establishment on the Pumice Plain, Mount St. Helens, Washington. American Jour- nal of Botany. 77:1411-1418. Madrono, Vol. 47, No. 1, pp. 10-20, 2000 ENVIRONMENTAL GRADIENTS AND VEGETATION STRUCTURE ON SOUTH TEXAS COASTAL CLAY DUNES Kern Ewing College of Forest Resources, Box 354115, University of Washington, Seattle, Washington 98195 Abstract Clay dunes are unusual geological features that occur near playas, lagoons, or flats that are sometimes wet but dry out annually. If the sediment in these ephemeral bodies of water contains clay, and if there are strong prevailing winds, flakes or granules of clayey material are transported during the dry season and are caught by edge vegetation. The clay particles moisten in the dew or rain and stick together, eventually creating dunes that support vegetation. Known locally as lomas, the clay dunes along the Gulf coast of Texas and Mexico reach their greatest stature near the mouth of the Rio Grande River, where this study was carried out. These dunes support ecologically unique vegetation assemblages. They sit, like islands, in hypersaline lagoons. Sharp environmental gradients separate halophytes from typical coast- al thornscrub vegetation. Endangered animal species such as ocelots live in the thornscrub. Development pressures along the border threaten their existence, and the construction of artificial lomas has been proposed. In this paper I characterize four loma plant communities. The first community is found in the adjacent hypersaline Flats, and is limited to halophytes. The second community is found in lower but still elevated salinities at the Edge of the lomas. At low salinities atop the lomas are the dense Thornscrub community, and a Mixed Halophyte and Thornscrub community that is hypothesized to be the result of disturbance. Analysis of elevation and salinity at plots along transects through the lomas allows me to correlate individual plant species with salinity preferences and community membership. An interesting outcome is that while a number of species have fidelity to one community type, there are quite a few bridging species that are found in two community types. This information has important implications for the degree of precision required when attempting to restore or create the clay dune ecosystem. The clay dunes that occur along the Laguna Ma- dre on the southern part of the Texas Gulf coast are interesting both geologically and biologically. Clay dunes are found in association with playas in west Texas and New Mexico, and are reported in Aus- tralia and Africa. The coastal clay dunes described in this paper reach their greatest stature and number near the mouth of the Rio Grande River and de- crease to the north and south. Similarly, the vege- tation of the south Texas delta of the Rio Grande is unique and in some places luxuriant, but dimin- ishes as the distance from both the river and the Gulf increases. The clay dunes, or lomas as they are known lo- cally, are rendered more exotic by the fact that they exist as low hills with non-halophytic vegetation, sitting in the middle of extensive hypersaline wind tidal flats or lagoons. The lagoons are periodically inundated by wind tides or hurricanes, but are sub- jected to long periods of drying in the hot south Texas climate. Salts concentrate and the surface of the lagoons may become dry. In addition to the salinity and periodic droughts, there is a persistent brisk wind out of the southeast for much of the year. In this semiarid, semitropical climate, the warm months may be generally described as any but January and February, and even during these months, daytime temperatures above 30°C are com- mon. Lomas near Boca Chica, the beach north of and adjacent to the mouth of the Rio Grande River, are covered with thornscrub vegetation, some of which is commonly found over much of the Tamaulipan biotic province. In addition, species characteristic of the lomas and other coastal areas are found, in- cluding Citharexylum berlandieri Robins., Mayten- us texana Lundell, Prosopis reptans Benth., Echeandia chandleri (Greenm. & Thomps.) M. C. Johnst., Monanthochloe littoralis Engelm., and Yucca treculeana Carr. Most of the lomas in the Boca Chica area are named, and the site for this work is called Loma Tio Alejos. It has a roughly north-south orientation, is 200 to 300 meters wide by 600 meters long, and rises about 7.5 m out of a lagoon, which is at an elevation of about 1.5 m. The south end of the loma is 300 meters north of a bend in the Rio Grande River; the east side is about 12 km from the Gulf of Mexico. This site and much of the surrounding land is now part of the Lower Rio Grande Valley National Wildlife Refuge. It has long been known that non-halophytic veg- etation grows on lomas, and that hypersaline marshes and flats surround them. They are recog- nized as a unique biotic community by the U.S.F.W.S., and are included in a proposed wildlife corridor running down the Rio Grande and up the coast. One impetus for this study is the extensive restoration program in place on the L.R.G.V. Na- tional Wildlife Refuge and the potential for con- 2000] EWING: VEGETATION OF CLAY DUNES 11 structing or restoring lomas to create habitat. Off- refuge, there are significant pressures from en- croaching conunercial and residential development and from the proposed construction of a new inter- national bridge and its infrastructure 8-km west of this site. Knowledge of typical vegetation composition on lomas and the relationships among vegetation, loma elevation and soil salinity are critical to understand- ing what controls the vegetation structure. Loma vegetation is known to be dense, and often much shorter than it would be in other locations. In this study I measured vegetation composition, woody plant height and density, canopy cover, elevation and soil salinity in quadrats along transects across Loma Tio Alejos. Ordination and classification anal- yses were performed on cover data. The nature of the relationship between loma elevation and salinity was determined, and species affinity to sites was related to salinity and elevation. Clay dunes. The clay dunes along the Gulf Coast in southern Texas have been remarked upon almost since the first accounts of the exploration of the area, probably because these explorations sought river mouths and disembarked from coastal areas. Coffey (1909), while on a soil survey of the region for U.S.D.A., saw the dunes and hypothesized that they were formed by granules of clay, which were blown off the surface of dried lagoons during hot, windy summers. The particles blew to the edge of the lagoon, were caught by vegetation or by wrack or debris, and began to accumulate. Rainfall or the humidity of the nights caused the particles to coa- lesce. Coffey further noted that they were found near the Rio Grande because the rains are seasonal and lagoons dry out; in more humid climates such conditions do not occur. Foscue (1932) noted that the dunes looked like small islands covered with brush. Huffman and Price (1949) and Price and Komicker (1961) compared clay dunes all along the Texas and Mexican coast and determined that they existed along the mainland coast from Soto la Marina River in Tamaulipas, Mexico to St. Charles Bay (at the Aransas National Wildlife Refuge) in Texas. The dunes are highest at the Rio Grande (10 m), and become lower (1 m) in the more humid climates to the north and south. These authors es- sentially agreed with Coffey about the formation of the dunes, adding that they probably grow only dur- ing hot months (March to November), retain a loosely porous structure, and represent about 5000 years of growth since beginning of the current still- stand of sea level. During a seven year period of drought in the fifties, about a foot of loosely con- solidated pellets accumulated. Their height made the dunes attractive camp sites for the coastal In- dian tribes that fished in the area. Aboriginal arti- facts occur from about mid-dune to near the top foot, and European artifacts occur near the top. In addition to their use by humans, the endangered ocelot (Tewes 1982) also uses clay dunes. Tamaulipan Thornscrub. Brown (1994) de- scribes the Tamaulipan biotic province as being one of several provinces that are semidesert scrublands. Such systems are dominated by thorny shrubs and small trees, and characterize much of the world's tropic -subtropic zones. They are found in Australia (mulga), southern Africa (bush). South America (chaco-seco), Mexico (matorral) and Texas (chap- arral). They are drought-deciduous communities that occupy a position on a moisture gradient some- where between desert scrub and woodland or forest. They often have an irregularly layered overstory between 2 and 8 m in height, and are typically com- posed of spinose, microphyllous, and succulent life forms. Thornscrub is often in competition with grassland, and may increase under grazing pres- sures, with fire suppression, or on poorer soils. MuUer (1947) observed that east central Coahui- la, southern Texas, northern Nuevo Leon and north- em Tamaulipas all have a vegetation form that is similar. Shreve (1917) called it Texas semi-desert. MuUer proposed that it be called Tamaulipan thorn shrub. The more luxuriant and tree-dominated forms found in south Texas and Tamaulipas were called Tamaulipan thorn forest. These environments differ from the adjacent Chihuahuan desert shrub in that they are found at lower elevations, have more rainfall, and are exposed to winds from the Gulf of Mexico. With these habitat differences are also found more thorny shrubs, an abundance of grasses, more luxuriant growth of shrubs, a richer flora, and more numerous characteristic species. With the increase in species there is also a greater number of variants of the vegetation formation. Blair (1950) included the area in Texas south of the Balcones fault line (which runs from Austin through San Antonio) in his Tamaulipan province. He described the biota of the province as neotropic, strongly diluted by Sonoran biota characteristic of the southwestern U.S. and parts of Mexico, and by biota characteristic of the forests blanketing the Gulf coastal plain. The climate is semiarid and megathermal. From the coast westward, the brush thins as available moisture declines. In Cameron County, at the southern tip of the state, average annual precipitation is just above 25 inches. Mean maximum temperature is 95° in July, mean mini- mum is 51° in January. Rainfall peaks during trop- ical storm season (centered on September). Long periods of drought, during which there is little or no rain for 4-6 months, are conmion; periods dur- ing which drought years occur for 3-5 years in suc- cession are also conamon. A strong, persistent hot wind blows out of the southeast for much of the year. Probably the earliest exhaustive description of the thornscrub vegetation of the Rio Grande delta was given by Clover (1937). Later Blair (1950), in 12 MADRONO [Vol. 47 his delineation of the biotic provinces of Texas, would call the area on the floodplain the Matamo- ran district of the Tamaulipan Biotic Province. Clo- ver justified the use of the term chaparral for the shorter vegetation of the area, saying that it referred to chaparro prieto {Acacia rigidula Benth.). Mes- quital is the term used for Prosopis glandulosa Tor- rey-dominated communities, and sacatal for grass- lands. Currently, two general types of brush habi- tats are recognized in the area. The first is referred to as riparian and scrub forests (associated with the Rio Grande, and producing taller vegetation); the second is upland thornscrub and thorn woodland (Jahrsdoerfer and Leslie 1988). Clover (1937) described the vegetation of the clay dunes near the coast as being similar to salt- affected thornscrub nearby, but being composed of shrubs twisted by the heavy winds. Dominants list- ed were Pithecellobium ebano (Berl.) C. H. Mull., Leucophyllum frutescens (Berl.) I. M. Johnst., Ziz- iphus obtusifolia (Torrey & A. Gray) A. Gray, Cas- tela texana (T. & G.) Rose, Randia rhagocarpa Standi., Forestiera angustifolia Torr., Prosopis glandulosa Torr. and Celtis pallida Torr. Between the clay dune "islands" and the main chaparral- mesquital was a transition zone and sacahuistal (dominated by Spartina spartinae (Trin.) Merr.). USFWS (1997) added Citharexylem berlandieri, Erythrina herbacea, Dalea scandens (Mill.) R. T. Clausen, Echeandia chandleri and Sporobolus thar- pii Hithc. as being found exclusively in or near the /om«-coastal brushland community. Johnson (1963) added that the windward sides of some of the dunes were covered with a thick growth of Sporobolus wrightii Scribn. (sacaton). Methods In late October 1998, I began the establishment of two transects across Loma Tio Alejos. The first transect ran approximately east-west. It started in an unvegetated area of hypersaline lagoon to the west of the loma, crossed 200-m of halophytic veg- etation, encountered the southern part of the loma and entered thornscrub vegetation. It climbed for the next 100 m to the ridgeline and then dropped, over the following 100-m, to the edge of halophytic vegetation on the east of the loma. The transect then ended after it traversed 30 m of the halophytic vegetation. Rather than being symmetrical, the loma is kidney-bean shaped, so that transect two could be oriented in a north-south direction and cross the north end of the loma almost perpendic- ular to its axis. The second transect was finished by the begin- ning of December. It started on the south side of the north end of the loma and traveled for 30 m in halophytic vegetation, then entered the brush and traveled 80 m to the ridgeline. It then went down through a depression and up to another ridgeline, traversing extremely dense brush for about 70 m. The transect descended through thornscrub for 70 | meters and was terminated about 50 m into the hal- ophytic vegetation to the north of the loma. j Distance along transects was measured with sur- veying tapes, and station stakes were placed every ten meters (stations 0+00, 0+10, etc.). Differential | leveling, employing a Keuffel and Esser optical | level, was used to determine relative elevations along each transect. Elevations on transect one were tied to elevations on transect two by closing a leveling loop from one transect to the other along a trail which ran down the ridge of the loma. The elevation of the transects was then fitted to a USGS contour map of the loma and elevations from that map were used to register the high and low points surveyed. Vegetation was sampled in 10 X 20 m quadrats along most of both transects vegetation was so thick from station 0+20 to station 2+00 along tran- sect 2 that it was sampled using 5 X 20 m quadrats oriented with their long axis parallel to the ma- chete-cut line through the brush. By the conclusion of the investigation, 57 quadrats were measured and 76 plant species were found. In each quadrat the percent cover of herbaceous species was recorded, as well as the number of species, the number of individuals of each woody species, height and two crown- width dimensions for woody species. Cover for woody species was calculated by averaging the two dimensions and calculating the area of a circle with this average as the diameter. Soil samples for salinity measurements (Abbott 1967) were taken from each quadrat using a 3.8- cm diameter corer. Cores were generally 10-15 cm long, and were taken after removing organic matter and debris from the soil surface. Cores were ex- truded onto a tray and the length of each sample was measured, allowing the calculation of soil vol- ume. Wet weight of each core was measured, all were oven-dried at 80°C, and dry weights were measured. Dried cores were placed in flasks and a volume of water twice the original volume of the cores was added to each. The flasks were sealed with rubber stoppers and agitated for three days. They were then allowed to settle in a cold room for three days and the clear supernatant was removed with a syringe. Osmolality of the soil extract was measured with an Advanced Instruments Model 3300 Micro Os- mometer (Advanced Instruments, Inc, Two Tech- nology Way, Norwood, MA 02062). From the os- molality of the soil extract, values were calculated for the osmotically active solutes per unit volume of soil ("a" below) and the apparent salinity of the soil solution at the time the sample was taken ("e" below) (Mahall and Park 1976). The following cal- culations were made to arrive at these two values. b X c b X c ; 2000] EWING: VEGETATION OF CLAY DUNES 13 where a = osmotically active solutes per unit volume of soil (m-osm. ml '); b — osmolality (m-osm. ml ' water) of soil extract (from freezing point depression); c = volume of water added to dry soil core (ml); d = original volume of soil core (ml); e = apparent salinity of soil solution (m-osm. g~' water); f = weight of water in soil core (g). The soil solution salinity was converted to ppt for figures. Two data sets, one containing species cover in- formation for each quadrat and the second contain- ing elevation and soil salinity data at each quadrat were entered as the primary and secondary matrices in the multivariate software package PC-Ord (McCune and Mefford 1997). Classification was performed using TWINSPAN, and ordination was performed using DCA (detrended correspondence analysis), which uses the stand species matrix; the joint plot option was used to overlay environmental variable vectors over the DCA-generated ordination plot. TWINSPAN is a divisive classification technique that divides all initial stands in an analysis into two groups using an ordination, then iteratively refines the division. Each group formed is then divided into two new groups. (Jongman et al 1987). The mechanical division of existing groups can go on until some stopping rule is triggered (maximum number of divisions or minimum number of stands in a group may be specified). The actual selection of what divisions to accept may depend upon whether further divisions add to the explanatory power of the analysis (Gauch 1982). In this anal- ysis, four groups were used because each of the four groups represented a homogeneity of species composition and was characterized by similar dom- inant species. Results Elevation. The elevation of the saline flats around the lomas in this area is about 1.5 m, USGS datum. On this loma, over a distance of 100-150 m, the measured transects rose to about 6.5-7 m at the height of land, then dropped again (Fig. 1). The elevation of the highest point on the loma was es- timated at about 9 m. Vegetation. Seventy-six species were found with- in the sampling quadrats along the two transects across Loma Tio Alejos. Twenty-seven of them were erect, woody plants. There were also 8 grass- es, and a number of halophytes, weedy herbaceous species, and herbaceous understory species. Echeandia chandleri, a lily limited to clay soils in south Texas and described as rare (USFWS 1997), was common here. Ophioglossum vulgatum L., a vascular cryptogam, was found growing under Pro- sopis glandulosa trees. The vegetation gradient from the saline flats to the thomscrub of the lomas was short and steep, but there was surprising over- lap of thomscrub and halophytic species. Vegetation composition and cover data were an- alyzed using multivariate analysis. Classification of stands and species was accomplished using TWIN- SPAN. One of the products of a TWINSPAN anal- ysis is a joint ordination of stands and species called a two-way table. The species list from a two- way table is ordered, i.e., the species that are on either end of the list are usually found in complete- ly different environments and species which are next to one another on the list are usually found together. On the list in Table 1, species at the top are from sites near the highest elevations of the loma, while species at the bottom occur mostly in the adjacent flats and at the loma edges. I elected to stop the TWINSPAN procedure after four groups of stands had been generated because each of the four groups represented a relatively ho- mogeneous association of species and was charac- terized by a unique dominant species or group of dominants. I will describe the important species, the location of stands, and the general environment in which each group occurs. The first division segregated the high-diversity, non-halophytic thomscrub vegetation into one group. Woody thomscrub species and their associ- ated herbaceous understory or gap species domi- nate it. The other group from this division includes all of the species known to be halophytic, but hal- ophytes are not limited to the second group. The second division divided stands in the thom- scmb vegetation group into two smaller groups, one of which is made up of species characteristic of more widely distributed coastal upland sites. All of the sites in this subgroup are found on the higher elevations of the north transect; the vegetation there is characterized by a dense canopy and shortened stature. On Figure 2, this association is called "'Thomscrub''. The other subgroup in this division includes more salt-tolerant thomscmb which is found on the lower ends of the north transect and on the south transect. On Figure 2 this association is called ''Mixed Thomscrub and Halophytes''. In- dicator species for the group are Prosopis reptans, Maytenus texana Lundell. and Ericamera austro- texana M. C. Johnst. Almost all of the Yucca tre- culeana Carr. and Prosopis glandulosa are also found in this grouping. This second subgroup could be further divided into closed canopy and open can- opy groupings, which would have slight species differences. The third division divides the stands containing mostly halophytic vegetation into two subgroups. The first subgroup is the association that makes up the low shmbby vegetation around the edges of the loma. On Figure 2 this is called "Edge" . Indicator species for this subgroup are Borrichia frutescens (L.) DC and Lycium carolinianum Walt. The sec- 14 MADRONO [Vol. 47 NORTH TRANSECT SALINITY SOUTH TRANSECT SALINITY 80.0 a ^ <9 <^ 'P ^ >^ '§> ^ ^ ^ "b^ n" C)" cT c> cT cT "jT o" v v n k n n s n S> ef^ r& Fig. 1. Elevation and salinity along the north transect (which runs north-south) and the south transect (which runs east- west) across Loma Tio Alejos, near the mouth of the Rio Grande River in south Texas. Salinity is represented by a dashed line, elevation by a solid line. The sahnity is the apparent soil water salinity obtained by measuring the osmolality of a known volume of water into which a dried soil sample was mixed; salinity at the water content of the soil when the sample was taken was then calculated. ond subgroup is dominated by three species {Sali- comia virginica L., Monanthochloe littoralis En- gelm. and Batis maritima L.) and little else; sites making up this subgroup occur in the extremely salty flats surrounding the lomas. On Figure 2 this association is referred to as "'Flats''. The same data set that was used to obtain a clas- sification of stands and species with TWINSPAN was subjected to an ordination analysis. Ordination allows the investigator to plot stands or species in a multi-dimensional space that can be interpreted in terms of environmental variation; Detrended Correspondence Analysis (DCA), an indirect ordi- nation technique, was used. Since an environmental matrix of stands by environmental data (salinity, elevation) was available, the joint plot option of PC-Ord was used to plot environmental vectors on the DCA ordination plots. The first DCA axis had and an eigenvalue of 0.859, indicating that a sub- stantial amount of the variation in the data set was accounted for by this axis. When plotted as a joint plot, the arrow representing salinity was almost 2000] EWING: VEGETATION OF CLAY DUNES 15 identical to axis one, and the arrow representing elevation, while negatively correlated with salinity, was very close to axis one. Canonical Correspondence Analysis (CCA) al- lows the investigator to constrain the ordination axes to some combination of directly measured en- vironmental variables (Jongman et al. 1987). When this was done using elevation and salinity as the environmental variables, the eigenvalue for the first axis was 0.662, which is still high (Jongman et al. 1987). In the CCA analysis the arrows for elevation and salinity were highly correlated with axis 1 . The results of these ordinations indicate that elevation and salinity are environmental factors that account for most of the variation in vegetation structure on the lomas. They are inversely correlated; as eleva- tion increases, salinity decreases. Salinity. Salinity was high in the flats but dropped very quickly as the loma elevation rose above that of the surrounding flats. (Fig. 1). Salinity per volume of soil and apparent salinity of soil wa- ter were highly correlated (r = 0.995), so apparent salinity in ppt will be used to discuss the relation- ship among elevation, vegetation and salinity. Av- erage soil water salinity in plots in the ''Flats'' veg- etation association was 64.1 ± 11.8 ppt. This salin- ity is significantly (P = 0.05) greater than that in any of the other associations. Mean soil water sa- linity in the ''Edge'' association, 6.3 ±1.9 ppt, is higher than that in the "Mixed" or "Thornscrub" groups, though not significantly so. Mean salinity in the "Mixed" group is 2.1 ± 0.4, and in the "Thornscrub" group is 2.0 ± 0.3 ppt. Discussion The ecotone between the hypersaline flats of La- guna Madre and the coastal thornscrub in south Texas has created interesting and unexpected veg- etation associations. The lomas or clay dunes are a microcosm of this contact, and they reach their maximum expression near the mouth of the Rio Grande River. Coastal thornscrub on the lomas and on the nearby mainland is valued because of its importance as wildlife habitat (ocelots, birds, and butterflies). It is part of a breathtaking diversity of vegetation; Jahrsdoerfer and Leslie (1988) indicat- ed that there were 265 native woody species in the thornscrub of southern Texas. It is the site of scarce and unusual plant species including the species of concern Echeandia chandleri, Citharexylum ber- landieri Robins, and the endemic Sporobolus thar- pii Hitchc. (Jahrsdoerfer and Leslie 1988). The La- guna Madre is unique in that it is a huge lagoon, which has little freshwater input except from trop- ical storms. The hot climate, persistent winds and a tendency to experience prolonged periods of drought have created a water body that becomes hypersaline. Many of the lomas near the site of this study rise out of wind flats at the edge of the La- guna Madre; these flats are sometimes inundated but are very salty during most years. Many of the lomas near the mouth of the Rio Grande and behind Boca Chica beach are now pro- tected and part of the Lower Rio Grande Valley National Wildlife Refuge. This refuge and the La- guna Atascosa National Wildlife Refuge on the coast 20 km to the north are both sites of active vegetation restoration programs. Because of the sharp environmental gradients in areas adjacent to salt flats and lagoons, and because of unique tol- erances and preferences of plant species which would normally be selected for restoration plant- ings (in general, dominant and secondary woody species), information about the sorting of species along environmental gradients is important. Since the lomas rise like small islands out of low and level salt flats, their elevation was predicted to be an important environmental axis; this proved to be the case. Since lomas support species common to sites that are not salt-affected, and because they are surrounded by halophytic vegetation that is tolerant to high concentrations of salts, a salinity gradient was predicted. This also proved to be true. Other environmental factors may also shape veg- etation structure. Winds off the nearby gulf are per- sistent and may result in dwarfing of vegetation. Soils sampled were generally silty clays or clayey silts, with more clay in soils in the flats. Lomas have a history of use by people, and there are roads leading to them, around them and across them. There are excavation sites, dumps, and disturbed areas with weedy vegetation (primary weeds are the introduced pasture and lawn grasses Cenchrus cil- iaris L., Cynodon dactylon (L.) Pers., Dichanthium annulatum Stapf. and Panicum maximum Jacq.); some of the vegetation plots occurred in these ar- eas. Multivariate analysis indicated that elevation and salinity change explains a very high amount of the variation in vegetation on Loma Tio Alejos. The analysis of data from 57 quadrats along two transects across the loma was carried out by per- forming a classification procedure (TWINSPAN) and an ordination procedure (DC A). As a result of the TWINSPAN analysis, four vegetation group- ings or associations were identified. The groups were called "Thornscrub" , "Mixed Thornscrub and Halophytes" , "Edge" and "Flats". The 16 plots in the Thornscrub group contained 36 plant species. For Mixed Thornscrub and Halophytes the numbers were 20 plots and 67 species, for Edge 13 plots and 41 species, and for Flats 8 plots and 11 species. All except one of the plots assigned to the Thorn- scrub group were on the north transect; that one occurred in very thick brush on the south transect. This association is made up of plants in very dense, short (3-4 m) vegetation. Vegetation on the north transect may have been subjected to greater wind intensity, because the south part of the loma is par- tially protected from the wind by vegetation along 16 MADRONO [Vol. 47 Table 1 . List of Species Found along Transects at Loma Tio Alejos. Species have been ranked by TWINSPAN classification program so that those generally found at the higher elevation, lower salinity sites occur at the top of the list; those found at the lower elevations and in the saline flats are at the bottom. Frequency of occurrence in plots of each of the four TWINSPAN community types at the site (Thomscrub, Mixed Thomscrub and Halophytes, Edge and Flats) is shown for each species as a percentage. Numbers of plots of each of the community types are respectively n = 16, 20, 13 and 8. Thomscrub Mixed Edge Flats Caste la texana 44 10 0 0 Pithecellobium pallens 6 0 0 0 Malpighia glabra 25 5 0 0 Bastardia viscosa 31 5 0 0 Rivina humilis 50 20 0 0 Celtis pallida 63 30 8 0 Pithecellobium ebano 19 0 0 0 Phaulothamnus spinescens 81 45 8 0 Randia rhagocarpa 69 10 0 0 Lycium berlandieri 13 5 0 0 Aloysia gratissima 13 0 0 0 Citharexylum berlandieri 94 85 8 0 Zanthoxylum fagara 100 85 0 0 Karwinskia humboldtiana 63 35 8 0 Lantana horrida 38 60 0 0 Capsicum annum 6 20 0 0 Schaejferia cuneifolia 25 40 0 0 Passiflora foetida 6 5 0 0 Cissus incisa 44 60 8 0 Forestiera angustifolia 19 30 8 0 Verbesina microptera 38 30 0 0 Zisiphus obtusifolia 13 20 8 0 Isocoma drummondii 19 80 15 0 Allowissadula lozani 19 55 8 0 Yucca treculeana 13 60 15 0 Prosopis glandulosa 25 85 8 0 Eupatorium azureum 50 70 8 0 Leucophyllum frutescens 38 65 8 0 Condalia hookeri 6 10 0 0 Echeandia chandleri 13 25 0 0 Gymnosperma glutinosum 6 10 0 0 Ericameria austrotexana 0 80 8 0 Eupatorium incarnatum 0 20 0 0 Cenchrus incertus 0 10 0 0 Atriplex acanthocarpa 0 5 0 0 Wedelia hispida 0 15 0 0 Physalis cinerascens 0 20 0 0 Ophioglossum vulgatum 0 20 0 0 Acacia farnesiana 0 5 0 0 Trixis inula 0 5 0 0 Dichanthium annulatum 0 40 0 0 Croton cortesianus 0 5 0 0 Sida ciliaris 0 10 0 0 Malvastrum americanum 0 15 0 0 Ibervillea lindheimeri 0 10 0 0 Chenopodium ambrosoides 0 5 0 0 Cenchrus ciliaris 0 50 15 0 Sarcostema cynanchoides 0 30 15 0 Cynodon dactylon 0 10 8 0 Croton leucophyllus 0 10 8 0 Opuntia leptocaulis 6 30 31 0 Trandescantia micrantha 6 10 23 0 Acleisanthes obtusa 6 5 8 0 Borrichia frutescens 13 55 85 0 Oxalis drummondii 0 5 8 0 Solanum eleagnifolium 0 10 8 0 Sporobolus wrightii 0 10 15 0 Panicum maximum 0 5 0 13 Oxalis dichondrifolia 6 15 15 0 2000] EWING: VEGETATION OF CLAY DUNES Table 1. Continued. 17 Thornscrub Mixed Edge Flats EvoIvuIms ttlsinoidcs 0 5 8 0 Maytenus texcina 0 75 77 13 Pfosopis rcptans 0 70 85 25 Mdchciefcintherci phylloccphciln 0 45 46 0 Spciftinci spcirtincie 0 10 15 0 SuQedci lineciT'is 0 0 23 25 CressQ nudicciulis 0 10 23 25 jJCtttCC' ILlLt VII ^ If lll^lA- 0 0 23 100 Talinum paniculatum 0 5 8 25 AtrinlfY mntnmnrfriKi K 0 0 0 13 Lycium carolinianum 0 20 85 0 Limonium nashii 0 0 31 0 Echinocactus setispinus var. setaceus 0 0 8 0 Distichlis spicata 0 0 8 0 Monanthochloe littoralis 0 30 100 75 Opuntia engelmannii 0 10 38 13 Batis maritima 0 5 69 100 the river and by a road berm. Plots assigned to the Mixed group occurred in a more disturbed area on the north end of the north transect, and at the upper elevations of the south transect. There was erosion and open areas in both locations, so human distur- bance may have been partially involved in the cre- ation of such sites. Since salinity and elevation dif- ferences were inconsequential between plots in these two associations, the hypothesis that the Mixed association is generated by disturbance should be investigated more thoroughly. Any res- toration attempt would create a disturbed environ- ment, and so the Mixed association might be the expected mid-successional vegetation type on less- salty soils. Prosopis glandulosa, well-known as a self-seeder on open sites in South Texas (Archer et al. 1988), is a dominant species in the Mixed as- sociation but found in few plots in the more dense Thornscrub association. Plots in the Edge association occur in a band about 30-40 m wide around the loma, and are vi- sually different from the adjacent brush because they are open, and woody vegetation in them is either short or widely spaced. At the lower and salt- ier sites in this association, plsint diversity dimin- ishes to an average of 4 species per plot, and plants are generally less than a few decimeters tall. The saline Flats are characterized by a few species, and in places the salinity may become so great that no vascular plants occur. There was a substantial in- crease in soil salinity at the interface between the Edge association and the Flats (Fig. 1). At one un- vegetated quadrat in the flats, a soil water salinity of 187 ppt, roughly five times that of sea water, was measured. The mean soil water salinity of soil samples taken from plots in the Flats was 64 ppt, or almost twice the salinity of seawater. The Flats are truly the province of halophytes. Localized or general evaporation and concentration probably result in much higher localized salinity during drought pe- riods. Salinity in the Mixed and Thornscrub asso- ciations, on the other hand, fell under or around the 1.5-2 ppt salinity threshold that is generally con- sidered the point below which crop plants have no salinity problems (Hartman et al. 1990). Salinity in plots in the Edge association were generally in the range within which plants are likely to be affected, but not so high as to limit plant composition to halophytes. Each of the four vegetation associations identi- fied by classification analysis was characterized by a set of dominant species (based upon total cover). In analyzing the species composition of each as- sociation, it became evident that there were some species which preferred conditions found in sites limited to one vegetation association, but there were also many species that did quite well in two of the associations. For instance, Castela texana (T. & G.) Rose, and Randia rhagocarpa (Fig. 2a) were common in Thornscrub association and rare in the Mixed association. Citharexylum berlandieri (Fig. 2b) and Zanthoxylum fagara (L.) Sarg. were in both Thornscrub and Mixed associations. Yucca trecu- leana Carr. (Fig. 2c) preferred the Mixed associa- tion, while Borrichia frutescens (L.) DC (Fig. 2d), Maytenus texana Lundell. and Prosopis reptans were found in both Mixed and Edge associations. This combination of species with fidelity to an as- sociation and species which overlap associations continued with Lycium carolinianum Walt. (Fig. 2e) found in Edge, Monanthochloe littoralis and Batis maritima L. (Fig. 2f) found in Edge and Flats, and Salicornia virginiana L. (Fig. 2g) found in Flats. All of the species listed by Shindle and Tewes (1998) as recommended for the restoration of oce- lot habitat, as well as a broad variety of others, are found on the lomas. This paper has presented in- formation about the environmental preferences of loma species. The mixing of species which have 18 MADRONO [Vol. 47 1 1 ^ \i 1 THORNSCRUB T ! 1 " EDGE MIXED THORNSCRU AND HAiOPHYTES ' lu 1 1 i EDGE 1 SOUTH TRANSECT r MIXED THORNSCRUB Ann MAI nPMVTCt NORTH TRANSECT n, c [JUAfiBCl 'LUM BERLANDIERI - | ! 1 i 1 / ! ^ j THORNSCRUB IXED THORNSCRUB 1 Ann HAI nPHvn:* 1 EDGE i IXED EDGE C. YUCCA TRECULEANA ilORTH TRANSECT 1 0 1 i 0 0 MIXED THORNSCRUB - ANO >mOPWYTgS THORNSCRUB 0 EDGE ^IXED EDGE D. E 30RRICHI 1 ' ■ NORTH TRANSECT AFRUTESCENS i ! , 1 h. » IB' MIXED THORNSCRUB AND HALOPHYTES THORNSCRUB EDGE ^IXb 1 ^ 1 EDGE "A- SOUTH TRANSECT /\ j FEATS — EDO i MIXED THORNSCRUB ^^Mj 1 AWDWALOPHtTEb ^g.^^^^^^^^^^ — : EDGE SOUTH TRANSECT ^1 rlJVTS / ANO HALOPHYTES EDGE EDGE SOUTH TRANSECT 1 f Hh MIXED THORNSCRUB FLATS ^^^^y^ EDGE AND HALOPHYTES EDGE 1 1 Fig. 2. Presence of species in plots along transects. A solid vertical bar indicates that the species shown was present in a plot: a.) Randia rhagocarpa, occurring primarily in the Thornscrub association, b.) Citharexylum berlandieri, occurring in both the Thornscrub and Mixed associations, c.) Yucca treculeana, occurring primarily in the Mixed association, d.) Borrichia frutescens, occurring in both the Mixed and Edge associations, e.) Lycium carolinianum, occurring primarily in the Edge association, f.) Batis maritima, occurring in both the Edge and Flats associations, g.) Salicornia virginica, occurring primarily in the Flats association. narrow habitat ranges with species which have broader habitat ranges when planting a restoration project is wasteful of plant materials. For the plant installation phase of a restoration project, some spe- cies may be placed on the landscape with less pre- cision, but others require an exact understanding of the species preferences and the site conditions. Plant materials for restoration or creation of lomas 2000] EWING: VEGETATION OF CLAY DUNES 19 NORTH TRANSECT E. LYCIUM CAROLINIANUM J \/ IXEO THORNSCRU THORNSCRUB 3 EDGE AND HALOPHYTES. MIXED 1 EDGE SOUTH TRANSECT 1 1 — MIXED THORNSCRUB AND HALOPHYTES r t DGE FIATS -EOGC- \ NORTH TRANSECT F. BATIS MARITIMA SOUTH TRANSECT i 1 \ .ll 1 XED THORNSCRL THORNSCRUB 8 EDGE - AND halophyte; 1 mXEO ! ! i EDGE ! 1 i NORTH TRANSECT G. SALICORNIA VIRGINICA ' '7\ ^ — ' / MIXED THORNSCRUB ANn HAl nPHYTFS EDGE FIATS EDGE SOUTH TRANSECT ^ \ Mi / N i M / ,X'^^'lXED THCRNSCRl g THORNSCRUB ^^M. ^ FnnF AND HALOPHYTE MIXED 1 / \ 1 1 / \ MIXED THORNSCRUa AND HALOPHYTES EDGE H FLATS EDGE Fig. 2. Continued. will probably never be abundant, so placement of seedlings or seeds into the proper environmental zone will be critical to the success of a restoration project. In conclusion, this work has confirmed that coastal clay dunes or lomas are unique systems that are persistent over time. Unusual plant associations (thomscrub vegetation and halophytes in close proximity or mixed) and rare and threatened plants {Echeandia chandleri, Citharexylum berlandieri) are found on them; their individual vegetation structure can be complex. They are known to be valuable as wildlife habitat. Vegetation structure across lomas varies along environmental gradients, which can be predicted for the most part by mea- suring elevation and salinity. Reports in the litera- ture suggest that wind direction can also be an im- portant factor in vegetation composition and size (Clover 1937). For many centuries, these unique systems have been isolated and not greatly dam- aged. Population pressure and commercial devel- opment now pose a threat to the vegetation systems and the wildlife that they support. Restoration in other areas of south Texas Tamaulipan thornscrub has been undertaken successfully, and the core of an extensive wildlife corridor is being created along the coast and up the Rio Grande River. The resto- ration or creation of loma vegetation to augment habitat and add to the wildlife corridor is an im- portant and achievable element of this restoration. The ability to restore unique ecosystems like the clay dunes, if indeed we have that ability, does not mean that there is no need for conservation of such unusual habitats. Conservation is an integral part of the U.S.F.W.S. plan for development of a wildlife corridor in south Texas. Important parcels have been identified and a considerable acreage of land upon which dunes sit has been purchased. Resto- ration can augment the effectiveness of conserva- tion in a number of ways, including the creation of 20 MADRONO [Vol. 47 buffers, the increase in the effective size of a con- served parcel, the creation of corridors, and the ini- tiation of a successional trajectory that will even- tually result in an ecosystem that is not much dif- ferent from one at a conserved site. Acknowledgments I would like to thank Chris Best, Monica Monk and Frank Gonzales for help in selecting this site, collecting data and reviewing results. Support from the U.S.F.W.S was provided by Donna Howell, Ken Merritt, and Larry Ditto. Liz Van Volkenburgh kindly allowed me to use her new freezing point depression osmometer and Kari Stiles showed me how to use it. Literature Cited Abbot, W. 1967. Salinity determination by freezing-point depression osmometry. Lagunas Costeras, un Sim- posio. Mem. Simp. Intern. Lagunas Costeras. UNAM-UNESCO, Nov 28-30. Mexico, D.F.:341- 348. Archer, S., C. Scifres and C. R. Bassham. 1988. Au- togenic succession in a subtropical savanna: Conver- sion of grassland to thorn woodland. Ecological Monographs 58(2): 1 1 1-127. Blair, W. F 1950. The biotic provinces of Texas. The Texas Journal of Science 2(1):93-117. Brown, D. E. 1994. Biotic Communities: Southwestern United States and Northwestern Mexico. University of Utah Press, Salt Lake City. Clover, E. U. 1937. Vegetational survey of the Lower Rio Grande Valley of Texas. Madrono 4(2):41-66 and 4(3):77-100. Coffey, G. N. 1909. Clay dunes. The Journal of Geology 17:754-755. CoRRELL, D. S. AND M. C. JoHNSTON. 1970. Manual of the Vascular Plants of Texas. Texas Research Foundation, Renner, Texas. EvERiTT, J. H. AND D. L. Drawe. 1993. Trees, Shrubs and Cacti of South Texas. Texas Tech University Press, Lubbock. FoscuE, E. J. 1932. Physiography of the Lower Rio Gran- de Valley. Pan-American Geologist 57:263-267. Gauch, H. G. 1982. Muldvariate Analysis in Community Ecology. Cambridge University Press, Cambridge. Gould, F W. 1975. The Grasses of Texas. Texas Agri- cultural Experiment Station; Texas A&M University Press, College Station. Hartmann, H. T, D. E. Kester and F T. Davies. 1990. Plant Propagation Principles and Practices. Prentice Hall Career and Technology, Englewood Cliffs, New Jersey. Huffman, G. G. and W. A. Price. 1949. Clay dune for- mation near Corpus Christi, Texas. Journal of Sedi- mentary Petrology 19(3):1 18-127. Jahrsdoerfer, S. E. and D. M. Leslie. 1988. Tamaulipan brushland of the Lower Rio Grande Valley of Texas: Description, human impacts and management op- tions. U.S. Fish and Wildlife Service. Biological Re- port 88(36). U.S. Department of Interior. Johnston, M. C. 1963. Past and present grasslands of southern Texas and northeastern Mexico. Ecology 44(3):456-466. Jongman, R. H. G., C. J. F. TER Braak, and O. F. R. van ToNGEREN. 1987. Data analysis in community and landscape ecology. Pudoc, Wageningen, The Nether- lands. LONARD, R. I., J. H. EVERITT, AND F W. JUDD. 1991. Woody Plants of the Lower Rio Grande Valley, Tex- as. Number 7 Miscellaneous Publications, Texas Me- morial Museum, The University of Texas at Austin. LoNARD, R. I. 1993. Guide to Grasses of the Lower Rio Grande Valley, Texas. The University of Texas-Pan American Press, Edinburg, Texas. Mahall, B. E. AND R. B. Park. 1976. The ecotone be- tween Spartina foliosa Trin and Salicomia virginiana L., in salt marshes of northern San Francisco Bay. II. Soil water and salinity. Journal of Ecology 64:783- 809. McCuNE, B. AND M. J. Mefford. 1997. Multivariate Anal- ysis of Ecological Data. Version 3.05. MjM Software, Gleneden Beach, OR. Muller, C. H. 1947. Vegetation and climate of Coahuila, Mexico. Madrono 9:33-57. Price, W. A. and L. S. Kornicker. 1961. Marine and lagoonal deposits in clay dunes. Gulf Coast, Texas. Journal of Sedimentary Petrology 31(2):245-255. Richardson, A. 1995. Plants of the Rio Grande Delta. University of Texas Press, Austin. Shindle, D. B. and M. E. Tewes. 1998. Woody species composition of habitats used by ocelots (Leopardus pardalis) in the Tamaulipan Biotic Province. South- western Naturahst 43(2):273-279. Shreve, F. 1917. A map of the vegetation of the United States. Geogr. Rev. 3:119-125. Tewes, M. E. and D. D. Everitt. 1982. Study of the en- dangered ocelot occurring in Texas. Year-end Report, U.S. Fish and Wildlife Service, Albuquerque, NM. U.S. Fish and Wildlife Service. 1997. Lower Rio Grande Valley and Santa Ana National Wildlife Refuges. Fi- nal interim comprehensive management plan and draft environmental assessment. U.S. Department of the Interior. Vines, R. A. 1974. Trees, shrubs and woody vines of the Southwest. University of Texas Press, Austin. .7 Madrono, Vol. 47, No. 1, pp. 21-28, 2000 CHROMOSOME COUNTS AND TAXONOMIC NOTES ON DRABA (BRASSICACEAE) OF THE INTERMOUNTAIN WEST 1: UTAH AND VICINITY Michael D. Windham Utah Museum of Natural History, University of Utah, Salt Lake City, UT 84112-0050 Abstract Of the 350+ species ascribed to Draba, nearly one quarter occur in the Intermountain Region of the western United States. Most of these Draba species have not been examined cytologically. This paper presents a total of 18 chromosome counts for 1 1 different taxa occurring in Utah, Wyoming, and Arizona. The chromosome numbers of D. juniperina, D. kassii, D. maguirei var. maguirei, D. rectifructa, D. sobolifera, D. spectabilis var. spectabilis, and D. subalpina are reported here for the first time. Counts differing from published reports are documented for D. asprella var. stelligera and D. cuneifolia var. cuneifolia. The taxonomic significance of the new chromosome counts is discussed for each species. Counts of n = 11 and n = 13 appear to be the first reports of those numbers in the genus, and they complete the continuous series of aneuploid base numbers extending from 8 to 16. It is suggested that the Intermountain West may be a center of diversity for aneuploid Draba, and that this assemblage of species provides a unique opportunity to study chromosomal evolution and speciation. Species assigned to Draba, considered to be the largest genus in the Brassicaceae (Rollins 1993), occupy a variety of habitats and occur on all con- tinents except Australia and Antarctica. The group achieves its greatest diversity in topographically complex, mountainous regions where the disjunct occurrence of suitable habitats seems to favor iso- lation and speciation (Pay son 1917). A prime ex- ample of this is seen in the Intermountain Region of the western United States, broadly defined here as the territory extending from the continental di- vide to the Pacific Crest (Sierran-Cascade axis). Of the 350+ species attributed to Draba by Rollins (1993), nearly one quarter occur in this region and more than 50% of those are endemic to it. The Intermountain West is terra incognita as far as the cytology of Draba is concerned. Of the 57 taxa confined to this region, only 1 1 have been ex- amined chromosomally. Half of these are known from single counts, and none can be considered ad- equately sampled. By comparison, 37 of the 40 Draba species found in Canada and Alaska have been studied cytologically, thanks in large part to the diligent efforts of G. A. Mulhgan (1966, 1970a, b, 1971a, b, 1972, 1974, 1975, 1976). Mulligan's work on the high-latitude North American species of Draba (summarized in the 1976 paper) led to major advances in our taxonom- ic understanding of the genus. In addition to clari- fying species boundaries in several groups, his data provided the basis for the only modem infrageneric classification of North American Draba. Setting aside Draba (Erophila) verna L., a Eurasian intro- duction unrelated to the native species. Mulligan (1976) recognized three informal groups based on a combination of chromosome number, flower col- or, breeding system, and hybridization studies. All 17 of the white-flowered species studied by Mulligan exhibit euploid chromosome numbers based on x = 8. They clearly are related to Eurasian boreal species assigned by Schulz (1927) to the sec- tion Leucodraba DC. Another nine Canadian spe- cies were assigned to his yellow-flowered euploid alliance, which also frequents boreal habitats and has representatives in Eurasia. The remaining 13 Canadian taxa were placed in a yellow-flowered group characterized by aneuploid chromosome numbers of « = 9, 10, 12, 14, 15, and 37 (Mulligan 1976). Apparently restricted to North and South America, this assemblage of species appears more tolerant of the warm/dry conditions that prevail in much of the western United States. Mulligan's (1976) informal classification of North American Draba is a vast improvement over the patently unnatural sections proposed by Schulz (1927). However, it can neither be used nor eval- uated phylogenetically until the chromosome num- bers of local Draba species have been determined. The goals of this study were: 1) to collect crucial chromosome data for Intermountain Draba species, 2) to critically assess current taxonomic treatments for the species sampled, and 3) to develop a set of chromosomally vouchered samples for a DNA analysis (Beilstein and Windham in prep.) designed to test the monophyly of Mulligan's (1976) infor- mal species groups. Materials and Methods Chromosome counts were made from flower buds of wild plants fixed in Farmer's solution (3 22 MADRONO [Vol. 47 parts 95% ethanol: 1 part glacial acetic acid). Fixed materials were stored at -20°C for up to five years and transferred to 70% ethanol immediately before making slides. Buds (or dissected anthers in larger- flowered species) were macerated in a drop of 1% acetocarmine stain, which was mixed 1 : 1 with Hoy- er's solution prior to setting the cover slip and squashing. Slides were examined with an Olympus BH-2 phase contrast microscope, and representative cells were photographed using Kodak Technical Pan 2415 film. A full set of voucher specimens was deposited at the Garrett Herbarium, Utah Museum of Natural History (UT). Duplicate vouchers were deposited at the herbaria listed in Table 1 . To guide the discussion, I produced a compendium of pub- lished chromosome counts for all taxa studied and their putative relatives. This list was assembled by running all accepted names and synonyms from Rollins (1993) and Kartesz (1994) through Chro- mosome Numbers of Flowering Plants (Federov 1974) and a complete set of the Index to Plant Chromosome Numbers spanning the period 1966- 1995 (Omduff 1967, 1968; Moore 1973, 1974, 1977; Goldblatt 1981, 1984, 1985, 1988; Goldblatt and Johnson 1990, 1991, 1994, 1996 & 1998). The primary literature was consulted to verify critical taxonomic and geographic information for each North American count identified by this search. Results My chromosome studies of Utah, Wyoming, and Arizona Draba species yielded a total of 18 counts for 11 different taxa (Table 1). Seven of these taxa have not been counted previously. These include D. juniperina Dom {n = 11), D. kassii Welsh {n = 11), D. maguirei C. L. Hitchc. var. maguirei {n — 16), D. rectifructa C. L. Hitchc. {n = 12), D. so- bolifera Rydb. {n = 13), D. spectabilis Greene var. spectabilis {n — 10), and D. subalpina Goodman & C. L. Hitchc. {n = 13). Counts for two of the re- maining taxa, D. asprella Greene var. stelligera O. E. Schulz {n — 15), and D. cuneifolia Nutt. ex. T. & G. var. cuneifolia (n = 15), differ from numbers previously reported in the literature. Unexpected counts, especially those that disagree with the lit- erature, are documented photographically in Fig- ures 1-6. Determinations of = 11 and « = 13 appear to be the first reports of those numbers in the genus, and they complete the continuous series of aneuploid base numbers extending from 8 to 16. In fact, this small sample of taxa includes every step in that aneuploid series except n = 9 and n = 14. Discussion The plants herein referred to Draba albertina Greene originally were identified as D. stenoloba Ledeb. based on the treatment in A Utah Flora (Welsh 1993). Because D stenoloba has a chro- mosome number of n = 20 (Mulligan 1975), I was surprised when samples from two widely separated Utah populations yielded counts of n = 12 (Fig. 1) and 2n = 24. These determinations agree with pre- vious reports for D. albertina, including four counts from Alberta and one from the Northwest Territo- ries (Mulligan 1975). An additional count of n = 12 from Wyoming originally attributed to D. sten- oloba (Mulligan 1966) was reassigned to D. alber- tina in a subsequent paper by Mulligan (1975). Prior to detailed studies of the group (Mulligan 1975), Draba albertina was treated as a synonym or variety (nana) of D. stenoloba. After discovering that the two taxa had different chromosome num- bers. Mulligan recognized them as separate species based on correlated morphological and geographi- cal differences. The decision to classify these taxa as species also is supported by artificial hybridiza- tion experiments (Mulligan 1975), which indicate that any hybrids formed are completely sterile. According to Mulligan (1975) and Rollins (1993), D. stenoloba, with a chromosome number of = 20 and mostly dendritic trichomes on the upper leaf surfaces, is rarely encountered south of the Canadian border. They assign most collections identified as D. stenoloba from the western United States to D. albertina, characterized by a chromo- some number ofn = 12 and simple or once-forked adaxial leaf trichomes. My morphological studies of Utah specimens concur that typical D. stenoloba is not present in the state, and all collections iden- tified as such represent D. albertina. Both taxa be- long to Mulligan's (1976) yellow-flowered aneu- ploid group. Draba asprella, a species endemic to Arizona and southern Utah, is represented by few herbarium collections and a confusing chromosome literature. A single count of n = ±16 appears in the primary literature and the Indexes to Plant Chromosome Numbers. This count derives from a population in Coconino Co., AZ studied by Rollins and Riiden- berg (1971), which was not identified to variety in the original paper. Rollins (1993) attributes this count to var. asprella and reports an additional, ap- parently undocumented count of n = 16 for var. stelligera. The latter count is critical because it seems to place D. asprella in Mulligan's (1976) yellow-flowered euploid assemblage, whereas my count of n = 15 (Fig. 2) would suggest an affilia- tion with his aneuploid group. I am confident of my determination, which is based on at least 40 cells from eight individuals. At this point, I am inclined to discount the undocumented euploid report and assign D. asprella to the yellow-flowered aneuploid group. In the upcoming field season, I hope to ob- tain accurate counts for all four varieties and de- termine whether var. stelligera is truly polymorphic with regard to chromosome number. The available literature provides two chromo- some counts for Draba cuneifolia. Rollins and Rii- denberg (1971) report a count ofn= 16 from Pecos Co., TX. Although not identified to variety, this col- 2000] WINDHAM: CHROMOSOME COUNTS ON DRAB A 23 Table 1. Chromosome Counts on Draba from Utah and Vicinity. Counts differing from previously published reports are marked by an asterisk. Apparent first counts for a taxon are marked by a double asterisk following the relevant name. Letters before collection numbers identify the following collectors: ER = Eric Rickart; RS = R. Douglas Stone; JT = James Therrien; W = Michael Windham; TW = Theresa Windham; MEW = Maria Windham; MKW = Molly Windham. Herbaria housing voucher specimens are identified by upper case abbreviations (based on Holmgren et al. 1990) following the collection numbers. Draba albertina Greene 2n = 24 UT 12 UT Emery Co. Salt Lake Co. Draba asprella Greene var. stelligera O.E. Schulz n = 15, 2« = 30* AZ Coconino Co. in South Hughes Canyon on the Wasatch Plateau (T14S, R7E, S30); W & ER 95-185 (UT) E of Guardsman Pass along State Route 152 in the Wasatch Mts. (T2S, R3E, S25); W 98-320 (MO, NY, UT) along tributary of Bear Wallow Canyon E of Sedona (T17N, R6E, SIO); W 95-250 (ASU, BRY, COLO, UT, UTC); W, TW & MKW 98-002 (MO, NY, UT) Draba cuneifolia Nutt. ex Torr. & A. Gray var. cuneifolia 15* 15* AZ UT Draba juniperina Dorn** n = n UT 11 11 Draba kassii Welsh*' n = 11 UT WY UT Yavapai Co. Washington Co. Daggett Co. Daggett Co. Sweetwater Co. Tooele Co. Draba maguirei C.L. Hitchc.** var. maguirei 16 Draba nemorosa L. UT Cache Co. var. nemorosa UT Summit Co. Draba rectifructa C.L. Hitchc' 12 UT Draba sobolifera Rydb.** n = n UT Juab Co. Piute Co. Draba spectabilis Greene var. spectabilis** n = 10 UT San Juan Co. 10 UT San Juan Co. Draba subalpina Goodman & C.L. Hitchc.** 13 13 13 UT UT UT Garfield Co. Garfield Co. Iron Co. WNW of Sedona on the SW side of Fay Canyon (T18N, R5E, S30); W, JT & MEW 97-005 (MO, UT) NE of Pinto on low hills overlooking road to Cedar City (T37S, R15W, S26); 99-008 (MO, UT) along Browns Park-Clay Basin road in upper Jesse Ewing Can- yon (T2N, R24E, SI); W 96-152 (MO, NY, UT) along State Route 44 on N side of Spring Creek (T2N, R20E, S\9);W 99-073 (COLO, MO, UT) just E of Richards Gap at S edge of Red Creek Basin (T12N, R105W, S22); W 00-012 (ASU, BRY, MO, UT) in Goshute Canyon on E slope of the Deep Creek Range (TIOS, R18W, S36); W 98-211 (ASU, COLO, MO, NY, UT) SE slope of Mt. Magog in the Bear River Range (T14N, R3E); W95-161 (ARIZ, ASU, BRY, COLO, CPH, DAO, ISTC, MO, NY, OGDF, UC, US, UT, UTC) N base of Windy Ridge on NE slope of the Uinta Mts. (T2N, R19E, S24); W 99-072 (COLO, MO, NY, UT) N of Mount Nebo near head of Gibson Creek (Tl IS, R2E, SI 9); W 96-204 (UT) S side of Bullion Canyon in the Tushar Mts. (T28S, R5W, Sll); W cfe /?5 95-201 (ASU, BRY, COLO, MO, NY, OGDF, UT) SE of Gold Basin in the La Sal Mts. (T27S, R24E, S15); W95- 170 (ASU, BRY, COLO, CPH, MO, NY, OGDF, UT, UTC) W & ER 97-188 (ISTC, UT) NW slope of South Peak in the Abajo Mts. (T34S, R22E); W95-182 (ASU, BRY, COLO, CPH, MO, NY, OGDF, UT, UTC) along tributary of Red Canyon on the Paunsaugunt Plateau (T36S, R4V2W, SI); W & MKW 92-037 (COLO, MO, UT); W 96-036 (DAO) near headwaters of Coyote Hollow on the Paunsaugunt Plateau (T36S, R4V2W, Siy,W 98-129 (MO, NY, UT) NW slope of Blowhard Mtn. on the Markagunt Plateau (T37S, R9W, S15); W 92-135 (BRY, MO, NY, UT, UTC) 24 MADRONO [Vol. 47 2 3 ' P ^^^^ " 4 % 0 0 e 6 Q Figs. 1-6. Meiotic chromosome squashes for various Draba species. Solid spherical bodies in Figs. 1, 4, and 6 = nucleoli. Arrows identify overlapping pairs. 1. Diakinesis in D. albertina (n = 12). 2. Late diakinesis in D. asprella van stelligera (n = 15). 3. Metaphase I in D. cuneifolia var. cuneifolia (n = 15). 4. Diakinesis in D. juniperina (n = 11). 5. Late prophase II in D. maguirei var. maguirei {n = 16 at each pole). Faint spherical body near the center of each cluster = nucleolus. 6. Diakinesis in D. spectabilis var. spectabilis (n = 10). lection is presumed to represent var. cuneifolia based on geographic location. Hartman et al. (1975) also report n = 16 for a collection of the typical variety from Dallas Co., TX. Given this history, I was surprised to obtain clear preparations of = 15 (Fig. 3) for two populations of D. cuneifolia var. cuneifolia from Arizona and Utah. These counts were confirmed in at least five cells from three dif- ferent plants in each population, so it seems likely that the apparent chromosomal polymorphism is real. It is interesting to note that my counts derive from the northwestern portion of the species distri- bution, whereas the two reports of n = 16 represent the southeastern portion of the native range. Further sampling is needed to determine whether chromo- some number truly is correlated with geography in D. cuneifolia. Such an investigation also should en- compass Draba reptans, (Lam.) Fern, which is con- sidered closely related (Hitchcock 1941) or inter- gradient (Welsh 1993) and apparently displays par- 2000] WINDHAM: CHROMOSOME COUNTS ON DRAB A 25 allel variation in chromosome number (Mulligan 1966; Love and Love 1982). Although D. reptans is placed in the aneuploid group by Mulligan (1976), the taxon is white-flowered and probably should be assigned to a separate group (Beilstein personal communication). Draba juniperina is endemic to pinyon-juniper woodlands at the northeastern edge of the Uinta Mountains near "Three Comers", the point where Utah, Wyoming, and Colorado meet. The taxon, long thought to be related to D. oligosperma Hook., because of the shared occurrence of doubly pecti- nate trichomes, has a complex nomenclatural his- tory. It was first separated from the yellow-flowered £). oligosperma under the name D. pectinipila (Rol- lins 1953), a taxon typified on white-flowered spec- imens from alpine habitats in northwestern Wyo- ming. Dom (1978) pointed out that the petals of D. pectinipila truly are white, but the flowers of pop- ulations from southwestern Wyoming and north- eastern Utah are yellow when fresh. Additional morphological features were found to correlate with flower color, geography, and habitat, which led Dom (1978) to describe the Uinta populations as a new species, D. juniperina. Subsequent studies by Lichvar (1983) seemed to reinforce the distinctions among D. oligosperma, D. pectinipila, and D. juniperina but, in his most re- cent work, Rollins (1993) abandoned this taxono- my. Stating that designating "deviant types as in- dependent taxa . . . has done little to clarify the na- ture of the species as a whole" (Rollins 1993), he once again synonymized the segregate taxa under D. oligosperma. Kartesz (1994) followed suit, though Welsh (1986a) maintained juniperina as a variety of D. oligosperma without further comment. There has been little use of this combination, how- ever, because var. juniperina is described as having "petals evidently white" (Welsh 1993), a character state not found in Utah specimens. The chromosome counts presented here for Dra- ba juniperina (Fig. 4) provide valuable insight into the taxonomy of this contentious species complex. Studies at two widely separated localities in Dag- gett Co., UT and one site in Sweetwater Co., WY revealed that D. juniperina is a sexually-reproduc- ing taxon with a chromosome number of n — 11. This is one of two numbers not previously docu- mented in Mulligan's (1966, 1976) aneuploid se- ries, and clearly establishes this taxon as a member of the yellow-flowered aneuploid group. Draba oli- gosperma, on the other hand, is an apomictic taxon (Mulligan and Findlay 1970) with three reported chromosome numbers: 1) 2n = 32 from Alberta (Chinnappa and Chmielewski 1987), 2) 2n = ±60 from Wyonung (Rolhns 1966), and 3) 2n = 64 from seven populations in Alberta and one in Yu- kon Territory (Mulligan 1972). These numbers in- dicate that D. oligosperma belongs to Mulligan's (1976) yellow-flowered euploid group. The difference in chromosome base numbers be- tween D. oligosperma (x = 8) and D. juniperina (x = 11) is not trivial. The former is not a simple polyploid derivative of the latter and, if Mulligan (1976) is right in his assessment of relationships, they may belong to different major lineages. The two taxa are easily distinguished using the charac- ters hsted by Dom (1978) and Lichvar (1983), even where their ranges overlap. Even if they grew to- gether, which they apparently do not, there would be no opportunity for hybridization because D. oli- gosperma is apomictic and apparently does not pro- duce functional gametes (Mulligan and Findlay 1970). All of this provides a strong argument for maintaining Draba juniperina as a distinct species. Draba kassii is a very rare species endemic to a few canyons in the Deep Creek Mountains of west- em Utah. Its relationships are obscure, with Rollins (1993) stating that it "is not closely enough related to any known species of Draba to allow inferences as to its phylogeny." Comparisons have been drawn to D. asprella (Welsh 1986b) and D. stan- dleyi J. F Macbr. & Payson (Rollins 1993), though both authors suggest that the similarities may be superficial. Chromosome numbers have the poten- tial to play a cmcial role in determining the rela- tionships of this species. At least 20 cells from five different plants clearly establish that the chromo- some number of D. kassii is n = 11 (Table 1). This number, which establishes the taxon as a member of Mulligan's yellow-flowered aneuploid group, would seem to mle out a direct phylogenetic link to D. asprella (n = 15, 16?). The possibility of a relationship to D. juniperina, the only other species known to have n — 11, is intriguing. However, the two taxa do not appear closely related morpholog- ically, and any hypothesis of relationships will re- main speculative until additional Draba species (in- cluding D. standleyi) have been sampled chromo- somally. The phylogenetic affinities of Draba maguirei are as contentious as those of D. kassii, and it ap- pears that no recent author has ventured to discuss its possible relationships. On first describing the species, Hitchcock (1941) stated that it "is very striking and quite unlike any of the other Drabas from its immediate vicinity. Its closest relatives are probably those of the ventosa group . . .". Of the eight taxa comprising Hitchcock's ventosa group, chromosome counts have been published for two {D. ventosa Gray and D. ruaxes Payson & St. John) and a third {D. sobolifera) is reported here. All three belong to Mulligan's (1976) yellow-flowered aneuploid group with chromosome numbers based on X = 12 and 13. Thus, it is surprising to find that D. maguirei var. maguirei shows a euploid count of n = 16 (Fig. 5). Additional chromosome counts on D. maguirei and other members of Hitchcock's ventosa group are needed to resolve this apparent conflict. Draba nemorosa L., a species of widespread oc- currence in both North America and Eurasia, was 26 MADRONO [Vol. 47 assigned by Mulligan (1976) to his yellow-flowered euploid group. All populations analyzed chromo- somally have shown ^ = 8, regardless of geograph- ic origin. In North America, there have been four counts from Alberta (Packer 1964; Mulligan 1966, 1975), one from Manitoba (Love and Love 1982), two from Ontario (Mulligan 1975), and two from Saskatchewan (Mulligan 1966, 1975). It appears that my determination ofn = S (Table 1) from Dag- gett Co., UT is the first report for the United States. None of the previous North American reports spec- ify variety, though most are surely var. nemorosa, the taxon to which my count is assigned following the taxonomy of Kartesz (1994). Although the gla- brous-fruited form (var. leiocarpa) is considered taxonomically insignificant by many authors, there does appear to be some geographic integrity to its occurrence. Therefore, it seems wise to maintain the distinction until North American populations are studied adequately. Although Hitchcock (1941) considers Draba rec- tifructa to be a close relative of euploid D. nemo- rosa, little evidence is cited to support such an as- sociation. Instead, it appears to be very closely re- lated to D. albertina, distinguished from that spe- cies mainly by its pubescent upper stems and pedicels. Ongoing studies of populations in north- em Utah suggest that D. rectifructa and D. alber- tina hybridize when growing in close proximity. Thus, it is not surprising to find that D. rectifructa has a chromosome number of n = 12 (Table 1), identical to that of its putative aneuploid relative. Draba sobolifera, endemic to the Tushar Moun- tains of southern Utah, is considered a member of Hitchcock's (1941) ventosa group with close affin- ities to D. cusickii Robinson & D. E. Schulz (Rol- lins 1993). The chromosome number of the latter species is unknown, but the two members of the ventosa complex previously reported, D. ventosa and D. ruaxes, show 2n = 36 and 2n = 72 (Mul- ligan 1971). They are considered to be triploid and hexaploid respectively, with a base number of x = 12. Cytological studies on a population of D. so- bolifera from Piute Co., UT reveal that it is a sex- ually-reproducing taxon with a chromosome num- ber of n = 13. This is the last number to be doc- umented in Mulligan's (1966, 1976) aneuploid se- ries, and it firmly establishes this species as a member of the yellow-flowered aneuploid group. There are three previous chromosome counts for Draba spectabilis, all from Colorado and all as- signed to var. oxyloba (Greene) Gilg. & O. E. Schulz by Price (1980). The earhest report (Mulh- gan 1966) of n = 10 seemed to indicate that the species belonged in the aneuploid group. However, two subsequent counts of n = 16 and n = 16 ± 2 by Price (1980) suggest an affinity to the yellow- flowered euploid assemblage. My determinations, apparently the first for var. spectabilis, are from two widely separated populations in San Juan Co., UT (Fig. 6). They agree with Mulligan's (1966) report of n = 10 and point out the need for further sam- pling to determine the relationships and proper tax- onomy of D. spectabilis. Draba subalpina generally is restricted to a sin- gle geologic stratum, the Claron Formation of Bryce Canyon National Park and vicinity. Although recent authors have said little regarding its probable relationships, Hitchcock (1941) states that its clos- est relative is D. oreibata J. F. Macbr. & Payson, a species under which it was subsumed prior to 1932. The latter taxon is endemic to central Idaho in its typical form, is similarly white-flowered, and shows a chromosome number of n = 16 (Hender- son et al. 1980). In light of its proposed relation- ships and the assumption that D. subalpina was a member of the white-flowered euploid group, the actual chromosome number was unexpected. Based on at least ten cells from five individuals in each of three populations, the chromosome count of Draba subalpina proves to be « = 13 (Table 1). Whether D. subalpina belongs to a relatively rare, white-flowered aneuploid group or is more closely related to some of its yellow-flowered congeners remains to be deternuned. The close proximity (ca. 60 km) of D. sobolifera, the only other species known to exhibit « = 13, raises intriguing possi- bilities regarding the relationships of white- and yellow-flowered aneuploids in Draba. Even with the small sample size of this nascent effort, it is clear that the taxonomic composition of the Intermountain Draba flora is quite different from the intensively studied assemblage of Canada and Alaska. In the latter. Mulligan (1976) assigned 17 species to his white-flowered euploid group, nine to the yellow-flowered euploid assemblage, and 13 to his yellow-flowered aneuploid group. In my sample from Utah, Wyoming, and Arizona, white-flowered euploids are not represented (unless D. cuneifolia belongs here) and yellow-flowered euploids are rare, comprising only D. nemorosa and possibly D. maguirei. Seven of the Intermountain taxa belong to the yellow-flowered aneuploid as- semblage and the remaining taxon {D. subalpina) is a white-flowered aneuploid of uncertain affinity. A growing number of chromosome counts for the region suggests that the Intermountain West may be a center of diversity for aneuploid Draba. With the discovery of both n = W and /i = 13 among local endenaics, a complete series of base numbers extending from 8 to 16 has been docu- mented. Only n = 9 and « = 14 are missing from my sample, and those numbers have been con- firmed in other taxa from the region. This means that every major step in the process of aneuploid evolution is preserved among the Draba species of the Intermountain West. In this assemblage of Dra- ba species, we have an unprecedented opportunity to study the processes of chromosomal evolution and speciation in plants. With further cytological sampling and concurrent DNA studies of the group, we soon may be in a position to elucidate the evo- 2000] WINDHAM: CHROMOSOME COUNTS ON DRABA 11 lutionary history of this interesting and diverse set of organisms. Acknowledgments I wish to express my gratitude to the Utah Museum of Natural History (UMNH) and the staff of Wasatch-Cache National Forest (especially Wayne Padgett) for providing funding that made this work possible. The job of locating the rarer taxa and estimating the dates of meiosis was greatly facilitated by access to the extensive collections of the S. L. Welsh Herbarium (Brigham Young University). Special thanks are due R. Douglas Stone, who piqued my interest in Draba and whose help and dedication to the project (and ability to find buds undergoing meiosis) was a major factor in its success. I also thank Erin Kincaid (Exhibits Department, UMNH) for help with the figures, and Dr. Christopher Haufler (University of Kansas) for helpful comments on an earlier version of this manuscript. Literature Cited Chinnappa, C. C. and J. G. Chmielewski. 1987. Docu- mented chromosome numbers 1987: 1. Miscellaneous counts from western North America. Sida 12:409- 417. Dorn, R. D. 1978. A new species of Draba (Cruciferae) from Wyoming and Utah. Madrono 25:101-103. Federov, a. (ed.). 1974. Chromosome Numbers of Flow- ering Plants. Otto Koeltz, Koenigstein. GoLDBLATT, P. (ed.). 1981. Index to plant chromosome numbers 1975-1978. Monographs in Systematic Bot- any from the Missouri Botanical Garden 5:1-553. . (ed.). 1984. Index to plant chromosome numbers 1979-1981. Monographs in Systematic Botany from the Missouri Botanical Garden 8:1-427. . (ed.). 1985. Index to plant chromosome numbers 1982-1983. Monographs in Systematic Botany from the Missouri Botanical Garden 13:1-224. . (ed.). 1988. Index to plant chromosome numbers 1984-1985. Monographs in Systematic Botany from the Missouri Botanical Garden 23:1-264. AND D. E. Johnson, (eds.). 1990. Index to plant chromosome numbers 1986-1987. Monographs in Systematic Botany from the Missouri Botanical Gar- den 30:1-243. AND . (eds.). 1991. Index to plant chro- mosome numbers 1988-1989. Monographs in Sys- tematic Botany from the Missouri Botanical Garden 40:1-238. AND . (eds.). 1994. Index to plant chro- mosome numbers 1990-1991. Monographs in Sys- tematic Botany from the Missouri Botanical Garden 51:1-267. AND . (eds.). 1996. Index to plant chro- mosome numbers 1992-1993. Monographs in Sys- tematic Botany from the Missouri Botanical Garden 58:1-276. AND . (eds.). 1998. Index to plant chro- mosome numbers 1994-1995. Monographs in Sys- tematic Botany from the Missouri Botanical Garden 69:1-208. Hartman, R. L., J. D. Bacon, and C. F. Bohnstedt. 1975. Biosystematics of Draba cuneifolia and D. platycar- pa (Cruciferae) with emphasis on volatile and flavo- noid constituents. Brittonia 27:317-327. Henderson, D., A. Cholewa, and N. Reese. 1980. in A. Love (ed.), lOPB Chromosome Number Reports LXVIII. Taxon 29:534. Holmgren, P. K., N. H. Holmgren and L. C. Barnett. (eds.). 1990. Index Herbariorum. Part I: The Herbaria of the World, 8th ed. New York Botanical Garden, Bronx, NY. Kartesz, J. T. 1994. A Synonymized Checklist of the Vas- cular Flora of the United States, Canada, and Green- land, 2nd ed. Timber Press, Portland, OR. LiCHVAR, R. W. 1983. Evaluation of Draba oligosperma, D. pectinipila, and D. juniperina complex (Crucifer- ae). Great Basin Naturalist 43:441-444. Love, A. and D. Love. 1982. Pp. 120-126 in A. Love (ed.). lOPB Chromosome Number Reports LXXIV. Taxon 31:119-128. Moore, R. J. (ed.). 1973. Index to plant chromosome numbers for 1967-1971. Regnum Vegetabile 90:1- 539. . (ed.). 1974. Index to plant chromosome numbers for 1972. Regnum Vegetabile 91:1-108. . (ed.). 1977. Index to plant chromosome numbers for 1973/74. Regnum Vegetabile 96:1-257. Mulligan, G. A. 1966. Chromosome numbers of the fam- ily Cruciferae. III. Canadian Journal of Botany 44: 309-319. . 1970a. Cytotaxonomic studies of Draba glabella and its close allies in Canada and Alaska. Canadian Journal of Botany 48:1431-1437. . 1970b. A new species of Draba in the Kanan- askis Range of southwestern Alberta. Canadian Jour- nal of Botany 48:1897-1898. . 1971a. Cytotaxonomic studies of the closely al- lied Draba cana, D. cinerea, and D. groenlandica in Canada and Alaska. Canadian Journal of Botany 49: 89-93. . 1971b. Cytotaxonomic studies of Draba species in Canada and Alaska: D. ventosa, D. ruaxes, and D. paysonii. Canadian Journal of Botany 49:1455-1460. . 1972. Cytotaxonomic studies of Draba species in Canada and Alaska: D. oligosperma and D. incerta. Canadian Journal of Botany 50:1763-1767. . 1974. Cytotaxonomic studies of Draba nivalis and its close allies in Canada and Alaska. Canadian Journal of Botany 52:1793-1801. . 1975. Draba crassifolia, D. albertina, D. nemo- rosa, and D. stenoloba in Canada and Alaska. Ca- nadian Journal of Botany 53:745-751. . 1976. The genus Draba in Canada and Alaska: key and summary. Canadian Journal of Botany 54: 1386-1393. and J. N. Findlay. 1970. Sexual reproduction and agamospermy in the genus Draba. Canadian Journal of Botany 48:269-270. Ornduff, R. (ed.). 1967. Index to plant chromosome num- bers for 1965. Regnum Vegetabile 50:1-128. . (ed.). 1968. Index to plant chromosome numbers for 1966. Regnum Vegetabile 55:1-126. Packer, J. G. 1964. Chromosome numbers and taxonomic notes on western Canadian and Arctic plants. Cana- dian Journal of Botany 42:473-494. Payson, E. B. 1917. The perennial scapose Drabas of North America. American Journal of Botany 4:253-267. Price, R. A. 1980. Draba streptobrachia (Brassicaceae), a new species from Colorado. Brittonia 32:160-169. Rollins, R. C. 1953. Draba on Clay Butte, Wyoming. Rhodora 55:229-235. . 1966. Chromosome numbers of Cruciferae. Con- tributions of the Gray Herbarium 197:43-65. 28 MADRONO [Vol. 47 . 1993. The Cruciferae of Continental North Amer- ica: Systematics of the Mustard Family from the Arctic to Panama. Stanford University Press, Stanford, CA. AND L. RuDENBERG. 1971. Chromosome numbers of Cruciferae. II. Contributions of the Gray Herbari- um 201:117-133. ScHULZ, O. E. 1927. Cruciferae -Dra^a et Erophila. in: A. Engler (ed.), das Pflanzenreich IV. 105(Heft 89): 1-395. Welsh, S. L. 1986a. New taxa and combinations in the Utah Flora. Great Basin Naturalist 46:254-260. . 1986b. New taxa in miscellaneous families from Utah. Great Basin Naturalist 46:261-264. . 1993. Cruciferae. Pp. 275-325 in S. L. Welsh, N. D. Atwood, S. Goodrich & L. C. Higgins (eds.), A Utah Flora, 2nd ed. Jones Endowment Fund, Brigham Young University, Provo, UT. Madrono, Vol. 47, No. 1, pp. 29-42, 2000 A FLORISTIC STUDY OF TSEGI CANYON, ARIZONA Susan Holiday Northern Arizona University, Flagstaff, AZ 86011 Abstract The purpose of this study is to list the vascular flora present in Tsegi Canyon, Arizona, and to describe any change in flora that may have happened during the past hundred years. Plants were collected during the years 1994-1997. Three hundred and ten species representing seventy three families are reported to occur within the Tsegi drainage. Three percent of these species are endemic to the Colorado Plateau and twelve percent of the species found are non-native. A change in floristic composition is found to have occurred in the last century, correlated with a shift in habitat types in the canyon. The canyon of the late 1800's had a slowly moving stream and marshes along a continuous alluvium. The present canyon has a faster moving stream that has eroded much of the alluvium to bedrock. One species, Cymopterus beckii is a new report for Arizona and also is listed as a candidate for rare and endangered species status. Floristics on the Colorado Plateau have not been widely studied, with perhaps the exception of the Grand Canyon (Phillips et al. 1987). The reason for this may be the remoteness of the area from civi- lization and difficulty in traveling to many parts of the plateau due to lack of roads and extremes in temperature. Although most of the plateau vegeta- tion consists of pinion and juniper in higher ele- vations and desert scrub in the lower, there are also canyons which harbor a very different flora, in- cluding some relict populations, as well as seeps and alcoves with their unique flora. The purpose of this paper is to describe one of the canyons on the plateau that includes year-round, running water as a small stream, as well as seeps and alcoves. Although this is the first compilation of flora done of Tsegi Canyon, the flora has been studied and described by various researchers previously. One of the first descriptions of the flora of the can- yon was done by Clute (1920). J. T. Hack (1945) also studied the canyon as part of his documenta- tion of the erosion/deposition cycles in northern Ar- izona. He, along with Dean (1969) and Weatherill (1953) describe the erosion of this area that oc- curred within the last hundred years, which could possibly be related to the environmental changes that occurred in the southwest during the first part of the century (Hastings & Turner 1965). An in- ventory was done for the park service in the 1970's, with Brotherson et al. (1978) publishing a flora of Navajo National Monument. Historical collections were examined at the Walter B. McDougall Her- barium, the Deaver Herbarium, the U. of A. Her- barium, and at the herbarium at Navajo National Monument. Study area description. The Tsegi Canyon drain- age system is a complex of canyons that forms the headwaters of Laguna Creek, one water supply for the town of Kayenta, AZ. Tsegi Canyon, whose name in the Navajo language means 'in the rocks' or canyon, is in Navajo County in the northeast comer of Arizona, 540000E, 4054000N, 36°40'N, 110°30'W. Tsegi Canyon includes in its boundaries two of the three sections of Navajo National Mon- ument, Betatakin and Keet Seel. The head of Tsegi Canyon is located on the Or- gan Rock Monocline. This is an uplift that is fol- lowed by Highway 160, in Long Valley between the Shonto Plateau and Black Mesa. The canyon is a complex drainage cut into the Shonto Plateau on the east and Skeleton Mesa to the west. Six differ- ent geological formations are visible in the canyon. The top of the plateau is made of Navajo Sand- stone, the formation responsible for the magnificent sandstone cliffs at the top of the canyon. This layer is of Jurassic age and made of wind-blown sand and dunes. The layer just under the Navajo Sand- stone is the Kayenta formation consisting of gray- red sandstone and some clay shale. This layer is about 61 meters thick at the head of the canyon. Because the Navajo sandstone is porous, it allows percolation of water onto the top of the less porous Kayenta formation. The water moves laterally over the Kayenta formation to flow out and form seeps on the canyon walls. Exfoliation of the sandstone above the seeps causes the formation of the alcoves that were utilized by the Anasazi. The layers under the Kayenta formation are a part of the Glen Can- yon group. In the canyon, it is represented by the Lukachukai member of the Wingate sandstone. This is a reddish-brown, cliff-forming sandstone of- ten responsible for the rockfall in the canyon. Un- der this is the Chinle formation. It is represented by two strata, the Churchrock member and the Owlrock member. The Churchrock member con- sists of brownish-red siltstone, mudstone, and fine- grained sandstone with small, white spots and streaks. Below is the Owlrock member which con- sists of reddish-brown siltstone and mudstone, and greenish-gray claystone laid down in the Triassic Age. There is also a limited exposure located north of the Tsegi Hotel, of Petrified Forest member with red, purple, and green/gray betonite claystone. On 30 MADRONO [Vol. 47 the bottom of the canyon is alluvial fill (Beaumont & Dixon 1964). There are three layers of alluvial deposition in Tsegi Canyon. The oldest layer is the Jeddito for- mation that was laid down before 3500 BC. This is overlaid by the Tsegi formation that was laid down between 3500 BC and AD 1300. The youngest lay- er of alluvium is the Naha formation that was laid down between 1450 and 1880 (Hack 1945). Al- though there is a small amount of post- 1900 allu- vium, at the present there is more erosion happen- ing than deposition. The reasons for the deposition and erosion cycles is not definitely known. It has been suggested that rainfall, climate fluctuations, and land management practices may all be contrib- utors to this cycle, with perhaps, the climate being the most influential factor (Clay-Poole 1989). How- ever there is also evidence that human activity has affected arroyo cutting in the canyon. The cutting of the Tsegi-Naha arroyo in Keet Seel was preceded by the clearing of an aspen forest at the bottom of the canyon in the 1200's. After the abandonment of Keep Seel, the area was redeposited with alluvium and recolonized by Quercus gambelii (Dean 1969). However, with the new arroyo cutting of the present century, the distribution of oaks have retreated to the upper side alluvium. The climate at Tsegi Canyon is arid with cold winters and hot sununers. The daily average tem- perature at Tsegi is Celsius. Temperatures vary from highs of 340 to 380° C in July to lows of -230 to 130° C in the winter. The frost-free season averages about 155 days. Precipitation in the can- yon is variable from year to year. Over a 17-year span, the rainfall at the Betatakin Monument ranged from a low to 17.3 centimeters to a high of 47.7 centimeters (U.S.D.C. 1979-1996). The variability is caused by differences in winter precipitation and is also enhanced by the fact that monsoon rains are very spotty and usually do not equally wet all parts of the canyon (Dean 1969). Methods Seventeen collecting trips to Tsegi Canyon were made between 1994 and 1997. The main focus of collecting was to include as many species as pos- sible for the floristic list. The collections were done between the months of April and October, as most of the vascular flora is dormant during the winter. Lower Tsegi Canyon was visited April 23, 1994; May 30, 1994; July 16, 1994; and August 12, 1994. Wildcat Canyon and Lone Cottonwood Canyon were visited June 19, 1995, July 26, 1995, and Au- gust 11, 1995. Upper Tsegi Canyon, Fir and Beta- takin Canyon were visited June 3, 1994; August 7, 1994; September 16, 1994; May 21, 1995; May 29, 1996; and August 8, 1996. Dowozhiebito and Keet Seel Canyons were visited June 18, 1994; May 5, 1996; September 28, 1996; and June 26, 1997. Four of the trips. May 29, 1996; August 8, 1996; Sep- tember 28, 1996; and June 26, 1997 were made at the request of the National Park Service as a survey for rare and endangered species. At this time, col- lections were made at Betatakin National Monu- ment and Keet Seel National Monument. However, the majority of the collecting was done on Navajo Tribal land. All specimens were pressed, dried, and stored at the Deaver Herbarium (ASC) at Northern Arizona University, with duplicates sent to the Navajo Trib- al Heritage Herbarium. Specimens were named fol- lowing Kartesz (1994). Previous collections located at the Deaver Herbarium (ASC), the museum of Northern Arizona Herbarium (MNA), and the Uni- versity of Arizona Herbarium (ARIZ) were used for comparison. A classification model, modified from Rowlands' Colorado Plateau Vegetation Assessment and Clas- sification Manual was used to describe the various vegetation assemblages in the canyon. I determined dominant/co-dominants by using the largest sized plants that appeared to be the most abundant (Bon- ham 1989; Rowlands 1994). The other notable spe- cies are included to help describe the assemblages. Boundaries were determined using physical bound- aries, such as terrace levels and natural altitudinal separations. The assemblages were mapped out us- ing USGS topographic maps of the Betatakin, Keet Seel, and Marsh Pass quadrangles with additional information relating to location and size taken from aerial photos. The maps were drawn by hand and the areas of the vegetation assemblages determined using a Mackintosh scanner and NIH Image area analysis. Results There were 310 species found in Tsegi Canyon during the study years. In comparison, 518 species were found in Canyon deChelly National Monu- ment (Halse 1973; Harlan 1976), 293 species were found at Navajo National Monument (Brotherson 1978), 376 species were found in Volunteer and Sycamore Canyon (ShiUing 1980), and 326 species were found in the Walnut Canyon National Monu- ment (Amberger 1947; Spangle 1953; Joyce 1976). There are eight vegetation assemblages described in this study for Tsegi Canyon which include the Pseudotsuga assemblage, Populus tremuliodes as- semblage, the Pinus edulislJuniperus osteosperma assemblage, Quercus gambelii assemblage, Atri- plexl Artemisia assemblage, the Juncus marshland assemblage, Betula occidentalis assemblage, the Gutierrezia assemblage, and the Pucinnellia bad- lands assemblage. This classification is split into more detail than the USFS digitized classification system (Brown 1980) and some of the assemblages are combined compared to Rowlands (1994). Be- cause there is no standardized way to classify the complex systems of a riparian canyon on the Col- orado Plateau, this scheme was created based on both classification systems. 2000] HOLIDAY: A FLORISTIC STUDY OF TSEGI CANYON, ARIZONA 31 Pseudotsuga occurred in shaded, mostly north and west facing areas in the side canyons of Tsegi. This assemblage consists of about 8% of the total cov- erage of vegetation sampled in the canyon. Most Pseudotsuga menziesii individuals encountered were older trees. A study of whether or not seedlings are present in high enough numbers to replace older trees would be of value for this area. The Pseudo- tsuga often graded down into populations of Populus tremuliodes in the moist side canyons, and some- times an individual fir could be found in the upper regions of a stream bed. Rarely, Pinus ponderosa could be found growing among the firs. Li the upper Keet Seel canyon, few individuals of Abies concolor were found among the fir. Shrubs growing in this assemblage include Ssymphoricarpos oreophilius, Ribes cereum, Ribes inerme, and Amelanchier alni- folia. Herbaceous plants include Antennaria parvi- folia, Mahonia repens, Corydalis aurea, Galium aparine, and Valeriana acutiloba. This assemblage is included in the Cold Temperate Forest and Wood- lands, Rocky Mountain Montane Conifer Forest, Douglas fir- White Fir series, 122.311. Pseudotsuga menziesii Association in the Digitized Systematic Classification used by the Forest Service (Brown 1980). In Rowlands (1994), this is part of the Mon- tane Zone, Forest and Woodland Formation, and Pseudotsuga menziesii Series. Populus tremuloides creek bottoms include plants adapted to a shady, moist environment. Pop- ulus tremuloides covers less than 1% of the area studied, occurring in the upper areas of the Beta- takin-Fir Canyon side canyons. This assemblage is bounded above mostly by Pseudotsuga menziesii on the shady sides of the canyons and Quercus gambelii on the more sunny sides. Betula occiden- talis can be found further down the creek bed if there is running water, otherwise it usually ends abruptly in PinuslJuniperus or sandy creek bottom vegetation. Other trees that can be found in this assemblage include Pseudotsuga menziesii and Prunus virginiana. Shrubs present include Rhus aromatica, Symphoricarpos oreophilus, Cornus sericea, Arctostaphylos pungens, Ribes leptanthum, Rosa woodsii, Salix exigua, and Salix lasiolepis. Herbaceous plants include Equisetum arvense, Car- ex athostachya, Eleocharis palustris, Juncus arcti- cus, Fritillaria atropurpurea, Smilacina stellata, Poa praatensis, Erigeron speciosus, Silene menzie- sii, Lathyrus brachycalyz, Androsace septentrion- alis, Clematis ligusticfolia, Thalictrum fendleri, Heuchera parvifolia, and Mimulus rubellus. The closest classification of this assemblage found in the USFS classification is the Great Basin Interior Strand, which is very non-specific (Brown 1980). Rowlands (1994) has a Populus tremuloides Series in his Forest and Woodlands Formation. However, I do not feel that this quite fits as it describes aspen in a pine forest where the pines will eventually suc- ceed the aspen. Here the aspen are the climax spe- cies, with aspen saplings replacing older trees. The most common assemblage found on the Shonto Plateau is the Pinus edulislJuniperus osteos- perma. This assemblage forms 50% of the plant conmiunities mapped. In the canyon, this assem- blage is found in many of the upper ledges and south facing canyon sides. This assemblage surrounds the Monument headquarters and is common on the top of Skeleton Mesa. Shrubs that can be found in as- sociation with this assemblage include Ephedra vir- idis, Chrysothamnus nauseosus, Shepherdia rotun- difolia, Fendleera rupicola, Amelanchier utahensis, Cercocarpus intricatus, Cerccocarpus montanus, Holodiscus dumosus, and Yucca angustissima. Her- baceous plants include Allium macropetalum, Calo- chortus aureus, Bouteloua gracilis, Cymopterus acaulis, Artemisia dracunculus, Psilostrophe spar- siflora, Heterotheca villosa, Arabis perennans, Streptanthus cordatus, Echinocereus triglochidiatus, Opuntia polycantha. Astragalus ceramicus, Mirabi- lis multiflora, Ipomopsis aggregata, Castilleja lina- riifolia, Cordylanthus wrightii, and the parasitic Phoradendron juniperinum. The Pinus edulislJunip- erus osteospermus assemblage is part of the Forest and Woodland Formation (Rowlands 1994) and clas- sified by the USFS as part of the Great Basin Conifer Woodland, Pinion-Juniper Series (Brown 1980). The area described as the Quercus gambelii as- semblage occurs on the upper terraces in side can- yons draining from west to east, though there are exceptions in the upper areas included in this study. This assemblage is estimated to cover about 2% of the study area. The main component of this assem- blage are thickets of Quercus gambelii with occa- sional larger trees of Quercus included. This area's shading consists of thick leaf litter and generally is not as diverse as the more expansive PinuslJunip- erus assemblage. Many plants found here grow in spaces in the thicket where more light can pene- trate. Other shrubs included in this assemblage are Ribes cereum and Prunus virginiana. The herba- ceous cover includes Juncus arcticus, Smilacinia stellata, Bouteloua curtipendula, Opuntia phaea- cantha, Erigeron utahensis, and Lathyrus brachy- calyx. Rowlands (1994) classifies this as the Tall Shrubland Formation, Quercus gambilii Series. The USFS Digitized Classification (Brown 1980) in- cludes this in the Cold Temperate Scrublands, Great Basin Montane Scrub, Oak-Scrub Series. The Atriplexl Artemisia assemblage dominate on the lower terraces above the main creek and at the mouths of the side canyons. This is the second larg- est vegetation assemblage covering about 22% of the total canyon. Although there may be an occa- sional Pinion or Juniper tree associated with this assemblage, the dominant larger plants are the shrubs Atriplex canescens and Artemisia tridentata. There are also occasional lone Elaegnus angusti- folia individuals of unknown origin, possibly plant- ed by members of the family that use the canyon (Melberg 1988), or distributed by birds. There is also a small stand of about four Ulmus pumila at 32 MADRONO [Vol. 47 one site, planted along the side of the dirt road. Shrubs associated with this area include chryso- thamnus viscidiflorus, Sarcobatus vermiculaatus, Poliomintha incana, and Cercocarpus intricatus. Herbaceous plants include Elymus smithii, Artemis- ia frigida, Helianthus peetiolaris, Senecio multilo- batus, Cryptantha crassisepala, Descurainia pin- nata, Salsola iberica, Astragalus amphioxys, Pha- celia ivesiana, Spaeralcea parvifolia, Mirabilis ox- ybaphoides, Gayophytum racemosum, Orobanche multiflora, Erigonum cemuum, Ranunculus testi- culatus, and Verbeena bracteata. Rowlands (1994) split these two shrub species into two separate se- ries in his classification scheme. However, in Tsegi Canyon, the two species were found in many places together, and thus would be hard to separate into separate assemblages. The closest classification found in the USPS classification manual (Brown 1980) is defined as the Cold Temperate Desertlands, Great Basin Desertscrub, Mixed Scrub Series. The Juncus marshland assemblages were usually found at the bottoms of most small side creek drainages and flattened areas below seeps. The Jun- cus assemblage is a minor component of the can- yon, consisting of less than 1% of the canyon sur- veyed. These areas were often heavily used by cat- tle and damaged by trampling. This assemblage had only occasional Elaeagnus angustifolia or Tamarix ramosissima. There was also one example of a Populus fremonti tree at the edge of one marshland. Herbs found in this assemblage include Juncus arc- ticus, Equisetum arvense, Polypogon monspelien- sis, Scirpus pungens, Aster frondosus, Conyza can- adensis, Taraxacum officinale, Crypthantha inae- quata, Lepidium virginicum, Epilobium ciliatum, Plantago major, and Ranunculus cymbalaria. Row- lands (1994) classifies this as a Marshland Forma- tion, Juncus arcticus Series. Using the USPS clas- sification this assemblage would be included in the Cold Temperate Marshlands, Great Basin Interior Marshland, Rush Series (Brown 1980). Betula occidentalis creek bottom assemblage was found only in the areas of Betatakin-Pir Canyon drainages and some north-draining side canyons of Keet Seel and upper Dowozhiebito Canyon. This assemblage accounted for about 1% of the area mapped. In the Betatakin area it was bounded above by Populus tremuoides and below by Elaeagnus angustifolia. In other side-canyons, Bet- ula was the uppermost tree species on the drainage floor. The Betula commonly grew along both sides of drainages that included year-round running wa- ter. Another tree species associated with this assem- blage was Acer negundo. Salix monticola and S. lasiolepis were also found growing among the birches. Herbaceous plants found in this assem- blage include Corallorhiza maculata. Toxicoden- dron rydbergii, and Chenopodium album. Like the Quercus assemblage, the density of the trees tend to shade the floor of the creek. This, along with occasional high water levels, tend to limit the num- ber of herbs present. Rowlands (1994) classifies this in the Montane Zone, Tall Shrubland Porma- tion, Betula occidentalis series. However, this would be more applicable if it was placed in a ri- ; parian formation, which is not included in this clas- sification scheme. In the USPS (Brown 1980) there I is a Cold Temperate, Great Basin Interior Strand, i which includes all riparian vegetation in the Great ! Basin Biome. | The bottoms of the major drainages are classified i as the Gutierrezia stream bottom assemblage. This | assemblage covers about 18% of the canyon. The j soils in this area are characterized by sandy depos- its that are typically scoured at least once a year by flooding. There is also quicksand after floods and other areas that are devoid of vegetation because of animal or human (automobile) use. This assem- blage also includes areas of bare sandstone and low, dry, sandy dunes. Shrubs that grow here include Gutierrezia sarothrae, Chrysothamnus depressus, and Artemisia frigida. Herb species that grow here include Equisetum hyemale, Chenopodium lepto- phyllum, Salsola iberica. Astragalus amphioxys, Nama retrorsum, Tripterocalyx micranthus, Oeno- thera pallida, and Verbascum thapsus. Rowlands (1994) has a Gutierrezia sarothrae series in his classification under the Low Shrubland Pormation. The USPS (Brown 1980) includes this area under the Great Basin Interior Strand. Downstream the Betula individuals in the Beta- takin drainage is a stand of Elaeagnus angustifolia. Here, E. angustifolia is growing along both sides of the creek. Whether E. angustifolia is displacing the birch or growing in an area that is for some reason too low in the drainage for the birch is un- known. However, since E. angustifolia is an intro- duced species, and there is an area of interface be- tween the two species, and because the Elaeagnus seems to be spreading in the canyon (Melberg, 1994), the former possibility needs to be examined. Elsewhere in the canyon, E. angustifolia is mostly present as individual trees or young plants. Young plants were found growing in the Keet Seel drain- age, about seven miles from Betatakin, and other young plants were found in side canyons across and above the Betatakin drainage. This is one species that needs to be watched closely because it seems to be able to colonize some local areas and possibly out-compete native growth. The Puccinellia assemblage consists of few plant species growing on betonite clays. This assemblage covers less than 1% of the area mapped. The largest example of this assemblage is in upper Wildcat Canyon on the north-west facing side of the can- yon. Most of the plants are concentrated near a small seep. The area grades into the PinuslJunip- erus assemblage above it on the canyon sides. Plants included here are Apocynum cannabinum and Puccinellia distans. The closest classification in Rowlands (1994) would probably be the Sub- montane Barren Pormation 1408.03. In the USPS 2000] HOLIDAY: A FLORISTIC STUDY OF TSEGI CANYON, ARIZONA 33 Table 1. A Comparison of the Ten Largest Families of Vascular Flora of Tsegi Canyon, Petrified Forest, N.R, Capitol Reef N.R, Grand Canyon N.R, and the Navajo Nation. *Family not in top ten % flora, 'Kierstead, 1981, 2Heil et al., 1993, Phillips et. al., 1987, "^Mayes and Rominger, 1994. Petrified' CapitaP Grand^ Family Tsegi Canyon Forest N.P. Reef N.P Canyon N.P. Navajo** Nation Asteraceae 19.0 19.9 20.4 16.6 17.6 Poaceae 11.2 19.6 14.0 11.3 11.7 Brassicaceae J.J A 1 O.J A Q Fabaceae 5.2 5.6 8.4 4.8 8.1 Scrophulariacea 3.3 * 3.5 3.1 3.3 Chenopodiaceae 3.5 6.8 3.9 * 3.1 Rosaceae 3.2 * * 2.3 2.5 Boraginaceae 2.9 * * 3.0 2.3 Cactaceae 2.5 * * * * Salicaceae 2.2 * * * Total in Top Ten % of Flora 58.5% 72.4% 68.8% 53.6% 59.9% classification is it closest to the Cold Temperate Grasslands under the Great Basin Shrub-Grasslands (Brown, 1980). Alcoves, hanging gardens, and seeps are very specialized and variable components of the canyon. These plant communities vary according to direc- tional aspect, amount of sunlight received, depth of alcove, soil type, and amount and duration of water flow. Alcoves at Tsegi are created out of sandstone eroded by vertical movement of water across rock seams. Alcoves range from less than a meter to many meters in size, and may also include prehis- toric housing ruins. The species most common to alcove seeps include Mimulus eastwoodiae and Adiantum capillus-veneris. Other species found in cave seeps include Mentha arvensis, Selaginella mutica, Platanthera zothecina, Pragmites australis, Carex aurea, Carex lanuginosa, Carex specuicola, Oenotheera elata, Aquilegia micrantha, Epipactis gigantea, and Mimulus guttatus. Another type of seep found in the canyon comes out of a vertical canyon side, usually in loose, sandy soil. This type of seep receives a greater amount of sunlight in comparison to the hanging gardens in the alcoves. The plant composition in this type of seep includes grasses such as Avena fatua, Elymus canadensis, Elymus elymoides, Glyceria striata, Hordeum jubatum, Poa annua, Polypogon monspe- liensis, Schizachyrium scoparium, Secale cereale, and Sphenooopholis obtusata. Other plants at these sites include Cymopterus beckii, Apocynum can- nabinum, Artemisia ludoviciana. Aster frondosus, Solidago sparsiflora, Lithospermum incisum, and Glycyrrhiza lepidota. There is not a dominant spe- cies listed because of the differences of plant com- ponents present among various seeps. Seeps and alcoves in this area can harbor endemic, rare, and endangered plant species, such as Carex specuicola and Platanthera zothecina. Discussion The floristic composition of representative plant families of the canyon is similar to the floristic composition of the Navajo Nation as last reported by Mayes and Rominger (1994; Table 1). The can- yon has a similar number of Asteraceae and Po- aceae when compared to other similar sites. This is surprising as Tsegi, and the Navajo Nation, are grazed while the national parks used for compari- son are not. Perhaps a more detailed study is need- ed to determine whether or not the grazing affects floristic composition. The differences in the fami- lies showing a small percentage of the total prob- ably can be accounted for by the concentration of specialized habitat types, such as alcoves and ele- vation at each site. Species considered endemic to the Colorado Pla- teau by Welsh (1993) include Calochortus aureus, Platanthera zothecina. Astragalus zionis. Astraga- lus cottamii. Astragalus sesquiflorus, and Cymop- terus beckii. The rare and endangered plants in- clude Platanthera zothecina, a candidate species, Carex specuicola, a listed threatened plant, and Nama retrosum, Penstemon pseudoputus. Astraga- lus cottami, and Cymopterus beckii, whose popu- lations are being watched. Aletes sessiliflorus, iden- tified by L. Constance, and Cymopterus beckii were new reports for the state of Arizona. Introduced species are listed in Table 2. These exotic plants species comprise about 10% of the total plant population of the canyon. Many of the weedy herbs may have been introduced by domes- tic grazing animals, whose feed is supplemented with commercial hay, and also by the disturbing of the land by off-road vehicles. Three introduced plants, Tamarix ramosissima, Elaeagnus angusti- folia, and Ulmus pumila were introduced purposely in the Southwest as shade trees and to aid in erosion control (Welsh 1993). Of these, E. angustifolia is considered harmful in the canyon by the National Park Service. Attempts are currently being made to keep it out of Betatakin National Monument (Mel- berg 1996). There is evidence that suggests that the inner canyon has changed in the last 150 years. In 1916, L.C. Whitehead (MNA) collected Epipactis gigan- 34 MADRONO [Vol. 47 Table 2. Exotic Plants Found in Tsegi Canyon. 'Exotics found in Navajo Monument not listed in Brotherson et. al., 1978. ^State of Arizona Designated exotic plant species. Species Name Common Name Origin Monocotyledoneae Poaceae Avena fatua Bromus tectorum Dactylis glomerata Erenopyrm triticeum Polypogon monospeliensis Polypogon semiverticillatus Secale cere ale Dicotyledoneae Asteraceae Artemisia absinthium Cirsium vulgare Latuca serriola Sonchus asper Tagetes patula Taraxacum officinale Tragopogon dubius^ Xanthium strumarium Brassicaceae Capsella bursa-pastoris Corisppora tenella^ Descurainia sophia Sisymbrium altissimum Chenopodiaceae Kochia scoparia Salsola iberica Elaegnaceae Elaeagnus angustifolia Fabaceae Trifolium repends^ Medicago lupulina Medicago staiva Melilotus album Geraniaceae Erodium cicutarium Lamiaceae Draccocephalum tymiflorum Marrubium vulgare Plantaginaceae Plantago lanceolata Plantago major^ Ranunculaceae Ranunculus testiculatus^ Scrophulariaceae Verbascum thapsus Tamaricaceae Tamarix ramosissima Ulmaceae Ulmus pumila^ Zygophyllaceae Tribulus terrestris^ Wild Oats Cheat Grass Orchard Grass Annual Wheat Grass Rabbitfoot Grass Water Polypogon Cultivated Rye Absinthe Bull Thistle Prickly Lettuce Spiny Sow Thistle Marigold Dandelion Cocklebur Shepherds Purse Musk mustard Tumble Mustard Summer Cypress Tumble Weed Russian Olive White Clover Hop Clover Alfalfa White Sweet Clover Storkbill Horehound Broadleaf Plantain Bur Buttercup Wooly Mullein Tamarack Siberian Elm Puncture Vine Eurasia Eurasia Eurasia/ Africa Central Asia Eurasia/ Africa Eurasia Eurasia Europe Eurasia Europe Europe Mexico Eurasia Eurasia Eastern U.S. Europe Asia Europe Europe Eurasia Asia Europe Europe Europe Europe Europe Europe Eurasia Eurasia Eurasia Europe Eurasia Eurasia Eurasia Asia Eurasia tea, Pinus edulis, Abies concolor, Equisetum hie- male, Salix exigua, Populus tremuliodes, and Quer- cus gambelii from Tsegi Canyon. All of these spe- cies are now present in the canyon, though the Abies is now only found in one side canyon north- east of the Keet Seel ruin. Of these, Epipactis, Eq- uisetum, Salix, and Populus are usually found in more mesic soils. Clute (1920) described the can- yon as containing desert plants and some meso- phytes. He also stated, that at the time, there was a layer of "peat two feet thick" that contained snail shells. He felt that this was evidence of the previ- ously reported lakes and swamps. The most exten- sive historical collection was done by the Wetherill family in the 1930's. Most of the plants in that col- lection are also found at the present and are indi- cated as such in the floristic appendix of this paper. Plants included in the Wetherill collection (Wyman 2000] HOLIDAY: A FLORISTIC STUDY OF TSEGI CANYON, ARIZONA 35 1951) that were not found in the present study or listed by Brotherson et al. (1978) in their survey of Navajo National Monument include Abronia fra- grans, Cologonia angustifolia, Amaranthus retro- flexus, Juniperus communis, Bromus anomalus, Bromus vulgaris, Pachystema myrsinites, Oeno- thera lavandulifolius, Eriophyllum lanosum, Achil- lea lanulosa, Setaria viridis, Pedicularis centranth- era, Bouteloua eriopoda, Sporobolus pulvinatus, and Nicotiana trigonophylla. These particular plant species could possibly survive in the microhabitats of the present Tsegi drainage. Why they were not found during my survey periods is unknown. Al- though One Side Canyon was in the past called "Water lily Canyon", I could find no evidence of water lilies collected in the early 1900's. Conclusion Tsegi Canyon is a complex canyon of numerous drainages. It has experienced a large amount of ero- sion in the last hundred years. This has changed the nature of the bottom of the canyon from marshy pools to a faster moving creek. Erosion has lowered the bottom of the creek and that lowered the water table and affected small seeps along the sides of the canyon. Some marshy areas still exist, but only in limited areas in protected side-canyons. There is a good possibility that changes in the fauna and flora have occurred along with local extinctions. Hopefully, more studies will be done here to doc- ument the flora and fauna in this remote area of Arizona. Information needs to be gathered to create an effective means of preserving this riparian area while still allowing usage by local inhabitants. Acknowledgments I would like to thank Dr. Tina Ayers and Dr. Randy Scott for their guidance in preparing this paper. Dr. H. David Hammond for his help with identifying plants I found so frustrating, Bruce Melberg and the staff at Navajo National Monument for their help and support. Dr. Joseph E. Laf- erriere for his review and helpful comments, and my hus- band, Gordon, my sons, Lyle and Curtis, and fellow bota- nist Abril Perez, who accompanied me on several trips. Literature Cited BoNHAM, C. D. 1989. Measurements for Terrestrial Veg- etation. John Wiley and Sons. Brotherson, J. D., G. Nebeken, M. Skougard, and J. Fairchild. 1987. Plants of Navajo National Monu- menL Great Basin Naturahst 38(1): 19-30. Brotherson, J. D., Rushforth, S. R., and Johansen, J. R. 1983. Effects of Long-term Grazing on Cryptogam Crust Cover in Navajo National Monument, Ariz. Journal of Range Management 36(5):579-581. Brown, D. E., Lowe, C. H., and Pase, C. R 1980. A Digitized Systematic Classification for Ecosystems with an Illustrated Summary of the National Vege- tation of North America. USDA Forest Service Gen- eral Technical Report RM-73. Clay-Poole, S. T. 1980. Pollen Flora and Paleoecology of Holocene Alluvial Tsegi and Naha Formations, Northeast Arizona. Northern Arizona University. Flagstaff, AZ. Clute, W. N. 1920. Notes on the Navajo Region. The American Botanist. 26(2):38-47. Dean, J. S. 1969. Chronological Analysis of Tsegi Phase Sites in Northeastern Arizona. University of Arizona Press. Tucson, AZ. Hack, J. T. 1945. The Changing Physical Environment of the Hopi Indians of Arizona. Papers of the Peabody Museum of American Archaeology and Ethnology, Havard University. 35(1). Hastings, R. H. and Turner R. M. 1965. The Changing Mile. University of Arizona Press. Heil, K. D., Porter, J. M., Fleming, R., and Romme, W. H. 1993. Vascular Flora and Vegetation of Capitol Reef National Park, Utah. USDI, National Park Service Technical Report NPS/NAUCARE/NRTR-93/01. Kartesz, J. T, AND Kartesz, R. 1890. A Synonymized Checklist of the Vascular Flora of the United States, Canada, and Greenland, Vol. II, The Biota of North America. The University of North Carolina Press. Chapell Hill, NC. KiERSTEAD, J. R. 1981. Flora of Petrified National Park, Arizona. Thesis. Northern Arizona University. Mayes, V. O. and Rominger, J. M. 1994. Navajoland Plant Catalog. National Woodlands Publishing Company. Phillips, B. G., Phillips, A. M. and Bernzott, M. S. 1987. Annotated Checklist of Vascular Plants of Grand Canyon National Park. Grand Canyon Natural History Association, Monograph 7. Rowlands, P. G. 1994. Colorado Plateau Vegetation Ass- esment and Classification Manual. USDI, National Park Service, Technical Report NPS/NAUCPRS/ NRTR-94/06. U.S. Department of Commerce, 1996 Climatological Data. Welsh, S. L., Atwood, N. D., Higgins, L. C, and Good- rich, S. 1993. A Utah Flora, Great Basin Naturahst Memooirs No. 9. Wyman, L. C. 1951. The Ethnobotany of the Kayenta Navajo. University of New Mexico Press. Albuquer- que NM Appendix 1 Flora of Tsegi Canyon The nomenclature in this flora follows Kartesz (1980). The letter designation for endemic species is 'A, exotic species is 'B', federal category C2 is 'C2', fegeral category C3 is 'C3', federally listed threatened and endangered is 'T', those included in the Wetherill collection are 'W'. Selaginellaceae Equisetaceae Selaginella mutica D. C. Eaton. Perennial; moist Equisetum arvense L. Perennial; sandy, moist areas areas under cave seeps, Apr. -June. by springs and seeps, June-Aug. 36 MADRONO [Vol. 47 Equisetum hyemale L. Perennial; sandy, moist area near upper Laguna Creek. June- Aug., W. Adiantaceae Adiantum capillus-veneris L. Perennial; cave seeps and hanging gardens, May-Aug. W. Aspleniaceae Woodsia oregana D.C. Eaton. Perennial; seep run- ning over crack in sandstone cliff, Sept. Cupressaceae Juniperus osteospenna (Torr.) Little. Evergreen Tree; widespread on sandy flats, W. Ephedraceae Ephedra viridis Cov. Evergreen shrub; widespread on sandy flats, W. Pinaceae Abies concolor (Gord. & Glend.) Lindl. Evergreen tree; shady, upper areas of Keet Seel Canyon. Pinus edulis Engelm. Evergreen tree; widespread on sandy flats, W. Pinus ponderosa Dougl. Evergreen tree; sometimes found in side canyons. W. Pseudotsuga menziesii (Mirbel) Franco. Evergreen tree; found in shady side canyons, W. Agavaceae Yucca angustissima Engelm. ex Trel. Perennial; found on sandy flats, flowers in June, W. Yucca baccata var. baccata Torr. Perennial; north slope of Betatakin canyon, W. Commelinaceae Tradescantia occidentalis (Britt.) Smyth. Perennial herb; trail to Betatakin ruins and sandy areas in side canyons, May-June, W. Cyperaceae Carex aurea Nutt. Perennial herb; seeps and springs, June-Aug. Carex lanuginosa Michx. Perennial herb; seep near Keet Seel Ruins, June-Aug. Carex rossii F. Boott. Perennial herb; Betatakin canyon by trail bench, June-Aug. Carex specuicola J. T. Howell. Perennial herb; hanging gardens and seeps, June-Aug., A, T. Eleocharis palustris (L.) R. & S. Perennial herb; Betatakin creek, June-Aug. Scirpus pungens Vahl. Perennial herb; moist creek bottoms, June-Aug. Juncaceae Juncus arcticus Willd. Perennial herb; moist creek bottoms, May-Aug., W. Juncus bufonius L. Perennial herb; moist sand, low- er sidecanyons, June-Aug. Juncus saximontanus A. Nels. Perennial herb; by small creek, May-June, W. Liliaceae Allium macropetalum Rybd. Perennial herb; sandy flats, Apr. -May, W. Androstephium breviflorum Wats. Perennial herb; trail to Betatakin, Apr. -May. W. Calochortus aureus Wats. Perennial herb; sandy flats, June-July. W. Fritillaria atropurpurea Nutt. Perennial herb; creekside. Fir Canyon, June-July W. Smilacina stellata (L.) Desf. Perennial herb; moist canyon bottoms, June-Aug. W. Orchidaceae Corallorhiza maculata Raf. Perennial herb; lower Betatakin canyon, under trees, June-July, W. Epipacis gigantea Dougl. ex. Hook. Perennial herb; hanging gardens and seeps, June-Aug., W. Platanthera zothecina Higgins & Welsh. Perennial herb; hanging gardens and cave seeps, including Betatakin ruin, July-Aug. A, C2, W Poaceae Agrostis exarata Trin. Perennial herb; beside small stream in Keet Seel Canyon, June. Arista purpurea Nutt. Perennial herb, dry sandy soil, May-June, W Avena fatua L. Annual herb; found in moist sand and flat seeps, July-Sept., B. Bouteloua curtipendula (Michx.) Torr. Perennial herb; tree shade in side canyons, Aug.-Sept., W. Bouteloua gracilis (H.B.K.) Lag. ex Steudel. Pe- rennial herb; found on sandy flats with pinion trees, June-Sept., W. Bromus carinatus H. & A. Perennial herb; sandy soil in side canyon, May-June, W Bromus tectorum L. Annual herb; found on all sandy flats and near streams, April-Sept., B, W Dactylis glomerata L. Perennial herb; sandy soil beside creek, July-Aug., B. Elymus canadensis L. Perennial herb; moist sand near seep, Aug.-Sept. Elymus cinereus Scribn. & Merr. Perennial herb; sandy soil beside creek, Aug.-Sept. Elymus elymoides (Raf.) Swezey. Perennial herb; wet sand of seep, April-May, W. Elymus smithii (Rybd.) Gould. Perennial herb; sandy areas of lower canyon, April-June. Elymus trachycaulus (Link) Gould ex Shinners. Pe- rennial herb; sandy soil, May-June. Eremopyrum triticeum (Gaetm.) Nevski. Annual herb; dry sandy soil beside creek, July-Aug., B. Glyceria striata (Lam.) A. S. Hitch. Perennial herb; wet sandy soil by seep, June-July. Hilaria jamesii (Torr.) Benth. Perennial herb; dry sandy soil, June-July. 2000] Hordeum jubatum L. Perennial herb; wet sandy soil by seep, June-July. Hordeum pus ilium Nutt. Annual herb; dry sandy soil, June-July. Muhlenbergia andina (Nutt.) A. S. Hitch. Perennial herb; found in sandy soil of dry creek bed, Aug.- Sept., W. Muhlenbergia pungens Thurben in Gray. Perennial herb; dry sandy soil, Aug.-Sept., W. Monroa squarrosa (Nutt.) Torr. Annual herb; found on sandy soil, July- Aug., W. Pragmites australis (Cav.) Trin. ex Steudel. Tall perennial herb; found at Betatakin ruin, June- Sept., W. Poa pratensis L. Perennial herb, moist soil in shade, June-Sept. Polypogon semiverticillatus (Forsskal) Hylander. Perennial herb, moist, sandy soil, June-July, B. Puccinellia nuttalliana (Schultes) A. S. Hitch. Pe- rennial herb, sand by seep, June- July. Schizachyrium scoparium (Michx.) Nash in Small. Perennial herb, sand by seep, July-Aug. Secale cereale L. Annual herb, moist sand by seeps, July-Aug., B. Sphenopholis obtusata (Mitchx.) Scribn. Annual herb, wet sandy soil, July-Aug. Sporobolus cryptandrus (Torr.) gray. Perennial herb, dry sand, Aug.-Sept., W. Stipa comata Trin. & Rupr. Perennial herb, sandy soil, May-June, W. Stipa hymenoides R. & S. Perennial herb, dry sand. May- June, W. Vulpia octoflora Walter. Annual herb, sandy soil creekside, May-June, W. Typhaceae Typha domingensis Pers. Perennial, below Keet Seel ruin and side canyons, June-Sept. Aceraceae Acer negundo L. Tree, near streams and seeps, May-Oct., W. Amaranthaceae Amaranthus blitoides Wats. Annual herb, sandy soil near streams, July-Sept. Anacardiaceae Rhus aromatica var. trilobata (Nutt.) Gray. Shrub, upper side canyon, June- Aug., W Toxicodendron rydbergii (Small) Greene. Small shrub, upper side canyons, June-Sept., W. Apiaceae Aletes sessiliflorus Theobald & Tseng. Perennial herb, sand by streams, June-July, new report. Cymopterus acaulis (Pursh) Raf. Perennial herb, sand under trees, June-July, W. 37 Cymopterus beckii Welsh & Goodrich. Perennial herb near seeps, June-July, new report, A, C2. Cymopterus purpureus Wats. Perennial herb, clay soil near trees, June- July. Apocynaceae Apocynum cannabium L. Perennial herb, moist sand by seeps, June-Aug. Apocynum x medium Greene. Perennial herb, moist sand by stream, June-July. Asclepiadaceae Asclepias asperula (Decne.) Woodson. Perennial herb, sand, May-June, W. Asclepias latifolia (Torr.) Raf. Perennial herb, sun- ny sand, June-July. Asclepias speciosa Torr. Perennial herb, beside stream, July-Aug., W. Asclepias subverticillata (Gray) Vail. Perennial herb, sunny canyon sides, June-Aug. Asteraceae Ambrosia acanthicarpa Hook. Annual herb, sandy soil, Aug.-Sept., W. Antennaria neglecta Greene. Perennial herb, shade, sand, May-June, W. Antennaria parvifolia Nutt. Perennial herb, shade, June, W. Artemisia absinthium L. Perennial herb, dry sand, July-Aug. Artemisia campestris L. Perennial herb, dry sandy soil, shade, July-Aug., W. Artemisia carruthii Wood ex Carruth. Perennial herb, sandy soil, July-Aug. Artemisia dracunculus L. Perennial herb, shade, June-Aug., W. Artemisia frigida Willds. Perennial herb, sandy soil, June-Sept., W. Artemisia ludoviciana Nutt. Perennial herb, moist sand, July-Aug., W. Artemisia tridentata var. tridentata Nutt. Shrub, common in canyon, June-Aug., W. Aster frondosus (Nutt)T. & G. Annual herb, moist sand, Aug.-Sept., W. Aster glaucodes Blake. Perennial herb, creekbed, July-Sept. Brickellia californica Gray. Perennial subshrub, moist sand, July-Aug., W. Brickellia microphylla (Nutt.) Gray. Small shrub, sandy soil, Aug.-Oct., W. Brickellia oblongifolia Nutt. Perennial subshrub, dry clays, May-June. Chaetopappa ericoides (Torr.) Nesom. Perennial herb, dry sand, June-July, W. Chrysothamnus depressus Nutt. Low shrub, dry creek bottom, July-Sept. Chrysothamnus nauseosus (Pallas)Britt. Shrub, common in canyon, July-Sept. Chrysothamnus viscidiflorus (Hook.)Nutt. Shrub, lower canyon, July-Sept., W. HOLIDAY: A FLORISTIC STUDY OF TSEGI CANYON, ARIZONA 38 MADRONO [Vol. 47 Cirsium calcareum var. pulchellum (Greene)Welsh. Perennial herb, creek bottoms, July-Sept. Cirsium vulgare (Savi)Ten. Biennial herb, creek bottom, July-Sept. Conyza canadensis (L.)Cronq. Annual herb, creek bottom, June-Sept. Erigeron bellidiastrum Nutt. Annual herb, sandy soil, July-Aug. Erigeron compactus Blake. Perennial herb, dry sand. May- June. Erigeron eatonii Gray. Perennial herb, sand, shade, May-June, W. Erigeron flagellaris Gray. Perennial herb, near seep, April-May. Erigeron lonchophyllus Hook. Perennial herb, moist sand, July-Aug. Erigeron pumilis Nutt. Perennial herb, sandy soil, June- July. Erigeron speciosus (Lindl.) D.C. Perennial herb, sandy soil July-Aug. Erigeron utahensis Gray. Perennial herb, sandy soil, sun, May-June. Gnaphalium chilense Sprengel. Annual herb, moist sand, July-Aug. Gutierrezia sarothrae (Prush)Britt. & Rusby. Small shrub, sand, sun, July-Sept., W. Haplopappus ameriodes (Nutt.)Gray. Perennial herb, dry clays, May-June. Haplopappus spinulosus (Prush)D.C. Perennial herb, sand, shade, July-Aug. Helianthus petiolaris Nutt. Annual herbs, sand, sun, July-Aug., W. Heterotheca villosa (Pursh)Shinn. Perennial herb, sandy soil, conmion, July-Sept., W. Hymenopappus filifolius Hook. Perennial herb, west facing walls, June-July, W. Hymenoxys acaulis (Pursh)Parker. Perennial herb, sandy soil. May- June, W. Lactuca serriola L. Biennial herb, sandy soil, Aug.-Sept., B, W. Lygodesmia grandiflora (Nutt.)T.&G. Perennial herb, clay soil, June-July. Machaeranthera canescens (Pursh)Gray. Biennial herb, sandy soil, July-Aug., W. Machaeranthera grindeliodes (Nutt.)Shinn. Peren- nial herb, dry sand, June-July, W. Petradoria pumila (Nutt.)Greene. Perennial herb, dry sand, June-July. Psilostrophe sparsiflora (Gray)W.Nels. Perennial herb, sandy soil, June- Aug., W. Senecio douglasii DC. Perennial herb, sandy soil below seep, July-Aug. Senecio multilobatus T.&G. Perennial herb, sandy soil, shade, May-June, W. Senecio spartioides T.&G. Perennial herb, sandy soil, Aug.-Sept., W. Solidago canadensis L. Perennial herb, dry creek bottom, July-Sept., W. Sonchus asper (L.)Hill. Annual herb, moist sand, July-Sept., B. Stephanomeria exigua Nutt. Annual herb, sandy soil, July-Aug. Stephanomeria tenuifolia (Torr.)Hall. Perennial herb, sandy soil, July-Aug. Tagetes patula L. Annual herb, moist sandy soil, July-Aug., B. Taraxacum officinale Weber ex Wiggers. Perennial herb, moist soil, May-Sept., B. Thelesperma subnudum Gray. Perennial herb, sandy soil, sun, June-July, W. Tragopogon dubius Scop. Biennial herb, dry sand, June- Aug., B. Verbesina enceliodes (Cav.)Benth.&Hook. Annual herb, sandy soil, July-Aug. Wyethia scabra Hook. Perennial herb, creekbed sand, June-July. Xanthium strumarium L. Annual herb, moist sand, June-Aug., B. Berberidaceae Mahonia repens (Lindl.)G.Don. Small evergreen shrub, shade, W. Betulaceae Betula occidentalis Hook. Small trees, along creek sides. May- Aug. Boraginaceae * Cryptantha bakeri (Greene)Payson. Biermial herb, sandy soil, May-June. Cryptantha crassisepala (T.&G.)Greene. Annual herb, sand, sun, May-June, W. Cryptantha flava (A.Nels.)Payson. Perennial herb, sandy soil, May-June. Cryptantha fulvocanescens (Wats.)Payson. Peren- nial herb, dry sand, June-July. Cryptantha inaequata Johnston. Perennial herb, moist sand below seep, June- July. Cryptantha cinerea (Torr.)Cronq. Perennial herb, sandy soil, sun, May-June. Cryptantha circumscissa (H.&A.)Johnston. Annual herb, sandy soil, May-June. Lappula occidentalis (Wats.)Greene. Annual herb, dry sand, April-May, W. Lithospermum insicum Lehm. Perennial herb, moist sand, June-Aug. Brassicaceae Arabis perennans Wats. Perennial herb, sand, shade, June-Aug., W Arabis pulchra var. pallens Jones. Perennial herb, sandy soil, May-June. Capsella bursa-pastoris (L.)Medicus. Annual herb, near trails, May-June, B. Chorispora tenella (Pallas)DC. Annual herb, along trails, June-July, B. Descurainia pinnata (Walter)Britt. Annual herb, sandy soil, May-June, W. 2000] Descurainia sophia (L.) Webb ex Prantl. Annual herb, trail side, May-June, W. Dithyrea wislizenni Engelm. in Wisliz. Annual herb, sand, May-June, W. Lepidium montanum var. spathulatum (Rob- ins.)C.L. Hitch. Perennial herb, sandy soil, July- Aug., W. Lepidium virginicum L. Annual herb, moist sand by seep, July- Aug. Lesquerella intermedia (Wats.)Heller. Perennial herb, sandy soil, April-June, W. Pysaria newberryi Gray. Perennial herb, sandy soil, June-July. Sisymbrium altissimum L. Annual herb, sand dunes, July-Sept., B. Stanley a pinnata (Pursh)Britt. Perennial herb, clays, June-Aug. Streptanthella longirostris (Wats.)Rydb. Annual herb, sandy soil, shade, May-June. Strepthanthus cordatus Nutt. ex T.&G. Perennial herb, sand, shade, June-Aug., W. Thelypodium integrifolium (Nutt.)Britt.&Rose. Bi- ennial herb, sandy soil, June-Aug. Thlaspi montanum L. Annual herb, moist sand be- low seep, April-May. Cactaceae Coryphantha vivipara (Nutt.)Britt.&Rose. Perenni- al, sandy soil, shade, June. Echinocereus triglochidiatus Englem. Perennial, sandy soil, shade, June. Opuntia erinacea Englem. Perennial, sandy soil, sun, June. Opuntia fragilis (Nutt.) Haw. Perennial, sandy soil, shade, June, W. Opuntia phaeacantha var. discata (Griffiths)Benson & Walkington. Perennial, below Betatakin ruin, June. Opuntia polycantha Haw. Perennial, sandy soil, sun, June. Opuntia whipplei Engelm. Perennial, sandy soil, sun, June. Sclerocactus whipplei (Engelm.)Britt.&Rose. Pe- rennial, sandy soil, shade, June. Cannabaceae Humulus americanus Nutt. Perennial herb, dry creek bottom, June-July, W. Capparaceae Cleome serrulata Pursh. Annual herb, sandy soil, sun, May-July, W. Caprifoliaceae Symphoricarpos oreophilius Gray. Shrub, sandy soil, shade, June-Aug., W. Caryophyllaceae Arenaria fendleri Gray. Perennial herb, sandy soil, sun, May-June, W. 39 Silene menziesii Hook. Perennial herb, sandy soil, shade, May-June, W. Chenopodiaceae Atriplex canescens (Pursh)Nutt. Shrub, sandy soil, sun, July-Aug. Atriplex confertifolia (Torr.&Frem)Wats. Shrub, sandy soil, July-Aug. Atriplex powellii Wats. Annual herb, sandy soil be- low seep, July-Aug. Chenopodium album L. Annual herb, sandy soil, shade, June-Aug., B. Chenopodium leptophyllum (Moq.)Wats. Annual herb, sandy soil, July-Aug. Chenopodium rubrum L. Annual herb, sandy soil, July-Aug. Corispermum villosum Rybd. Annual herb, sandy soil, July-Sept. Kochia scoparia (L.)Schrader. Annual herb, sandy soil, July-Aug., B. Salsola iberica Sennen & Pau. Annual herbs, sandy soil, sun, July-Sept., B, W. Sarcobatus vermiculatus (Hook.)Torr in Emory. Shrub, sandy soil, July-Sept. Suaeda torreyana Wats. Perennial subshrub, sandy soil, July-Aug. Comaceae Cornus sericea L. Shrub, streamside, shade, May- Sept., W. Elaegnaceae Elaeagnus angustifolia L. Small tree, near creeks, May-Sept., B. Shepherdia rotundifolia Parry. Evergreen shrub, cliffsides, W. Ericaceae Arctostaphylos patula Greene. Shrub, sand, near creek, May-Sept., W. Euphorbiaceae Euphorbia lurida Engelm. Perennial herb, sand, sun, June-Aug. Euphorbia micromera Boiss. Annual herb, sand, sun, July-Sept. Fabaceae Astragalus amphioxys Gray. Perennial herb, dry sand, sun, June-July, W. Astragalus ceramicus Sheldon. Perennial herb, sandy soil, June-July. Astragalus cottamii Welsh. Perennial herb, sandy soil, shade, June-July, A, C3. Astragalus flavus Nutt. ex T.&G. Perennial herb, sandy soil, May-June. Astragalus lentiginosus Dougl. ex Hook. Perennial herb, sandy soil, sun, May-June. HOLIDAY: A FLORISTIC STUDY OF TSEGI CANYON, ARIZONA 40 MADRONO [Vol. 47 Astragalus mollissimus Torr. Perennial herb, sandy soil, June- July, W. Astragalus sesquiflorus Wats. Perennial herb, sand on canyon sides, June-Sept., A, W. Astragalus zionis Jones. Perennial herb, along trail, May-June, A. Glycyrrhiza lepidota Pursh. Perennial herb, moist sand, shade, July-Aug. Lathyrus brachycalyx Rybd. Perennial herb, sand, shade, July-Aug., W. Lupinus argenteus Pursh. Perennial herb, sand, sun, June- Aug. Medicago lupulina L. Annual herb, sand near creek, June-July, B. Medicago sativa L. Annual herb, sandy soil, sun, July-Sept., B. Melilotus alba Medic. Annual herb, sand, shade, June- July, B. Psoralidium lanceolatum (Pursh)Rydb. Perennial herb, sandy soil, June-July. Trifolium repens L. Perennial herb, sand near creek, June-July, B. Fagaceae Quercus gambelii Nutt. Small trees, sandy soil, canyon sides, May-Sept., W. Fumariaceae Corydalis aurea Willd. Annual herb, sandy soil, shade, June- July, W. Gentianaceae Swertia radiata (Kellogg) Kuntze. Perennial herb, sandy soil; June-Sept. Geraniaceae Erodium cicutarium (L.)L'Her. Annual herb, sand, sun, May-Sept., W. Geranium caespitosum James. Perennial herb, sandy soil, shade, June- Aug., W. Grossulariaceae Ribes cereum Dougl. Shrub, sand, shade, May- Sept., W. Ribes inerme Rydb. Shrub, side canyons, shade, June-Sept. Ribes leptanthus Gray. Shrub, side canyons, shade, June-July, W. Hydrangeaceae Fendlera rupicola Gray. Shrub, sandy soil, sun, May-July, W. Hydrophyllaceae Nama retrosum J.T. Howell. Annual herb, sand dunes, sun, June-July, A, C3, W. Phacelia ivesiana Torr. in Ives. Annual herb, sandy soil, sun, May-June. Lamiaceae Dracocephalum thymiflorum L. Annual herb, sandy soil, near creek, June- July, B. Hedeoma drummondii Benth. Annual or perennial herb, sand, June-July. Marrubium vulgare L. Perennial herb, sandy soil, Aug.-Sept., B. Mentha arvensis L. Perennial herb, moist sand, shade, Aug.-Sept., W. Poliomintha incana (Torr.)Gray. Shrub, sandy soil, May-June, W. Linaceae Linum aristatum Engelm. Annual herb, sandy soil, Aug.-Sept. Linum perenne L. Perennial herb, sandy soil, shade, July-Aug. Loasaceae Mentzelia albicaulis Dougl. ex Hook. Annual herb, sandy soil, June-July. Malvaceae Sphaeralcea parvifolia A. Nels. Perennial herb, sand, sun, May-July. Nytaginaceae Mirabilis linearis (Pursh)Heimerl. Perennial herb, sandy soil, July-Aug., W. Mirabilis multiflora (Torr.)Gray in Torr. Perennial herb, sand, shade, July-Aug. Mirabilis oxybaphoides (Gray)Gray in Torr. Peren- nial herb, sandy soil, Aug.-Sept. Tripterocalyx carneus (Greene)Galloway. Annual herb, sand, sun, July-Aug., W. Onagraceae Epilobium ciliatum Raf. Perennial herb, moist sand, shade, June- Aug., W. Gayophytum racemosum T.&G. Annual herb, dry sand, sun, May-June. Oenothera caespitosa Nutt. Perennial herb, sand, beside trail, June-July. Oenothera elata H.B.K. Biennial herb, sandy, shade, July-Sept. Oenothera pallida Lindl. Annual herb, sand, sun, June-July, W. Orobanchaceae Orobanche multiflora Nutt. Perennial herb, near A. tridentata Nutt., June- July, W. Plantaginaceae Plantago lanceolata L. Perennial herb, sandy soil, beside trails, May-June, B. Plantago major L. Perennial herb, moist sand, June- Aug., B. 2000] HOLIDAY: A FLORISTIC STUDY OF TSEGI CANYON, ARIZONA 41 Plantago patagonica Jacq. Annual herb, dry sand, near trail, May-June, W. Polemoniaceae Gilia aggregata (Pursh)Sprengel. Perennial herb, sandy soil, sun, June- Aug., W. Gilia leptomeria Gray. Annual herb, sandy soil, sun, June-July. Gilia longiflora (Torr.)D.Don. Annual herb, sandy soil, shade, June-July, W. Leptodactylon pugens (Torr.)Nutt. Subshrub, sand, sun, May-July, W. Polygonaceae Erigonum alatum Torr. in Sitg. Perennial herb, sandy soil, sun, May-July, W. Erigonum cemuum Nutt. Annual herb, sandy soil, Aug.-Sept. Erigonum microthecum Nutt. Small shrub, sandy soil, Aug.-Sept., W. Polygonum aviculare L. Annual herb, sandy soil, Aug.-Sept. Polygonum douglassi Greene. Annual herb, sandy soil, June-July. Portulaceae Portulaca oleraceae L. Annual herb, moist sand, June- July. Portulaca retusa Engelm. Annual herb, sand be- hind trail, June-July. Talium parviflorum Nutt. Perennial herb, sandy de- pressions, May-June. Primulaceae Androsace septentrionalis L. Annual herb, sandy soil, shade, June-July, W. Ranunculaceae Aquilegia micrantha Eastw. Perennial herb, hang- ing gardens, June-July, W. Clematis lingusticifolia Nutt. Woody vine, canyon sides, shade, June-July, W. Delphinium andersonii Gray. Perennial herb, sand along trail, June-July, W. Ranunculus cymbalaria Pursh. Perennial herb in marshy areas, May-Sept., W. Ranunculus testiculatus Crantz. Annual herb, sandy soil, sun, May-June, B. Thalictrum fendleri Engelm. Perennial herbs, sand, shade, June- Aug., W. Rhamnaceae Rhamnus betulifolia Greene. Shrub, above pool, shade, June. Rosaceae Amelanchier alnifolia (Nutt.)Nutt. Shrub, sandy soil, shade, May-June. Amelanchier utahensis Koehne. Shrub, sandy soil, sun, May-June, W. Cercocarpus intricatus Wats. Shrub, sand, sun, May-June, W. Cercocarpus montanus Raf. Shrub, sandy soil, par- tial shade, June-July. Holodiscus dumosus (Nutt.)Heller. Shrub, sandy soil, June-July, W. Prunus angustifolia Marsh. Small trees, Keet Seel ruin, May-July. Prunus virginiana L. Small tree, beside streams, June-July, W. Pursia mexicana (D.Don)Welsh. Shrub, sandy soil, sun, May-June, W. Pursia tridentata (Pursh)DC. Shrub, sandy soil, sun, May-June, W. Rubiaceae Galium aparine L. Annual herb, moist sand, shade, June- Aug. Salicaceae Populus angustifolia James. Tree, sandy soil, by streams. May. Populus fremontii Wats. Tree, sand, along streams. May. Salix exigua Nutt. Shrub, sand, sun, along streams. May- June. Salix laevigata Bebb. Small tree, sand, along stream, May-June, W. Salix lasiolepis Benth. Shrub, sand, along stream, May- June. Salix monticola Bebb. ex Coult. Shrub, moist sand, May-June. Santalaceae Comandra umbellata (L.)Nuatt. Perennial herb, sandy soil, sun, June-July, W. Saxifragaceae Heuchera parvifclia Nutt. in T.&G. Perennial herb, sandy soil, shade, June-July, W. S crophul ariaceae Castilleja chromosa A. Nels. Perennial herb, sandy soil, June-July. Castilleja linariifolia Benth. Perennial herb, sandy soil, shade, May-June, W. Cordylanthus wrightii Gray. Annual herb, sandy soil, shade, July-Aug., W. Mimulus eastwoodiae Rydb. Perennial herb, hang- ing gardens, July-Aug. Mimulus guttatus DC. Perennial herb, moist sand, shade, June-July. Mimulus rubellus Gray. Annual herb, sandy soil, shade, July-Aug. Penstemon barbatus (Cav.)Roth. Perennial herb, sand, July-Aug., W. 42 MADRONO [Vol. 47 Penstemon comarrhenus Gray. Perennial herb, sand, sun, June-July, W. Penstemon eatonii var. undosus Jones. Perennial herb, sand, shade, May-June, W. Penstemon pseudoputus (Crosswhite) N.Holmgren. Perennial, sand, June- July, A, C3. Penstemon rostriflorus Kellogg. Perennial, sand, shade, June-July. Verbascum thapsus L. Biennial, sand, sun, July- Aug., B. Veronica pergrina L. Annual herb, moist sand, June-July. Solanaceae Chamaesaracha coronopus (Dunal)Gray. Perennial herb, sandy soil, Aug.-Sept., W. Datura wrightii Regel. Annual herb, sandy soil, partial shade, July- Aug. Lycium pallidum Miers. Shrub, sand, sun, May- June, W. Pysalis hederifolia Gray. Perennial herb, sand, sun, July- Aug. Solanum jamesii Torr. Perennial herb, sand, by trail, July- Aug., W. Tamaricaceae Tamarix ramosissima Ledeb. Shrub, sand, along ; washes, May-June, B. ' Ulmaceae i Ulmus pumila L. Tree, along sides of creek, May- June, B. i I Valerianaceae Valeriana acutiloba Rybd. Annual herb, sand, shade, July- Aug., W. Verbenaceae Verbena bracteata Lag.&Rodr. Perennial herb, sand, sun, Aug.-Sept. Viscaceae Phorodendron juniperinum Gray. Parasitic peren- nial, found on juniper trees. Zygophyllaceae Tribulus terrestris L. Annual herb, sand, sun, July- Aug., B. Madrono, Vol. 47, No. 1, pp. 43-52, 2000 MIXED CONIFER AND RED FIR FOREST STRUCTURE AND USES IN 1899 FROM THE CENTRAL AND NORTHERN SIERRA NEVADA, CALIFORNIA Scott L. Stephens ' - 'Natural Resources Management Department, California Polytechnic State University, San Luis Obispo, CA 93407 -Current Address: Division of Forest Sciences, Department of Environmental Science, Policy, and Management; University of California, Berkeley, CA 94720-3114 stephens@nature.berkeley.edu Abstract Historical data collected from five "average" mixed conifer stands, four large mixed conifer stands, and four red fir stands from the central and northern Sierra Nevada by George Sudworth in 1899 were analyzed to determine historic forest structure including diameter distributions, basal areas, and snag and live tree densities. The effects of early logging operations on stand composition and structure is quantified by comparing characteristics of the trees that were harvested versus those unharvested in four mixed conifer stands. Average diameter at breast height (DBH) was 86 cm (34 inches) in the "average" mixed conifer stands, 110 cm (43 inches) in the large mixed conifer stands (this was equal to the average DBH of 8 mixed conifer stands sampled by Sudworth in the southern Sierra Nevada), and 77 cm (30 inches) in the red fir stands for trees greater than 30.5 cm DBH. Shade intolerant tree species dominated the "average" mixed conifer stands, shade intolerant, intermediate, and shade tolerant species were abundant in the large mixed conifer stands, and Abies magnifica Andr. Murray dominated the red fir stands. Mean tree density for the "average" mixed conifer, large mixed conifer, and red fir stands was 229 trees/ha, 235 trees/ha, and 433 trees/ha, respectively. Average tree density was higher in Sudworths southern Sierra Nevada mixed conifer stands when compared to the central and northern Sierra Nevada. Snag density averaged 5/ha in the large mixed conifer stands and 17.5/ha in the red fir stands. Early logging operations removed the majority of the Pinus spp. and Pseudotsuga menziesii (Mirbel) France leaving large amounts of Calocedrus decurrens (Torrey) Florin and Abies concolor (Gordon & Glend.) Lindley. Information from this study can assist in the characterization of historic stand structure in these forest types. The absence of fire in the 20th century and past harvesting operations have modified the structure and ecosystem processes in the coniferous forests of the Sierra Nevada. An increase in the density of small shade tolerant trees has been produced in many forest types (Leopold et al. 1963; Hartes veldt and Harvey 1967; Vankat and Major 1978; Parsons and DeBendeetti 1979; Bonnicksen and Stone 1982) and this increase has resulted in a decrease in forest sustainability (Weatherspoon and Skinner 1996; van Wagtendonk 1996; Stephens 1998). Changes in climate over the last centaury may have also contributed to the changes in forest structure (Millar and Woolfenden 1999). Historical and prehistoric information on the structure (density, size distribution, and species composition) of mixed conifer forests are relatively rare and they have been reviewed elsewhere (Ste- phens and Elliott-Fisk 1998). One of the methods that can be used to determine prehistoric forest structure is the analysis of data from early forest inventories. These data provide quantitative infor- mation on historic forest structure, however, the re- sults from the analyses can be biased because the methods used to select the stands were frequently not recorded (Stephens and EUiott-Fisk 1998). Analysis of historical data have been done for the Stanislaus and Lake Tahoe Forest Reserves (Sudworth 1900), portions of the northern Sierra Nevada and the Transverse Ranges of southern Cal- ifornia (McKelvey and Johnston 1992), and por- tions of the southern Sierra Nevada (Stephens and Elliott-Fisk 1998). All of these studies discuss early logging operations but no work has been done that quantifies the effects of early logging at the stand level, quantifies the amount of hardwoods present historically in mixed conifer forests, determines historic snag densities and sizes, or differentiates between average and mature mixed conifer stands. Early logging operations affected the composi- tion and structure of Sierra Nevada forests, es- pecially between 1860 and 1950 (Laudenslayer and Darr 1990). In 1899, approximately 45 percent of the trees harvested in California were either Pinus ponderosa Laws (ponderosa pine) or Pinus lam- bertiana Douglas (sugar pine). Most early logging operations in the Sierra Nevada harvested all trees that were considered to be merchantable at the time of the harvest (Laudenslayer and Darr 1990). The viability of the California spotted owl (Strix occidentalis occidentalis) is receiving major atten- 44 MADRONO [Vol. 47 tion in California. The owl prefers to nest in mixed conifer forests with 80 percent of the nesting sites occurring in this forest type followed by 10 percent in Abies magnifica Andr. Murray (red fir) and 7 percent in Pinus ponderosa hardwood forests (Ver- ner et al. 1992). The remaining 3 percent of nests occur in eastside pine forests and foothill riparian- hardwood habitats in the western Sierra Nevada foothills (Vemer et al. 1992). The habitat requirements of the California spot- ted owl have been investigated and it nests in old- growth forests with high canopy cover (Gutierrez et al. 1992). A relatively high number of snags and down logs are also correlated to the current nesting sites of the California spotted owl (Gutierrez et al. 1992) but no prehistorical data exist on the abun- dance of snags or fuel loads in this forest type mak- ing it difficult to describe the composition of the prehistorical habitat. The objective of this paper is to analyze mixed conifer and red fir forest inventory data acquired by George Sudworth in 1899 from the central and northern Sierra Nevada to further our understanding of forest conditions and their management in the late 19th century. Analysis includes snag and live tree densities, basal areas, diameter distributions, and quantification of the effects of early logging operations on stand composition and structure. Study Site and Methods The historic data analyzed in this paper were ob- tained from the area of the central and northern Sierra Nevada that now includes the southern por- tion of the Tahoe National Forest, the El Dorado National Forest, and northern portion of the Stan- islaus National Forest. Mixed conifer and red fir forests were surveyed in 1899 by George B. Sudworth while employed by the United States Geological Survey. The purpose of this survey was to inventory the forest reserves of the Sierra Nevada. The original unpublished field notebooks (Sudworth 1899) were the source of the inventory data analyzed in this paper. Sierra Nevada mixed conifer forests sampled by Sudworth were composed of white fir Abies con- color (Gordon & Glend.) Lindley (white fir), Abies magnifica, Pinus ponderosa, Pinis lambertiana, Pinis jejfreyi Grev. and Balf (Jeffrey pine), Calo- cedrus decurrens (Torrey) Florin (incense cedar), Pseudotsuga menziesii (Mirbel) Franco (Douglas- fir), and Quercus kelloggii Newb. (California black oak). The red fir forests were composed of Abies magnifica, Pinus jejfreyi, Pinus monticola Douglas (western white pine), Pinus contorta spp. murray- ana (Grev. & Balf.) Critchf. (lodgepole pine), and Tsuga mertensiana (Bong.) Carriere (mountain hemlock). Red fir forests are widely distributed and they can be found on both the west and east sides of the Sierra Nevada (Rundel et al. 1977). All stand data recorded by Sudworth in 1899 are analyzed in this paper with the exception of one stand located in a pure Pinus jejfreyi forest because of no replication in this forest type. Exact stand locations are not given in the field notebooks but references to rivers, mountains, and landmarks are included (Sudworth 1899). Five "average" mixed conifer stands, four large mixed conifer stands, and four red fir stands were recorded in the 1899 field notebooks (Sudworth 1899). Mixed conifer stand data were stratified into two classes (average and large) whereas this was not done in the southern Sierra Nevada analysis (Stephens and Elliott-Fisk 1998) because no stands were identified by Sudworth in his notebooks as having "average" characteristics. Sudworth recorded the species, diameter at breast height (DBH), and number of 4.9 m (16 ft) logs for each tree greater than 30.5 cm (12 inches) DBH (one 28 cm DBH tree was recorded in a red fir stand). Each stand was sampled with one 0.1 ha (0.25 acres) plot. He recorded notes on regeneration (estimate of density by species, not a complete seedling inventory), forest floor depth, and other information such as the revenue generated by early logging operations. He also frequently commented on the effects of early grazing and burning on the Sierra Nevada and his conunents are summarized below. The following values were calculated by aver- aging all stand data for each of the 3 forest types ("average" mixed conifer, large mixed conifer, red fir): number of snags per hectare, snag basal area, diameter and species of trees removed by early log- ging operations, basal area per hectare by species, number of trees per hectare by species (density), quadratic mean diameter by species, percent total basal area by species, and percent total stocking by species. Stand data are summarized and discussed, but a statistical analysis was not performed. Selection of an appropriate analysis method requires informa- tion on sampling procedures which are unknown for this early forest inventory (Stephens and Elliott- Fisk 1998). Results ''Average'' mixed conifer stands. The five mixed conifer stands denoted as "average" by George Sudworth were dominated by moderate sized trees of several species. The average quadratic mean di- ameter for all trees over 30.5 cm DBH was 86 cm (34 inches). Average tree density was 229 trees/ha (92 trees/acre) (range 150-300 trees/ha). Average basal area was 130 m^/ha (558 ftVacre) (range 94- 186 m^/ha). Table 1 summarizes all stand calcula- tions for the "average" mixed conifer stands. The largest trees in the "average" mixed conifer stands were Pinus lambertiana with an average DBH of 108 cm (42 inches). The largest Pinus lam- bertiana recorded in the inventory had a DBH of 2000] STEPHENS: MIXED CONIFER AND RED FIR FOREST STRUCTURE AND USES IN 1899 45 Table 1. Average Calculations of George Sudworth's 5 "average" Mixed Conifer Stands in the Central and Northern Sierra Nevada in 1899 (Standard Error). * Average value for all stands. Basal area DBH Percent (m^/ha) Trees/ha (cm) of total Percent of Tree [130]* [229]* [86]* basal area trees/ha Abies concolor 5.3 6.0 105.7 4 3 (5.2) (6.0) (0) Calocedrus decurrens 26.0 54.0 80.6 20 24 (6.6) (11.2) (10.5) Pifius IdfTihefticinci 8.8 12 107.8 7 5 (4.2) (7.4) (22.8) Pinus ponderosa 56.6 106.0 83.9 43 46 (23.1) (37.0) (12.1) Pseudotsuga menziesii 30.9 38.0 101.2 24 16 (20.2) (24.6) (1.8) Quercus kelloggii 2.4 13.3 58.0 2 6 (1.4) (3.7) (7.2) 152 cm (60 inches). Pinus ponderosa was the most common species comprising 46 percent of total stocking and 44 percent of total basal area (Table 1). Abies concolor was rare in the stands accounting for only 3 percent of total stocking and 4 percent of total basal area. Calocedrus decurrens and Pseu- dotsuga menziesii were the next most common spe- cies, after Pinus ponderosa, respectively. The av- erage DBH of the Quercus kelloggii was the small- est of the species found in the mixed conifer stands, the conifer with the smallest average DBH was Calocedrus decurrens (Table 1). Pinus lambertiana, Pseudotsuga menziesii, and Abies concolor all had similar average DBH's of approximately 105 cm whereas Pinus ponderosa and Calocedrus decurrens had average DBH's of approximately 82 cm. Quercus kelloggii accounted for an average of 6 percent of stand stocking. No snags were recorded in the average mixed conifer stands (Table 2). Four of the "average" mixed conifer stands in- ventoried by Sudworth were in the process of being harvested in 1899. Sudworth's notebooks recorded the diameter and species of all trees harvested and also recorded the same information on all trees that remained after the harvesting operation. All of the Pseudotsuga menziesii trees in these Table 2. Characteristics of Snags Found in Mixed Conifer and Red Fir Stands in the Central and North- ern Sierra Nevada in 1899. Density Average Average range Average basal density (snags/ DBH area Stand type (snags/ha) ha) (cm) (m^/ha) Average mixed conifer Large mixed conifer Red fir four stands were harvested along with 88 percent of the Pinus lambertiana trees (Table 3). The ma- jority of the wood harvested from these stands was from Pinus ponderosa trees with an average of 64 m^/ha (275 ft^/acre) removed and this was 2.4 times greater than Pseudotsuga menziesii which was the next most common species harvested. The amount of Calocedrus decurrens and Abies concolor trees harvested was low, averaging 13 percent and 33 percent, respectively (Table 3). The following comments were written by George Sudworth in the original field notebooks and in- clude information about regeneration and impacts from early European settlers (Sudworth 1899). September 27, 1899. Near Beech Sawmill (above Placerville) on Big Iowa Canyon. No reproduction (manzanita brush) but abundant a few yards distant. Grazed, no humus, all trees fire marked. September 28, 1899. South of Blair Sawmill (near Sly Park) on summit of ridge. All touched with fire, humus 1-2 inches in spots. Ample repro- duction of all species in patches. September 30, 1899. Ssimple on big hill south west of Grizzly Flat 0.5 mile. Humus all burned off. October 5, 1899. 2 miles east of Whitmore's IVIill Table 3. Average Amount Harvested in 4 "average" Mixed Conifer Stands Located in the Central and Northern Sierra Nevada in 1899. Tree Trees/ ha cut (per- cent) Basal area cut (per- cent) Basal area DBH of cut trees (mV cut ha) (cm) 0 0 0 0 5 0-10 108.7 4.6 17.5 0-60 57.3 4.5 Abies concolor 33.3 48.2 3.9 127 Calocedrus decurrens 12.5 12.5 1.6 88.9 Pinus lambertiana 87.5 92.8 17.1 104.3 Pinus ponderosa 57.9 64.2 64.3 103.3 Pseudotsuga menziesii 100.0 100.0 27.2 101.7 Quercus kelloggii 25.0 38.1 2.9 86.4 46 MADRONO [Vol. 47 Fig. 1. El Dorado county, 1899. Opposite Snyder and Sherman's Ranch. Yellow pine (mixed conifer) forest on south slope of Silver Fork. Ponderosa pine 91-193 cm (36-76 inches) in diameter, 46-50 m high (150-165 feet), clear 8- 11m (25-35 feet), ten in 0.1 ha (0.25 acre), 3-5 white fir (Abies concolor) same size. Cattle grazed. (Mill Creek, near Volcano), representing no cut stumpage, rolling flat 1000 feet above creek bot- tom. No humus, frequent burning destroyed all. Abundant reproduction of pines and cedar 5-8 years old, mostly under 4. October 5, 1899. Near Whitmore's Mill but in shallow ravine. Taken as a whole mill operator es- timates output 10-20 thousand per acre. Abundant reproduction of all species. Taxus brevifolia Nutt., dogwood, and Acer macrophyllum Pursh abundant. Humus in part 3-6 inches. Large mixed conifer stands. The four large mixed conifer stands were dominated by large trees of several species and the average quadratic mean diameter at breast height was 110 cm (43 inches) for all trees above 30.5 cm DBH. Average tree den- sity was 235 trees/ha (94 trees/acre) (range 160- 300 trees/ha). Average basal area was large 215 mV ha (923 ftVacre) (range 188-232 m^/ha). The stands inventoried by Sudworth were relatively open and dominated by large trees (Fig. 1). Table 4 sunmia- rizes all stand calculations for the large mixed co- nifer stands. Abies concolor was the most common species comprising 46 percent of total stocking, but only accounting for 34 percent of total basal area be- cause of their relatively small diameters. The larg- est tree inventoried in these stands was a Pseudo- tsuga menziesii and it had a DBH of 188 cm (74 inches). Pseudotsuga menziesii made up only 16 percent of the trees/ha but contributed 24 percent of the basal area of the stands because of their large size. Abies concolor trees were much more com- mon in the large mixed conifer stands when com- pared to the "average" mixed conifer stands (46 percent stocking versus 3 percent stocking, respec- tively). Abies concolor, Abies magnifica, and Calocedrus decurrens were the smallest trees with average qua- dratic mean diameters of approximately 93 cm. Pi- nus ponderosa and Pinus lambertiana were larger with average diameters of approximately 112 cm. I 2000] STEPHENS: MIXED CONIFER AND RED FIR FOREST STRUCTURE AND USES IN 1899 47 Table 4. Average Calculations of George Sudworth's 4 Large Mixed Conifer Stands in the Central and Northern Sierra Nevada in 1899 (Standard Error). * Average value for all stands. Basal area Percent of (m^/ha) Trees/ha DBH (cm) total basal Percent of Tree [215]* [235]* [110]* area trees/ha Abies concolor 1 1.5 1 AT C y /. J "2/1 4o (ZU.U) (J8.Z) (0.4) Calocedrus decurrens ZD. / yz.u IZ 1 /I 14 (9--5) ( \ n A\ • Pinus lambertiana z / . J 1 no 1 1 A lO iZ (is. J) (ZU.5) Pinus ponderosa 10.5 10.0 115.3 5 4 (10.5) (10.0) (0) Pinus jejfreyi 13.8 10.0 120.4 6 4 (11.4) (7.1) (21.4) Pseudotsuga menziesii 51.8 37.5 123.7 24 16 (45.4) (31.2) (11.8) Abies magnifica 6.2 10.0 88.9 3 4 (6.2) (10.0) (0) and the largest trees were Pinus jejfreyi and Pseu- dotsuga menziesii with average diameters of ap- proximately 122 cm. Quercus kelloggii was not re- corded in any of the large mixed conifer stands. Snag density averaged 5/ha with a range 0-10/ ha (Table 2). Average snag quadratic mean diam- eter was 109 cm and snag average snag basal area was 4.6 m^/ha. The following comments were written by George Sudworth in the original field notebooks and in- clude information about regeneration and impacts from early European settlers (Sudworth 1899). September 3, 1899. 12-15 miles west of Bloods, north slope of Mokelumne River. 30 concolor 2-8 inches diameter, 100 under 6 inches. 5 sugar pine under 2 feet high. Thickets of Acer oblusifobium iglabruml). September 8, 1899. South slope of Bear River, one half way up slope. Seedlings of all in spots near blue ceonothus when protected from tramping of cattle. No sheep here, but no humus. Abundant blue ceonothus chaparral. September 9, 1899. South lower slope of Silver Fork (American River) in rich bottom bench (at point where a little stream enters Silver Fork). Dense fir and cedar on outskirts, no seedlings with- in. Humus 2-3 inches, cattle grazing and sheep. September 21, 1899. 1.5 miles south of IVIerzns, across (west) of Dark and IVIulton Canyons (where Georgetown road crosses). South slope of South Fork of the Consumnes River (?) (Sudworth in- cluded the? mark and was probably referring to the American River). Abundant reproduction of all spe- cies 1-12 years old, all fire marked 15 years back. Humus 1.5 inches deep, soil sandy loam with rock. Red fir stands. Sudworth sampled four red fir stands during this inventory and all stands were dominated by Abies magnifica. The average qua- dratic mean diameter at breast height was 77 cm (30 inches) for all trees inventoried. Average tree density was 433 trees/ha (173 trees/acre) (range 180-610 trees/ha) for trees greater that 28 cm DBH. Average basal area was 202 m^/ha (867 ftV acre) (range 98-286 m-/ha). Table 5 summarizes all stand calculations for the red fir stands. Table 5. Average Calculations of George Sudworth's 4 Red Fir Stands in the Central and Northern Sierra Nevada in 1899 (Standard Error). * Average value for all stands. Basal area DBH Percent of (m^/ha) Trees/ha (cm) total basal Percent of Tree [202]* [433]* [77]* area trees/ha Pinus jejfreyi 32.2 25.0 128.1 16 6 (32.2) (25.0) (15.1) Abies magnifica 136.2 272.5 80.1 67 63 (55.3) (93.5) (10.9) Pinus monticola 11.3 30.0 66.4 6 7 (8.0) (19.2) (19.4) Pinus contorta 12.6 70.0 47.8 6 16 (12.6) (70.0) (12.0) Tsuga mertensiana 9.3 35.0 58.0 5 8 (9.3) (35.0) (14.5) 48 MADRONO [Vol. 47 Fig. 2. Amador county, 1899. Near sawmill 5-6 km (3-4 miles) below dam on Bear River. Forest fire in fir and pine, killed all seedlings, just started. The largest trees in the red fir stands were Abies magnifica and Pinus jejfreyi with DBH's of 160 cm (63 inches). Abies magnifica was the most common tree in the stands accounting for 63 percent of all trees inventoried and 68 percent of average stand basal area (Table 5). The next most common tree found was Pinus contorta which accounted for 16 percent of all trees but only contributed to 6 percent of basal area because of the smallest DBH of any species in this forest type. Snag density averaged 17.5 per ha with a range 0-60 per ha (Table 2). Average snag quadratic mean diameter was 57 cm (22 inches) and average snag basal area was 4.5 m^/ha. The following comments were written by George Sudworth in the original field notebooks and in- clude information about regeneration and impacts from early European settlers (Sudworth 1899). September 2, 1899. On foothill (above) Bear Meadow, north fork of Stanislaus River. No graz- ing, 40 young trees under 10 inches diameter. Hu- mus 4-6 inches deep, no herbaceous growth. 75- 100 seedlings 2-10 inches. September 7, 1899. On south slope 4-5 miles down on Silver Fork (near Silver Lake and Kirk- wood). Sheep grazing, no reproduction. Scattered bunches of blue ceonothus. Earth bare, rock and gravel. September 7, 1899. North side of Thimble Peak (west of Kirkwood Meadow). On volcanic and granite. No humus, grazed by sheep. Dense shade in part, no reproduction. 2 Abies magnifica down. September 13, 1899. On Rocky flat between Ly- ons and Blakley (south fork Silver Creek, west side of Pyramid Peak). Abundant 1 year fir seedlings, 50 fir under 10 feet, 20 Murr {Pinus contorta) pines 2-10 feet, 5 Pimo (Pinus monticola) 1-3 feet. Hu- mus 1 inch. Cattle grazed, no sheep within 5 years. Discussion Sudworth 's noted recent evidence of fire in many stands and believed fires were ignited by sheep herders to increase forage production and by log- gers to consume slash fuels (Fig. 2). This burning apparently did not spread extensively because fire scar analysis in the Sierra Nevada have documented the almost complete removal of surface fires in 2000] STEPHENS: MIXED CONIFER AND RED FIR FOREST STRUCTURE AND USES IN 1899 49 many mixed conifer forests in the 1 860-1 870's (Kilgore and Taylor 1979; Swetnam et al. 1990; Swetnam et al. 1992; Caprio and Swetnam 1995) at the same time burning was reportedly being used by loggers and sheep herders (Sudworth 1900; McKelvey and Johnston 1992; Stephens and El- liott-Fisk 1998). Regeneration was noted in the majority of "av- erage" mixed conifer stands. More site resources (light, water, nutrients) were probably available for regeneration in the "average" mixed conifer stands because of their lower stocking and basal areas. Re- generation in mixed conifer forests probably oc- curred prehistorically when small gaps were created by the interaction of fire and locally high fuel loads (Stephens et al. 1999). Regeneration was noted in half of the red fir stands and livestock grazing was noted in all stands. Sudworth noted that in some high elevation sites sheep were actually grazing on conifer seed- lings (Sudworth 1899). Many photos in the collec- tion show complete bare mineral soils (Fig. 3) and seedlings were reportedly also trampled by live- stock. Early logging operations had a dramatic effect on the species composition and diameter distribu- tions of mixed conifer stands sampled by Sudworth (Table 3). The majority of the Pinus spp. and Pseu- dotsuga menziesii were harvested in the stands leaving large amounts of Calocedrus decurrens and Abies concolor (Fig. 4). This type of logging op- eration has been described as "high-grading" be- cause of the preference for large trees of particular genera. In this period it was common for all mer- chantable trees to be removed during logging op- erations (Laudenslayer and Darr 1990). Abies con- color and Calocedrus decurrens were therefore left because they were of relatively low economic value late in the 19'^ centaury. Early logging operations coupled with a national fire suppression policy that began in the early 20* century favored shade tolerant species such as Cal- ocedrus decurrens and Abies concolor. Climate changes over this period (wetter than average) may 50 MADRONO [Vol. 47 Fig. 4. El Dorado county, 1899. Forest logged out 5-6 years ago. Sugar pine (Pinus lambertiana), large ponderosa pine {Pinus ponderosa), Douglas-fir {Pseudotsuga menziesii) taken out, Kellogg oak (Quercus kelloggii) and incense cedar {Calocedrus decurrens), 12-25 per ha., remain (5-10 per acre). Reproduction of incense-cedar, ponderosa pine, white fir, and Douglas-fir abundant 2-9 m high (6-30 feet), 2-8 years old. Ground grazed. Half mile south of Blairs Mill at Sly Park. Humas 4-10 cm deep (2-4 inches). Soil deep brown, sandy loam. have also led to an increase in tree densities in Si- erra Nevada forests. The management of Quercus kelloggii is receiv- ing increased attention in the Sierra Nevada Frame- work Project (SNFP) Environmental Impact State- ment because several rare species such as the Cal- ifornia spotted owl and Pacific fisher use this spe- cies for foraging and denning habitat (USDA 2000). Quercus kelloggii is shade intolerant, and therefore, would have difficulty living in areas dominated by large mixed conifers because it can be over-topped and killed which is one explanation of why it was not recorded in any large mixed co- nifer stands. Quercus kelloggii did contribute to 6 percent of average stand stocking on the less stocked "average" mixed conifer stands because these stands were composed by smaller trees, and therefore, more site resources were probably avail- able for the oaks. There was a great deal of variability in snag den- sities in the stands, particularly in the red fir forest type. Tree density was also much higher in the red fir forest type when compared to the mixed conifer forests. Snag basal area was almost identical in the large mixed conifer stands and red fir stands (4.6 m^/ha and 4.5 m^/ha, respectively). Snag densities found in the large mixed conifer stands are on the lower end of the current requirements for California spotted owls (Vemer et al. 1992). Very little snag information exists for red fir forests making it dif- ficult to compare this historic data to contemporary data. The average basal area recorded in the mixed conifer stands is high, even for those labeled as "average." The SNFP Environmental Impact Re- port is defining desired conditions in mixed conifer forests as having basal areas below 70 m^/ha (300 ft^/acre). The large mixed conifer stands Sudworth inventoried had over three times this basal area and the "average" mixed conifer stands had double the 2000] basal area. Some areas of mixed conifer forest in the Sierra Nevada have the abihty to produce much larger trees. The average quadratic mean diameter of the large mixed conifer stands from this study (110 cm) is equal to those recorded in the 8 mixed conifer stands in the southern Sierra Nevada (110 cm) for all trees greater than 30.5 cm DBH. (Stephens and Elliott-Fisk 1998). Omitting Sequoiadendron gi- ganteum (Lindley) Buchholz (giant sequoia) data from the four Sequoiadendron giganteum -mixed conifer stands in the southern Sierra Nevada pro- duced an average DBH of the remaining trees of 111 cm which is also very similar to those recorded above. Average tree density was higher in the southern Sierra Nevada when compared to the central and northern Sierra (278 trees/ha compared to 235 trees/ha, respectively). Average stand basal area was also higher in the mixed conifer stands from the southern Sierra Nevada when compared to the large mixed conifer stands from this study (27 1 wri ha versus 215 m^/ha, respectively). Since the av- erage DBH was equal in the mixed conifer stands the increase in basal area is a result of increased stocking in the southern Sierra Nevada (18 percent higher). Abies concolor was very rare in the "average" mixed conifer stands but was the most common tree in the large mixed conifer stands. Low amounts of Abies concolor in the "average" mixed conifer stands may have occurred because these stands were probably less developed (younger) or they may have been in drier locations which would have favored pines over true fir species. In the southern Sierra Nevada Abies concolor contributed to 40 percent of average stand stocking and 28 percent of average stand basal area (Stephens and Elliott- Fisk 1998). In the large mixed conifer stands in this study, Abies concolor contributed to 46 percent of average stand stocking and 34 percent of average stand basal area indicating that Abies concolor was slightly more conmion in the central and northern Sierra Nevada stands sampled by George Sudworth. Pinus lambertiana was much more common in the southern Sierra Nevada when compared to the central and northern Sierra Nevada (19 percent of stocking, 36 percent of basal area versus 12 percent of stocking, 16 percent of basal area, respectively). This difference can be partially explained by the presence of Pseudotsuga menziesii in relatively large amounts (16 percent of stocking, 24 percent of basal area) in the northern Sierra Nevada where- as Pseudotsuga menziesii is not native to the south- em Sierra Nevada. Both Pseudotsuga menziesii and Pinus lambertiana are classified as shade interme- diate (in between shade tolerant and shade intoler- ant) and therefore, Pseudotsuga menziesii may have occupied areas that Pinus lambertiana could have also dominated. 51 Conclusion The mixed conifer stands sampled by George Sudworth in 1899 were dominated by large trees at relatively low densities. Shade intolerant species, particularly Pinus ponderosa, dominated the "av- erage" mixed conifer stands whereas the large mixed conifer stands were composed of shade tol- erant, intermediate, and shade intolerant species. Early harvesting operations removed the major- ity of the economically viable species (Pinus spp. and Pseudotsuga menziesii) and left a large amount of Calocedrus decurrens and Abies concolor. This practice coupled with fire suppression policies ini- tiated at the beginning of the 20''' century promoted the establishment and growth of shade tolerant spe- cies. There was a large amount of variability in snag densities, particularly in the red fir stands. The red fir stands had the highest tree densities and Abies magnifica dominated in these stands. Acknowledgments I posthumously thank George Sudworth and his inven- tory crew for collecting the original data analyzed in this paper. I am grateful to Craig Olsen for introducing me to George Sudworth's field notebooks. I wish to give special recognition to Norma Kobzina, Librarian at the University of California Biosciences and Natural Resources Library, for her assistance on this project. Literature Cited BoNNiCKSEN, T. M. AND E. C. Stone. 1982. Reconstruction of a presettlement giant sequoia-mixed conifer forest community using the aggregation approach. Ecology 63(4): 1134-1 148. Caprio, a. C. AND T. W. Swetnam. 1995. Historic fire regimes along an elevational gradient on the west slope of the Sierra Nevada, California, in J. K. Brown, (ed.). Proceedings: Symposium on fire in wil- derness and park management. USDA Forest Service Gen. Tech. Rep. INT-320. Intermoutain Research Sta- tion, Ogden, UT. Gutierrez, R.J., J. Verner, K. S. McKelvey, B. R. Noon, G N. Steger, D. R. Call, W. S. LaHaye, B. B. Bingham, and J. S. Senser. 1992. Habitat relations of the California spotted owl. USDA Forest Service Gen. Tech. Rep. PSW-133. Pacific Southwest Re- search Station, Albany, CA. Hartesveldt, R. J. AND H. T. Harvey. 1967. The fire ecology of sequoia regeneration. Proceedings of the Tall Timbers Fire Ecology Conference, Tallehassee, FL. 7:65-77 KiLGORE, B. M. AND D. Taylor. 1979. Fire history of a sequoia mixed conifer forest. Ecology 60(1): 129- 142. Laudenslayer, W. F. and H. H. Darr. 1990. Historical effects of logging on forests of the Cascade and Sierra Nevada Ranges of California. Transactions of The Western Section of the Wildlife Society 26:12-23. Leopold, S. A., S. A. Cain, C. A. Cottam, 1. N. Gabriel- son, AND T L. Kimball. 1963. Wildlife management in the national parks. American Forestry 69:32-35, 61-63. STEPHENS: MIXED CONIFER AND RED FIR FOREST STRUCTURE AND USES IN 1899 52 McKelvey, K. S. and J. D. Johnston. 1992. Historical perspectives on forests of the Sierra Nevada and the Transverse Ranges of southern California: Forest con- ditions at the turn of the century, in J. Vemer et al. (eds.), USDA Forest Service Gen. Tech. Rep. PSW- 133. Pacific Southwest Research Station, Albany, CA, Millar, C. I. and W. B. Woolfenden. 1999. The role of climate change in interpreting historical variability. Ecological Applications, vol. 9:1207-1216. Parsons, D. J. and S. H. DeBendeettl 1979. Impact of fire suppression on a mixed-conifer forest. Forest Ecology and Management 2(l):21-33. RuNDEL, P W., D. J. Parsons, and D. T. Gordon. 1977. Montane and subalpine vegetation of the Sierra Ne- vada and Cascade Ranges, in M. F. Barbour and J. Maor (eds.). Terrestrial vegetation of California, John Wiley and Sons. Stephens, S. L. 1998. Effects of fuels and silvicultural treatments on potential fire behavior in mixed conifer forests of the Sierra Nevada, CA. Forest Ecology and Management 105:21-34. , D. DuLiTZ, and R. E. Martin. 1999. Giant se- quoia regeneration in group selection openings of the southern Sierra Nevada. Forest Ecology and Manage- ment 120:89-95. AND D. E. Elliott-Fisk. 1998. Sequoiadendron giganteum-mixed conifer forest structure in 1900- 1901 from the southern Sierra Nevada, CA. Madrorio, vol. 45(3):221-230. SuDWORTH, J. B. 1899. Unpublished field note books of Sierra Nevada forest reserve inventory. (University of California, Berkeley, Bioscience and Natural Re- sources Library) . 1900. Stanislaus and Lake Tahoe forest reserves, California, and adjacent territory. In Annual reports of the Department of the Interior, 21st report of the U.S. Geological Survey, 56th Congress, 2nd session. [Vol. 47 senate document #3. Washington D.C. Government printing office. SwETNAM, T. W, C. H. Baisan, P M. Brown, A. C. Ca- PRio, AND R. ToucHAN. 1990. Late Holocene fire and climate variability in giant sequoia groves. Bulletin of the Ecological Society of America 71(2): 342. , Baisan, C. H., Caprio, A. C, Touchan, R., and P. M. Brown. 1992. Tree ring reconstruction of giant sequoia fire regimes. Final report to Sequoia-Kings Canyon and Yosemite National Parks. Cooperative agreement DOI 8018-1-002, Laboratory of tree ring research. University of Arizona, Tucson, AZ. United States Department of Agriculture. 2000. In Press. United States Forest Service Sierra Nevada Framework Project Draft Environmental Impact Statement. Pacific Southwest Region, Mare Island, CA. VAN Wagtendonk, J. W. 1996. Use of a deterministic fire growth model to test fuel treatments. Sierra Nevada Ecosystem Project, final report to congress, vol. II, Assessments and Scientific Basis for Management Options. Davis: University of California, Centers for Water and Wildland Resources. Vankat, J. L. AND J. Major. 1978. Vegetation changes in Sequoia National Park, California. Journal of Bioge- ography 5:377-402. Verner, J., K. S. McKelvey, B. R. Noon, R. J. Gutier- rez, G. I. Gould, and T. W. Beck. 1992. The Cali- fornia Spotted Owl: A technical assessment of its cur- rent status. USDA Forest Service Gen. Tech. Rep. PSW-133. Pacific Southwest Research Station, Alba- ny, CA. Weatherspoon, C. p. and C. N. Skinner. 1996. Fire sil- viculture relationships in Sierra forests. Sierra Nevada Ecosystem Project, final report to congress, vol. II, Assessments and Scientific Basis for Management Options. Davis: University of California, Centers for Water and Wildland Resource. MADRONO Madrono, Vol. 47, No. 1, pp. 53-60, 2000 OLD-GROWTH FOREST ASSOCIATIONS IN THE NORTHERN RANGE OF COASTAL REDWOOD Thomas M. Mahony College of Natural Resources and Sciences, Humboldt State University, Areata, CA 95521 John D. Stuart^ Department of Forestry, Humboldt State University, Areata, CA 95521, (707) 826-3823 Abstract Old-growth Sequoia sempervirens (D. Don) Endl. (redwood) forests occurring in northwestern Cali- fornia and southwestern Oregon were classified and described using data from 206 systematically placed plots. Data were collected from Jedediah Smith Redwoods State Park, Del Norte Coast Redwoods State Park, northern Redwood National Park, and the southwestern portion of the Siskiyou National Forest. Plot data were analyzed using TWINSPAN and polar ordination. Six associations within the redwood series were classified: Sequoia sempervirenslPolystichum munitum (Kaulf) C. Presl (SESE/POMU), Se- quoia sempervirens-Pseudotsuga menziesii (Mirbel) Franco/Rhododendron macrophyllum D. Don (SESE- PSME/RHMA), Sequoia sempervirens-Tsuga heterophylla (Raf.) Sarg./Vaccinium ovatum Pursh (SESE- TSHEA^AOV), Sequoia sempervirens-Tsuga heterophyllalPolystichum munitum (SESE-TSHE/POMU), Sequoia sempervirens-Tsuga heterophyllalRubus spectabilis Pursh (SESE-TSHE/RUSP), and Sequoia sempervirens-Alnus rubra Bong./Rubus spectabilis (SESE-ALRU/RUSP). Discriminant analysis was used to assess the relationships between abiotic site variables and classified fioristic associations. Elevation and coastal proximity explained 81.1 percent of the variation among associations. Aspect and topographic position explained 14.2 percent of the remaining variation. Moisture was the primary environmental variable controlling the distribution of classified forest associations. Sequoia sempervirens (D. Don) Endl. (redwood) forests are endemic to coastal margins and mesic inland sites from central California to southern Or- egon. Along this broad latitudinal gradient, S. sem- pervirens is limited to a narrow belt 10 to 50 ki- lometers wide (Roy 1966; Fox 1989). The extreme northern range of S. sempervirens has not been ad- equately classified and described. Vast tracts of old- growth forest in Jedediah Smith Redwoods State Park, Del Norte Coast Redwoods State Park, and northern sections of Redwood National Park have been virtually ignored in the S. sempervirens liter- ature. The difficult access, steep terrain, and huge volume of coarse woody debris characterizing in- terior portions of these parks may explain the dearth of botanical information in the region. As a result of this relative isolation, these parks contain some of the most primeval and undisturbed old- growth (Helms 1998) redwood vegetation in exis- tence. Southwestern Siskiyou National Forest con- tains a patchy network of old-growth representing the northernmost natural S. sempervirens stands. Since they exist at the terminus of the redwood range, these stands are ecologically significant. They may give insight into processes affecting oth- er parts of the range, including gradients in soil moisture and temperature that affect species com- position and stand dynamics. ' To whom correspondence should be addressed. Methods Study Area. The northern range of redwood, as defined in this study, includes Jedediah Smith Red- woods State Park, Del Norte Coast Redwoods State Park, and northern Redwood National Park, all lo- cated in northern California, and portions of Sis- kiyou National Forest located in southwestern Or- egon. It extends from 41°47'N to 42°10'N, and 124°4'W to 124°12'W. The study area is topograph- ically diverse — elevations range from sea level to over 490 m. Rocks of the Franciscan Formation, a subduction complex consisting of accreted frag- ments of oceanic crust and forearc sediments, un- derlay most of the region (Aalto and Harper 1989). Soils were mapped as predominantly Melbourne and Empire series by the California State Cooper- ative Soil- Vegetation Survey (Smith et al. 1977; Delapp et al. 1978). Crescent City, CA is the closest weather station to the study area. Precipitation data (1948-2000) indicated that maximum precipitation fell during December and January, averaging 27.7 cm and 29.6 cm, respectively. The least amount fell during July and August, averaging 1.0 cm and 2.0 cm, respec- tively. Annual average precipitation was 168.1 cm. The highest mean temperatures occurred in Cres- cent City during August and September, at 14.9°C and 14.6°C, respectively. The lowest mean temper- atures occurred in December and January, at 8.8°C 54 MADRONO [Vol. 47 Table 1 . Cover Abundance Scale and Midpoints used IN Ocular Estimates. C_ovcr class rs-£ingc oi cover \ /C) V^lass IlllCipuiIlls \ /V } 8 75-100 87.5 7 50-75 62.5 6 25-50 37.5 5 5-25 15 4 1-5 3 3 0.1-1 0.6 2 0.01-0.1 0.06 1 0.001-0.01 0.006 and 8.7°C, respectively (Western U.S. Climate His- torical Summaries 2000). Study area vegetation conforms to the Society of American Foresters redwood forest cover type (Eyre 1980). East and north of the study area, the Douglas-fir forest cover type dominates. Sampling. For Jedediah Smith Redwoods State Park, Del Norte Coast Redwoods State Park, and Redwood National Park, old-growth forest was identified on 1:12,000 color infrared aerial photo- graphs. Two hundred plots were stratified based on three elevation classes (0-105 m, 106-215 m, and >215 m), and placed onto USGS topographic quad- rangles in a systematic grid 485 meters apart. For Siskiyou National Forest, four old-growth polygons were identified on maps obtained from the USDA Forest Service GIS database, and trans- ferred to 1:24,000 USGS topographic quadrangles. Five plots were placed in each polygon via a sys- tematic sampling grid, with plot spacing propor- tional to polygon area. Riparian zones within Jedediah Smith Redwoods State Park were sampled separately to best char- acterize this unique and diverse vegetation. Fifteen sample plots were systematically placed approxi- mately 1500 m apart (total stream length/ 15) along Cedar Creek, Mill Creek, Clarks Creek, and several other unnamed perennial drainages within park boundaries. Of 235 plots slated for sampling, 206 were even- tually field checked. The remaining plots were not sampled because of difficult or dangerous access, or the plot was not in an old-growth forest. The 206 circular 0.05 ha (500 m^) plots were thoroughly searched and all vascular plant species identified and recorded with an ocular cover estimate using a modified Braun-Blanquet cover abundance scale (Table 1; Mueller-Dombois and Ellenberg 1974). Tree species were tallied based on stem density in three height classes: 0-3 m, 3-10 m, and >10 m. Basal area, taken from plot center, was estimated using a "cruise angle" sighting device for canopy (dominant, co-dominant, and intermediate crown classes) species. Elevation was determined with a pocket altimeter and topographic map. Slope angle was recorded in percent using a clinometer. Aspect was assessed with a hand compass. Distance from the ocean was estimated using a topographic map. Topographic position was recorded for each plot. Data Analysis. Two- Way Indicator Species Anal- ysis (TWINSPAN) (Hill 1979) was used to simul- taneously classify species and samples. Only spe- cies occurring in greater than 5 percent of plots were used in the analysis (Gauch 1982). Plots were analyzed in TWINSPAN with species cover cut levels of 0.6, 3.0, 15.0, 37.5, 62.5, and 87.5 percent. The 15.0 and 37.5 cut levels were weighted to em- phasize dominance (Stuart et al. 1996). TWIN- SPAN groupings were analyzed using a polar (Bray-Curtis) ordination, to further analyze and re- fine the TWINSPAN output. Species richness was determined by randomly selecting 5 plots from each association and calculating the mean number of species per plot (Stuart et al. 1996). In addition, stem density per hectare in three height classes and canopy species basal area were averaged for each association. Discriminant analysis was performed using NCSS 2000 (Hintze 1998) to relate floristic asso- ciations with abiotic site characteristics. Elevation, slope angle, coastal proximity, and a Moisture Equivalency Index (MEI) were used as abiotic vari- ables in the discriminant analysis. The MEI was adapted from Sawyer and Thomburgh (1974) and Matthews (1986). It incorporates topographic po- sition and aspect, two variables important to soil moisture. A lower index number (1-15) assumes greater soil moisture available to plants. Results and Discussion TWINSPAN and polar ordination analysis pro- duced six groups that were interpreted as associa- tions (Fig. 1). Groups were consistent with vege- tation units observed in the field. All associations were in the Sequoia sempervirens series, with Pseu- dotsuga menziesii (Mirbel) Franco, Tsuga hetero- phylla (Raf.) Sarg., and Alnus rubra Bong, sub-se- ries. The first TWINSPAN division separated groups based on understories dominated by either Vaccinium ovatum Pursh or Polystichum munitum (Kaulf) C. Presl. Within these broad groupings, subsequent TWINSPAN division levels reflected groupings based on other indicator understory spe- cies such as Lithocarpus densiflorus (Hook. & Am.) Rehder, Rhododendron macrophyllum D. Don, and Rubus spectabilis Pursh. The following association descriptions are presented from rela- tively dry types to wet types. A more detailed treat- ment of the associations can be found in Mahony (1999). The Sequoia sempervirens-Pseudotsuga menzie- siil Rhododendron macrophyllum Association. Total vegetation cover averaged 85 percent, and total overstory cover averaged 68 percent. Over- stories were dominated by Sequoia sempervirens and Pseudotsuga menziesii, with mean cover values 2000] MAHONY AND STUART: OLD-GROWTH FOREST ASSOCIATIONS 55 Level Understories dominated by Vaccinium ovatum Understories dominated by Polystichum munitum Tsuga heterophylla 1 sparse 206 Tsuga heterophylla dense 77 Inland/non-riparian Lithocarpus densiflorus common Coastal/riparian 129 Lithocarpus densiflorus uncommon AInus rubra uncommon 112 AInus rubra common 17 27 50 SESE-PSME/RHMA SESE-TSHEA/AOV 64 SESE/POMU 48 10 SESE-TSHE/POMU SESE-TSHE/RUSP SESE-ALRU/RUSP Fig. 1. Dendrogram of TWINSPAN classification. Numbers beneath lines represent the number of plots prior to division. Numbers above association acronyms are the number of plots in each classified association. Association acronyms are: SESE-PSME/RHMA = Sequoia semperx'irens-Pseudotsuga menziesiil Rhododendron macrophyllum, SESE-TSHEA^AOV = Sequoia sempervirens-Tsuga heterophyllalVaccinium ovatum, SESE/POMU = Sequoia sem- pervirens/ Polystichum munitum, SESE-TSHE/POMU — Sequoia sempervirens-Tsuga heterophylla! Polystichum muni- tum, SESE-TSHE/RUSP = Sequoia semper\'irens-Tsuga heterophyllalRubus spectabilis, SESE-ALRU/RUSP = Se- quoia sempervirens-Alnus rubral Rubus spectabilis. of 43 and 31 percent, respectively, and mean con- stancies of 96 and 100 percent, respectively (Table 2). Tsuga heterophylla was occasionally present but contributed minimal cover. Lithocarpus densiflorus dominated the sub-canopy. Basal area averaged 123 m2/ha (Table 3). The shrub layer was extremely dense. Vaccinium ovatum and Rhododendron macrophyllum dominat- ed, with mean cover values of 47 and 35 percent, respectively, and mean constancies of 100 percent each. Berberis nervosa Pursh, Gaultheria shallon Pursh, Rhamnus purshiana DC, and Vaccinium parvifolium Smith each had greater than 30 percent constancy but less than 3 percent cover. The herb layer was virtually absent. Polystichum munitum was the most dominant species in this lay- er with 7 percent cover and 93 percent constancy. Disporum hooker (Torrey) Nicholson, Galium tri- florum Michaux, Oxalis oregana Nutt., Trillium ovatum Pursh, and Viola sempervirens E. Greene were common but contributed negligible cover. The Sequoia sempervirens-Pseudotsuga menzie- siil Rhododendron macrophyllum association was generally found on upper slopes and ridges in Sis- kiyou National Forest and Del Norte Coast Red- woods State Park. Elevations ranged from 58-470 m, averaging 312 m. Distance from the ocean av- eraged 8.5 km. Slopes averaged 43 percent, and Moisture Equivalency Index (MEI) scores averaged 9.8. Species richness averaged 13.6 species (Table 4). Vegetation dynamics. Lithocarpus densiflorus and Sequoia sempervirens dominated reproduction. Veirs (1979) suggested S. sempervirens and L. den- siflorus were components of the "climax" vegeta- tion and would remain in the stand regardless of disturbance such as fire. The presence of S. sem- pervirens in all height classes represents an uneven age structure for redwood. Pseudotsuga menziesii is a serai species that will disappear from stands without major disturbance (Daubenmire 1975; Veirs 1979; Eyre 1980). Low P. menziesii stem densities for the 0-3 m and 3-10 m height classes, and high density of trees >10 m, suggested that a cohort resulted from disturbance, and additional disturbance will be necessary for continued pres- ence of P. menziesii in this association. Relationships to previous classifications. The Se- quoia sempervirens-Pseudotsuga menziesiil Rhodo- dendron macrophyllum association closely resem- bled the midslope stands encountered by Dymess et al. (1972) in Wheeler Creek Research Natural Area, and Tanoak-coast redwood association stands described by Atzet and Wheeler (1984) for south- western Oregon. Other similar types include the Se- quoia sempervirens-Pseudotsuga menziesiilVaccin- ium ovatum association described by Matthews 56 MADRONO [Vol. 47 Table 2. Average Cover and Constancy for Species Used in TWINSPAN Analysis. Species reported are those with >50 percent constancy. See Figure 1 for plant association acronyms. Cov = average cover (%). Con = constancy (%). SESE- PSME/ SESE-TSHE/ SESE-TSHE/ SESE-TSHE/ SESE- RHMA VAOV SESE/POMU POMU RUSP ALRU/RUSP Cov Con Cov Con Cov Con Cov Con Cov Con Cov Con Maianthemum dilatatum Menziesia ferruginea Pseudotsuga menziesii 31 Rhododendron macrophyllum 35 Viola sempervirens <1 Disporum hookeri Lithocarpus densiflorus 45 Vaccinium ovatum 47 Tsuga heterophylla Trillium ovatum <1 Vaccinium parvifolium 1 Blechnum spicant Polysdchum munitum 7 Sequoia sempervirens 43 Oxalis oregana <1 Disporum smithii Rhamnus purshiana Gaultheria shallon 2 Galium triflorum Vancouveria hexandra Dryopteris expansa Rubus spectabilis Acer circinatum Acer macrophyllum Adiantum aleuticum Corylus cornuta Ribes bracteosum Asarum caudatum Athyrium filix-femina Tolmiea menziesii Alnus rubra Claytonia sibirica Marah oreganus Polypodium scouleri Rubus parviflorus Sambucus racemosa Stachys ajugoides 100 17 78 100 13 76 74 <1 90 <1 61 <1 54 <1 66 100 21 94 16 95 100 51 98 16 97 41 100 24 75 93 <1 98 <1 98 70 2 70 3 73 3 72 4 73 93 12 100 55 100 96 37 100 60 100 52 <1 78 13 98 78 2 74 5 80 <1 54 <1 61 <1 55 <1 52 3 69 8 57 14 98 5 57 2 70 39 88 21 86 <1 69 <1 86 5 94 3 86 1 1 96 10 100 <1 50 67 100 30 100 24 100 53 98 22 86 36 80 9 100 16 100 <1 67 <1 100 <1 60 1 57 1 75 6 71 10 80 <1 57 <1 60 <1 57 <1 10 2 75 1 86 2 50 5 63 25 100 22 100 19 71 10 57 <1 71 14 57 1 57 <1 71 <1 70 <1 58 2 100 2 60 <1 71 2 60 38 100 <1 80 5 80 <1 50 3 90 3 70 <1 90 (1986) and the Sequoia sempervirens! Arbutus men- ziesii Pursh association described by Lenihan (1986). This association might be considered an ex- tension of the Pseudotsuga-h2ii&w ood forests de- scribed by Sawyer et al. (1977). The Sequoia sempervirens-Tsuga heterophylla/ Vaccinium ovatum Association. Total vegetation cover averaged 88 percent, and total overstory cover averaged 74 percent. Sequoia sempervirens, Tsuga heterophylla, and Pseudotsu- Table 3. Mean Basal Area (mVha) for Canopy Species by Association. SESE- PSME/RHMA SESE- TSHE/VAOV SESE/POMU SESE- TSHE/POMU SESE- TSHE/RUSP SESE- ALRU/RUSP Sequoia sempervirens 86.0 114.0 165.0 170.0 73.0 87.0 Pseudotsuga menziesii 37.0 21.0 10.0 2.0 2.0 7.0 Tsuga heterophylla 0.4 23.0 15.0 23.0 11.0 0.0 Picea sitchensis 0.0 0.0 0.0 4.0 6.0 10.0 Abies grandis 0.0 3.0 1.0 0.2 0.0 3.0 Total basal area 123.4 161.0 191.0 199.2 92.0 107.0 2000] MAHONY AND STUART: OLD-GROWTH FOREST ASSOCIATIONS 57 Table 4. Environmental Characteristics, Tree Density in Three Height Classes, and Species Richness for Each Forest Association. PSME/RHMA TSHE/VAOV SESE/POMU SESE- TSHE/POMU SESE- TSHE/RUSP ALRU/RUSP Elevation (m) 312.0 161.0 143.0 114.0 67.0 136.0 Distance (km) 8.5 7.4 6.4 5.5 6.8 3.6 olope \ /o) A1 Q jO.Z jD.Z 'XA Q JO.U ZIQ 7 MEI (1-15) 9.8 9.0 7.5 7.0 1.3 6.7 Stems/ha: 0-3 m 71.8 87.6 114.0 54.0 68.6 74.0 3-10 m 127.6 76.8 86.6 66.2 51.6 168.0 >10 m 180.8 206.0 172.4 165.8 72.0 170.0 Sp. Richness 13.6 16.6 19.0 15.6 26.8 19.4 ga menziesii dominated the canopy, with mean cov- ers of 37, 41, and 17 percent, respectively, and mean constancies of 100, 100, and 78 percent, re- spectively (Table 2). Abies grandis (Douglas) Lind- ley appeared occasionally in the canopy. Lithocar- pus densiflorus was common in the subcanopy. Ba- sal area averaged 161.0 m^/ha (Table 3). The shrub layer was dense. Vaccinium ovatum dominated, averaging 5 1 percent cover and 98 per- cent constancy. Rhododendron macrophyllum had 13 percent cover and 76 percent constancy. Ber- beris nervosa, Gaultheria shallon and Vaccinium parvifolium each had greater than 40 percent con- stancy but less than 2 percent cover. Corylus cor- nuta Marsh, and Rhamnus purshiana occurred spo- radically. The sparse herb layer was dominated by Polys- tichum munitum, averaging 12 percent cover and 100 percent constancy. The Sequoia sempervirens-Tsuga heterophyllal Vaccinium ovatum association was usually found on inland upper slopes and ridges in Jedediah Smith Redwoods State Park. Elevations ranged from 40- 460 m, averaging 161 m. Distance inland averaged 7.4 km. Slopes averaged 36 percent, and MEI scores averaged 9. Species richness averaged 16.6 species (Table 4). Vegetation dynamics. Tsuga heterophylla and Lithocarpus densiflorus dominated reproduction. Tsuga heterophylla seedlings were particularly abundant on downed logs. Combs (1984) noted a similar pattern of T. heterophylla regeneration in the Little Lost Man Creek Research Natural Area in Redwood National Park. He suggested that few seedlings would reach maturity because of vulner- ability to fire and disease. Daubenmire (1975) noted extensive T. heterophylla in all size classes in Je- dediah Smith Redwoods State Park, but believed the species would decline without disturbance. Veirs (1979) suggested that light ground fires, un- affecting the canopy, will favor T. heterophylla re- generation. The high density of T. heterophylla and the complete absence of P. menziesii seedlings sug- gested a light fire regime, sufficient for S. semper- virens and T. heterophylla regeneration, but not for regeneration of P. menziesii. Relationships to previous classifications. The Se- quoia sempervirens-Tsuga heterophyllalVaccinium ovatum association was unique compared to other redwood types described in the literature due to the importance of Tsuga heterophylla. While other red- wood classifications have noted the presence of T. heterophylla (Dyrness et al. 1972; Atzet and Whee- ler 1984; Lenihan 1986), none have shown such dominance by this mesic conifer. The Sequoia sem- pervirenslBerberis nervosa association described by Lenihan (1986), and the Tsuga phase of the Pseudotsuga-hdiidyNood forests described by Saw- yer et al. (1977) were similar in composition to this association. The Sequoia sempervirenslPolystichum munitum Association. Total vegetation cover averaged 90 percent. Total overstory cover averaged 76 percent. Sequoia sem- pervirens dominated the canopy with 60 percent cover and 100 percent constancy (Table 2). Tsuga heterophylla was common, and Pseudotsuga men- ziesii appeared occasionally in the canopy. Abies grandis, Cupressus lawsoniana A. Murray and Um- bellularia californica (Hook. & Am.) Nutt. oc- curred sporadically, contributing minimal cover. Lithocarpus densiflorus was ubiquitous in the sub- canopy. Basal area averaged 191.0 m^/ha (Table 3). Vaccinium ovatum dominated the relatively sparse shrub layer, averaging 16 percent cover and 97 percent constancy. Gaultheria shallon. Rhodo- dendron macrophyllum, and Vaccinium parvifolium each had greater than 40 percent constancy but less than 5 percent cover. Acer circinatum Pursh, Ber- beris nervosa, Corylus cornuta, and Rubus spec- tabilis occurred sporadically, contributing minimal cover. Herbaceous cover and species diversity was moderately high. Polystichum munitum dominated, averaging 55 percent cover and 100 percent con- stancy. Oxalis oregana was extremely conmion. The Sequoia sempervirenslPolystichum munitum 58 MADRONO [Vol. 47 association was found throughout the study area, generally on lower and middle slopes at moderate distances from the ocean. Elevations ranged from 21-369 m, averaging 143 m. Distance from the ocean averaged 6.4 km. Slopes averaged 36 per- cent, and MEI scores averaged 7.5. Species rich- ness averaged 19 species (Table 4). Vegetation Dynamics. Lithocarpus densiflorus and Sequoia sempervirens dominated reproduction. The moderate levels of Abies grandis, Tsuga het- erophylla, and L. densiflorus reproduction may be indicative of the light fire regime in intermediate to mesic sites referred to by Veirs (1979). However, he noted that these species exhibited an all aged pattern and can reproduce regardless of fire. Relationships to previous classifications. The Se- quoia sempervirenslPolystichum munitum associa- tion contained elements of the Sequoia sempervi- rens! Blechnum spicant (L.) Smith association de- scribed by Lenihan (1986), though Lenihan's as- sociation appeared wetter. The dominance of Sequoia sempervirens, the sparse shrub layer, and the well-developed herb layer related this associa- tion to Becking's (1967) Redwood-oxalis alliance. The Sequoia sempervirens-Tsuga heterophyllal Polystichum munitum Association. Total vegetation cover averaged 92 percent, and total overstory cover averaged 75 percent. Sequoia sempervirens and Tsuga heterophylla dominated the canopy, with mean covers of 53 and 39 percent, respectively, and mean constancies of 98 and 88 percent, respectively (Table 2). Abies grandis, Lith- ocarpus densiflorus, Picea sitchensis (Borg.) Car- riere and Pseudotsuga menziesii occurred sporadi- cally, contributing minimal cover Thuja plicata D. Don appeared occasionally in mesic sites. Basal area averaged 199.2 m^/ha (Table 3). The shrub layer was generally not well devel- oped. Vaccinium ovatum was the most abundant shrub, averaging 14 percent cover and 98 percent constancy. Menziesia ferruginea Smith, Rubus spec- tabilis, Vaccinium parvifolium, Gaultheria shallon, and Rhamnus purshiana each had greater than 60 percent constancy but less than 6 percent cover The herbaceous layer was dense. Polystichum munitum dominated, averaging 67 percent cover and 100 percent constancy. Blechnum spicant and Oxalis oregana were common. The Sequoia sempervirens-Tsuga heterophyllal Polystichum munitum association was generally found at lower slopes and elevations, especially in southwestern areas of Jedediah Smith Redwoods State Park exposed to maritime influence. Eleva- tions ranged from 40-274 m, averaging 114 m. Distance inland averaged 5.5 km. Slopes averaged 35 percent, and MEI scores averaged 7. Species richness averaged 15.6 species (Table 4). Vegetation Dynamics. Tsuga heterophylla and Sequoia sempervirens dominated reproduction. Se- quoia sempervirens had fewer stems in the lower height classes relative to T. heterophylla, but the longevity and resilience of S. sempervirens makes abundant individuals in the reproduction layers un- necessary to ensure continued dominance. Relationships to previous classifications. The Se- quoia sempervirens-Tsuga heterophyllal Polysti- chum munitum association, like Sequoia sempervi- rens-Tsuga heterophyllalVaccinium ovatum, ap- peared unlike any previously described redwood types. It was similar in many respects to the mesic Tsu gal Polystichum association described by Frank- lin and Dymess (1973) for Oregon Coast Range forests in the Tsuga heterophylla Zone. Addition- ally, it contained elements of the Tsuga-picealOplo- panax horridumlAthyrium filix-femina association of Picea sitchensis Zone forests described by Franklin and Dymess (1973). It related tangentially to Lenihan's (1986) Sequoia sempervirensi Blech- num spicant association. The Sequoia sempervirens-Tsuga heterophyllal Rubus spectabilis Association. Total vegetation cover averaged 94 percent, and total overstory cover averaged 55 percent. Sequoia sempervirens and Tsuga heterophylla were canopy dominants, averaging 22 and 21 percent cover, re- spectively. Both species had 86 percent constancy (Table 2). Picea sitchensis and Thuja plicata were occasional to common in mesic sites. Pseudotsuga menziesii occurred sporadically. Acer macrophyl- lum was common, especially near stream channels. Lithocarpus densiflorus was common in the sub- canopy. Alnus rubra and Sambucus racemosa L. appeared occasionally. Basal area averaged 92.0 m2/ha (Table 3). Rubus spectabilis dominated the dense shrub lay- er, averaging 25 percent cover and 100 percent con- stancy. Acer circinatum and Corylus comuta were abundant in this layer having 71 percent and 57 percent constancy and 19 percent and 14 percent cover, respectively. Other conamon shrubs having greater than 40 percent constancy but less than 6 percent cover included Gaultheria shallon, Menzie- sia ferruginea, Rhamnus purshiana, Ribes bracteo- sum Douglas, Rubus parviflorus, Vaccinium ova- tum, and V. parvifolium. The herbaceous layer was dense and floristically diverse. Polystichum munitum dominated with 31 percent cover and 100 percent constancy. Oxalis oregana and Blechnum spicant were abundant. The Sequoia sempervirens-Tsuga heterophyllal Rubus spectabilis association was restricted to in- terior perennial drainages in Jedediah Smith Red- woods State Park. Elevations ranged from 37-122 m, averaging 67 m. Distance from the ocean aver- aged 6.8 km. Slopes averaged 38 percent, and MEI scores averaged 1.3. Species richness averaged 26.8 species (Table 4). Vegetation Dynamics. Tsuga heterophylla and Lithocarpus densiflorus dominated reproduction. 2000] MAHONY AND STUART: OLD-GROWTH FOREST ASSOCIATIONS Table 5. Standard Canonical Coefficients Used in Discriminant Analysis. 59 Variable Variate 1 Variate 2 Variate 3 Variate 4 Elevation -0.874009 0.717388 0.045912 -0.376631 Slope 0.189664 0.275203 0.477507 0.858742 Distance -0.828487 -0.300679 -0.599034 0.465420 MEI -0.376680 -1.100974 0.339726 0.143424 Riparian conditions produced the wettest and most floristically diverse association encountered in the study area. Conifer basal area was greatly reduced compared to other associations. Sequoia sempervi- rens attained its lowest basal area, but still domi- nated conifer basal area. The streamside environ- ment allowed mesic woody species such as Acer macrophyllum, A. circinatum, Corylus cornuta, and Rubus spectabilis to thrive. Relationships to previous classifications. The Se- quoia sempervirens-Tsuga heterophyllal Rubus spectabilis association appeared much wetter than any redwood association previously described. It shared many of the same riparian components, such as high cover of herbaceous and hardwood species, described by Dyrness et al. (1972) for lower slopes in the Wheeler Creek Research Natural Area in southwestern Oregon, and appeared similar in many respects to the Tsuga heterophyllal Acer cir- cinatum! Poly stichum munitum-Oxalis oregana as- sociation described by Franklin and Dyrness (1973) for alluvial terrace vegetation in the Tsuga hetero- phylla Zone of Oregon. The absence of Sequoia sempervirens in Tsuga heterophylla Zone forests makes comparison difficult, however. The Sequoia sempervirens-Alnus rubral Rubus spectabilis Association. Total vegetation cover averaged 93 percent, and total overstory cover averaged 71 percent. Sequoia sempervirens dominated the canopy, averaging 36 percent cover and 80 percent constancy (Table 2). Picea sitchensis was common in coastal sites. Pseu- dotsuga menziesii and Abies grandis occurred spo- radically. Alnus rubra dominated the subcanopy. Acer macrophyllum, Lithocarpus densiflorus, Rhamnus purshiana, Sambucus racemosa, and Um- bellularia califomica appeared occasionally in the subcanopy. Basal area averaged 107.0 m^/ha (Table 3). Rubus spectabilis dominated the moderately dense shrub layer, averaging 22 percent cover and 100 percent constancy. Gaultheria shallon had 10 percent cover and 80 percent constancy. Other conmion shrubs included Acer circinatum, Corylus cornuta, Rubus parviflorus, R. ursinus Cham & Schldl., and Vaccinium ovatum with constancies greater than 20 percent but with less than 8 percent cover. The herbaceous layer was diverse. Polystichum munitum dominated, averaging 24 percent cover and 100 percent constancy. Oxalis oregana oc- curred sporadically, but was generally abundant when it did occur. The Sequoia sempervirens-Alnus rubralRubus spectabilis association was generally found along the Smith River, or on coastal bluffs in Del Norte Coast Redwoods State Park. Elevations ranged from 18-299 m, averaging 136 m. Distance from the ocean averaged 3.6 km. Slopes averaged 50 per- cent, and MEI scores averaged 6.7. Species rich- ness averaged 19.4 species (Table 4). Vegetation dynamics. Sequoia sempervirens dominated reproduction. Picea sitchensis was very conmion on coastal bluffs, and Pseudotsuga men- ziesii was conamon along the Smith River. Alnus rubra achieved high cover in the subcanopy. The mesic, high light environments of the Smith River floodplain and exposed coastal bluffs provided fa- vorable conditions for this shade intolerant hard- wood (Hibbs et al. 1994; Harlow et al. 1996). Ad- ditionally, natural disturbance from Smith River flooding likely enhanced the competitive ability of A. rubra, which is more tolerant of flooding and poor drainage than its associates (Hibbs et al. 1994). Tolerance of salt spray and resistance to windthrow allowed A. rubra to thrive along the coastal bluffs of Del Norte Coast Redwoods State Park. Periodic disturbances likely benefited the ser- ai Pseudotsuga menziesii. Alnus rubra showed high stem densities in the 3-10 and >10 m height class- es, but minimal density in the 0-3 m class, indi- cating many stands may be recovering from distur- bance. Relationships to previous classifications. The Se- quoia sempervirens-Alnus rubralRubus spectabilis association was similar to coastal sections of the Wildcat Hills transect described by Zinke (1977), as well as the red alder series described in Sawyer and Keeler-Wolf (1995). It should be noted that a pure Picea sitchensis forest type may exist imme- diately adjacent to the coast in Del Norte Coast Redwoods, but was not sampled. Discriminant Analysis. Discriminant analysis re- vealed that elevation, coastal proximity, and topo- graphic position/aspect (MEI) were statistically sig- nificant (P < 0.01) in discriminating among floristic associations. Elevation and coastal proximity had the greatest influence on the first discriminant func- tion (Table 5). This function explained 81.1 percent of the variation between groups. MEI had the great- est influence on the second discriminant function. 60 MADRONO [Vol. 47 which explained 14.2 percent of group variation. Together, the first two discriminant functions, influ- enced by elevation, distance to the ocean, and MEI, explained 95.3 percent of group variation. The physiographic factors influencing floristic associa- tions, in decreasing order of importance, were ele- vation, coastal proximity, and aspect/topographic position (MEI). Acknowledgments We thank John Sawyer and Larry Fox for their help in project design and manuscript review. Thanks to Jim Belsher for field assistance, and Stephen Underwood at Redwood National and State Parks for logistical support. Partial funding came from a Mclntire-Stennis grant. Literature cited Aalto, K. R. and G. D. Harper. 1989. Geologic evolu- tion of the northernmost coast ranges and western Klamath mountains, California. American Geophysi- cal Union, Washington, D.C. Atzet, T. and D. L. Wheeler. 1984. Preliminary plant associations of the Siskiyou Mountain Province. U.S.D.A. Forest Service, Pacific Northwest Region. Becking, R. W. 1967. The ecology of the coastal redwood forest and the impact of the 1964 floods upon red- wood vegetation. Final Report. NSF Grant GB No. 3468.91 p. (on file at Humboldt State University li- brary. Areata, CA). Combs, W. E. 1984. Stand structure and composition of the Little Lost Man Creek Research Natural Area, Redwood National Park. M.S. thesis, Humboldt State University, Areata, CA. Daubenmire, R. W. 1975. The community status of the coastal redwood. Sequoia sempervirens: A report pre- pared for the National Park Service, Areata, CA. DeLapp, J. A., B. F Smith, and W. R. Powell. 1978. Soil- vegetation map and legend-NVi of Klamath quadran- gle (9D-1,2). California Cooperative Soil-Vegetation Survey. U.S.D.A. Forest Service, Pacific Southwest Forest and Range Experiment Station, Berkeley, CA. Dyrness, C. T, J. F Franklin, and C. Maser. 1972. Wheeler Creek Research Natural Area, federal re- search natural areas in Oregon and Washington: A guidebook for scientists and educators. U.S.D.A. For- est Service, Pacific Northwest Forest and Range Ex- periment Station, Portland, OR. Supplement No. 1. Eyre, F. H. (ed). 1980. Forest cover types of the United States and Canada. Society of American Foresters, Washington D.C. Fox, Lawrence III. 1989. A classification, map and vol- ume estimate for the coast redwood forest in Califor- nia. Final report. Interagency agreement number 8CA52849. The Forest and Rangeland Resources As- sessment Program. California Department of Forestry and Fire Protection, Sacramento, CA. Franklin, J. F. and C. T. Dyrness. 1973. Natural vege- tation of Oregon and Washington. U.S.D.A. Forest Service General Technical Report PNW-8. Pacific Northwest Forest and Range Experiment Station, Portland, OR. Gauch, H. G. 1982. Multivariate analysis in community ecology. Cambridge University Press, New York. Harlow, W. M., E. S. Harrar, J. W. Hardin, and F. M. White. 1996. Textbook of dendrology. McGraw-Hill, Inc., New York. Helms, J. A. (ed). 1998. The dictionary of forestry. So- ciety of American Foresters, Washington D.C. HiBBS, D. E., D. S. DeBell, and R. F Tarrant. 1994. The biology and management of red alder. Oregon State University Press, Corvallis, OR. Hill, M. O. 1979. TWINSPAN: A FORTRAN program for arranging multivariate data in an ordered two-way table by classification of the individuals and attri- butes. Section of Ecology and Systematics, Cornell University. Ithaca, NY. HiNTZE, J. L. 1998. NCSS 2000. NCSS, Kaysville, UT Lenihan, J. M. 1986. The forest associations of the Little Lost Man Creek Research Natural Area, Redwood National Park, CA. M.S. thesis, Humboldt State Uni- versity, Areata, CA. Mahony, T. M. 1999. Old-growth forest associations in the northern range of redwood. M.S. thesis, Humboldt State University, Areata, CA. Matthews, S. C. 1986. Old-growth forest associations of the Bull Creek watershed Humboldt Redwoods State Park, California. M.S. thesis, Humboldt State Uni- versity, Areata, CA. Mueller-Dombois, D. and H. Ellenberg. 1974. Aims and methods of vegetation ecology. John Wiley and Sons, New York. Roy, D. F. 1966. Silvicultural characteristics of redwood. U.S.D.A. Forest Service research paper PSW-28. Pa- cific Southwest Forest and Range Experiment Station. Berkeley, CA. Sawyer, J. O. and T. Keeler-Wolf. 1995. A manual of California vegetation. California Native Plant Soci- ety, Sacramento, CA. Sawyer, J. O. and D. A. Thornburgh. 1974. Subalpine and montane forests on granodiorite in the central Klamath mountains of California. Report to Pacific Southwest Forest and Range Experiment Station and Humboldt State University, Areata, CA. (on file at Humboldt State University library. Areata, CA). Sawyer, J. O., D. A. Thornburgh, and J. R. Griffin. 1977. Mixed evergreen forest. Pp. 359-381 in M. G. Barbour and J. Major (eds.). Terrestrial vegetation of California. John Wiley and Sons, New York. Smith, B. F, J. A. Delapp, and W. R. Powell. 1977. Soil- vegetation map and legend-SVa Crescent City quad- rangle (9A-3,4). California Cooperative Soil- Vegeta- tion Survey. U.S.D.A. Forest Service, Pacific South- west Forest and Range Experiment Station, Berkeley, CA. Stuart, J. D., T. Worley, and A. C. Buell. 1996. Plant associations of Castle Crags State Park, Shasta Coun- ty, CA. Madrono 43:273-291. Veirs, S. D., Jr. 1979. The role of fire in northern coast redwood forest dynamics. Proceedings from the Sec- ond Conference on Scientific Research in the Nation- al Parks, San Francisco, CA. Western U.S. Climate Historical Summaries. 2000. Crescent City 1 N, California (042147), July 1948- April 2000. in Program for climate, ecosystem, and fire application. Desert Research Institute and USDI Bureau of Land Management, Reno, Nevada. Ac- cessed July 17, 2000. Available: www.wrcc.dri.edu. Zinke, p. J. 1977. The redwood forest and associated north coast forests. Pp. 679-698 in M.G. Barbour and J. Major (eds.). Terrestrial vegetation of California. John Wiley and Sons, New York. Madrono, Vol. 47, No. 1, pp. 61-67, 2000 ESTIMATED AGES OF SOME LARGE GIANT SEQUOIAS: GENERAL SHERMAN KEEPS GETTING YOUNGER Nathan L. Stephenson U.S. Geological Survey, Western Ecological Research Center, Sequoia and Kings Canyon Field Station, Three Rivers, CA 93271-9651 Abstract Using a method that combines information on tree size with growth rates determined from relatively short increment cores, I estimated the ages of several of the largest living Sequoiadendron giganteum (Lindley) Buchholz. Compared to the longest-lived S. giganteum known, which was at least 3266 years old, most of the large sequoias analyzed here were relatively young, with estimated ages of only 1650 to 2150 years. Thus, contrary to common supposition, the largest S. giganteum generally owe their great size to rapid growth, not to exceptional age. However, two of the largest S. giganteum were substantially older, with estimated ages of 2850 and 2890 years. There is a high probability that some S. giganteum living today are older than the oldest S. giganteum yet discovered. People have long been fascinated by the great size and longevity of Sequoiadendron giganteum (Lindley) Buchholz (giant sequoias), which grow naturally only in isolated groves on the western slope of California's Sierra Nevada. Sequoiaden- dron giganteum are the world's largest trees, reach- ing a maximum known bole volume of nearly 1500 m^ (Hartesveldt et al. 1975; Flint 1987 and in press). Precise cross-dating of tree rings on cut stumps has shown that sequoias can reach at least 3266 years in age (R. Touchan personal commu- nication), making 5. giganteum the third longest- lived, non-clonal tree species known, exceeded only by Pinus longaeva Bailey (bristlecone pine, 4844 years) of western North America's Great Ba- sin (Currey 1965) and Fitzroya cupressoides (Mo- lina) Johnston, (alerce, 3613 years) of Chile and Argentina (Lara and Villalba 1993). Here I present age estimates for some large, well- known S. giganteum, thereby addressing one of the most frequently-asked questions about famous S. giganteum — namely, "how old is this tree?" I ad- ditionally address two questions regarding S. gi- ganteum sizes and ages. First, are the largest S. gi- ganteum so massive because they are exceptionally old, as is often presumed, or because they have grown particularly rapidly? Second, are there likely to be any S. giganteum alive today that are older than the longest-lived S. giganteum yet known, which is known only from a cut stump? These questions are difficult to answer because the only way to precisely determine the age of liv- ing S. giganteum is to crossdate tree rings on in- crement cores that intersect the tree's pith (Stokes and Smiley 1968). However, the tremendous girth of large S. giganteum usually makes it impossible to reach their piths with hand-driven increment bor- ers. Power increment borers with very long bits can sometimes be used to obtain cores that reach the pith (Echols 1969; Johansen 1987), but have sev- eral disadvantages, which include unacceptably large holes left in the trees, poor quality of many of the cores extracted, and unacceptable use of noisy power tools on and around popular and fre- quently-visited S. giganteum. I therefore estimated the ages of several large S. giganteum using a method that takes advantage of information from partial increment cores (cores that fall well short of a tree's pith). The derivation and testing of the method is described in detail else- where (Stephenson and Demetry 1995). Unlike pre- vious attempts to estimate the ages of large S. gi- ganteum (e.g., Douglass 1946; Hartesveldt et al. 1975), this method has been tested on hundreds of S. giganteum stumps, does not systematically over- or underestimate tree ages, and offers confidence intervals on the final age estimates. Methods Choice of individual Sequoiadendron giganteum for analysis. The primary criteria for choosing in- dividual S. giganteum for analysis were (1) the S. giganteum were among the largest known, and (2) the cores and other data needed for age estimation were already available (that is, no S. giganteum was to be cored solely for the purpose of this study). Specifically, for a given S. giganteum to be includ- ed, original increment cores or the necessary mea- surements from those cores had to be available, along with measurements of the tree's bark thick- ness and diameter at the height at which the cores were taken. These data requirements limited the pool of S. giganteum available for analysis. While many large S. giganteum have been cored for stud- ies of human impacts (Hartesveldt 1962, 1965), ring-width chronology development (Brown et al. 1992; Hughes et al. 1996), climatic reconstructions (Hughes and Brown 1992), forest dynamics studies 62 MADRONO [Vol. 47 Table 1 . Sequoiodendron giganteum Selected for Analysis (Size Ranks and Bole Volumes are from Flint in Press and Personal Communication). Bole Size rank volume Tree name (by volume) (m^) Location General Sherman 1487 Giant Forest, Sequoia National Park Washington 2 1355 Giant Forest, Sequoia National Park General Grant 3 1320 General Grant Grove, Kings Canyon National Park Boole 7 1202 Converse Basin Grove, Giant Sequoia National Monument Grizzly Giant 27 963 Mariposa Grove, Yosemite National Park Cleveland 36 887 Giant Forest, Sequoia National Park Sentinel Not ranked 790 Giant Forest, Sequoia National Park NOTE: Future discoveries of previously unrecognized large sequoias will probably change the ranking of sequoias smaller than the Boole tree. For example, the fourteenth largest sequoia known (the Ishi Giant of Kennedy Grove) was identified only in 1993 (Willard 1994; Flint personal communication). (Stephenson 1994), and fire history reconstruction (Swetnam 1993), only a limited subset of those S. giganteum have associated records of diameter at core height. Diameter at core height is essential for age estimation (Stephenson and Demetry 1995), and cannot be estimated readily from published di- ameters at breast height of individual S. giganteum. Cores are rarely taken exactly at breast height, and sequoia bole diameter usually changes rapidly with increasing distance from breast height. The following seven large S. giganteum were se- lected for analysis (Table 1). The General Sherman, Washington, and General Grant trees are the world's three largest trees, with the General Sher- man and General Grant trees being among the most heavily visited of all S. giganteum. The Boole tree is the seventh largest, and is well-known as being the largest sequoia on lands managed by the U.S. Forest Service. The Grizzly Giant is heavily visited because of its craggy appearance and status as one of the two largest S. giganteum in Yosemite Na- tional Park, whereas the Cleveland tree is a lesser- known and seldom- visited tree in Sequoia National Park. Finally, the Sentinel tree is a well-known se- quoia beside the road at the southern entrance to Giant Forest in Sequoia National Park. The General Sherman, Washington, General Grant, Grizzly Giant, and Cleveland trees all were cored by R. J. Hartes veldt and his colleagues for various studies during the late 1950's and early 1960's. All cores and data sheets for these trees are archived at Sequoia National Park, except I was unable to locate the original core for the Washing- ton tree, and therefore relied exclusively on Har- tes veldt's ring measurements for that tree. The Boole tree was cored by researchers from the Uni- versity of Arizona in 1992; those data were kindly supplied by L. S. Mutch. Finally, the Sentinel tree was cored by V. G. Pile and me in 1998 at the request of National Park Service staff, who wished to have an age estimate for displays near the tree. Estimating tree ages. I estimated ages of these seven S. giganteum following Stephenson and De- metry's (1995) approach, which combines knowl- edge of tree size with information gained from par- tial increment cores. The derivation and biological basis of this approach are too lengthy to repeat here; interested readers are therefore referred to Stephenson and Demetry (1995). When tested on 231 sequoia stumps up to 3200 years old and 6.5 m in diameter, this approach gave age estimates that were within 10% of actual age 62% of the time, and within 25% of actual age 98% of the time, assuming that two 60-cm increment cores are avail- able for analysis; fewer or shorter cores gave less precise estimates. This level of precision is a sub- stantial improvement over that of previously pub- lished methods, which estimated tree age from di- ameter alone, by assuming that basal area incre- ment is constant through time, or by linear extrap- olation of growth rates from the innermost portion of an increment core (Stephenson and Demetry 1995). Sequoia age in years, a, was estimated according to the following equation, lOOr^ a = {c- 100) + [1] - (r - gY where c is the full ring count of a partial increment core; g is the length of the innermost 100 rings of the increment core; r is the length g plus the length of the section of bole radius (extending to the tree's pith) that was not sampled by the increment core; and d is given by the following equation: d = 0.230 + 0.759(100/^^ J + 1.27r - 0.848r2 -f- 0.159r^ [2] Units for g and r are meters, whereas g^^n, is the length of the innermost 100 rings of the increment core in mm. For reasons discussed in Stephenson and Demetry (1995), if r exceeded 3 m, r = 3 m was substituted into eq. 2 for calculating d. A sequoia's pith usually is not at the geometric center of its bole. However, we typically have no way of determining the location of a living tree's pith, and therefore cannot directly measure the val- 2000] STEPHENSON: SEQUOIA AGES 63 Table 2. Confidence Intervals for S. giganteum Age Estimates Based on Different Numbers and Lengths of Increment Cores (from Stephenson and Demetry 1995). Two 60-cm cores One 60-cm core Two 30-cm cores One 30-cm core 50% 95% confidence interval confidence interval -6.9 to 9.0 -23.7 to 19.5 -8.4 to 9.4 -36.7 to 19.7 -14.1 to 11.1 -45.8 to 26.4 -13.0 to 11.8 -48.2 to 27.5 NOTE: The intervals are expressed as percentage of estimated sequoia age. For example, the -23.7% listed as one endpoint of the 95% confidence interval for two 60-cm cores means that 2.5% of the time, actual tree age will be more than 1.237 times estimated tree age. (Rephrased, 2.5% of the time estimated sequoia age will be at least 23.7% less, expressed in terms of estimated sequoia age, than actual sequoia age.) The 19.5% listed as the other endpoint of the interval means that 2.5% of the time, actual tree age will be less than 0.805 times estimated tree age. ue of r associated with a particular increment core. Therefore r was estimated as described by Ste- phenson and Demetry (1995). First, tree radius was calculated as half of tree diameter (determined by diameter tape) at the height at which the increment core was taken. Average bark thickness, determined by probes at several location around the bole, was then subtracted to determine tree radius inside the bark. From this, the length of the increment core, excluding the core's innermost 100 rings, was sub- tracted, yielding an estimate of r. Because increment cores shrink as they dry, the wet length of a core must be known for the most accurate application of eqs 1 and 2. However, for most of the S. giganteum analyzed here (the Sen- tinel tree being the one exception), wet lengths of cores were not recorded. My colleagues and I (un- published data) have found that the average shrink- age of hundreds of sequoia cores was about 2%. Thus, when the wet length of a core was not re- corded, it was estimated by multiplying the core's dry length by 1.02. To improve accuracy, when several cores were available from a sequoia, a given core's location on the bole had to be separated from that of the other cores by at least 90° of circumference to be includ- ed in the age estimation (Stephenson and Demetry 1995). Tree age at height cored was estimated by averaging the age estimates based on the individual cores (Stephenson and Demetry 1995). Some of the data used to estimate sequoia ages came from S. giganteum cored several decades ago. It was therefore necessary to account for the num- ber of years that have passed since a sequoia was cored. Because, for convenience, I wished to esti- mate all sequoia ages relative to the year 2000, I subtracted the year in which a core was taken from 2000, then added the result to estimated tree age. The method outlined above only estimates se- quoia age at the height at which the cores were taken. However, accounting for the time it took a tree to grow to the height cored potentially can add decades to the tree's estimated age. To account for height growth, I multiplied the height of the core above ground level (in m) by 178x"^^^^, where x is the (estimated) cumulative width, in mm, of the 10 rings that abut the tree's pith. This empirical factor scales height growth to radial growth, and was de- rived from ring measurements of 41 smaller S. gi- ganteum which my colleagues and I cored to the pith both near ground level and near breast height (see Agee et al. 1986 for a similar approach). How- ever, because there is no way of knowing the actual cumulative width of the 10 rings that abut the pith of the large S. giganteum analyzed here, I assumed that the width was 27.5 mm, based on the average from measurements of more than 450 sequoia stumps (Table A in Huntington 1914). Thus, I as- sumed that large S. giganteum took 178 X (27. 5)"^^^^ = 7.5 years to grow each meter taller until core height was reached. However, with the exception of the Sentinel and Grizzly Giant trees, core heights were not recorded. I therefore esti- mated core heights for the other trees based on con- versations and correspondence with individuals in- volved in the corings (H. S. Shellhammer for the General Sherman, Washington, General Grant, and Cleveland trees, and R. Adams and L. Mutch for the Boole tree). Confidence intervals. Stephenson and Demetry (1995) showed that as both the number and length of increment cores increase, confidence in sequoia age estimates also increases (Table 2). However, the numbers and lengths of cores used did not always fall neatly into the categories in Table 2. To deter- mine confidence intervals, core lengths were there- fore rounded to the nearest category shown in Table 2 (either 30 or 60 cm). In two cases (the General Sherman and General Grant trees), three cores rath- er than two were used. However, since confidence is improved relatively little by increasing core num- ber (it is improved more by increasing core length; Table 2), confidence intervals for only two cores were used. The number of years elapsed between the year in which a tree was cored and the year 2000 was then added to the endpoints of the tree's confidence intervals, as was the estimated number of years it took each sequoia to grow to the height at which it was cored. Admittedly, the latter step does not change a sequoia's age confidence intervals to re- flect the uncertainty associated with estimating the number of years it took a sequoia to grow to the height cored. However, uncertainty added at this 64 MADRONO [Vol. 47 Table 3. Data Used to Estimate the Ages of the Selected S. giganteum. ^ Side of tree from which core was taken. Confidence intervals (see Table 2) are: 1 X 30, one 30-cm increment core; 2 X 30, two 30-cm cores; 1 X 60, one 60-cm core; 2 X 60, two 60-cm cores. '^Estimated from length of innermost 154 rings. ''Estimated from length of innermost 280 rings. Diameter Bark at core thick- height ness Tree Core^ (m) (m) Wet Wet length of length innermost of full 100 rings core of core (m) (m) Height Ring of core count above of full ground core (m) Confidence Year interval cored used^ General Sherman South 7.325 0.127 0.387 0.148 317 1.6 1964 2 X 30 Northwest 7.325 0.127 0.365 0.156 249 1.6 1964 East 7.325 0.127 0.352 0.120 315 1.6 1964 Washington 7.858 0.152 0.291 0.091 325 1.4 1963 1 X 30 General Grant Southeast 6.705 0.203 0.375 0.259 146 2.0 1964 2 X 30 West 6.705 0.203 0.376 0.180 233 2.0 1964 North 6.705 0.203 0.378 0.139 293 2.0 1964 Boole B (Northwest?) 7.45 0.090 0.418 0.124^ 259 1.4 1992 1 X 60 C (Northeast?) 7.45 0.090 0.639 386 1.4 1992 Grizzly Giant Southwest 6.621 0.127 0.289 0.175 206 3.05 1958 2 X 30 East 6.621 0.127 0.266 0.148 175 3.05 1958 Cleveland 5.613 0.127 0.347 0.045 598 1.6 1964 1 X 30 Sentinel Northeast 6.399 0.073 0.515 0.099 366 2.56 1998 2 X 60 Southwest 6.399 0.073 0.556 0.128 333 2.04 1998 stage is small compared to the uncertainty of esti- mating the tree's age at core height. Statistics on the longest-lived sequoia known. As a yardstick for interpreting results, I used the age and size of the longest-lived sequoia known — a cut stump in Converse Basin, Giant Sequoia National Monument, designated CBR26 by its discoverers (R. Touchan and E. Wright of the University of Ar- izona's Laboratory of Tree-Ring Research). To- General Sherman Washington General Grant Boole Grizzly Giant Cleveland Sentinel IZ> I— □: i-cn — I 2500 3000 Age (years) 4000 4500 Fig. 1. Estimated ages of selected S. giganteum in the year 2000, with associated confidence intervals. The ver- tical line within each horizontal box indicates that tree's estimated age. The ends of each box delimit the 50% con- fidence interval for that tree's age, whereas the "whisk- ers" extending from each box delimit the 95% confidence interval. The dotted vertical line at 3266 years indicates the age of the oldest sequoia yet discovered (see the text). Because the innermost ring of a long core taken within a fire scar cavity at the base of the Boole tree has been crossdated to A.D. 143 by E. Wright of the University of Arizona (L. Mutch personal communication), the Boole tree is at least 1858 years old, as indicated by the asterisk. uchan has precisely crossdated 3207 rings on the stump. It is missing much of its sapwood, so the outermost ring dates to 1834. However, the exten- sive logging of Converse Basin Grove occurred be- tween 1893 and 1908 (Johnston 1983; Willard 1994). Thus, at least 59 years of sapwood are miss- ing, and the tree therefore was at least 3266 years old when it was cut. (It is unlikely that the tree exceeded 3290 years old, including the time it took the tree to grow to the height sampled by Touchan and Wright.) The stump is relatively small: 5.8 m in diameter near ground level and 4.3 m in diameter at the cut surface 2.2 m above ground level (R. Touchan personal communication). Even with sap- wood and bark intact, the tree's diameter at 2.2 m above ground level was probably less than 5 m when it was cut, much smaller than any of the trees analyzed here (Table 3). While we will never know the volume of the living CBR26, it is clear that many hundreds of S. giganteum alive today (prob- ably well over one thousand) are larger than CBR26 was before it was cut (e.g., see Appendix 1 in Stohlgren 1991). Results Table 3 presents the data used to estimate the ages of the seven large S. giganteum. Estimated ages ranged from 1650 years for the General Grant tree to 2890 years for the Cleveland tree (Fig. 1), averaging 2230 years. Though all of these S. gi- ganteum were much larger than CBR26, the lon- gest-lived sequoia known, five had estimated ages at least 1000 years younger than CBR26 (Fig. 1). In fact, the third-largest living sequoia (the General Grant tree) is estimated to be little more than half 2000] STEPHENSON: SEQUOIA AGES 65 i as old as CBR26. Additionally, CBR26's age lies well outside of the high end of the 95% confidence I intervals of the five S. giganteum (Fig. 1). j While there are exceptions (namely, the Wash- ington and Cleveland trees), the largest living S. giganteum generally owe their great bulk to rapid growth, not to extraordinary age. For example, av- erage ring width from the cores of the (estimated) youngest sequoia (the General Grant tree, 1.82 mm) was more than three times that of the (estimated) oldest sequoia (the Cleveland tree, 0.58 mm). This notion is further supported by Huntington's (1914) age data from more than 450 sequoia stumps (the accuracy of which is discussed in Stephenson and Demetry 1995). Huntington's ten largest stumps av- eraged 6.0 m in diameter inside the bark, but only 1842 years old by direct ring count (the largest was 6.5 m in diameter but only 1347 years old). In sharp contrast, his ten oldest stumps averaged only 4.9 m in diameter inside the bark, but 2822 years old — 1 m less in diameter but nearly 1000 years older. Membership in the two groups of stumps was al- most mutually exclusive; only one stump was both one of the ten largest and one of the ten oldest (see Fig. 1 in Stephenson and Demetry 1995). Thus, for whatever reason, S. giganteum that reach great age tend to have grown relatively slowly. Figure 1 indicates that there is sl 25% probability that the Cleveland tree is older than CBR26, and a similar probability that the Washington tree is older. The probability that at least one of these two living trees (Cleveland or Washington) is older than CBR26 therefore is roughly 1 - (0.75)% or 44%— nearly even odds. Given that the seven S. gigan- teum examined here are only a small sample of all potentially old, living S. giganteum (likely candi- dates would number well over one thousand), it seems highly likely that some S. giganteum living today exceed the age of CBR26. Discussion There has been a long-standing belief that the largest S. giganteum are the oldest. This is well illustrated by tracing the history of age estimates for the General Sherman tree, the world's largest tree. By the early 20'*' century, careful ring counts and crossdating had identified a handful of sequoia stumps more than 3000 years old, the oldest being about 3200 years old (Huntington 1914; Douglass 1919, 1945). (John Muir's reported count of 4000 rings on the "Muir Snag" in 1875 has not been repeated and was almost certainly in error [Flint 1987], and other early claims of up to 11,000 rings counted on stump tops [Jordan 1907] cannot be taken seriously.) Since none of these old stumps approached the great size of the General Sherman tree, most natural historians concluded that the General Sherman tree must be more than 3500 years old (e.g.. Fry and White 1930). Stewart (1930) believed that the General Sherman tree was about 4000 years old, though he reported that an estimate based on "average number of rings count- ed ... in charred fragments from parts of the [Gen- eral Sherman tree's] burned trunk, in connection with the actual counts of rings of felled trees . . . which have grown under conditions and situation similar to those of the Sherman tree" yielded an age of 5200 years. Popular publications, such as a 1931 program for a play performed among the se- quoias not far from the General Sherman tree, tend- ed to be more extravagant, proclaiming the tree to be 6000 years old (see also Hartesveldt et al. 1975). Ironically, the aforementioned play took place less than two months before the first quantitative esti- mate of the General Sherman tree's age based on increment cores, by A. E. Douglass. Douglass, the founder of the modem science of dendrochronology, obtained six short cores from the General Sherman tree in 1931 (the year is mis- takenly given as 1935 in Douglass [1946]). He deemed two of the cores to be good enough to use for age estimation, finding that average ring width at 4.6 m above ground level was 0.81 mm. This ring width is less than that of Hartesveldt's cores (Table 3) because it comes from a height where the General Sherman tree's bole is narrower. Douglass stated that "[t]hese are ring sizes which, in relation to the total size of the tree and the probable rate at which rings increase in size toward the center, sup- plied an estimate of the age of the tree of 3500 years plus or minus 500 years" (Douglass 1946). I have found no quantitative description of how Douglass accounted for "the probable rate at which rings increase in size toward the [tree's] center." To shed light on Douglass' age estimate, I ap- plied the approach outlined in this paper to his data. Douglass' data yield an age of only 2380 years for the General Sherman tree in 1931, or 2450 years in 2000 (rounded to the nearest decade). This latter estimate is only 300 years older than the estimate based on Hartesveldt's cores (Fig. 1), and is well within that estimate's 95% confidence interval. However, I judge the estimate based on Hartes- veldt's cores to be much more reliable than that based on Douglass' cores. Specifically, the estimate based on Hartesveldt's cores required that fewer key parameters be estimated (such as the diameter of the General Sherman tree at 4.6 m above ground level in 1931, needed for using Douglass' data), and was based on three cores widely spaced around the tree's bole, each of which was nearly twice as long as the longest of Douglass' two adjacent cores. In contrast, an age estimate based on linear ex- trapolation of Douglass' ring- width data, assuming no change in ring width toward the General Sher- man tree's center (an unrealistic assumption), would yield an age of 3790 years in 1931. Thus, Douglass' estimate of 3500 (±500) years apparent- ly was little different from an estimate based on a simple linear extrapolation, and did not adequately consider the increase in ring widths toward the pith. 66 MADRONO [Vol. 47 Douglass' age estimate was widely quoted (and sometimes exaggerated) from 1931 until the 1960's, when Hartesveldt et al. (1975) radically revised the estimate downward. Unlike Douglass, Hartesveldt and his colleagues explicitly stated their assumption as to how ring widths change within a tree: they assumed that basal area increment is constant (that is, trees add a constant amount of basal area each year). This is equivalent to substituting d = 2 into eq. 1 (Stephenson and Demetry 1995). Hartes- veldt's notes (archived at Sequoia National Park) show that when he strictly adhered to this assump- tion, he estimated that in 1964 the General Sherman tree was only about 1600 years old. However, Har- tesveldt's examination of growth patterns on se- quoia stumps measured by Huntington (1914) in- dicated that strict adherence to this assumption sometimes underestimated the ages of S. giganteum (Hartesveldt et al. 1975). Thus, apparently based on a combination of assumed constant basal area in- crement and judicious comparisons with Hunting- ton's data, Hartesveldt and his colleagues (1975) cautiously stated that the General Sherman tree ". . . is less than 2500 years old." According to my calculations using their original cores and data, their statement has a more than 75% probability of being true (Fig. 1). As careful as Hartesveldt et al. (1975) may have been in stating that the General Sherman tree was less than 2500 years old, the National Park Service, perhaps unable to bear such a precipitous decline in the tree's age, instead adopted 2500 years as the midpoint for a range encompassing the tree's esti- mated age. At the time of this writing. Park litera- ture and the plaque at the General Sherman tree stated that the tree's estimated age was "2300- 2700 years." Additionally, a popular book authored by Hartesveldt's colleagues (Harvey et al. 1981) dropped the qualifier "less than," stating instead that the tree ". . . is about 2,500 years old" (though a table on the same page gives the General Sher- man tree's age as "2,500-3,000" years!). The most recent estimate of the General Sherman tree's age — 2150 years (Fig. 1) — is most closely aligned with Hartesveldt et al.'s (1975) original statement that the tree is less than 2500 years old. The relative youth of other famous S. giganteum may come as a disappointment to some. For ex- ample, the decline in the estimated age of the Griz- zly Giant tree has been even more precipitous than that of the General Sherman tree. Clark (1910) re- ported that the Grizzly Giant had been growing so slowly over the last few centuries that its rings (pre- sumably observed inside of a fire scar cavity) were "as thin as wrapping paper, too fine to be counted with the unaided eye." (On the contrary, measured ring widths [Table 3] and measured tree volume changes [W. Flint personal communication] both in- dicate that the tree has been growing quite rapidly.) Comparing these purported ring widths with those of some fallen S. giganteum, Clark concluded that "the Grizzly Giant must be not less than six thou- sand years old," and that the tree was probably the oldest living thing on earth. Other early age esti- mates placed the Grizzly Giant at a more modest 3800 years old, while Hartesveldt et al. (1975) later suggested that the tree "... is perhaps only 2500 years old." At the time of this writing, the National Park Service reported the age of the Grizzly Giant as 2700 years. However, I estimate the tree to be only about 1790 years old (Fig. 1), and that the probability of it being at least 2700 years old is less than 2%. Hartesveldt and his colleagues (1975) of- fered solace to those disappointed by the suggestion that certain large S. giganteum might be younger than expected: "... this [discovery] effects a change only in superlatives; the world's largest trees are the world's fastest-growing trees." Some readers may be disappointed by the broad confidence intervals associated with age estimates in Figure 1. There is a great deal of uncertainty in estimating the ages of individual large S. gigan- teum, largely due to relatively abrupt and sustained changes in ring widths in the part of the bole not sampled by increment cores, and therefore invisible to us (Stephenson and Demetry 1995). Such changes in growth rates are due to unpredictable, site- specific events in the past, such as occasional, localized high-intensity fires (e.g.. Mutch and Swet- nam 1995). Thus, though Figure 1 suggests that the General Sherman and Sentinel trees are the same age (2150 years), the broad confidence intervals ad- ditionally suggest that this correspondence is most likely a meaningless coincidence. However, most of the confidence intervals in Figure 1 are based on relatively short cores. Confidence intervals could be tightened somewhat in the future by taking longer cores and, in the case of the Washington and Cleve- land trees, more cores. Acknowledgments I thank R. Touchan of the University of Arizona's Lab- oratory of Tree-Ring Research for graciously supplying data on CBR26, and L. S. Mutch for supplying data on the Boole tree. Additional thanks go to V. G. Pile (USGS), who helped me core the Sentinel tree, and D. J. McGraw (University of San Diego), who sent copies of Douglass' original notes and correspondence on the General Sher- man tree, and who inspired me to complete this work. A. Caprio, M. Crapsey, W. Fhnt, J. Keeley, D. McGraw, H. Shellhammer, D. Shenk, R. Touchan, W. Tweed, and two reviewers provided helpful comments on the manuscript. Literature Cited Agee, J. K., M. Finney, and R. de Gouvenain. 1986. The fire history of Desolation Peak. Cooperative Parks Study Unit, University of Washington, Seatde. Final contract report to the National Park Service, Coop- erative Agreement CA-9000-3-0004. Brown, R M., M. K. Hughes, C. H. Baisan, T. W. Swet- NAM, AND A. C. Caprio. 1992. Giant sequoia ring- width chronologies from the central Sierra Nevada, California. Tree-Ring Bulletin 52:1-14. 2000] STEPHENSON: SEQUOIA AGES 67 Clark, G. 1910. The big trees of California. Yosemite Valley, CA. CuRREY, D. R. 1965. An ancient bristlecone pine stand in eastern Nevada. Ecology 46:564-566. Douglass, A. E. 1919. Climatic cycles and tree growth, Vol. I. Carnegie Institute of Washington Publication No. 289, Washington D.C. . 1945. Survey of sequoia studies. Tree-Ring Bul- letin 11:26-32. . 1946. Sequoia survey — III: miscellaneous notes. Tree-Ring Bulletin 13:5-8. Echols, R. M. 1969. Powered drive for large increment borers. Journal of Forestry 67:123-125. Flint, W. D. 1987. To find the biggest tree. Sequoia Nat- ural History Association, Three Rivers, CA. . In press. To find the biggest tree, 2nd ed. Sequoia Natural History Association, Three Rivers, CA. Fry, W. and J. R. White. 1930. Big trees. Stanford Uni- versity Press, CA. Hartes VELDT, R. J. 1962. The effects of human impact upon Sequoia gigantea and its environment in the Mariposa Grove, Yosemite National Park, California. Ph.D. dissertation. University of Michigan, Ann Ar- bor. . 1965. An investigation of the effect of direct hu- man impact and of advanced plant succession on Se- quoia gigantea in Sequoia and Kings Canyon Na- tional Parks, California. Report on contract number 14-10-0434-1421, USDI National Park Service, San Francisco, CA. , H. T. Harvey, S. H. Shellhammer, and R. E. Stecker. 1975. The giant sequoia of the Sierra Ne- vada. U.S. Department of the Interior, National Park Service, Washington D.C. Harvey, H. T, H. S. Shellhammer, R. E. Stecker, and R. J. Hartesveldt. 1981. Giant sequoias. Sequoia Natural History Association, Three Rivers, CA. Hughes, M. K., R. Touchan, and P. M. Brown. 1996. A multimillennial network of giant sequoia chronolo- gies for dendroclimatology. Pp. 225-234 in J. S. Dean, D. M. Meko, and T. W. Swetnam (eds.). Tree rings, environment, and humanity. Proceedings of the International Conference. Radiocarbon, Department of Geosciences, The University of Arizona, Tucson, AZ. AND P. M. Brown. 1992. Drought frequency in central California since 101 B.C. recorded in giant sequoia tree rings. Climate Dynamics 6:161-167. Huntington, E. 1914. The climatic factor as illustrated in arid America. Carnegie Institute of Washington Pub- lication No. 192, Washington D.C. Johansen, R. W. 1987. Taking increment cores with pow- er tools. Southern Journal of Applied Forestry 1 1 : 151-153. Johnston, H. 1983. They felled the redwoods. Trans- An- glo Books, Glendale, CA. Jordan, D. S. 1907. The alps of the King-Kern divide. A. M. Robertson, San Francisco, CA. Lara, A. and R. Villalba. 1993. A 3620-year tempera- ture record from Fitzroya cupressoides tree rings in southern South America. Science 260:1104-1106. Mutch, L. S. and T. W. Swetnam. 1995. Effects of fire severity and climate on ring-width growth of giant sequoia after burning. Pp. 241-246 in J. K. Brown, R. W. Mutch, C. W. Spoon, and R. H. Wakimoto, (technical coordinators). Proceedings: symposium on fire in wilderness and park management, 30 March- 1 April 1993, Missoula, Montana. USDA Forest Ser- vice General Technical Report INT-GTR-320. Stephenson, N. L. 1994. Long-term dynamics of giant sequoia populations: implications for managing a pi- oneer species. Pages 56-63 in P. S. Aune, technical coordinator. Proceedings of the Symposium on Giant Sequoias: their place in the ecosystem and society, 23-25 June 1992, Visalia, California. USDA Forest Service Gen. Tech. Rep. PSW-151. AND A. Demetry. 1995. Estimating ages of giant sequoias. Canadian Journal of Forest Research 25: 223-233. Stewart, G. W. 1930. Big trees of the Giant Forest. A. M. Robertson, San Francisco, CA. Stohlgren, T J. 1991. Size distributions and spatial pat- terns of giant sequoia {Sequoiadendron giganteum) in Sequoia and Kings Canyon National Parks, Califor- nia. Technical Report No. 43, Cooperative National Park Resources Study Unit, University of California, Davis, CA. USDI National Park Service. Stokes, M. A. and T. L. Smiley. 1968. An introduction to tree-ring dating. University of Chicago Press, Chi- cago. Swetnam, T. W. 1993. Fire history and climate change in giant sequoia groves. Science, 262:885-889. WiLLARD, D. 1994. Giant sequoia groves of the Sierra Ne- vada: a reference guide. Privately published by D. Willard, R O. Box 7304, Berkeley, CA, 94707. Madrono, Vol. 47, No. 1, pp. 68-69, 2000 REVIEW A natural history of the Sonoran Desert. Edited by S. J. Phillips and P. W. Comus. 2000. Arizona-So- nora Desert Museum Press, Tucson AZ, and Uni- versity of California Press, Berkeley CA. 628 pp. Cloth $55.00 ISBN 0-520-22029-3 Paper $24.00 ISBN 0-520-21980-5. The staff and associates of the Arizona-Sonora Desert Museum wrote this book as a compilation of their training courses, research, and personal ex- periences in the Sonoran Desert of Arizona, Cali- fornia, and Mexico. It evolved from the Museum's docent handbook, developed over 30 years as a training document for volunteer interpreters. It pro- vides a summary of numerous biotic and abiotic patterns and processes, emphasizing adaptations of desert organisms and the interrelationships between nature and humans, both past and present. It an- swers in nontechnical prose typical questions of visitors to the Sonoran Desert. As such, this book covers a wide range of topics within its 628 pages. The book begins by briefly describing what a desert is, and how the Sonoran Desert differs from other deserts of North America. The regional sub- divisions and biomes are nicely summarized based on the original work of Forrest Shreve. The book then launches into two chapters describing a cal- endar of natural events and ten nature watching hot- spots. Although these two chapters are informative and will undoubtedly be of great use to those plan- ning trips to this region, they would make more sense if read after the other chapters and should have been placed at the end of the book, possibly as appendices. A chapter on desert storms gets the book gets back on track. This is a key chapter appropriately placed near the beginning of the text. It introduces rainfall as a significant factor influencing the evo- lution of desert organisms and the limitations to human settlement in the Sonoran Desert. Most oth- er chapters that follow presuppose some informa- tion contained in this chapter. The next chapter on desert air and light breaks with the main theme of the book by presenting ex- planations couched in basic physics. Phenomena such as mirages, atmospheric shimmer, and dust devils are described using simple descriptions of light refraction and the influence of temperature on the behavior of air. Although it is not directly re- lated to the other chapters, it presents an entertain- ing, effective, and succinct summary and does not significantly detract from the flow of the book. Consecutive chapters on deep history, geologic origins, soils, human ecology, and biodiversity frame the current Sonoran Desert in the perspective of evolutionary and recent time scales, and set the stage for the meat of the book which is the ecology of plants and animals. The deep history chapter de- scribes changes in flora and fauna over geologic time and discusses relationships between their past and current distributions. The geologic origins chapter describes the geomorphological develop- ment of major geologic features. The soils chapter describes the development and physical and biolog- ical properties of soils, and their implications for plants and animals. The human ecology chapter de- scribes the influence of humans over the past 12,000 years, ranging from native American hunt- ing, gathering, and farming, to Anglo-American ag- riculture, mining, and dam building. The chapter on biodiversity defines different indices and scales of biological diversity, describes natural centers of di- versity, and explains how humans have reduced di- versity by introducing invasive species and con- verting native plant communities to monoculture farms. Plants are discussed in separate treatments of plant ecology, flowering plants and grasses. The chapter on plant ecology covers a wide range of topics in its brief 23 pages. Topics include rudi- mentary descriptions of flower anatomy and pho- tosynthesis, and more detailed descriptions of drought adaptations, pollination, seed dispersal, and flowering seasons. The chapter on flowering plants includes a sampling of the most conmion and in- teresting angiosperms (other than grasses) in the Sonoran Desert. It is organized by taxonomic fam- ily, within which a few representative species are presented, covering their description, range, and comments on ethnobotany and natural history. The chapter on grasses includes a relatively detailed ac- count of different grassland types and dominant species, including a few dominant aliens, and some original natural history accounts. Animals are presented in separate sections cov- ering invertebrates, birds, mammals, fishes, reptiles, and amphibians. Adaptations to life in the desert by these groups are sunMnarized. Species accounts of the common or otherwise interesting taxa include descriptions of distinguishing characteristics, habi- tat, range, life history, and in some cases feeding behavior. The glossary is very brief, but only includes terms which are not referenced in the extensive in- dex which lists a wide variety of items including conmion and scientific names, geographic places, and descriptive terms. This book provides a comprehensive introduc- tion to the natural history of the Sonoran and Mo- jave deserts, because many examples presented and 2000] REVIEW 69 species discussed are common to these two deserts. It would be an ideal text for a conmiunity college or undergraduate course on desert ecology. Upper division and graduate students would not find much new information in this book. The strength of the book lies in the natural history descriptions for in- dividual species. This information is typically given very short treatment in the floras and field guides of the region, and all readers should find this in- formation interesting and useful. Matthew L. Brooks. United States Geological Survey, Western Ecological Research Center, Box Springs Field Station, 6221 Box Springs Blvd. Riv- erside CA 92507 California Botanical Society — Meeting Program 2000-2001 Academic Year All Meetings are held at 7:30 pm in room 2063 in the Valley Life Sciences Building on the UC Berkeley campus. September 21, 2000 Predicting the future of Sierran conifer forests: no lessons from the past. John Battles, Professor, University of California, Berkeley October 19, 2000 Diversity in California's serpentine plants: the roles of patchiness, grazing, and burning. Susan Harrison, Professor, University of California, Davis November 16, 2000 Restoration of oak woodlands and grasslands in California: an evolutionary perspective. Kevin Rice, Professor, University of California, Davis January 18, 2001 Explosive beauty: rare plant research and management at Lawrence Livermore National Lab's high explosive test facility, site 300. Tina Carlsen, Project Leader and Ecologist, Lawrence Livermore National Lab February 21, 2001 ANNUAL BANQUET and SEMI-ANNUAL GRADUATE STUDENT MEETINGS **NOTE CHANGE OF LOCATION: California State University, Chico The role of geology in molding the California flora Arthur Kruckeberg, Professor Emeritus, University of Washington March 15, 2001 Molecular phylogenetic studies in Rosaceae. Dan Potter, Professor, University of California, Davis April 19, 2001 Defenders or pretenders? Interactions between an African acacia tree and four symbiotic ants. Maureen Stanton, Professor, University of California, Davis May 17, 2001 Using DNA fingerprinting to study Sequoia sempervirens populations in Big Basin Redwoods State Park. Chris Brinegar, Professor, San Jose State University Volume 47, Number 1, pages 1-70, published 8 December 2000. 1 j i I Subscriptions- — Membership Membership in the California Botanical Society is open to individuals ($27 per year; family $30 per year; emeritus $17 per year; students $17 per year for a maximum of 7 years). Late fees may be assessed. Members of the Society receive Madrono free. Institutional subscriptions to Madrono are available ($60). Membership is based on a calendar year only. Life memberships are $540. Applications for membership (including dues), orders for sub- scriptions, and renewal payments should be sent to the Treasurer. Requests and rates for back issues, changes of address, and undelivered copies of Madrono should be sent to the Corresponding Secretary. Information for Contributors Manuscripts submitted for publication in Madrono should be sent to the editor. It is preferred that all authors be members of the California Botanical Society. Manuscripts by authors having outstanding page charges will not be sent for review. Manuscripts may be submitted in English or Spanish. English-language manuscripts dealing with taxa or topics of Latin America and Spanish-language manuscripts must have a Spanish Resumen and an English Abstract. Manuscripts and review copies of illustrations must be submitted in triplicate for all articles and short items (NOTES, NOTEWORTHY COLLECTIONS, POINTS OF VIEW, etc.). Follow the format used in recent issues for the type of item submitted. Allow ample margins all around. Manuscripts MUST BE DOUBLE-SPACED THROUGHOUT. For articles this includes title (all caps, centered), author names (all caps, centered), addresses (caps and lower case, centered), abstract and resumen, 5 key words or phrases, text, acknowledgments, literature cited, tables (caption on same page), and figure captions (grouped as consecutive paragraphs on one page). Order parts in the sequence listed, ending with figures. Each page should have a running header that includes the name(s) of the author(s), a shortened title, and the page number. Do not use a separate cover page or 'erasable' paper. Avoid footnotes except to indicate address changes. Abbreviations should be used sparingly and only standard abbrevia- tions will be accepted. Table and figure captions should contain all information relevant to information presented. All measurements and elevations should be in metric units, except specimen citations, which may include English or metric measurements. Authors of accepted papers will be asked to submit an electronic version of the manuscript. Microsoft Word 6.0 or WordPerfect 6.0 for Windows is the preferred software. Line copy illustrations should be clean and legible, proportioned to the Madrono page. Scales should be in- cluded in figures, as should explanation of symbols, including graph coordinates. Symbols smaller than 1 mm after reduction are not acceptable. Maps must include a scale and latitude and longitude or UTM references. In no case should original illustrations be sent prior to the acceptance of a manuscript. Illustrations should be sent flat. No illustrations larger than 27 X 43 cm will be accepted. Presentation of nomenclatural matter (accepted names, synonyms, typification) should follow the format used by Sivinski, Robert C, in Madrono 41(4), 1994. Institutional abbreviations in specimen citations should follow Holmgren, Keuken, and Schofield, Index Herbariorum, 8th ed. Names of authors of scientific names should be abbreviated according to Brummitt and Powell, Authors of Plant Names ( 1992) and, if not included in this index, spelled out in full. Titles of all periodicals, serials, and books should be given in full. Books should include the place and date of publication, publisher, and edition, if other than the first. All members of the California Botanical Society are allotted 5 free pages per volume in Madrono. Joint authors may split the full page number. Beyond that number of pages a required editorial fee of $40 per page will be assessed. The purpose of this fee is not to pay directly for the costs of publishing any particular paper, but rather to allow the Society to continue publishing Madrono on a reasonable schedule, with equity among all members for access to its pages. Printer's fees for illustrations and typographically difficult material @ $35 per page (if their sum exceeds 30 percent of the paper) and for author's changes after typesetting @ $4.50 per line will be charged to authors. At the time of submission, authors must provide information describing the extent to which data in the manu- script have been used in other papers that are published, in press, submitted, or soon to be submitted elsewhere. VOLUME 47, NUMBER 2 APRIL-JUNE 2000 ^mADRONO A WEST AMERICAN JOURNAL OF BOTANY Locations of Endangered Spruce Populations in Mexico and the Demography of Picea chihuahuana F. Thomas Ledig, Manuel Mdpula-Larreta, Basilio Bermejo-Veldzquez, Valentin Reyes -He rndndez, Celestino Flores-Lopez, and Miguel A. Capd-Arteaga 71 Revision of Corethrogyne (Compositae: Astereae) J. Phillip Saroyan, Dennis R. Parnell, and John L. Strother 89 Correspondence between Ni Tolerance and Hyperaccumulation in Streptanthus (Brassicaceae) Robert S. Boyd, Michael A. Wall, and James E. Watkins, Jr. 97 Annual Variation in Xylem Water Potential in California Oaks Johannes M. H. Knops and Walter D. Koenig 106 Molecular Evidence for the Hybrid Origin of Opuntia prolifera (Cactaceae) Michael S. Mayer, Laura M. Williams, and Jon P. Rebman 109 Floral Variation in Delphinium variegatum (Ranunculaceae) Shana C. Dodd and Kaius Helenurm 1 16 Crown Structure of the World's Second Largest Tree Stephen C. Sillett, James C. Spickler, and Robert Van Pelt 127 Eriogonum spectabile (Polygonaceae): A New Species from Northeastern California Beth Lowe Corbin, James L. Reveal, and Robin Barron 134 ARIZONA 138 CALIFORNIA 138 COLORADO 142 OREGON 144 PUBLISHED QUARTERLY BY THE CALIFORNIA BOTANICAL SOCIETY Madrono (ISSN 0024-9637) is published quarterly by the California Botanical Society, Inc., and is issued from the office of the Society, Herbaria, Life Sciences Building, University of California, Berkeley, CA 94720. Subscription information on inside back cover. Established 1916. Periodicals postage paid at Berkeley, CA, and additional maihng offices. Return requested. Postmaster: Send address changes to Madrono, Roy Buck, % University Herbarium, University of California, Berkeley, CA 94720. Editor — Kristina A. Schierenbeck California State University, Chico Department of Biology Chico, CA 95429-05 15 kschierenbeck ©csuchico.edu Editorial Assistant — David T. Parks Book Editor — Jon E. Keeley Noteworthy Collections Editors — Dieter Wilken, Margriet Wetherwax Board of Editors Class of: 2000 — Pamela S. Soltis, Washington State University, Pullman, WA John Callaway, University of San Francisco, San Francisco, CA 2001 — Robert Patterson, San Francisco State University, San Francisco, CA Paula M. Schiffman, California State University, Northridge, CA 2002 — Norman Ellstrand, University of California, Riverside, CA Carla M. D' Antonio, University of California, Berkeley, CA 2003 — Frederick Zechman, California State University, Fresno, CA JoN E. Keeley, U.S. Geological Service, Biological Resources Division, Three Rivers, CA 2004 — David M. Wood, California State University, Chico, CA Ingrid Parker, University of California, Santa Cruz, CA CALIFORNIA BOTANICAL SOCIETY, INC. Officers for 2000-2001 President: Bruce Baldwin, Jepson Herbarium and Dept. of Integrative Biology, 1001 Valley Life Sciences Bldg. #2465, University of California, Berkeley, CA 94720. First Vice President: Rod Myatt, San Jose State University, Dept. of Biol. Sciences, One Washington Square, San Jose, CA 95192. rmyatt@email.sjsu.edu Second Vice President: Rob Schlising, Cahfomia State University, Chico, Dept. of Biol. Sciences, Chico, CA 95424. rschlising@csuchico.edu Recording Secretary: Dean Kelch, Jepson and University Herbarium, University of California, Berkeley, CA 94720. dkelch @ sscl.berkeley.edu Corresponding Secretary: Susan Bainbridge, Jepson Herbarium, University of California, Berkeley, CA 94720. suebain@SSCL.berkeley.edu Treasurer: Roy Buck, % University Herbarium, University of California, Berkeley, CA 94720. The Council of the California Botanical Society comprises the officers listed above plus the immediate Past President, R. John Little, Sycamore Environmental Consultants, 6355 Riverside Blvd., Suite C, Sacramento, CA 95831; the Editor of Madrono; three elected Council Members: Bian Tan, Strybing Arboretum, Golden Gate Park, San Francisco, CA 94122; James Shevock, USDI National Park Service, Pacific West Region, 600 Harrison Street, Suite 600, San Francisco, CA 94107; Diane Elam, U.S. Fish and Wildlife Service, 3310 El Camino Avenue, Sacramento, CA 95825; Graduate Student Representative: Kjrsten Johanus, Jepson Herbarium, University of California, Berkeley, CA 94720. @ This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Madrono, Vol. 47, No. 2, pp. 71-88, 2000 LOCATIONS OF ENDANGERED SPRUCE POPULATIONS IN MEXICO AND THE DEMOGRAPHY OF PICEA CHIHUAHUANA F Thomas Ledig* Institute of Forest Genetics, Pacific Southwest Research Station, USDA Forest Service, 2480 Carson Road, Placerville, CA 95667 USA Manuel Mapula-Larreta, Basilio Bermejo- Velazquez and Valentin Reyes-Hernandez Centro de Genetica Forestal, Universidad Autonoma Chapingo, Apartado Postal No. 37, Chapingo, Mexico, CP. 56230, Mexico Celestino Flores-Lopez and Miguel A. Capo-Arteaga Departamento Forestal, Universidad Autonoma Agraria Antonio Narro, Buenavista, Saltillo, Coahuila, Mexico Abstract Picea A. Dietr. (spruce) is an essentially boreal genus, but three endemic taxa occur in Mexico. Co- ordinates were determined for all known stands to accurately map their range and stimulate their protec- tion, conservation, and study. Thirty-nine stands of Picea chihuahuana Martinez (Chihuahua spruce) in the Sierra Madre Occidental were found in three clusters, each separated by 2 to 2.5° of latitude. The southernmost stands occur just south of the Tropic of Cancer. The entire north-south range is 687 km. Mean elevation of southern and central clusters was 2675 m, but stands in the northern cluster averaged over 350 m lower in elevation. Picea chihuahuana was associated with steep-sided arroyos or at the base of barrancas (cliffs or gorges). Picea martinezii T F. Patterson (Martinez spruce) was found in six stands in the Sierra Madre Oriental, at an elevation of about 2250 to 2650 m and all within 147 km of each other. Picea mexicana Martinez (Mexican spruce) occurred on two of the highest ridges in the Sierra Madre Oriental, about 5 km apart and at an elevation of about 3500 m, and on the highest point (3185 m) in Chihuahua in the Sierra Madre Occidental, 676 km to the west. It is probable that P. mexicana will be found on one or two other high ridges in the Sierra Madre Oriental. Every P. chihuahuana over 0.3 m in height was counted, measured, and scored for mistletoe infection, fire damage, and crown dieback from unknown cause(s) in 21 stands. Similar observations were made for another 18 stands by Narvaez et al. (1983) about 15 years earlier. The combined count was 42,610 P. chihuahuana, which includes 24,221 trees and saplings over 2 m tall and 18,389 seedlings under 2 m but over 0.3 m in height. The distribution of diameter classes in our sample of 21 stands was a reverse- J, suggesting that the species is reproducing. However, the ratio of seedlings to saplings and trees was less than 1.0 in all except four of the 39 stands, indicating that the species may actually be in jeopardy. Based on ring counts from increment cores and stumps, P. chihuahuana can reach 272 years of age. This is a relatively short life span compared to other North American spruces. The largest trees were 51 m tall and 125 to 150 cm in diameter-breast-high, and size was about average compared to its congeners in the United States and Canada. Many trees were in poor condition or damaged from cutting, mistletoe, top dieback, and fire. Contrary to expectations, the southern stands were in no poorer condition than the northern, and in fact, the incidence of mistletoe was highest in the north; this seems to be the first report of mistletoe on P. chihuahuana. Trees in southern stands were larger and older, and the ratio of seedlings to saplings and trees was highest in the southern and central stands and lowest in the north. Reslimen Los abetos {Picea A. Dietr.) son un genero esencialmente boreal. A pesar de lo anterior, tres taxa endemicos ocurren en Mexico. Las coordenadas geograficas fueron determinadas para todos los rodales conocidos con el proposito de ubicar de una manera precisa su area de distribucion en un mapa y promover su proteccion, conservacion y estudio. Treinta y nueve rodales de Picea chihuahuana localizados en la Sierra Madre Occidental estuvieron agrupados en tres clusters, cada uno separado por 2 a 2.5° de latitud. Los rodales mas sureiios ocurrieron justamente al sur del Tropico de Cancer. El area entera de distribucion de norte a sur fue de 687 km. El promedio de elevacion de los clusters del sur al centro fue de 2675 m, sin embargo, en los rodales pertenecientes a los clusters del norte, estos promediaron 350 m mas bajos en elevacion. Picea chihuahuana estuvo asociada con arroyos en pendientes pronunciadas y barrancas. * To whom correspondence should be addressed. 72 MADRONO [Vol. 47 Picea martinezii fue encontrada en seis rodales en la Sierra Madre Oriental, a una elevacion de cerca de 2250 m a 2650 m y localizados todos a una separacion de 147 km uno del otro. Picea mexicana ocurrio en dos de los puntos mas altos en la Sierra Madre Oriental separados 5 km uno del otro a una elevacion cerca de 3500 m, y sobre uno de los puntos mas altos (3185 m) en Chihuahua en la Sierra Madre Occidental, 676 km al oeste. Es posible que Picea mexicana pudiera ser encontrada sobre uno u otros dos de los puntos mas elevados en la Sierra Madre Oriental. Cada una de las plantulas de Picea chihuahiiana arriba de 0.3 m de altura fue contada, medida, y registrada por presencia de infestacion de muerdago, daiio por incendio, y muerte por despunte de la copa debido a causas desconocidas en 21 rodales. Observaciones similares fueron hechas para otros 18 rodales por Narvaez et al. (1983) con 15 anos de anterioridad. La contabilizacion combinada fue de 42 610 plantas de Picea chihiiahuana, la cual incluyo 24 221 arboles adultos y jovenes fustales arriba de 2 m de alto y 18 389 plantulas abajo de 2 m. La distribucion de clases de diametro en nuestra muestra de 21 rodales tuvo una distribucion de una jota invertida, sugiriendo que la especie se esta reproduciendo. Sin embargo, la proporcion de plantulas a jovenes fustales y arboles adultos fue menos de 1.0 en tados, excepto en cuatro de los 39 rodales, indicando que la especie podria realmente estar en peligro de extincion. Sobre la base del conteo de los anillos de crecimiento en virutas de madera y tocones, Picea chihu- ahiiana puede alcanzar hasta 272 anos de edad. Esto es un ciclo de vida corto comparado con otras piceas norteamericanas. Los arboles mas grandes tuvieron 51 m de alto y de 125 cm a 150 cm de diametro a la altura del pecho y su tamaiio estuvo cerca del promedio comparado con sus congeneres en los Estados Unidos y Canada. Muchos arboles presentaron una condicion pobre o estuvieron daiiados por corta, muerdago, despunte de copa por causas indeterminadas, y fuego. Los datos pueden ser utilizados para monitorar la especie para cambios futuros. La evidencia de fosiles indica que el area de distribucion de Picea chihuahuana estuvo confinada hacia el norte durante el Holoceno, y nosotros especulamos que los rodales localizados mas hacia el sur podrian estar en una condicion mas pobre que aquellos localizados hacia el norte. Contrario a las expectaciones, los rodales localizados hacia el sur no estuvieron en una condicion muy pobre comparados con aquellos localizados hacia el norte, y de hecho, la incidencia de infestacion por muerdago fue mas alta en el norte. Los arboles en los rodales del sur fueron mas grandes y viejos, y la proporcion de plantulas a jovenes fustales y arboles adultos fue mas alta en los rodales del sur y del centro, y mas baja en el norte. The occurrence of Picea A. Dietr. (spruce) in the subtropical latitudes of Mexico is surprising. Spruce is a largely boreal genus which, depending on the taxonomist, includes 31 to 50 species (Dal- limore and Jackson 1923; Wright 1955; Bobrov 1970; Everett 1981). Three species of spruce occur in cool, temperate, montane forests of Mexico. In the Sierra Madre Occidental, spruce forest occurs less than 38 km from the Rio Urique at the bottom of the Barranca del Cobre (the Copper Canyon), where bananas and citrus are grown. Palynological evidence suggests that the spruces of Mexico are relicts stranded by a warming climate during the current interglacial. They may serve the same func- tion as canaries in a coal mine; i.e., by monitoring Mexican spruce populations, we may have an early signal of climate change projected for the next cen- tury (Mahlman 1997). The disappearance of cool temperate conifer forest in Mexico is projected un- der three climate change scenarios (Villers and Tre- jo 1997). The most common spruce in Mexico is Picea chihuahuana Martinez (Chihuahua or prickly spruce), which occurs in the Sierra Madre Occiden- tal, Mexico's western cordillera (Fig. 1). Picea chi- huahuana was first reported in 1942 from a site called Talayotes in the State of Chihuahua (Marti- nez 1953). Locally, it is called cahuite, cahuite es- pinoso, cahuite bravo, and pinabete espinoso, or, by the native Tarahumara Indians of the Barranca del Cobre, matego or mategoco. Vegetation, soils, and climate at sites where it occurs have been described by Gordon (1968), Narvaez, Sanchez, and Olivas (1983), and Narvaez (1984). Picea mexicana Martinez (Mexican spruce) was discovered in 1961 (Martinez 1961) on Sierra La Marta at the border between the States of Nuevo Leon and Coahuila (Fig. 1). Cerro El Morro on Si- erra La Marta is the highest point (ca. 3700 m) in the Sierra Madre Oriental, the range that parallels the east coast of Mexico. A year earlier, spruce had been reported on Cerro Mohinora (Correll 1960), the highest point in Chihuahua (ca. 3300 m), and these were referred to as P. "hybrida" or P. "in- determinada". This spruce was very different from P. chihuahuana but similar to P. engelmannii En- gelm. (Engelmann spruce) that reaches its southern limit in the Chiricahua Mountains of southern Ar- izona about 700 km distant (Little 1971). Using morphological data and multivariate techniques, Taylor and Patterson (1980) found that the trees from Cerro Mohinora clustered with P. mexicana. Subsequently, P. mexicana and P. "indeterminada" were treated as a variety of P. engelmannii by Tay- lor, Patterson, and Harrod (1994); i.e., P. engel- mannii Parry var. mexicana (Martinez) Silba. In 1983 two stands of spruce, which Miiller-Us- ing and Alanis (1984) identified as P. chihuahuana, were discovered in the Sierra Madre Oriental in the State of Nuevo Leon (Fig. 1). However, the nearest populations of P. chihuahuana were at least 510 km distant in the Sierra Madre Occidental, separat- ed from the Sierra Madre Oriental by the arid Mes- eta Central of Mexico. Subsequently, Patterson 2000] LEDIG ET AL.: SPRUCE IN MEXICO 73 and some geographic features mentioned in the text. (1988) decided that these trees were sufficiently dif- ferent from P. chihuahuana to merit species status, and named them P. martinezH T. F. Patterson (Mar- tinez spruce). Needles of P. martinezii are flat and flexible and its cone scales have denticulate mar- gins, but the needles of P. chihuahuana are four- sided and stiff and its cone scales are rounded. The cones and seeds of P. martinezii are larger than those of P. chihuahuana. The ecology of P. mar- tinezii and P. mexicana in the State of Nuevo Leon was described by Capo et al. (1997). The endemic spruces are a minor element in the flora of Mexico, yet potentially important from the standpoint of science, their unique contribution to the biodiversity of Mexico, and their value as ge- netic resources. Picea chihuahuana and P. mexi- cana were included on a list of endangered arboreal taxa prepared for the Instituto Nacional de Inves- tigaciones Forestales y Agropecuarias (INIFAP) by Vera (1990), and all three species qualify as threat- ened under the guidelines of the International Union for the Conservation of Nature and Natural Resources (Sanchez 1984; Sanchez and Narvaez 1990) . Picea chihuahuana occupies sites with some of the richest arboreal species diversities in the Sierra Madre Occidental (Gordon 1968), or in all of tem- perate North America, and for that reason its habitat should be a crucial focus for protection. The Sierra Madre Occidental was nominated by the lUCN as a global center of plant diversity (Anonymous 1991) . Spruce in Mexico may have retreated northward several times, most recently during Holocene warming. Spruce occurred at least as far south as the Isthmus of Tehuantepec (18°09'N) in the mid- Pliocene, 5 million years ago (Graham 1993). Pol- len in the ancient bed of Lake Texcoco, under Mex- ico City, and in Lake Chalco in the basin of Mex- ico, show that spruce occurred in the surrounding uplands at the end of the Pleistocene (Clisby and Sears 1955) and at least as recently as 7000 to 8000 74 MADRONO [Vol. 47 yr before present (B.P.; Lozano-Garcia et al. 1993; M.d.S. Lozano-Garcia, personal communication 1997). The nearest P. chihuahuana are now about 700 km northwest of Mexico City in the Sierra Ma- dre Occidental, and P. martinezii occurs about 500 km north in the Sierra Madre Oriental (Patterson 1988). All P. mexicana may have had ranges as far south as Mexico City, but P. chihuahuana is most likely to have occurred there. The topography of Mexico is more conducive to migration of high el- evation taxa between Mexico City and the Sierra Madre Occidental than between Mexico City and the Sierra Madre Oriental, where P. martinezii oc- curs. In addition, the high endemism of the subal- pine habitats in the Sierra Madre Oriental suggests that they were not linked during the Pleistocene with the Transverse Volcanic Belt in which Mexico City lies (McDonald 1993). In any case, palynolog- ical and genetic studies indicate that the range of spruce retreated northward since the Pleistocene and all Mexican spruces are now characterized by small, fragmented populations (Ledig et al. 1997). The southernmost stands of P. chihuahuana, Ar- royo de la Pista and Arroyo del Chino, lie a few kilometers south of the Tropic of Cancer (23°30'N). Only P. morrisonicola Hayata (Morrison spruce) of Taiwan grows at such southerly latitudes (Wright 1955). No comprehensive description exists to docu- ment the location of the Mexican spruces. Previ- ously published information is incomplete, incon- sistent, and, in some cases, incorrect. Confusion ex- ists in the spelling of place names, which appear in several variants. Coordinates are sometimes wrong. Many important publications on spruce in Mexico are agency reports or proceedings and not easily obtainable. A census provides data important to conservation and a baseline against which to monitor the effects of climate change. The increase in temperature at the end of the last glacial was about equal to that projected after a doubling of atmospheric carbon dioxide, which may occur in less than half a cen- tury (Mahlman 1997). Spruce was so sensitive to the post-glacial warming that it was reduced from a widespread taxon or group of taxa that occurred as far south as Lake Texcoco to isolated relicts in the highest mountain ranges (Ledig et al. 1997). An excellent start was made on censusing P. chi- huahuana by Narvaez et al. (1983), and on P. mar- tinezii by Muller-Using and Alams (1984). In Chi- huahua, Narvaez et al. (1983) counted every P. chi- huahuana then known, but a census had never been undertaken for Durango, where the southernmost stands occur, and where the continued existence of P. chihuahuana may be most in jeopardy because of climatic warming. Our objective was to accurately map the location of all stands of spruce in Mexico, to make them more accessible for scientific study, and to identify sites worthy of conservation. We anticipate that more stands will be found in time, because bota- nists have not yet completely explored the rugged mountains of Mexico. Nevertheless, the present in- ventory is a beginning, and will probably stimulate additional exploration. Our second goal was to complete the census of P. chihuahuana begun by Narvaez et al. (1983) to provide a baseline against which to measure future change, and to characterize the health of the stands as an aid in making man- agement decisions. Methods Each known location of spruce in Mexico was visited in 1997 or 1998 and its coordinates (lati- tude, longitude, and elevation) were determined with a geographic positioning system (GPS, Trim- ble's GeoExplorer). No base stations were available in Mexico, so the data were not corrected for sat- ellite signal degrade introduced by the U.S. De- partment of Defense through selective availability. However, even with selective availability, the co- ordinates should not be off more than 100 m and are probably within 50 m of the actual center of the stand. Elevation may vary substantially within stands, and in most cases we took readings at the center and the extremes. Observations were also made of exposure, slope, and land use. Distance between stands was calculated with an online program (Kindred 1997) and used to group them into geographically coherent clusters using Statistica's clustering module (StatSoft 1995a). Municipio and land ownership (usually an ejido) were determined in the field. Municipios are divi- sions of states, roughly equivalent in size and po- sition in the political hierarchy to counties in the United States. Ejidos occur within municipios and are lands given to groups of peasants after the 1910 revolution (Vargas 1996). Ejido lands are often held and used communally. However, parcels may be owned individually if members of the ejido decide that this is appropriate. By reference to GPS coordinates and topographic features, stand locations were plotted on 1:50,000 maps variously published by: the Comision de Es- tudios del Territorio Nacional (CETENAL), Secre- taria de la Presidencia; the Departamento Carto- grafico, Secretaria de la Defensa Nacional; the In- stituto Nacional de Estadistica, Geografia e Infor- matica (INEGI); or the Coordinacion General de los Servicios Nacionales de Estadistica, Geografia e In- formatica. Duplicate maps are in the files of the Institute of Forest Genetics, Placerville, CA, and the Centro de Genetica Forestal, Chapingo, Mexico. We counted every P. chihuahuana in the State of Durango greater than or equal to 0.3 m tall. Dur- ing the census, the spruces were marked with paint to ensure that none was counted twice and none overlooked. The stands in Durango were censused during April and May, 1997, except for Arroyo del Chino and Arroyo del Agua, which were only dis- 2000] LEDIG ET AL.: SPRUCE IN MEXICO 75 covered in 1997, and were censused in August and September 1998. In 1998, we also counted spruces at La Luisiana, Arroyo de las Ranas, La "Y", Las Lajas, and Llano Grande in the State of Chihuahua. In June 1999, we counted spruce at Arroyo de Que- brada near El Vergel, Chihuahua. Narvaez (1984; Narvaez et al. 1983) censused every stand in Chi- huahua known in 1983, but La Luisiana, Arroyo de las Ranas, La "Y", and Las Lajas were not known until 1997, and he omitted Arroyo de Quebrada be- cause he believed that it was in Durango. Our cen- sus of Llano Grande unintentionally provided a comparison with Narvaez' (1984) counts. In each of the 21 stands censused, we measured height of all spruces <3 m but >0.3 m tall with a measuring pole, and estimated height of the rest. Diameter-breast-high (dbh at 1.4 m) of spruce over 2 m tall was measured with a graduated scale, called a Biltmore stick. In most stands, more ac- curate measurements were made on 1000 m^ plots in the stand interior. The number of plots in a stand varied from three to 13 to provide a sample equal to approximately 10% of the number of trees >10 cm dbh. On the plots, we measured dbh to the near- est millimeter with a diameter tape and height to the nearest meter with a Haga altimeter. In a few stands, plotless samples were measured, and in stands with 15 or fewer trees, we measured every tree with diameter tape and altimeter. We aged one or two of the large trees in a stand using increment cores taken from standing trees. The very largest trees inevitably had a rotten heart, and it was not possible to age them. We supple- mented the increment cores by opportunistically counting rings on stumps. Although P. chihuahu- ana are nominally protected, stumps were observed in many stands. In all, 29 trees were cored, and ring counts were made on 40 stumps. Finally, we scored condition of the P. chihuahu- ana. We recorded the number with mistletoe, and visually scored the severity of infection on a scale from one to three: 1) one-third of the crown in- fected, 2) two-thirds of the crown infected, 3) all of the crown infected. Many spruce had dead tops of unknown cause, and we classified these into three categories: 1) one-third of the crown dead, 2) two-thirds of the crown dead, 3) nearly 100% of the crown dead. To summarize these observations, we constructed weighted scores for mistletoe and for crown dieback: the score was the number of spruce in each class multiplied by the severity rat- ing (1 to 3), divided by the total number of spruce. We also presented the data as the percentage of spruce infected with mistletoe or with crown die- back, regardless of severity. If the tree was fire- scarred, that was also recorded. In reporting the census data, we excluded dead spruce. For regressions of log(height) on log(dbh) and in calculating the means of height and dbh for the interior sample, we excluded dead trees, burned trees, trees with 100% of their crown dead, trees with missing tops, and trees of sprout origin. Statistical analyses, including regressions, anal- yses of variance, and post-hoc comparisons of means, were made with Statistica modules for cor- relation matrices, breakdown and one-way ANO- VA, and tables and banners (StatSoft 1995b). How- ever, most of the data presented here for P. chihu- ahuana represent the entire metapopulation, not samples, so significance testing is not actually nec- essary. All differences are real. Where we present the results of significance tests, it is only to suggest the level of confidence had these stands been a ran- dom sample. Results Thirty-nine stands of P. chihuahuana were lo- cated, six of P. martinezii, and three of P. mexi- cana. We could not locate the stand of 15 P. chi- huahuana trees listed in Narvaez (1984) and San- chez and Narvaez (1990) as Rio Verde, Ejido Ca- tedral. Martinez (1953) noted P. chihuahuana at Arroyos de Urichique, Cuervo, and Meguachic, but we are not certain where he meant. The location of all stands of spruce in Mexico that were known and confirmed as of June 1999 are presented in Tables 1-4 and Fig. 1. Synonyms for place names previ- ously published are listed in the footnotes to Tables 1-4. Picea chihuahuana occurs between 23°20'N (south of the Tropic of Cancer) and 28°39'N. The distance between the northernmost and the south- ernmost stand is 687 km. The stands are grouped in three major areas with large gaps between, al- though Arroyo de Chachamori and Faldeo de Ce- bollitas are sufficiently isolated that they might be considered their own unique "clusters" (Fig. 1). The largest, northern cluster consists of almost all the stands in Chihuahua immediately north of the Barranca del Cobre, from Arroyo de Chachamori south to Rio Vinihueachi. The southern cluster con- sists of four stands in Durango: Arroyo de la Pista, Arroyo del Chino, Arroyo de las Lagunas, and Ar- royo del Infiemo. The central cluster is largely in Durango with the exception of one stand, Arroyo de Quebrada, which is in Chihuahua near its border with Durango. For the 21 stands in Durango and Chihuahua at which we recorded site conditions, all were rocky, 16 abundantly so. In every case, the slope aspect was north to northeast, in only two cases varying as much as 20° west of north. The slope averaged 65%, and was less than 50% only at La Estancia Agua-Amarilla and Arroyo del Infiemo, where some of the trees grew on level ground. The spruce generally occurred near the bottoms of canyons or barrancas, sometimes extending nearly to the ridge. Grazing animals were present in all stands except Arroyo de la Pista, Piedra Rayada, Arroyo del Chi- no, and La 'Y'. Stumps were observed in some MADRONO [Vol. PRU »nal a < 'u cd D X 3 vn _o cd ILHi var fen X U a De ti. cd S o z de o H ph v2 < cd cd Loc Sod cret o 10 cm dbh, and regeneration (seedlings and saplings), which were <10 cm dbh but >0.3 m in height. Our census of the 15 known stands of P. chihuahuana in Durango and six in Chihuahua totaled 10,943 living trees >10 cm dbh and a regeneration of 18,687 (Table 5). In Durango, stands varied from a low of 36 to a high of 6005 spruce (i.e., trees plus regeneration). Several stands in Durango were substantially larger than those in Chihuahua (Table 6), previously censused by Nar- vaez (1984). Apparently, Narvaez (1984) used the criterion of 2 m in height to separate his categories of "rege- neracion" from that of saplings and mature trees ("arbolado joven y adultos"). To make the data sets comparable, we used his criterion, with the result that there were 16,491 trees or saplings over 2 m tall and 13,139 seedlings (or regeneration) in the 15 stands in Durango and the six stands that we count- ed in Chihuahua (Table 6). Combining our data with that of Narvaez (1984), the total number of P. chihuahuana is at least 24,221 saplings and mature trees and 18,389 seedlings, or a total count of 42,610 spruce over 0.3 m tall. We unintentionally included Llano Grande in our census, a stand previously counted by Narvaez (1984), who mistakenly called it Arroyo Ancho. The duplication provided a comparison between his count and ours. The total number of spruce was very close in the two counts, but the number in individual size classes differed somewhat. Because Narvaez' (1984) count was made about 15 years earlier than ours, most trees would have grown in the interim, moving from one class (<2 m tall) to the next (>2 m). We tabulated our data into various height categories, and a division at 3 m in height resulted in excellent agreement between Narvaez' (1984) count and ours; his 545 saplings and trees versus our 528, and his 370 seedlings versus our 378. Actually, our census also counted 15 dead or burned trees, and if these are added to our tree class, the total is 543, almost the same as Narvaez' (1984) number of 545. The difference between counts suggests that there has been a slight reduc- LEDIG ET AL.: SPRUCE IN MEXICO 78 MADRONO [Vol. 80 J3 >. o o o .2 fc 2So 03 3 C/5 C _o > £ 2 cj o 1=! 3 11 P Z c 1-1 !^ O ,0 C/3 Z o - g ^ ^ (u o (U ° ^ N c 5 ^ 2 ^ H B r- C g o 3 ON O N -S . S2 -a u c« O o Oh I 4-i ^ 3 «^ 3 IJ N ^ 2 O IS 2 3 cd (+1 o O 3 2 cd o o- •3 OJ ^§1 to O ' c/3 >^ o (u CQ -3 .. 2 S II o ^ (U cd o f-H On b2 < : DS ON 'S >^ O _ 00 - C ON • 3 g CJ 3 XI Qh ^ o o (u 03 O w (U O =^ ^ o r >> o n o 3 cd GO P -3 w ^ <-» pq (U 2Z (N m 0 in 0 0 in 0 (N m in 0 (N 00 in in in in in in 0^ in >n in in in in in in in in in 0 0 0 0 0 0 0 0 0 0 0 0 r- r- r- r- (N (N (N (N (N (N (N (N (N (N O in \D O in in 00 ^ m m (N CM 000 (N 00 in (N r*-) (N (N (N o > a o 3 QJ t~-~ vflj (N ■I o o o o o o <-c<-c< |< |<-c< |< |< |< |<-c5-c 5|S|5 go g5|5 = 5 go go §5|o| ^ ^ ^ ^ so o o o 3 O o o PQ o o o OQ -O ^3 > 'S 3 o S !/3 O ■^^ II •c ^ J o 03 3 o o PQ 3 O CJ O PQ 03 3 O o o CQ ^ g ^ ^ ^ (u o (U PLh PLh cd (U a o 2 o w w w w :3 O 2000] LEDIG ET AL.: SPRUCE IN MEXICO 79 IT) o in (N 'vO (N (N r*-) (N (N (N 10 cm dbh, mean height of sampled trees (i.e., trees in the interior sample plots) varied among stands from 9.8 m at La Lou- isiana to 25.9 m at Arroyo del Infiemo and mean diameter varied from 22.5 cm at Piedra Rayada to 55.6 cm at Arroyo del Infiemo (Table 5). In most cases these values were larger than the mean heights and diameters for the entire stand, but are emphasized here because they were taken with greater precision (Haga altimeter and diameter-tape versus ocular estimate and Biltmore stick). Mean height and diameter tended to be greatest in the southern cluster (20.5 m and 42.3 cm), less in the central cluster (16.2 m and 31.7 cm), and least in the four stands that we measured in the northern cluster (12.2 m and 31.7 cm). If these were sam- ples, the differences, between the southern and northern clusters would be statistically significant according to Scheffe's test. Maximum height and diameter (using data for the entire stand) also de- creased south to north: mean maximum heights were 45.0, 37.3, and 24.5 m, and mean maximum diameters were 105.9, 98.8, and 59.8 cm, respec- tively, for the southern, central, and northern clus- ters. Again, if these were samples, the southern and central clusters would differ significantly from the northern cluster by Scheffe's test. The logarithms of height and diameter were closely related in all 21 stands in which we took data (Fig. 4). Because the number of observations was so large, tests for homogeneity of regression coefficients (p. 319 in Steel and Torrie 1960) for height on diameter indicated that the slopes differed both between and within clusters. Analysis of vari- ance also showed that the intercepts differed among clusters, and Scheffe's test indicated that the inter- cepts for the northern cluster were significantly dif- ferent from those for the central and southern clus- ters. Although the differences were small, the pat- tern was consistent; the smallest intercepts (most negative) and largest slopes occurred in the north and the greatest intercepts and smallest slopes in the south (Table 7). For a given diameter, trees were 80 MADRONO [Vol. 47 o ON in 3 O in in m m o c ON 0 ON o ON o ON o ON o o o ON On ON ON ON o .5^ in o o o o o ^ (N O (N 00 in 00 ^ ^ e3 1 -2 2 m ^ 2 ^ o C o Oh W 24.5 25,5 26 5 LATITUDE (degrees) Fig. 2. Scatterplot of elevational coordinates versus lat- itudinal coordinates for 39 stands of Chihuahua spruce. coordinates for Arroyo de Chachamori are about 2' in error for latitude and 1.5' for longitude; Narvaez (1984) and Narvaez et al. (1983) gave the coordi- nates for Arroyo de Chachamori in degrees and minutes only, suggesting that they could not locate it with much precision on topographic maps. For P. chihuahuana, stands in the northern clus- ter were lower in elevation, on average, than those in the central cluster. A species' elevational bounds generally decrease with increasing latitude. In Cal- ifornia, species' ranges decrease 172 m in elevation for each increase of a degree in latitude, and in the Rocky Mountains, the change is 77 m (Barbour and Minnich 1999). According to Hopkins' (1938) Law, a change of 1° in latitude is equivalent to 122 m in elevation. Therefore, we expected the central cluster to be 305 m lower in elevation than the southern cluster, however, stands occur at virtually the same elevation in the south and the central clus- ters. A decrease of 244 m should be expected be- tween the central and northern clusters, but the de- crease is actually 352 m. The total decrease in el- evation between southern and northern clusters was 379 m, much less than the expected value of 549 m predicted by Hopkins' Law. The explanation may be that P. chihuahuana is restricted to sites that attenuate the effects of latitude. More detailed analysis is not possible because climate data are not available for the remote sites where P. chihuahuana are found. Our census is very accurate and we doubt that the counts deviate from actual numbers by more than one percent. Our count at Llano Grande de- viates from that of Narvaez (1984) by less than 1%, despite being separated in time by about 15 years. Fifteen of the trees counted by Narvaez (1984) now seem to be dead, although still standing. In contrast to the agreement between Narvaez' (1984) count 82 MADRONO [Vol. 47 Table 5. Number of Living Trees >10 cm Diameter Breast High (DBH), Regeneration <10 cm DBH and >0.3 M Tall, Mean Height of Sampled Trees, Mean DBH of Sampled Trees, Mean Density of Trees on Sample Plots, AND Condition of all 15 Stands of Chihuahua Spruce in Durango and 6 in Chihuahua. ' Infection by mistletoe, recorded in three severity classes. ^ Dead top from unknown cause(s), recorded in three severity classes. ^ Damaged by fire. The number of affected spruce in each severity class weighted by the class value (1-3) and divided by the total number of trees plus regeneration. ^ The number of spruce affected expressed as a percentage of the total number of spruce >0.3 m in height. XX • . x^x^xx x^ • Mistletoe^ Top dieback- Height DBH Density 1 Scar^ Stand Trees Regen. (m) (cm) (ha"') Score'* %^ Score^ %^ (%) r^Lixjyyj \j.c icL r i?>Lci 599 809 24.7 40 10 15.3 185 1930 16.1 Arrovo HpI Tnfiprno 107 174 25.9 Faldeo de CeboUitas 83 172 15.5 Arroyo de los Angeles 1570 4435 16.7 Estancia-Agua Amarilla 834 855 17.9 La Medalla 694 1174 16.7 Arroyo del Agua 375 420 15.5 La Medallita 264 218 16.0 El Saltito 472 732 19.9 Arroyo de Rosales 14 22 19.4 Arroyo del Indio Ignacio 1563 2587 13.5 Piedra Rayada 2342 3204 12.5 Arroyo de Enmedio 316 832 15.4 Arroyo de Quebrada 765 475 15.4 La "Y" 11 10 11.7 Llano Grande 480 407 13.5 Las Lajas 15 11 10.8 La Luisiana 92 90 9.8 Arroyo de las Ranas 122 120 15.3 46.7 63.0 0.001 0.0 0.351 19.5 5.0 36.4 36.6 0.340 26.0 0.500 44.0 46.0 30.6 97.5 0.00 0.0 0.010 0.9 0 Q 55.6 21.7 0.00 0.0 0.125 6.8 1.4 32.3 22.0 0.00 0.0 0.122 6.3 ^ 5 J.J 26.0 161.1 0.190 13.3 0.100 7.9 4.5 32.7 115.7 0.139 11.2 0.377 19.3 11.0 31.8 112.9 0.054 3.4 0.065 4.8 3.5 33.4 78.3 0.486 26.5 0.211 13.8 5.0 32.2 62.5 0.224 16.4 0.145 8.5 6.2 38.3 84.0 0.151 10.6 0.114 7.1 4.2 44.0 0.361 30.6 0.306 25.0 13.9 25.5 130.0 0.114 7.9 0.118 8.0 3.2 22.5 187.7 0.111 8.2 0.097 7.2 4.3 27.2 126.7 0.072 3.7 0.083 3.3 0.6 34.0 122.9 0.348 25.7 0.379 26.0 17.3 27.7 0.476 33.3 0.286 28.6 4.8 30.0 0.306 21.5 0.268 20.4 14.5 33.5 0.462 26.9 0.346 19.2 11.5 28.4 0.253 18.7 0.275 22.0 9.9 38.8 0.182 14.9 0.161 12.4 22.3 and ours, our count of 107 stems >10 cm dbh, or even 56 stems >30 cm, dbh is in excess of the 36 trees and 9 seedlings recorded at Arroyo del Infier- no by Gordon (1968). Combining our data with that of Narvaez (1984), the total number of P. chihuahuana >2 m in height is a sizeable figure, 24,221, and the number of seed- lings <2 m and >0.3 m is about 18,389. Density of P. chihuahuana >10 cm dbh, derived from the 1000 m^ plots located in the centers of the stands, averaged 94.8 ha ' with a range for 15 stands from 21.7 to 187.7 ha ' (Table 5). Although the data are not completely comparable, density was slightly lower in the 1 8 stands reported by Nar- vaez et al. (1983), which had a mean of 72.8 and a range of 25.4 to 1 17.0 trees >2 m tall ha '. Picea chihuahuana stands are similar to but, perhaps, slightly less dense than stands of P. breweriana S. Watson, (Brewer spruce) which may range from about 125 to 320 trees >10 cm dbh ha ' (Thorn- burgh 1990), or stands of P. sitchensis (Bong.) Car- riere, (Sitka spruce) which have about 188 trees ha ' (Harris 1990). However, density of boreal and other north temperate spruces is much higher: 815 to 1324 trees ha ' for well stocked P. glauca (Moench) Voss (white spruce) stands in Alaska and Saskatchewan (Nienstaedt and Zasada 1990); 1110 to 1780 trees ha ' of P. mariana (Mill.) B.S.P. (black spruce) in Ontario (Viereck and Johnston 1990); 140 to 780 P. engelmannii >4 inches (ap- proximately 10 cm) dbh in mixed spruce-fir forests in Wyoming (Costing and Billings 1951); and 121 to 480 trees ha ' P. rubens Sarg. (red spruce) in the Maritime Provinces and Maine (Blum 1990; Costing and Reed 1942). Though we have not counted the trees of P. mex- icana, they are numerous, perhaps in the thousands, at all three confirmed sites. Rushforth (1986) mis- takenly reported that the stand at Sierra La Marta was reduced to six trees at the top of the range. A fire in 1975 destroyed the type locality on the lower slopes of Sierra la Marta, and most of the spruces, but many hundreds escaped damage in a Canada (a precipitous cleft smaller than a canyon) high on the mountain. Muller-Using and Alanis (1984) counted 68 P. martinezii ^10 cm dbh at La Tinaja and 350 at Canon el Butano. The number at Agua Frfa, re- cently discovered, may exceed the number at Ca- non el Butano. Nevertheless, P. martinezii is ex- tremely rare. Cur data from increment-cored trees was insuf- ficient to construct precise height-age and dbh-age relationships because of the limited range in age of the trees we bored. However, Gordon (1968) pre- sented relationships based on stem analyses and Narvaez (1984) presented scatterplots of height and diameter on age for 37 trees spanning a range from about 20 to 120 yr. Based on Narvaez' (1984) scat- terplots, spruce on the ten sites he sampled in Chi- 2000] LEDIG ET AL.: SPRUCE IN MEXICO 83 Table 6. Numbers of Trees and Saplings (T > 2 m Tall), Seedlings (S < 2 m Tall), and the Ratio Seedlings/ Trees + Saplings for all 39 Confirmed Stands of Chihuahua Spruce. ' We did not include seedlings <0.3 m tall; data for Narvaez (1984) apparently does. - Counts made in April and May 1997. ^ Counts made in August and Sep- tember 1998. ^ Count made in June 1999. ^Counts reported in Narvaez (1984), which differ from counts in Narvaez et al. (1983). During seed collections, we found only 17 trees, most with dead tops, and stumps, indicating that the population has declined in the last 15 years. ^ Called Llano Grande by Narvaez (1984) and Narvaez et al. (1983). Stand T 1 C 1 Total Arroyo de la Pista^ 919 489 1408 0.532 Arroyo del Chino^ 46 4 50 0.087 Arroyo de las Lagunas- 505 1610 2115 3.188 Arroyo del Infierno- 148 133 281 0.899 Faldeo de Cebollitas- 172 83 255 0.483 Arroyo de los Angeles- 2507 3498 6005 1.395 La Estancia-Agua Amarilla- 1195 494 1689 0.413 La Medalla- 1012 856 1868 0.846 Arroyo del Agua- 510 285 795 0.559 La Medallita- 356 126 482 0.354 El Saltito- 656 548 1204 0.835 Arroyo de Rosales- 21 15 36 0.714 Arroyo del Indio Ignacio- 2628 1522 4150 0.579 Piedra Rayada- 3564 1982 5546 0.556 Arroyo de Enmedio- 465 683 1148 1.469 Arroyo de Quebrada"^ 877 363 1240 0.414 Rio Vinihueachi^ 1785 1579 3364 0.885 El Pinabetal^ 455 267 722 0.587 Las Trojas^ 874 780 1654 0.892 Napahuichi I'' 1064 921 1985 0.866 Napahuichi II-'' 209 150 359 0.718 Talayotes^ 291 299 590 1.027 La "Y"^ 13 8 21 0.615 Situriachi"^ 389 286 675 0.735 Las Aguilas'' 548 168 716 0.307 El Realito^ 587 210 797 0.358 El Cuervo^ 140 96 236 0.686 El Ranchito^ 217 162 379 0.747 La Tinaja^ 99 37 136 0.374 Cerro de la Cruz^-^ 20 5 25 0.250 Llano Grande*^ 545 370 915 0.679 Llano Grande^ 614 273 887 0.445 Arroyo Ancho^^ 127 8 135 0.063 Las Lajas^ 19 7 26 0.368 La Luisiana^ 127 55 182 0.433 Mategoina P 124 29 153 0.234 Mategoina IP 448 164 612 0.366 Mategoina III"^ 207 65 272 0.314 Arroyo de las Ranas"* 137 105 242 0.766 Arroyo de Chachamori^ 146 24 170 0.164 Total 24,221 18,389 42,610 0.759 huahua would be about 29 m tall at 100 yr of age. Gordon's (1968) stem analyses suggest a height of 28 m at 100 yr on a moist site at Arroyo del In- fiemo in Durango, but only 12 m for a poorer site. Based on a regression of height on age for the 29 trees we bored (r = 0.49), we would expect a height of about 24 m at age 100. Thus, P. chihuahuana probably grows slowly, about one-quarter to one- third of a meter per year, on average, over its first 100 years. Compared to other North American spruces, P. chihuahuana does not seem particularly long-lived. Most of the large P. chihuahuana that we cored had between 100 and 200 rings at breast height. How- ever, the largest trees had heart rot, which made it impossible to age them. Sanchez and Narvaez (1990) reported finding wood rot in 40% of adult trees, possibly caused by fungi belonging to one or more of the genera Alternaria, Helminthosporum, Nigrospora, Sporotrichum, or Trichoderma. The oldest trees that we bored were 272- and 244-yr- old at breast height, and several stumps had over 200 rings. By contrast to P. chihuahuana, the oldest P. breweriana might be 900 yr (Waring, Emming- ham, and Running 1975); P. sitchensis may live to 700 or 800 yr (Harris 1990); trees of P. engelmannii between 500 and 600 yr of age are "not uncom- mon" (Alexander and Shepperd 1990) and can sur- 84 MADRONO [Vol. 47 80 70 50 40 30 20 10 1 1 1 Southern Central Northern iiiiiili^^ I lOiOiOiOiOLOLOiO T- CO -"^ CD r-- C5) o in LO ■t- CO lo m in to m c\i csj 1^ (D r-- o) o in in in in in eg c\i CD cn o DBH Classes Fig. 3. Frequency of observations in 5 cm-dbh classes plotted over the class mid-point for all Chihuahua spruce >0.3 m tall in the southern, central, and northern clusters of stands. vive to 680 yr (Brown et al. 1995); P. pungens Engelm. (blue spruce) may survive to 600 yr or more (Fechner 1990); and the maximum age in P. rubens is about 400 yr (Blum 1990). The oldest P. glauca on good sites may be 250 to 300 yr, similar to P. chihuahuana, but above the Arctic Circle, slow-growing P. glauca may reach 1000 yr (Nien- staedt and Zasada 1990). Picea mariana, alone of the North American spruces, may be slightly short- er-lived than P. chihuahuana, but P. mariana of 280 yr have been reported (Viereck and Johnson 1990). Muller-Using and Lassig (1986) found a P. martinezii with a ring count of 279 at breast height. With regard to ultimate size, P. chihuahuana is average among the North American spruces. The largest trees were in the southern cluster and were about 50 m in height and 125 cm dbh. Size seemed to decrease from the southern to the northern clus- ter. The largest P. martinezii were only 32 m tall and 62 cm dbh (Miiller-Using and Lassig 1986). The largest P. breweriana, P. engelmannii, and P. pungens are very close in height and diameter to P. chihuahuana (Thomburgh 1990; Waring et al. 1975; Alexander and Shepperd 1990; Fechner 1990). Picea glauca may be slightly smaller, the largest being about 55 m by 120 cm (Nienstaedt and Zasada 1990). Picea mariana and P. rubens are substantially smaller, reaching 27 m height by 46 cm dbh and 35 m height by 61 cm dbh, respectively (Viereck and Johnson 1990; Blum 1990). On the other extreme, P. sitchensis is a giant, the largest being about 66 m by 510 cm (Harris 1990). Many P. chihuahuana were in poor condition or damaged. Within the 21 stands that we studied, from 1% to 44% of the trees had dead tops from unknown causes. The severity varied from a third to almost the complete crown, and incidence tended to be higher in smaller stands (Table 5). Dead tops were also observed by Narvaez (1984) and Narvaez et al. (1983), but only in 1.3% of the spruces in the State of Chihuahua. We found that 19% of the spruces had some degree of dieback in the five stands we observed in the northern cluster, includ- ing 181 of 887 spruce (20%) at Llano Grande, which was also scored by Narvaez (1984). Narvaez (1984) recorded only 4 of 915 spruce with top die- back, "puntisecos". Perhaps, the condition has be- come substantially worse in the 15 years between observations. Or, perhaps, Narvaez (1984) scored only the most severely affected trees. In our sample of five stands in Chihuahua, only 2.2% of the trees >10 cm dbh had tops with complete dieback, but 33.2% had dieback of less severity. In any case, damage from top dieback and other causes is extensive, and further threatens this rare species. Fire scars were noted in every stand, and reached high proportions, up to 46% in some small- er stands. In the survey of Sanchez and Narvaez (1990), only 3.4% of the spruce were damaged by fire. Even though P. chihuahuana is nominally pro- 2000] LEDIG ET AL.: SPRUCE IN MEXICO 85 0.0 0.6 1.2 1.8 2.4 0.0 0.6 1.2 1.8 2.4 Log(DBH) Fig. 4. Regression of log(height in m) on Iog(dbh in cm) for a sample of 190 Chihuahua spruces from the southern cluster of stands (Arroyo del Infierno, Arroyo de la Pista, Arroyo de las Lagunas, and Arroyo del Chino). tected, clandestine cutting occurs, as evidenced by the 40 stumps on which we made ring counts. Many other trees were topped for Christmas deco- rations. Our observation of mistletoe seems to be the first report on P. chihuahuana (Hawksworth and Wiens 1996), and we have not identified the species. We observed mistletoe on spruce in some stands, but not in all. Mistletoe was least common and the fre- quency of infected trees was least in the southern cluster and greatest in the northern cluster, where it exceeded 19%. Sanchez and Narvaez (1990) did not mention mistletoe. Dwarf mistletoes are dam- aging to all other North American spruces north of Mexico. Arceuthobium microcarpum causes heavy mortality to P. engelmannii in Arizona and New Mexico, but is not found farther north (Alexander and Shepperd 1990). It also colonizes P. pungens (Fechner 1990). In some stands, 36% of P. brew- eriana are parasitized by A. campylodum (Thom- burgh 1990), which may destroy infected trees. A. pusillum is common and destructive on black spruce in eastern North America, but absent in the West (Viereck and Johnston 1990), and it also col- onizes P. glauca (Nienstaedt and Zasada 1990), and, occasionally, P. rubens (Blum 1990). Some stands seemed to be in poorer health than others, and these included some of the smallest stands. For example, the number of spruce at Ar- royo del Chino was only 50, and 44% of these had tops affected to various degrees by dieback, 46% were scarred by fire, and 26% had mistletoe. In addition, the ratio of seedlings to trees and saplings was only 0.087. It may be very difficult to conserve such stands, and we suggest that seeds or cuttings be collected as soon as possible to conserve their genetic resources. The most threatened stands can be identified in Tables 5 and 6. The most critical question with regard to the con- servation of P. chihuahuana is whether the stands are regenerating adequately. The distribution of spruce into 5 cm diameter-size classes forms a re- verse J-shape (Fig. 3) in most stands, particularly the large ones. This suggests that the older trees will be replaced. However, this may be misleading. The ratio of seedlings <2 m tall to trees and sap- lings >2 m is less than 1.0 in all except four of 39 stands for which data exist. One of these, Arroyo Table 7. Regression Coefficients (B) and Intercepts (A) of Log(Height in m) on Log(dbh in cm), with Standard Errors (in Parentheses), and Correlation Coefficients (R) in Chihuahua Spruce from Northern, Central, and Southern Clusters. ' Five stands are represented in the northern cluster, 12 in the central, and 4 in the southern. - Number of trees, from the interior plots in each stand. Cluster' n-^ b a r northern 341 0.9524 (0.0221) -0.3145 (0.0306) 0.920 central 1235 0.8704 (0.0097) -0.0944 (0.0125) 0.938 southern 193 0.8458 (0.0188) -0.0461 (0.0247) 0.956 86 MADRONO [Vol. 47 de las Lagunas, was heavily logged somewhat over a decade ago, removing many of the large trees, which may explain the high ratio of 3.19 seedlings to saplings and trees. In many stands, the number of trees >10 cm dbh exceeds the number of seed- lings <2 m tall, suggesting that the stands are in trouble. We believe that the ratio of seedlings to trees is significant because P. chihuahuana probably regen- erates in the shade. Although ecological studies of P. chihuahuana are lacking on this point, other North American spruces, such as P. breweriana, P. engelmannii, P. glauca, P. rubens, P. mariana, and P. sitchensis are tolerant of shade. In fact, P. brew- eriana and P. engelmannii seedlings cannot survive strong sunlight (Thomburgh 1990; Alexander and Shepperd 1990; Ronco 1975), and regeneration of P. pungens, P. rubens, and P. sitchensis benefits from shade (Fechner 1990; Blum 1990; Harris 1990). All spruces north of Mexico are thin-barked and highly susceptible to fire, and do not normally require disturbance for regeneration, with the ex- ception of the partially serotinous P. mariana. The ratio of seedlings to saplings and trees in some other North American spruces seems at least twice that observed in P. chihuahuana. However, comparison with other studies is difficult because of differences in methodology and presentation of data. In P. rubens, the ratio of stems <10 ft tall to stems >10 ft was 1.87 for a 60-yr-old stand in Maine (Costing and Reed 1942). In eight stands of P. engelmannii in Wyoming, the ratio of stems <8 ft tall to stems >8 ft averaged 2.34, ranging from 0.24 to 7.33 (Costing and Reed 1952). For seven stands in the Smoky Mountains of North Carolina and Tennessee, the ratio of P. rubens stems <2 inches dbh to stems >2 inches was 8.45, and in four stands in the White Mountains of New Hamp- shire, the ratio was 23.81, with a range of 22.26 to 26.85 (Costing and Billings 1951). A 2-inch di- ameter limit is roughly equivalent to a 3 m tall P. chihuahuana (from Table 7). If Costing and Bill- ings' (1951) 1-inch diameter limit is used (equiv- alent to a 1.5 m tall P. chihuahuana), the ratios of P. rubens regeneration to trees was 5.24 in the Smoky Mountains and 13.34 in the White Moun- tains. Global warming is a threat to the cool temperate conifer forest of Mexico (Villers and Trejo 1997). Therefore, we might expect the stands in the south- em cluster to be in greater decline than those in the central cluster, and those in the central cluster to be in greater decline than those in the northern cluster. This does not yet seem to be the case. The ratio of seedlings to saplings and trees was 1.18, 0.72, and 0.53 for the southern, central, and northern clusters, respectively. However, trees were larger on average in the south and smallest in the north, suggesting that northern stands are younger and will survive longer without replacement. Acknowledgments This study was an undertaking of the Forest Genetic Resources Study Group/North American Forestry Com- mission/Food and Agricultural Organization of the United Nations. It was completed with the help of National Re- search Initiatives Competitive Grant Program award no. 95-37101-1916 and PSW Global Change Research Grant 95-02. We are grateful to Jesus Sanchez-Cordova for shar- ing his great knowledge of Chihuahua spruce with us, and thank James A. Baldwin, Michael G. Barbour, G. E. Reh- feldt, and two anonymous reviewers for helpful comments that greatly improved the manuscript. Literature Cited | j Alexander, R. R. and W. D. Shepperd. 1990. Picea en- ' gelmannii Parry ex Engelm. Engelmann spruce. Pp. 187-203 in R. M. Burns and B. H. Honkala, tech. | coords., Silvics of North America. Vol. 1, Conifers. Agriculture Handbook 654. USDA, Forest Service, | Washington, DC. Anonymous. 1991. NS/S nominates Sierra Madre as a global center of plant diversity. The Seedhead News No. 34:4. Native Seeds/SEARCH, Tucson, AZ. j Barbour, M. G. and R. A. Minnich. 1999. Californian j upland forests and woodlands. Pp. 159-200 in M. G. I Barbour and W. D. Billings (eds.). North American | terrestrial vegetation, 2nd ed. Cambridge University \ Press, NY. j Blum, B. B. 1990. Picea rubens Sarg. red spruce. Pp. [ 250-259 in R. M. Burns and B. H. Honkala (tech. i coords), Silvics of North America. Vol. 1, Conifers. Agriculture Handbook 654. USDA, Forest Service, Washington, DC. BoBROV, E. 1970. History and systematics of the genus ! Picea A. Dietr. [in Russian]. Novosti Sistematiki Vys- 1 shikh Rastenii 7:5-40. I Brown, P. M., W. D. Shepperd, C. C. Brown, S. A. Mata, j AND D. L. McClain. 1995. Oldest known Engelmann | spruce. U.S. Forest Service Research Note RM-RN- 534. I Burns, R. M. and B. H. Honkala (tech. coords.). 1990. Silvics of North America. Vol. 1, Conifers. Agricul- j ture Handbook 654. USDA, Forest Service, Washing- J ton, DC. j Capo-Arteaga, M., S. Valencia-Manzo, C. Flores- |I Lopez, and S. Braham-S. 1997. Informe final de ac- \ tividades del proyecto: autoecologia del genero Picea \ en Nuevo Leon. Presentado a: Consejo Consultivo Estatal para la Preservacion y Fomento de la Flora y Fauna Silvestre de Nuevo Leon. Universidad Auton- oma Agraria Antonio Narro, Buenavista, Saltillo, Coahuila. Clisby, K. H. and p. B. Sears. 1955. Palynology in south- em North America, part III: microfossil profiles under Mexico City correlated with the sedimentary profiles. Bulletin of the Geological Society Amer 66:51 1-520. CoRRELL, D. S. 1960. A mule-train trip to Sierra Mohi- nora. Chihuahua. American Fern Journal 50:66-78. i Dallimore, W. and a. B. Jackson. 1923. A handbook of Coniferae including Ginkgoaceae. Edward Arnold and Company, London. Everett, T. H. 1981. The New York Botanical Garden illustrated encyclopedia of horticulture. Vol. 9 Par-Py. Garland Publishing, NY-London. Fechner, G. H. 1990. Picea pungens Engelm. blue spruce. Pp. 238-249 in R. M. Burns and B. H. Honkala (tech. 2000] LEDIG ET AL.: SPRUCE IN MEXICO 87 coords.), Silvics of North America, Vol. 1, Conifers. Agriculture Handbook 654. USDA, Forest Service, Washington, DC. Gordon, A. G. 1968. Ecology of Picea chihuahuana Mar- tinez. Ecology 49:880-896. Graham, A. 1993. Historical factors and biodiversity in Mexico. Pp. 109-127 in T. P. Ramamoorthy, R. Bye, A. Lot, and J. Fa (eds.). Biological diversity of Mex- ico: origins and distribution. Oxford University Press, NY-Oxford. Harris, A. S. 1990. Picea sitchensis (Bong.) Carr. Sitka spruce. Pp. 260-267 in R. M. Burns and B. H. Hon- kala (tech. coords.), Silvics of North America. Vol. 1, Conifers, Agriculture Handbook 654. USDA, Forest Service, Washington, DC. Hawksworth, F. G. and D. Wiens. 1996. Dwarf mistle- toes: biology, pathology, and systematics. Agricultur- al Handbook 709. USDA, Forest Service, Washing- ton, DC. Hopkins, A. D. 1938. Bioclimatics: a science of life and climate relations. Miscellaneous Publication No. 280. USDA, Washington, DC. Kindred, D. "How far is it?" Distance calculator, (http:/ /www.indo. com/distance/, accessed 1997). Ledig, F. T, V. Jacob-Cervantes, P. D. Hodgskiss, and T. Eguiluz-Piedra. 1997. Recent evolution and di- vergence among populations of a rare Mexican en- demic. Chihuahua spruce, following Holocene cli- matic warming. Evolution 51:1815-1827. Little, E. L., Jr. 1971. Atlas of United States trees. Vol. 1. Conifers and important hardwoods. Miscellaneous Publication No. 1146. Forest Service, USDA, Wash- ington, DC. Lozano-Garcia, M. S., B. Ortega-Guerrero, M. Cab- allero-Miranda, and J. Urrutia-Fucugauchi. 1993. Late Pleistocene and Holocene paleoenvironments of Chalco Lake, central Mexico. Quaternary Res. 40: 332-342. Mahlman, J. D. 1997. Uncertainties in projections of hu- man-caused climate warming. Science 278:1416- 1417. Martinez, M. 1953. Las Pinaceas Mexicanas. Subsecre- taria de Recursos Forestales y de Caza, Secretaria de Agricultura y Ganaderia, Mexico. Martinez, M. 1961. Una nueva especie de Picea in Mex- ico. Anales del Instituto de Biologi'a de la Universi- dad Nacional de Mexico 32:137-142. McDonald, J. A. 1993. Phytogeography and history of the alpine-subalpine flora of northeastern Mexico. Pp. 681-703 in T. P Ramamoorthy, R. Bye, A. Lot, and J. Fa (eds.). Biological diversity of Mexico: origins and distribution. Oxford University Press, NY-Ox- ford. Muller-Using, B. and G. Alanis-Flores. 1984. Nuevos registros del pinabete de Chihuahua {Picea chihuahu- ana Martinez) en Nuevo Leon propuesta para la pro- teccion legas de dos areas de especial interes ecolo- gico. Pp. 130-132 in Reunion Regional de Ecologia Norte, 25, 26 y 27 de abril 1984, Monterrey, Nuevo Leon, Mexico. Secretaria de Desarrollo Urbano y Ecologia, Subecretaria de Ecologia, Mexico. MCller-Using, B. and R. Lassig. 1986. Zur Verbreitung der Chihuahua-Fichte {Picea chihuahuana Martinez) in Mexiko. Mitt. Deutsch. Dendrol. Ges. 76:157-169. Narvaez-Flores, R. 1984. Contribucion al conocimiento de la ecologia de Picea chihuahuana Martinez. Un- published thesis. Universidad Autonoma de Nuevo Leon, Monterrey, Nuevo Leon. Narvaez-Flores, R., J. Sanchez-Cordova, and A. Oli- vas-Meza. 1983. Distribucion y poblacion de Picea chihuahuana [in Spanish]. Nota Tecnica No. 6, Cen- tro de Investigaciones Forestales del Norte, Instituto Nacional de Investigaciones Forestales, Secretaria de Agricultura y Recursos Hidraulicos. NiENSTAEDT, H. AND J. C. Zasada. 1990. Picea i^lauca (Moench) Voss white spruce. Pp. 204-226 in R. M. Burns and B. H. Honkala, tech. coords., Silvics of North America, Vol. 1, Conifers. Agriculture Hand- book 654. USDA, Forest Service, Washington, DC. OosTiNG, H. J. AND W. D. BiLLiNGS. 1951. A Comparison of virgin spruce-fir forest in the northern and southern Appalachian system. Ecology 32:84-103. OosTiNG, H. J. AND J. F. Reed. 1942. The establishment of a white birch community on cutover pulpwood land in northwestern Maine. Bull. Torrey Bot. Club 69:647-660. OosTiNG, H. J. AND J. F. Reed. 1952. Virgin spruce-fir of the Medicine Bow Mountains, Wyoming. Ecological Monographs 22:69-91. Patterson, T. F. 1988. A new species of Picea (Pinaceae) from Nuevo Leon, Mexico. SIDA 13:131-135. RONCO, F. 1975. Diagnosis: "sunburned" trees. Journal of Forestry 73:31-35. RusHFORTH, K. 1986. Mexico's spruces — rare members of an important genus. Kew Magazine 3:119-124. Sanchez-Cordova, J. 1984. Picea chihuahuana Mrtz., una conifera en peligro de extincion. Ciencia Forestal, Revista del Instituto Nacional de Investigaciones Fo- restales, Secretaria de Agricultura y Recursos Hidrau- licos 9(no. 51):51-63. Sanchez-Cordova, J. and R. Narvaez-Flores. 1990. Plan integral para la proteccion y fomento de Picea chihuahuana Martinez. Ecosistemas Forestales, Insti- tuto Nacional de Investigaciones Forestales y Agro- pecuarias, Secretaria de Agricultura y Recursos Hid- raulicos 1:37-55. StatSoft. 1995a. Statistica, Vol. Ill: Statistics II. StatSoft, Inc., Tulsa, OK. StatSoft. 1995b. Statistica, Vol. I: General Conventions and Statistics I. StatSoft, Inc., Tulsa, OK. Steel, R. G. D. and J. H. Torrie. I960. Principles and procedures of statistics with special reference to the biological sciences. McGraw-Hill, NY. Taylor, R. J. and T. F. Patterson. 1980. Biosystematics of Mexican spruce species and populations. Taxon 29: 421-469. Taylor, R. J., T. F Patterson, and R. J. Harrod. 1994. Systematics of Mexican spruce — revisited. Syst. Bot. 19:47-59. Thornburgh, D. 1990. Picea hreweriana Wats. Brewer spruce. Pp. 181-186 in R. M. Burns and B. H. Hon- kala (tech. coords.), Silvics of North America, Vol. 1, Conifers. Agriculture Handbook 654. USDA, Forest Service, Washington, DC. Vargas-Hernandez, J. J. 1996. Forest land ownership and management in Mexico. P. 21 in D. L. Rogers and F. T. Ledig (eds.). The status of temperate North American forest genetic resources. Report no. 16, University of California Genetic Resources Conser- vation Program, Davis, CA. Vera, G. 1990. Algunas especies arboreas raras o amen- azadas en Mexico. Secretaria de Agricultura y Re- cursos Hidraulicos (SARH), Instituto Nacional de In- vestigaciones Forestales y Agropecuarias (INIFAP), Mexico. ViERECK, L. A. AND W. F. JoHNSTON. 1990. Picca mariana 88 MADRONO [Vol. 47 (Mill.) B.S.P. black spruce. Pp. lll-l^l in R. M. Bums and B. H. Honkala (tech. coords.), Silvics of North America. Vol. 1, Conifers. Agriculture Hand- book 654. USDA, Forest Service, Washington, DC. ViLLERS-Ruiz, L. AND I. Trejo-Vazquez. 1997. Assess- ment of the vulnerability of forest ecosystems to cli- mate change in Mexico. Climate Research 9:87-93. Waring, R. H., W. H. Emmingham, and S. W. Running. 1975. Environmental limits of an endemic spruce, Picea breweriana. Canadian Journal of Botany 53: 1599-1613. Wright, J. W. 1955. Species crossability in spruce in re- lation to distribution and taxonomy. Forest Science 1 : 319-349. Madrono, Vol. 47, No. 2, pp. 89-96, 2000 REVISION OF CORETHROGYNE (COMPOSITAE: ASTEREAE) J. Phillip Saroyan 1838 Indiana, Vallejo, California 94590 Dennis R. Parnell Department of Biology, Santa Clara University, Santa Clara, California 95053-0001 John L. Strother University Herbarium, 1001 Valley Life Sciences Building #2465, University of California, Berkeley, California 94720-2465 Abstract Review of variation in morphological characters w^ithin and among natural populations and comparisons of greenhouse-grown plants, coupled with high pollen stainabilities in progeny from crosses of plants from different populations, led to the conclusion that Corethrogyne should be treated taxonomically as comprising a single species with two varieties. The revised taxonomy results in the new combination Corethrogyne filaginifolia var. californica. Corethrogynes occur, as scattered, local popula- tions, from Coos Co., Oregon, south through much of California into northern Baja California, from sea level along the immediate coast to ca. 2500 m in the Sierra Nevada. Keck (1959) recognized three species of Corethrogyne DC; Ferris (1960) recog- nized seven species. Lane (1992, 1993) treated the type species of Corethrogyne and Lessingia Cham- isso as congeneric and, based in part on the work summarized here, treated the corethrogynes of Keck and Ferris as two varieties in one species of Lessingia. Morphological variation within and among pop- ulations of these plants is considerable and is re- flected in the 33 basionyms that have been linked to the generic name Corethrogyne. The 30 'taxa' referable here to Corethrogyne have been distin- guished mainly by traditional morphological char- acteristics such as habit, size and shape of leaves, number of heads per flowering stem, size of heads, and shape of involucres. Aspects of indument such as relative amounts of tomentum and/or stipitate- glandular hairs have also been used in drawing cir- cumscriptions. Field observations by Saroyan indicated that plants with characteristics of reputedly allopatric 'taxa' sometimes grow together in local popula- tions. Plants with heads (including rays) 15 vs. 30 mm wide (assignable to two distinct 'species'), plants with leaves 5 vs. 19 mm wide (different 'spe- cies'), or plants with involucres 7 vs. 12 mm long (different 'species' or different 'varieties') are found in local populations, represent extremes of continua, and exemplify the kinds of problems that are found in past taxonomies. Such problems of reconciling taxonomy and plants did not go unnot- iced by Keck or Ferris. "This imperfectly known sp. [C. leucophylla] recombines the characters of the other two [C. californica and C. filaginifolia sensu Keck] and needs further study." — Keck (1959). "Relatively few individual collections in any given range completely conform to the type and the original description. Intermediate forms abound . . .." — Ferris (1960). On transfer of Corethrogyne to Lessingia, Lane (1992) emphasized morphological and chemical (chloroplast DNA) similarities. We emphasize the differences, as did Jones (1977), who compared corethrogynes and lessingias and concluded, "It would serve no useful purpose to argue for a con- generic status." The genera are readily distinguished: Plants perennial; heads radiate Corethrogyne Plants annual; heads discoid or ± radiant (corollas of peripheral florets often strongly zygomorphic) Lessingia Materials and Methods This paper is based on an unpublished thesis (Sa- royan 1974). In addition to traditional field observations made up and down California and traditional review of specimens in herbaria (ca. 1500 sheets from 17 her- baria - see Acknowledgments), Saroyan (1974) studied morphological variation in samples from 20 populations of Corethrogyne from northern to southern California (A-T in Appendix I). Each population was associated with one of three vege- tation types: type I, grasslands; type II, coastal scrub and chaparral; and type III, forest. For each of the study populations, one to five plants grown from seed in greenhouses in Hayward, CA were 90 MADRONO [Vol. 47 Fig. 1. Characters of a generalized Corethrogyne. used in tests of self-compatibility and in inter-pop- ulation crosses and served as tests for phenotypic plasticity in the study populations. Saroyan (1974) chose characters that have been used in taxonomy of Compositae in general and/or in Corethrogyne in particular and that could be eas- ily assessed statistically (cf. Fig. 1). He determined mean, standard error of the mean, and range of val- ues for: 1) length of stem, 2) length of leaf, 3) width of leaf, 4) length of floral stem, 5) number of heads per flowering stem, 6) width of head, 7) number of ray florets per head, 8) length of involucre, 9) di- ameter of involucre, 10) number of series of phyl- laries, 11) number of serrations per leaf, 12) depth of leaf serrations, 13) length of cypsela, 14) width of cypsela, 15) length of pappus, 16) ratio length/ width of leaf, and 17) ratio length/diameter of in- volucre. To determine minimum adequate sample size for representative statistics for morphological charac- ters in wild populations, Saroyan selected samples of 25, 50, and 100 plants from one such population; data from those three samples were pooled to give a fourth sample of 175 plants. Comparisons of F- test statistics for the four samples indicated that re- liable results could be achieved from a sample size of 50 individuals. Fifty plants were randomly se- lected in the field from each of the 20 study pop- ulations by use of a grid and a table of random numbers, yielding a grand total of 1000 plants sam- pled. Sampling of individual structures was also ran- domized. For some characters, all like structures (e.g., heads at mid anthesis) were removed from a plant and tossed into a paper bag, then one was extracted blindly for measurement. For other char- acters, the structure to be measured was chosen by association with a table of random numbers. Mean, 2000] standard error of the mean, and range were deter- mined for each set of 50 measurements. Chromosome counts were made from microspo- rocytes fixed in acetic ethanol (3:1, v:v) and stained in aceto-carmine for 2-13 plants from each study population. Pollen was stained in lactophenol-cot- ton blue for 1-13 plants (300 grains counted for each plant) from each study population. Vouchers for each study population and repre- sentative greenhouse-grown plants (including prog- eny from crosses made in the greenhouse) were de- posited in UC. Results and Discussion Character by character comparisons across the 20 study populations confirmed our impression of gen- eral morphological continuity within Corethrogyne. Not only ranges of absolute values but even mea- sures of two standard errors of the mean for any one population overlapped with the same values or measures for some or all of the other 19 popula- tions (e.g., length of stem and width of head; cf. Fig. 2). Nevertheless, the five northern populations (A-E in Appendix 1) showed coherence and were somewhat distinct from the southern ones (F-T in Appendix 1) for some characters (e.g., number of heads per flowering stem and the ratio length/di- ameter of involucre; cf. Fig. 3). Variation in single characters seldom showed clear correlations with environmental or habitat pa- rameters. Plants from coastal populations usually had shorter, more prostrate stems; plants from in- land populations were usually more erect. The four populations with longest and narrowest leaves were the four associated with chaparral or coastal scrub vegetation; they were not distinctive for other char- acters. Morphologies of greenhouse-grown progeny from each of the study populations were similar to those of parental plants. Evidently, phenotypic plas- ticity is not strong for any of the characters exam- ined. All published reports of chromosome number (see standard indices; e.g., Goldblatt and Johnson 1992) for Corethrogyne have given 2n — \0 (some as « = 5; some as 2« = 5 II). Similarly, we found meiosis to be regular with 2^2 = 5 II in all of our samples, including the interpopulational hybrids. Pollen stainabilities in our study populations ranged from 80 to 97 percent (Appendix 1). One or more plants grown from seed collected from each study population were self-pollinated; none set viable seeds. Inter-population crosses of one to five pairs of plants in the combinations E X H, N X H, J X D, and M X C (cf. Appendix 1) yielded hybrids with no obvious irregularities at meiosis and with pollen stainabilities of 82 to 96 percent. In general aspect, some local populations of Cor- ethrogyne are strikingly different from others. Nev- ertheless, overlap and continuity in expression of 91 most morphological characters, coupled with cross- compatibility between plants from quite disjunct and dissimilar populations, have led us to treat Cor- ethrogyne as comprising a single species. Plants from the northwestern part of the range of the spe- cies usually have fewer heads on each flowering stem and have larger heads than do plants from other areas; we have recognized them as a variety. Taxonomy Corethrogyne DC, Prodr. 5:215. 1836. — Type: Corethrogyne californica DC. Herbaceous or suffrutescent perennials, primary stems decumbent to ascending or erect, mostly 1- 10 dm long, usually densely white-tomentose, sometimes becoming glabrate and/or glandular dis- tally. Leaves alternate, often crowded at bases of stems, sessile or with bases of blades ± decurrent on petioles, blades ovate to spatulate, oblanceolate, or linear, 1-7+ cm long, 3-19+ mm wide, becom- ing smaller, sessile, and bractlike distally, margins entire or variously toothed. Heads pedunculate or sessile, 1-20+ per floral stem. Involucres hemi- spheric to campanulate, turbinate, or cylindric, 6- 14 mm long, 3-10 mm in diam. Phyllaries 30-90+, strongly graduated in 3-9 series, narrowly lanceo- late to linear, cartilaginous to scarious with herba- ceous, often spreading to squarrose tips, becoming deflexed as cypselae are shed. Ray florets 10-43 in 1 series, neutral, corollas purplish through violet and pink to white, laminae ± linear. Disc florets 12-120+, bisexual, corollas yellow, actinomorphic, 4-8 mm long, tubes 0.6—1.4 mm long, glabrous, throats very narrowly cylindric, 2.8-5.5 mm long, often sparsely puberulent, lobes equal, narrowly lanceolate, 0.7-1.2 mm long, sparsely to densely glandular-puberulent abaxially, papillate-ciliolate on margins and/or adaxially; style branches linear with blunt to subulate appendages, ± hispid with rigid yellow hairs, the appendages to half as long as the stigmatic lines. Cypselae cuneiform to linear, mostly 2-5 mm long, 5-7-ribbed, puberulent to pi- lose. Pappi of 35-65 coarse, unequal, brownish to reddish bristles 3-8 mm long. Chromosomes: 2n = 10. As treated here the genus is monotypic. Corethrogyne eilaginifolia (Hook. & Am.) Nutt., Trans. Amer. Phil. Soc, ser. 2. 7:290. 1840 [1841]. = Aster? filaginifolius Hook. & Arn., Bot. Beechey voy. 146. 1833. = Corethrogyne californica DC. [var.] filaginifolia (Hook. & Am.) Kuntze, Rev. gen. pi. 1:330. 1891 [illegit., oldest sp. epithet not used]. = Corethrogyne fi- laginifolia (Hook. & Am.) Nutt. var. typica M. L. Canby, Bull. S. Calif. Acad. Sci. 26:10. 1927. = Lessingia filaginifolia (Hook. & Am.) M. A. Lane, Novon 2:213. 1992. —Type: Califomia, F. W. Beechey et al. s.n. (holotype: E!). SAROYAN ET AL.: CORETHROGYNE 92 MADRONO [Vol. 47 90 ^ 80 e' w 70 E • 60 «D ^ 50 o f 40 • 30 -i 20 10 ABCDEFGH I J KLMNOPQRST Population 40 • 30 20 10 ABCDEFGH I J KLMNOPQRST Population Fig. 2. Dice diagrams for variation (mean, ± two standard errors of the mean, and absolute range) in stem length and in width of head (cf. Fig. 1) in 20 populations (cf. appendix 1) of Corethrogyne (A-E = C. filaginifolia van californica; F-T ^ C. f. var. filaginifolia). As indicated in discussion above, we recognize two ± allopatric varieties within C. filaginifolia. Key to varieties of Corethrogyne filaginifolia 1. Floral stems usually branched, each with 3-6(1- 20 + ) heads; lengths of involucres (fresh, at anthe- sis) 6-14 mm, mostly twice the diameters . . . C. filaginifolia var. filaginifolia r Floral stems usually unbranched, each with l(-5) heads; lengths of involucres (fresh, at anthesis) 5- 10 mm, about equal to diameters C. filaginifolia var. californica 1. Corethrogyne filaginifolia (Hook. & Am.) Nutt. var. filaginifolia. Asterl tomentellus Hook. & Am., Bot. Beechey voy. 146. 1833. = Corethrogyne tomentella (Hook. & Am.) Torrey & A. Gray, Fl. N. Amer. 2:99. 1841. = Corethrogyne californica DC. [var.] tomentella (Hook. & Am.) Kuntze, Rev. 2000] SAROYAN ET AL.: CORETHROGYNE 93 20 15 10 *• 5 ll ABCDEFGH I J KLMNOPQRST Popalation ABCDEFGH 1 J KLMNOPQRST Population Fig. 3. Dice diagrams for variation (mean, ± two standard errors of the mean, and absolute range) in numbers of heads per flowering stem and in ratio of length/diameter of involucre in 20 populations (cf. Appendix 1 ) of Coreth- rogyne (A-E = C. filaginifolia var. californica; F-T = C. f. var. filaginifolia). gen. pi. 1:330. 1891 [illegit., oldest sp. epithet not used]. — Type: California, F. W. Beechey et al. s.n. (holotype: K!). Diplopappus incanus LindL, Edward's Bot. Reg. 20:1693. 1835. = Corethrogyne incana (Lindl.) Nutt., Trans. Amer. Phil. Soc, ser 2. 7:290. 1840 [1841], = Corethrogyne californica DC. [var.] incana (Lindl.) Kuntze, Rev. gen. pi. 1:330. 1891 [illegit., oldest sp. epithet not used]. = Corethro- gyne filaginifolia (Hoolc. & Am) Nutt. var. in- cana (Nutt.) IVl. L. Canby, Bull. S. CaHf. Acad. Sci. 26:14. 1927. —Type: California, D. Douglas s.n. (no specimen located; the plate confirms ap- plication of the name). Aplopappusl haenkei DC, Prodr. 5:349. 1836. — Type: "inter plantas Regiomontanas herb. Haen- keani ab ill. de Sternberg ad studium missi con- servatur"; specimens came from near IVIonterey, California, fide A. Gray (1876). Holotype may be in P; specimen in G-DC (microfiche!) con- firms application of the name. Diplopappus leucophyllus Lindl. ex DC, Prodr. 5: 94 MADRONO [Vol. 47 278. 1836. = Corethrogyne californica DC. [van] leucophylla (Lindl. ex DC.) Kuntze, Rev. gen. pi. 1:330. 1891 [illegit., oldest sp. epithet not used]. = Corethrogyne leucophylla (Lindl. ex DC.) Jeps. [attributed to Menzies], Fl. w. Calif. 564. 1901. —Type: California, Monterey Co., near Monterey, A. Menzies s.n. (holotype: K!). Corethrogyne virgata Benth., Bot. voy. Sulphur 23. 1844. = Corethrogyne filaginifolia (Hook. & Arn.) Nutt. var. virgata (Benth.) A. Gray in S. Watson, Bot. Calif. 1:321. 1876. = Corethrogyne californica DC. [var.] virgata (Benth.) Kuntze, Rev. gen. pi. 1:330. 1891 [illegit., oldest sp. ep- ithet not used]. — Type: California, San Pedro, R. Hinds s.n. (holotype: K!). Corethrogyne incana (Lindl.) Nutt. var.? rigida Benth., PI. hartweg. 316. 1849. = Corethrogyne filaginifolia (Hook. & Am.) Nutt. var. rigida (Benth.) A. Gray, Synop. fl. N. Amer. 1(2): 170. 1884. = Corethrogyne californica DC. [var.] rig- ida (Benth.) Kuntze, Rev. gen. pi. 1:330. 1891 [illegit., oldest sp. epithet not used]. = Corethro- gyne rigida (Benth.) A. Heller, Muhlenbergia 2: 256. 1906. — Type: California, Monterey Co., "In collibus arenosis juxta Monterey," K. T. Hartweg ''1771(130)'' (holotype: K!). Corethrogyne filaginifolia (Hook. & Arn.) Nutt. var. robusta Greene, Pittonia 1:89. 1887. — Lec- totype (here designated): California, Santa Bar- bara Co., San Miguel Island, Sep 1886, E. L. Greene s.n. (lectotype: CAS!). Corethrogyne viscidula Greene, Fl. francisc. 378. 1897. = Corethrogyne filaginifolia (Hook. & Am.) Nutt. var. viscidula (Greene) D. D. Keck, Aliso 4:105. 1958. — Type: Califomia, Monterey Co., "Monterey," 22 Jul 1888, C. C. Parry s.n. (holotype: ND-G!). Corethrogyne viscidula Greene var. greenei Jeps., Fl. w. Calif. 564. 1901. —Lectotype (here des- ignated): Califomia, Alameda Co., "Niles," 25 Jun 1896, W. L. Jepson ''14614'' (lectotype: JEPS!). Corethrogyne virgata Benth. var. bernardina Abrams, Fl. Los Angeles 401. 1904. = Corethro- gyne filaginifolia (Hook. & Am.) Nutt. var. ber- nardina (Abrams) H. M. Hall, Univ. Calif. Publ. Bot. 3:71. 1907. —Type: Califomia, San Ber- nardino Co., Mentone, 10 Aug 1903, L. Abrams 2931 (holotype: DS!). Corethrogyne filaginifolia (Hook. & Am.) Nutt. var. glomerata H. M. Hall, Univ. Calif. Publ. Bot. 3:72. 1907. —Type: California, San Ber- nardino Co., "Oak Glen, Yucaipe Ranch, near Redlands," Nov 1903, G. Robertson 117 (holo- type: UC!). Corethrogyne filaginifolia (Hook. & Am.) Nutt. var. latifolia H. M. Hall, Univ. Calif. Publ. Bot. 3:70. 1907. —Type: Califomia, Ventura Co., Ox- nard, 1901, 7. B. Davy 7815 (holotype: UC!). Corethrogyne filaginifolia (Hook. & Am.) Nutt. var. linifolia H. M. HaU, Univ. Calif. Publ. Bot. 3:71. 1907. = Corethrogyne linifolia (H. M. Hall) Ferris, Contr. Dudley Herb. 5:100. 1958. — Type: Califomia, San Diego Co., ca. 1 km south of Del Mar, 5 Aug 1906, K. Brandegee s.n. (holotype: UC!). Corethrogyne filaginifolia (Hook. & Am.) Nutt. var. pacifica H. M. Hall, Univ. Calif. Publ. Bot. 3:73. 1907. —Type: Califomia, San Diego Co., "Pacific Beach, near San Diego," May-Oct 1899, C. A. Purpus s.n. (holotype: UC!; isotype: US!). Corethrogyne brevicula Greene, Leafl. Bot. Ob- serv. Crit. 2:26. 1910. = Corethrogyne filagini- folia (Hook. & Arn.) Nutt. var. brevicula (Greene) M. L. Canby, Bull. S. Calif. Acad. Sci. 26:12. 1927. —Type: Califomia, "Mountains of San Diego Co.," Oct 1889 (label) or 1899 (pro- tologue), C. R. Orcutt s.n. (holotype: US!). Corethrogyne flagellaris Greene, Leafl. Bot. Ob- serv. Crit. 2:26. 1910. —Type: Califomia, Los Angeles Co., "Along the seaboard at Redondo," 25 May 1902, E. Braunton 280 (holotype: US!; isotype: DS!). Corethrogyne fioccosa Greene, Leafl. Bot. Observ. Crit. 2:25. 1910. —Type: Califomia, Santa Bar- bara Co., "Elwood, near Santa Barbara," Sep 1908, A. Eastwood s.n. (holotype: US!). Corethrogyne lavandulacea Greene, Leafl. Bot. Observ. Crit. 2:27. 1910. —Type: Califomia, Santa Catalina Island, Sep 1898, B. Trask s.n. (holotype: US!). Corethrogyne racemosa Greene, Leafl. Bot. Ob- serv. Crit. 2:26. 1910. —Type: California, "Mountains of San Diego Co.," Oct 1889, C. R. Orcutt s.n. (holotype: US!). Corethrogyne scabra Greene, Leafl. Bot. Observ. Crit. 2:25. 1910. —Type: Califomia, Los Ange- les Co., Sep 1890, H. E. Hasse s.n. (holotype: US!). Corethrogyne sessilis Greene, Leafl. Bot. Observ. Crit. 2:25. 1910. = Corethrogyne filaginifolia (Hook. & Am.) Nutt. var. sessilis (Greene) M. L. Canby, Bull. S. Calif. Acad. Sci. 26:15. 1927. — Type: Califomia, "San Bemardino Mountains," 23 Oct 1891, 5. 5. Parish 2233 (holotype: US!; isotype: UC!). Corethrogyne filaginifolia (Hook. & Am.) Nutt. var. pinetorum I. M. Johnst., Bull. S. Calif. Acad. Sci. 18:21. 1919. —Type: Califomia, Los An- geles Co., San Antonio Mts., Brown's Flat, 1 Sep 1918, /. M. Johnston 2137 (holotype: POM!; iso- type: DS!). Corethrogyne filaginifolia (Hook. & Am.) Nutt. var. peirsonii M. L. Canby, Bull. S. Calif. Acad. Sci. 26:14. 1927. —Type: Califomia, Los An- geles Co., Newhall, 7 Oct 1923, F. W. Peirson 4159 (holotype: POM!; isotype: DS!). Corethrogyne filaginifolia (Hook. & Am.) Nutt. var. hamiltonensis D. D. Keck, Aliso 4:105. 1958. —Type: Califomia, Santa Clara Co., Mt. 2000] SAROYAN ET AL.: CORETHROGYNE 95 , Hamilton, Aug 1914, H. M. Hall 9865 (holotype: I NY!). Primary stems predominately erect, mostly 1-10 dm long, usually branched distally. Leaf blades nar- rowly spatulate to linear, entire or toothed. Heads (l-)5-20+ per floral stem. Involucres cylindric to turbinate, lengths (6-14 mm) mostly twice diame- ters (3-8 mm) at anthesis in living plants. Phyllaries mostly 4-9-seriate. Disc florets mostly 12-40, co- rollas mostly 4-6 mm long. Widespread through much of cismontane Cali- fornia from Placer Co. south through Sierra Nevada and in interior mountains of Contra Costa and Al- ameda COS. south through coast ranges to transverse ranges and Channel Islands and into northern Baja California. 2. Corethrogyne filaginifolia (Hook. & Am.) Nutt. var. californica (DC.) Saroyan, comb. nov. = Corethrogyne californica DC, Prodr. 5:215. 1836. = Lessingia filaginifolia (Hook. & Am.) M. A. Lane var. californica (DC.) M. A. Lane, Novon 2:213. 1992. — Type: "Nova Califomia," 1833, D. Douglas s.n. (holotype: G-DC, micro- fiche!; isotypes: BM!, K!). Corethrogyne obovata Benth., Bot. voy. Sulphur 22. 1844. = Corethrogyne californica DC. [van] obovata (Benth.) Kuntze, Rev. gen. pi. 1:330. 1891 [illegit., oldest sp. epithet not used]. — Type: Califomia, "Bodegas," 1841, R. Hinds s.n. (holotype: K!). Corethrogyne spathulata A. Gray, Proc. Amer. Acad. Arts 7:351. 1868. = Corethrogyne cali- fornica DC. [var.] spathulata (A. Gray) Kuntze, Rev. gen. pi. 1:330. 1891 [illegit., oldest sp. ep- ithet not used]. — Type: California, Humboldt Co., Shelter Cove, 1867, H N. Bolander 6505 (holotype: GH!; isotypes: BM!, K!, US!). Corethrogyne caespitosa Greene, Fl. francisc. 378. 1897. — Type: Califomia, San IVIateo Co., Crys- tal Springs, 22 Jun 1886, E. L. Greene s.n. (ho- lotype: ND-G!; isotypes: PH!, US!). Corethrogyne californica DC. var. lyonii S. F. Blake, J. Wash. Acad. Sci. 33:267. 1943. — Type: Califomia, IMerced Co., Cathedral Peak, 4 Jun 1941, G. S. Lyon 1572 (holotype: US!, for- merly in NA; isotypes: DS!, UC!). Primary stems predominately decumbent to as- cending, mostly less than 6 dm long, mostly un- branched. Leaf blades mostly obovate to spatulate, toothed. Heads mostly l(-5) per floral stem. Invo- lucres hemispheric to campanulate, lengths (5-10 mm) mostly equalling diameters at anthesis in liv- ing plants. Phyllaries mostly 3-5-seriate. Disc flo- rets 30-120+, corollas mostly 6-8 mm long. Common as discrete, very local, usually dense populations from northem Monterey, San Benito, and westem Merced cos. north through the North Coast Ranges of Califomia into western Klamath Range in Coos Co., Oregon. Questionable Name and Excluded Taxa Corethrogyne californica DC. [var.] recurva Kun- tze, Rev. gen. pi. 1:330. 1891 [illegit., oldest sp. epithet not used]. — Type: We have not seen type material; the description, "Involucri bracteae ap- ice recurvae," is insufficient to allow confident application of the name. Corethrogyne cana (A. Gray) Greene, Bull. Calif. Acad. Sci. l(no. 4): 223. 1885. = Diplostephium canum A. Gray, Proc. Amer. Acad. Arts 11:75. 1876. = Hazardia cana (A. Gray) Greene, Pit- tonia 1:29. 1887. = Haplopappus canus (A. Gray) S. F Blake, Contr. U.S. Natl. Herb. 24:86. 1922. — Type: Mexico, Baja Califomia, Guada- lupe Island, 28 Mar 1875, E. Palmer s.n. (holo- type: GH). Corethrogyne detonsa Greene, Bull. Torrey Bot. Club 10:41. 1883. = Hazardia detonsa (Greene) Greene, Pittonia 1:29. 1887. = Haplo- pappus detonsus (Greene) P. H. Raven, Aliso 5: 343. 1963. — Type: Origin and collector un- known (holotype: CAS!). Acknowledgments We thank B. Baldwin, T. Duncan, M. Lane, S. Markos, G. Nesom, R. Ornduff, R. Pimentel, R. Price, J. Semple, A. Smith, and S. Sundberg for sharing information and/or for helpful comments on earlier versions of this paper. We thank staff at BM, CAS, DS, E, G, GH, HAY, JEPS, K, ND, ND-G, NY, PH, POM, RSA, UC, and US for loans and/or for making specimens available for study. JPS ex- tends special thanks to Mary Lou Wilcox for guidance, support, and encouragement throughout the course of his thesis research. The original for Figure 1 was drawn by Randall E. Tribo. Literature Cited Ferris, R. S. 1960. Corethrogyne. Pp. 337-342 in R. S. Ferris, Illustrated flora of the Pacific States, vol. 4. Stanford University Press, Stanford. GoLDBLATT, P. AND D. E. JoHNSON. 1992. Index to chro- mosome numbers in plants. Monographs in System- atic Botany from the Missouri Botanical Garden. 40: vi-viii, 1-238. Gray, A. 1876. Corethrogyne. Pp. 320-321 m S. Watson et al.. Geological survey of California . . . Botany . . .. [Bot. Calif.], vol. 1. Welch, Bigelow, & Co., Univer- sity Press, Cambridge, Mass. Jones, D. T. 1977. Relationships of Lessingia Cham, and Corethrogyne DC. (Asteraceae). M.A. thesis, Califor- nia State University, Hayward. Keck, D. D. 1959. Corethrogyne. Pp. 1204-1207 in P. A. Munz, A California flora. University California Press, Berkeley. Lane, M. 1992. New combinations in Californian Lessin- gia (Compositae: Astereae). Novon 2:213. . 1993. Lessingia. Pp. 304-306 in J. D. Hickman (ed.). The Jepson manual. University California Press, Berkeley. Saroyan, J. P. 1974. Variation in the genus Corethrogyne DC. (Astereae - Asteraceae). M.A. thesis, California State University, Hayward. 96 MADRONO [Vol. 47 Appendix 1. Study populations of Corethrogyne. Order of entries is north to south. Each entry follows the form: alphabetic identifier; latitude; longitude; elevation (m); distance from ocean (km); vegetation (grasslands, coastal scrub, chaparral, forest); geographic location (all in California); and average pollen stainability (number of plants sampled for pollen stainability). Chromosome counts for one or more plants from each of the populations all yielded 2n = 5 II. Voucher collections are in UC. A. 41°47'; 124°08'; 75 m; 0.4 km; grasslands; Humboldt Co., 2.5 miles north of Patricks Point; 90%(4). B. 38°57'; 123°43'; 75 m; 0.04 km; grasslands; Mendocino Co., 500 yards east of Point Arena; 88%(7). C. 38°08'; 122°53'; 75 m; 4.8 km; grasslands; Marin Co., along Pierce Point Road, 1 mile south of Tomales Bay State Park; 87%(5). D. 38°05'; 122°45'; 200 m; 9.5 km; grasslands; Marin Co., on Inverness Ridge, 3 miles west of Inverness; 93%(7). E. 37°30'; 122°20'; 100 m; 1 1.25 km; grasslands; San Mateo Co., east bank of Upper Crystal Springs Reservoir 0.25 mile south of highway 92; 97%(11). F. 36°42'; 121°48'; 25 m; 0.4 km; grasslands; Monterey Co., 0.25 miles east of ocean, 1 mile south of animal shelter on dunes of Marina Beach; 93%(7). G. 36°38'; 12r46'; 125 m; 4.8 km; chaparral; Monterey Co., Fort Ord, break area of M-79 grenade range 83%(1). H. 36°37'; 12r56'; 10 m; 0.02 km; grasslands; Monterey Co., 1 mile north of Hmit of Asilomar Beach; 93%(13). I. 36°35'; 12r58'; 10 m; 0.03 km; grasslands; Monterey Co., opposite Seal Rock on 17-Mile Drive; 91%(2). J. 35°34'; 121°59'; 25 m; 0.03 km; forest; Monterey Co., opposite Cypress Point on 17-Mile Drive; 94%(9). K. 36°30'; 12r55'; 10 m; 0.02 km; grasslands; Monterey Co., vacant lot at Yankee Point, 1 mile south of Point Lobos; 80%(4). L. 36°23'; 12r30'; 40 m; 35.5 km; chaparral; Monterey Co., Hastings Reservation, along trail between bunk houses; 84%(5). M. 36°17'; 121°5r; 25 m; 0.8 km; grasslands; Monterey Co., along highway 1, 10 miles south of Point Lobos; 93%(5). N. 36°16'; 12r50'; 250 m; 0.8 km; coastal scrub; Monterey Co., coastal bluff, 0.25 mile west of highway 1, 15 miles south of Carmel; 92%(12). O. 35°52'; 121°27'; 300 m; 0.5 km; grasslands; Monterey Co., 1.2 miles north of Gorda; 94%(2). P. 35°39'; 12ri4'; 10 m; 0.02 km; grasslands; San Luis Obispo Co., 3 miles north of San Simeon; 91%(3). Q. 34°56'; 120°38'; 150 m; 3 km; grasslands; Santa Barbara Co., 3 miles east of Point Sal; 85%(1). R. 34°35'; 120°25'; 100 m; 14.5 km; grasslands; Santa Barbara Co., 1 mile south of Lompoc, 0.5 mile west of highway 1; 98%(2). S. 34°13'; 117°12'; 2000 m; 100 km; forest; San Bernardino Co., 3 miles south of Lake Arrowhead, 0.5 miles west of highway 18; 86%(4). T. 33°00'; 117°15'; 200 m; 0.8 km; coastal scrub; San Diego Co., coastal bluff near Torrey Pines State Reserve; 90%(3). sMadrono, Vol. 47, No. 2, pp. 97-105, 2000 CORRESPONDENCE BETWEEN NI TOLERANCE AND HYPERACCUMULATION IN STREPTANTHUS (BRASSICACEAE) Robert S. Boyd, Michael A. Wall, and James E. Watkins, Jr. Department of Biological Sciences and Alabama Agricultural Experiment Station, Auburn University, AL 36849-5407 Abstract Nickel hyperaccumulation may be associated with increased Ni tolerance for some plant species that grow on serpentine soils. We contrasted the Ni tolerance of three species: a Ni hyperaccumulator {Strep- tanthus polygaloides A. Gray) endemic to serpentine soil, a congeneric non-hyperaccumulator also en- demic to serpentine soil {S. breweri A. Gray), and a species from the same family but not adapted to serpentine soil {Brassica oleracea L.). We assessed Ni tolerance by measuring germination and radicle elongation in test solutions varying in Ni^- content. By both approaches, Ni tolerance was greatest for the hyperaccumulator, intermediate for the non-hyperaccumulator, and least for the unadapted species. A soil-based test of root elongation, using S. polygaloides and B. oleracea with two serpentine soils and one non-serpentine soil, showed a significant species-by-soil interaction. Root elongation of B. oleracea was inhibited in serpentine soil, whereas S. polygaloides showed reduced root elongation in non-serpentine soil. We concluded that these results are consistent with the hypothesis that Ni hyperaccumulation is a metal tolerance mechanism adopted by some species native to serpentine soils. These results also are consistent with other ecological functions of Ni hyperaccumulation, such as the elemental allelopathy or microsite tolerance hypotheses. Plant tissues vary widely in heavy metal concen- trations, although most plant species contain very low levels. Pais and Jones (1997) reported that spe- cies not adapted to high-Ni soils typically contain 0.3-3.5 |jLg Ni/g dry wt. For these species, tissue Ni in the range 8-50 |xg/g dry wt usually denotes a toxic Ni concentration (MacNicol and Beckett 1985). Serpentine soils often contain elevated levels of Ni (Kruckeberg 1984; Brooks 1987). Many plants native to these soils also contain elevated levels of Ni (Reeves 1992), often ranging from 10-100 |jLg Ni/g (Brooks 1987). Baker and Walker (1990) called these species "accumulators". A small pro- portion of plant species native to serpentine soils accumulate Ni to an extraordinary degree over a wide range of soil Ni concentrations (Morrison et al. 1980). Brooks et al. (1977) termed these plants "hyperaccumulators", defining them as containing at least 1000 jxg Ni/g. Many workers have suggested that metal hyper- accumulation has an adaptive function. Metal hy- peraccumulators belong to a number of evolution- ary lines of dicotyledonous plants (Brooks 1987), suggesting multiple independent evolution of this trait and therefore that metal hyperaccumulation has positive selective value. Boyd and Martens (1992) suggested four functions of metal hyperac- cumulation: metal tolerance, drought tolerance/ avoidance, defense against herbivores/pathogens, and interference with neighboring plants. To date most research has focused on defensive explana- tions (Boyd 1998). This work has shown that hy- peraccumulated metals can defend plants against herbivores (Boyd and Martens 1994; Martens and Boyd 1994; Pollard and Baker 1997; Sagner et al. 1998; Boyd and Moar 1999; Davis 1999; Jhee et al. 1999) and pathogens (Boyd et al. 1994; Ghad- erian et al. 2000). The remaining hypothesized functions of metal hyperaccumulation are relatively unexplored. Metal hyperaccumulation has been suggested to function as a mechanism for tolerating elevated soil metal contents (Boyd and Martens 1992). Metals could be removed from metabolically sensitive ar- eas of a plant's cells or tissues by concentrating them in less sensitive locations (e.g., vacuoles, cell walls, epidermal cells, and trichomes). In some cases (e.g., Ernst 1972; Wild 1978), it has been pro- posed that abscission or loss of high-metal plant parts serves to dispose metals from the plant body. The difficulty with this "tolerance hypothesis" {sensu Boyd and Martens 1992) is that, although metal hyperaccumulators must surely be able to tol- erate very high tissue metal levels, there is no ev- idence that tolerance is an adaptive function of met- al hyperaccumulation. Thus, it is important to com- pare the metal tolerances of hyperaccumulator and non-hyperaccumulator species native to serpentine soils. If both have equivalent metal tolerance abil- ities, then this would be evidence contrary to the tolerance hypothesis. Increased metal tolerance of hyperaccumulator species relative to other species able to grow on serpentine soil is also crucial to the "interference hypothesis" {sensu Boyd and Martens 1992). Some authors (e.g., Gabbrielli et al. 1991; Baker et al. 1992) have suggested that the elevated metal con- tent of leaf litter produced by hyperaccumulators can lead to increased metal content of the surface 98 MAE soil beneath individual plant canopies. Less metal- tolerant species may be prevented from growing in these metal-enriched areas, resulting in lessened competition for the hyperaccumulator. Wilson and Agnew (1992) further suggested that metal hyper- accumulators might suppress the growth of less metal-tolerant species through surface soil metal enrichment and thus create areas dominated by rel- atively pure stands of the hyperaccumulator spe- cies. This interaction was termed "elemental alle- lopathy" by Boyd and Martens (1998), due to the similarity of this phenomenon with allelopathy (Rice 1984). Boyd and Martens (1998) pointed out that experimental confirmation of elemental allelop- athy must demonstrate two facts. First, soil Ni lev- els must be significantly elevated in the vicinity of hyperaccumulator plants, relative to other micro- sites in serpentine habitats. Second, it must be dem- onstrated that hyperaccumulator species are more metal tolerant than potentially competing species. Thus, demonstration of differential metal tolerance between hyperaccumulator and non-hyperaccumu- lator serpentine species is one of the two principles essential to uphold the elemental allelopathy hy- pothesis. As explained above, both the tolerance hypoth- esis and the elemental allelopathy (interference) hy- pothesis require that metal hyperaccumulators be more metal-tolerant than other serpentine soil spe- cies. Some studies of hyperaccumulators (e.g., Kra- mer et al. 1997) have contrasted them with a con- gener that is not native to metalliferous soils and several studies have compared the metal tolerance of hyperaccumulator species with other species. The early work of Morrison et al. (1980) examined the Ni tolerances of seven Ni-hyperaccumulating species and two non-hyperaccumulating species of Alyssum. Root elongation tests showed the hyper- accumulators to be more Ni tolerant than non-hy- peraccumulators. Gabbrielli et al. (1990) used root elongation tests to contrast the Ni tolerance of the hyperaccumulator A. bertolonii Desv. with that of the non-hyperaccumulator serpentine species, Sile- ne italica L., and showed that the hyperaccumulator was much more Ni tolerant than the non-hyperac- cumulator. Homer et al. (1991) contrasted a Ni hy- peraccumulator {Alyssum troodii Boiss) and a non- hyperaccumulator {Alyssum saxatilis = Aurinia saxatilis (L.) Desv.), finding greater Ni tolerance for the hyperaccumulator by examining both bio- mass production and germination experiments. Kra- mer et al. (1996) obtained similar results, using a non-hyperaccumulating Alyssum species {A. mon- tanum) and contrasting its Ni tolerance with the Ni hyperaccumulator A. lesbiacum. Biomass produc- tion of the Ni hyperaccumulator was much greater than the non-hyperaccumulator when plants were grown in a series of Ni-containing solutions. Shen et al. (1997) studied Zn tolerance of two serpentine soil species of Thlaspi by measuring biomass ac- cumulation as a function of Zn concentration in the ONO [Vol. 47 growth medium. They found that the Zn hyperac- cumulator T. caerulescens was more Zn tolerant than the non-hyperaccumulator T. ochroleucum. The research reported here was conducted to compare the Ni tolerance of two annual species of Streptanthus, both of which are endemic to serpen- tine soils, but only one is a Ni hyperaccumulator. This New World genus in the Brassicaceae contains a single Ni hyperaccumulator species, S. polyga- loides A. Gray, and a number of non-hyperaccu- mulating taxa endemic to serpentine soils (Kruck- eberg 1984). Streptanthus polygaloides also is unique among hyperaccumulators due to its being an obligate annual. Other Ni hyperaccumulators, in- cluding those in Alyssum and Thlaspi, are peren- nials and thus the present work extends our under- standing of the relationship between hyperaccu- mulation and tolerance to annual species in another genus. Methods Study species. Streptanthus polygaloides is an annual Ni hyperaccumulator endemic to serpentine soils in the western foothills of the Sierra Nevada in California (Munz and Keck 1968). Studies by Reeves et al. (1981) and Kruckeberg and Reeves (1995) have documented >1000 fxg Ni/g dry wt in all parts of this species. We collected seeds from the Red Hills of Tuolumne County, California (Fa- vre 1987), approximating sample #6737 of Kruck- eberg and Reeves (1995). We selected S. breweri A. Gray to represent a non-hyperaccumulating species of Streptanthus. This species also is an annual and is also endemic to serpentine soils (Kruckeberg 1984). The ranges of these two species do not overlap because S. breweri is native to the Coast Ranges of California (Munz and Keck 1968). Analysis of leaves of this species by Kruckeberg and Reeves (1995) showed a maximum of 13 |jLg Ni/g. Specimens analyzed by Reeves et al. (1981) contained <9 fxg Ni/g. We col- lected seeds from a population in Napa County, California, near sample #6742 of Kruckeberg and Reeves (1995). The third species used was broccoli, Brassica oleracea L. This species is also in the Brassicaceae, but is not adapted to serpentine soil. Seeds were obtained from a commercial source. Germination tests. Seeds were placed in 5-cm di- ameter petri dishes under 24-hr illumination at room temperature. Germination solutions were add- ed in sufficient quantity to completely cover seeds, and were replenished as needed during the experi- ments. Nickel concentrations used for seeds of S. polygaloides and S. breweri were 0, 4.3, 8.5, 13, 17, 26, 34, 43, 51, and 60 mmol/L. The experiment using B. oleracea seeds used fewer Ni+^ concentra- tions, 0, 8.5, 13, 17, 34, and 51 mmol/L, because we assumed the Ni tolerance of this unadapted species would be easier to characterize. Nickel was 2000] BOYD ET AL.: NI TOLERANCE AND HYPERACCUMULATION 99 supplied as NiCU (Fisher Scientific). The use of NiCls presented the possibiHty that inhibition could occur due to increased CI ' ion concentration rather than Ni+2. To test for this effect (as well as osmotic effects) on seed germination, solutions containing 0, 12, 19, 25, 50, 75, 100, 125, 150, 175, 200, 225, and 250 mmol Ca+^/L (as CaCl., Fisher Scientific) were used. Seeds of two species, S. polygaloides and B. oleracea, were used for Ca^^ experiments. Due to insufficient seed availability, these experi- ments were not conducted using S. breweri seeds. Seeds (ranging in number between 10 and 15) of each species were placed in each petri plate and monitored for germination (defined as emergence of the radicle from the seed coat) until germination declined to zero during a several-day period. Ten petri plates (replicates) were used for each combi- nation of species, Ni^-^, and Ca^-^ concentration. Germinating seeds were removed from petri plates to minimize their influence on the chemistry of the germination solutions. Percent germination was cal- culated for the seeds in each petri plate and these data were analyzed by one-way ANOVA for each species and each ion (Ni+^ or Ca+^) used, after transformation (arcsine square root) so data would better fit ANOVA assumptions (Zar 1996). Fisher's Protected Least Significant Difference (PLSD) test (Abacus Concepts 1992) was used for post-hoc means separation (a < 0.05). Root elongation tests in solution. Seeds of all three species were placed on moistened filter paper in petri dishes and allowed to germinate. Once rad- icles reached lengths of 1-3 mm, seedlings were removed and transferred to root elongation test so- lutions. At least three seedlings (up to five if more were available) were used in each petri plate. Three petri plates (replicates) of each species were used for each concentration of Ca+^ or Ni+^. Radicles of transplanted seedlings were allowed to elongate for 7 d at room temperature, after which they were re- moved from test solutions and the length of the primary root measured. All test solutions contained a background level of ions important for normal seedling development (Baker 1987). The background solution contained 0.77 mmol Ca^'/L as Ca(N03)2, 0.82 mmol Mg+V L as MgS04, and 0.28 mmol K+'/L as KNO3 (all obtained from Fisher Scientific). NiCl, was then added to the background solution to create concen- trations of 0, 0.085, 0.17, 0.34, 0.51, 0.68, 0.85, 1.4, and 1.7 mmol Ni+^/L. As in the germination experiment, the addition of CI ' along with Ni+^ presented the possibility that inhibition could occur due to increased CI ' concentration rather than Ni+2. To test for this effect (and general osmotic effects), three Ca^^ solutions were created with CslCU, using the same back- ground solution used for the Ni+- experiments, to create solutions containing 2.5, 5.0, or 10 mmol Ca-2/L. Root lengths from seedlings in each petri plate were averaged, and the data analyzed by one-way ANOVA for each species and solution type (Ni+- or Ca+-) after log-transformation so data would bet- ter fit ANOVA assumptions (Zar 1996). Fisher's PLSD test was used for post-hoc means separation (a < 0.05). Root elongation tests in soil. This experiment de- termined if the tolerance results from Ni^^ solutions had relevance to root performance in serpentine soils. Three soils were used for testing root elon- gation: two from serpentine sites in California and one from a non-serpentine site in Alabama. The first serpentine soil was collected from the Red Hills in Tuolumne County. The second was col- lected from a serpentine site along U.S. Highway 101 near the southern city limit of San Luis Obispo, San Luis Obispo County, California. The non-ser- pentine soil was collected from Auburn, Lee Coun- ty, Alabama. Approximately 1 L of soil was col- lected to 10 cm depth at each location and sieved to remove stones >2 mm dia. A subsample of each soil was used for elemental analysis. Soil samples were double-acid extracted using 20 mL of extractant (0.05 M HCl/0.025 M H2SO4) shaken with 5 g of dry soil for 5 min. The extract was analyzed for Ni using an atomic ab- sorption spectrophotometer (IL 251, Instrumenta- tion Laboratory, Franklin, MA), and for Ca, K, Mg, P, Cu, Fe, Mn, Cr, Pb, Co and Zn using an induc- tively-coupled argon plasma spectrometer (ICAP 9000, Jarrell-Ash, Franklin, MA). Soil from each location was put into small (5 cm diam.) petri plates to approx. 5 mm depth and moistened with deionized water. Seeds of two spe- cies, the unadapted B. oleracea and the hyperac- cumulator S. polygaloides, were used (not enough seeds of S. breweri were available). Seeds were germinated on moist filter paper. Germinating seeds (radicles 1-3 mm long) were transferred to the sur- face of the soil plates and allowed to grow for 4 d. Three seedlings were used for each replicate, and radicle lengths were averaged for each plate. Seed- lings of these two species differed in size, so that direct comparisons of root length were confounded by this factor. We minimized the influence of this innate size difference by relativizing mean root lengths for each species within a replicate. Means were expressed as a decimal fraction of the mean for that treatment which produced the largest mean root length. Mean root lengths for each of the three treatments were each divided by the largest value of those three data, resulting in a value of 1 for the largest mean and lesser values for the other two means. These relativized data were analyzed by two-way ANOVA, using transformed (arcsine square root) data to better meet the assumptions of ANOVA (Zar 1996). Soil collection site and spe- cies were main effect factors and the interaction term was included in the ANOVA model. Fisher's 100 MADRONO [Vol. 47 Protected Least Significant Difference (PLSD) test (Abacus Concepts 1992) was used for post-hoc means separation (a < 0.05). Results Germination tests. Germination of all three spe- cies was significantly affected by Ni+^ concentra- tion. For S. polygaloides, ANOVA indicated a sig- nificant Ni+2 effect (F990 = 21, P < 0.0001). Ger- mination was relatively high (>50%) for even the most concentrated (60 mmol/L) solution (Fig. 1), and Fisher's PLSD test indicated that germination was equivalent to that of the control for all solu- tions containing <17 mmol/L. Streptanthus breweri germination also was significantly affected by Ni+^ (ANOVA: F989 = 22, P < 0.0001). Germination was less than 50% for the higher Ni+^ concentra- tions (>34 mmol/L), and a significant decline in germination relative to the control (Fisher's PLSD test) was first observed at a lesser concentration than for S. polygaloides (13 mmol/L vs. 17 mmol/ L: Fig. 1). Brassica oleracea germination also de- clined significantly as Ni^^ concentration increased (ANOVA: F554 = 38, P < 0.0001). Germination declined to <50% at concentrations >8.5 mmol Ni^-^/L, with Fisher's PLSD test showing that ger- mination first declined significantly relative to the control at 8.5 mmol Ni+-/L. Experiments with Ca^^ solutions also showed in- hibition of germination for all species (data not shown), but at greater concentrations than with Ni^^ solutions. Germination of S. polygaloides was sig- nificantly affected by Ca+- concentration (ANOVA: F12.117 ^ 110, P < 0.0001). Mean germination was >75% for those solutions containing up to 50 mmol Ca+^/L, but declined significantly compared to the control (Fisher's PLSD test), reaching 37% at 75 mmol/L. Germination of B. oleracea also was sig- nificantly affected by Ca+^ concentration (ANOVA: Fi2 n7 = 67, P < 0.0001). Mean germination was >72% for concentrations up to 75 mmol Ca+^/L, and then declined significantly relative to the con- trol (Fisher's PLSD test) to 64% at 100 mmol/L. These results show that CI ' did not produce the decreased germination observed with the Ni^^ so- lutions, and that Ni^^ played a significant role in decreasing seed germination in this experiment. Root elongation tests in solution. Solution Ni+^ concentration significantly affected root elongation for all species tested. For S. polygaloides, ANOVA showed a significant Ni+2 effect (F^^j = 29, P < 0.0001). Mean root length declined significantly relative to the control at 0.085 mmol Ni+^/L (Fish- er's PLSD test. Fig. 2). Maximum inhibition (small- est mean root elongation) was observed at the high- est Ni+2 concentration used (1.7 mmol Ni+^/L). Comparing means to that value, the lowest concen- tration of Ni+^ that resulted in maximum inhibition was 0.85 mmol/L (Fisher's PLSD test). This value begins the maximum inhibition zone noted on Fig- §100 I 80 60 40 20i 0 Q) Streptanthus polygaloides i 1 0 I 0 4.3 8.5 1 31 7 26 34 43 51 60 §100 i g 80 60 40 20 0 Streptanthus breweri h 17 0 4.3 8.51 31 7 26 34 43 51 60 0 8.5 1 3 1 7 34 51 Solution Ni"*"^ content (mmol/L) Fig. 1. Percent germination of the three experimental species as influenced by Ni+- content of the germination test solution. Note that the x-axes are not linear. The hatched bars in each graph indicate treatments for which germination was lowered significantly from that of the control (0 mmol Ni'^'/L) treatment. Error bars denote the upper 95% confidence limit of each mean. ure 2, and represents the lowest Ni+^ level that causes maximum root growth inhibition. Root growth of S. breweri also declined with increasing Ni+2 concentration (ANOVA: Fg.e, = 84, P < 0.0001). Again, means declined significantly rela- tive to the control at 0.085 mmol/L (Fisher's PLSD test. Fig. 2). In this case, however, the maximum inhibition zone began at a lesser Ni concentration than for S. polygaloides. For S. breweri, the maxi- mum inhibition zone started at 0.51 mmol Ni+^/L 2000] BOYD ET AL.: NI TOLERANCE AND HYPERACCUMULATION 101 0 n-T 1 6- Ui QJ - ■*—' o O \_ c (0 4- (1) 0-^ S. polygaloides critical concentration 0 0.17 0.51 0.85 1.7 0.085 0.34 0.68 1.4 2 5 2 0 1 5 1 0 5 0 2 5- S. brewer! T critical C T ;oncentration ■,,,T ,,. 0 0.17 0.51 0.85 1.7 0.085 0.34 0.68 1.4 B. oleracea critical concentration 0 0.17 0.51 0.85 1.7 0.085 0.34 0.68 1.4 Solution Ni"*"^ content (mmol/L) Fig. 2. Mean root lengths (cm) of seedlings of the three experimental species as affected by solution Ni^^ concen- tration. The hatched bars denote treatments that resulted in maximum inhibition of root elongation (all hatched bars are not statistically different from the 1.7 mmol Ni+^/L solution, using Fisher's PLSD test). Error bars denote the upper 95% confidence limit of each mean. (Fisher's PLSD test comparison with mean value at 1.7 mmol/L, Fig. 2). Root growth of B. oleracea also was significantly depressed by Ni^^ solutions (ANOVA: F8 27 = 4.3, P = 0.0021). Mean root elongation again declined significantly at 0.085 mmol Ni+2/L (Fisher's PLSD test. Fig. 2), but the maximum inhibition zone began at the lowest Ni concentration for all three species tested. For B. oleracea, the maximum inhibition zone extended ^ ^ E o 2 0- c \ D 0 - ^— ' o 1 0- o 5- Q) o-" t 4 0" _ o SI D) 3 0" C _ CD O 2 0" O i_ c (0 1 0- 0) S. polygaloides L j_ 2.5 5.0 1 0 S. breweri t 2.5 5.0 1 0 _ 60 1, 50- D>401 f 30- o 2 20- I 10- B. oleracea 0 2.5 + 2 5.0 1 0 Solution Ca content (mmol/L) Fig. 3. Mean root lengths (cm) of seedlings of the three experimental species as affected by solution Ca^-^ concen- tration. Error bars denote the upper 95% confidence limit of each mean. from 0.17-1.7 mmol/L (Fisher's PLSD test com- parison with mean at 1.7 mmol/L, Fig. 2). Effects of the Ni^^ solutions on root elongation did not result from CI ' or overall osmotic concen- trations of the test solutions. For all species tested, root elongation in CaCU solutions was not signifi- cantly depressed relative to control solutions at the highest concentration tested (10 mmol/L, Fig. 3). The CI"' concentration of the 10 mmol Ca+^/L so- lution was much higher than that of the highest (1.7 mmol/L) Ni+^ solution. Thus, a CI ' effect cannot explain the significant declines of root elongation observed for all species tested with the 0.085 mmol Ni+2/L solution. However, a significantly positive 102 MADRONO [Vol. 47 1.0 0 03 Ui c _o CD O O ^ 0.4 -S 0.2 jSerpentine 1 ^Serpentine 2 I I Nonserpentine Brassica oleracea Streptanthus polygaloides Fig. 4. Mean relative root elongation of B. oleracea and S. polygaloides on three soil treatments (two serpentine and one non-serpentine soil). Error bars denote the upper 95% confidence limit of each mean. effect of Ca+^ was detected for two of the three species. Streptanthus breweri was the one species that lacked a significant Ca+^ effect (ANOVA: F3 28 = 2.5, P = 0.0803). Streptanthus polygaloides showed a significant Ca^-^ effect (ANOVA: F328 ~ 8.2, P = 0.0004). Fisher's PLSD test revealed that root elongation for the 2.5 mmol Ca+^/L solution was greater than for any other test solution. Bras- sica oleracea also showed a significant effect (ANOVA: F3 ,2 = 12, P = 0.006). In this case. Fish- er's PLSD test showed that root elongation for the control treatment was significantly lower than for the solutions with elevated Ca+^. In all cases, a sig- nificant decline in root elongation (either due to CI ' or osmotic effects) was not observed with the Ca+2 solutions used. Thus, we conclude that the in- hibitory effects of the Ni+2 solutions were attrib- utable to Ni+2. Root elongation tests in soil. Relativized root elongation values were not significantly affected by either of the two main effect factors (soil and spe- cies) in the ANOVA (for soil: F272 = 0.40, P = 0.67, and for species: F, 72 = 0.86, P = 0.36). How- ever, the interaction term was highly significant (^2,72 = 21, P < 0.0001). Inspection of mean rela- tive root elongation values revealed that the two species reacted in opposite ways to the soils (Fig. 4). Brassica oleracea root elongation was greatest in the non-serpentine soil (mean = 0.78), and less in both of the serpentine soils (Serpentine Soil 1 mean = 0.32; Serpentine Soil 2 mean = 0.55). Fisher's PLSD test showed that the mean for roots from the non-serpentine soil was significantly greater than for roots from Serpentine Soil 1, but the other pairwise comparisons were only margin- ally significant (P = 0.072 for each). On the other Table 1 . Elemental Analysis of the Three Soils Used IN THE Soil Root Elongation Tests. Serpentine Soil 1 was collected from the Red Hills, Tuolumne County, Cal- ifornia, Serpentine Soil 2 was collected from San Luis Obispo County, California, and the Non-serpentine Soil came from Auburn, Lee County, Alabama. Parameter Serpentine Serpentine Non-serpentine (M^g g ') Soil 1 Soil 2 Soil Ca 200 1960 484 K 33.2 15.6 16.8 Mg 493 742 59.0 P 5.10 17.8 60.2 Cu 0.23 0.991 1.41 Mn 27.6 56.1 20.4 Zn 0.92 4.82 53.0 Co 1.72 4.47 0.14 Cr 0.12 1.21 0.085 Pb 0.443 11.4 12.8 Ni 80 128 <4 Ca/Mg ratio 0.41 2.6 8.2 hand, S. polygaloides showed depressed root elon- gation in the non-serpentine soil (mean = 0.29) and higher root elongation in both serpentine soils (Ser- pentine Soil 1 mean = 0.83; Serpentine Soil 2 mean = 0.73). Fisher's PLSD test showed that the mean for roots from non-serpentine soil was significantly less than for both serpentine soils (P < 0.0001 for both comparisons) and that the means for roots from the serpentine soils did not differ from each other (P = 0.31). Elemental analysis of the soils used in the above experiment showed several differences between the serpentine soils and the non- serpentine soil (Table 1). Notable were the elevated Ni levels in the two serpentine soils, along with the lower Ca/Mg ratios for those soils relative to the non- serpentine soil. Also, Serpentine Soil 2 had an unusually high Ca content, giving it a higher Ca/Mg ratio than is usual for serpentine soils (Brooks 1987). Discussion Our experiments showed that the Ni hyperaccu- mulator species, S. polygaloides, was more Ni tol- erant than either the congeneric serpentine soil spe- cies or the unadapted species. This contrast was consistent for both the germination and the root elongation experiments. This finding for Streptan- thus is consistent with earlier work using species of Alyssum and Thlaspi. Both Morrison et al. (1980) and Kramer et al. (1996) reported greater Ni tolerance, as measured by root elongation or bio- mass comparisons, for Ni-hyperaccumulating Alys- sum species relative to non-hyperaccumulators that grew on serpentine soil. Shen et al. (1997) reported greater Ni tolerance in T. caerulescens relative to the non-hyperaccumulating serpentine species T. ochroleucum. Thus, we can extend the generality of the correlation between Ni hyperaccumulation ability and enhanced Ni tolerance to yet another 2000] BOYD ET AL.: NI TOLERANCE AND HYPERACCUMULATION 103 genus within the Brassicaceae, in this case to in- clude an annual hyperaccumulating species. We hope that additional tests of this hypothesis may be undertaken using congeneric taxa from another family. The recent discovery by Reeves et al. (1996; 1999) of a large number of Ni hyperaccu- mulators from Cuba, many within the genera Phyl- lanthus and Leucocroton (Euphorbiaceae) and Bux- us (Buxaceae), provides an excellent opportunity for such research. These genera would also allow extension of these questions to include woody (shrub, tree) growth forms. As pointed out earlier, equivalent metal tolerance between hyperaccumulator and non-hyperaccumu- lator species from serpentine soils would constitute evidence contrary to both the tolerance and ele- mental allelopathy hypotheses. However, our re- sults, plus those reported earlier (Morrison et al. 1980; Gabbrielh et al. 1990; Homer et al. 1991; Kramer et al. 1997; Shen et al. 1997), showed greater metal tolerance by the hyperaccumulator species studied. These results are consistent with both the tolerance and the elemental allelopathy hy- potheses. More information is needed to decide whether either of these hypotheses provides an evolutionary rationale for metal hyperaccumulation. For exam- ple, the elemental allelopathy hypothesis requires that soil Ni levels under hyperaccumulating plants are significantly higher than in other microsites. If this occurs, then the lesser Ni tolerance of co-oc- curring plant species might put them at a compet- itive disadvantage relative to hyperaccumulating species. Our (and other) Ni tolerance tests indicate that elemental allelopathy may indeed provide an adaptive rationale for metal hyperaccumulation. Unfortunately, little information on the microdis- tribution of soil Ni around hyperaccumulator plants is available. Two preliminary reports (Baker et al. 1992; Schlegel et al. 1992) indicated that localized Ni enrichment might occur A third study (Boyd and Jaffre in review) also documented significantly greater surface soil Ni concentrations under cano- pies of the Ni-hyperaccumulating tree, Sebertia ac- uminata Pierre ex Baillon, compared to surface soil under the canopies of nearby non-hyperaccumula- tor tree species. In contrast, a study by Boyd et al. (1999), using the New Caledonian Ni hyperaccu- mulating shrub Psychotria douarrei (Beauvis.) Daniker, provided information contrary to this hy- pothesis. They analyzed soil Ni content under shrubs ranging in size from saplings to full-sized adults. No correlation of soil Ni with shrub size was detected, indicating that Ni enrichment was not oc- curring in that case. We also point out that enhanced tolerance of soil metals by hyperaccumulators may result in another ecological advantage apart from the elemental al- lelopathy hypothesis discussed above. Greater met- al tolerance would allow hyperaccumulators to ex- ploit relatively high-metal soil microsites that might exist on serpentine sites. If other serpentine soil species are unable to grow (or are unable to grow well) in these microsites, hyperaccumulators might avoid competition for soil water/nutrients and thus gain a survival advantage. We call this the "mi- crosite tolerance" hypothesis, to separate it from the elemental allelopathy hypothesis discussed pre- viously. Even for an annual species like S. polygaloides, enhanced Ni tolerance could allow colonization of relatively high-Ni microsites that are not exploited by other species, resulting in a survival advantage. The variability of metal levels within and between serpentine soil sites has been noted before (Kruck- eberg 1984), including recent studies on serpentine soils from California (Nicks and Chambers 1998). Field investigations of the relationship between S. polygaloides density and soil Ni levels would pro- vide evidence pertinent to this hypothesis, but to our knowledge such studies have not been con- ducted. We should note that the microsite tolerance function of hyperaccumulation might intergrade with elemental allelopathy. For example, seedlings of a metal hyperaccumulator that become estab- lished on a microsite containing elevated soil met- als might, over time, further elevate surface soil metal content and thus extend the boundaries of the high-metal microsite. The concentration of metals into surface soil, along with the expanded area of the microsite, could then provide a survival advan- tage via elemental allelopathy. The tolerance hypothesis is more difficult to completely falsify. Hyperaccumulators must, by def- inition, possess the ability to tolerate high tissue metal levels. Therefore, demonstration of greater Ni tolerance by hyperaccumulators does not contradict this hypothesis but does not provide definitive ev- idence that tolerance is an adaptive function of hy- peraccumulation. Our experimental result, demon- strating greater Ni tolerance of S. polygaloides rel- ative to S. breweri, is consistent with the tolerance hypothesis. However, several authors have suggest- ed that tolerance and hyperaccumulation are not strongly linked traits. Ingrouille and Smirnoff (1986) first suggested this for Thlaspi caerulescens, stating that Zn tolerance and Zn hyperaccumulation in that species may be independently inherited. Meerts and Van Isacker (1997) compared Zn hy- peraccumulation and tolerance among populations of Thlaspi caerulescens collected from high- and low-metal soil sites. They found that the low-metal populations were able to hyperaccumulate Zn to a greater extent, but the high-metal populations were more Zn tolerant. We suggest that other approaches can more clear- ly address this question. Perhaps the creation of non-hyperaccumulating mutants of a metal hyper- accumulator species, that can then be used to com- pare metal tolerances, can provide another way to test this hypothesis. Until that point, the tolerance hypothesis must be regarded as a possibly viable 104 MADRONO [Vol. 47 explanation for the adaptive value of metal hyper- accumulation. Our soil-based experiment showed that solution- based root elongation tests may mirror the effects of Ni in soils, but suggested that factors other than Ni content may affect root elongation in serpentine soils. That Brassica root elongation was less in ser- pentine soils was not surprising, as this species does not possess adaptations that allow it to tolerate ser- pentine soil conditions. The result of the soil-based experiment with S. polygaloides (greater elongation in serpentine soils) was unexpected. One explana- tion for this result might be that S. polygaloides requires elevated soil Ni for optimum root growth. A higher metal requirement for hyperaccumulators has been suggested by some experiments (e.g., Shen et al. 1997) but not others (e.g., Morrison et al. 1980, Kramer et al. 1996). Results of our root elongation experiment do not show a Ni^^ require- ment for S. polygaloides, as all species we tested showed a decline in root elongation at the lowest Ni+^ level tested (0.085 mmol/L) relative to the control solution. In contrast, Nicks and Chambers (1995) reported lower biomass for S. polygaloides individuals grown in very low-Ni+^ nutrient solu- tions, and suggested that some level of Ni+- in the growth medium was needed for optimum growth. We have noted mixed results in our own studies with Streptanthus polygaloides. In one experiment (Martens and Boyd 1994), plants grown on high- Ni greenhouse soil had less biomass than plants grown on low-Ni greenhouse soil. A second exper- iment (Boyd et al. 1994) showed the opposite re- sult. A second explanation may involve interactions between a hyperaccumulator and the soil microflo- ra. For S. polygaloides, this could be a positive (in the serpentine soil) or a negative (in the non-ser- pentine soil) interaction that could produce the dif- ference in root growth that we observed. Soil pathogens have been reported to limit plant growth when serpentine soil species are grown in low-Ni soil. For example, Tadros (1957) reported that soil- borne pathogens caused damping-off of seedlings of a serpentine soil non-hyperaccumulator Emmen- anthe species when seedlings were grown on non- serpentine soil. Also, Brooks (1987) reported that Ni hyperaccumulators in the genus Alyssum could be difficult to grow on low-Ni soil due to apparent pathogen attack. Certainly, it seems likely that the performance of serpentine soil species on their na- tive soils will be affected by similar organismal in- teractions, and that these interactions might be af- fected by the elevated metal contents of hyperac- cumulators. The possible consequences of these hy- peraccumulator/soil microflora interactions are only now being articulated (e.g., Boyd and Martens 1998). Acknowledgments We thank M. Davis, R. Dute, D. Folkerts, and R. Locy for constructive comments on an earlier version of this manuscript. This paper is Alabama Agricultural Experi- ] ment Station Journal No. 6-996053. Literature Cited ' Abacus Concepts. 1992. StatView. Abacus Concepts, ' Inc., Berkeley, CA. i Baker, A. J. M. 1987. Metal tolerance. Pp. 93-111 in I. H. Rorison, J. R Grime, R. Hunt, G. A. E Hendry, and D. H. Lewis (eds.). Frontiers of comparative plant j ecology. Academic Press, London. Baker, A. J. M., J. Proctor, M. M. J. van Balgooy, and R. D. Reeves. 1992. Hyperaccumulation of nickel by the flora of the ultramafics of Palawan, Republic of i the Philippines. Pp. 291-304 in A. J. M. Baker, J. Proctor, and R. D. Reeves (eds.). The vegetation of ultramafic (serpentine) soils. Intercept, Andover Baker, A. J. M. and P. L. Walker. 1990. Ecophysiology | of metal uptake by tolerant plants. Pp. 155-177 in A. J. Shaw (ed.). Heavy metal tolerance in plants: Evo- lutionary aspects. CRC Press, Boca Raton. Boyd. R. S. 1998. Hyperaccumulation as a plant defensive strategy. Pp. 181-201 in R. R. Brooks (ed.). Plants i that hyperaccumulate heavy metals. CAB Internation- al, Oxford. , Boyd, R. S., T. Jaffre, and J. Odom. 1999. Variation in nickel content of the nickel-hyperaccumulating shrub Psychotria douarrei (Rubiaceae) from New Caledo- nia. Biotropica 31:403-410. Boyd, R. S. and T. Jaffre. In review. Phytoenrichment of soil Ni content by Sebertia acuminata in New Cale- donia and the concept of elemental allelopathy. South African Journal of Science. j Boyd, R. S. and S. N. Martens. 1992. The raison d'etre for metal hyperaccumulation by plants. Pp. 279-289 i in A. J. M. Baker, J. Proctor, and R. D. Reeves (eds.), j The vegetation of ultramafic (serpentine) soils. Inter- ; cept, Andover ] Boyd, R. S. and S. N. Martens. 1994. Nickel hyperac- cumulated by Thlaspi montanum var montanum is i acutely toxic to an insect herbivore. Oikos 70:21-25. Boyd, R. S. and S. N. Martens. 1998. The significance of metal hyperaccumulation for biotic interactions. Chemoecology 8:1-7. Boyd, R. S. and W. J. Moar. 1999. The defensive func- tion of Ni in plants: Response of the polyphagous herbivore Spodoptera exigua (Lepidoptera: Noctui- dae) to hyperaccumulator and accumulator species of Streptantlnis (Brassicaceae). Oecologia 118:218-224. Boyd, R. S., J. Shaw, and S. N. Martens. 1994. Nickel hyperaccumulation defends Streptantlms polygalo- ides (Brassicaceae) against pathogens. American Journal of Botany 81:294-300. Brooks, R. R. 1987. Serpentine and its vegetation: a mul- tidisciplinary approach. Dioscorides Press, Portland, OK. Brooks, R. R., J. Lee, R. D. Reeves, and T. Jaffre. 1977. Detection of nickeliferous rocks by analysis of her- barium specimens of indicator plants. Journal of Geo- chemical Exploration 7:49-57. ' Davis, M. A. 1999. Characterization of the metal-based defense of Ni-hyperaccumulating plants. Ph.D. dis- sertation. Auburn University, Auburn, Alabama. Ernst, W. H. O. 1972. Ecophysiological studies on heavy j metal plants in South Central Africa. Kirkia 8:125- 145. Favre, R. M. 1987. A management plan for rare plants in the Red Hills of Tuolumne County, California. Pp. 2000] BOYD ET AL.: NI TOLERANCE AND HYPERACCUMULATION 105 425-427 in T. S. Elias (ed.). Conservation and man- agement of rare and endangered plants. California Native Plant Society, Sacramento. Gabbrielli, R., T. Pandolfini, O. Vergnano, and M. R. Palandri. 1990. Comparison of two serpentine spe- cies with different nickel tolerance strategies. Plant and Soil 122:271-277. Gabbrielli, R., C. Mattioni, and O. Vergnano. 1991. Accumulation mechanisms and heavy metal tolerance of a nickel hyperaccumulator. Journal of Plant Nutri- tion 14:1067-1080. Ghaderian, S. M., a. J. E. Lyon, and A. J. M. Baker. 2000. Seedling mortality of metal hyperaccumulator plants resulting from damping-off by Pythiiim spp. New Phytologist 146:219-224. Homer, F. A., R. S. Morrison, R. R. Brooks, J. Clemens, AND R. D. Reeves. 1991. Comparative studies of nickel, cobalt and copper uptake by some nickel hy- peraccumulators of the genus Alyssum. Plant and Soil 138:195-205. Ingrouille, M. J. AND N. Smirnoff. 1986. Thlaspi cae- rulescens J & C Presl {T. alpestre L.) in Britain. New Phytologist 102:219-233. Jhee, E. M., K. L. Dandridge, A. M. Christy Jr., and A. J. Pollard. 1999. Selective herbivory on low-zinc phenotypes of the hyperaccumulator Thlaspi caeru- lescens (Brassicaceae). Chemoecology 9:93-95. Kramer, U., J. D. Cotter-Howells, J. M. Charnock, A. J. M. Baker, and J. A. C. Smith. 1996. Free histidine as a metal chelator in plants that accumulate nickel. Nature 379:635-638. Kramer, U., R. D. Smith, W. W. Wenzel, I. Raskin, and D. E. Salt. 1997. The role of metal transport and tolerance in nickel hyperaccumulation by Thlaspi goesingense Halacsy. Plant Physiology 115:1641- 1650. Kruckeberg, a. R. 1984. California serpentines: flora, vegetation, geology, soils, and management prob- lems. University of California Press, Berkeley. Kruckeberg, A. R. and R. D. Reeves. 1995. Nickel ac- cumulation by serpentine species of Streptanthus (Brassicaceae): field and greenhouse studies. Madro- no 42:458-469. MacNicol, R. D. and P H. T. Beckett. 1985. Critical tissue concentrations of potentially toxic elements. Plant and Soil 85:107-129. Martens, S. N. and R. S. Boyd. 1994. The ecological significance of nickel hyperaccumulation: a plant chemical defense. Oecologia 98:379-384. Meerts, p. and N. Van Isacker. 1997. Heavy metal tol- erance and accumulation in metallicolous and non- metallicolous populations of Thlaspi caerulescens from continental Europe. Plant Ecology 133:221- 231. Morrison, R. S., R. R. Brooks, and R. D. Reeves. 1980. Nickel uptake by Alyssum species. Plant Science Let- ters 17:451-457. Munz, R a. and D. D. Keck. 1968. A California flora. University of California Press, Berkeley. Nicks, L. J. and M. F. Chambers. 1995. Farming for met- als? Mining Environmental Management 3:15-18. Nicks, L. J. and M. F. Chambers. 1998. A pioneering study of the potential of phytomining for nickel. Pp. 313-325 in R. R. Brooks (ed.). Plants that hyperac- cumulate heavy metals. CAB International, Oxford. Pais, I. and J. B. Jones Jr. 1997. The handbook of trace elements. St Lucie Press, Boca Raton, FL. Pollard, A. J. and A. J. M. Baker. 1997. Deterrence of herbivory by zinc hyperaccumulation in Thlaspi cae- rulescens (Brassicaceae). New Phytologist 132:113- 118. Reeves, R. D. 1992. The hyperaccumulation of nickel by serpentine plants. Pp. 253-277 in A. J. M. Baker, J. Proctor, and R. D. Reeves (eds.). The vegetation of ultramafic (serpentine) soils. Intercept, Andover, MA. Reeves, R. D., A. J. M. Baker, A. Borhidi, and R. Ber- ezain. 1996. Nickel-accumulating plants from the an- cient soils of Cuba. New Phytologist 133:217-224. Reeves, R. D., A. J. M. Baker, A. Borhidi, and R. Ber- ezain. 1999. Nickel hyperaccumulation in the serpen- tine flora of Cuba. Annals of Botany 83:29-38. Reeves, R. D., R. R. Brooks, and R. M. Macfarlane. 1981. Nickel uptake by Californian Streptanthus and Caulanthus with particular reference to the hyperac- cumulator S. polygaloides Gray (Brassicaceae). American Journal of Botany 68:708-712. Rice, E. L. 1984. Allelopathy. Academic Press, Orlando, FL. Sagner, S., R. Kneer, G. Wanner, J. -P. Cosson, B. Deus- Neumann, and M. H. Zenk. 1998. Hyperaccumula- tion, complexation and distribution of nickel in Se- hertia acuminata. Phytochemistry 47:339-347. Schlegel, H. G., M. Meyer, T. Schmidt, R. D. Stoppel, AND M. PiCKHARDT. 1992. A Community of nickel- resistant bacteria under nickel-hyperaccumulating plants. Pp. 305-317 in A. J. M. Baker, J. Proctor, and R. D. Reeves (eds.). The vegetation of ultramafic (ser- pentine) soils. Intercept, Andover, MA. Shen, a. G., F J. Zhao, and S. R McGrath. 1997. Uptake and transport of zinc in the hyperaccumulator Thlaspi caerulescens and the non-hyperaccumulator Thlaspi ochroleucum. Plant, Cell and Environment 20:898- 906. Tadros, T. M. 1957. Evidence of the presence of an eda- pho-biotic factor in the problem of serpentine toler- ance. Ecology 38:14-23. Wild, H. 1978. The vegetation of heavy metal and other toxic soils. Pp. 1301-1332 in M. J. A. Werger (ed.), Biogeography and ecology of southern Africa. Junk, The Hague. Wilson, J. B. and A. D. Q. Agnew. 1992. Positive-feed- back switches in plant communities. Advances in Ecological Research 23:263-336. Zar, J. H. 1996. Biostatistical analysis. Prentice-Hall, En- glewood Cliffs, NJ. Madrono, Vol. 47, No. 2, pp. 106-108, 2000 ANNUAL VARIATION IN XYLEM WATER POTENTIAL IN CALIFORNIA OAKS Johannes M. H. Knops School of Biological Sciences, University of Nebraska, 348 Manter Hall, Lincoln, NE 68588 Walter D. Koenig Hastings Natural History Reservation, University of California, Berkeley, 38601 E. Carmel Valley Road, Carmel Valley, CA 93924 Abstract We measured late-summer predawn, daytime and the overnight recovery of xylem water potential for six years in Querciis lobata Nee, Q. douglasii Hook. & Arn. and Q. agrifolia Nee. Predawn xylem water potential was positively correlated with the rainfall in the previous year, indicating that low rainfall years are experienced as dry years by these oaks. Querciis douglasii had consistently lower xylem water po- tential than the other two species. Predawn values were consistently different among individuals trees and species, but the daytime and recovery values converged in the wet years. These results indicate that one- time measurements of predawn xylem water potential are a good indicator of species and individual tree differences in access to soil moisture. The pressure bomb technique is an easy, reliable method of measuring xylem water potential of trees in the field (Koide et al. 1991; Ritchie and Hinckley 1975; Turner 1981) and is widely used as an indi- cation of water stress of individuals (Callaway et al. 1991; DeLucia et al. 1988; Donovan and Ehler- inger 1994; Knops and Koenig 1994; Kolb and Da- vis 1994; Stringer et al. 1989). Implicit assumptions of these studies are that inter-annual variation in environmental conditions have a minimal effect on either the ranking of individuals or species-specific water stress such that differences in water stress are consistent from year to year. We tested this as- sumption in three species or California oaks by measuring xylem water potential in the same trees in six different years. Based on the large number of trees measured during the first year of the study, we previously reported differences among the species: Quercus douglasii Hook. & Arn. (blue oak), a winter decid- uous species, exhibited low xylem water potential values, indicative of little access to ground water; Q. lobata Nee (valley oak), a second winter decid- uous species, exhibited significantly higher values indicating good access to ground water; and Q. agrifolia Nee (coast live oak), an evergreen species with high daytime and predawn xylem water po- tential values, indicating limited transpiration dur- ing the dry part of the year (Knops and Koenig 1994). Methods This study was conducted at the Hastings Natural History Reservation in central coastal California. The landscape consists of a mosaic of Mediterra- nean grasslands, oak woodlands, chaparral and ri- parian areas (Griffin 1988; Knops et al. 1995). Rainfall was measured daily and we used the an- nual total from September 1 of the previous year through August 30 of the current year. We measured xylem water potential of 14 Q. agrifolia, 13 Q. douglasii and 13 Q. lobata trees. These trees were selected from 250 trees of five species that are part of a long-term study examining acorn productivity (Knops and Koenig 1994, 1997; Koenig et al. 1994, 1996). Trees were selected to represent a gradient in acorn productivity within each species and are located throughout Hastings over a distance of approximately 3.5 km. Xylem water potential was measured using a pressure bomb (PMS Instruments Co.) in Septem- ber of each year, the end of the dry season when temperatures are hottest and water stress the great- est (Knops and Koenig 1994). Daytime measure- ment were made between 1300 and 1700, whereas predawn measurements on the same trees were made on the subsequent night between 0200 and 0600. Recovery was calculated as daytime minus predawn. Shoots with a minimum of three leaves and approximately 5 cm long were cut and imme- diately measured in the field. Shoots were not bagged, because we found no differences with sam- ples pre-bagged in plastic. All daytime shoots were cut in direct sunlight. We measured 2 shoots per tree, except if the values were more than 10% dif- ferent, in which case we measured a third shoot. Measurements were taken in 1991 and 1994 through 1998. In each year trees were measured at the end of the dry season in September Results There was a significant positive relationship be- tween the amount of rainfall and the predawn xy- 2000] KNOPS AND KOENIG: OAK WATER POTENTIALS 107 Rainfall (cm) Fig. 1. Predawn, daytime and recovery (calculated as daytime - predawn) xylem water potential of Q. lobata (n = 14), Q. douglasii (n = 13) and Q. ognfolia (n = 13). Given are the means ±1 S.E. for the years (from left to right) 1994, 1996, 1991, 1997, 1995, 1998. Predawn xy- lem water potential, Q. agrifolia F = 184, = 0.98, P < 0.001; Q. douglasii F = 45, R- = 0.92, P < 0.003; Q. lobata F = 47, R- = 0.92, P < 0.003; midday xylem water potential Q. douglasii F = 11, R^ = 0.73, P < 0.03; all other regressions are P > 0.05. lem water potential of all three species, but there was a significant relationship between daytime xy- lem water potential and rainfall only in Q. douglasii (Fig. 1). None of the species exhibited a significant relationship between the overnight recovery and rainfall (Fig. 1). Although all three xylem water potential mea- surements were significantly concordant among the individual trees over the six-years of the study, pre- dawn values were consistantly more similar from year to year than either of the other measures (Ken- dall's coefficient of concordance, predawn 0.653, day 0.298, recovery 0.184, all Chi-Square >39, all P < 0.000). Quercus douglasii and Q. agrifolia had consis- tently lower predawn xylem water potential values than Q. lobata (one way ANOVA with Scheffe's post hoc comparison; Q. douglasii P < 0.05 in all 6 years; Q. agrifolia P < 0.05 in all years, except 1995). Q. douglasii had consistently the lowest daytime values (significant from Q. lobata in 4 out of 6 years, not in 1995 and 1998), with Q. lobata having intermediate values (significantly different from Q. agrifolia 2 out of 6 years, e.g., 1996 and 1997). Recovery was greater for Q. douglasii and Q. lobata than for Q. agrifolia (significant 4 out of 6 years, not in 1995 and 1998). Discussion Does xylem water potential reflect rainfall? Pre- dawn xylem water potential values were signifi- cantly correlated with rainfall in all three species. This supports the assumption that predawn xylem water potential measured in the driest period of the year reflects relative water availability and that these oak species experience lower water status in dry years. Predawn values were consistently sig- nificantly different among the species, and the dif- ferences were largest in drier years. In contrast, daytime and recovery xylem water potential values converge in wet years (Fig. 1) and have only a lim- ited value in characterizing differences among spe- cies. Are individuals and species different over time? Measurements for individual trees were significant- ly concordant from year to year, more so for pre- dawn than daytime or recovery values. Thus, pre- dawn xylem water potential apparently reflects real and consistent differences among individuals in ei- ther their access to water, in their genetic ability to acquire water or in their ability to conserve water. Our data also support the hypothesis that pre- dawn xylem water potential values are consistently significantly different among the species, with Q. douglasii being more water stress tolerant, because of its low predawn xylem water potential and the significant relationship between rainfall and day- time xylem water potential. However, this scenario does not fit the other two species. These differences are consistent with the limited data on root distri- bution of these three species. Quercus lobata is re- ported as having a deep root system connected to the ground water (Griffin 1973), which might make it less sensitive to the previous 12 months rainfall and more sensitive to long term changes in ground water levels. Quercus agrifolia has an extensive shallow system (Canon 1914a, 1914b) and presum- ably lacks access to the previous winter precipita- tion, which is likely stored in deeper soil levels and Q. douglasii, which does not have consistent access 108 MADRONO [Vol. 47 to groundwater (Griffin 1973). This also raises the alternative hypothesis that these oak species might differ in critical water potential for cavitation (e.g., the formation of unreversible air bubbles within the xylem vessels). Cavitation can be a significant cause of hydraulic conductivity loss within oaks due to water stress (Tognetti et al. 1996, 1998) and species specific differences in vulnerability of cav- itation have been reported for oak species (Cocard et al. 1996; Tognetti et al. 1996). Lastly, differences in water conservation caused by differences in phe- nology and physiology among the species may also have contributed to these patterns. The lack of a relationship for Q. agrifolia and Q. lobata correlation between midday xylem water po- tential and the rainfall and the lack between recov- ery and rainfall for all three species indicates that the degree to which xylem water potential recovers overnight is not dependent on rainfall. Instead, in- dividual trees may be able to lower their predawn xylem water potential, and in the case of Q. doug- lasii daytime xylem water potential, thereby in- creasing water uptake in the driest years. This does not imply that the activity of the trees is the same in each year, as the time that the stomates are open in the daytime might be correlated with the amount of water available for transpiration. Alternatively, this might also indicate a strict regulation for water loss for Q. agrifolia and Q. lobata, via stomatal conductance or adjustment in hydraulic conduc- tance to avoid cavitation and Q. douglasii might have a lower critical threshold for cavitation. Test- ing this would require measuring daily patterns of xylem water potential, hydraulic conductivity and cavitation, which we did not do as part of this study. Acknowledgments Thanks to Louise Johnson for statistical advice, Sarah Hobbie, Bryan Foster, Lars Pierce, Bill Schlesinger for comments, Mark Stromberg and Mark Johnson for logistic help. Literature Cited Callaway, R. M., N. M. Nadkarni and B. E. Mahall. 1991. Facilitation and interfering of Quercus doug- lasii on understory productivity in central California. Ecology 72:1484-1499. Canon, W. A. 1914a. Specialization in vegetation and in environment in California. Plant World 17:223-237. Canon, W. A. 1914b. Tree distribution in central Califor- nia. Popular Science Monthly 85:417-424. Cochard, H., N. Breda and A. Garnier. 1996. Whole tree hydraulic conductance and water loss regulation in Quercus during drought: Evidence for stomatal control of embolism? Annales des Sciences Foresti- eres 53:197-206. DeLucia, E. H., W. H. Schlesinger and W. D. Billings. 1988. Water relations and the maintenance of Sierran conifers on hydrothermically altered rock. Ecology 69:303-311. Donovan, L. A. and J. R. Ehleringer. 1994. Water stress and use of summer precipitation in a Great Basin shrub community. Functional Ecology 8:289-297. Griffin, J. R. 1973. Xylem sap tension in three woodland oaks in Central California. Ecology 54:152-159. Griffin, J. R. 1988. A natural history of Hastings Reser- vation, Carmel Valley, California. Unpublished manuscript on file at the Hastings Reservation. Uni- versity of California, Berkeley. Knops, J. M. H., J. R. Griffin and A. C. Royalty. 1995. Introduced and native plants of the Hastings reser- vation, central coastal California, A comparison. Bi- ological Conservation 71:115-123. Knops, J. M. H. and W. D. Koenig. 1994. Water use strat- egies of five sympatric Quercus species in central coastal California. Madrofio 41:290-301. Knops, J. M. H. and W. D. Koenig. 1997. Site fertility and leaf nutrients of sympatric evergreen and decid- uous species of Quercus in central coastal California. Plant Ecology 130:121-131. Koenig, W. D., J. M. H. Knops, W. J. Carmen, M. T. Stanback and R. L. Mumme. 1996. Acorn produc- tion by oaks in central coastal California: influence of weather at three levels. Canadian Journal of Forest Research 26:1677-1683. Koenig, W. D., R. L. Mumme, W. J. Carmen and M. T. Stanback. 1994. Acorn production by oaks in cen- tral, coastal California: variation within and among years. Ecology 75:99-109. KoiDE, R. T, R. H. RoBicHAUX, S. R. Morse and C. M. Smith. 1991. Plant water status, hydraulic resistance and capacitance. Pages 161-184 in R. W. Pearcy, J. Ehleringer, H. A. Mooney and P. W. Rundell (eds.). Plant Physiological Ecology. Chapman and Hall, London. KoLB, K. J. and S. D. Davis. 1994. Drought tolerance and xylem embolism in co-occurring species of coastal sage and chaparral. Ecology 75:648-659. Ritchie, G. A. and T. M. Hinckley. 1975. The pressure chamber as an instrument for ecological research. Ad- vances in Ecological Research 9:165-254. Stringer, J. W, R J. Kalisz and J. A. Volpe. 1989. Deep tritiated water uptake and predawn xylem water po- tentials as indicators of vertical rooting extent in a Quercus-Carya forest. Canadian Journal of Forest Research 19:627-631. Tognetti, R., A. Longobucco and A. Rashi. 1998. Vul- nerability of xylem to embolism in relation to plant hydraulic resistance in Quercus puhescens and Quer- cus ilex. New Phytologist 139:437-447. Tognetti, R., A. Rashi, C. Beres, A. Fenyvesi and H. W. Ridder. 1996. Comparison of sap flow, cavitation and water status of Quercus patraea and Quercus cerris trees with special reference to computer to- mography. Plant, Cell & Environment 19:928-938. Turner, N. C. 1981. Techniques and experimental ap- proaches for the measurement of plant water status. Pant and Soil 58:339-366. Madrono, Vol. 47, No. 2, pp. 109-115, 2000 MOLECULAR EVIDENCE FOR THE HYBRID ORIGIN OF OPUNTIA PROLIFERA (CACTACEAE) Michael S. Mayer* and Laura M. Williams Department of Biology, University of San Diego, 5998 Alcala Park, San Diego, CA 92110-2492 Jon R Rebman Department of Botany, San Diego Natural History Museum, Balboa Park, P.O. Box 121390, San Diego, CA 92112-1390 Abstract Opimtia prolifera Engelm., (Coastal Cholla) is common to the coastal sage scrub community extending from Ventura County, California to El Rosario, Baja California. On the basis of morphological interme- diacy, O. prolifera is suspected to have originated through hybridization between O. alcahes and O. cholla, both species of coastal and inland deserts of Baja California and Baja California Sur. For an independent test of this hypothesis, we generated RAPD banding patterns from exemplars of different populations of O. prolifera and the putative parents. In order to exclude other potential parents and to distinguish species-specific RAPD bands we included O. bigelovii Engelm., O. ganderi, O. tesajo, and O. wolfii (L. Benson) M. Baker in the screening. The results provide support for the hybridization hy- pothesis as well as some insight into the speciation process. Twenty-nine primers revealed 44 bands shared only between O. prolifera and one or the other putative parent. No other species included in the screening proved to be comparable alternatives to O. alcahes or O. cholla as the parents of O. prolifera. Unique bands are rare ( = 2) in O. prolifera compared with O. alcahes (=19) and O. cholla ( = 23). Trends in the degree of band sharing between O. prolifera and representatives of O. alcahes and O. cholla suggest a central Baja California origin of the species. The dynamic geologic and climatic history of Baja California has fostered a diverse and highly endemic flora on the peninsula, and one of the most speciose genera is Opuntia (Cactaceae). The spe- ciation routes taken by Opuntia have also been di- verse: many species are proven hybrids and many more exhibit multiple ploidal levels (D. Pinkava pers. comm.). One suspected hybrid, Opuntia pro- lifera Engelm., was until recently considered a spe- cies derived through cladogenesis. Opuntia prolifera (Coastal Cholla) occurs in the coastal sage scrub community adjacent to the Pa- cific Ocean between Cedros Island, Baja California and Ventura County, California. This taxon is trip- loid (Pinkava et al. 1992) and reproduces almost exclusively asexually, usually through dispersal of detached stem segments. Morphological interme- diacy of O. prolifera between O. alcahes F. A. C. Weber and O. cholla F. A. C. Weber in several char- acteristics (Table 1) has prompted speculation that O. prolifera may have arisen through hybridization of these species (Rebman 1995). Opuntia alcahes and O. cholla are desert taxa of Baja California and typically diploid (Pinkava et al. 1992; Rebman 1995). The two species commonly grow sympatri- cally without hybridizing (Fig. 1), however a hybrid swarm involving the two species exists near El Ro- sario (Rebman 1995) — which is in the southern part of the range of O. prolifera, an area of overlap be- * Author for correspondence. tween the Sonoran Desert and the California Flo- ristic Province (Fig. 1). Despite the general inter- mediacy of O. prolifera, phenotypic plasticity of the putative hybrid and parent species prevents a strong case for hybridization to rest on morpholog- ical data alone. To subject the hybridization hypothesis to further scrutiny, we surveyed patterns of Random Ampli- fied Polymorphic DNA (RAPD) markers obtained from O. prolifera and its putative parents. The RAPD technique allows relatively quick assessment of a large number of highly polymorphic loci, largely of the nuclear genome (Welsh and Mc- Clelland 1990; Williams et al. 1990). Recent stud- ies have successfully applied the RAPD approach to questions of interspecific hybridization (Pham and Smith 1995; Barker et al. 1996; Daehler and Strong 1997) and hybrid speciation (Smith et al. 1995; Lifante and Aguinagalde 1996; Allan et al. 1997; Padgett et al. 1998). We used RAPD data to test if Opuntia prolifera exhibits the classic genetic expectations of hybrid taxa. If the putative parent species, O. alcahes and O. cholla, were sufficiently divergent genetically prior to a hybridization event, then the hybrid, i.e., O. prolifera, should exhibit additivity of genetic markers specific to the parent species as well as a lack of unique markers (Gallez and Gottlieb 1982). Additionally, the sterile triploid nature of O. pro- lifera suggests the possibility that "fixed" hetero- zygosity (sensu Roose and Gottlieb 1976) in O. 110 MADRONO [Vol. 47 Table 1. Selected Morphological Characteristics of Opuntia prolifera and its Hypothesized Parents, O. al- CAHES and O. CHOLLA, IN REGIONS OF SyMPATRY (FROM REBMAN 1995). Characteristic O. alcahes O. prolifera O. cholla Inner tepal color yellow, green, or red- magenta magenta to deep red light to dark pink Stem segment shape long and narrow (3.5- intermediate (7.5-12.6 short and wide (6-11.5 13 X 1.5-4.5 cm) X 3.0-4.1 cm) X 3-5.5 cm) Tubercule length 4-22 mm 12-24 mm 20-35 mm Tubercule height 2-9 mm 5-9 mm 10-20 mm Spine shape short and thin (4-20 X intermediate (14-18 X long and thick (20-35 0.3-0.5 mm) 0.7-0.9 mm) X 0.8-1.3 mm) Areole size 3-5 X 2-4 mm 5-8 X 3-5 mm 6-1 1 X 3-5 mm Proliferating fruit rare yes yes prolifera could endow it with a higher overall num- ber of RAPD markers relative to its putative par- ents. Finally, patterns in the degree of band sharing between hybrid and parents can also be used to make preliminary inferences regarding the geo- graphic region in which the hybrid taxon arose, as well as the possibility that this event occurred mul- tiple times. Methods Field collection and DNA extraction. Stem seg- ments were gathered from a single plant (exemplar) at each location (Table 2). DNA was extracted from fresh or frozen stem tissue following a modification (Doyle and Doyle 1987) of the hot CTAB method of Saghai-Maroof et al. (1984). Initial RAPD screening. DNA extracts from Opuntia alcahes, O. prolifera, and O. cholla were subjected to DNA amplification via the polymerase chain reaction (PGR) using the 100 10-mer primers of RAPD Oligo Set 3 (Nucleic Acid-Protein Ser- vice Unit of the University of British Columbia). Each 25 |jlL reaction contained 1 unit of Promega (Madison, WI) Taq polymerase, 1 X reaction buffer. Table 2. Collections of Opuntia from California and Mexico Analyzed in the Present Study; Precise Lat./ Long. Data are Available Upon Request. Exemplars are given abbreviated names for reference in text, tables, and figures; B.C. = Baja California, B.C.S. = Baja California Sur; asterisk denotes collections used in initial screening only. Species Collection Location O. alcahes F. A. C. Weber ale 1 ale 2 ale 3 ale 4 O. higelovii Engelmann var. Bigelovii O. cholla F. A. C. Weber cho 1 cho 2 cho 3 cho 4 O. ganderi (C. B. Wolf) J. Rebman & D. J. Pinkava O. prolifera Engelmann pro 1 pro 2 pro 3 pro 4 O. tesajo Engelmann O. wolfii (L. D. Benson) M. A. Baker Rebman s.n.* Voss 1174 Rebman 4157 Rebman 4835 Rebman 5183 Rebman 4956 Rebman 4158 Rebman 4501 Rebman 4827 Rebman 5184 Rebman 4973 Mayer 591 Rebman 3951 Rebman 3977 Rebman 5119 Rebman 4972 Rebman 3820 CA., San Diego Co., Quail Botanical Gar- dens B.C.S. , Cape Region B.C., southwest of Catavina B.C.S. near Rt. 1 and rd. to Punta Abreo- jos B.C. Sur, Sierra Guadalupe CA., San Diego Co., Hwy S-2 at Cane- brake B.C., southwest of Catavina B.C.S., Sierra San Francisco B.C.S., Isla Margarita B.C.S., Sierra Guadalupe B.C., San Felipe Desert, n. of Laguna Diablo CA., San Diego Co., U.S.D. campus. West Point B.C., between La Bocana and Puerto San- to Tom as B.C., s. of Punto Canoas B.C., near La Mision B.C., San Felipe Desert, n. of Laguna Diablo CA., Imperial Co., along 1-8 at Mountain Springs Grade 2000] MAYER ET AL.: ORIGIN OF OPUNTIA PROLIFERA 111 California A Fig. 1. Ranges of Opuntia prolifera, O. alcahes, and O. cholla\ locations of collections used for population-level comparisons are noted by abbreviated names listed in Ta- ble 2. 1.5 mM MgCl., 0.1 mM of each dNTP, 0.2 ^jlM of one primer, and 1 fjil dilute DNA extract. After 2 min at 94°C, the following cycle was repeated 40 times: denaturing at 94°C for 15 s, annealing at 36°C for 1 min, and elongation at 72°C for 1 min. A final elongation segment was held at 72°C for an additional 6 min. The PCR products were separated electrophoretically in 2% agarose gels and banding patterns were visualized by staining with ethidium bromide and inspection under ultraviolet light. Of the 100 primers, twenty-one showed banding poly- morphism and the sharing of bands between ex- emplars of O. prolifera and either O. alcahes or O. cholla; therefore these primers were used in sub- sequent screening experiments. To replicate the patterns observed in the first round of screening and to identify bands shared be- tween O. prolifera and only O. alcahes or O. chol- la, we included other related species (O. bigelovii, O. ganderi, O. tesajo) in new screening experi- ments using the primers identified in the first round. We assumed that any marker that was also present in one of these additional species was a symple- siomorphic characteristic and not helpful in a rig- orous test of the hybridization hypothesis. Opuntia prolifera exhibited a total of five bands that it shared only with both the putative parents, six bands that it shared only with O. alcahes, and eight bands that it shared only with O. cholla. When O. prolifera was compared in the same way with O. bigelovii, O. ganderi, and O. tesajo, the numbers of exclusively-shared markers were zero, two, and one, respectively. Primary RAPD screening. The results of the ini- tial rounds of screening increased our confidence that Opuntia prolifera was a hybrid derivative of O. alcahes and O. cholla. We then examined the distribution of RAPD markers among populations within these species. We employed the same prim- ers that had proven useful in previous rounds of screening and increased our sample sizes of O. pro- lifera, O. alcahes, and O. cholla to include an ex- emplar from four populations of each taxon (Table 2) . In addition, one exemplar each was included from O. bigelovii, O. ganderi, O. tesajo, and O. wolfii. This allowed us to (1) replicate previously observed patterns and identify additional bands shared only between the putative hybrid and its par- ents, (2) get a cursory look at intraspecific RAPD polymorphism, and (3) assess the degree of band sharing on a pairwise population level and, subse- quently, compare these data to the geographic dis- tribution of the populations represented. Results Banding patterns derived from screening 16 ex- emplars using 29 RAPD primers revealed a greater number of markers in support of the hybridization hypothesis than did the initial comparisons (Table 3) , presumably because more of the total variation 112 MADRONO [Vol. 47 Table 3. Primers that Resolve Markers Shared Ex- clusively Between Opuntia prolifera and its Putative Parents. A = O. alcahes, C = O. cholla, A + C = both species. Markers shared with O. prolifera Primer Sequence A C A + C UBC 202 GAGCACTTAC 2 0 0 UBC 204 TTCGGGCCGT 1 0 0 UBC 218 CTCAGCCCAG 1 0 0 UBC 219 GTGACCTCAG 1 1 2 UBC 220 GTCGATGTCG 3 1 0 UBC 225 CGACTCACAG 1 2 1 UBC 226 GGGCCTCTAT 1 2 0 UBC 227 CTAGAGGTCC 0 1 0 UBC 228 GCTGGGCCGA 1 1 0 UBC 238 CTGTCCAGCA 0 1 0 UBC 245 CGCGTGCCAG 1 0 0 UBC 246 TATGGTCCGG 1 1 0 UBC 247 TACCGACGGA 0 2 0 UBC 250 CGACAGTCCC 0 1 1 UBC 253 CCGTGCAGTA 1 2 0 UBC 259 GGTACGTACT 2 1 0 UBC 260 TCTCAGCTAC 1 0 0 UBC 269 CCAGTTCGCC 1 2 0 UBC 270 TGCGCGCGGG 1 2 0 UBC 275 CCGGGCAAGC 0 1 0 UBC 281 GAGAGTGGAA 3 0 1 UBC 283 CGGCCACCGT 1 0 0 within each species was assessed and more primers were successful. Pairwise comparisons between ex- emplars of O. prolifera and O. alcahes, or O. pro- lifera and O. cholla revealed 23 and 21 bands, re- spectively, present in at least one population of the two species compared, and found in no other spe- cies (Tables 3, 4). Of these 44 marker loci, the group of O. prolifera exemplars is polymorphic for at least 31 (>70%). A comparison between O. al- cahes and O. cholla detected just one shared band, which was unique to just one exemplar of each spe- cies. A comparison of O. prolifera exemplars against representatives of O. bigelovii, O. ganderi, O. tesajo, and O. wolfii revealed one, one, zero, and zero bands, respectively, that were exclusively shared. Of the aforementioned markers of hybrid- ization, only a small number are fixed in all ex- emplars of O. prolifera and O. alcahes ( = 3) or O. prolifera and O. cholla ( = 5) (Table 4). Five addi- tional bands were shared exclusively among O. prolifera and both putative parents. Opuntia prolifera did not possess a significantly greater number of bands (P = 0.97) compared with its putative parents: 167 bands total versus 164 in both O. alcahes and O. cholla (Table 4). Compar- ison of O. prolifera with its putative parents also revealed significantly (P < 0.01) fewer unique bands in O. prolifera (n = 2) than in either O. alcahes (n = 19) or O. cholla (n = 23) (Table 4). A factor analysis (Statview 5.0, SAS Institute, Inc. 1998) of Table 4. Summary Data from Primary Screening of RAPD Patterns in Opuntia prolifera (P) and its Pu- tative Parents O. alcahes (A) and O. cholla (C). Characteristic P A C Total bands examined 167 164 164 Unique bands 2 19 23 Bands shared only with P 23 21 Bands shared only with P, fixed for both taxa 3 5 Bands shared only between A and C — 1 Bands shared only among A, C, and P 5 the RAPD data provided the means to assess over- all similarity among the exemplars included in the study. The first two factors account for 41.5% and 16.8% of the variance in the data set. Plotting the exemplars by their scores along factors one and two places O. prolifera clearly intermediate between O. alcahes and O. cholla (Fig. 2). Estimates of banding pattern similarity between pairs of populations of O. prolifera and O. alcahes or O. cholla were made in two ways: using the Simple Matching Coefficient (Sokal and Michener 1958) and the Coefficient of Jaccard (Sneath 1957). We tallied presence or absence of marker bands for all pairwise comparisons of exemplars of O. pro- lifera vs. O. alcahes or O. cholla. We ignored bands that were fixed for all exemplars of the two taxa being compared in an effort to minimize the effect of symplesiomorphies on the coefficient. The Sim- ple Matching Coefficient (SMC) was calculated by adding the matches (shared absences plus shared presences of markers) and dividing by the total number of matches and mismatches. The Coeffi- cient of Jaccard (CJ) omits shared absences from the numerator and denominator. We were con- cerned that the SMC would be biased by artefacts i •• • ® o 1 « ^ • = ale ® = pro 1 O = cho - 1 -.75 -.5 -.25 0 .25 .5 .75 1 Factor 1 Fig. 2. Unrotated factor plot showing position of ex- emplars along factors one and two; refer to Table 2 for key to abbreviations. UBC 202 GAGCACTTAC 2 0 UBC 204 ± J- \_ \jj\jr\jiv_- ^ vjj J. 0 UBC 218 1 0 UBC 219 GTGACCTCAG 1 1 UBC 220 GTCGATGTCG 3 1 UBC 225 rnACTCACAG 1 2 UBC 226 GGGCCTCTAT 1 2 UBC 227 pTAHAnnTrc 0 1 UBC 228 GCTGGGCCGA 1 1 UBC 238 CTGTCCAGCA 0 1 UBC 245 CGCGTGCCAG 1 0 UBC 246 TATGGTCCGG 1 1 UBC 247 TACCGACGGA 0 2 UBC 250 CGACAGTCCC 0 1 UBC 253 CCGTGCAGTA 1 2 UBC 259 GGTACGTACT 2 1 UBC 260 TCTCAGCTAC 1 0 UBC 269 CCAGTTCGCC 1 2 UBC 270 TGCGCGCGGG 1 2 UBC 275 CCGGGCAAGC 0 1 UBC 281 GAGAGTGGAA 3 0 UBC 283 CGGCCACCGT 1 0 2000] MAYER ET AL.: ORIGIN OF OPUNTIA PROLIFERA 113 Table 5. Pairwise Similarity Coefficients Between Exemplars Measuring Opuntia frolifera (pro) x O. alcahes (alc) and O. frolifera X O. CHOLLA (CHO). Simple Matching Coefficients before slash. Coefficient of Jaccard after; see Table 2 for key to abbreviations. pro 1 pro 2 pro 3 pro 4 alc 1 0.539/0.143 0.524/0.231 0.359/0.242 0.225/0.184 alc 2 0.583/0.000 0.568/0.111 0.250/0.069 0.189/0.063 alc 3 0.475/0.160 0.512/0.259 0.525/0.424 0.342/0.308 alc 4 0.436/0.154 0.475/0.250 0.539/0.455 0.400/0.368 cho 1 0.528/0.150 0.421/0.120 0.297/0.212 0.447/0.364 cho 2 0.650/0.263 0.600/0.200 0.263/0.125 0.250/0.167 cho 3 0.528/0.105 0.526/0.182 0.243/0.152 0.368/0.273 cho 4 0.514/0.182 0.462/0.192 0.324/0.242 0.487/0.412 arising from poor amplification, thereby inflating estimates of similarity between two populations. As expected, the SMC values were uniformly greater than the CJ values (Table 5), but in some cases the two approaches yield different patterns of relation- ships among the populations. For example, the two sets of coefficients comparing pro 2 with the four populations of O. alcahes display almost opposite rankings by magnitude (Table 5), perhaps indicat- ing that the inclusion of shared absences does in- deed bias the SMC in this application. Considering, therefore, only the CJ values we see that among all the exemplars of O. alcahes and O. cholla, two exemplars of O. prolifera (pro 1 and 2) exhibited greater similarity, albeit by narrow mar- gins, to alc 3 and cho 2 (Table 5). In contrast, pro 3 and pro 4 exhibited greater similarity to alc 4 and cho 4. These relationships also had geographic sig- nificance: exemplars alc 4 and cho 4 were collected from the same vicinity in northern Baja California Sur, and alc 3 and cho 2 were collected just 40 km apart, also in northern Baja California Sur (Fig. 1). Discussion Molecular evidence supports the proposition that hybridization between Opuntia alcahes and O. cholla gave rise to O. prolifera. Forty-four RAPD markers are shared only between O. prolifera and one or the other parent species; no other candidates emerge as comparable alternatives to O. alcahes or O. cholla as the parents of O. prolifera. As ex- pected for a hybrid, O. prolifera exhibits signifi- cantly fewer unique RAPD markers than its parent species. Moreover, multivariate analysis of marker distribution places exemplars of O. prolifera inter- mediate between those of O. alcahes and O. cholla. Some results of this study, however, were con- trary to early expectations. First, O. prolifera band- ing patterns did not exhibit the greater numbers of loci predicted for a sterile hybrid or allopolyploid (Table 4). This observation may indicate a relative- ly low degree of divergence between O. alcahes and O. cholla, or a low amount of variation derived from the actual hybridization event, or it may ex- pose a limitation of RAPD markers in this appli- cation: RAPDs are dominant, diallelic markers and thus may not show the same patterns of additivity as codominant markers. Another surprising out- come was the RAPD polymorphism evident among exemplars of O. prolifera, indicating interpopula- tional genetic diversity. Because O. prolifera is only known to reproduce asexually, this variation may signify one or more of the following: (1) mul- tiple independent hybrid origins of O. prolifera, (2) undetected sexual reproduction, or (3) genetic di- vergence via somatic mutations. We introduce these alternative processes briefly below, but leave a crit- ical analysis to future studies specifically targeted to discriminating among these phenomena. First, recurrent origin of a triploid O. prolifera would require either that multiple diploid-level hy- bridizations must each have been followed by the production of triploid offspring, or that a pairing of a diploid parent with a tetraploid parent must have occurred multiple times. Because both Opuntia al- cahes and O. cholla are diploid with rare exception (Rebman 1995), the latter scenario seems unlikely. If the former scenario occurred, diploid hybrids should be common and widespread in the zone of sympatry. However, only one diploid count has been documented for O. prolifera (Pinkava and Parfitt 1982). Despite the apparent obstacles to re- curring origins of O. prolifera, RAPD-based rela- tionships among exemplars employed in the present study provide some evidence in its support. Two exemplars of O. prolifera (pro 1 and 2) are more closely related to alc 3 and cho 2 than to the other representatives of O. alcahes and O. cholla. In con- trast, the other two exemplars of O. prolifera (pro 3 and 4) are more closely related to alc 4 and cho 4. Furthermore, specimens alc 4 and cho 4 were collected from the same vicinity, and the locations of alc 3 and cho 2 were separated by just 40 km. Next, for sexual reproduction to be the source of interpopulational variation, triploid O. prolifera plants must give rise to triploid offspring. If meiosis could occasionally generate viable gametes of vary- ing ploidy in O. prolifera, we should expect more ploidal levels than just triploid in these populations. Currently, only a hybrid swarm of the El Rosario 114 MADRONO [Vol. 47 area (Fig. 1) has yielded counts in O. prolifera that exceed triploidy, including a hexaploid — presum- ably an autopolyploid that formed through the fu- sion of two unreduced gametes (Rebman 1995). Lastly, reproduction in O. prolifera relies per- haps exclusively on establishment of detached stem segments (Rebman 1995). Long-term clonal growth allows for the possibility that somatic mutations in branch primordia could generate RAPD variation among populations of O. prolifera. The importance of somatic mutations in clonal species has long been suspected and is gaining more experimental support (Ellstrand and Roose 1987, reviewed in de Kroon and van Groenendael 1997). Origin of Opuntia prolifera. Although it is al- most uniformly triploid across its range, O. prolif- era could have originated as a diploid, through hy- bridization of diploid O. alcahes and O. cholla. Meiotic irregularities in this diploid hybrid allowed the production and subsequent fusion of a reduced and unreduced gamete, generating a triploid off- spring. This route from diploidy to triploidy has been seen repeatedly among cactus species (D. Pin- kava pers. comm.). A notable example is O. bige- lovii, a close relative of O. prolifera, which appar- ently arose as a diploid but is now predominantly triploid (D. Pinkava pers. comm.). If indeed O. pro- lifera originated in this way, some set of factors then allowed the triploid to surpass its diploid pro- genitor and thrive in the coastal sage scrub of the Califomias, a habitat to which few other chollas are well-adapted. All exemplars of O. prolifera showed the great- est similarity to representatives of O. alcahes (ale 3 and 4) and O. cholla (cho 2 and 4) collected from the northern end of Baja California Sur, indicating a possible region of origin of O. prolifera. Surpris- ingly, this region is greatly disjunct from the pres- ent range of O. prolifera (Fig. 1). However, the repeated shifts in climate and vegetation in the his- tory of Baja California cautions us from excluding this proposition prior to further investigation. Establishment of Opuntia alcahes and O. cholla as the parents of O. prolifera now sets the stage for further population genetic studies, which should be aimed towards testing for recurrent origins of O. prolifera and the route by which it attained triplo- idy. Acknowledgments We thank Douglas Gilbert and Michelle Darnell for as- sistance in the lab. Rick Gonzalez for help with statistical tests, and two anonymous reviewers for comments on an earlier version of this paper. This research was supported in part by a Faculty Research Grant from the College of Arts and Sciences of the University of San Diego to M. S. M., and by the Undergraduate Research Fund of the Associated Students of the University of San Diego. Literature Cited Allan, G. J., C. Clark, and L. H. Reiseberg. 1997. Dis- tribution of parental DNA markers in Encelia virgi- nensis (Asteraceae: Heliantheae), a diploid species of putative hybrid origin. Plant Systematics and Evolu- tion 205:205-221. Barker, N. P., L. Y. Kruger, J. H. Vlok, and E. H. Har- LEY. 1996. RAPD data suggest the Gamkapoort Aloe to be of hybrid origin. South African Journal of Bot- any 62:292-295. Daehler, C. W. and D. R. Strong. 1997. Hybridization between introduced smooth cordgrass {Spartina al- terniflora; Poaceae) and native California cordgrass (S. foliosa) in San Francisco Bay, California, USA. American Journal of Botany 84:607-611. Doyle, J. J. and J. L. Doyle. 1987. A rapid DNA isola- tion procedure for small amounts of fresh leaf tissue. Phytochemical Bulletin 19:11-15. De Kroon, H. and J. Van Groenendael (eds.). 1997. The Ecology and Evolution of Clonal Plants. Backhuys, Leiden, The Netherlands. Ellstrand, N. C. and K. L. Roose. 1987. Patterns of genotype diversity in clonal plant species. American Journal of Botany 74:123-131. Gallez, G. p. and L. D. Gottlieb. 1982. Genetic evidence for the hybrid origin of the diploid plant Stephano- meria diegensis. Evolution 36:1158-1167. LiFANTE, Z. D. AND I. Aguinagalde. 1996. The use of Random Amplified Polymorphic DNA (RAPD) markers for the study of taxonomical relationships among species of Asphodelus sect. Verinea (Aspho- delaceae). American Journal of Botany 83:949-953. Padgett, D. J., D. H. Les, and G. C. Crow. Evidence for the hybrid origin of Nuphar X rubrodisca (Nympha- ceae). American Journal of Botany 85:1468-1476. Pham, T V. and J. F. Smith. 1995. Genetic variation and hybridization in the narrow endemic Allium aaseae based on RAPD data. American Journal of Botany 82 (supplement: abstracts): 155-156. AND B. D. Parfitt. 1982. Chromosome numbers in some cacti of western North America-IV. Bulletin of the Torrey Botanical Club 109:121-128. , , M. A. Baker, and R. D. Worthington. 1992. Chromosome numbers in some cacti of western North America-Vl, with nomenclatural changes. Ma- drono 39:98-113. Rebman, J. P. 1995. Biosystematics of Opuntia subgenus Cylindropuntia (Cactaceae), the chollas of Lower California, Mexico. Ph.D. dissertation. Arizona State University. Roose, M. L. and L. D. Gottlieb. 1976. Genetic and biochemical consequences of polyploidy in Trago- pogon. Evolution 30:818-830. Saghai-Maroof, M. a., K. M. Soliman, R. A. Jorgensen, AND R. W. Allard. 1984. Ribosomal spacer lengths in barley: Mendelian inheritance, chromosomal lo- cation and population dynamics. Proceedings of the National Academy of Sciences, U.S.A. 81:8014- 8018. Smith, J. F, C. C. Burke, and W. L. Wagner. 1995. In- terspecific hybridization in natural populations of Ha- waiian Cyrtandra: evidence from RAPD markers. American Journal of Botany 82 (supplement: ab- stracts): 162. 2000] Sneath, p. H. a. 1957. Some thoughts on bacterial clas- sification. Journal of General Microbiology 17:184- 200. SoKAL, R. R. AND C. D. MiCHENER. 1958. A statistical method for evaluating systematic relationships. Uni- versity of Kansas Science Bulletin 38:1409-1438. Welsh, J. and M. McClelland. 1990. Fingerprinting ge- 115 nomes using PGR with arbitrary primers. Nucleic Ac- ids Research 18:7213-7218. Williams, J. G. K., A. R. Kubelik, K. J. Livak, J. A. Rafalski, and S. V. TiNGEY. 1990. DNA polymor- phisms amplified by arbitrary primers are useful as genetic markers. Nucleic Acids Research 18:6531- 6535. MAYER ET AL.: ORIGIN OF OPUNTIA PROLIFERA Madrono, Vol. 47, No. 2, pp. 116-126, 2000 FLORAL VARIATION IN DELPHINIUM VARIEGATUM (RANUNCULACEAE) Shan A C. Dodd' Department of Biology, San Diego State University, San Diego, California 92182 Kaius Helenurm- Department of Biology, University of South Dakota, Vermillion, South Dakota 57069 Abstract Delphinium variegatum is subdivided into three subspecies distinguished by three floral characters. Delphinium v. variegatum is found in central and northern California, while D. v. kinkiense (an endangered taxon) and D. v. thornei are endemic to San Clemente Island off the coast of southern California. Broad variation is documented in most natural populations for all three floral characters. Our results indicate that the two metric characters, lateral sepal length and lower petal blade length, provide no clear distinction between the taxa. Sepal color is the least ambiguous for differentiating the subspecies, but is problematic in distinguishing between D. v. kinkiense and D. v. thornei. Sepal color exhibits a complex pattern of variation on San Clemente Island in which northern populations generally contain primarily light-flowered individuals, southern populations generally contain primarily dark-flowered individuals, and central pop- ulations may contain substantial numbers of both light- and dark-flowered individuals as well as inter- mediates. However, one southern population contains primarily light-flowered individuals, and almost half of the populations contain individuals having sepal colors considered to represent the two different sub- species. Further taxonomic study including additional characters is recommended to determine whether D. V. kinkiense and D. v. thornei should be considered distinct taxa. Delphinium variegatum Torrey & A. Gray (Ran- unculaceae) is a perennial larkspur that is found in grassland and open woodlands of mainland Cali- fornia and San Clemente Island, the southernmost of the Channel Islands off the coast of southern California (Wamock 1990b). One subspecies, D. v. ssp. variegatum (Royal larkspur), is found exclu- sively on the mainland and ranges approximately from northern to central California, from the coast to the Sierra Nevada ?? foothills (Fig. 1). The other two subspecies, D. v. ssp. kinkiense (Munz) M. J. Wamock (San Clemente Island larkspur; Wamock 1990a) and D. v. ssp. thornei Munz (Thome's lark- spur; Munz 1969) are insular endemics found only on San Clemente Island. The Channel Islands are thought to provide refuge for a number of species with northem affinities, including D. variegatum (Raven and Axelrod 1978), that once extended far- ther south on the mainland during Pleistocene plu- vial cycles (Raven 1963). The island endemic subspecies of D. variegatum are vulnerable to extinction because of their limited distribution (Skinner and Pavlik 1994). Delphinium V. kinkiense is listed as endangered by the U.S. Fish & Wildlife Service (USFWS) and by the Califomia Department of Fish and Game. However, the rarest of the subspecies, D. v. thornei, has no special legal 'Current address: S. C. Dodd Biological Consulting, 3786 Dana Place, San Diego, California 92103. ^Author for correspondence: Department of Biology, University of South Dakota, Vermillion, South Dakota 57069. E-mail: helenurm@usd.edu. status, although the USFWS considers it to be a species of concern. Both of these taxa are on the Califomia Native Plant Society List IB (plants rare, threatened, or endangered in Califomia and else- where; Skinner and Pavlik 1994). The subspecies of D. variegatum are distin- guished primarily by three floral characters: sepal color, lateral sepal length and lower petal blade length (Wamock 1990b, 1993, 1997; summarized in Table 1; Fig. 2). However, there is overlap among the subspecies. The mainland subspecies, D. V. variegatum, is differentiated from the two island subspecies by its deep royal blue flowers, as the ranges for the two metric characters encompasses the variation observed in the entire species. The two island subspecies are differentiated from each other by all three characters, in spite of consider- able overlap, with D. v. kinkiense having mainly white, smaller flowers and D. v. thornei having mainly bright blue, larger flowers. Munz (1974), interestingly, had described D. v. thornei as having smaller flowers (sepals ca. 12 mm long) than D. v. kinkiense (which he recognized as a separate spe- cies, D. kinkiense; sepals 16-18 mm long). Current keys use sepal color to identify taxa (Wamock 1990b, 1993, 1997), although the most recent also uses density of hairs on the base of the stem to distinguish between D. v. variegatum and the island subspecies (Wamock 1997). Casual observation of natural populations of D. variegatum on San Clemente Island suggests that hybridization may be occurring between D. v. kin- kiense and D. v. thornei in some populations. At 2000] DODD AND HELENURM: FLORAL VARIATION IN DELPHINIUM VARIEGATUM 1 17 Fig. 1. Distribution and sampled populations of Delphinium variegatum from mainland California (ssp. vahegatum) and from San Clemente Island (ssp. kinkiense and ssp. thornei). the time this study was initiated, fewer than 15 pop- ulations of D. variegatum were known on San Cle- mente Island (along with scattered individuals), with D. V. kinkiense occurring in the northern half of San Clemente Island and D. v. thornei in the southern half. However, some populations in the central part of San Clemente Island include indi- viduals exhibiting white, bright blue or intermediate flower colors. Natural hybridization has been doc- umented to regularly occur among other taxa in the 118 MADRONO [Vol. 47 Fig. 2. Diagram of (a) side view and (b) front view of a D. variegatum flower. genus Delphinium (Wamock 1990b, 1997); natural hybrids are known between D. v. variegatum and D. hansenii, D. hesperium, D. parryi, and D. re- curvatum (Lewis and Epling 1954). The goals of this study were (1) to document variation in floral morphology in natural popula- tions of the subspecies of D. variegatum, (2) to evaluate the utility of the three floral characters for distinguishing D. v. kinkiense and D. v. thomei, and (3) to identify populations of D. variegatum on San Clemente Island as D. v. kinkiense, D. v. thomei or mixed populations. Materials and Methods Study sites. Twenty-four populations of D. v. kin- kiense and D. v. thomei were located and sampled from San Clemente Island in 1996 (Fig. 1; Table 2). This represents all known populations and prob- ably all of the populations on the island; subsequent surveys have failed to reveal additional locations (Junak and Wilken 1998; S. Burckhalter, University of South Dakota, pers. comm.; K. Helenurm, per- sonal observation). Seven populations of D. v. var- iegatum were sampled across its range, from Marin County in the north to southern Monterey County and east to Tuolumne and Mariposa Counties. All populations of the three subspecies occurred in open grassland habitat. Island populations were found only on west or northerly aspects, probably due to moister, cooler conditions in these areas. Flower collection and measurements. Thirty to forty-four flowering individuals were haphazardly sampled from large populations (Table 2). In small- er populations, all flowering individuals were sam- pled. Two flowers from each sample individual were measured for sepal color, lateral sepal lengths, and lower petal blade lengths (Fig. 2). Sepal color was measured by matching lateral sepals to a color chart (Royal Horticultural Society 1986). Colors were quantified by matching the color chart patches to colors in Adobe Photoshop (1995) computer software using a calibrated monitor. We recorded their hue, saturation, and brightness values using the same computer system for all measurements. Values for brightness were used for analysis be- cause brightness (quantifying the degree of light- ness or darkness, ranging from 0 representing black to 100 representing white) best reflects Wamock's (1990b) descriptions of the subspecies and the range of variation in flower color we observed. Al- though differences in hue (the attribute of colors that permits them to be classed as blue versus lav- ender or purple, for example) and saturation (the degree of difference from a gray having the same lightness) occur, the quantifiable difference between white, light blue, bright blue and deep royal blue sepals is reflected in brightness values rather than hue or saturation. In all, 775 individuals were measured for floral characters in all 24 San Clemente Island popula- tions, and 242 individuals were measured in 7 mainland populations, for a total of 1017 individ- uals. Analysis. Measurements of brightness and metric characters were averaged for different flowers of the same individual. Associations among the dif- ferent floral characters in D. v. kinkiense and D. v. thomei were addressed in three ways. First, t-tests were used to test differences in lateral sepal and lower petal blade lengths in individuals with light versus dark sepal color. Second, correlations among the floral characters were tested using Pearson's correlation analysis. Third, grouping of floral char- acters was investigated using principal components analysis (PC A). All analyses were performed using SYSTAT (1992). Results Variation in floral morphology. Box plots of flo- ral variation in D. variegatum illustrate broad vari- ation in most natural populations on San Clemente Island (Fig. 3). Sepal color is invariant, or nearly so, in some populations (populations 1-7, 10). The metric characters, lateral sepal length and lower Table 1 . Floral Characters used to Distinguish the Three Subspecies of Delphinium variegatum (Summarized FROM Warnock 1990b, 1993, 1997). Delphinium variegatum Floral character ssp. kinkiense ssp. thomei ssp. variegatum Sepal color white to light blue (or light blue to bright deep royal blue, rarely lavender) blue white or lavender Lateral sepal length 11-18 mm 17-21 mm 10-25 mm Lower petal blade length 4-9 mm 6-1 1 mm 4-1 1 mm 2000] DODD AND HELENURM: FLORAL VARIATION IN DELPHINIUM VARIEGATUM 119 Table 2. Population Number, Subspecies Designation, Collection Locations, Approximate Population Sizes (1996), AND Sample Sizes of Delphinium Variegatum. San Clemente Island populations are listed north to south. Population Population Sample number Subspecies Location size size Island 1 kinkiense Flasher Canyon ZUU ZO z kinkiensc Nots Drive 900 i kinkiense Pelican Canyon 9^nn ZjUU 44 A 4 kinkiense Larkspur Canyon 1 JU /in 4U C J kinkiense Stone Canyon J / z: O kinkiense Burns-Horton Canyon /in 4U 1 kinkiense Lower Twin Dams Canyon iO Q Q O mix Boulder ZUU /in 4U Q mix Upper Twin Dams Canyon 1 OOO lU kinkiense VVctllCIl \_^tlliyUli 900 ZwO J 1 1 1 i 1 thornei Upper Middle Ranch Canyon ID 19 jZ 1 9 mix ^SO ^9 JZ 1 ^ ID kinkiense Waynuk Canyon 1 OOO 41 1 A thornei North Norton Canyon AO 1 7 1 / 1 < 1 J IrlUi fit: I ^r^ntVi Mr»i*tr\Ti ^r»r»\/r^n oUULll INUiliJll v^dliyuil soo 1 o ItlUI fie I \j 1 7 iflur fie I JjiJA v^dHyiJll ^0 1 8 1 o thornei v_.cive v^diiyoii 4.00 D 1 1 Q iflur fie I 1 '^O JO ifiuf fie I 7 / VJ 21 thnrnci Bryce Canyon 200 39 22 thornei Malo 300 31 23 thornei Canchalagua Canyon 3000 40 24 kinkiense Guds 75 35 Mainland 1 variegatum Edgewood County Park 200 33 2 variegatum China camp State Park 100 38 3 variegatum Green Springs Road 40 22 4 variegatum Chinese Station 200 40 5 variegatum Route 49 150 39 6 variegatum Nacimiento-Ferguson Road 200 31 7 variegatum G14 250 39 petal blade length, are highly variable in most pop- ulations, with largely overlapping ranges. Mainland populations show a similar pattern (Fig. 3). Sepal color is relatively invariant in main- land populations with the exception of China Camp State Park, in which many white-flowered individ- uals occurred (18 of the 38 sampled). Mainland populations also show variation in metric charac- ters, but they generally have narrower distributions with fewer outside values. Edgewood County Park has longer lateral sepals and lower petal blades than the other mainland populations. Histograms of the three floral characters indicate lighter-colored and larger flowers for the island populations (treated together because of the broadly overlapping distributions noted above) than for mainland populations (Fig. 4). Sepal color is dis- tributed bimodally on the mainland only because of white-flowered individuals in China Camp State Park. The distribution of sepal color on San Cle- mente Island is clearly bimodal, with 375 of the 775 individuals sampled (48.4%) having white or very light blue flowers (henceforward "Hght-flow- ered"; brightness values from 88-100), 72 (9.3%) being intermediate (brightness values 56-87), and 328 (42.3%) having bright blue flowers (hencefor- ward "dark-flowered"; brightness values 28-55). In contrast, the metric characters have unimodal distributions. The overall bimodal distribution of sepal color on San Clemente Island shows a geographic pat- tern. Northern populations generally contain pri- marily light-flowered individuals and southern pop- ulations generally contain primarily dark-flowered individuals (Fig. 5). Central populations may con- tain substantial numbers of both flower types as well as intermediates. Association among floral characters. The lateral sepal lengths of San Clemente Island individuals with light (brightness values 88-100) and dark (brightness values 28-55) sepal colors are signifi- cantly different (t = -5.78, df = 698, P < 0.0001), but their means (16.31 mm and 17.07 mm, respec- tively) and ranges (12.0-21.75 mm and 10.75- 24.25 mm, respectively) are very similar. Likewise, lower petal blade lengths are significantly different for the two brightness classes (t = -6.1 1, df = 689, 120 MADRONO [Vol. 47 100 90 80 Sepal 70 brightness (%) 60 50 40 30 Lateral sepal length (mm) 16 T ^ 1 J I L J I I I L 12 1 1 10 Lower petal g blade length (mm) 8 7 6 5 a J I I L 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 1 2 3 4 5 6 7 San Clemente Island Mainland California Fig. 3. Box plots of (a) brightness, (b) lateral sepal length, and (c) lower petal blade length in populations of D. variegatum. Median values (central line, defining the 50th percentile), upper and lower hinges (edges of the central box, defining the 25th and 75th percentiles), whiskers (extending to the farthest observation from the hinges not farther than 1.5 times the distance between the hinges), outside values (asterisks, observations farther from the hinges than 1.5 times the distance between the hinges), and far outside values (open circles, observation farther from the hinges than 3.0 times the distance between the hinges) are illustrated. P < 0.0001), but their means (6.86 mm and 7.22 mm, respectively) and ranges (5.0-9.5 mm and 5.0-10.12 mm, respectively) are almost identical. Three of the four means fall within the overlapping portion of the ranges described for the two subspe- cies (Wamock 1990b, 1993, 1997). A strong correlation exists between lateral sepal length and lower petal blade length (Pearson's r = 0.619, P < 0.0001; Fig. 6). Weaker correlations ex- ist between brightness and lateral sepal length (Pearson's r = -0.213, P < 0.0001) and between brightness and lower petal blade length (Pearson's r = -0.216, P < 0.0001). PCA groups individuals primarily by flower col- or with a broad range of lateral sepal and lower petal blade lengths for each color class (Fig. 7). The first two axes account for 58.11% and 29.20% of the total variation, for a total of 87.31%. Plots of the first and third and of the second and third axes (not illustrated) are dense clouds of points showing no structure. The deep royal blue sepal color of D. v. varie- gatum is significantly different from both light- flowered and dark-flowered island plants (mean = 38.19, range = 30.0-100.0; F = 1778.9, df = 2, r^ = 0.798, P < 0.0001; Tukey HSD multiple com- parison P < 0.0001 for both comparisons). Lateral sepals in D. v. variegatum are significantly shorter than in light-flowered and dark-flowered island plants (mean = 15.66, range = 11.25-20.5; F = 2000] DODD AND HELENURM: FLORAL VARIATION IN DELPHINIUM VARIEGATUM 121 Number of individuals 400 300 - 200 100 - Mainland SCI 28 32 36 40 44 48 52 56 60 64 68 72 76 80 84 88 92 96 100 Sepal brightness (%) Number of individuals 200 100 Mainland SCI 1 i 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Lateral sepal length (mm) Number of individuals 300 200 100 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 1 0 10.5 1 1 1 1.5 1 2 Lower petal blade length (mm) Fig. 4. Histograms of brightness, lateral sepal length, and lower petal blade length for populations of D. variegatum on San Clemente Island. X-axis values represent the minimum of a class. 37.2, df = 2, r2 = 0.076, P < 0.0001; Tukey HSD multiple comparison P < 0.0001 for both compar- isons). Lower petal blades are also significantly shorter in D. v. variegatum (mean = 6.34, range = 4.00-12.15; F = 47.7, df = 2, r^ = 0.097, P < 0.0001; Tukey HSD multiple comparison P < 0.0001 for both comparisons). Discussion Floral variation. All three floral characters ex- hibit substantial variation within populations. The metric characters, lateral sepal length and lower petal blade length, exhibit unimodal distributions both on the mainland and on San Clemente Island. Mainland populations have smaller flowers than is- land populations, with the exception of Edgewood County Park in which lower petal blades lengths even exceed those of San Clemente Island plants. Sepal color is relatively invariant on the main- land, although it shows a bimodal distribution due to the high proportion of white-flowered individuals in China Camp State Park. In contrast, sepal color is clearly bimodally distributed on San Clemente Island. Most island populations of D. variegatum are highly variable in sepal color, although some consist primarily of white-flowered individuals. Floral characters and taxonomy. Warnock (1990b) divided D. variegatum into three subspe- cies primarily on the basis of sepal color, lateral sepal length, and lower petal blade length. The 122 MADRONO [Vol. 47 Fig. 5. Pie charts of flower color in San Clemente Island populations of D. variegatum. White areas represent pro- portion of individuals with white or light blue flowers (brightness values from 88 to 100), black areas represent individuals with bright blue flowers (brightness ^55), and stippled areas represent individuals with intermediate colors (brightness values from 56 to 87). mainland populations we sampled generally fit Wamock's (1990b) descriptions of D. v. variegatum (Table 1), although some individuals in Edgewood County Park have lower petal blade lengths ex- ceeding the described taxonomic range. Delphinium V. variegatum differs from the two island subspe- cies in generally having darker (deep versus bright blue) flowers (except for many individuals in China Camp State Park) and shorter lower petal blades (except in Edgewood County Park). Considerable population differentiation appears to exist within D. v. variegatum. Of the seven pop- ulations we sampled, two are morphologically dis- tinct: China Camp State Park has a high proportion of white-flowered individuals (absent in the other populations we sampled), and Edgewood County 2000] DODD AND HELENURM: FLORAL VARIATION IN DELPHINIUM VARIEGATUM 123 25 Lateral sepal length (mm) 20 15 10 I I • 4 A A % A I'Ml' r- • * 1. • • i 30 — I— 40 — 1— 50 — I— 70 60 70 80 Sepal brightness (%) — r- 90 1 00 10- ' Lower petal blade length (mm) 4 H 1 1 1 1 1 1 1 1 1 1 1 1 1 30 40 50 60 70 80 90 100 Sepal brightness (%) Fig. 6. Scatterplots of sepal brightness, lateral sepal length, and lower petal blade length for individuals of all pop- ulations of D. variegatum on San Clemente Island. Taxonomic designations shown are from Warnock (1990b, 1993, 1997). Park has larger flowers than other populations. Sub- sequent to sampling, we discovered that Edgewood Park is the only population we sampled that occurs on serpentine soils. Warnock (1990b) considers ser- pentine soil populations of D. v. variegatum to be not well marked morphologically and did not rec- ognize them as distinct taxa. Instead, Warnock (1997) comments that plants with large flowers are common in the San Francisco Bay area, either as scattered individuals or as populations made up largely of such individuals. In other species, plants growing on serpentine soils have often been docu- mented to be morphologically and genetically dis- tinct from plants growing on non-serpentine soils (Kruckeberg 1954; Mayer et al. 1994). Intensive sampling of additional natural populations of D. v. 124 MADRONO [Vol. 47 Factor 1 Fig. 7. Scatterplot of the first and second PC A axes for floral characters in D. variegatum on San Clemente Island. Individuals are coded as having Hght (brightness values 88-100), intermediate (brightness values 56-87) or dark (brightness values 28-55) flower color. variegatum may clarify whether serpentine soil populations are differentiated morphologically from non-serpentine soil populations or whether varia- tion is geographic in pattern. Delphinium v. kinkiense and D. v. thornei also generally conform to Wamock's (1990b, 1993, 1997) descriptions, in spite of some individuals having lateral sepals exceeding the described tax- onomic range (Fig. 6). However, our data do not support the separation of D. v. kinkiense and D. v. thornei on the basis of all three floral characters. Specifically, our results indicate that the two metric characters, lateral sepal length and lower petal blade length, provide no clear distinction between these taxa. Both exhibit a unimodal distribution on San Clemente Island so that any attempt to use them for delineating taxa is necessarily arbitrary. In addition, the majority of individuals fall within or near the area of overlap for these characters. Both lateral sepal length and lower petal blade length show a statistically significant association with sepal color, indicating that light-colored flow- ers tend to be smaller than dark-colored flowers. These results are in agreement with Wamock's (1990b) descriptions. However, their significance should be considered an artifact of large sample size. The almost identical means and almost com- pletely overlapping ranges of the metric characters for light- versus dark-colored flowers indicate that the statistical differences have little taxonomic sig- nificance. The remaining floral character, sepal color, is the least ambiguous for differentiating between D. v. kinkiense and D. v. thornei but it is also problem- atic, exhibiting a more complex pattern of variation on San Clemente Island than previously suspected. Although northern populations generally contain primarily light-flowered individuals and southern populations generally contain primarily dark-flow- ered individuals, central populations may contain substantial numbers of both flower types as well as intermediates. Moreover, this is only a general trend as the southernmost population (Guds) is predom- inantly, although not exclusively, light-flowered. In addition, nearly half of the populations ( 1 1 of 24) contain both light- and dark-flowered individuals, thus having individuals with sepal colors consid- ered to represent different subspecies. The complex pattern of variation observed for sepal color on San Clemente Island may be due to hybridization and subsequent introgression between the taxa. Genetic data may provide evidence regarding this possibil- ity. The discrepancy between our results and the tax- onomic separation of D. v. kinkiense and D. v. thor- nei is probably due to our intensive sampling of all natural populations of D. variegatum on San Cle- mente Island. The taxonomic descriptions of these 12000] DODD AND HELENURM: FLORAL VARIATION IN DELPHINIUM VARIEGATUM 125 ^ taxa are based on examination of herbarium spec- imens (M. Warnock, University of Missouri, Co- lumbia, pers. comm.) that may have represented , only a fraction of the variation found in natural populations. This is clearly a potential problem with any taxon endemic to remote locations, es- pecially in cases where access is highly restricted (permission of the U.S. Navy is required to visit San Clemente Island). I Classification of San Clemente Island popula- i tions. Because D. v. kinkiense is listed as endan- gered and D. v. thornei is merely considered a spe- cies of concern by the USFWS, it is necessary for ! practical reasons to identify populations on San Clemente Island. Since our analyses show poor sep- aration between subspecies for lateral sepal and lower petal blade lengths, sepal color was used to classify populations as D. v. kinkiense, D. v. thor- nei, or mixed. Populations having at least 80% light-flowered individuals (brightness values be- tween 88 and 100) were classified as D. v. kin- kiense, and populations having at least 80% dark- flowered individuals (brightness values below 56) were classified as D. v. thornei. Populations with less than 80% of its individuals in either brightness range were classified as mixed. This criterion is based on the observed bimodal distribution of sepal color. Using this criterion, there are 10 populations of D. V. kinkiense, 1 1 populations of D. v. thornei, and 3 mixed populations (Table 2, Fig. 5). Although an 80% criterion seems to be a weak basis for distinguishing taxa, it may be preferable to a stricter classification. If Warnock's (1990b, 1993, 1997) descriptions are interpreted in con- junction with the bimodal distribution we have doc- umented, then individuals with brightness values from 88 to 100 may be classified as D. v. kinkiense (white to light blue flowers) and individuals with brightness values below 88 may be classified as D. V. thornei (light blue to bright blue flowers). Using this criterion, there are 7 populations of D. v. kin- kiense, 5 populations of D. v. thornei, and 1 1 mixed populations (Fig. 5). The results of this study indicate that D. v. kin- kiense and D. v. thornei are, at best, currently sep- arable only on the basis of sepal color. They may be more appropriately classified as varieties rather than subspecies or classified together as one sub- species (as defined by Stuessy 1990). However, it is not uncommon for plant taxa to be separated on the basis of morphological characters controlled by only one or two loci, such as flower color (Bach- mann 1983; Gottheb 1984; Hilu 1983). Moreover, other characters such as flowering time may clearly distinguish D. v. kinkiense and D. v. thornei. The northern populations of predominantly light-flow- ered individuals flower earlier than the southern, dark-flowered populations (S. Junak, Santa Barbara Botanic Garden, pers. obs.), although this may have an environmental rather than a genetic basis. Fur- ther taxonomic study using additional characters should be conducted to decide whether the island taxa have been appropriately designated as separate subspecies. Correct taxonomic designation has practical implications for the survival of these taxa because only D. v. kinkiense has legal protection. Acknowledgments The authors thank Steve Burckhalter, Steve Junak, Liz Kellogg and Jane Rombouts for providing many popula- tion locations on San Clemente Island, Toni Corelli and Mike Warnock for providing population locations on the mainland, and Jan Larson and Jennifer Stone for logistic support. This research was funded by the Natural Re- sources Office, Staff Civil Engineer, Naval Air Station, North Island, San Diego, California. Literature Cited Adobe Photoshop. 1995. Adobe Photoshop computer software, version 3.0. Adobe Systems Incorporated, San Jose, CA, USA. Bachmann, K. 1983. Evolutionary genetics and the ge- netic control of morphogenesis in flowering plants. Evolutionary Biology 16:157-208. Gottlieb, L. D. 1984. Genetics and morphological evo- lution in plants. American Naturalist 123:681-709. Hilu, K. W. 1983. The role of single-gene mutations in the evolution of flowering plants. Evolutionary Bi- ology 16:97-128. Junak, S. A. and D. H. Wilken. 1998. Sensitive plant status survey. Naval Auxiliary Landing Field, San Clemente Island, California. Santa Barbara Botanic Garden Technical Report No. 1, Santa Barbara, CA, USA. Kruckeberg, a. R. 1954. The ecology of serpentine soils. III. Plant species in relation to serpentine soils. Ecol- ogy 35:267-274. Lewis, H. and C. Epling. 1954. A taxonomic study of Californian Delphiniums. Brittonia 8:1-22. Mayer, M. S., P. S. Soltis, and D. E. Soltis. 1994. The evolution of the Streptanthus glandulosus complex (Cruciferae): genetic divergence and gene flow in ser- pentine endemics. American Journal of Botany 81: 1288-1299. MuNZ, p. A. 1969. California miscellany VII. Aliso 7:65- 71. . 1974. A flora of southern California. University of California Press, Berkeley, California, USA. Raven, R H. 1963. A flora of San Clemente Island, Cal- ifornia. Aliso 5:289-347. AND D. I. AxELROD. 1978. Origin and relationships of the California flora. University of California Press, Berkeley, CA, USA. Royal Horticultural Society. 1986. Royal Horticultur- al Society colour chart (editions 1,2). Royal Horti- cultural Society, London. Skinner, M. W. and B. M. Pavlik. 1994. California Na- tive Plant Society Inventory of rare and endangered vascular plants of California. California Native Plant Society, Sacramento, CA, USA. Stuessy, T. F. 1990. Plant taxonomy: the systematic eval- uation of comparative data. Columbia University Press, New York. SYSTAT. 1992. SYSTAT: Statistics, Version 5.2 Edition. SYSTAT, Inc., Evanston, Illinois, USA. 126 MADRONO [Vol. 47 Warnock, M. J. 1990a. New taxa and combinations in North American Delphinium (Ranunculaceae). Phy- tologia 68:1-6. . 1990b. Taxonomic and ecological review of Cal- ifornia Delphinium. Collectanea Botanica 19:45-74. . 1993. Delphinium. Pp. 916-922 in J. C. Hickman ed. The Jepson manual: higher plants of California, 916-922. University of California Press, Berkeley, CA. . 1997. Delphinium. Pp. 196-240 in Flora of North America Editorial Committee ed. Flora of North America, Volume 3. Oxford University Press, NY. Madrono, Vol. 47, No. 2, pp. 127-133, 2000 CROWN STRUCTURE OF THE WORLD'S SECOND LARGEST TREE Stephen C. Sillett and James C. Spickler Department of Biological Sciences, Humboldt State University, Areata, CA 95521 Robert Van Pelt College of Forest Resources, Box 352100, University of Washington, Seattle, WA 98195 Abstract We studied the crown structure of the Washington Tree {Sequoiadendron giganteum (Lindley) Buch- holz) in Sequoia National Park, CaHfornia. The tree was 77.3 m tall and 9.1 m basal diameter. Its total volume was 1403.2 m\ including the main trunk (1357.3 m^) and 46 reiterated trunks (45.8 m^). The main trunk was hollow, and 133.2 m^ of wood volume was missing. A 35-m deep, 2-3-m wide pit extended into the heart of the main trunk below 58 m. The microclimate at the bottom of the pit was dark, cool, and humid. Fire and fungal decay apparently contributed to the formation of the pit. Some charred wood was evident throughout the pit, but most of this had fallen away and been replaced by decayed wood. The walls of the pit in the lower 17 m were spongy, wet, and covered by fungal mycelia. Sequoiadendron giganteum (Lindley) Buchholz (giant sequoia) is an awe-inspiring species restrict- ed to 66 groves in California's Sierra Nevada (Wil- lard 1995). These are the world's largest living trees (Flint 1987), and some individuals are over 3000 years old (Hartesveldt et al. 1975). As such, they have attracted a great deal of scientific interest. They were among the earliest trees in North Amer- ica to be studied from a canopy perspective; rope techniques and an elevator were used to sample cones and arboreal arthropods in the 1970's (Har- vey et al. 1980). However, the crown structure of ancient S. giganteum trees has never been studied. The crown structure of ancient Sequoia semper- virens (D. Don) Endl. trees, the closest living rel- atives of S. giganteum, has recently been the focus of research (Sillett 1999). Like many other conifers, including S. giganteum, S. sempervirens grows via a simple architectural model consisting of a vertical trunk that supports numerous horizontal branches. Ancient trees, which have endured centuries of wind and fire, develop highly individualized crowns consisting of multiple, resprouted trunks (i.e., vertically oriented stems) arising from other trunks and branches. One very complex S. semper- virens tree, for example, has a crown with 148 res- prouted trunks accounting for over 14 percent of its total wood volume (Sillett and Van Pelt 2000). Such extra trunks are reiterations of the tree's ar- chitectural model (Halle et al. 1978). They are in- distinguishable from free-standing trees except for their locations within the crown of a larger tree. Each reiterated trunk supports its own system of horizontal branches. This is the first rope-based study of crown struc- ture in an ancient S. giganteum tree. We used meth- ods developed in 5. sempervirens to map the crown of the Washington Tree, the second largest S. gi- ganteum tree (Flint 1987). Our objective was to cal- culate the tree's total volume, including reiterations. During our exploration of the tree, we discovered a deep pit extending into the heart of the tree's main trunk. We compared microclimatic conditions in- side the pit with those on top of the crown. Study Area The Washington Tree is located at 2085 m ele- vation near the center of Giant Forest in the south- ern Sierra Nevada of California (36°33'N, 118°45'W). At 1212 hectares. Giant Forest is the second largest of the unlogged S. giganteum groves (Willard 1995). The average temperature ranges from 0.2°C in January to 17.9°C in July with an annual average of 8.1°C. Much of the area's 110 cm precipitation falls as snow (482 cm average ac- cumulation), and a snowpack persists into spring. The summer is very dry; only 2 cm of precipitation falls between July and September (National Park Service). The Washington Tree grows near a large granite outcrop in a forest dominated by S. giganteum, Abies concolor (Gordon & Glend.) Lindley, and Pi- nus lambertiana Douglas. The area has been burned within the last 15 years as part of a prescribed burn- ing program carried out by the National Park Ser- vice. Abundant regeneration of S. giganteum and P. lambertiana is visible in the vicinity of the Washington Tree and the nearby Franklin Tree. Methods Tree access. We accessed the Washington Tree by shooting a rubber-tipped fiberglass arrow trailing 10 kg test strength Fireline® filament over sturdy branches with a compound bow mounted to a spin- ning reel. A 3 mm nylon cord, followed by a 10 mm static kemmantle climbing rope, was then hauled over the branches. We anchored one end of MADRONO [Vol. 47 the rope at ground level and climbed the other us- ing mechanical ascenders. We used a 20 m long arborist's rope lanyard to access progressively high- er branches and to move laterally through the crown. We worked in the tree crown during a two- week period from May to June 1999. Crown mapping. We mapped crown structure of the Washington Tree by measuring dimensions of the main trunk and all reiterated trunks (both living and dead) over 5 cm basal diameter. All trunk di- ameters were measured directly using a graduated tape. The main trunk's diameter was measured at the highest ground surface, which was 0.65 m above true ground level (i.e., the average of the highest and lowest ground surfaces), at 2.5 m in- tervals from true ground level to 15 m, and at 5 m intervals above 15 m. Branches and reiterated trunks prevented us from obtaining diameter mea- surements at every height interval. In these cases, we measured trunk diameter as close to the regular interval as possible. Since the main trunk was hol- low (see RESULTS), we also measured the maxi- mum and minimum diameters of the hollow cavity at 5 m intervals. A series of reiterated trunks extended well above the broken top of the main trunk. The tree's total height was measured by lowering a tape from the topmost foliage to true ground level. We recorded the following data for each reiterated trunk: top height, height of origin, basal diameter, and diam- eter at 5 m intervals along the length of the trunk. For reiterated trunks arising from branches, we also recorded horizontal distance to main trunk, branch height, branch basal diameter, and branch diameter at reiteration. We only surveyed branches giving rise to reiterated trunks; no other branches were measured. Reiterated trunks were referenced to the main trunk by recording azimuths and distances be- tween them at 5 m height intervals. We used an Impulse® laser range finder (Laser Technology, Inc.) to measure horizontal distances between trunks. All other measurements were made with the aid of a compass, clinometer, and graduated tape. Since large trunks often gave rise to smaller trunks, we sketched crown structure and noted physical connections between trunks and branches. We also noted whether trunks were monopodial (i.e., con- sisting of a single axis), sympodial (i.e., consisting of successive axes), or otherwise broken. All in- formation was used to generate a tree crown dia- gram. Crown illustrations. We made an illustration of the entire Washington Tree from the ground. Major branches, clumps of foliage, burls, and kinks were Fig. 1. Illustration of the Washington Tree prepared by Robert Van Pelt on the basis of crown structure data and photographs taken from the ground. 2000] SILLETT ET AL.: SECOND LARGEST TREE'S CROWN STRUCTURE 129 noted on a sketch, and heights to these landmarks were measured with the Impulse® laser. We then took photographs of all portions of one side of the crown from as far away from the tree as possible. The illustration itself started with a "skeleton," which was based on the height and diameter mea- surements taken both from within the crown and from the ground. We used the photographs to pro- vide details of foliage, branches, and bark texture. The illustration was drawn at '/120 scale, and human figures were added for additional scale. We also made a detailed illustration of the upper crown, in- cluding the entrance to the pit. Photographs and sketches made from a nearby hill supplemented sketches made from within the tree crown. Microclimate sampling. We measured light, air temperature, and relative humidity at three posi- tions in the tree's crown. Sensors were lashed to the tree's uppermost branches with nylon cord to obtain measurements from the top of the crown (77 m). Sensors were suspended on a rope to obtain measurements from the top of the pit (57 m) and the bottom of the pit (23 m). Microclimate mea- surements were made at 4-minute intervals over a 23-hour period (11 a.m. June 11 to 10 a.m. June 12). We used HOBO® RH/Temp Loggers to mea- sure air temperature and relative humidity and Stow Away® Light Intensity Loggers to measure light intensity (Onset Computer Corporation). Calculations. We used Wendell Flint's ground- level survey data for Washington Tree (Flint un- published) to determine the basal diameter of the main trunk. Direct tape measurements overestimate basal area by failing to account for missing wood in fire cavities and spaces between buttresses. We calculated the surface area of the tree's "footprint" and converted this to a true basal diameter (diam- eter = 2 [footprint area/ir]'')- Trunk volume was cal- culated by applying two different equations to the trunk diameter data. We used the equation for a parabolic frustum (i.e., volume = length/2* [A 1 + A2], where Al and A2 are the upper and lower trunk cross sectional areas) for sections of trunks that tapered slowly. We used the equation for a reg- ular conic frustum (i.e., volume = length*Tr/ 3* [lower diameter^ -I- (lower diameter)* (upper di- ameter) + upper diameter^]) for sections of trunks that tapered more rapidly, such as within the crown. Using the latter equation, we also calculated the volume of branches supporting reiterations based on the limited data we collected on these branches (see above). Results Volume. The Washington Tree is 77.3 m tall and 9.1 m basal diameter (Fig. 1). A large fire cave has been burned away from the main trunk near the ground. It consists of an inner portion that is 3.7 m high, 2.4 m wide, and 2.6 m deep as well as an outer portion that is 5.8 m high, 4.2 m wide, and 0.5 m deep. Above 58 m, the main trunk is broken (Fig. 2). A shield of wood extends 10 m above this break and terminates two structures: a partially dead reiteration and the splintered remains of the main trunk. According to our calculations, the Washington Tree's main trunk has a volume of 1357.3 m\ including the volume occupied by the pit (see below). Reiterated trunks (see below) add an additional 45.8 m^ of volume, the three largest reiterations accounting for 79.6 percent of this total. Thus, the Washington Tree's total trunk volume is 1403.2 m\ Branches bearing reiterations on the Washington Tree (n = 22) add an additional 25.6 m^ volume, but we emphasize that our measure- ments of branch volume are incomplete. Reiterations. The Washington Tree has 46 reit- erated trunks arising from the main trunk (n = 3), other trunks (n = 10), and branches (n = 33) (Fig- ure 2). The largest reiteration, which is 1.7 m basal diameter and 16 m long, sprouts from the main trunk at the mouth of the pit and terminates in a dead broken stump that is 0.8 m diameter. It sup- ports 19 other trunks, including two dead and seven sympodial trunks. Like the main trunk, the largest reiterated trunk is hollow with a lesser pit that is 5.1 m deep. Another large reiteration sprouts from the end of a 1.1 m diameter branch emanating from the main trunk at the mouth of the pit. Its top is dead, and it supports four other trunks. Eight trunks emerge from the backside of the shield above the pit. Two of these trunks are completely dead, and two have dead tops. All of the rest of the Washing- ton Tree's reiterations arise from branches emanat- ing from the main trunk. The largest of these trunks is 1 m basal diameter and 25.2 m long. It sprouts from the end of a 4 m long, 1 .2 m diameter branch at 35.7 m above the ground. The Washington Tree's largest branch, which supports three trunks, is 1.4 m basal diameter, 10 m long, and located 45.4 m above the ground. The pit. A 35-m-deep pit extends down into the heart of the Washington Tree from the break in its main trunk below the shield (Figs. 2 and 3). The pit is over 2 m diameter at the mouth, and it en- larges to over 3 m diameter farther down. At the bottom (i.e., 22.8 m above ground level), the pit is 3.05 by 0.62 m wide. Charred wood is evident along the walls of the pit throughout most of its length, and fungal decay becomes pronounced with increasing depth. Below 40 m, fungal mycelia are evident along the walls, and the wood is soft and wet. Humus has accumulated on ledges of rotting wood. Massive protrusions of dead wood extend from the walls of the main trunk into the pit, evi- dence of ancient branch bases. At the bottom of the pit, these branch bases are small, and the humus is deep, rich, and filled with rotting seed cones. An- nual rings are visible in the spongy wood, some of them perhaps 3000 years old (Stephenson in press). The pit occupies 133.2 m\ or about 10 percent, of 130 MADRONO [Vol. 47 80 70 60- 50 c 2 « 40 > o X) X3 30 20 10- 10 4 3 2 1 0 1 2 3 4 distance from central axis (m) 10 11 12 Fig. 2. Crown diagram of the Washington Tree. All trunk diameters are drawn to the scale of the x-axis, which is expanded relative to the scale of the y-axis. Circles correspond to the basal diameters of reiterated trunks. Serrated edges indicate broken trunks. Branches bearing reiterated trunks are depicted with single straight lines. No other branches are shown. Gray lines indicate dead branches and trunks. Gray lines also indicate a basal fire cavity as well as maximum and minimum diameters of the pit. the main trunk's volume. Thus, the Washington Tree's total wood volume is 1270.0 m\ There are two narrow fissures in the main trunk far above the bottom of the pit (i.e., 59.5-51.7 m and 36.3-33.4 m). The upper fissure is located in a large area of dead wood on the outside of the main trunk that extends from the shield to 49.4 m above the ground. Some light and wind pass into the pit through the opening, which is up to 10 cm wide. The lower fissure is much narrower. It affords no breeze, but there is enough light to support an epix- ylic green alga on the inner surface of the opening. Microclimate. At the bottom of the pit, the mi- croclimate was relatively dark, cool, and humid during the 23 -hour sampling period. Light intensity during the day was less than 0. 1 percent as high as on top of the crown and less than 1 percent as high as on top of the pit. Temperatures at the bottom of the pit remained a constant 4°C for the duration of the sampling period. The other locations were 4.5 to 1 1°C warmer than the bottom of the pit. The top of the pit was 0.5 to 3°C cooler during the day and up to 2°C warmer during the night than the top of the crown. Relative humidity at the bottom of the 2000] SILLETT ET AL.: SECOND LARGEST TREE'S CROWN STRUCTURE 131 Fig. 3. Detailed illustration of the Washington Tree's upper crown, including the mouth of the pit, prepared by Robert Van Pelt on the basis of crown structure data and photographs. This view is from a location 120° clockwise of the view in Figure 1. 132 MADRONO [Vol. 47 pit remained a constant 99 percent for the duration of the sampling period. The other locations were 23 to 60 percent less humid than the bottom of the pit. The top of the pit was 5 to 25 percent more humid during the day and up to 10 percent less humid during the night than the top of the crown. Discussion Tree size. The Washington Tree has been known as the world's second largest tree for over a decade (Flint 1987), and our direct crown-level measure- ments confirm this fact. Among living trees, only the General Sherman Tree, whose main trunk has a volume of 1489 m^, is larger (Van Pelt in press). The main trunk of the General Grant Tree, whose volume is 1357.3 m\ is identical in volume to the Washington Tree's main trunk, but the General Grant Tree has a smaller volume of reiterated trunks (Van Pelt in press). However, the Washing- ton Tree's total wood volume, including reitera- tions, is actually less than that of the General Grant Tree, the President Tree (1318.0 m\ Van Pelt in press), and the Lincoln Tree (1275.2 m\ Van Pelt in press) because its main trunk is hollow. Among living S. giganteum, many trees are taller than the Washington Tree, and several trees (e.g., Ishi, Grant, and Boole) have larger bases (Van Pelt in press). The Washington Tree, however, is the largest living tree from 3.2 to 4.3 m (main trunk diameters 7.6 to 7.0 m) and from 50.0 to 58.5 m (main trunk diameters 4.6 to 3.8 m) above the ground. Prior to losing the upper part of its main trunk, the Washington Tree may have been the only living tree larger than the General Shennan Tree (Flint 1987). Crown structure. Like their close relative. Se- quoia sempervirens, ancient Sequoiadendron gi- ganteum trees can have complex crowns consisting of multiple reiterated trunks. Ancient trees of both redwood species frequently possess branches that gradually curve upwards and become increasingly trunk-like over time. Perfectly vertical reiterated trunks, however, appear to be less common on S. giganteum than on S. sempervirens. And unlike S. sempervirens, old S. giganteum branches tend not to be heavily buttressed; they appear almost circular in transverse section. Furthermore, flagelliform branches and fusions between trunks and branches, which are common in ancient S. sempervirens crowns (Sillett 1999; Sillett and Van Pelt 2000), are rarely encountered on S. giganteum. It is difficult to compare the Washington Tree's crown structure with that of other ancient S. gigan- teum trees because no others have been thoroughly mapped using rope-based methods of access. Ground-level surveys permit a few preliminary comparisons. Unlike the Washington Tree, several of the largest S. giganteum trees (e.g.. Grant, Lin- coln, Stagg, Boole, Genesis) have very few (if any) reiterated trunks (Van Pelt in press). The General Sherman Tree's crown, however, is highly reiterat- ed, and a few of these reiterations are larger than any on the Washington Tree (Van Pelt in press). The Franklin Tree (1222.7 m^ volume. Van Pelt in press), which grows within 500 m of the Washing- ton Tree, also has a highly reiterated crown, but nearly all of its large reiterations are dead. Fire and fungi. Tops of the main trunks on many ancient S. giganteum trees are dead (Rundel 1973), and most trees have fire scars throughout their crowns. The Washington Tree is no exception. Trunks that have been hollowed out by fire are also commonly observed, but the Washington Tree's pit may be unique. In reference to hollow trunks, John Muir wrote, "All of these famous hollows are burned out of solid wood, for no Sequoia is ever hollowed by decay" (1909). He clearly never ob- served the Washington Tree's pit! There is no doubt that fungal decay has played a major role in creating the pit; the heartwood is rotting and spongy, and fungal mycelia are abun- dant. Indeed, microclimatic conditions in the lower portion of the pit promote fungal decay. After the top of the main trunk broke away, fire probably initiated formation of the pit, perhaps via a mech- anism similar to the one Muir observed on fallen logs during a fire (1909): After the great glowing ends fronting each other have burned so far apart that their rims cease to bum, the fire continues to work on in the centres, and the ends become deeply concave. Then heat being radiated from side to side, the burning goes on in each section of the trunk independent of the other, until the diameter of the bore is so great that the heat radiated across from side to side is not sufficient to keep them burning. It ap- pears, therefore, that only very large trees can receive the fire-auger and have any shell rim left. But precipitation accumulated in the newly formed pit, and the moist wood was ultimately colonized by decay fungi that increased its size over many years. Subsequent fires probably consumed much of the decaying wood (Piirto et al. 1984). We ob- served some charred wood on the walls of the pit to within 0.7 m of the bottom, so fire has clearly contributed to the hollowing of the pit to a great depth. However, most of the charred wood has fall- en away and been replaced by decaying wood. Thus, both fire and fungal decay were directly in- volved in the formation of the pit. Acknowledgments Funding from Global Forest supported Sillett during this research (OF- 18-2000-48). MacGillivray-Freeman Films and the Adventures in Wild California film crew provided logistical support. Dana Laughlin, Brett Love- lace, and Shanti Revotskie assisted with the field work. Erik Jules, Nathan Stephenson, and two anonymous re- 2000] SILLETT ET AL.: SECOND LARGEST TREE'S CROWN STRUCTURE 133 i! viewers provided constructive comments on the manu- 1 script. Finally, Wendell Flint graciously provided us with his ground-level survey data for the Washington Tree. Literature Cited Flint, W. D. 1987. To find the biggest tree. Sequoia Nat- ural History Association. Three Rivers, CA. Halle, E, R. A. A. Oldeman, and R B. Tomlinson. 1978. Tropical trees and forests: an architectural analysis. Springer- Verlag, New York. Hartesveldt, R. J., H. T. Harvey, H. S. Shellhammer, j AND R. E. Stecker. 1975. The giant sequoia of the Sierra Nevada. U.S. Department of the Interior, Na- tional Park Service, Washington, DC. Harvey, H. T, H. S. Shellhammer, and R. E. Stecker. I 1980. Giant sequoia ecology. U.S. Department of the Interior, National Park Service, Scientific Monograph Series 12. Washington, DC. MuiR, J. 1909. Our National Parks. The Riverside Press, Cambridge. PiiRTO, D. D., W. W. Wilcox, J. R. Parmeter Jr., and D. L. Wood. 1984. Causes of uprooting and breakage of specimen giant sequoia trees. Bulletin 1909. Division of Agriculture and Natural Resources, University of California, Berkeley. RuNDEL, P. W. 1973. The relationship between basal fire scars and crown damage in giant sequoia. Ecology 54:210-213. SiLLETT, S. C. 1999. Tree crown structure and vascular epiphyte distribution in Sequoia sempervirens rain forest canopies. Selbyana 20:76-97. SiLLETT, S. C. AND R. Van Pelt. 2000. A redwood tree whose crown is a forest canopy. Northwest Science 74:34-43. Stephenson, N. L. 2000. Estimated ages of some large giant sequoia: General Sherman keeps getting youn- ger. Madrono (in press). Van Pelt, R. In press. Forest giants of the Pacific coast. University of Washington Press, Seattle. WiLLARD, D. 1995. Giant sequoia groves of the Sierra Ne- vada: a reference guide. Dwight Willard, Berkeley, CA. Madrono, Vol. 47, No. 2, pp. 134-137, 2000 ERIOGONUM SPECTABILE (POLYGON ACEAE): A NEW SPECIES FROM NORTHEASTERN CALIFORNIA Beth Lowe Core in Lassen National Forest, 2550 Riverside Dr., Susanville, CA 96130 James L. Reveal Norton-Brown Herbarium, University of Maryland, College Park, MD 20742-5815 Robin Barron 1731 Country Lane, Placerville, CA 95667 Abstract Eriogonum spectabile, a new species of the subgenus Eucycla, is described from northeastern Plumas County in northeastern California, USA. It differs from the related E. pendulum of northwestern California and adjacent southwestern Oregon in being a shorter, more compact plant with more numerous branches at the base, narrower leaves with the pubescence equally distributed on both surfaces, longer petioles, reduced umbellate inflorescences, broadly campanulate involucres, and densely pubescent flowers with gland-tipped hairs among the silky-white ones. The new species is currently known only in an extremely limited area of glaciated andesite southeast of Lassen Peak. A new species of Eriogonum Michx. was en- countered during project field surveys on the Las- sen National Forest in northeastern California on 30 July 1997. This plant was immediately recog- nizable as distinctly different from other known Er- iogonum species in this part of the state, and indeed from all other known Eriogonum by the combina- tion of its low shrubby habit and densely pubescent flowers and fruits. Subsequent surveys have result- ed in a total of only three occurrences with about 250 plants total, all within about 1 km of each oth- er. Eriogonum spectabile B. L. Corbin, Reveal, & R. Barron, sp. nov. (Fig. 1). — TYPE: USA, Califor- nia, Plumas Co., ca. 13 km north of Chester, ca. 1.9 km west-southwest of Hay Meadows trail- head to the Caribou Wilderness, Lassen National Forest, T30N, R7E, sect. 28 NE Va of SW Va, MDM, ca. 40°25'N, 12ri2'W, 18 Aug 1998, Corbin and Earll 910 (Holotype: US; Isotypes DAV, JEPS, K, MARY, NY, RSA, Lassen Na- tional Forest herbarium.) Planta perennis, suffrutex, usque 2.5 cm alta; caules patentes dense ramosi fragiles; petioli breves 0.6—0.9 cm longi; lamina foliaris anguste elliptica, 0.5-1.7 cm longa, 0.4-0.7 cm lata, in superficiebus ambabus sericea cinereo-tomentosa, marginibus in- tegris planis vel revolutis praedita. Caules florentes scaposi, primo albi-tomentosi deinde glabrati; inflo- rescentia umbellata; bracteae semifoliaceae, angus- ti-oblongae vel-ellipticae, albi-tomentosae; involu- cra solitaria, extus sericeo-tometosa atque dense gladularia intus glabra; dentes involucrales 5-7, acute triangulares, 1.0-1.2 mm longi; bracteolae numerosae lineares. Flores sub anthesi albi, sub gemmascentia frutescentiaque rosei vel subrubri, nervo medio fuscato omati; tepala extus dense hir- suta intus subglabra, per anthesin 4.0-4.5 mm lon- ga; stamina inclusa vel exserta; antherae rubrae vel purpureae. Achenia 3-4 mm longa, tomentosa, ad basem subglobosam in rostrum crassum 3-angulum contracta. Plants low shrubs 1-1.5 dm high vegetatively and 1.7-2.5 dm high in flower, mostly 3-5.5 dm across; stems spreading, densely branched, brittle, arising from a stout, woody taproot (up to 5 cm across at the top of the taproot), the older branches with reddish-brown bark exfoliating in wide strips. Leaves arranged in open rosettes mostly at the base of the flowering stem or at the tips of exposed caudex branches, others sheathing shortly up the herbaceous stems; leaf blades narrowly elliptic, (0.7-)1.2-1.7(-2.2) cm long, (0.2-)0.4-0.7(-0.9) cm wide, equally densely gray and somewhat silky tomentose on both surfaces; leaf margins entire and plane to re volute; leaf apex broadly acute, the base cuneate. Petiole short, (0.2-)0.6-0.9(-1.3) cm long, silky tomentose. Petiole base elongate triangular, 3- 5 mm long, 2-4 mm wide, densely white tomentose without, glabrous within. Flowering stems scapose, erect (1.5-)6.1-13.3(-17) cm long, white tomen- tose, becoming glabrate at maturity. Inflorescences umbellate. Bracts mostly semifoliaceous, 2-4(-6), narrowly oblong to narrowly elliptic, 2-5 mm long, 1-2 mm wide, white tomentose. Peduncles slender, slightly spreading, (1.5-)2-8.5(-9.5) cm long, silky tomentose to glabrate at maturity, occasionally some peduncles have an extra whorl of bracts mid- length. Involucres solitary, broadly campanulate, the tube 2-3 mm long, (2-)3-4(-5) mm wide, silky tomentose and densely glandular without, glabrous 2000] CORBIN ET AL.: ERIOGONUM SPECTABILE 135 Fig. 1. Eriogonum spectabile. Left-habit. Right-involucre. Drawn from Corbin 906. 136 MADRONO [Vol. 47 within; involucral teeth 5-7, acutely triangular, 1- 1.2 mm long; bractlets numerous, linear, 2-3 mm long, fringed with long, silky hairs and minute gland-tipped cells. Pedicels 2.5—4 mm long, gla- brous below, glandular and slightly hairy above. Flowers white (in anthesis) to pink or reddish (in fruit, also in bud). Stipe essentially lacking, 0.1-0.2 mm long. Tepals 6, with slightly darker greenish to reddish bases and midribs (midribs greenish within and dark pink without), 4-6 mm long (mostly 4- 4.5 mm in flower, longer in fruit), densely hairy without with long, slender, silky-white hairs and short, capitate glands, essentially glabrous within except for minute glands and some hairs mainly along the midrib; tepals essentially similar, obovate, those of the inner whorl slightly longer than the outer whorl, united for less than V4 of their length. Stamens ? included to exerted, 2.5-3(-6.5) mm long, the filaments sparsely pilose at the base; an- thers red when fresh, purplish-red to purple when dried, 0.6-0.7 mm long, broadly ovate. Gynoccium with a style 1-1.3(-1.5) mm long. Achenes light brown, 3-4 mm long, tomentose, the subglobose base tapering to a stout, 3-angled beak. Paratypes. Topotypes — 30 Jul 1997, Barron s.n. (Lassen National Forest herbarium), 18 Aug 1997, Barron s.n. (CHSC, JEPS, MARY, Lassen National Forest herbarium), 8 Sep 1997, Corbin et al. 861 (MARY), 8 Aug 1998 Corbin et al. 906 (Lassen National Forest herbarium). USA, California: Plu- mas Co., ca. 2.2 km SW of Hay Meadows trailhead to the Caribou Wilderness, T30N, R7E, sect. 28 SW 1/4 of SW 1/4, MDM, 23 Sep 1997 Corbin 882 (MARY, Lassen National Forest herbarium); ca. 2.2 km WSW of Hay Meadows trailhead, T30N, R7E, sect. 28 NW 1/4 of SW 1/4, MDM, 23 Sept 1997 Barron s.n. (Lassen National Forest herbarium). Eriogonum spectabile is most closely related to E. pendulum S. Watson; both are members of an as yet undescribed section of the subg. Eucycla (Nutt.) Kuntze. The new species is a shorter, more compact plant than E. pendulum with more numerous branches at the base, narrower leaves with the pu- bescence equally distributed on both surfaces, lon- ger petioles, reduced umbellate inflorescences, broadly campanulate involucres, and densely pu- bescent flowers with gland-tipped hairs among the silky-white ones. Eriogonum pendulum is found in extreme northwestern California (Del Norte Co.) and adjacent southwestern Oregon (Josephine Co.) in dry sandy soil in mixed evergreen forests not unlike that of E. spectabile (see below). The spe- cific epithet refers to the spectacular appearance of this small shrub, which is quite attractive. We sug- gest the common name "Barron's buckwheat" to acknowledge the first collector. Eriogonum spectabile appears to be limited to three small occurrences, all within one quarter sec- tion. The first discovered occurrence (and the type locality) is the largest; 194 plants were counted on 8 Sep 1997. The second occurrence had 54 plants on 23 Sep 1997, and the third only three (also on 23 Sep 1997). Numbers were similar in 1998 and 1999 visits. All known locations are on the Lassen National Forest in northeastern California within about 5 km of Lassen Volcanic National Park, and about 1 .2 km from the Forest's Caribou Wilderness Area. This region is considered part of the southern limit of the Cascades Range. Extensive searches in adjacent Lassen National Forest areas were made in 1997, 1998, and 1999. Lassen Volcanic National Park has been fairly well botanized in the past (Os- wald et al. 1995, Gillett et al. 1961), but no collec- tions from Lassen Volcanic National Park are known. The Caribou Wilderness Area contains ex- tensive apparently suitable habitat, much of which has not been surveyed, so Eriogonum spectabile may occur there as well. The new species grows in open areas on minor ridges within a Pinus contorta Loudon, subsp. mur- rayana (Grev. & Balf.) Critchf., Abies magnifica Andr. Murray var. magnifica, and A. concolor (Gor- don & Glend.) Lindley forest, at 2010 to 2025 m elevation. The general area is Quaternary glacial deposits (Lydon et al. 1976), with moraines form- ing low ridges interspersed by several small kettle lakes. Glaciation is particularly evident at the type locality, as shown by glaciated andesite bedrock or large boulders with smooth and striated surfaces, and chatter marks on the larger surfaces. Arctostaphylos nevadensis A. Gray, which is abundant in this general area, is the species most closely associated with Eriogonum spectabile; how- ever, E. spectabile occurs only in the less common openings between individuals of Arctostaphylos nevadensis. Other less abundant associates include Achnatherum occidentale (Thurbes) Barkworth subsp. californicum (Merr. & Burtt Davy) Bark- worth, Arctostaphylos patula E. Greene, Ceanothus prostratus Benth., Cymopterus terebinthinus (Hook.) M. E. Jones var. californicus (J. Coulter & Rose) Jepson, and Helianthella californica DC. var. nevadensis (Greene) Jeps. The area receives about 60 inches (150 cm) of precipitation per year, mostly as snow (Ranz 1969). Plants were in late flower on 30 Jul 1997 and 18 Aug 1997 visits, and in fruit in early Sep 1997. The spring of 1998 was much colder. The site was snow-covered on 17 Jun 1998; the plants were mostly vegetative on 22 Jul 1998, in bud on 8 Aug 1998, and in fuU bloom on 18 Aug 1998. Plants were in full flower on 26 Jul 1999, and still flow- ering and in early fruit on 31 Aug 1999. At all three occurrences, most individuals ap- peared to be mature shrubs, and many had dead wood about the base. Few seedlings (only four at the largest occurrence in 1998, and none at the smallest) or apparently young plants were ob- served. Not all plants at an occurrence flowered: only 18 percent of the second occurrence flowered in 1997, but 60 percent of the adult plants at the 2000] CORBIN ET AL.: ERIOGONUM SPECTABILE 137 type locality flowered in 1998. The apparent low rate of recruitment may indicate an uncertain future for this species. Although Eriogonum spectabile occurs relatively close to a wilderness area and national park, it is on national forest land without special designation. No human disturbance was observed at the three sites where the plant was found, but adjacent areas have been logged extensively. Firewood cutting is also common and numerous skid trails and wood- cutter roads criss-cross the general area. Besides potential human impacts, the low numbers and lim- ited distribution of E. spectabile suggest it is at risk of extinction from natural habitat changes. One change may be an increase in competing vegetation (particularly Arctostaphylos nevadensis, but also other shrubs and the coniferous overstory), perhaps due to climate change and/or a change in the fire regime. Wildfire effects are not certain; given the plant's shrubby, presumably non-sprouting, long- lived growth form, it is likely that fire would kill existing plants. The species would then depend on seedling recruitment from the soil seed bank (of which we have no information) or from seeding in from adjacent areas, which is highly unlikely given its rarity and lack of obvious seed dispersal mech- anism(s). Another potential effect is browsing of flowering stems. On a 30 Sep 1998 visit to the second oc- currence, nearly all flowering stems and some of the tips of the leafy shoots had been browsed, pre- sumably by deer. The result is a virtual lack of in- tact seeds produced from this occurrence in 1998. Acknowledgments Special thanks to Vernon H. Oswald and Pete Figura for their help in recognizing the distinctiveness of this plant. Thanks to P. M. Eckel for the Latin translation. Thanks to Shannon Workman for the illustration. Literature Cited GiLLETT, G. W., J. T. Howell, and H. Leschke. 1961. A flora of Lassen Volcanic Park. Wasmann Journal of Biology 19(1):1-185. Lydon, p. a., T E. Gay, Jr., and S. W. Jennings. 1960. Geologic Map of California, Olaf P. Jenkins Edition, Westwood Sheet. Third printing, 1976. [Map at a scale of 1:250,000.] Oswald, V. H., D. W. Showers, & M. A. Showers. 1995. A Flora of Lassen Volcanic National Park, California. California Native Plant Society, Sacramento, Califor- nia. [Revision of the original flora by Gillett et al 1961.] Ranz, S. E. 1969. Mean annual precipitation in the Cali- fornia region. U.S. Geologic Survey, Water Resources Division, Menlo Park, California. Reprinted in 1972. [Set of two maps at approximate scale 1:990,000.] Madrono, Vol. 47, No. 2, pp. 138-145, 2000 NOTEWORTHY COLLECTIONS Arizona Hexalectris revoluta Correll (ORCHIDACEAE). — Pima County, Baboquivari Canyon, and McCleary Can- yon; Santa Cruz County, Sawmill Canyon. Between 1371 and 1524 meters elevation in canyon bottoms and sides of canyons, under oaks and mesquite, in association with Arizona walnut. Previous knowledge. Previously known range was lim- ited to portions of northern Mexico, and the Big Bend area of Texas. Voucher specimens of H. revoluta deposited at the University of Arizona Herbarium (ARIZ), Tucson, AZ collected in Baboquivari Canyon by Toolin in 1981 and McCleary Canyon by McLaughlin in 1986 were orig- inally identified as H. spicata. Studies of fresh material in the field by the author indicated the plants are correctly H. revoluta. Significance. First record of this species in Arizona, and represents a western range extension of approximately 290 miles (483 km) and a northern range extension of approx- imately 210 miles (350 km) from Big Bend National Park. Not known from New Mexico. Hexalectris revoluta is not currently a candidate for Federal Endangered Species sta- tus, but should be considered for listing due to rareness across its range. The McCleary Canyon location was re- cently included within the boundaries of land being con- sidered for trade from the Forest Service to a mining de- veloper. That trade is not currently under consideration. — Ronald A. Coleman, University of Arizona, 11520 E. Calle Del Valle, Tucson, AZ 85749. California EscHSCHOLZiA RHOMBIPETALA E. Greene (PAPAVERA- CEAE). — Alameda County: Lawrence Livermore Nation- al Laboratory, Site 300, T3S R4E, SW Va Sec. 29, elev. 850 ft, on N-facing crumbling clay bank, with Poa secun- da J. S. Presl, Bromus madritensis L. subsp. rubens (L.) Husnot, Avena harhata Link, Stylomecon heterophylla (Benth.) G. C. Taylor, Microseris douglasii (DC.) Schultz- Bip., Blepharizonia plumosa (Kellogg) E. Greene, 06 May 1997, E. Preston 1028 (DAV). Previous knowledge. Historically known from the in- terior foothills of the Hamilton and Diablo Ranges, with disjunct occurrences on the Carrizo Plains (W. Ernst, Ma- drofio 17:281-294, 1964). Believed extinct (M. Skinner and B. Pavlik, Inventory of Rare and Endangered Vascular Plants of California, 1994) until rediscovered at Carrizo Plains in 1993 by David Keil and in 1995 by Curtis Clark (C. Clark, The genus Eschscholzia: California poppies and their relatives, http://www.intranet.csupomona.edu/ -jcclark/poppy/, 2000). Significance. First east Bay Area record since 1949. Site 300 is near Corral Hollow, where the species was last collected by Peter Raven. Subsequent attempts by Raven and Clark to relocate the Corral Hollow occurrences were unsuccessful (California Natural Diversity Database, Rar- efind 2, Version 2.1.2, March 24, 2000 update; C. Clark, personal communication). Because the plants are small, they may be easy to overlook, and the plants may only appear in favorable years (C. Clark, personal communi- cation). Trichocoronis wrightii (A. Gray) A. Gray ( ASTER A- CEAE). — Merced County: Merced National Wildlife Ref- uge, S of Mariposa Bypass, T9S, R12E, SW Va Sec. 3, elev. 100 ft, 21 May 1997, R. E. Preston 1031 (DAV. CAS). Previous knowledge. Native to Mexico, Texas. In Cal- ifornia, known from four occurrences in Riverside County and four scattered locations in the Central Valley. Cali- fornia populations are presumed to be introductions (A. M. Powell in J. C. Hickman [ed.]. The Jepson Manual: Higher Plants of California, 1993), although Skinner & Pavlik (1994) suggest that the species may be native to California. Previously thought to be extirpated in the Cen- tral Valley (Skinner and PavHk 1994). Significance. First Central Valley record since 1953. Found growing in the bypass floodplain, with Eleocharis macrostachya Britton, Xanthium strumarium L., Malvella leprosa (Ortega) Krapov., Phyla nodifiora (L.) E. Greene, Polygonum arenastrum Boreau, and Frankenia salina (Molina) 1. M. Johnston. Senecio aphanactis E. Greene (ASTERACEAE). — Al- ameda County: Corral Hollow, 0.5 mi NW of Tesla town site, T3S R3E, SE Va of NE Va S26, elev. 1500 ft, scattered on barrens, with Plantago erecta E. Morris, Bromus mad- ritensis L. subsp. rubens (L.) Husnot, Erodium cicutarium (L.) LHes, Hypochaeris glabra L., Erodium botrys (Cav.) Bertol., Medicago polymorpha, Avena fatua L., 21 April 1998, Robert E. Preston 1097 (DAV, CAS). Previous knowledge. Widely but infrequently collected in the California Coast Ranges south of San Francisco Bay; the Transverse Ranges; southwest California, includ- ing Santa Cruz Island; and Baja California. This species is included in list IB of the CNPS Inventory (Skinner and Pavlik 1994). Significance. First record for Alameda County and first San Francisco Bay Area collection since 1933. In the Jep- son Manual, this species is reported to occur on drying alkali flats (T. M. Barkley in Hickman, 1993). However, information obtained from herbarium specimens indicates that it occurs on various substrates: clay; coarse sand; rock outcrops, including serpentinite; and soils with high gyp- sum content or high alkalinity. Common to all occurrences is a conspicuous absence of vegetative cover. — Robert E. Preston, Jones & Stokes Associates, 2600 V Street, Suite 100, Sacramento, CA 95616. California Boehmeria cylindrica (L.) Sw. (URTICACEAE). — Sacramento Co., widely scattered colonies in riparian zone along both sides of Georgiana Slough separating Andrus and Tyler Islands, from approx. 2 mi SW of Georgiana slough divergence from Sacramento River to approx. 1 1 mi along the slough to near The Oxbow; less common in Snodgrass Slough immediately upstream off the Sacra- mento Riven Elev. ~ sea level. G. F. Hrusa 14879, J. A. Hart, 11 Oct. 1998, 2 mi SW divergence of Georgiana 2000] NOTEWORTHY COLLECTIONS 139 Slough from Sacramento River on W side. Rhizomatous colony in opening at waters edge at or slightly below high tide level, also in shade beneath adjacent Alnus rhombi- folia Nutt. 38°07'50.4"N; 121°34'55.6"W (CDA and to be distributed.); G. F. Hrusa 15277, J. A. Hart, M. J. Hooper, 24 Nov. 1999. Snodgrass Slough at Delta Meadows State Park, on exposed and partially submerged logs. Elev. ~ sea level. (CDA). Previous knowledge. Native throughout the region east of the Rocky Mountains occurring in bogs, marshes and other wet places. Collected in Arizona near the turn of the century but only recently redocumented there (J. Bouf- ford, Ariz-Nevada Acad. Sci. 26:42-43, 1992). Significance. First records for California. In addition to the collected sites listed above, the species has been ob- served near the town of Rio Vista on the Sacramento Riv- er. It is assumed here that Boehmeria cylindrica is intro- duced to California, based primarily on its general occu- pation of disturbed and rip-rapped riverbanks. Moreover, the heavy boat traffic throughout the Delta region would appear to provide ample opportunity for introduction and spread of this species; however, it has also been found growing in less disturbed conditions and it may be a pre- viously overlooked native opportunistically occupying disturbed situations. — G. Frederic Hrusa, Herbarium CDA, California Dept. of Food & Agriculture, Plant Pest Diagnostics Cen- ter, 3294 Meadowview Rd., Sacramento, CA 95832; Jef- frey A. Hart, H. A. R. T. Inc. (Habitat Assessment & Restoration Team) 13737 Grand Island Rd., Walnut Grove, CA 95690. California Eryngium constancei Y. Sheikh (APIACEAE). — Son- oma Co., dense populations in two seasonal pools, one draining into the other, on summit of Diamond Mtn. 4 km SSW of the town of Calistoga. Elev. 685 meters, 38°32'30"N; 122°35'00"W. G. F. Hrusa 13582a (lower pool center), 13582b (lower pool periphery), and 13582c (upper pool), A. Buckmann. Oct. 05, 1996. Verified by L. Constance, November 1997. (CDA, UC/JEPS and to be distributed). Previous knowledge. Previously known only from two sites, one the type locality, at and near Loch Lomond in Lake Co., approx. 35 km NNE. Described in 1983 (L. Sheikh, Madrono 30:93-101, 1983). Significance. First record for Sonoma Co. Two sites are known near Loch Lomond in Lake County; the type lo- cality immediately N of the Loch Lomond townsite and a second approximately 3 km to the east. The habitat on Diamond Mtn. appears similar to that at the type locality, but is dominated by Que reus garryana Hook., Q. lobata Nee and Pseudotsuga menziesii (Mirbel) Franco rather than the Pinus ponderosa Laws., Quercus kelloggii Newb. mix at Loch Lomond. The Diamond Mtn. pools have been variously disturbed, and apparently a permanent spring which in past times fed both pools was closed some de- cades before the current owners took up residence. How this affected the local hydrology and flora is unknown. The plants at Diamond Mtn. do not match exactly the form at the type locality, the divergence most noticeable in the larger number of flowers per capitulum and the variable habit, ranging from slender and upright on the pool margins, to prostrate and stout in the deepest center of the lower pool. These and other similar populations in the Sonoma-Lake County region are currently under study to assess their relationship to both E. constancei and E. aristulatum Jepson. The type locality is currently listed in Title 14 of the Fish & Game Code as the Loch Lomond Vernal Pool Eco- logical Reserve. At present such safety cannot be claimed for the Diamond Mtn. locality as the pools are in an area under active viticulture development and forest harvest. However, the current landowners are aware of the pools' botanical importance in addition to DFG and California Department of Forestry regulations concerning their pro- tection. DFG plans to pursue some form of permanent protection. — G. Frederic Hrusa, Herbarium CDA, California Dept. of Food & Agriculture, Plant Pest Diagnostics Cen- ter, 3294 Meadowview Rd., Sacramento, CA 95832. Al- lan Buckmann, California Dept. of Fish & Game, Region 3. RO. Box 47, YountviUe, CA 94599. California Ononis alopecuroides L. (FABACEAE). — San Luis Obispo Co., Temettate Rd. approx. 1 km by road NW of intersection with Suey Creek Rd., in NW corner of S6, TUN; R33W, SB meridian. 35°03'58.2"N; 120°23'59.9"W. Elev. 380 m. Occupying approximately V2 acre in open grazed woodland-savanna among Quercus agrifi)lia Nee, Pinus sabiniana Douglas, extending east- ward down a dry arroyo to edge of riparian zone. Perez & Parks s.n., 9 July 1998 (CDA), Hrusa 14732 a-g, 21 July 1998 (CDA and to be distributed). Hrusa 14732c, 14732d, and 14732h confirmed by R. B. Ivimey Cook (EXR) & S. Jury (RNG). Previous knowledge. Native to southwest Europe, North Africa. Adventive in central Europe. Significance. First record for North America. The prop- agule source(s) is unknown. First noticed by the landown- er 2 years previous, the population apparently expanded rapidly into a dense but currently still more or less local- ized colony. In 1998 it was found spreading downslope along a dry drainage leading to Suey Creek and so may be expected more widely, at least locally, in the near fu- ture. The plants are unpalatable to the horses and burros that graze through the local area and because the species may form a dense stand capable of excluding more pal- atable vegetation it is currently the target of active erad- ication efforts by the San Luis Obispo County Agricultural Commissioners office. As of July 2000, an intensive out- reach program to local residents by the Ag. Commission- ers office has not revealed additional populations; how- ever, the current one remains active. — G. Frederic Hrusa, Herbarium CDA, California Dept. of Food & Agriculture, Plant Pest Diagnostics Cen- ter, 3294 Meadowview Rd., Sacramento, CA 95832. California Ambrosia pumila (Nutt.) A. Gray ( ASTER ACE AE).— Riverside Co., Riverside, moist area along Arlington Av- enue, La Sierra Heights, 26 Aug 1940, Ruth Cooper s.n. (Riverside Community College Herbarium); Nichols Road wetlands area, northwest of Nichols Road and west of 140 MADRONO [Vol. 47 Alberhill Creek, southwest of the 1-15 Freeway, Lake El- sinore 7.5' Quad., T5S R5W NW/4 S25, alt. 384 m, ca. 1300 plants in Arbuckle loam soil, annual grassland with Vulpia myuros (L.) C. Gmelin, Isocoma menziesii (Hook. & Am.) G. Nesom, Nassella pulchra, Bromus rubens, Er- odium botrys (Cav.) Bertol., Hirschfeldia incana (L.) Lagr.-Fossat, Hemizonia paniculata A. Gray, and Avena fatua L., 18 Jul 1997, D. B ramie t 2575 (UCR); Warm Springs Valley, Nichols road, 0.3 km by road W junction 1-15 near head of Walker Canyon, ca. 2.8 km NW Lake Elsinore, alt. 383 m, 29 Jun 1997, Fred M. Roberts Jr. 5043 (RSA); Warm Springs Valley, NW of Lake Elsinore, about Alberhill Creek at the head of Walker Canyon near Durant Siding, SE base of Alberhill Mountain, along Nichols Road, 0.2 miles SW of junction with Collier Av- enue, near 33°42'N, 1 17°21'W, T5N R5W NW Va S25, alt. 381 m, 22 Jul 1997, Steve Boyd 10017 (RSA). Previous knowledge. Reported only from San Diego County by The Jepson Manual (J. C. Hickman, ed., 1993, U.C. Press) and from "sw San Diego Co." by Munz (A California Flora, 1959, U.C. Press) but a locality at Skunk Hollow, Riverside County is also known (Madrono, 1992, 39(2): 157). This species is considered rare and endangered and is reported to be declining by the CNPS Inventory of Rare and Endangered Vascular Plants of California (Skin- ner and Pavlik, 1994, CNPS Inventory of Rare and En- dangered Vascular Plants of California, 5th Ed.). Significance. These specimens provide second and third localities for Riverside County and range extensions of 27 km WNW and 53 km NW from the previously reported site at Skunk Hollow. The La Sierra Heights plants, only now being reported 60 years after their collection, were from the wet alkaline areas that formerly existed from near the intersection of Arlington Ave. and Van Buren Blvd., and the Riverside Airport, to the vicinity of California Ave. at Jackson Street. Unfortunately, there is little chance the plants persist there as the area is now largely paved and urbanized, though a few small pockets of marginal habitat do remain. The record at RCC was represented by three replicate specimens! It is unfortunate that two of these were not distributed to larger herbaria where they would have more rapidly come to the attention of the bo- tanical community. Perhaps something could have been done to salvage part of this population if the location had been generally known. The Nichols Road population is under considerable threat due to its location in a rapidly urbanizing area. These discoveries highlight both the need to look for this species in other moist alkaline places in western Riverside County and the importance of checking small and inactive herbaria for significant records of other rare plants. DicoRiA CANESCENS A. Gray (ASTERACEAE). — San Bernardino Co., Rialto [actually Colton, near the city boundary], degraded sand dunes on Slover Ave. just E of the tank farms and E of Riverside Ave., S of I- 10 Freeway, very common, 14 Oct. 1993, Chet McGaugh s.n. (UCR); same location, 34°04'N, 117°22'W, alt. 320 m, ca. 200 plants seen in loose blowsand at summit of dune, in and near a disturbed OHV area, very local at this site and at another ca. 700 m further west, 14 Sept. 1999, A. C. Sand- ers, S. Boyd and M. Provance 23073 (RSA, UCR). Previous knowledge. A native plant of desert dunes, previously unrecorded on the coastal slope of California, except as a waif along the railroad near Elysian Park in Los Angeles (A. Davidson s.n. in 1892 and 1893, POM; H. M. Hall, Compositae of Southern California, Univ. of California Publ. Botany, Vol. 3, 1907). Significance. First records of the species in natural hab- itat west of the deserts and a range extension of 55 km WNW from the Cabazon area. The species is well estab- ^ lished at this site, which is a remnant of the formerly extensive Colton Dunes. We cannot be certain that the ! species was not introduced, but it appears native and is , present in an arid interior valley in the best available nat- ural habitat on the coastal slope. It is surprising that this i species has so long escaped detection if it is native at this \ site, but other typically desert species are also found near- | by, including Camissonia campestris (E. Greene) Raven, Encelia farinosa Torrey & A. Gray, Eriophyllum wallacei (A. Gray) A. Gray, Malacothrix glabrata A. Gray, and Prosopis glandulosa Torrey. While this area has long had an active botanical community, the total number of col- lectors has never been great, and it's probable that the dunes have just not been thoroughly surveyed in the fall, when few other plants are flowering. Erodium malacoides (L.) Willd. (GERANIACEAE).— San Bernardino Co., Blue Mtn., Grand Terrace, 34°01'14"N, 117°18'14"W, alt. 396 m, abundant on the pediment near a water tank on the W side of the mountain, associated with Heterotheca grandifiora Nutt., 4 Apr 1998, M. C. Provance 350 (UCR). | Previous knowledge. Weed introduced from Europe. Previously reported in California only from the northern San Joaquin Valley and San Francisco Bay areas. Significance. First record for San Bernardino County and southern California and a range extension of over 600 km from the San Francisco Bay region. This weed should be sought in other areas in southern California and its status and spread carefully monitored. KoELERiA phleoides (ViUars) Pers. (POACEAE). — Riv- erside Co., Jurupa Mtns., Sunnyslope near Rattlesnake Mtn., Armstrong St., 0.1 mi. south of San Bernardino Co. line, 34°01'59"N, 117°24'53"W, T2S R5W NW/4 S4, alt. 1050 ft, in hard dry soil in a disturbed field, 7 Mar 1998, M. C. Provance 174 (ARIZ, UCR). Det. by J. & C. Reed- er, 1999. Previous knowledge. Uncommon introduction, reported from scattered locations from Santa Barbara and Kern Counties through northern California. Significance. First record for Riverside County and a range extension of 250 km southeast from Santa Barbara County. MoNARDELLA PRiNGLEi A. Gray (LAMIACEAE). — San Bernardino Co., sand hills west of Colton, 17 May 1941, J. C. Roos 2472 (La Sierra College Herbarium). Previous knowledge. This species, a very local endemic of the Colton Dunes, has generally been thought extinct since 1921 (Skinner and Pavlik, 1994). Significance. This collection extends the known chro- nological range of this species by 20 years, but unfortu- nately we still have no evidence that the species has per- sisted until today. Like the Ambrosia record above, this record emphasizes the need to examine all herbaria for informative collections. This population was reported on the label to have been "mutilated by grasshoppers", though the specimen preserved was not too badly dam- aged. The recent discovery of other noteworthy species on the remnants of the Colton Dunes offers hope that this species may yet be rediscovered. Nama stenocarpum Gray. (HYDROPHYLLACEAE).— Orange Co., San Joaquin Hills, Emerald Canyon, 3 km up canyon from Pacific Coast Highway, 2.75 km SW of in- tersection of Laguna Canyon Rd. and El Toro Rd., Laguna 2000] NOTEWORTHY COLLECTIONS 141 Beach, 33°34'25"N, 1 17°47'10"W, a few plants on sandbar ''in steep incised channel with Typha sp., Chenopodium ambrosioides L., Juncus xiphoides E. Meyer and Bac- charis salicifoUa (Ruiz Lopez & Pavon) Pers., alt. 122 m, 21 July 1998, A. L. Wolf 402 (UCR); Laguna Lakes (northernmost lake), W of Laguna Canyon Rd., Laguna Beach, 33°36'50"N, 117°45'30"W, alt. 118 m, 30+ plants on south edge of lake on drying margin with Petunia par- viflora A. L. Juss., Rorippa ciirvisiliqua (Hook.) Britten and Lythrum californicum Torrey & A. Gray, 26 June 1998, A. L. Wolf 358 (UCR); Lambert Reservoir, N of El Toro Marine Corps Air Station, 33°41 '32"N, 1 17°42'40"W; alt. 134 m, few plants on mudflat on S edge of reservoir with Ammania rohusta Heer & Regel, Lythrum hyssopi- folium L., and Juncus hufonius L., 15 May 1998, A. L. Wolf 276 (UCR); Peters Canyon Channel, E side of chan- nel between Alton Pkwy and Barranca Pkwy, Tustin, 33°41'30"N, 117°49'17"; alt. 15 m, 2 plants in sediment basin adjacent to channel in the southwest portion of re- cently bladed field, 1 July 1998, A. L. Wolf 414 (UCR); Riverside Co., Mystic (San Jacinto) Lake, 1.4 miles SE of Jackrabbit Trail on Oilman Springs Rd., 1.4 mi S of Glen Eden Hot Springs, 33°52'31'N, 117°03'W, alt. 433 m, common on receding lakeshore with numerous herbs, 26 Sep 1999, A. C. Sanders & M. Provance 23120 (UCR, and to be distributed); same location, 15 Oct 1999, A. C. Sanders, D. Bramlet, M. Costea & T. Salvato 23173 (UCR, and to be distributed). Previous knowledge. A rare plant of seasonally moist areas from scattered counties, mainly coastal in southern California and extending south into Mexico. This species is considered rare in California but more or less common elsewhere (Skinner and Pavlik 1994). Not recorded on the mainland of southern California since 1939, nor anywhere in California since then, except for two collections on San Clemente Island (California Natural Diversity Database), the last in 1991. Significance. First records for Riverside County and first records for California, except for San Clemente Is- land, in over 60 years. This rare plant is still a member of the mainland California flora. QuERCUS PALMER! Eugclm. (FAGACEAE). — Riverside Co., Jurupa Mountains, Rattlesnake Mt. near Crestmore Heights, ca. 4.2 km NW of downtown Riverside, 34°01.59'N, 117°23.81'W, T2S R5W S3, alt. 400 m, ca. 50 individual shrubs forming a dense and nearly homog- enous vegetation over an area of approximately 25 m X 8 m, growing in a nook on a rocky north-facing slope, associated with Prunus illicifolia (Nutt.) Walp., Ribes in- decorum Eastw. and Phacelia ramosissima Lehm., 14 Apr 1998, Mitchell C. Provance 441 (UCR); Jurupa Moun- tains, Rattlesnake Mt. (hill 1452) above Crestmore, NW of Riverside, Fontana 7.5' Quad., 34°02'N, 117°23.5'W, T2S R5W center of W/2 S3, alt. 365-400 m, dense colony in a notch on the ridge and in adjacent rocky gully be- tween outcrops on the N-facing granitic slope, coastal sage scrub with chaparral elements, Prunus illicifolia, Ri- bes indecorum, Eriogonum fasciculatum (Benth.) Torrey & A. Gray, Rhamnus crocea Nutt., Mimulus aurantiacus Curtis, Rhus trilobata Torrey & A. Gray, Salvia mellifera E. Greene, Toxicodendron diversilobum (Torrey & A. Gray) E. Greene, etc., 14 May 1998, A. C. Sanders and Mitch Provance 21848 (CAS, DAV, RSA, SD, UC, UCR). Previous knowledge. Occurrences of this species are patchy from Colusa Co., California south to Baja Califor- nia, Mexico and to the east in Arizona. In southern Cali- fornia, the species is most common in the Peninsular Range, San Jacinto Mountains and south, but also occurs on the desert slopes of the San Gabriel and San Bernar- dino Mountains. There is a record of a single plant in the Little San Bernardino Mountains. This species has been considered for inclusion in the CNPS Inventory of Rare and Endangered Vascular Plants of California. Significance. First known record of this species for the South Coast subregion (Hickman 1993) and a range ex- tension of approximately 38 km SSW from the nearest known populations on the north side of the San Bernar- dino Mtns., near Mojave River Forks. This population is well separated from the other known populations, and oc- curs far below the lowest elevation previously known for the species in the region. The rate of sexual reproduction in this population appears to be extremely low, although some of the plants do appear to be making a few healthy acorns. It is amazing that this conspicuous species has escaped detection on the outskirts of a large city that has had an active botanical community for over 100 years. For example, some early botanists and plant collectors who were Riverside residents, prior to the establishment of UC Riverside, include: Charlotte M. Wilder (whose house was in the Jurupa Mtns.), Fred M. Reed, Harvey Monroe Hall, David D. Keck, and Edmund C. Jaeger. In addition, Samuel B. Parish's residence in San Bernardino was less than 15 km NE of the Quercus palmeri site. This discovery points up the need to continue searching for new species and important range extensions even in "well known" areas: an unexplored ridge can hide something new or interesting, and there are many unexplored ridges. — Mitchell C. Provance and Andrew C. Sanders, Herbarium, Dept. of Botany & Plant Sciences, University of California, Riverside, CA 92521-0124, Valerie Soza and Steve Boyd, Herbarium, Rancho Santa Ana Botanic Garden, 1500 N. College Avenue, Claremont, CA 91711, David Bramlet, 1691 Mesa Dr. A-2, Santa Ana, CA 92707 and Adrian L. Wolf, Harmsworth Associates, 36 Bluebird Lane, Aliso Viejo, CA 92656. California All the following collections are from the San Bernar- dino Mountains. Brickellia knappiana Drew ( ASTER ACEAE). — San Bernardino Co., northeast side of Blackhawk Mountain, east of Blackhawk Canyon, 34°21'06"N 116°47'24"W, elev. 1372 m, shrub 1.8 m tall in alluvium of minor can- yon, northeast-facing wash, 10 Aug 1998, Valerie Soza et al. 409 (RSA, and to be distributed). Previous knowledge. This probable hybrid (treated as a species in all available manuals) between B. desertorum Cov. and B. multiflora Kellogg is known from only a few locations in the northern and eastern mountain ranges of the Mojave Desert, almost always at sites where both par- ents are present. Previously collected from the Argus, Fu- neral, Panamint, and Kingston Mountains, but with the type locality at an undefined site along the Mojave River. The type locality is given simply as "in the neighborhood of the Mohave River" (Pittonia, 1888, 1:260). The vague- ness of the collection data inspires little confidence that the site was particularly near the Mojave River, especially since the plant has not, in the 110 years since, been re- corded there. Significance. First record for the San Bernardino Moun- 142 MADRONO [Vol. 47 tains, which extends the range of this taxon 90 km south- ward from the closest known locality in the Argus Range. Brickellia multiflora Kellogg ( ASTER ACE AE). — San Bernardino Co., canyon running from west to east along the northern edge of Horsethief Flat, 34° 1 9.487 'N 116°45.84rw, elev. 1372 m, rocky drainage area near mouth of canyon, seasonally moist, subjected to carbonate scree slides from along canyon walls, 29 Jun 1998, Mitch- ell C. Provance & Valerie Soza 792 (UCR). Previous knowledge. Uncommon shrub occurring from the northern and eastern Mojave Desert of California to Nevada, particularly in the mountains of Inyo County, the White Mountains of Mono County, at Little Lake in Kern County, and the Kingston, Clark and Granite Mountains of San Bernardino County. Significance. First record for the San Bernardino Moun- tains, which extends the range of this species about 100 km southwestward from the nearest known occurrence in the Granite Mtns. near Kelso. It is noteworthy that this species turned up at about the same time that its hybrid progeny, B. knappiana Drew, was also discovered in the range. It is interesting that several species traditionally known from the northern and eastern Mojave Desert (e.g., Baileya multiradiata A. Gray, Madrono 43 (4):524, and the plants reported here) have recently been found on the northern slopes of the San Bernardino Mountains, as that area has begun to be explored away from the major routes of travel. Camissonia pterosperma (S. Watson) Raven (ONA- GRACEAE). — San Bernardino Co., west of Horsethief Flat, above road 3N03A, 34°19'06"N 116°47'03"W, elev. 1768 m, scarce annual in open WNW-facing slope, 16 Jul 1998, Valerie Soza & Tasha LaDoux 390 (RSA); San Ber- nardino Mountains, northwest of Tip Top Mountain, east of Arrastre Creek, 34°15'59"N 116°43'45"W, elev. 1920 m, rare annual on open rocky north-facing lower slope, 30 Jul 1998, Valerie Soza & Tasha LaDoux 404 (RSA). Previous knowledge. Rare annual in northern mountain ranges of the Mojave Desert, e.g., Panamint and Clark mountains, and Inyo and White mountains east to Last Chance Range and Fish Lake Valley, to Utah and Oregon. Significance. First record for the San Bernardino Moun- tains, range extension of about 100 km southwestward from Clark Mountain. CoRNUS glabrata Benth. (CORNACEAE). — Riverside Co., Morongo Indian Reservation, very locally common at a seep along a gully at the west end of Burro Flat, 33°59'N, 116°52'W, alt. 1160 m, 14 Nov 1997, A. C. Sanders & T. Pennant 21596 (DAV, RSA, SD, TEX, UC, UCR, UTC). Previous knowledge. Scarce in southern California and known from Riverside County only from a single collec- tion from the San Jacinto Mountains made in 1922 (R A. Munz 5806, alt. 1500 m, Hemet Valley, frequent along banks of Pipe Creek) based on specimens at UCR and RSA. Significance. First record for the San Bernardino Moun- tains, second record from Riverside County and the first collection of this species from that county in 77 years. Cynanchum utahense (Engelm.) Woodson (ASCLE- PIADACEAE).— San Bernardino Co., E. of Horsethief Flat, 0.5 km N of the Arrastre Creek Dam, 34°19.44'N 1 16°45.77'W, elev. 1433 m, on a steep, barren, sandy, SE- facing slope, 23 Jun 1998, Mitchell C. Provance & Val- erie Soza 744 (UCR). Previous knowledge. Uncommon perennial occurring on the Mojave Desert of California and to Utah and Ari- i zona. ! Significance. First record for the San Bernardino Moun- j tains and the Transverse Ranges and extends the range of this species slightly (10 km southwest) from the nearest ! known occurrence of Old Woman Spring on the southern Mojave Desert. Glyceria occidentalis (Piper) J. C. Nelson. (PO- I ACEAE). — Riverside Co., Morongo Indian Reservation, \ very locally common in mud around the sag pond at the southeast end of Burro Flat, 33°59.5'N, 116°51'W, T2S RIE SE/4 S14, alt. 1150 m, 24 Apr 1996, A. C. Sanders & S. Hawkins 18088 (RSA, UCR, and to be distributed). Pet. by Travis Columbus. Previous knowledge. Northwestern California and north to Idaho and British Columbia, the furthest south previ- ously known populations are apparently in San Mateo County (P. A. Munz, 1968, Supplement to A California Flora, University of California Press). Significance. First record for southern California and a range extension of 650 km from the San Francisco Bay area. The site where this species was collected is a shallow but permanently wet sag pond on the San Andreas Fault from which, reportedly, peat was formerly harvested. This bizarre disjunction in the distribution of a native plant suggests that the species should be sought in other wet areas in central and southern California. Travis Columbus notes that this species is possibly not distinct from the Eurasian species Glyceria declinata Brebiss. That species has also been reported from northern California (e.g., A. S. Hitchcock and A. Chase, 1935, Manual of the Grasses of the United States; P. A. Munz, 1959, A California Flora, University of California Press), but is not mentioned in the Jepson Manual. Glyceria declinata was reported in the literature for California in 1957 by Beecher Crampton from Stanislaus County, south of Oakdale, further ESE from the San Francisco Bay area (Leaflets of Western Bot- any 8 (6): 160). Nicotian A acuminata Hook var. multiflora (Philippi) Reiche (SOLANACEAE).— Riverside Co., Morongo In- dian Reservation, lower Hathaway Creek Cyn., 33°58'N, 116°52'W, elev. 838 m, disturbed roadside especially in moist areas, margins of riparian forest, 20 Aug 1998, A. C. Sanders 22232 (UCR). Previous knowledge. An introduced weed from South America, long known from northern California, but only first reported in southern California in 1996 from San Ber- nardino County (Madrono 43(2):334-336). Significance. First record for Riverside County; extends its southern California range 18 km southeast from Mill Creek Cyn., and further documents the establishment and spread of this introduced weed in southern California. — Valerie Soza, Herbarium, Rancho Santa Ana Botan- ic Garden, 1500 N. College Avenue, Claremont, CA 91711; Mitchell C. Provance and Andrew C. Sanders, Herbarium, Dept. of Botany & Plant Sciences, University of California, Riverside, CA 92521-0124; and Steve Boyd, Herbarium, Rancho Santa Ana Botanic Garden, 1500 N. College Avenue, Claremont, CA 91711. Colorado Aliciella sedifolia (Brandegee) J. M. Porter [Gilia sed- ifolia Brandegee] (POLEMONIACEAE).— Hinsdale Co., 2000] NOTEWORTHY COLLECTIONS 143 San Juan Mts., Half Peak; 19 km SW of Lake City, T42N R6W SEC25 SEV4; 4110 m. South facing slope, gravelly patches with no other vegetation, common locally. 5 Au- gust 1995, Susan Komarek 478 (COLO 457147). Previous knowledge. Evidently a rare endemic; thought extinct since original collection by Purpus in 1893 [Gun- nison County, Uncompahgre Range, Sheep Mt., 11,800', July 1893. Purpus 697 (GH)]. Significance. First collection in 103 years. AscLEPiAS INVOLUCRATA Engclmanu ex Torrey (ASCLE- PIADACEAE).— Bent Co., 18 km NE of Las Animas, T21S R51W SEC23 NW^A; 1260 m. Transitional zone be- tween shortgrass and sandsage prairie, with Buchloe dac- tyloides, Psoralidium tenuiflorum, and Oligosporus filifol- ius. 1 June 1998, Dina Clark 686 (COLO 471324). Previous knowledge. Southern Great Plains, New Mex- ico, Arizona, and Mexico; one collection from Las Ani- mas County, Colorado in 1948 [Rogers 5834, May 31, 1948. (COLO 55248)] (Great Plains Flora Association, Flora of the Great Plains 1986). Significance. First Colorado collection in 50 years. The documented distribution of this species may become better known with increased field work in the eastern part of the state. AsTROLEPis INTEGERRIMA (Hooker) Beuham & Windham (PTERIDACEAE).— Las Animas Co., Mesa de Maya, Je- sus Mesa, near Colorado - New Mexico State Line, ca. 85 km ESE of Trinidad; T35S R54W SEC 16 NEV4; 1540 m. Dry, southwest facing slope in crevice of Dakota sand- stone outcrop. 19 September 1994, Dina Clark 582 & Car- olyn Crawford (COLO 455480). Previous knowledge. Arizona, Nevada, New Mexico, Oklahoma, Texas, and Mexico (Great Plains Flora Asso- ciation, Flora of the Great Plains 1986). Significance. First Colorado record. Apparently the northernmost record of this New World genus (Flora of North America Association, Flora of North America Vol. 2 1993). BoTHRiocHLOA SPRiNGEiELDii (Gould) Parodi [Andropo- gon springfieldii Gould] (POACEAE). — Las Animas Co., Mesa de Maya, vicinity of upper Gotera Canyon, ca. 70 km ESE of Trinidad; T34S R55W SEC27 SEV4; 1700 m. South facing slope in grassy breaks. 5 July 1993, Dina Clark 204 & C. Deihl (COLO 455362); 4 August 1993, Dina Clark 216 & P Deihl (COLO 455213). Previous knowledge. West Texas to Arizona (Great Plains Flora Association, Flora of the Great Plains 1986). Significance. First Colorado record. A range extension of ca. 100 km and the northernmost record for this species (Great Plains Flora Association, Atlas of the Flora of the Great Plains 1977). Chenopodium cycloides Nelson (CHENOPODIACE- AE).— Weld Co., ca. 25 km ENE of Greeley; T6N R63W SEC34 SEV4; 1430 m. Andropogon hallii-Calamovilfa lon- gifolia grassland on eolian deposited sand. 3 September 1997, Dina Clark 634 (COLO 469818). Previous knowledge. Southwest Kansas south through west Texas, west to southern New Mexico; known in Col- orado from Las Animas and Pueblo counties (Spackman et al., Colorado Rare Plant Field Guide 1997). Significance. Northward range extension of ca. 250 km for this rare, or perhaps overlooked, species. DiPLACHNE DUBIA (Kuuth) Scribuer [Leptochloa dubia Humboldt, Bonpland, & Kunth] (POACEAE).— Las An- imas Co., Mesa de Maya, ca. 1 15 km E of Trinidad; T33S R51W SEC25 SEV4; 1450 m. Bottom of dry, rocky, S facing slope. 4 September 1993, Dina Clark 283, T. Ho- gan, & R. Brune (COLO 455445). Previous knowledge. Texas, western Oklahoma, Arizo- na, and Mexico; also south Florida and Argentina (Correll and Johnston, Manual of the Vascular Plants of Texas 1970). Significance. First Colorado record. A range extension of ca. 500 km from previously known location (Neil Snow personal communication). Eleocharis xyridiformis (Fernald) Brackett (CYPER- ACEAE). — Logan Co., Vicinity of Sterling, Peetz Table, 40°56'36"N, 103°10'19"W; 1300 m. Low swale, on clay- sand substrate. 31 August 1997, V^. A. Weber 19273 & Ron Wittmann (COLO 467546). Cheyenne Co., Three km E of Kit Carson on Hwy 40; 1250 m. Low swale in sand hills of high plains. 7 September 1997, W. A. Weber 19382, Ron Wittmann, & Dina Clark (COLO 467358). Previous knowledge. North Dakota and Montana, south to Texas and Mexico (Great Plains Flora Association, Flo- ra of the Great Plains 1986). Significance. First Colorado record. Probably over- looked previously, and to be expected more frequently with increasing field work in eastern Colorado. Festuca subulata Trin. (POACEAE). — Rio Blanco Co., North Fork of White River near North Fork Camp- ground, White River N.E, ca. 20 km NE of Buford; TIN R90W SEC13; 2400 m. Riverbank with Alnus incand (L.) Moench and Sali.x drummondiana Hook. 9 August 1992, Nan Lederer 3661 (COLO 451683); 3 September 1992, Gwen Kittel (COLO 451607). Previous knowledge. Alaska to Alberta, south to Cali- fornia, east to Idaho, Montana, Utah, and Wyoming. Re- ported from northern tiers of Utah (Welsh, A Utah Flora 1993) and Wyoming (Dorn, Vascular Plants of Wyoming 1992). Significance. First Colorado record. Approaching the southern extent of its range in North America (known from the central Sierra Nevada of California). Helenium microcephalum De Candolle (ASTERA- CEAE). — Las Animas Co., Mesa de Maya, ca. 105 km E of Trinidad; T33S R51W SEC22 NWV4; 1500 m. Muddy edge of stock pond in shortgrass prairie. 16 September 1994, Dina Clark 554 (COLO 455583). Previous knowledge. Southwest Oklahoma, Texas, southern New Mexico, to northern Mexico (Great Plains Flora Association, Flora of the Great Plains 1986). Significance. First Colorado record. A species of the southern Great Plains, this record represents a range ex- tension of nearly 300 km. Heterosperma pinnatum Cavanilles (ASTERA- CEAE). — Las Animas Co., Mesa de Maya, ca. 75 km E of Trinidad; T34S R56W SEC20 NEV4; 1700 m. Among sandstone slabs near drainage; shortgrass prairie. 17 Sep- tember 1994, Dina Clark 566 (COLO 455475). Previous knowledge. Mexican highlands, north to Ari- zona and Texas (Correll and Johnston, Manual of the Vas- cular Plants of Texas 1970). Significance. First Colorado record. A Chihuahuan spe- cies at its northernmost point in North America; an exten- sion of its previously known range from Santa Fe County, New Mexico and the Davis Mountains of Texas. Reverchonia arenaria Gray (EUPHORBIACEAE).— Bent Co., S shore of John Martin Reservoir; T23S R49W SEC 17 SW1/4 of NWy4; 1 150 m. Sandsage prairie on dune. 144 MADRONO [Vol. 47 6 June 1998, Dina Clark 683 (COLO 471323); 30 July 1998, Dina Clark 740 (COLO 471325). Previous knowledge. Oklahoma, Kansas, Texas, New Mexico, Utah, Arizona, and northern Mexico (Great Plains Flora Association, Flora of the Great Plains 1986). Significance. First Colorado record. Northernmost rec- ord for this species; a range extension of ca. 200 km. Triteleia grand/flora Lindl. [Brodiaea douglasii Wats.] (ALLIACEAE).— Montezuma Co., Boggy Draw near Peel Reservoir; T38N R14W SEC7 SWy4; 2400 m. Pine-oak vegetation; population of about 2000 plants in 8 ha area. 22 June 1998, Leslie Stewart 4 (COLO 470260, 470261). Previous knowledge. A showy species of the northwest; nearest reports from the Wasatch Range of northeast Utah (Welsh, A Utah Flora 1993) and northwestern Wyoming (Dorn, Vascular Plants of Wyoming 1992). Significance. First Colorado record. There is some spec- ulation this may have been transported by indigenous peo- ple as a food source. Specimen was collected by Charlotte Thompson, a seasonal employee of the U.S. Forest Ser- vice, in the process of clearing a timber sale. Population was reportedly protected from lumbering operation (Leslie Stewart personal communication). —Dina Clark and Tim Hogan, University of Colorado Herbarium (COLO), Campus Box 350, Boulder, CO 80309. Oregon Carex chordorrhiza Ehrh. ex L. f. (CYPERA- CEAE). — Clatsop Co., weedy cranberry fields, with C. arcta Boott, W of CuUaby Lake, elev. 4 m, T7N RlOW S22, 2 Oct 1999, P. F. Zika 14455 WTU; Coos Co., well- established weed in cranberry fields, with Vaccinium ma- crocarpon Aiton, Aster subspicatus, Lysimachia terrestris, between Spruce Hollow and Coquille River, elev. <50 m, T28S R14W S20 EV4, 20 June 1997, P. F. Zika 13217 & B. Wilson (OSC, MICH); Curry Co., weed in cranberry fields, with Agrostis exarata, 2.5 km W of Route 101, SW of Floras Lake, elev. ca. 15 m, T31S R15W S20, 21 Au- gust 1997, P. F. Zika 13365 & B. Wilson (OSC). Previous knowledge. Creeping sedge is a circumboreal sedge native in eastern and northern North America. The nearest native population is 380 km northeast in Okanog- an Co., Washington {Wooten 1334 WS). Significance. First report for Oregon. All sites are as- sociated with wetland cranberry agriculture, which began with the introduction of Vaccinium macrocarpon to Ore- gon in 1885. The cranberry vines are cut and hauled be- tween farms to establish new cranberry fields. The process also transports weed seeds from one farm to the next. Carex chordorrhiza is known as a weed in cranberry fields in Wisconsin (Eck, The American Cranberry, 1990), where it is a native and invades from adjacent wetlands (A. A. Reznicek, personal communication). Creeping sedge is not reported as a weed from cranberry farms else- where in North America (Eck 1990). Thus we suspect C. chordorrhiza arrived in Oregon during the transport of cranberry vines from Wisconsin to Oregon. There are sev- eral other eastern native wetland plants that are reported below as new weeds in our area. All probably originated as propagules attached to cranberry vines transported from the eastern United States. EscALLONiA rubra (Ruiz & Pavon) Pers. (GROSSU- f LARIACEAE). — Coos Co., naturalized weed on muddy ' shore and on disturbed ground in adjacent fields, with ! Gaultheria shallon Pursh, Myrica californica Cham., SE j end of Lost Lake, NW of McTimmonds Road, elev. 7 m, T29S R15W S36 NWV4, 20 June 1997, P. F. Zika 13207' & B. Wilson (OSC). i Previous knowledge. Native to Chile. Commonly cul- 1 tivated west of the Cascade Mts. in Oregon and Washing- j ton. j Significance. First report for Oregon as a garden escape. I Naturalized in a wild setting in the New River Area of Critical Environmental Concern, Coos Bay District of the Bureau of Land Management. \ Fuchsia magellanica Lam. (ONAGRACEAE). — Coos Co., steep weedy roadside bank, with Rubus armeniacus, Lathyrus latifolius L., Route 101, Coos Bay, near North Bend town line, elev. ca. 10 m, T25S R13W S23, 18 June 1997, P. F. Zika 13139 & B. Wilson (OSC); Curry Co., weed in hedge rows, with Rubus spectabilis Pursh, R. ar- meniacus, R. laciniatus Willd., Pseudotsuga menziesii (Mirbel) Franco, near waste treatment plant. Wharf Street, , Brookings, elev. ca. 20 m, T41S R13W S6, 19 May 1997, P. F. Zika 13083 (WTU); roadside weed. Route 635 at 0.3 km NW of Yorke Creek, elev. 24 m, T37S R14W S18 ' NWV4 of SEy4; 19 June 1997, P. F. Zika 13163, B. Wilson, & V. Stansell (OSC). Previous knowledge. Native to Chile and Argentina. Common ornamental in Oregon and Washington, occa- sionally fruiting west of the Cascade Mtns. Significance. First report in Oregon as a garden escape. Most likely dispersed by fruit-eating birds. Hypericum boreale (Britton) E. Bickn. (CLUSI- ACEAE). — Clatsop Co., weedy cranberry fields, W of Cullaby Lake, elev. 4 m, T7N RlOW S22, 2 Oct 1999, P. F. Zika 14449 OSC, WTU; Coos Co., common weed in moist sand, in and near cultivated cranberry fields, with Vaccinium macrocarpon Aiton, Lysimachia terrestris, Juncus ejfusus. Lower Fourmile Road, 1.5 km W of Route 101, elev. ca. 10 m, T30S R15W SI NWy4 of NWV4, 20 August 1997, P. F. Zika 13338, B. Rittenhouse, B. New- house, & B. Wilson (OSC); sunny depression between cranberry fields, and nearby ditches, with Sisyrinchium californicum (Ker Gawler) Dryander, Lotus corniculatus L., Rubus ursinus Cham. & Schldl., Panicum occidentale, Croft Lake Road, 1.5 km W of Route 101, elev. ca. 15 m; T30S R15W Sll; 20 August 1997, P. F. Zika 13340, 13346 (OSC); Curry Co., cranberry fields and ditch mar- gins, well established weed, with Juncus bufimius L., 2.5 km W of Route 101, SW of Floras Lake, elev. ca. 15 m, T31S R15W S20, 21 August 1997, P. F. Zika 13367 & B. Wilson (OSC). Previous knowledge. Native to eastern North America, west to Illinois. Significance. First record for Oregon. At present it ap- pears to be restricted to cultivated cranberry fields, and adjacent moist, disturbed ground. Juncus brevicaudatus (Engelmann) Fernald (JUNCA- CEAE). — Clatsop Co., weedy cranberry fields, W of Cul- laby Lake, elev. 4 m, T7N RlOW S22, 2 Oct 1999, P. F. Zika 14454 OSC, WTU; Curry Co., sandy banks near cranberry fields, with J. bufonius L., /. canadensis, J. planifiAius, 2.5 km W of Route 101, SW of Floras Lake, elev. ca. 15 m, T31S R15W S20, 21 August 1997, P. F. Zika 13366 & B. Wilson (OSC). 2000] NOTEWORTHY COLLECTIONS 145 Previous knowledge. Native to eastern North America, west to Minnesota. Significance. First Oregon record. Another weed asso- ciated with cranberry agriculture. J UNCUS CANADENSIS J. Gay ex Laharpe (JUNCA- CEAE). — Clatsop Co., weedy cranberry fields, W of Cul- laby Lake, elev. 4 m, T7N RlOW S22, 2 Oct 1999, P. F. Zika 14452 OSC, WTU; Coos Co., mucky shoreline, with Carex aquatilis Wahlenb. var. aquatilis, Dulichium, Lysi- machia terrestris, Juncus supiniformis Engelm., Scirpus subterminalis Torrey, Muddy Lake, elev. 5 m, T30S, R15W Sll SWV4, 20 August 1997, P. F. Zika 13353 & B. Rittenhouse (OSC); shallow water emergent, with Po- tentilla palustris (L.) Scop., Gentiana sceptrum Griseb., Hypericum anagalloides Cham. & Schldl., Vaccinium ma- crocarpon Alton, E shore of unnamed pond, W of Lost Lake, elev. 7 m, T29S R15W S36 NWy4 of NWy4, 21 August 1997, B. Newhouse 97055 (OSC); Curry Co., well established weed in ditches and cranberry fields, with J. effusus L., J. falcatus E. Meyer, Scirpus setaceus L., 2.5 km W of Route 101, SW of Floras Lake, elev. ca. 15 m, T31S R15W S20, 21 August 1997, P. F. Zika 13360 & B. Wilson (OSC). Previous knowledge. Native to eastern North America, west to Minnesota. Reported as a weed in cranberry fields in New England (Sears et al.. An Illustrated Guide to the Weeds of Cranberry Bogs in Southeastern New England, 1996). Significance. First report for Oregon. Another weed as- sociated with the cranberry industry. Our collections from undisturbed wetlands indicate this species is successfully invading native plant communities in southwestern Ore- gon. Juncus pelocarpus E. Meyer (JUNCACEAE). — Coos Co., sandy soil in swamps and cranberry bogs, Bandon; 2 Sept 1958; G. Scott s.n. (OSC); sandy shoreline of ir- rigation pond, with Lysimachia terrestris, Salix sitchensis Bong., Viola lanceolata, near cranberry fields. Croft Lake Road, 1.5 km W of Route 101, elev. ca. 15 m; T30S R15W Sll; 20 August 1997, P. F. Zika 13339 (OSC). Previous knowledge. Native to eastern North America, west to Minnesota. Noted as a weed in cranberry fields in New England (Sears et al. 1996). Significance. First report for Oregon. Another weed as- sociated with the water systems for cranberry farming. Spiraea tomentosa L. (ROSACEAE). — Coos Co., weed in cranberry fields, with Vaccinium macrocarpon Alton, Lysimachia terrestris. Aster subspicatus, between Spruce Hollow and Coquille River, elev. <50 m, T28S R14W S20 EV4, 20 June 1997, P. F. Zika 13213 & B. Wilson (OSC). Previous knowledge. Native to eastern North America, west to Minnesota. A weed in cranberry fields in New England (Sears et al. 1996). Significance. First record for Oregon. Another weed as- sociated with the cranberry industry. — Peter F. Zika, Keli Kuykendall and Barbara Wil- son, Herbarium, Department of Botany and Plant Pathol- ogy, Oregon State University, Corvallis, OR 97331. Volume 47, Number 2, pages 71-146, published 15 March 2001 Subscriptions — Membership Membership in the Cahfomia Botanical Society is open to individuals ($27 per year; family $30 per year; emeritus $ 1 7 per year; students $ 1 7 per year for a maximum of 7 years). Late fees may be assessed. Members of the Society receive Madrono free. Institutional subscriptions to Madrono are available ($60). Membership is based on a calendar year only. Life memberships are $540. Applications for membership (including dues), orders for sub- scriptions, and renewal payments should be sent to the Treasurer. Requests and rates for back issues, changes of address, and undelivered copies of Madrono should be sent to the Corresponding Secretary. Information for Contributors Manuscripts submitted for publication in Madrono should be sent to the editor. It is preferred that all authors be members of the California Botanical Society. Manuscripts by authors having outstanding page charges will not be sent for review. Manuscripts may be submitted in English or Spanish. English-language manuscripts dealing with taxa or topics of Latin America and Spanish-language manuscripts must have a Spanish Resumen and an English Abstract. Manuscripts and review copies of illustrations must be submitted in triplicate for all articles and short items (NOTES, NOTEWORTHY COLLECTIONS, POINTS OF VIEW, etc.). Follow the format used in recent issues for the type of item submitted. Allow ample margins all around. Manuscripts MUST BE DOUBLE-SPACED THROUGHOUT. For articles this includes title (all caps, centered), author names (all caps, centered), addresses (caps and lower case, centered), abstract and resumen, 5 key words or phrases, text, acknowledgments, literature cited, tables (caption on same page), and figure captions (grouped as consecutive paragraphs on one page). Order parts in the sequence listed, ending with figures. Each page should have a running header that includes the name(s) of the author(s), a shortened title, and the page number. Do not use a separate cover page or 'erasable' paper. Avoid footnotes except to indicate address changes. Abbreviations should be used sparingly and only standard abbrevia- tions will be accepted. Table and figure captions should contain all information relevant to information presented. All measurements and elevations should be in metric units, except specimen citations, which may include English or metric measurements. Authors of accepted papers will be asked to submit an electronic version of the manuscript. Microsoft Word 6.0 or WordPerfect 6.0 for Windows is the preferred software. Line copy illustrations should be clean and legible, proportioned to the Madrono page. Scales should be in- cluded in figures, as should explanation of symbols, including graph coordinates. Symbols smaller than 1 mm after reduction are not acceptable. Maps must include a scale and latitude and longitude or UTM references. In no case should original illustrations be sent prior to the acceptance of a manuscript. Illustrations should be sent flat. No illustrations larger than 27 X 43 cm will be accepted. Presentation of nomenclatural matter (accepted names, synonyms, typification) should follow the format used by Sivinski, Robert C, in Madrono 41(4), 1994. Institutional abbreviations in specimen citations should follow Holmgren, Keuken, and Schofield, Index Herbariorum, 8th ed. Names of authors of scientific names should be abbreviated according to Brummitt and Powell, Authors of Plant Names ( 1992) and, if not included in this index, spelled out in full. Titles of all periodicals, serials, and books should be given in full. Books should include the place and date of publication, publisher, and edition, if other than the first. All members of the California Botanical Society are allotted 5 free pages per volume in Madrono. Joint authors may split the full page number. Beyond that number of pages a required editorial fee of $40 per page will be assessed. The purpose of this fee is not to pay directly for the costs of publishing any particular paper, but rather to allow the Society to continue publishing Madrono on a reasonable schedule, with equity among all members for access to its pages. Printer's fees for illustrations and typographically difficult material @ $35 per page (if their sum exceeds 30 percent of the paper) and for author's changes after typesetting @ $4.50 per line will be charged to authors. At the time of submission, authors must provide information describing the extent to which data in the manu- script have been used in other papers that are published, in press, submitted, or soon to be submitted elsewhere. VOLUME 47, NUMBER 3 JULY-SEPTEMBER 2000 MADRONO A WEST AMERICAN JOURNAL OF BOTANY Allozyme Data Support a Eurasian Origin for Carex viridula subsp. viridula VAR. VIRIDULA (CyPERACEAE) Shannon D. Kuchel and Leo P. Bruederle 1 47 The Taxonomic History, Identity, and Distribution of the Nevada Endemic, Plagiobothrys glomeratus (Boraginaceae) Arnold Tiehm 1 59 Genetic Variation in Pinus ponderosa, Purshia tridentata, and Festuca iDAHOENSis, Community Dominant Plants of California's Yellow Pine Forest Safiya Samman, Barbara L. Wilson, and Valerie D. Hipkins 164 The Effect of Climatic Variability on Growth, Reproduction, and Population Viability of a Sensitive Salt Marsh Plant Species, Lasthenia glabrata subsp. coulteri (Asteraceae) Lorraine S. Parsons and Adam W. Whelchel 1 74 Pleistocene Macrofossil Records of Four-Needled Pinyon or Juniper Encinal in the Northern Vizcaino Desert, Baja California del Norte Philip V Wells 189 Seed Banks of Long-Unburned Stands of Maritime Chaparral: Composition, Germination Behavior, and Survival with Fire Dennis C. Odion 195 Hedosyne (Compositae, Ambrosiinae), a New Genus for Iva ambrosiifolia John L. Strother 204 2^" Interface Between Ecology and Land Development in California Edited by J. E. Keeley, M. Baer-Keeley, and C. J. Fotheringham. U.S. Geological Survey Open-file Report 00-62. Sacramento, Ca Richard J. Hobbs 206 Synthesis of the North American Flora. Version 1.0 By John T. Kartesz and Christopher A. Meacham Ronald L. Hartman 207 California 209 Arizona and Mexico 211 Oregon 213 Washington 214 PUBLISHED QUARTERLY BY THE CALIFORNIA BOTANICAL SOCIETY Madrono (ISSN 0024-9637) is published quarterly by the California Botanical Society, Inc., and is issued from the office of the Society, Herbaria, Life Sciences Building, University of California, Berkeley, CA 94720. Subscription information on inside back cover. Established 1916. Periodicals postage paid at Berkeley, CA, and additional mailing offices. Return requested. Postmaster: Send address changes to Madrono, Roy Buck, % University Herbarium, University of California, Berkeley, CA 94720. Editor — KrISTINA A. ScHffiRENBECK California State University, Chico Department of Biology Chico, CA 95929-0515 kschierenbeck@csuchico.edu Editorial Assistant — Jeannie Trizzino Book Editor — Jon E. Keeley Noteworthy Collections Editors — Dieter Wilken, Margriet Wetherwax Board of Editors Class of: 2001— Robert Patterson, San Francisco State University, San Francisco, CA Paula M. Schiffman, California State University, Northridge, CA 2002 — Norman Ellstrand, University of California, Riverside, C A Carla M. D' Antonio, University of California, Berkeley, CA 2003 — Frederick Zechman, California State University, Fresno, CA JoN E. Keeley, U.S. Geological Service, Biological Resources Division, Three Rivers, CA 2004 — David M. Wood, California State University, Chico, CA Ingrid Parker, University of California, Santa Cruz, CA CALIFORNIA BOTANICAL SOCIETY, INC. Officers FOR 2000-2001 President: Bruce Baldwin, Jepson Herbarium and Dept. of Integrative Biology, 1001 Valley Life Sciences Bldg. #2465, University of California, Berkeley, CA 94720. First Vice President: Rod Myatt, San Jose State University, Dept. of Biol. Sciences, One Washington Square, San Jose, CA 95192. rmyatt@email.sjsu.edu Second Vice President: Roe Schlising, California State University, Chico, Dept. of Biol. Sciences, Chico, CA 95424. rschlising@csuchico.edu Recording Secretary: Dean Kelch, Jepson and University Herbarium, University of California, Berkeley, CA 94720. dkelch @ sscl.berkeley.edu Corresponding Secretary: Susan Bainbridge, Jepson Herbarium, University of California, Berkeley, CA 94720. suebain@SSCL.berkeley.edu Treasurer: Roy Buck, % University Herbarium, University of California, Berkeley, CA 94720. The Council of the California Botanical Society comprises the officers listed above plus the immediate Past President, R. John Little, Sycamore Environmental Consultants, 6355 Riverside Blvd., Suite C, Sacramento, CA 95831; the Editor of Madrono; three elected Council Members: Bian Tan, Strybing Arboretum, Golden Gate Park, San Francisco, CA 94122; James Shevock, USDI National Park Service, Pacific West Region, 600 Harrison Street, Suite 600, San Francisco, CA 94107; Diane Elam, U.S. Fish and Wildlife Service, 3310 El Camino Avenue, Sacramento, CA 95825; Graduate Student Representative: Kirsten Johanus, Jepson Herbarium, University of California, Berkeley, CA 94720. @ This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Madrono, Vol. 47, No. 3, pp. 147-158, 2000 ALLOZYME DATA SUPPORT A EURASIAN ORIGIN FOR CAREX VIRIDULA SUBSR VIRIDUlJi VAR. VIRIDULA (CYPERACEAE) Shannon D. Kuchel' and Leo P Bruederle Department of Biology, Campus Box 171, University of Colorado at Denver, PO. Box 173364, Denver, CO 80127-3364 Abstract Carex viridula Michaux subsp. viridida van viridula (Cyperaceae), the green sedge, occurs in wetland habitats distributed throughout northern and central North America. Its distribution also extends to the southern Rocky Mountain region in several disjunct sites, including alpine wetlands in Colorado, where it is rare. Populations of C. viridida from Colorado were investigated using starch gel electrophoresis of soluble enzymatic proteins coupled with substrate specific staining in order to describe genetic diversity and structure. The objective was to determine if Colorado populations exhibited the reduced genetic diversity expected of marginal populations when compared to other populations from North America and Europe. Genotypic data were collected for 15 enzyme systems encoded by 21 putative loci in 350 indi- viduals from seven populations in Colorado and in 179 individuals from eight populations from elsewhere throughout the range in North America. Data from all North American populations were compared with data previously reported from European populations of this species by Bruederle and Jensen (1991). No variation, either within or among North American populations, was detected at any of the loci. However, North American populations were genetically differentiated from European populations, with significantly more diversity maintained by European populations. The surprising lack of genetic diversity in North American populations is probably the combined result of high levels of selfing and inbreeding, restricted ecological amplitude, and genetic drift. Genetic bottlenecks are presumed to have occurred as a result of climate changes associated with Pleistocene glaciation or founding events associated with colonization of North America by proposed ancestral European populations. Since 1986, starch gel electrophoresis and allo- zyme analysis have been used to study genetic di- versity in no fewer than 43 species representing nine sections of the genus Carex. These studies have been useful not only in elucidating systematic relationships (e.g., Whitkus 1992; Ford et al. 1998), but also in providing indirect estimates of mating systems (e.g.. Waterway 1990), identifying hybrid origins (e.g., Standley 1990), and revealing corre- lations between genetic diversity and structure and certain life-history traits (e.g., Jonsson et al. 1996). However, few studies have examined genetic di- versity in disjunct populations of a broadly distrib- uted species of Carex. Carex section Ceratocystis Dumort. (Cypera- ceae) comprises seven species worldwide, which collectively occur throughout much of the northern hemisphere, particularly in boreal latitudes and the subalpine (Crins and Ball 1988). Carex viridula Michaux subsp. viridula var viridula, the green sedge, is the only representative of section Cera- tocystis in Colorado. Carex viridula is putatively one of the most re- cently derived members of section Ceratocystis, al- though it is not entirely clear when it diverged from its closest relative, C. viridula subsp. oedocarpa (N. ' Author to whom correspondence should be addressed. Current address: Environmental, Population, and Organ- ismic Biology, Campus Box 334, University of Colorado, Boulder, CO 80309-0334. J. Anderson) B. Schmid (Schmid 1984b; Crins and Ball 1989; Bruederle and Jensen 1991). It has been hypothesized that their common ancestor differen- tiated in West Europe. Thereafter, C viridula is pre- sumed to have colonized the remainder of the tem- perate and boreal northern hemisphere, perhaps be- fore Pleistocene glaciation, by way of the Bering land bridge (Crins and Ball 1989). Carex viridula is a short-lived perennial with a densely caespitose habit. It is monoecious, charac- terized by a single terminal staminate spike and several sessile pistillate spikes. It has been sug- gested that C. viridula is a dispersal generalist, with possible transport by biotic, e.g., birds and mam- mals, and abiotic agents, e.g., wind and water (Schmid 1984a; Crins and Ball 1989). While there are no apparent impediments to outcrossing, the breeding system is predominantly selfing. In a study by Schmid (1984a) examining the life history of C. viridula, tests for self-compatibility in the field and in experimental gardens using fine mesh bags to control pollination were positive. Addition- ally, inflorescences from which the staminate spike had been removed had maximum seed sets of only 10% when growing in the immediate vicinity of other fertile plants. Similarly, Bruederle and Jensen (1991) attributed low genetic diversity (e.g., pro- portion of polymorphic loci and observed hetero- zygosity) and deviations from Hardy- Weinberg equilibrium in West European populations of C. viridula to selfing. Genetic diversity was appor- 148 MADRONO [Vol. 47 tioned among populations with relatively little vari- ation found within populations. Ecologically, C. vihdula is characterized by rap- id growth and development, small stature, short life-span, early reproduction, large reproductive ef- fort, and small population size (Schmid 1984a, b). As such, C. viridula is an early successional, r- selected species, or in Grime's (1979) classification, a ruderal species. Although C. viridula typically oc- cupies moist, early successional sites characterized by fluctuating and unpredictable water levels, these can vary from calcareous, acidic, sandy, or organic shorelines; runnels in limestone barriers; wet mead- ows; marshes; on borders of streams, ponds, and lakes; and fens. This species' success in colonizing is likely due to its tolerance of diverse and fluctu- ating environments, high phenotypic plasticity, and ability to reproduce quickly and profusely (Schmid 1984a, b; Crins and Ball 1989). Geographically, C. viridula is the most wide- spread taxon in section Ceratocystis, with a near circumboreal distribution. It is common throughout northern Europe, and much of northern and central North America; it is also scattered across the cen- tral and eastern parts of temperate Asia to the Pa- cific Ocean (Crins and Ball 1989). In North Amer- ica, its range extends south in the Rocky Mountain region to several disjunct sites in Colorado, Wyo- ming, Utah, and Nevada. In the Southern Rocky Mountains, C. viridula occupies an uncommon hab- itat, alpine wetlands, with habitat specificity con- tributing to rarity in this region (Rabinowitz 1981). The most significant threat to these rare populations may be habitat alteration and loss, as a result of peat mining and the draining of wetlands for irri- gation of surrounding ranchlands and diversion to municipal drinking water supplies. Although C. vir- idula has been assigned a state ranking of SI, in- dicating that it is critically imperiled in Colorado, it has received a global ranking of G5, indicating that it is demonstrably secure globally (Spackman et al. 1997). Extant Colorado populations of C. viridula are geographically marginal, occurring at the edge of the species distribution in North America. Further- more, North American populations, in general, are peripheral relative to West Europe, the putative center of diversity and origin for C. viridula (Crins and Ball 1989). Genetic theory predicts differenti- ation of marginal populations with respect to cen- tral populations, with reduced levels of genetic variation and greater population differentiation (Bruederle 1999). While marginal populations are expected to maintain a subset of the genetic varia- tion observed in central populations as a result of reduced gene flow (Yeh and Layton 1979), both random genetic drift and selection may cause the fixation of alleles that are rare in central popula- tions (Blows and Hoffman 1993). However, other factors in addition to distribution may influence levels and apportionment of genetic diversity in species. Geographical range, succes- sional status, population size, life form, breeding system, and seed dispersal mechanism have all been demonstrated to have significant effects on ge- netic diversity and structure (Brown 1979; Hamrick et al. 1979; Loveless and Hamrick 1984; Karron 1987; Hamrick and Godt 1989; Hamrick et al. 1991; Barrett and Kohn 1991). Furthermore, levels and apportionment of genetic variation could be the consequence not only of life-history characteristics, but also of historical and evolutionary events such as genetic bottlenecks resulting from founder effect, glaciation, migration, and speciation (Lewis and Crawford 1995). The purpose of this research is to describe ge- netic diversity and structure in populations of C. viridula from Colorado relative to other North American and West European populations. Collec- tively, the aforementioned influential factors lead to three specific predictions regarding population ge- netic diversity and structure in C. viridula. First, life-history characteristics, such as self-compatibil- ity, restricted habitat, short-lived perennial growth, herbaceous habit, ruderal strategy, and small pop- ulation sizes, are expected to confer low levels of within-population genetic variation (Schmid 1984a, b; Crins and Ball 1988, 1989; Bruederle and Jensen 1991). Second, deviations from Hardy- Weinberg equilibrium and heterozygote deficiency are ex- pected as a result of the caespitose habit, which has been correlated with high levels of inbreeding and genetic substructuring. Finally, pronounced differ- entiation among populations is expected as a result of reduced gene flow, isolation, inbreeding, and ge- netic drift. It is expected that Colorado populations will be genetically differentiated from other North American populations, and that North American populations, in general, will be genetically differ- entiated from West European populations. Materials and Methods Fifteen populations of C. viridula were sampled during the summers of 1998 and 1999 from the Pacific Northwest, Rocky Mountain, and Great Lakes regions of the United States and Canada (Fig. 1, Table 1). Sites were typically peatlands or other wetlands, with C. viridula occupying early succes- sional microsites along the shores of streams, springs, ponds, creeks, and swamps; or along road- sides, ditches, and ruts (Table 1). Population sam- ples ranged in size from six to 50 individuals. Be- cause C. viridula is caespitose, samples obtained from discrete, well-spaced clumps were assumed to represent different individuals. At each site, whole vegetative culms were harvested, placed in separate plastic bags with moist paper towels, and kept re- frigerated until extraction of soluble enzymatic pro- teins. Extraction procedures followed those previ- ously reported in Bruederle and Fairbrothers (1986). Voucher specimens for each population 2000] KUCHEL AND BRUEDERLE: EURASIAN ORIGIN FOR CAREX VIRIDULA 149 Fig. 1. Map of North America showing the locations of 15 Carex viridula subsp. viridula van viridula populations sampled for allozyme analysis. Detailed locations are provided for seven Colorado populations. Population numbers correspond to those in Table 1. have been deposited at the University of Colorado at Denver Herbarium and the University of Colo- rado Herbarium (COLO). Electrophoresis and staining followed Bruederle and Fairbrothers (1986) and Bruederle and Jensen (1991). Three gel-buffer systems and 15 enzyme stains were used to resolve 21 putative loci using 10.5% starch gels (Sigma-Aldrich, Inc.). Isocitrate dehydrogenase (IDH), malate dehydrogenase (MDH), 6-phosphogluconate dehydrogenase (6PGD), phosphoglucose isomerase (PGI), phos- phoglucomutase (PGM), and shikimic acid dehy- drogenase (SkDH) were stained on a discontinuous histidine-HCl system (Gottlieb 1981). Aspartate aminotransferase (AAT), acid phoshatase (ACP), and glyceradehyde-3-phosphate dehydrogenase (G3PDH) were stained on a tris-citrate system (Sol- tis et al. 1983). Alcohol dehydrogenase (ADH), dia- phorase (DIA), malic enzyme (ME), menadione re- ductase (MNR), superoxide dismutase (SOD), and triose phosphate isomerase (TPI) were stained on a discontinuous lithium-borate system (Soltis et al. 1983). Staining followed Soltis et al. (1983) with minor modifications for IDH, MDH, PGI, PGM, SkDH, ACP, G3PDH, ME, and TPI; Gottlieb (1973) for PGD and ADH; and Cardy et al. (1981) for AAT. Electrophoretic allozyme phenotypes were inter- preted genetically on the basis of segregation pat- terns, know substructure and intracellular compart- mentalization of enzymes, and previously observed electrophoretic patterns (e.g., Bruederle and Fair- brothers 1986). Data were collected as individual genotypes; these data have been deposited at the University of Colorado at Denver Herbarium and are available upon request (LPB). Standards rep- resenting most of the common electrophoretic var- iants for the section were incorporated into each of the gels to facilitate allele identification. Data were analyzed using BIOSYS-1 (Swofford and Selander 1981) to obtain common measures of genetic diversity including proportion of polymor- phic loci {P), mean number of alleles per locus (A) and per polymorphic locus (A^,), observed hetero- zygosity (//,,), and expected heterozygosity (//J. The distribution of genetic variation within and among populations was calculated using Nei's (1973) gene diversity statistics and GENESTAT-PC (Lewis and Whitkus 1989). In order to assess dif- ferences in genetic diversity, data from North American populations were compared with data previously reported for European populations of this species using a /-test (Table 1; Bruederle and Jensen 1991). Results The fifteen enzymes assayed for C. viridula are encoded by 21 putative loci: AAT-1, AAT-2, ACP- 1, ADH, DIA-1, DIA-3, G3PDH, IDH-1, IDH-2, MDH-1, MDH-2, ME, MNR, 6PGD, PGI -2, PGM- 1, PGM-2, SkDH, SOD, TPI- J, and TPI-2. Four ad- ditional loci {MDH-3, PGI-1, ACP-2, and ACP-3) were not included in the analysis, because they did not exhibit consistent activity or clearly interpreta- ble banding patterns. North American C. viridula maintains no genetic diversity based upon this sample of 15 populations and 21 loci. None of the loci examined were poly- morphic. At every locus assayed, each of the 529 individuals was homozygous for the same allele; no heterozygosity or allozyme variation was observed 150 MADRONO [Vol. 47 Table 1. Locations and Site Information for 15 North American Carex viridula Michx. subsp. viridula var. viRjDULA Populations Sampled for Allozyme Analysis, as well as Three West European Populations (Bruederle AND Jensen 1991). Country, Estimated Habitat and Pop. state, and Sample popula- microhabitat no. province size Location Latitude Longitude tion size description 1 /^/^ T TO A 1 CD, USA 50 Park Co., Colorado, jy 05 N 105 55 W 1000 scattered throughout High Creek Fen, well-developed I J Km {o mj o or peatland; coloniz- Fairplay on U.S. ing alongside the Rte. 285 banks ot a tew streams 2 CO, UoA 50 Park Co., Colorado, 39 03 JN 1 ncoco 'WT 105 5o W 1200 scattered throughout Sweet Water well-developed Kancn, 1 o km (11 peatland; coloniz- m) S of Fairplay ing alongside on U.S. Rte. 285 large, deep ditch dug through peat- land O /^/^ T TO A 3 CO, USA 50 Park Co., Colorado, 39 09 N 106 03 W 100 growing on shore of Warm Springs spring and on Ranch, 5 km (3 moist areas adja- m) S of Fairplay cent to spring, on U.S. Rte. 285 alongside snore or pond A /^/~\ T T O A 4 CO, USA 50 Jackson Co., Colora- A r\0 /I /I ' TVT 40 44 N 106 34 W 50 shore ot pond, colo- do, Lone Pine, 23 nizing edge of one Km (I'M- m^ w or bank and on top Walden on Co. of a few hum- Rd. 16 mocks 5 CO, USA 50 Jackson Co., Colora- 40 45 IN 106 35 W 50 shore of creek, colo- do, Bear Creek, nizmg along fallen 24 km (15 m) W log and within ad- of Walden on Co. jacent rut Rd. lo ^ /^/^ T TO A 6 CO, USA 50 Grand Co., Colora- 39 55 N 106 19 W 75 near outlet of small do, Haystack alkaline spring, Mountain, 16 km some peat accu- (10 m) N ol Sil- mulation, growing verthorne on St. on top of small Rte. 9 hummocks ^ /~^/^ T TO A 7 CO, USA 50 San Juan Co., Colo- 37 43 N 1 r\^0 A'^ f\\T 107 42 W 1000 scattered throughout rado, Andrew's well-developed Lake, 10 km (o peatland; coloniz- m) S of Silverton ing shores of a on U.S. Rte. 550 few ponds 8 OH, USA 28 Ottawa Co., Ohio, 41°31'N 82 45 W 1000 scattered through Quarry Rd., 1 km moist areas in (0.6 m) SW or floor of old lime- Lakeside on St. stone quarry Rte. 163 9 Ml, USA 27 Iosco Co., Michigan, A /I O 1 O ' XT 44 12 N O Oo ^ f\\7 83 37 W 500 adjacent to swamp. 6 km (4 m) W or colonizing burrow T T O T~» A. O XT V U.S. Rte. 23, N of pit Alabaster Rd. 10 MI, USA 25 Mackinac Co., Mich- 46°12'N 85°21'W 50 well-drained edge of igan, 13 km (8 m) sandy road N of U.S. Rte. 2, W of Borgstrom Rd. 11 WI, USA 27 Waushara Co., Wis- 44°09'N 89°09'W 100 shore of lake; grow- consin, Hills Lake, ing in rows of re- 6 km (4 m) E of cent, successive St. Rte. 22, S of colonizations Co. Rd. H 2000] KUCHEL AND BRUEDERLE: EURASIAN ORIGIN FOR CAREX VIRIDULA 151 Table 1. Continued. Country, • Estimated Habitat and Pop. state, and Sample popula- microhabitat no. province size Location Latitude T f A Longitude tion size description 12 WA, USA 28 King Co., Washing- 47°40'N 121°37'W 1000 scattered throughout ton, Snoqualmie large bog/fen sys- Bog, 21 km (13 tem underlain by m) E of St. Rte. peat soil 203, N of N. Fork Co. Rd. 13 ONT, CAN 25 Norfolk Co., Ontar- 42°34'N 80°25'W 100 colonizing middle io, Long Point on and edges of Lake Erie, 48 km raised vehicular (30 m) S of Hwy. track 401 14 ONX CAN 6 Peterborough Co., 44°40'N 78°80'W 100 shore of lake Ontario, Belmont Lake, 24 km (15 m) N of Hwy. 7 15 ONT, CAN 13 Nipissing Dist., On- 45°60'N 78°30'W 100 shore of lake tario, Radiant Lake, Algonquin Prov. Park, 16 km (10 m) N of Rte. 60 16 AUS 3 Trunnahiitte, Austria, 47°30'N ir70'E na growing in seepage 3.8 km (2.4 m) at base of slope ^ SSW of Trin 17 SWE 26 Asa, Sweden, 0.9 58°65'N ir80'E na colonizing along km (0.56 m) hummocks and in WSW depressions in sat- urated soils of meadow 18 SWE 28 Skanor, Sweden, 1 56°10'N 12°90'E na colonizing around km (0.62 m) E rocks, hummocks. and in depressions of low, pastured meadow (Table 2). As such, for all populations, the propor- tion of loci polymorphic (P), observed heterozy- gosity (//^), and expected heterozygosity (//J were all zero. Similarly, the number of alleles per locus (A) was one, the minimum value for this statistic (Table 3). While low. West European populations exhibited higher levels of genetic diversity than North Amer- ican populations (Bruederle and Jensen 1991). On average, two of the 20 loci examined for West Eu- ropean populations (10.0%) were polymorphic (Ta- ble 2). Mean number of alleles per locus was 1.15, while mean number of alleles per polymorphic lo- cus was 2.0. Observed heterozygosity was 0.014 and expected heterozygosity was 0.039 (Table 3). Genetic diversity in West European populations was significantly higher than that in North Ameri- can populations when compared using a two sam- ple r-test assuming unequal variances for proportion of polymorphic loci {P < 0.10), mean number of alleles per locus (P < 0.05), observed heterozygos- ity (P < 0.05), and expected heterozygosity {P < 0.10). Levels of genetic diversity for the species across its sample range in North America and West Europe were extremely low. The mean proportion of poly- morphic loci was 1.7%. Mean number of alleles per locus was 1.03, while mean number of alleles per polymorphic locus was 2.0. Observed heterozygos- ity was 0.002 and expected heterozygosity was 0.007 (Table 3). Despite the differences in levels of genetic di- versity. West European and North American pop- ulations are, in fact, very similar. Mean genetic identity obtained from pairwise comparisons of populations from. North America and West Europe was 0.987. Mean genetic identity among North American populations was 1.000, and among Eu- ropean populations was 0.974, ranging from 0.960 to 1.000 for the latter. Of the small amount of di- versity maintained by populations of C. viridula, the majority (G,, = 0.650) was due to differences among populations, both among West European and between North American and West European populations. 152 MADRONO [Vol. 47 3 Oh o Oh a o W 3 o ^ r<-) "vO O O On O ON ON O O d d d d d d d o o -vO ^ o o o O O O On O O O d> d> 1^ n-) O O O O O ^ O O O O O n-i o d d d d o o o o o o o o o o o o o o o d — ' d d o o o o o o o o o o o o o o o d — — d d o o o o o o o o o o o o o o o d — — ' d d o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o ^' d — d d o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o d — ■ d d o o o o o o o o o o o o o o o — ■ d " — ■ d d o o o o o o o o o o o o o o o d d d o o o o o o o o o o o o o o o — d — d d o o o o o o o o o o o o o o o d — " d d o o o o o o o o o o o o o o ^ Q Discussion As expected. North American populations of C. viridula do maintain extremely low levels of ge- netic diversity within populations. However, due to the low levels of genetic diversity within popula- tions and subsequent lack of allozyme markers, no genetic differentiation was observed among popu- lations in Colorado or North America. Neverthe- less, North American populations were genetically differentiated from West European populations, with significantly more diversity maintained by West European populations. A number of factors could have contributed to the paucity of genetic di- versity observed in populations of this species. First, the plant allozyme literature reveals strong associations between genetic diversity and breeding system, with those species characterized by out- crossed breeding systems having significantly high- er levels of genetic diversity apportioned among in- dividuals within populations. On average, species characterized by selfing breeding systems maintain significantly lower levels of genetic diversity, in- cluding proportion of polymorphic loci, number of alleles per locus and per polymorphic locus, and observed heterozygosity (Brown 1979; Hamrick et al. 1979; Loveless and Hamrick 1984; Hamrick and Godt 1989; Hamrick et al. 1991). Carex viridula has been shown to exhibit popu- lation genetic structure and seed set suggestive of selfing (Schmid 1984a; Bruederle and Jensen 1991). The extremely low levels of genetic varia- tion found in North American populations of C. viridula in this study may be the result, in part, of such selfing. High levels of inbreeding attributable to selfing are expected to result in homozygosity and decreased genetic variability. Of the large num- ber of species of vascular plants that have been ex- amined similarly, at least 14 other taxa have been reported to maintain no detectable allozyme diver- sity (Table 4). Although an exact comparison be- tween these taxa and C. viridula is not possible, it is noteworthy that almost all of these taxa also show substantial levels of selfing. In graminoids, high levels of inbreeding have also been correlated with the caespitose growth form. Stebbins (1950) proposed a relationship be- tween growth form and breeding system among grasses, suggesting that species with a rhizomatous growth form are predominantly outcrossing due to the intermingling of genets. In contrast, the caes- pitose habit results in a growth form in which the nearest neighbor of a flowering culm is another culm from the same plant (e.g., ramet), thus pro- moting inbreeding, and specifically, selfing. Genetic evidence substantiating this phenomenon in the gra- minoid genus Carex was first reported by Bruederle and Fairbrothers (1986) and Bruederle (1987). Ad- ditional evidence supporting the relationship be- tween growth form and genetic variability in Carex 2000] KUCHEL AND BRUEDERLE: EURASIAN ORIGIN FOR CAREX VIRIDULA 153 Table 3. Summary of Genetic Diversity for 18 Populations of Carex viridula Michx. subsp. viridula var. viri- dula: Sample Size (N), Mean Number of Alleles Per Locus (A) and Per Polymorphic Locus (A^), Proportion of Polymorphic Loci (P), Observed Heterozygosity (//„), and Expected Heterozygosity (//,.). Population numbers correspond to those in Table 1 . A locus was considered polymorphic if the frequency of the most common allele did not exceed 0.95. Pop. No. N A A r P 1 50 1.0 — 0.0 0.000 0.000 2 50 1.0 — 0.0 0.000 0.000 3 50 1.0 — 0.0 0.000 0.000 4 50 1.0 — 0.0 0.000 0.000 5 50 1.0 — 0.0 0.000 0.000 6 50 1.0 — 0.0 0.000 0.000 7 50 1.0 — 0.0 0.000 0.000 8 28 1.0 — 0.0 0.000 0.000 9 27 1.0 — 0.0 0.000 0.000 10 25 1.0 — 0.0 0.000 0.000 27 1 .0 \J.\J 0.000 0.000 12 28 1.0 0.0 0.000 0.000 13 25 1.0 0.0 0.000 0.000 14 6 1.0 0.0 0.000 0.000 15 13 1.0 0.0 0.000 0.000 Mean-N. America 529 1.0 0.0 0.000 0.000 16 3 1.15 2.0 15.0 0.017 0.070 17 26 1.05 2.0 5.0 0.006 0.006 18 28 1.25 2.0 10.0 0.018 0.042 Mean-W. Europe 57 1.15 2.0 10.0 0.014 0.039 Mean-Species 586 1.03 2.0 1.7 0.002 0.007 was subsequently provided by Ford et al. (1991, 1998). A survey of the population genetic literature for the genus Carex revealed data for 29 taxa including six rhizomatous and 23 caespitose carices (Kuchel, unpubl.). On average, populations of rhizomatous species harbor high levels of genetic diversity, e.g., Ap = 2.26 ± 0.12, P = 44.5 ± 4.08%, and //, = 0.171 ± 0.038, while caespitose species have sig- nificantly less, e.g.. A,, = 2.03 ± 0.09 (P < 0.05), P = 13.4 ± 12.0% (P < 0.001), and = 0.042 ± 0.04 {P < 0.001). Furthermore, whereas popu- lations of rhizomatous species are poorly differen- tiated (G^f = 0.159 ± 0.053), caespitose species are well-differentiated with nearly half of all genetic diversity attributable to differences among popula- tions (G,, = 0.462 ± 0.272). Although exceptions exist (Ford et al. 1998), it would appear that rhi- zomatous species maintain more variation within and less differentiation among populations, presum- ably due to outcrossing. Conversely, caespitose species have less variation within and more differ- entiation among populations, presumably due to in- breeding. As such, the extremely low levels of ge- netic variation found in North American popula- tions of C. viridula in this study may be the result, in part, of the caespitose growth form. Second, narrowly distributed plant species tend to maintain lower levels of genetic variation than more widespread species (Karron 1987; Karron et al. 1988; Hamrick and Godt 1989). The lower lev- els of genetic variation observed in narrowly dis- tributed species may be due to changes in allele frequencies due to chance (genetic drift and found- er effect) or strong, directional selection toward ge- netic uniformity in a limited habitat type (Karron 1987). Almost all of the aforementioned genetically invariable taxa are narrowly distributed (Table 4). Wolff and Jefferies (1987) hypothesized that the lack of diversity in one of these taxa, Salicornia europaea L., could be due to its restricted ecolog- ical distribution, despite the fact that it is geograph- ically widespread. Ecologically, this species is con- fined to chronically disturbed, early successional open habitats in coastal and inland salt marshes where individuals and populations are subject to considerable turnover and population re-establish- ment. The narrow habitat requirements, founding events, small population sizes, and possible selec- tion pressures experienced in such an environment could have contributed to the observed paucity of genetic diversity in S. europaea. Even though Carex viridula is distributed throughout boreal North America, its ecological distribution also appears to be narrow. Carex viri- dula is a habitat specialist, occurring only in highly disjunct wetland habitats. Additionally, C. viridula is confined to early successional microsites, which tend to be small, ephemeral, highly variable, and subject to repeated local extinction and coloniza- tion. As in Salicornia europaea, it is possible that the extremely low levels of genetic variation found in populations of Carex viridula in this study may be, in part, the result of this narrow ecological dis- tribution. It is interesting to note that a number of species 154 Table 4. MADRONO [Vol. 47 Summary of Allozyme Literature for those Species Having No Detectable Allozyme Variation. Species Breeding system Geographic range/ Ecological amplitude Inferred historical mechanisms Reference Carex viridula Howellia aqua- tilis Gray Lespedeza lep- tostachya En- gelm. Pedicularis fur- bishiae S. Wats Pinus resinosa Ait. selfing selfing Bensoniella ore- gona (Abrams & Bacig) Morton Chrysosplenium selfing iowense Rydb. ipproaches obli- gate selfing selfing Oenothera hook- selfing eri Torr. and Gray pollinator re- quired for pol- lination, but possibly self- compatible highly self-com- patible narrow distribution ecologically narrow distribution geographically narrow distribution geographically and ecologically narrow distribution geographically and ecologically narrow distribution geographically narrow distribution geographically narrow distribution geographically and ecologically widespread geographi- cally Genetic drift: inbreeding and genetic bottleneck associated with founding events and/or climate changes during glaciation Rapid colonizer Disturbed habitats Genetic drift: inbreeding and genetic bottleneck associated with range re- strictions during glacia- tion Clonal growth Small population sizes Genetic drift: inbreeding and genetic bottleneck associated with chmate changes during glaciation Clonal growth Small population sizes Genetic drift: inbreeding and genetic bottleneck associated with range re- strictions during glacia- tion Age of populations, not enough time to accumu- late variability and het- erozygosity Genetic drift: inbreeding and genetic bottleneck associated with range re- strictions during glacia- tion Genetic drift: inbreeding Age of populations, not enough time to accumu- late variability and het- erozygosity Rapid colonizer Permanent translocation heterozygosity Genetic drift: inbreeding and genetic bottleneck associated with range re- strictions during glacia- tion and/or founding events Local population extinctions Disturbed habitats Genetic drift: inbreeding and genetic bottleneck associated with range re- strictions during glacia- tion Age of populations, not enough time to accumu- late variability and het- erozygosity Soltis et al. 1992 Schwartz 1985 Lesica et al. 1988 Cole and Biesboer 1992 Levy and Levin 1975 Waller et al. 1987 Fowler and Morris 1977; Allendorf et al. 1982; Simon et al. 1986; Mosseler et al. 1991 2000] KUCHEL AND BRUEDERLE: EURASIAN ORIGIN FOR CAREX VIRIDULA 155 Table 4. Continued. Geographic range/ Species Breeding system Ecological amplitude Inferred historical mechanisms Reference Salicornia eiiro- selfing narrow distribution Genetic drift: inbreeding Jefferies and Gottlieb paea L. (s.I.) ecologically and genetic bottleneck 1982; Wolff and associated with range re- Jefferies 1987 strictions during glacia- tion and founding events Rapid colonizer Senecio mohav- obligate selfing narrow distribution Genetic drift: genetic bottle- Liston et al. 1989 ensis Gray geographically neck associated with founding events (recent colonization of North America) and/or climate changes during glaciation Rapid colonizer SuUvantia ore- selling narrow distribution Genetic drift: inbreeding Sohts 1982 gana S. Wats. geographically and and genetic bottleneck ecologically associated with range re- strictions during glacia- tion Taraxacum obli- agamospermous narrow distribution Post-glacial range expan- Van Oostrum et al. quum (Fr.) geographically and sion through hybridiza- 1985 Dahlst. ecologically tion to form polyploid agamospermous popula- tions Selection in more severe and less diverse environ- ments Thuja pUcata self-compatible narrow distribution No explanation given Copes 1981 Donn ex D. geographically Genetic bottleneck associat- Don ed with range restrictions unlikely; no evidence of physical barriers and as- sociated species show abundant genetic varia- Tragopogon selfing widespread geographi- Genetic drift: inbreeding Roose and Gottlieb pratensis cally Rapid colonizer 1976 Ownbey Typha domin- selfing widespread geographi- Genetic drift: inbreeding Mashburn et al. 1978; gensis Pers. cally and ecologi- Clonal growth Sharitz et al. 1980 cally Rapid colonizer Disturbed habitats reported to maintain no detectable allozyme diver- sity are rapid colonizers of disturbed habitats (Table 4). Carex vihdula has also been described as a rap- id colonizer and ruderal species with rapid growth and development, small size, short life-span, early reproduction, large reproductive effort, and small population size (Schmid 1984a, b). In Switzerland, C. viridula often occupies newly disturbed sites, with many small and isolated populations. It is a pioneer in open, wet habitats, but is quickly ex- cluded successionally (Schmid 1986). Not surpris- ingly, most populations sampled for this study ap- pear to occupy early successional microsites, grow- ing along pond shores, stream banks, roadsides, ditches, and ruts, comprising a part of larger, later successional communities (Table 1). Schmid (1984b) hypothesized that such ruderal species with early successional populations would have high ge- netic variability between, but low genetic variabil- ity and high plasticity within populations, as a re- sult of small population size, genetic drift, and di- rectional selection. This hypothesis is supported by the ecology of C. viridula (Schmid 1984b), as well as the present study of genetic diversity. Thus, low levels of genetic diversity found with- in populations of C. viridula may be attributed, in part, to effective breeding system and restricted ecological distribution. However, populations of this species in North America show substantially lower levels of genetic variability, even when com- pared to those means reported for selfing or nar- rowly distributed species — it should be reiterated that all putative loci examined were monomorphic. The most likely explanation for such low levels of polymorphism and heterozygosity is genetic drift. It is possible that a genetic bottleneck occurred at 156 MADRONO [Vol. 47 some point in the history of these populations that eliminated all or most of the allozyme polymor- phism. One possibility is that a genetic bottleneck re- sulted from the founding of North American pop- ulations. During migration and dispersal, new pop- ulations may be formed by a small number of initial colonists. The genetic material of such populations is limited to those alleles introduced by these few founders and may not be representative of the spe- cies as a whole (Schwaegerle and Schaal 1979). Since European populations are the proposed pro- genitors of North American populations (Crins and Ball 1989), it would be expected that the allozymes present in North American populations would largely comprise a subset of those alleles present in European populations (Crawford 1983; Cole and Biesboer 1992). Indeed, North American popula- tions are genetically differentiated from West Eu- ropean populations, with West European popula- tions of C. viridula harboring higher levels of ge- netic diversity, and North American populations harboring a subset of that genetic diversity. These data suggest that a small number of individuals from the putatively ancestral European populations founded North American populations. Limited gene flow between populations would likely maintain the lower levels of genetic diversity observed in North American populations. Although populations of Bromus tectorum L. have only recently been intro- duced to North America from Eurasia, a number of similarities can be seen between this species and C. viridula. Both species exhibit low genetic variabil- ity within and high differentiation among popula- tions possibly as a result of founding events and a selfing breeding system. Ecologically, both are characterized by routinely disturbed habitats and high phenotypic plasticity (Novak et al. 1991). Another possibility is that a genetic bottleneck occurred at some point after the founding of North American populations. This bottleneck could have resulted in genetic uniformity in an original popu- lation or populations with a reduced geographic distribution, followed by spread of this species to the range now occupied (Lesica et al. 1988; Cole and Biesboer 1992). Climatic changes during the Pleistocene, particularly the Xerothermic period about 8500 to 3000 y B.P., have been suggested as a possible cause for genetic bottlenecks in a number of other North American species exhibiting ex- tremely low levels of genetic diversity (Table 4), as well as in other species of Carex (Waterway 1990). Additionally, it has been suggested that these pop- ulations have not had enough time to accumulate variation and differentiate since these events (Levy and Levin 1975; Ledig and Conkle 1983; Lesica et al. 1988; Liston et al. 1989). It has been suggested that allozyme variation de- tectable by electrophoresis may not provide a com- plete measure of genetic diversity in the genome (Mosseler et al. 1991; Mosseler et al. 1992). A more direct analysis of variation in DNA, e.g., RAPDs, may provide the genetic markers necessary to infer genetic diversity and structure in C. viri- dula. Data from the present study do indicate that C. viridula is genetically depauperate over a large portion of its range. However, additional analyses of populations lying at the extreme northwestern, northeastern, and eastern edges of the distribution in North America, as well as in putative glacial re- fugia, should be carried out to confirm this. Anal- yses of additional Eurasian populations would also contribute to a reconstruction of the biogeography of C. viridula. Acknowledgments We are grateful to the following people and organiza- tions for their assistance with collections: P. W. Ball, Uni- versity of Toronto; W. J. Grins, Ontario Ministry of Nat- ural Resources; A. Cusick, Ohio State Department of Nat- ural Resources; J. Sanderson, Colorado Natural Heritage Program; F. Weinmann; and S. Komarek. We thank P. W. Ball, University of Toronto; B. A. Ford, University of Manitoba; K. A. Schierenbeck, California State Univer- sity, Chico; and D. F Tomback and T. M. Roane, Univer- sity of Colorado at Denver, for review of the manuscript. This work was supported, in part, by grants from the Sig- ma Xi Grants-in-Aid of Research Program and the Colo- rado Native Plant Society. Literature Cited Allendorf, F. W, K. L. Knudsen and G. M. Blake. 1982. Frequencies of null alleles at enzyme loci in natural populations of ponderosa and red pine. Ge- netics 100:497-504. Barret, S. C. H. and J. Kohn. 1991. Genetic and evo- lutionary consequences of small population size in plants: implications for conservation, p. 3-30. in D. A. Falk and K. E. Holsinger (eds.). Genetics and con- servation of rare plants. Oxford University Press, New York, N.Y. Blows, M. W. and A. A. Hoffman. 1993. The genetics of central and marginal populations of Drosophila serrata. 1. Genetic variation for stress resistance and species borders. Evolution 47:1255-1270. Brown, A. H. D. 1979. Enzyme polymorphism in plant populations. Theoretical Population Biology 15:1-42. Bruederle, L. p. 1999. Genetic differentiation of geo- graphically marginal populations in Carex mitchelli- ana (Cyperaceae): implications for conservation. Journal of the Torrey Botanical Society 126:1-8. . 1987. Allozyme variation in the genus Carex (Cyperaceae). XIV International Botanical Congress Abstracts 14:304. , and D. E. Fairbrothers. 1986. Allozyme vari- ation in populations of the Carex crinita complex (Cyperaceae). Systematic Botany 11:583-594. , AND U. Jensen. 1991. Genetic Differentiation of Carex flava and Carex viridula in West Europe (Cy- peraceae). Systematic Botany 16:41-49. Cardy, B. J., C. W. Stuber and M. M. Goodman. 1981. Techniques for starch gel electrophoresis of enzymes from maize {Zea mays L.). Institute of Statistics Mim- eograph Series No. 1317. North Carolina State Uni- versity, Raleigh, NC. Cole, C. T. and D. D. Biesboer. 1992. Monomorphism, 2000] KUCHEL AND BRUEDERLE: EURASIAN ORIGIN FOR CAREX VIRIDULA 157 reduced gene flow, and cleistogamy in rare and com- mon species of Lespedezci (Fabaceae). American Journal of Botany 79:567-575. Copes, D. L. 1981. Isoenzyme uniformity in western red cedar seedlings from Oregon and Washington. Ca- nadian Journal of Forest Research 1 1:451-453. Crawford, D. J. 1983. Phylogenetic and systematic in- ferences from electrophoretic studies, p. 257-87. in S. D. Tanksley and T. J. Orton (eds.). Isozymes in plant genetics and breeding, part A. Elsevier Science Publishers B. V., Elsevier, Amsterdam. Crins, W. J., AND P. W. Ball, 1988. Sectional limits and phylogenetic considerations in Corex Section Cera- tocystis (Cyperaceae). Brittonia 40:38-47. . 1989. Taxonomy of the Carex flava complex (Cy- peraceae) in North America and northern Eurasia. II. Taxonomic treatment. Canadian Journal of Botany 67:1048-1065. Fowler, D. P. and R. W. Morris. 1977. Genetic diversity in red pine: evidence for low genie heterozygosity. Canadian Journal of Forest Research 7:343-347. Ford, B. A., P. W. Ball and K. Ritland. 1991. AUozyme diversity and genetic relationships among North American members of the short-beaked taxa of Carex sect. Vesicciriae (Cyperaceae). Systematic Botany 16: 116-131. , D. A. R. McQueen, R. E C. Naczi and A. A. Reznicek. 1998. Allozyme variation and genetic re- lationships among species in the Carex willdeuowii complex (Cyperaceae). American Journal of Botany 85:546-552. Gottlieb, L. D. 1973. Genetic differentiation, sympatric speciation, and the origin of a diploid species of Stephanomeria. American Journal of Botany 69:545- 553. . 1981. Gene number in species of Astereae that have different chromosome numbers. Proceedings of the National Academy of Sciences, USA 78:3726- 3729. Grime, J. P. 1979. Plant Strategies and Vegetation Pro- cesses. John Wiley and Sons, New York, N.Y. Hamrick, J. L., Y. B. Linhart and J. B. Mitton. 1979. Relationships between life history characteristics and electrophoretically detectable genetic variation in plants. Annual Review of Ecology and Systematics 10:173-200. , and M. J. W. GoDT. 1989. Allozyme diversity in plant species, p. 43-63. in A. H. D. Brown, M. T. Clegg, A. L. Kahler and B. S. Weir (eds.). Plant pop- ulation genetics, breeding and genetic resources. Sin- auer, Sunderland, MA. , , D. A. MuRAWSKi AND M. D. Loveless. 1991. Correlations between species traits and allo- zyme diversity: implications for conservation biolo- gy, p. 75-86. in D. A. Ealk and K. E. Holsinger (eds.). Genetics and conservation of rare plants. Oxford Uni- versity Press, New York, N.Y. Jefferies, R. L. AND L. D. Gottlieb. 1982. Genetic dif- ferentiation of the microspecies Salicornia europaea L. {sensu stricto) and 5'. ramosissima J. Woods. New Phytology 92:123-129. JONSSON, B. O., I. S. JONSDOTTIR AND N. CRONBERG. 1996. Clonal diversity and allozyme variation in popula- tions of the arctic sedge Carex higelowii (Cypera- ceae). Journal of Ecology 84:449-459. Karron, J. D. 1987. A comparison of levels of genetic polymoiphism and self-compatibility in geographi- cally restricted and widespread plant congeners. Evo- lutionary Ecology 1 :47-58. , Y. B. Linhart, C. A. Chalilk, and C. A. Rob- ertson. 1988. Genetic structure of populations of geographically restricted and widespread species of Astragalus (Fabaceae). American Journal of Botany 75:1 1 14-1 1 19. Ledig, F. T. and M. T. Conkle. 1983. Gene diversity and genetic structure in a narrow endemic, Torrey pine {Finns torrevana Parry ex. Carr). Evolution 37:79- 85. Lesica, p., R. F. Leary, R. W. Allendorf and D. E. Bild- erback. 1988. Lack of genetic diversity within and among populations of an endangered plant, Howelia aquatilis. Conservation Biology 2:275-282. Levy, M. and D. A. Levin. 1975. Genie heterozygosity and variation in permanent translocation heterozy- gotes of the Oenothera biennis complex. Genetics 79: 493-512. Lewis, P. O. and D. J. Crawford. 1995. Pleistocene re- fugium endemics exhibit greater allozymic diversity than widespread congeners in the genus Polygonella (Polygonaceae). American Journal of Botany 82:141- 149. , and R. Whitkus. 1989. GENESTAT-PC. Amer- ican Society of Plant Taxonomists Newsletter 2:15- 16. LiSTON, A., L. H. Rieseberg and T. S. Elias. 1989. Ge- netic similarity is high between intercontinental dis- junct species of Senecio ( Asteraceae). American Jour- nal of Botany 76:383-388. Loveless, M. D. and J. L. Hamrick. 1984. Ecological determinants of genetic structure in plant populations. Annual Review of Ecology and Systematics 15:65- 95. Mashburn, S. J., R. R. Sharitz and M. H. Smith. 1978. Genetic variation among Typha populations of the southeastern United States. Evolution 32:681-685. MossELER, A., D. J. iNNES AND B. A. ROBERTS. 1991. Lack of allozymic variation in disjunct Newfoundland pop- ulations of red pine {Finns resinosa). Canadian Jour- nal of Forest Research 21:525-528. , K. N. Egger and G. a. Hughes. 1992. Low lev- els of genetic diversity in red pine confirmed by ran- dom amplified polymorphic DNA markers. Canadian Journal of Forest Research 22:1332-1337. Nei, M. 1973. Analysis of gene diversity in subdivided populations. Proceedings of the National Academy of Sciences. USA 70:3321-3323. Novak, S. J., R. N. Mack and D. E. Soltis. 1991. Genetic variation in Bromus tectoruni (Poaceae): population differentiation in its North American Range. Ameri- can Journal of Botany 78:1 150-1 161. Rabinowitz, D. 1981. Seven forms of rarity, p. 205-217. in H. Synge (ed.). The biological aspects of rare plant conservation. John Wiley and Sons, Chichester, En- gland. RoosE, M. L. AND L. D Gottlieb. 1976. Genetic and bio- chemical consequences of polyploidy in Tragopogon. Evolution 30:818-830. ScHMiD, B. 1984a. Life histories in clonal plants of the Carex flava group. Journal of Ecology 72:93-1 14. . 1984b. Niche width and variation within and be- tween populations in colonizing species {Carex flava group). Oecologia 63:1-5. . 1986. Colonizing plants with persistent seeds and persistent seedlings {Carex flava group). Botanica Helvetica 96:19-26. 158 MADRONO [Vol. 47 ScHWAEGERLE, K. E. AND B. A. ScHAAL. 1979. Gcnctic variability and founder effect in the pitcher plant Sar- racenia purpurea L. Evolution 33:1210-1218. Schwartz, A. O. 1985. Lack of protein polymorphism in the endemic relict Chrysosplenium iowense (Saxifra- gaceae). Canadian Journal of Botany 63:2031-2034. Sharitz, R. R., S. a. Wineriter, M. H. Smith and E. H. Liu. 1980. Comparison of isozymes among Typha species in the eastern United States. American Jour- nal of Botany 67:1297-1303. Simon, J.-R, Y. Bergeron and D. Gagnon. 1986. Isozyme uniformity in populations of red pine (Pinus resinosa) in the Abitibi region, Quebec. Canadian Journal of Forest Research 16:1133-1135. SoLTis, D. E. 1982. Allozymic variability in Sullvantia (Saxifragaceae). Systematic Botany 7:26-34. , C. H. Haufler, D. C. Darrow and G. J. Gas- tony. 1983. Starch gel electrophoresis of ferns: a compilation of grinding buffers, gel and electrode buffers, and staining schedules. American Fern Jour- nal 73:9-27. SoLTis, P. S., D. E. SoLTis, T. L. Tucker and F A. Lang. 1992. Allozyme variability is absent in the narrow endemic Bensoniella oregona (Saxifragaceae). Con- servation Biology 6:131-134. Spackman, S., B. Jennings, J. Coles, C. Dawson, M. Minton, a. Kratz and C. Spurrier. 1997. Colorado Rare Plant Field Guide. Prepared for the Bureau of Land Management, U. S. Forest Service and U. S. Fish and Wildlife Service by the Colorado Natural Heritage Program, Ft. Collins, CO. Standley, L. a. 1990. Allozyme evidence for the hybrid origin of the maritime species Carex salina and Car- ex recta (Cyperaceae) in Eastern North America. Sys- tematic Botany 15:182-191. Stebbins, G. L. 1950. Variation and evolution in plants. Columbia University Press, New York, N.Y SwoFFORD, D. L. AND R. B. Selander. 1981. BIOSYS-1. A Fortran program for the comprehensive analysis of electrophoretic data in population genetics and sys- tematics. Journal of Heredity 72:281-283. Van Oostrum, H., A. A. Sterk and H. J. W. Wijsman. 1985. Genetic variation in agamospermous microspe- cies of Taraxacum sect. Erythrosperma and sect. Ob- liqua. Heredity 55:223-228. Waller, D. M., D. M. O'Malley and S. C. Gawler. 1987. Genetic variation in the extreme endemic Ped- icularis furbishiae (Scrophulariaceae). Conservation Biology 1:335-340. Waterway, M. J. 1990. Genetic differentiation and hy- bridization between Carex gynodynama and C. men- docinensis (Cyperaceae) in California. American Journal of Botany 77:826-838. Whitkus, R. 1992. Allozyme variation within the Carex pachystachya complex (Cyperaceae). Systematic Bot- any 13:146-153. Wolff, S. L. and R. L. Jefferies. 1987. Morphological and isozyme variation in Salicornia europaea (s.l.) (Chenopodiaceae) in northeastern North America. Canadian Journal of Botany 65:1410-1419. Yeh, F C. and C. Layton. 1979. The organization of genetic variability in central and marginal populations of lodgepole pine {Pinus contorta spp. latifolia). Ca- nadian Journal of Genetics and Cytology 21:487-503. Madrono, Vol. 47, No. 3, pp. 159-163, 2000 THE TAXONOMIC HISTORY, IDENTITY, AND DISTRIBUTION OF THE NEVADA ENDEMIC, PLAGIOBOTHRYS GLOMERATUS (BORAGINACEAE) Arnold Tiehm 1550 Foster Dr., Reno, NV 89509 Abstract Plagiobothrys glomeratus A. Gray is a western Nevada endemic restricted to areas of altered andesite. It is morphologicaly close to P. hispidus A. Gray. Plagiobothrys hispidus is more widespread and shows much variation as to number and size of nutlets. Reports of P. glomeratus from California are based on misidentifications of P. hispidus. Illustrations of the nutlets and a distribution map of both species are included. Current floristic studies in the Pine Nut Moun- tains of western Nevada uncovered a problem in defining Plagiobothrys glomeratus A. Gray. This study was undertaken to clarify the identity and dis- tribution of P. glomeratus. Gray (1885) described Plagiobothrys glomeratus from two collections by Katharine Curran (later Brandegee) taken between Virginia City and Car- son City, Nevada. In the same article Gray de- scribed P. hispidus A. Gray based on a collection, again taken by Curran, from the streets of Truckee in nearby California. These two sites are approxi- mately 27 air miles apart. Both taxa are members of section Plagiobothrys characterized by alternate leaves, lateral nutlet scars placed near or above the center of the nutlets, and not growing in seasonally saturated soils. They share the characteristic of hav- ing rather broad upper cauline leaves with P. jonesii A. Gray and P. kingii (S. Watson) A. Gray. Both P. kingii and P. jonesii have elongated nutlet scars along the ventral keel and an earlier spring flow- ering time in contrast with nutlet scars about as wide as long, placed at the end of the ventral keel, and a later spring to summer flowering time in P. hispidus and P. glomeratus. As such P. hispidus and P. glomeratus are more similar to each other than to any other species. The first treatment of P. glomeratus is that of Greene (1887) who described the genus Sonnea to accomodate P. glomeratus, hispidus, jonesii, and kingii. He later described Sonnea foliacea from the geographic area between the type localities of P. glomeratus and P. hispidus (Greene 1888). Johnston (1923) published a synopsis of Pla- giobothrys placing glomeratus and hispidus in his Sonnea group and kingii and jonesii in his Amsinck- iopsis group. He also reduced Greene's Sonnea fol- iacea to a variety of P. hispidus and stated: "It is possible that the plant is a hybrid between P. his- pidus and P. glomeratus."" Tidestrom (1925) in his Flora of Utah and Ne- vada recognized the distinctiveness of P. hispidus and P. glomeratus but followed Greene in placing them in the genus Sonnea. He also maintained S. foliacea as a good species. Cronquist (1984) recognized P. glomeratus as an acceptable species with the comment "Reno south nearly to Carson City, rarely collected." The only other published references for the distribution of P. glomeratus in Nevada are from the south side of Peavine Mountain where it is reported as occurring in an open pine stand (Billings 1992; Williams et al. 1992). Plagiobothrys glomeratus is not included or mentioned in any flora covering California (Abrams 1951; Jepson 1925, 1943; Messick 1993; Munz 1968; Munz, and Keck 1959). Since none of the above references mention P. glomeratus or place it in synonymy I assume they did not have any evi- dence to believe it occurred in California. DeDecker (1990) reported P. glomeratus as new to California. Her records are Sweetwater Moun- tains, above Star City, DeDecker 5677 (RSA!) and Sierra Nevada, "The Bluffs," 0.6 miles NNE of Mammoth Rock, Bagley 3001 (personal herbarium of Mark Bagley, Bishop, California!). A check with Roxanne Bittman of the California Natural Diver- sity Data Base, in Sacramento, CA, revealed no other known specimens from California. The DeDecker and Bagley records are the basis for in- cluding P. glomeratus in the California Native Plant Society inventory of rare and endangered vas- cular plants of California (Skinner and Pavlik 1994). I find both of these specimens to be P. his- pidus. The misidentifications likely come from the lack of understanding of P. hispidus not from the true nature of P. glomeratus. Few California ref- erences provide the nutlet size for P. hispidus. Munz and Keck (1959) and Abrams (1951) fist the size as 1 mm while Messick (1993) lists the size as 1-1.5 mm. Cronquist (1959, 1984) twice has dealt with P. hispidus and his descriptions are es- sentially the same. The one sUght difference is nut- let length, 1-2 mm in 1959 and 1-2 (2.5) in 1984. I have found that the nutlets of Plagiobothrys hispidus vary in the number that mature. At the north end of its range many plants have four ma- 160 MADRONO [Vol. 47 turing nutlets while at the south end one or two is the norm. The number of maturing nutlets greatly influences their orientation, shape, and size. If four nutlets mature they are vertically oriented, less than 2 mm long, have a definite dorsal keel, and are unevenly tuberculate or rugose-tuberculate (see Fig. 1, illustration A). This is the nutlet type illustrated in Cronquist (1959, 1984) and represented by the type collection of P. hispidus. When one or two nutlets mature they are horizontally oriented, up to 2.4 mm long, flat-backed with a more obscure keel, the end farthest from the scar is greatly expanded, and the roughness is more evenly appressed and not as evident. It is this nutlet type that is represented by DeDecker 5677 and Bagley 3001, the basis of the reports of P. glomeratus from California, and by the type of Sonnea foliacea (see Fig. 1, illustra- tion C). On the other hand P. glomeratus is ex- tremely uniform with larger, mottled, shiny nutlets, and is edaphically restricted (see Fig. 1, illustration D). I can see how one could be misled in trying to identify the California specimens. The nutlet size does not fit the descriptions in Munz and Keck (1959) or Messick (1993). In checking Intermoun- tain Flora (Cronquist 1984) the illustration of P. hispidus is that of the smaller four nutlet type. The broad fat-ended illustration of P. glomeratus then becomes the logical choice. I agree with Cronquist (1984) in placing Sonnea foliacea in synonymy with P. hispidus. The ex- tremes seem distinctive but all stages of interme- diacy occur. For instance, many collections from the Truckee, CA area contain plants with nutlets of both the hispidus and foliacea type. Gray (1885) in describing P. glomeratus de- scribes its distributions as: "Western part of Ne- vada, between Carson and Virginia City, 1883 and 1884, Mrs. Layne-Curran."' There are two sheets in the Gray Herbarium that fit Gray's protologue. One is labeled "Geiger Grade, Aug. 1883, Curran s.n.'' and the other "between Carson and Virginia, [un- dated], Curran s.n.'" Selection of the "Carson to Virginia" sheet as a lectotype was effectively done by Cronquist (1984). The results of this study indicate that Plagiob- othrys glomeratus is a western Nevada endemic re- stricted to areas of altered andesite between 4860 and 6650 ft in elevation. These altered andesite ar- eas have shallow azonal soils nearly totally lacking in nutrients and with an acidic pH (3.7-4.0) (Bill- ings 1992). Soils are so nutrient poor that they are not able to support the ubiquitous sagebrush, Ar- temisia tridentata Nutt., or other shrubs in any number (Billings 1950, 1992). This lack of com- petition from shrubs has allowed relic stands of Si- erran conifers to persist in isolated pockets. The altered andesite areas are orangish light-brown in surface color and are dotted with dark green coni- fers. As such they are a conspicuous feature on the hills around Reno (Billings 1950, 1992). Although concentrated in the Reno area there are outliers of altered andesite as far northeast as the Pah Rah Range, east to Ramsey in the Virginia Range, and south to the Sweetwater and White Mountains of California and P. glomeratus may eventually be found at some of these sites (Billings 1992). Plagiobothryjs glomeratus is known from the Virginia Range in Storey and Washoe Counties, Carson Range of the Sierra Nevada, foothills north of Reno, and from nearby Peavine Mountain, all in Washoe County (see Fig. 2). Its distribution nearly matches that of the only other known altered an- desite endemic, Eriogonum robustum E. L. Greene (type also collected by Curran in 1884). Both occur less than six miles from California and eventually may be found there. Searches in the Truckee River canyon west of Reno, and near Markleeville south- southwest of Gardnerville, have so far proved fruit- less. Plagiobothry's hispidus occurs from south-central Oregon south and east through the eastern Sierra Nevada of California and Nevada to the Mammoth area in Mono and adjacent Madera Counties (see Fig. 2). There are outliers on Steens Mountain, Har- ney County, Oregon, Skeedaddle Mountain, Lassen County, California, Granite Range, Washoe County, Nevada, Pine Nut Mountains, Douglas County, Ne- vada, and the Masonic Hills and Sweetwater Moun- tains in Mono County, California. Key to plagiobothrys glomeratus and hispidus Nutlets smooth and shiny, mottled, 2.4-3.0 mm long, horizontally oriented P. glomeratus Nutlets unevenly tuberculate to pavemented with the roughness always readily discernable, up to 2.4 mm long, horizontally or vertically oriented P. hispidus Specimens of Plagiobothrys glomeratus examined, all from Nevada STOREY CO., Virginia Range, Six Mile Can- yon, 4.2 road miles E of highway 341, Tiehm 12544 (BRY, CAS, NY, OSC, RENO, RSA, UC, UTC); Virginia Range, 1.1 road miles SE of N junction of highways 341 and 342 on highway 341, Tiehm 12542 (ARIZ, BRY, CAS, GH, MONT, NY, OSC, RENO, RM, RSA, UC, UNLV, UT, UTC, WS): WASHOE CO., Dandini Blvd. N of Reno, Nachlin- ger 1375 (NY), Tiehm & Kelley 12522 (CAS, NY, OSC, RENO, UC, UTC); west slopes of Peavine Mountain., Nachlinger & Billings 1374 (NY); hiU east of Black Panther Mine, 3 miles N of Reno, Billings 1296 (RENO); Geiger Grade, Jul 1884, Curran s.n. (DS); Geiger Grade to Virginia City, Eastwood 14809 (CAS); Geiger Grade, Aug 1883, 2000] TIEHM: PLAGIOBOTHRYS GLOMERATVS 161 1 mm Fig. 1. A-C are nutlets of Plagiobothrys hispidus, D is nutlets of Plagiohothrys gloinenitus. A is drawn from Steward 6798, Deschutes Co., OR (NY); B from Sonne s.n., Truckee, Nevada Co., CA (NY); C from Tiehm 12244, Pine Nut Mountains, Douglas Co., NV (RENO); and D from Eastwood 14809, Geiger Grade to Virginia City, Storey Co., NV (CAS). 162 MADRONO [Vol. 47 Fig. 2. Map showing parts of Oregon, Idaho, Nevada, and California. The distribution of Plagiobothrys glomeratus is designated by solid circles and the distribution of P. hispidus is designated by open circles. Curran s.n. (GH); Virginia Range, Geiger Grade, 2.8 road miles E of highway 395 on highway 341, Tiehm 12540 (ARIZ, ASU, B, BRY, CAS, COLO, CS, DAO, GH, ID, K, KSC, LE, MICH, MO, MONT, MONTU, NY, OKL, OS, OSC, RENO, RM, RSA, SI, TEX, UC, UNLV, UTC, WIS, WS, WTU); Virginia Range, foothills E of the S end of Hidden Valley County Park, Tiehm 12547 (CAS, NY, OSC, RENO, RM, RSA, UC, UTC, WTU); Sierra Nevada, Carson Range, ridge on N side of N fork of Evans Creek, Tiehm 12548 (BRY, CAS, MICH, MO, MONT, NY, OSC, RENO, RM, RSA, UC, UNLV, UTC); Sierra Nevada, Carson Range, ridge divide between Hunter and Alum Creeks, Tiehm 12593 (CAS, NY, OSC, RENO, UC): COUNTY UNKNOWN, Nevada between Carson & Virginia, [undated], Curran s.n. (GH lectotype). Acknowledgments I am grateful to the following people for supplying me with advice and information: Mark Bagley, Janet Bair, Roxanne L. Bittman, Kenton L. Chambers, Ron Kelley, Sahy Manning, James D. Morefield, Janet L. Nachlinger, Kathleen Nelson, and Margriet Wetherwax. The map was prepared by Brian McMenamy. Comments from two anonymous reviewers added substantially to this work. I am also grateful to the curators of the following herbaria for access to their collections, for loans, or both: CAS, DS, GH, JEPS, NDG, NY, ORE, OSC, RENO, RSA, UC. Kathryn (Kay) Corbett aptly illustrated nutlets of P. glom- eratus and P. hispidus. Literature Cited Abrams, L. 1951. Illustrated flora of the Pacific States. 3: 1-866. Stanford University Press, Stanford, CA. Billings, W. D. 1950. Vegetation and plant growth as af- fected by chemically altered rocks in the western Great Basin. Ecology 31:62-74. . 1992. Islands of Sierran plants on the arid SLOPES OF Pea vine Mountain. Mentzelia 6, part 1:32-39. Cronquist, a. 1959. Boraginaceae, pp. 175-244. in C. L. Hitchock, A. Cronquist, M. Ownbey, and J. W. Thompson, Vascular Plants of the Pacific Northwest, Part 4. Univ. Wash. Publ. Biol. 17(4): 1-5 10. . 1984. Boraginaceae. Pp 207-293 in A. Cronquist, A. H. Holmgren, N. H. Holmgren, J. L. Reveal, and P. K. Holmgren, Intermountain Flora 4:1-573. New York Botanical Garden, Bronx, N.Y. 2000] TIEHM: PLAGIOBOTHRYS GLOMERATUS 163 DeDecker, M. 1990. Additions to our flora. California Native Plant Society Bristlecone Chapter Newsletter 9(6):3. Gray, A. 1885. Contributions to the botany of North America. Proc. Amer. Acad. Arts. 20:257-310. Greene, E. L. 1887. Some west American Asperifoliae. Pittonia 1:8-23. . 1888. New or noteworthy species 111. Pittonia 1: 215-225. Jepson, W. L. 1925. A manual of the flowering plants of California. University of California Press, Berkeley. . 1943. A flora of California. 3, part 11:129-464. Associated Students Store, University of California, Berkeley. Johnston, 1. M. 1923. Studies in the Boraginaceae. 4. A synopsis and redefinition of Plagiohothr-ys. Contr. Gray Herb. 68:57-80. Messick, T. C. 1993. Plagiohothrys, Pp. 386-390. //; J. C. Hickman (ed.). The Jepson Manual. Higher plants of California. University of California Press, Berkeley. MuNZ, P A. 1968. Supplement to A California Flora. Uni- versity of California Press, Berkeley. , and D. D. Keck. 1959. A California Flora. Uni- versity of California Press, Berkeley. Skinner, M. W. and B. M. Pavlik. 1994. Inventory of rare and endangered vascular plants of California. California Native Plant Society Special Publication No. 1, 5th ed. CNPS, Sacramento. Tidestrom, 1. 1925. Flora of Utah and Nevada. Contr. U.S. Natl. Herb. 25:1-665. Williams, M. J., J. T. Howell, G. H. True, Jr., and A. TiEHM. 1992. A catalogue of vascular plants on Peav- ine Mountain. Mentzelia 6, part 2:3-83. Madrono, Vol. 47, No. 3, pp. 164-173, 2000 GENETIC VARIATION IN PINUS PONDEROSA, PURSHIA TRIDENTATA, AND FESTUCA IDAHOENSIS, COMMUNITY-DOMINANT PLANTS OF CALIFORNIA'S YELLOW PINE FOREST Safiya Samman,', Barbara L. Wilson- and Valerie D. Hipkins^ 'USDA Forest Service, FHP, AB-2S, PO. Box 96090, Washington, D.C. 9009-6090 ^Department of Botany and Plant Pathology, 2064 Cordley Hall, Oregon State University, Corvalhs, OR 97331 3USDA Forest Service-NFGEL, 2375 Fruitridge Road, Camino, CA 95709 Abstract Genetic diversity of Finns ponderosa Laws., Piirshia tridentata (Pursh) DC, and Festuca idahoensis Elmer at Black's Mountain Experimental Forest in northwest California was evaluated using isozymes. This tree, shrub, and grass are all common, outcrossing, long-lived perennials that dominate their respec- tive layers of the same plant community. Genetic analyses were provided for diploid Finus ponderosa and Furshia tridentata. A phenotypic analysis of isozyme band patterns was provided for tetraploid F. idahoensis and, for comparison, previous reports of fescue isozyme variation were reanalyzed using this method. Finns ponderosa, Furshia tridentata, and Festuca idahoensis were highly genetically variable, with 75% to 92% polymorphic loci. For all three species, more than 90% of the genetic variation occurred within, rather than among, populations. This study compares genetic diversity in plant species of three life forms, while holding constant habitat, breeding system, and community domi- nance. The three plant species chosen for study are Pinus ponderosa Laws, Purshia tridentata (Pursh) DC, and Festuca idahoensis Elmer They represent three life forms, tree, shrub, and grass, respectively. All three are common, widespread, outcrossing, long-lived perennials. All dominate their respective layers in the plant community at the study site. They do have life history differences; F. idahoensis is insect pollinated while the other two species are wind pollinated, and F. idahoensis is tetraploid while the others are diploid. The three species af- fect one another in a complex web of competitive and commensal relationships (e.g.. Baron et al. 1966; Busse et al. 1996; Hall et al. 1995; vander Wall and vander Wall 1992). The study site. Black's Mountain Experimental Forest, was established in 1934 in the Lassen Na- tional Forest, Lassen County, California. More than 60 y of experimentation and careful record keeping make the 4050-ha forest a uniquely valuable re- source for investigating the effects of different tim- ber management practices on eastside pine type for- ests. In 1993 the Black's Mountain Interdisciplinary Research Program was established to study the ef- fects of forest management on various ecosystem components including vertebrates, insects, soil or- ganisms, and vegetation. This study provides base- Corresponding and proofs author: Valerie D. Hipkins, USDA Forest Service-NFGEL, 2375 Fruitridge Road, Camino, CA 95709. Phone: 530 642-5067. Fax: 530 642- 5093. E-mail: vhipkins@fs.fed.us line data for a long-term study of effects of silvi- cultural treatments on genetic biodiversity. Genetic and species biodiversity are elements of a healthy ecosystem. Little is known about the effects of for- est management on the genetics of forest plants, although some forestry practices can profoundly in- fluence tree genetics (Adams et al. 1998). Four plots similar in topography and vegetation were chosen for this study (Table 1). Genetic vari- ation in the three selected species was sampled in 1994 and 1995. Subsequently, three silvicultural treatments (a timber cutting regime, fire, and graz- ing) were applied to the plots (Table 1). Genetic diversity will be resampled in five to twenty years, to detect any effects from the silvicultural practices initiated in 1995. Methods In 1993, plots were chosen for an intensive, mul- tidisciplinary study of the effects of management practices on the entire ecosystem. Midpoints of the four plots were 2.2 to 4.0 km apart. These plots have been treated similarly in the past, and all were grazed lightly until 1996, when some were fenced to exclude cattle (Table 1). All four plots sampled in the genetics study consisted of dry forest domi- nated by P. ponderosa, and all plots were similar (Table 1). Half of each plot was subsequently burned. The presence of small, unburned, long-term control plots in some burned split plots is ignored in this analysis. Sample collection. Permanent markers were es- tablished on a 100-m grid within each plot. In each plot, fifty grid points were randomly selected as I I 2000] SAMMAN ET AL.: GENETICS OF PONDEROSA PINE, BITTERBRUSH, AND IDAHO FESCUE 165 Table 1. Characteristics and Management Plans of the Four Black's Mountain Plots Sampled in This Study. Pine trees are mostly Ponderosa Pine with some Jeffrey Pine; fir trees are White Fir, Abies concolor (Gordon and Glend.) Lindley. Perennial grass cover was predominantly Idaho Fescue cover. Structural vertical diversity is a man- agement practice imposed by thinning two of the plots to reduce structural diversity. Half of each plot was subsequently burned; burned and unburned halves are labeled "B'' and ''N," respectively, in other tables. Information on plot vegetation from W. W. Oliver (unpublished data). Plot Elevation (meters) Area (hectares) Live trees/hectare Pine Fir Idaho Fescue (frequency) Perennial grasses (foliar cover) Structural Vertical Diversity Grazing 38 523-540 136 440 9 59% 4% high no 39 526-543 120 333 0 77% 6% low no 41 575-580 108 440 16 22% 2% high yes 43 570-576 109 364 170 40% 2%' low yes collection sites. All three species were collected at each grid point, if all three were present. If one or more was missing, a replacement sample was col- lected at a different, ranodmly selected grid point. Finns ponderosa: Cones were collected from the tree nearest each selected grid point (207 total in- dividuals). Samples were collected in September 1994. Purshia tridentata: At each selected grid point, two individuals were flagged and sampled (404 to- tal individuals). The shrub closest to the grid point was the first sampled individual, and the second was the closest shrub on the opposite side of the gridpoint, on a line drawn from the first shrub through the grid point. A leafy shoot was cut from each individual, wrapped in wet paper towels, placed in a plastic bag, and stored on ice in the field. Samples were collected in June and July 1995. Festuca idahoensis: Near each flagged F. triden- tata shrub, a fescue individual was sainpled (385 total individuals). (If fescue was sparse, the fescue might be as much as 30 meters away from the F. thdentata.) If no fescue grew within 30 meters of a sampling point, another point was randomly se- lected for fescue sampling. From each fescue, a small rooted plug with about 20 leaves was col- lected, wrapped in wet paper towels, placed in a plastic bag, and stored on ice in the field. Samples were collected in July 1995. Sample preparation. Samples were prepared us- ing NFGEL standard operating procedures (Anon- ymous 1995). For F. ponderosa, seeds were ger- minated and the megagametophytes were ground in a 0.2 M phosphate buffer, pH 7.5. The slurry was absorbed onto 3 mm wide wicks prepared from Whatman 3MM chromatography paper, and stored at — 70°C. Furshia tridentata and F. idahoensis leaf tissue was ground in a Tris buffer pH 7.5 (Gotdieb 1981); liquid nitrogen was used to freeze F. triden- tata tissue before grinding. One hundred fifty mi- croliters of slurry per sample was transferred into each of two microliter plate wells, and plates were frozen at — 70°C. For electrophoresis, the slurry was thawed and absorbed onto wicks. Electrophoresis. Methods of electrophoresis are outlined in Anon. (1995), and follow the general methodology of Conkle et al. (1982) except that most enzyine stains are modified. The following en- zymes were examined: aconitase (ACO), catalase (CAT), diaphorase (DIA), florescent esterase (FEST), fructose- 1 ,6-diphosphate dehydrogenase (FDP), glutamate-oxaloacetate transaminase (GOT), glucose-6-phosphate dehydrogenase (G6PDH), glycerate-2-dehydrogenase (GLYDH), isocitrate dehydrogenase (IDH), leucine aminopep- tidase (LAP), malate dehydrogenase (MDH), malic enzyme (ME), phosphoglucomutase (PGM), phos- phogluconate dehydrogenase (6PGD), phosphog- lucose isomerase (PGI), shikimic acid dehydroge- nase (SKD), triosephosphate isomerase (TPI), and uridine diphosphoglucose pyrophosphorylase (UGPP). The enzymes examined, and the buffer systems used to resolve them, varied according to species (Table 2). All enzymes were resolved on 1 1% starch gels. Enzyme stain recipes for enzymes follow Anonymous (1995) except that GOT was stained using the recipe from Wendel and Weeden (1989). Two people independently scored each gel. When they disagreed, a third person resolved the conflict. For quality control, 10% of the individuals were run and scored twice. Both F. ponderosa and the morphologically sim- ilar Jeffrey Pine F. jejfreyi Grev. and Balf (Jeffrey Pine) occur in the research plots (Oliver, MS). For a few of the pine samples, alleles of multiple en- zymes differed from those seen before in F. pon- derosa (NFGEL, unpublished data). Such samples were omitted from analysis on the assuinption that they were F. jejfreyi. Finns ponderosa and F. tridentata are diploid. Most studied isozymes are known to show Men- delian inheritance in F. ponderosa (Linhait et al. 1989; O'Malley et al. 1979), but no such informa- tion is available for F. tridentata. Genetic interpre- tations were inferred directly from isozyme phe- 166 MADRONO [Vol. 47 Table 2. Isozymes Examined for Pinus ponderosa, Purshia tridentata, and Festuca idahoensis. LB = a lithium borate electrode buffer (pH 8.3) used with a Tris citrate gel buffer (pH 8.3) (Conkle et al. 1982). SB = a sodium borate electrode buffer (pH 8.0) used with a Tris citrate gel buffer (pH 8.8) (Conkle et al. 1982). MC6 = a morpholine citrate ■ electrode and gel buffer (pH 6.1) (Conkle et al. 1982). MC8 = a morpholine citrate electrode and gel buffer (pH 8.0) > (Anon. 1985). na = number of alleles per locus, ne = effective number of alleles per locus (Kimura and Crow 1964). np = number of patterns per locus. For enzymes abbreviations, see text. Numbers in isozyme abbreviations refer to different regions on gels, interpreted as different loci. Buffer: T R na ne na ne na ne na ne PiNUS PONDEROSA ACO 4 2.8 CAT 1 1.0 FEST 7 1.5 DIA 4 2.9 16 enzymes ADH 2 1.9 GOTl 4 1.1 MDHl 3 1.0 FDP 1 1.0 26 loci LAPl 3 1.1 GOT2 5 1.1 MDH2 3 1.9 IDHl 3 1.7 LAP2 4 1.1 GOT3 6 1.3 MDH3 4 1.1 IDH2 5 1.2 ME 4 1.2 G6PDH 4 1.1 6PGD1 5 2.0 pni 1 ryji 1 2 1 0 ■x 1 0 A 1 1 i . 1 PGI2 3 1 1 UGPP2 5 3.0 PGM 1 4 1 .9 PGM2 3 2.3 FEST 1 1.0 CAT 2 1.0 DIA 2 1.0 13 enzymes LAP 3 2.6 GOT 4 2.6 FDP 3 1.0 16 loci PGIl 1 1.0 6PGD 4 1.1 IDH 1 1.0 PGI2 3 1.0 TPIl 2 1.0 MDHl 2 1.1 PGM 3 2.0 TP12 1 1.0 MDH2 3 1.0 UGPP 3 1.9 Festuca idahoensis np np np ME 1 CAT 1 DIA 6 1 1 enzymes PGI 17 GLYDH 3 SKD 4 treated as 12 loci PGM 3 GOTl 3 GOT2 4 6PGD 2 TPI 4 UGPP 16 notypes, based on knowledge of the generally con- served enzyme substructure, compartmentalization, and isozyme number in higher plants (Gottlieb 1981, 1982; Weeden and Wendel 1990). Genotypes of P. ponderosa were inferred from the segregation patterns of 10 megagametophytes per tree. Festuca idahoensis has a chromosome number 2n = 28 and is thought to be autotetraploid (Dar- lington and Wylie 1955). Because of the compli- cated banding patterns observed, and because of lack of crossing studies to determine the inheritance of bands in this species, we were unable to identify specific alleles and loci for some enzymes, includ- ing highly variable MDH, PGI and UGPP. There- fore, a phenotypic instead of genotypic analysis was performed. Data analysis. In order to allow future compar- ison with genetic diversity after all silvicultural techniques (logging, grazing, and burning) are im- plemented, the data were analyzed in terms of eight plots, which represent the half plots subsequently burned (Table 3). For P. ponderosa and P. tridentata, results were analyzed using Popgene version 1.21 (Yeh et al. 1997). A locus was considered polymorphic if an alternate allele occurred at least once. Statistics cal- culated included unbiased genetic distances (Nei 1978), expected heterozygosity (Nei 1973), inferred gene flow (Nm = 0.25(1 /FJ/F,,; Slatkin and Barton 1989), and F statistics (F = (He-HJ/H^; Hartl and Clark 1989). The gene diversity statistics H^, (ex- pected heterozygosity at the species level) and H^p (expected heteozygosity at the population level) were calculated (Hamrick and Godt 1990). For F. idahoensis, phenotypic diversity measures were calculated from both band presence/absence and multi-band patterns. For presence/absence data, phenotypic diversity was measured by a polymor- phic index (PI), based on the frequency of occur- rence of each band. PI = the sum of f(l-f), where f = the frequency of a band in a population (Chung et al. 1991). For multi-band patterns, phenotypic diversity measures include: (1) the number of bands found in each plot, (2) percent of stains that yield more than one band pattern, (3) the average number of band patterns per stain in each plot, and (4) Shannon-Weaver Diversity Index values (Shannon and Weaver 1949). The Shannon-Weaver Diversity Index uses the frequency of each band pattern in each plot. The larger the Shannon-Weaver Index, 2000] SAMMAN ET AL.: GENETICS OF PONDEROSA PINE, BITTERBRUSH, AND IDAHO FESCUE 167 Table 3. Summary of Genetic Diversity Measures in Pinus ponderosa and Purshia tridentata by Species and by Plot. All = overall statistics for the study. Mean = average over the 8 plots. SD = standard deviation. N = mean number of individuals sampled per locus, per population. %P = percent of all loci that are polymorphic. A = average number of alleles per locus. Ap = the average number of alleles per polymorphic locus. H,, = observed frequency of heterozygotes. = frequency of heterozygotes expected under Hardy-Weinberg equilibrium conditions. F = the fixation index, = (H^.-H,,)/H,, (* = statistically significant difference (p < 0.05)). S-W = Shannon-Weaver diversity index. B and N following plot numbers indicated the half plots that were subsequently burned or not burned. N %P A Ap H, F S-W Pinus ponderosa (26 loci) All 207 92% 3.73 3.96 0.264 0.272 0.030 0.510 Mean 26 85% 2.79 3.09 0.265 0.271 0.021 0.483 SD 2.4 4.0 0.11 0.10 0.021 0.010 0.082 0.019 38B 24 88% 2.73 2.96 0.263 0.269 0.023 0.477 38N 26 81% 2.73 3.14 0.251 0.264 0.046 0.470 39B 27 92% 3.04 3.21 0.283 0.287 0.012 0.517 39N 29 85% 2.81 3.14 0.240 0.277 0.133 0.495 41B 26 85% 2.73 3.04 0.280 0.280 0.002 0.495 41N 23 81% 2.77 3.19 0.257 0.255 -0.010 0.455 43B 23 88% 2.73 2.96 0.301 0.264 -0.140 0.467 43N 29 81% 2.77 3.09 0.244 0.272 0.104 0.489 Purshia tridentata (16 loci) All 404 75% 2.38 2.83 0.138 0. 1 57 0.1 17* 0.263 Mean 50 55% 1.81 2.47 1.138 0.156 0.113 0.254 SD 8.0 9.6 0.12 0.26 0.006 0.005 0.049 0.01 1 38B 50 50% 1.69 2.38 0. 1 35 0.148 0.087 0.235 38N 50 56% 1.69 2.22 0.134 0.152 0.120* 0.247 39B 62 38% 1.75 3.00 0.140 0.156 0.101 0.254 39N 38 50% 1.75 2.50 0.138 0.152 0.091 0.246 41B 52 69% 1.94 2.36 0.130 0.155 0.165 0.260 41N 50 62% 2.00 2.60 0.147 0.157 0.059 0.262 43B 42 52% 1.75 2.20 0.132 0.166 0.204* 0.269 43N 60 62% 1.94 2.50 0.147 0. 1 59 0.076 0.261 the more diverse the plot. The distribution of the total variation within and among plots was deter- mined by partitioning the total Shannon-Weaver Diversity Index. The phenotypic relationships among plots were determined by calculating Hed- rick's phenotypic identities (Hedrick 1971) for mul- ti-band pattern data, and by cluster and principle coordinate analyses of Jaccard's Similarity Index for band presence/absence data (Chung et al. 1991; Rolf 1987). The hypothesis that all the plots had equal ge- netic diversity by species was tested by ANOVA using Excel (Microsoft 1997). In one analysis, plots were treated as blocks and the burned and unburned half plots as samples. In a second analysis, burned and unburned half plots were treated as blocks and the plots were treated as samples. Results Pinus ponderosa, P. tridentata, and F. idahoen- sis were all genetically variable (Tables 3 and 4). Both percent polymorphic loci/enzyme and the Shannon-Weaver diversity index indicate that P. tridentata is the least variable of the three species. In the two diploid species for which they could be calculated, observed heterozygosity nearly equaled that expected under Hardy-Weinberg con- ditions. Therefore, the fixation index (F) within each plot, calculated for P. ponderosa and P. tri- denta, was low (F < 0.113) (Table 3). Although heterozygosity could not be calculated for F. ida- hoensis, we observed the unequal band staining and complex band patterns characteristic of heterozy- gous tetraploids (Soltis and Riesbeig 1986). F„ values indicated that over 98% of the isozyme variation in P. ponderosa and P. tridentata was within, rather than between populations, and in- ferred gene flow was high (Table 5). G,, a measure of interpopulation diversity analogous to F,t but cal- culated from the Shannon-Weaver diversity index, was somewhat higher than the corresponding val- ues of F^t, but indicated that in all three species more than 92% of the variation was within, rather than between, populations (Table 5). Genetic similarities among plots were corre- spondingly high. Unbiased genetic identities (Nei 1978) between plots were greater than 0.99 for P. ponderosa and P. tridentata. Hedrick's distances (Hedrick 1971), calculated using band patterns, re- vealed a similarity greater than 0.98 for F. ida- hoensis. For F. idahoensis, band presence/absence data did not reveal differences between plots. No bands were unique to any plot. Two clusters ap- peared in a graph based on Jaccard's similarity in- dex (not shown), but all eight plots were repre- sented in both clusters. 168 MADRONO [Vol. 47 Table 4. Genetic Variation in Festuca idahoensis at Black's Mountain, for 12 Enzymes. All = overall statistics for the study. Mean = average over the 8 plots. SD = standard deviation. N = mean sample size/stain. #Bands = total number of bands (in all stains), in the population. %P* = percent of all presumed loci (regions on the gel that each probably represent a locus or set of homoeologos loci) that have more than one band pattern. A* = mean number of band patterns/stain. PI = polymorphic index based on band presence/absence data (see text). S-W = Shannon- Weaver diversity index is based on band patterns. B and N following plot numbers indicated the half plots that were subse- quently burned or not burned. N #Bands %P* A* PI S-W All 385 53 83% 5.17 3.804 0.563 Mean 48.1 45.1 66% 3.41 3.666 0.521 SD 8.6 2.5 8.3% 0.24 0.211 0.015 38B 61 46 75% 3.83 3.400 0.535 38N 47 45 67% 3.33 3.639 0.498 39B 59 45 67% 3.58 3.851 0.543 39N 36 42 58% 3.08 3.840 0.511 41B 49 43 50% 3.17 3.536 0.527 41N 47 49 75% 3.58 3.530 0.525 43 B 39 48 67% 3.33 4.013 0.516 43N 47 43 67% 3.42 3.520 0.510 With one exception, measures of genetic vari- ability in plots 38, 39, 41, and 43 did not differ significantly for any of the three species. The only exception was the Shannon-Weaver diversity index for P. tridentata. Its Shannon-Weaver diversity in- dex values differed significantly among plots (Table 6), and values for plot 43 are higher than those of plot 38. Before fire treatments were applied, mea- sures of genetic variability were the same in all the half plots, except that the percent polymorphic loci and observed heterozygosity for P. ponderosa were consistently higher on the plots that would later be burned (Table 6). The mean heterozygosity for the species (H^^) and the sampled populations (H^,p) (Hamrick and Godt 1990) were 0.271 and 0.266, respectively, for P. ponderosa, and 0.156 and 0.120 for P. triden- tata. Discussion Piniis ponderosa. Genetic variability found in this study is higher than previously reported in comparable studies of this species (that is, in stud- ies that involved at least twelve loci and including both polymorphic and monomorphic loci) (Allen- Table 5. Inter-population Diversity Statistics in Pi NUS PONDEROSA, PURSHIA TRIDENlArA AND FESTUCA IDAHOEN- SIS AT Black's Mountain. = a measure of inter-pop- ulational genetic differentiation derived from the Shannon-Weaver diversity index and analogous to F,,. F,, = Wright's fixation index (Weir 1990). S-W = mean Shannon-Weaver diversity index. Nm = inferred gene flow, = 0.25(I-FJ/F„. Species G, Nm Pinus ponderosa 0.0406 0.0188 13.0301 Purshia tridentata 0.0708 0.0148 16.6675 Festuca ida hoen s is 0.0637 dorf et al. 1982; Niebling and Conkle 1989; O'Malley et al. 1979; Woods et al. 1983; Yow et al. 1992). Expected heterozygosity (H^J in this study equals 0.272 while the average H^.^ of the oth- er studies equals 0.171. However, this level of ge- netic variability is consistent with previous NFGEL research on P. ponderosa (in previous NFGEL studies the average H^, = 0.231; NFGEL, unpubl.). The higher genetic variability reported by NFGEL for this species probably results from quality con- trol measures and highly standardized procedures that allow repeatable detection of small differences in enzyme migration distances. Including rare al- leles (those with frequencies lower than 0.05) in analyses may also contribute to the high genetic variability reported. Both NFGEL studies and pre- viously published work indicate that P. ponderosa subsp. ponderosa is more genetically variable than P. ponderosa subsp. scopidorum (average from NFGEL studies: H,., (subsp. ponderosa) = 0.247, (subsp. scopidorum) = 0.235; average from oth- er studies: H^, (subsp. ponderosa) = 0.161, Hg^ (subsp. scopulorum) = 0.151). Genetic variation in P. ponderosa was distributed within, rather than among, the plots. More than 90% of the isozyme variation often occurs within, rather than among, P. ponderosa populations (Hamrick et al. 1989, Linhart et al. 1981). Because P. ponderosa pollen can travel long distances (Latta et al. 1998), and calculated gene flow among plots in this study is high (Table 5), the short distances (2 to 4 kin) between plots may limit genetic differ- entiation. However, genetic differentiation has been detected previously over small distances in P. pon- derosa (Beckman and Mitton 1984; Mitton et al. 1977; Mitton et al. 1980). In general, the genetic similarily among plots provides a uniform back- ground against which the genetic effects, if any, of timber management practices will be detectable. However, the higher initial percent polymorphic 2000] SAMMAN ET AL.: GENETICS OF PONDEROSA PINE, BITTERBRUSH, AND IDAHO FESCUE 169 Table 6. Analysis of Variance of Measures of Genetic Variability in P/nus ponderosa, Purshia tridentata and Festuca idahoensis. The fixation index F = (H^,-H„)/H^.. The variance ratio F = Sf/So', where s' = the variance of the sample. * = statistically significant difference (p < 0.05). Half plots (half later burned) Plots 38, 39, 41 and 43 Measure of genetic variance F (variance ratio) probability F (variance ratio) probability Pinus ponderosa N = number of individuals/plot 1.0576 0.3434 0.7233 0.5886 P = % polymorphic loci 12.755 0.0118* 0.5460 0.6767 A = alleles/locus 0.2242 0.6526 2.3785 0.2107 Ap = alleles/polymorphic locus 2.4440 0. 1 690 0.9505 0.4964 Ho = observed heterozygosity 14.627 0.0087* 0. 1 305 0.9370 He = expected heterozygosity 1 .2605 0.3045 1 .0604 0.4588 F = fixation index 3.5968 0.1067 0.3551 0.7893 SW = Shannon Weaver diversity index 0.6778 0.4418 1.4773 0.3478 Purshia tridentata N = number of individuals/plot 0.1076 0.7540 0.0059 0.9993 P — % polymorphic loci 0.5627 0.4815 3.8774 0.1117 A = alleles/locus 0.4615 0.5223 6.0143 0.0579 Ap = alleles/polymorphic locus 0.0238 0.8824 1.5350 0.3355 H^ = observed heterozygosity 3.2858 0.1 198 0.1758 0.9076 He = expected heterozygosity 0.1197 0.7412 4.9700 0.0777 F = fixation index 2.9523 0.1366 0.2069 0.8868 SW = Shannon Weaver diversity index 0.0031 0.9574 6.6854 0.0489* Festuca idahoensis N = number of individuals/plot 1.8075 0.2274 0.41 15 0.7542 Number of bands/enzyme 0.1617 0.7015 0.2770 0.8401 %P* = % polymorphic enzyme 0.1005 0.7619 0.3481 0.7938 A* = number of patterns/enzyme 0.4906 0.5099 0.2970 0.8269 PI = polymorphic index 0.1811 0.6853 1 .4484 0.3542 SW = Shannon Weaver diversity index 5.6936 0.0543 0.3223 0.8104 loci and observed heterozygosity in burned than in unburned half-plots (Table 6) would have been con- sidered a treatment effect if this baseline study had not been done. Purshia tridentata. Bitterbrush has been the sub- ject of intense scrutiny, focused on interspecific re- lationships and management practices (e.g., Basile 1967) rather than genetic diversity. Secondary com- pounds have interferred with isozyme resolution in previous studies (S. Brunsfeld, pers. comm.). Ge- netic variation was less evenly distributed among plots in P. tridentata than in the other two species in this study (Table 3), possibly because P. triden- tata is pollinated by insects, rather than wind. Al- though plots were homogeneous for most measures, there were significant differences in the fixation in- dex (Table 3) and Shannon- Weaver diversity index (Table 6). Festuca idahoensis. Because polyploidy compli- cates gel interpretation, isozymes have been unde- rutilized for describing genetic variation in the fine- leaved fescues, Festuca subgenus Festuca, to which F. idahoensis belongs, and summary statistics are rarely reported. One should keep certain trends in mind when comparing a phenotypic analysis, like that performed for fescues, with genotypic analyses of isozymes. The proportion of polymorphic en- zymes (%PE in Tables 7, 8) is higher than the per- cent polymorphic loci (%?*) because stains may reveal both polymorphic and monomorphic loci for the same enzyme. %PE is reported because the sta- tistic is unambiguous. Percent polymorphic putative loci (%P*) is theoretically equal to %P of a genetic analysis, although for polyploid plants different re- searchers may parse band patterns into loci in dif- ferent ways. The number of patterns reported per putative locus (A*) should be somewhat greater than the number of alleles per locus, because any two alleles can produce three patterns. For exam- ple, AA homozygotes, BB homozygotes, and AB heterozygotes are counted as three different pat- terns. The polymorphic index (PI) is a rough mea- sure of heterozygosity, valid only for comparing populations within a study. Measures of similarity and diversity based on patterns (Hedrick's distance and the Shannon-Weaver diversity index, respec- tively) may overestimate differences, because AA homozygotes, BB homozygotes, and AB heterozy- gotes are considered three equally different pat- terns. On the other hand, measures of similarity and diversity based on bands (Jaccard's Similarity Index and the polymorphic index, respectively) produce uneven results among monomeric enzymes (which have two bands when heterozygous), dimers (which have three), and tetramers (which have five) (Got- tlieb 1977). The Gs statistic derived from partition- ing the Shannon-Weaver diversity index, like hier- archical F-statistics (Wright 1978), indicates wheth- er a greater proportion of variation resides within 170 MADRONO [Vol. 47 Table 7. Overall Isozyme Diversity Statistics for Taxa Festuca subgenus Festuca, from Studies that were ; NOT Limited to Polymorphic Alleles. Overall statistics are total values for the entire study. Chr. = Chromosome number (2X = 14, etc.). t = chromosome numbers from Markgraf-Dannenberg (1980); other chromosome numbers were provided in the source article. Pops. = number of populations. N = mean sample size per population. Enz = number of enzymes stained. Loci = number of regions on the gel that each probably represent a locus or set of homoelogous loci. Summary statistics calculated by authors: %PE = percent of polymorphic enzymes. %P = percent of polymorphic putative loci. AE = number of patterns per enzyme. A* = number of patterns per putative locus. Taxon Chr. Pops. N Enz. Loci %PE %P* AE A* Source (F. oviiui complex) F. iiiii'iculcitci 2X 4 20 10 15 90% 67% 1 F. baffensis 4X 3 26 10 15 90% 67% F. brachxphylla 6X 5 23 10 15 80% 67% 1 F. brevissiuia 2X 2 23 10 15 20% 13% 1 F. idahoeiisis 4X 8 48 1 1 12 82% 83% 5.17 5 F. idahoeusis 4X 8 41 1 1 18 91% 67% 2.67 4 F. niiuutilfora 2X 1 3 9 14 67% 21% 1 F. roemeri v. kicimatheiisis 4X 4 31 1 1 18 91% 67% 2.78 4 F. roemeri v. romeri 4X 8 27 1 1 18 91% 72% 4.00 4 F. valesiaca 2X 3 31 8 20 75% 50% 2.75 1.65 2 (F. rubra complex) F. amythestina 2Xt 2 58 8 20 62% 35% 2.25 1.40 2 F. asperifolia 1 40 8 20 75% 45% 2.25 1.65 2 F. diffusa 6X, 8Xt 1 40 8 20 50% 20% 1.65 1.35 2 F. heterophylla 4Xt 8 37 8 20 65% 40% 2.25 1.60 2 F. nigresceiis 4X, 6Xt 6 36 8 20 50% 35% 2.50 1.65 2 F. nigrescens 6X 3 27 1 1 18 82% 67% 2.33 4 F. peristerea 1 40 8 20 75% 30% 1.75 1.35 2 F. picturata 3 34 8 20 75% 35% 2.12 1.50 2 F. rubra 2X-10Xt 8 20 8 20 75% 45% 2.88 1.75 2 F. rubra 6X 6 38 10 100% 3 Sources: 1 2 3 4 5 Aiken et al. 1993 Angelov and Edreva 1987 (omitting anodal esterase), Angelov et al. 1988, Angelov 1992, 1993 Livesey and Norrington-Davies 1991 Wilson 1999 this study or among populations. Although a phenotypic anal- ysis is imprecise compared to genetic analyses of isozyme data, it does quantify isozyme variation and therefore allows comparisons among those polyploid taxa for which genetic interpretations are not possible. Available studies indicate that these fescues are highly polymorphic (Tables 7, 8), with the excep- Table 8. Mean Isozyme Diversity Statistics per Population for Festuca subgenus Festuca, from Studies that were not Limited to Polymorphic Alleles. Chr. = Chromosomes (2X = 14, etc.). Chromosome numbers were provided in the source article. Pops. = number of populations. N = mean sample size per population. Enz = number of enzymes stained. Loci = number of regions on the gel that each probably represent a locus or set of homoeologos loci. Summary statistics calculated by authors: %PE = percent of polymorphic enzymes. %P* = percent of polymorphic putative loci. A* = number of patterns per putative locus. S-W = Shannon-Weaver diversity index. Taxon Chr. Pops. N Enz. Loci %PE %P* AE A* Source {F. ovina complex) F. auriculata 2X 4 20 10 15 60% 43% 1 F. bracliyphylla 6X 5 23 10 15 67% 40% 1 F. brevissiiua 2X 2 23 10 15 68% 7% I F. idahoeusis 4X 8 48 1 1 12 68% 66% 3.41 3 F. ida/ioensis 4X 2 56 1 1 18 68% 50% 2.03 0.2032 2 F. niinutilfora 2X 1 3 9 14 67% 21% 1 F. roenun i v. klauiathensis 4X 4 31 1 1 18 80% 53% 1.89 0.2376 2 F. roemeri v. romeri 4X 6 33 1 1 18 67% 50% 1.94 0.2710 2 1 = Aiken et al. 1993 2 = Wilson 1999 3 = this study 2000] SAMMAN ET AL.: GENETICS OF PONDEROSA PINE, BITTERBRUSH, AND IDAHO FESCUE 171 tion of the uncommon F. brevissima Jurtzev. Vari- ability in Black's Mountain F. idahoensis was high but consistent with previously reported fescue ge- netic variation. Our reported number of band pat- terns per putative locus was particularly high but not anomalous. For example, we detected 17 PGI band patterns (Shannon- Weaver diversity index = 1.91), and the same number were found for PGI in a survey of 6 European Red Fescue (F. rubra L.) populations (S-W = 1.75) (Livesey and Norring- ton-Davies 1991). That study reported an average of 10 band patterns per enzyme in the three highly variable enzymes investigated. Genetic distances and identities among fescue populations are rarely reported. Genetic identities between populations of diploid arctic fescues are 0.934 {F. brevissima) and an average of 0.857 (F. auriculata Drobov aggregate) (Aiken et al. 1993). In both intra- and interspecific comparisons, Hed- rick's identities exceed 0.95 between populations in the tetraploid F. idahoensis and Roemer's Fescue (F. roemeri (Pavlick) E. B. Alexeev) species pair in northern California (Wilson 1999). The high (greater than 0.98) Hedrick's identities among Black's Mountain fescue populations are therefore expected. Because Black's Mountain populations of F. ida- hoensis were similar, any effects of burning, graz- ing, and logging regimes on genetic variability will be detectable. Such fine-scale adaptation to local habitat variables has been seen in grasses, particu- larly in self-pollinating species (Bradshaw 1959; Clary 1975; Clegg and Allard 1972; Hamrick and Allard 1972; Kahler et al. 1980; Lonn 1993; Nevo et al. 1983). However, coarse adaptation has also been observed in both self- and cross-pollinated in- troduced grasses (Rice and Mack 1991; Rapson and Wilson 1988). Summary. The gymnosperm tree P. ponderosa, the dicot shrub P. tridentata, and the monocot grass F. idahoensis are not phylogenetically close. How- ever, they have strikingly similar patterns of elec- trophoretically detected genetic variation. All are genetically variable, with well over 90% of the variation within, rather than among, populations in the area studied. All three are common, widespread, long-hved, perennial, outcrossing species that dom- inate late successional stages in their plant com- munity. Plants with this series of characteristics tend to be more genetically variable than average, and to have their genetic variation within, rather than among, populations (Hamrick and Godt 1990). The high level of genetic variability detected in the three plants is consistent with observed trends in genetic variability (Hamrick and Godt 1990). The markedly lower genetic variation in P. tri- dentata as compared to the other two species is consistent with the tendency for insect pollinated plants and dicots to have much less isozyme vari- ation than wind pollinated plants and gymnosperms or monocots (Hamrick and Godt 1990). The value of the genetic diversity statistics H^., and H^,p for P. ponderosa are more than one stan- dard deviation higher than comparable mean statis- tics reported (Hamrick and Godt 1990; Hamrick et al. 1992). Those for P. tridentata vary from that far above to somewhat below average, depending upon the comparison; they are low for woody angio- sperms (Hamrick et al. 1992), but most woody an- giosperms studied are wind-pollinated trees, and P. tridentata is an insect-pollinated shrub. For all three species, the percentage of genetic variability among populations is lower than the mean reported in comparable plants (Hamrick and Godt 1990), but few other studies compared populations growing in such close proximity in the same habitat. Acknowledgments We thank Chris Bailey and Dawn Cambell for collect- ing P. ponderosa cones. We thank Laura French and Des- sa Welty for collecting P. tridentata and F. idahoensis, and for providing insights into collection methods. We thank Snellen Carroll, Patricia Guge, and Randy Meyer for electrophoresis and gel scoring. We thank Dr. Steven J. Brunsfeld and Manuela Huso for their input. Dr. Georgi Angelov for providing unpublished information about fes- cue isozymes, Maria Ivantchenko for translation from the Bulgarian, Tanya Bray for text editing, and anonymous reviewers for critical comments on the manuscript. This article is contribution no. 18 of the Blacks Mountain Eco- logical Research Project, Pacitic Southwest Research Sta- tion, USDA Forest Service. Literature Cited Adams. W. T. J. Zuo, J. Y. Shimizu and J. C. Tappeiner. 1998. Impact of alternative regeneration methods on genetic diversity in coastal Douglas-Fir. Forest Sci- ence 44:390-396. Aiken, S. G., L. L. Consaul, J. I. Davis and P. S. Manos. 1993. Systematic inferences from variation in iso- zyme profiles of arctic and alpine cespitose Festuca (Poaceae). American Journal of Botany 80:76-82. Allendorf, F. W, K. L. Knudsen and G. M. Blake. 1982. Frequencies of null alleles at enzyme loci in natural populations of ponderosa and red pine. Ge- netics 100:497-504. Angelov. G. B. 1992. Biosystematic study of species from the genus Festuca L. IV. Electrophoretic anal- ysis of the enzymes peroxidase, acid phospliatase, and catalase. Fitologia 43:58-68. Angelov, G. B. 1993. A biosystematic study of species of genus Festuca L. V. An electrophoretic analysis of the enzymes glucose-6-phosphate dehydrogenase, 6- phosphogluconate dehydrogenase, isocitrate dehydro- genase, and superoxide dismutase. Fitologia 44:47- 55. Angelov, G. B. and A. Edreva. 1987. Electrophoretic study of species of genus Festuca L. I. Electropho- retic patterns of esterase. Trudove IV Natsionalna Konferentsiy Botanika, Bulgarska Akademiya na Naukite [Proceedings of the Fourth National Confer- ence of Botany; Bulgarian Academy of Science], So- fia, Bulgaria. 172 MADRONO [Vol. 47 Angelov, G. B., S. I. KojUHAROV AND A. Edreva. 1988. Biosystematic study of the species of genus Festuca L. II. Electrophoretic analysis of the seed proteins and the enzymes glutamate-oxaloacetate-transaminase and glutamate dehydrogenase. Fitologia 35:13-23. Anonymous. 1995. National Forest Genetic Electropho- resis Laboratory Standard Operating Procedures. NFGEL. USDA Forest Service, Camino, CA. Baron, F. J., E. C. Nord, A. B. Evanko and W. J. Makel. 1966. Seeding conifers and buffer crops to reduce deer depredation. U.S. Forest Service Res. Note PSW-100. Basile, J. V. 1967. An annotated bibliography of P. tri- denta (Purshia tridentata (Pursh) DC.). Intermoun- tain Forest and Range Experiment Station. USDA Forest Service Research Paper INT-44. Ogden, UT. Beckman, J. S. and J. B. MiTTON. 1984. Peroxidase allo- zyme differentiation among successional stands of P. poiiderosa. American Midland Naturalist 112:43-49. Bradshaw, a. D. 1959. Population differentiation in Agrostis tenuis Sibth: morphological differentiation. New Phytologist 58:208-227. Busse, M. D., R H. Cochran and J. W. Barrett. 1996. Changes in P. ponderosa site productivity following removal of understory vegetation. Soil Science So- ciety of America Journal 60:1614-1621. Chung, M. G., J. L. Hamrick, S. B. Jones and G. S. Derda. 1991. Isozyme variation within and among populations of Hosta (Liliaceae) in Korea. Systematic Botany 16:667-84. Clary, W. P. 1975. Ecotypic adaptation in Sitanion hys- trix. Ecology 56:1407-1415. Clegg, M. T. and R. W. Allard. 1972. Patterns of genetic differentiation in the slender wild oat species Avena barhata. Proceedings of the National Academy of Sciences USA 69:1820-1824. CONKLE, M. T, p. D. HODGSKISS, L. B. NUNNALLY AND S. C. Hunter. 1982. Starch Gel Electrophoresis of Co- nifer Seeds: A Laboratory Manual. Gen. Tech. Rep. PSW-64. Berkeley. CA: Pacific Southwest Forest and Range Experiment Station, Forest Service, U.S. De- partment of Agriculture. Darlington, C. D. and A. P. Wylie. 1955. Chromosome Atlas of Flowering Plants. University Press, Aber- deen, Great Britain. Gottlieb, L. D. 1977. Electrophoretic evidence and plant systematics. Annals of the Missouri Botanic Garden 64:161-180. Gottlieb, L. D. 1981. Gene number in species of Aster- aceae that have different chromosome numbers. Pro- ceedings of the National Academy of Sciences USA 78:3726-3729. Gottlieb, L. D. 1982. Conservation and duplication of isozymes in plants. Science 216:373-380. Hall, D. B., S. B. Monsen and V. J. Anderson. 1995. Influence of herbaceous understory on native shrub recruitment. Rangelands 17:12-14. Hamrick, J. L. and R. W. Allard. 1972. Microgeograph- ical variation in allozyme frequencies in Avena bar- hata. Proceedings of the National Academy Sciences USA 69:2100-2104. Hamrick, J. L., H. M. Blanton and K. J. Hamrick. 1989. Genetic structure of geographically marginal popu- lations of P. ponderosa. American Journal of Botany 76:1559-1568. Hamrick, J. L. and M. J. W. Godt. 1990. Allozyme di- versity in plant species. Pp. 43-63. in Brown, A. H. D. , M. T. Clegg, A. L. Kahler and B. S. Weir (eds.). Plant population genetics, breeding, and genetic re- sources. Sinauer Associates, Inc., Sunderland, MA. Hamrick, J. L., M. J. W. Godt and S. L. Sherman-Broy- LES. 1992. Factors influencing levels of genetic di- versity in woody plant species. Pp. 94-124. in W. T. Adams, S. H. Strauss, D. L. Copes and A. R. Griffin (eds.). Population genetics of forest trees. Kluwer Ac- ademic Publishers, Boston. Hartl, D. L. and a. G. Clark. 1989. Principles of Pop- ulation Genetics, 2nd ed. Sinauer Associates, Sun- derland, MA. Hedrick, p. W. 1971. a new approach to measuring ge- netic similarity. Evolution 25:276-280. Kahler, A. L., R. W. Allard, M. Krzakowa, C. F. Wehrhahn and E. Nevo. 1980. Associations between isozyme phenotypes and environment in the slender wild oat {Avena barbata) in Israel. Theoretical Ap- plied Genetics 56:31-47. KiMURA, M. and J. F. Crow. 1964. The number of alleles that can be maintained in a finite population. Genetics 49:725-738. Latta, R. G., Y. B. Linhart, D. Fleck and M. Elliot. 1998. Direct and indirect estimates of seed versus pollen movement within a population of P. ponder- osa. Evolution 52:61-67. Linhart, Y. B., M. C. Grant and P. Montazer. 1989. Experimental studies in P. ponderosa. I. Relationship between variation in proteins and morphology. Amer- ican Journal of Botany 76:1024-1032. Linhart, Y. B., J. B. Mitton, K. B. Sturgeon and M. L. Davis. 1981. Genetic variation in space and time in a population of P. ponderosa. Heredity 46:407-426. LiVESEY, V. and J. Norrington-Davies. 1991. Isozyme polymorphism in Festuca rubra L. Euphytica 55:72- 79. Lonn, M. 1993. Genetic structure and allozyme-micro- habitat associations in Bromus hordeaceus. Oikos 68: 99-106. Markgraf-Dannenberg, I. 1980. Festuca, Pp. 125-153. in Tugin, T. G., V. H. Heywood, N. A. Burges, D. M. Moore, D. H. Valentine, S. M. Walters and D. A. Webb (eds.). Flora Europea, Vol. 5. Monocotyledons. Cambridge University Press, Cambridge. Microsoft Corporation. 1997. Microsoft Excel SR-1. Redmond, Microsoft Corporation, Washington. Mitton, J. B., Y. B. Linhart, J. L. Hamrick, and J. S. Beckman. 1977. Observation on the genetic structure and mating system of P. ponderosa in the Colorado Front Range. Theoretical Applied Genetics 51:5-13. Mitton, J. B., K. B. Sturgeon and M. L. Davis. 1980. Genetic differentiation in P. ponderosa along a steep elevational transect. Silvae Genetica 29:100-103. Nei, M. 1973. Analysis of gene diversity in subdivided Populations. Proceedings of the National Academy of Sciences USA 70:3321-3323. Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89:583-590. Nevo, E., A. Beiles and N. Storch. 1983. Microgeo- graphic edaphic differentiation in hordein polymor- phisms of wild barley. Theoretical Applied Genetics 64:123-132. NiEBLiNG, C. R. AND M. T. CoNKLE. 1989. Diversity of Washoe pine and comparisons with allozymes of P. ponderosa races. Canadian Journal of Forest Re- sources 20:298-308. Oliver, W. W. MS. Black's Mountain Interdisciplinary Research: Vegetation. 2000] SAMMAN ET AL.: GENETICS OF PONDEROSA PINE, BITTERBRUSH, AND IDAHO FESCUE 173 O'Malley, D. M., F. W. Allendorf and G. M. Blake. 1979. Inheritance of isozyme variation and hetero- zygosity in Pinus ponderosa. Biochemical Genetics 17:233-250. Rapson, G. L. and J. B. Wilson. 1988. Non-adaptation in Agrostis capillaris L. (Poaceae). Functional Ecology 2:479-490. RoHLF, F. J. 1987. NTSYS-pc: numerical taxonomy and multivariate analysis system for the IBM PC micro- computer (and compatibles). Applied Biostatistics, Setauket, New York. Rice, K. J. and R. N. Mack. 1991. Ecological genetics of Bromiis tectorum III: The demography of reciprocally sown populations. Oecologia 88:91-101. Shannon, C. E. and W. Weaver. 1949. The mathematical theory of communication. University of Illinois Press, Urbana, IL. Slatkin, M. and N. H. Barton. 1989. A comparison of three indirect methods for estimating average levels of gene flow. Evolution 43:1349-1368. SoLTis, D. E. and L. H. Rieseberg. 1986. Autopolyploidy in Tolmiea meuziesii (Saxifragaceae): genetic insights from enzyme electrophoresis. American Journal of Botany 73:310-318. vander Wall, S. B. and S. B. vander Wall. 1992. Es- tablishment of Jeffrey pine seedlings from animal caches. Western Journal of Applied Forestry 7:14- 20. Weeden, N. F. and J. F. Wendel. 1990. Genetics of plant isozymes. Pp. 46-72. in D. E. Soltis and P. S. Soltis. (eds.). Isozymes in plant biology. Dioscorides Press, Portland, OR. Wendel, J. F. and N. F. Weeden. 1989. Visualization and interpretation of plant isozymes. Pp. 5-45 in D. E. Soltis and P. S. Soltis (eds.). Isozymes in Plant Bi- ology. Dioscorides Press, Portland, OR. Weir, B. S. 1990. Genetic Data Analysis. Sunderland, MA, Sinauer. Wilson, B. L. 1999. Fescue Taxonomy in the Pacific Coast States. Unpubl. Ph.D. Thesis, Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR. Woods, J. H., G. M. Blake and F. W. Allendorf. 1983. Amount and distribution of isozyme variation in P. ponderosa from eastern Montana. Silvae Genetica 32: 151-156. Wright, S. 1978. Evolution and the Genetics of Popula- tions, Vol. 4. Variability within and among natural populations. University of Chicago Press, Chicago. Yeh, E C, R.-C. Yang, T. Boyle, Z'H. Ye and J. X. Mao. 1997. Popgene, the userfiendly shareware for population genetic analysis. Version 1.20. Molecular Biology and Biotechnology Centre, University of Al- berta, Edmonton, Alberta. Yow, T. H., M. R. Wagner D. E. Wommack and G. A. TusKAN. 1992. Influence of selection for volume growth on the genetic variability of southwestern P. ponderosa. Silvae Genetica 41:326-333. Madrono, Vol. 47, No. 3, pp. 174-188, 2000 THE EFFECT OF CLIMATIC VARIABILITY ON GROWTH, REPRODUCTION, AND POPULATION VIABILITY OF A SENSITIVE SALT MARSH PLANT SPECIES, LASTHENIA GLABRATA SUBSP COULTERI (ASTERACEAE) Lorraine S. Parsons' and Adam W. Whelchel^-^ ^Point Reyes National Seashore, Point Reyes Station, CA 94956 2URS Corp., Box 290, 201 Willowbrook Blvd., Wayne, NJ 07474 Abstract As with many other sensitive species in California, the range of Lasthenia glabrata Lindley subsp. coulteri (A. Gray) Ornd. (Asteraceae; Coulter's goldfiields) has been dramatically reduced in recent de- cades by urbanization. Many populations are small, isolated, and seemingly unstable. In this study, we conducted an autecological assessment of a small L. glabrata subsp. coulteri population at San Diego County's San Dieguito Lagoon, using a large population at San Elijo Lagoon for comparison. The large population was not only more stable based on trends in seed production, but generally produced larger plants and more flowers and capitulescences than the small population. However, this relationship appears to be temporally variable and influenced significantly by climatic conditions, particularly rainfall totals and distribution. In a year with above-average rainfall, vegetative and reproductive yield of plants in the small population (San Dieguito Lagoon) matched or even exceeded that of plants in the large one (San Elijo Lagoon), which was subjected to prolonged inundation following heavy rains and a back-up of run- off and creek flows behind a dike system. Rainfall is linked not only to soil moisture, but to nutrient influx and cycling, variables that were strongly associated with group (marsh/monitoring year) separation and prediction in statistical analyses. When resources are sufficient, reproductive yield appears to be driven by other factors, the most probable of which is pollen supply. The relationship between rainfall and plant yield could prove integral to predicting long-term viability of L. glabrata subsp. coulteri pop- ulations, as above-average rainfall years are often sporadic and interspersed between lengthy periods of average or below-average rainfall in southern California. Conservation and enhancement of the remaining coastal salt marsh L. glabrata subsp. coulteri populations could perhaps be furthered by factoring this relationship into conservation and restoration projects and hydrologic regimes designed for managed wetland systems. Small populations of rare plant species face many genetic, demographic, and ecological chal- lenges. Small populations can suffer from reduced "fitness," often undergoing one or more genetic bottleneck events that reduce genetic variation (Nei et al. 1975; Hamrick et al. 1979; Hedrick 1983; Ledig 1986; Barrett and Kohn 1991 and oth- ers). Fewer numbers also create a greater chance for normally outbreeding species to inbreed and become less fit through concentration of deleteri- ous alleles (Charles worth and Charles worth 1987 and others). Opportunities for gene flow between populations — and the potential for infusion of new alleles — may be minimal due to the dwindling number of populations and the distance between them. Small plant populations are intrinsically less appealing to pollinators (Powell and Powell 1987; Morgan 1999), which can further reduce fecundity and the potential for even limited outbreeding be- tween more distant individuals. Reduced genetic variation can also increase small populations' sus- ceptibility to herbivory, pathogens, and stochastic factors such as floods and environmental and de- mographic variability (Shaffer 1981). In addition, population viability can be continually jeopardized by human-related disturbances or changes in wa- tershed or ecosystem conditions — sometimes the very changes believed to have made the species rare in the place. Even efforts to better manage, enhance, or restore systems in which rare plants occur can pose a threat if these activities do not balance their ecological requirements with those of other target plant and wildlife species and the ecosystem as a whole. Determining whether small populations are suc- cumbing to these challenges is not an easy task. Annual censuses are not only difficult, but often misleading unless conducted over several decades due to cryptic life history stages (i.e., seed banks) and normal fluctuations in population size that may have little impact on population stability or viability (Davy and Jefferies 1981; Schemske et al. 1994; Pavlik 1994 and others). A life table or population viability analysis (PVA) is often considered the op- timal approach for assessing population stability (Schemske et al. 1994; Pavlik 1994; Menges 1986). However, the probability of a long-lived seed bank immeasurably complicates performance of a life ta- ble or PVA for plant species (Pavlik 1994), despite arguments that there are ways to circumvent cal- culation of this unknown (Menges 1986). Some al- ternative approaches to assessing population stabil- 2000] PARSONS AND WHELCHEL: CLIMATIC VARIABILITY AND RARE PLANT 175 ity involve performance of non-integrated demo- graphic trend assessment, which focuses on overall i trajectories in survivorship, seed production, den- sity of viable seed, and frequency of establishment ' (Pavlik 1994). Morphological attributes associated with productivity or yield such as plant size/bio- I mass and flower number may be incorporated, as well (Menges 1986; Menges and Gordon 1996). These analyses are often improved through using either weedy congeners or large, more stable pop- ulations of the same species for comparison (Pavlik 1994). i Consistent with the major role that extrinsic dis- \ turbances or changes can play in population viabil- I ity, many studies on rare species include an eco- , logical, as well as demographic, component (Schemske et al. 1994). Some have criticized re- searchers for emphasizing autecology over demog- raphy in sensitive plant research, characterizing ecological research as premature in the absence of demographic information relevant to population vi- tal rates (Schemske et al. 1994). However, manag- ers of reserves and enhancement/restoration pro- jects often seek ecological information that might help them better manage reserves or design projects (Pavlik 1994). Moreover, ecological data can great- ly complement demographic assessments, particu- larly when the information is integrated to allow for identification of ecological constraints on key life history stages and variables associated with productivity (e.g., plant size, flower number) (Schemske et al. 1994; Pavlik 1994; Menges and Gordon 1986). San Dieguito Lagoon in San Diego County sup- ports a small population of a rare plant, Lasthenia glabrata Lindley subsp. coulteri (A. Gray) Ornd. (Asteraceae: Coulter's goldfields). For a sensitive species, L. glabrata subsp. coulteri has a remark- ably diverse distribution. This annual is found in alkali playas in southern California's arid inland ar- eas and salt marshes and vernal pools in the re- gion's more moderate coastal areas (NDDB 1998; Skinner and Pavlik 1994; Hickman 1993). This diverse distribution has not spared the spe- cies from the threat of extirpation, however. All of these habitats have been negatively impacted to some extent by California's extensive urbanization over the past 50 y (Skinner and Pavlik 1994). More than 90 percent of California's wetland habitats, in- cluding marshes, vernal pools, and alkali playas, have been destroyed by commercial and residential development, and despite regulatory efforts at ef- fecting a "no net loss" policy, this downward trend in wetland habitat acreage appears to be continuing. The wetland habitats that remain are often frag- mented, highly disturbed, and heavily impacted by outside influences such as nutrient and contaminant influx associated with watershed development. The toll these habitat losses and impacts has taken is apparent from the constriction of the species' his- toric range. In recent decades, its once extensive distribution throughout southern California has been reduced to a few marshes and vernal pools in San Diego, Ventura, and Santa Barbara counties and alkali playas in Riverside County (NDDB 1998). This precipitous decline in distribution prompted listing of L. glabrata subsp. coulteri as a species of concern (formerly C2) by the U.S. Fish and Wild- life Service and a species of limited distribution (List IB) by the California Native Plant Society (CNPS). While its cousin, Lasthenia glabrata Lind- ley subsp. glabrata, is relatively common and has even a larger range than subsp. coulteri, other Las- thenia species that occur in vernal pool habitats such as Lasthenia burkei (E. Greene) E. Greene (Burke's goldfields) and Lasthenia conjugens E. Greene (Contra Costa goldfields) are faced with similar threats in terms of potential extirpation (Skinner and Pavlik 1994). Within its historic coast- al range, L. glabrata subsp. coulteri often grows in high elevation areas of salt marshes — or the "high marsh" — alongside another sensitive species, Cor- dylanthus maritimus Benth. subsp. maritimus (salt marsh bird's beak), a state- and federally listed en- dangered species. San Dieguito Lagoon is one of six San Diego County coastal marsh systems that supports histor- ical and/or possibly reintroduced populations of L. glabrata subsp. coulteri. The species was once present at 10 San Diego County marshes (NDDB 1998), but probably in low abundance, as an early ecological study characterized it as only an "infre- quent" inhabitant (Purer 1942). Of the six remain- ing occurrences, three are believed to be small and relatively unstable populations, including the one at San Dieguito Lagoon. Over the past few decades, L. glabrata subsp. coulteri numbers at San Dieguito Lagoon have ranged from as low as zero in 1980 (Sea Science Services and Pacific Southwest Bio- logical Services, Inc. 1980) and six in 2000 (An- drea Thorpe personal communication) to as high as 1000 individuals during the mid- and late- 1990s (MFC Analytical, Inc. 1993; L. Parsons and A. Whelchel, personal observation). Neighboring marshes such as San Elijo Lagoon, as well as re- portedly Los Penasquistos Lagoon, support annual populations consistently numbering as many as 5000 to 10,000 individuals (L. Parsons and A. Whelchel personal observation). San Dieguito Lagoon is also one of seven coastal marshes in San Diego County, CA, for which res- toration and/or enhancement activities have been or are being conducted or are proposed. As with other San Diego County marshes, this coastal lagoon has been subject to a number of historic watershed changes and disturbances, including damming of its river, agricultural and residential development, dik- ing, and intermittent mouth closures that impound water and create hypoxic conditions. With restora- 176 MADRONO [Vol. 47 tion and enhancement plans for San Dieguito La- goon currently being developed, there appeared to be a strong and immediate need for information on demographic and ecological aspects of this small and possibly unstable population and its potential for conservation and even future enhancement. Few studies have actually assessed demography or aut- ecological relationships of this or other Lasthenia species. The research that exists deals primarily with upland species {Lasthenia califoniica Lindley; Rajakaruna and Bohm 1999; Vivrette 1999) or fo- cuses on salinity tolerance (L. glabrata subsp. coul- teri; Kingsbury et al. 1976; Callaway et al. 1990; Callaway and Sabraw 1994). In 1996, a study was implemented to assess de- mographic and ecological characteristics of the L. glabrata subsp. coiilteri population at San Dieguito Lagoon. For purposes of performing a comparative assessment, we broadened the scope of our study to include another L. glabrata subsp. coiilteri pop- ulation that appeared to be larger and more stable in terms of plant numbers, the population at San Elijo Lagoon, located directly north of San Diegui- to Lagoon. Through this study, we hoped to gain insight into differences in survivorship, reproduc- tive potential (plant size, flower number) and suc- cess (seed set and seed production) between pop- ulations. When possible, we also attempted to track trends in demographic results for purposes of as- sessing population stability. By comparing autecol- ogy of a small and unstable population with that of a large and stable one, we hoped to increase our understanding of the biotic and abiotic factors that might be influencing variations in plant yield and population viability. We believe that this informa- tion could prove invaluable to resource managers charged with planning or implementing complex restoration or conservation projects, particularly projects with multiple species and ecosystem ob- jectives. Study area. San Dieguito Lagoon is located in Del Mar, CA, approximately 25 km north of San Diego (Fig. 1). The lagoon's principal source of freshwater is the San Dieguito River, which has been dammed to create Lake Hodges, a reservoir in the inland area of the San Diego County. Water- shed of the lagoon totals 897 km% 785.2 km- of which is behind dams (California Wetlands Infor- mation System 1996). During the summer and fall, the lagoon mouth sometimes closes for several weeks to a month until it re-opens either naturally or manually (i.e., using bulldozers). Vegetation communities occurring within the lagoon include salt and brackish marshes and, at its eastern end, freshwater marsh and limited riparian habitats. Ap- proximately 104 of the 240 hectares of wetlands once present at San Dieguito Lagoon still remain (California Wetlands Information System 1996). In addition to damming, the watershed of the lagoon has been altered considerably by development of Fig. 1. Location of Lasthenia glabrata subsp. coulteri (Coulter's goldfields) study areas in San Diego County, CA. the adjacent floodplain and uplands for agriculture, commercial and residential structures, and a race- track/fairgrounds. The population of L. glabrata subsp. coulteri at this lagoon is located in a muted tidal basin in the southern portion of the lagoon, approximately 2 km from the mouth. San Elijo Lagoon is located in the Cardiff, CA, approximately 27.5 km north of San Diego and di- rectly north of San Dieguito Lagoon (Fig. 1). Sev- eral creeks flow into the lagoon, one of which is dammed by Lake Wohlford (Escondido Creek). The watershed for San Elijo Lagoon is 200.2 km- (Cal- ifornia Wetlands Information System 1996). The mouth of San Elijo Lagoon also closes during the summer and fall, typically with more frequency and longer duration than San Dieguito Lagoon. Current wetland acreage at San Elijo Lagoon totals 230.4 ha and is comprised of salt marsh, brackish marsh, freshwater marsh, and riparian habitat (California Wetlands Information System 1996). The water- shed of this lagoon has been altered considerably by development of the adjacent floodplain and up- lands for farming and commercial and residential structures. The lagoon has also been separated into a series of hydrologic "cells" by construction of a series of dikes and levees between 1880 and 1940. By closing flood gates at one of the eastern dikes, the area east of Interstate 5 is partially flooded from November through March for water- fowl enhancement (Susan Welker personal com- munication). The population of L. glabrata subsp. 2000] PARSONS AND WHELCHEL: CLIMATIC VARIABILITY AND RARE PLANT 177 Table 1 . Mean Annual Temperature and Rainfall and Number of Days that Lagoon Mouth was Open During THE Study Period (1996-1999). San Diego NWS-Lindbergh (NOAA; CDWR) San Dieguito Lagoon San Elijo Lagoon Temperature Departure No. of days (Jan-Dec) Departure Rainfall from mean No. of days mouth Year mean from mean (Oct-April) (% of mean mouth open* open** end (deg.C) (% of mean) total (cm) total) (Nov-Oct) % of year (Nov-Oct) % of year 1996 17.8 99.4 12.85 53.4 351 96 80 22 1997 18 100.5 17.48 72.7 358 98 153 42 1998 18.5 103.4 42.75 177.8 338 92 238 65 1999 15.6 87.2 16.2 67.4 2 ] 9*** 60 240 66 NOAA = National Oceanic Atmospheric Administration Center, National Climatic Data Center CDWR = California Department of Water Resources, California Data Exchange Center, CIMIS * = H. Elwany, unpublished data; San Diego County Department of Environmental Health ** = San Diego County Parks Department, unpublished data *** = All closure events occurred after April 1999 and plant senescence coulteri at this lagoon occurs at the eastern end of the lagoon in an area that receives little to no di- rect tidal flow, although some subsurface tidal in- flow may occur. Mean annual temperature and rainfall data and data for the number of days the lagoon mouths were open for the study period (1996-1999) are provided in Table 1. Rainfall was 53 to 73 percent of average during October-April in 1996, 1997, and 1999 and 178 percent of average during those months in 1998. Annual mean daily temperature showed less variation (87 to 103 percent of average between 1996-1999). Methods Annual Monitoring Demography. In general, demographic informa- tion was collected in 1996, 1997, 1998, and 1999 at San Dieguito Lagoon and in 1997, 1998, and 1999 at San Elijo Lagoon. To assess demography, 14 plants within each of 10 sampling locations (0.5 X 0.5 m plots) were haphazardly chosen and marked in late January or early February of each monitoring year, when the plants were 1 to 2 cm tall seedlings. The sampling locations were chosen as representative of the microhabitat diversity and environmental heterogeneity present with the pop- ulations' existing range at each marsh. Mortality and phenology were assessed on a monthly or twice monthly basis for three months: February, March, and April. Mortality was assessed as the number of marked plants that died between marking and mid- April. Phenology was broken into three basic stages: vegetative, in bud, and flowering. In addi- tion, plants were examined for signs of potential herbivory. In mid-April, when most plants had al- ready set seed, a minimum of 10 plants and a max- imum of 14 plants were harvested from each sam- pling location to determine aboveground plant height (cm), number of capitulescences (inflores- cences), capitulescence diameter (the diameter of the receptacle in mm excluding ray flowers), num- ber of flowers (total of disc and ray flowers), num- ber of seeds, and seed set (number of seeds/number of flowers). Seeds were also examined for signs of granivory or seed predation. There were some exceptions to the described de- mographic monitoring. Mortality was not assessed during the following marsh/years — San Dieguito Lagoon 1996 and 1999 and San Elijo Lagoon 1999 — but plants were harvested in mid- April for measurement of plant height, capitulescence num- ber and diameter, flower and seed number, and seed set. Efforts were made to assess survivorship at San Dieguito Lagoon and San Elijo Lagoon in 1998, but a sedimentation event associated with higher-than- average water levels in the study area at San Elijo Lagoon caused marking materials (colored rubber- bands) to be obscured. The 1998 data for San Die- guito Lagoon was incomplete, as information on mortality was not collected in April. Biotic variables. Density of L. glabrata subsp. coulteri was assessed biweekly in 1997 using a 0.5 X 0.5 m quadrat subdivided into 25 1-dm^ subquad- rats. One of the subquadrats was randoinly chosen, and the number of L. glabrata subsp. coulteri in- dividuals present within the subquadrat was count- ed. Also, vegetative cover within the sampling plot was assessed one time per inonitoring year using the 36 cross points of the subdivided 0. 5 X 0.5 in quadrat and recording the species or bareground oc- curring below the cross point. For analysis purpos- es, percent cover was calculated for L. glabrata subsp. coulteri, total vegetation cover, and cover of non-native species. Total vegetation cover included both native and non-native species. Native species were primarily coastal salt marsh inhabitants such as Salicornia virginica L. (pickleweed), Salicornia subterminalis Parish (glasswort), Frankenia salina (Molina) I. M. Johnston (alkali heath), Cressa trux- illensis Kunth (alkali weed), and Spergularia ma- rina (L.) Grisels. (sand-spurrey). Non-native spe- 178 MADRONO [Vol. 47 cies included Cotula coronopifolia L. (brass-but- tons), Mesembryanthemum crystallinum L. (crys- talline iceplant), Lolium multiflorum Lam. (Italian ryegrass), Parapholis incurva (L.) C. E. Hubb. (sickle grass), Polypogon monspeliensis (L.) Desf. (annual beard grass), and Poa annua L. (annual bluegrass). Abiotic variables. A total of 1 1 abiotic variables was assessed during each monitoring year, except 1999. The abiotic variables were soil pH, soil sa- linity, soil moisture, organic matter, ammonium, ni- trates + nitrites, phosphorous, cation exchange ca- pacity (CEC), calcium, magnesium, and potassium. Soil texture was assessed at both marshes in 1997. As this species grows in high marsh areas, which are only infrequently inundated, reduction-oxida- tion potential was not measured. Soil pH, soil sa- linity, and soil moisture was measured twice a month (1997) to monthly (1998) from 15-cm-deep soil core samples at all 10 sampling locations. Soil pH was measured by creating soil pastes in the field and measuring with an Oakton phTestr 3 (±0.01 pH resolution) field pH probe. Soil salinity was measured by expressing soil water from a syringe fitted with filter paper onto a refractometer, which reports salinities in grams per kilogram. Soil mois- ture was measured by removing 10- to 15-cm soil cores and assessing loss of mass on drying (Gard- ner 1986). For nutrient analysis, five of the 10 sam- pling locations at each marsh were randomly se- lected for subsampling twice each monitoring year (mid-February and mid-March). At these subsam- pling locations, approximately 100 g of soil was removed, air dried, and sent to A&L Western Ag- ricultural Laboratories (Modesto, California) for measurement of organic matter, ammonium, ni- trates + nitrites, phosphorous, cation exchange ca- pacity, calcium, magnesium, and potassium. In 1997, the laboratory also analyzed soil texture. The procedures described above were performed for both marshes during the years monitoring was per- formed, with the exception of San Dieguito Lagoon in 1996 and 1999 and San Elijo Lagoon in 1999. Soil salinity was not measured at San Dieguito La- goon in 1996, and pH, soil moisture, organic mat- ter, and other nutrients were only analyzed once during the 1996 monitoring year. Data for variables sampled more than once per season were averaged for analysis. Data analysis. Differences in plant population dynamics between marshes and sampling years were assessed by treating each "marsh/year" sam- pled independently and conducting a One-Way Analysis of Variance using the Systat computing package (SPSS, Chicago, IL). For the density com- parison, a t-test was conducted to test for differ- ences in plant density between marshes. When as- sumptions for parametric tests were not met, data were either transformed, or an equivalent, non- parametric procedure (e.g., Kruskal-Wallis) was conducted. If a significant difference was found, differences between particular means were ana-! lyzed further by using either Tukey, T'-method (So-i kal and Rohlf 1981), or non-parametric Tukey-typel (Zar 1984) multiple comparison procedures. The! dependent variable plant height was log-trans- j formed for analysis. Discriminant function analysis was used to ex- plore the association between plant population dy- namics and 13 biotic and abiotic factors within marshes. Quadratic discriminant function analyses were performed, because they are less sensitive to dissimilarities in covariance matrices between groups. Groupings used in analyses were based on! results from the Analysis of Variance tests, with' generally low yield plots separated from high yield plots. The analyses incorporated data from 1996- 1998: no biotic and abiotic data were collected in 1999. As salinity data were not available for Sani Dieguito in 1996, a preliminary analysis was per-! formed using models that incorporated the salinity, variable, but not the San Dieguito 1996 data. If sa- linity did not have a strong loading on any canon-, ical variable, a second or final analysis was per-; formed using models that incorporate San Dieguito 1996 data, but not the salinity variable. For the re- productive success analyses in which groups were: smaller, fewer than 13 variables were incorporated, using F-to-enter from the preliminary analyses as! the criterion. Discriminant function analyses were conducted using the Systat computing package. Thej following variables were log transformed for anal-' ysis: ammonium, nitrate + nitrite, calcium, and cat- ion exchange capacity. Results Comparison of plant yield between marshes. One of the intents of our study was to compare attributes of a small, and perhaps unstable, population with those of a large and stable one. A component of this study involved assessment of demographic variables associated with survivorship and yield to determine the degree to which these populations ac- tually differ other than in estimated population size. As shown in Figure 2, there were statistically sig- nificant differences between marsh/years for mean survivorship (per 0.25 m- plot; ANOVA, F = 8.67, n = 30, P = 0.001), mean plant height (ANOVA, F = 56.8, n = 69, P < 0.001), mean capitulescence number (Kruskal-Wallis, Test Stat. = 46.3, n = 69, P < 0.001), mean capitulescence diameter (ANO- VA, F = 19.6, n = 69, P < 0.001), mean flower number (Kruskal-WaUis, Test Stat. = 28.5, n = 49, P < 0.001), mean number of seeds produced (seed number) (Kruskal-WaUis, Test Stat. = 54.3, n = 69, P = 0.001), and mean seed set (mean number of seeds/mean number of flowers) (ANOVA, F = 25.1, n = 49, P < 0.001). The mean percent cover; of L. glabrata subsp. coulteri within 0.25m2 sam- pling plots also differed between marsh/years PARSONS AND WHELCHEL: CLIMATIC VARIABILITY AND RARE PLANT 179 Survivorship Plant Heiglit 5 75% I 25% i Incomplete; through bud stage only J 2 J Q Q W Capitulescence Number (Zl C« IZ) MarshAf ear Flower Number I I I ■iP t t 9P OP 2!' 70 percent). Based on biweekly censuses for marked individuals, most mortality oc- curred just prior to or after reproduction. In 1997, total survivorship at reproductive maturity (flow- ering stage) was 72 percent for San Dieguito La- goon and 95 percent for San Elijo Lagoon. As the San Dieguito Lagoon population matured more quickly than San Elijo Lagoon, it is probable that some of the mortality occurred after reproduction and therefore actually constituted senescence. The results suggest that, at least in 1997, the populations were following the Deevey Type I survivorship curve characteristic of stable populations (Pavlik 1994), in that the mortality inflection point fol- lowed onset of seed production. Furthermore, as the data for San Dieguito 1997 was recorded during a below-average rainfall year, low mortality cannot necessarily be ascribed to above-average environ- mental conditions. Means for plant height, capitulescence diameter and number, and flower number also showed some interesting relationships. Means were not only gen- erally lower at San Dieguito Lagoon than at San Elijo Lagoon, but remarkably similar between years within the respective marshes. There was one ex- ception. In 1998, yield of the San Dieguito Lagoon plants was actually closer to that of the 1997, 1998, and 1999 San Elijo Lagoon plants. Means for plant height and capitulescence diameter suggested that San Dieguito Lagoon 1999 might be intermediate between low and high yield marsh/years, but those for capitulescence and flower number were equiv- alent to means in low yield years. Results for seed set and seed number were some- what more complex. Plants at San Dieguito Lagoon set less seed in 1997 than in 1996, and seed set (number of seeds/number of flowers) was lower in both of these marsh/years than in San Dieguito La- goon 1999 and San Elijo Lagoon 1997 and 1999. For the total number of seeds produced, however, the 1996 San Dieguito Lagoon marsh/year was the least productive. The distinction between the re- maining marsh/years was less clearcut, but in terms of seed productivity, the ranking appeared to be, from lowest to highest, as follows: San Dieguito Lagoon 1997 and San Elijo Lagoon 1998; San Die- j guito Lagoon 1999; San Elijo Lagoon 1999; San I Elijo Lagoon 1997; and San Dieguito Lagoon 1998. [ Unlike the 1997 survivorship data, these results i suggest that the San Dieguito Lagoon population | might be less stable than that of the larger San Elijo ! Lagoon one. Based on non-integrated demographic trend analysis, seed production per individual of stable populations should consistently equal or ex- ceed that of a common congener or more stable population (Pavlik 1994). With the exception of i 1998, seed production of the San Dieguito Lagoon i plants was typically lower than those at San Elijo Lagoon. The large differences observed in annual ' population size may have only exacerbated this dis- parity in seed production between populations. Plant yield and influence of biotic and abiotic factors. The dissimilar patterns in sample means observed for mean seed number and seed set and the other plant variables may relate to an underly- ing difference in how biotic and abiotic factors af- fect various stages or aspects of plant development. Based on these patterns, we decided to analyze our results by dividing our results into two grouping ; structures — reproductive potential and reproductive success. Reproductive potential measures the po- tential of the plant to be more reproductively suc- cessful through survivorship to reproduction (mor- tality), being larger (mean plant height), and pro- ducing more capitulescences (mean capitulescence number) and more flowers (mean capitulescence di- ameter and mean flower number). All of these vari- ables relate to an individual's ability to outcompete another in terms of attracting pollinators or utilizing limited resources (e.g., water, nitrogen, etc.). Re- productive success measures the actual success of an individual in reproducing, as determined by seed number and seed set (percentage of flowers pro- ducing seed). For dependent variables such as plant height, capitulescence diameter and number, and flower number, marsh/years generally split into two groups based on yield, with San Dieguito Lagoon 1996, 1997, and 1999 in a low yield group (Reproductive Potential 1/RPl) and San Dieguito Lagoon 1998 and San Elijo Lagoon 1997, 1998, and 1999 in a high yield group (Reproductive Potential 2). For the dependent variables seed number and seed set, groupings were less distinct, but marsh/years were separated into three groups, with San Dieguito La- goon 1996 in a low yield group (Reproductive Suc- cess 1/RSl), San Dieguito Lagoon 1997 and San Elijo Lagoon 1998 in an intermediate yield group (Reproductive Success 2/RS2), and San Elijo La- goon 1997 and 1999 and San Dieguito Lagoon 1998 and 1999 in a high yield group (Reproductive i Success 3/RS3). | Reproductive potential of a germinated seedling is typically affected by herbivory, environmental factors, and intra- and inter-specific competition. | 2000] PARSONS AND WHELCHEL: CLIMATIC VARIABILITY AND RARE PLANT 181 Herbivory can negatively affect individuals through consumption of either vegetative tissue or flowers, which may weaken or kill the plant. Through the three years of study, no herbivory of vegetative tis- sue or flowers was observed at either marsh. The effect of intra- and inter-specific competition is not as directly observable and can be more complicat- ed. At high intra- or inter-specific densities, seed- lings can compete for resources or light or become more attractive to herbivores. At later stages, how- ever, high densities of synchronously flowering in- dividuals, including non-native neighbors such as Cotula coronopifolia, may also serve to attract pol- linators and thereby enhance reproductive success. Some estimates of L. glabrata subsp. coulteri density were collected in 1997, and densities ranged from three individuals (San Elijo Lagoon) to 140 individuals (San Dieguito Lagoon) per dm-. Overall, the 1997 sampling plots at San Dieguito Lagoon had higher densities per dm- (71.6 ± SE 13.5) than those at San Elijo Lagoon (28.4 ± SE 7.2) (Mest, t = 2.83, n = 20, P = 0.011). Densities of other species were not estimated, but cover of other native species ranged from 0 to 75 percent, and cover of non-native species ranged from 0 to 50 percent. Extremely low total mortality rates for vegetative and flowering individuals at both marsh- es in 1997 (—72 to 95 percent) suggests that either abundance of L. glabrata subsp. coulteri or other species was not high enough, or resources not lim- ited enough, to have induced either intra- and inter- specific competition at the seedling or vegetative stage during this year. While above-average rainfall may have increased seedling densities at San Die- guito Lagoon in 1998, the fact that 89 percent of the plants reached at least bud stage suggests that densities were not high enough to incur density- dependent mortality. Reproductive success is affected by all the same factors as reproductive potential, but other factors can limit reproduction, as well, specifically grani- vory (herbivory of unfertilized ovules or seed) and, for non-vegetative species such as L. glabrata subsp. coulteri, pollination success. Based on sta- tistical analyses, reproductive success was highest for the RS3 group (San Elijo Lagoon 1997 and 1999 and San Dieguito Lagoon 1998 and 1999) and lowest in the RSI group (San Dieguito Lagoon 1996). The fact that marsh/years with technically equivalent reproductive potential (San Dieguito La- goon 1996 and 1998) should have differing rates of reproductive success suggests that a different factor or suite of factors may be affecting seed number and seed set. As noted earlier, no consumption of entire flowers was observed during the three years of study, and low mortality rates indicate that most individuals survived to flowering and seed set. It is possible that competition among individuals for resources in- creased during the flowering stage, as ambient tem- peratures and rates of evaporation and evapotrans- piration typically climb during the warm spring months. Flowering often coincides with a neap tide series, a period of extremely low tides that often decrease soil moisture and increase evaporation rates and soil salinity in higher marsh elevations. Some granivory was actually observed in seeds of San Dieguito Lagoon individuals in 1998. The extent of granivory was not quantified, but in gen- eral, the number of individuals and/or number of seeds per individual that appeared to have been af- fected was relatively low. While viability of the damaged seeds was not tested, the damage ap- peared extensive enough to render the seeds invi- able. Granivory, or pre-dispersal predation, was not observed in the other study years at this marsh, nor was it observed in seeds of plants from San Elijo Lagoon. The presence of organisms that would re- move seeds after dispersal, including ground-dwell- ing insects such as ants, was sporadic, and even when present, abundance was low. Ground-dwell- ing organisms observed within L. glabrata subsp. coulteri patches included ants (Formicidae), thrips (Thysanoptera), and rove (Staphylinidae) and other beetles {Bembidion sp. and Dermestidae) (Wesley Maffei personal communication). With the excep- tion of ants, these invertebrates are considered un- likely post-dispersal seed predators. None of these organisms, including the ants, were observed re- moving fallen seeds, nor were birds observed for- aging in these areas. Pollination was not included within the scope of our study, but many insects were observed visiting flowers during our sampling efforts. The primary insect visitors appeared to be solitary bees {An- drena pallidofovea and A. cercocarpi), beetles (Dermestidae, Geocoris sp.), beeflies (Bombyli- idae), flies {Bufolucilia sp. and Nemotelus sp.), but- terflies {Coenonympha californica), and halictine or ''sweat" bees {Lasioglossum sp.) (W. Maffei per- sonal communication; Robbin Thorp personal com- munication). Several of these visitors have the po- tential to effect pollination either through collecting pollen (e.g., Andrenidae or halictine bees) or for- aging on pollen or other flower parts (e.g., Der- mestidae). Overall, visitor numbers and species di- versity appeared to be lower at San Dieguito La- goon than at San Elijo Lagoon, although no formal pollinator observations were conducted. Based on these observations, we hypothesize that resources such as nutrients and perhaps even pollen may be the primary determinants or reproductive potential and success. Discriminant function analysis. To explore fur- ther the association between biotic and abiotic fac- tors and the groupings of marsh/years suggested by results of multiple comparison testing, discriminant function analyses were performed. The question posed by these analyses was two-fold. Using the groups suggested by multiple comparison testing, was there some combination of biotic and abiotic 182 MADRONO [Vol. 47 Table 2. Results of the Discriminant Function Analyses for the Reproductive Potential and Success Models. Reproductive Potential Model Canonical Variable: 0.002*pH - 0. 193*organicmatter - 1 .330*soilmoisture + 1.092*phosphorous - 0.005*potassium — 0.189*magnesium — 0.443 *totalplantcover — 0.406*ammonium_log — 0.1 13*nitrates trites-log + 0.326*non-nativeplantcover - 0.487*calcium_log - 0.699*CEC_log Classification results: Actual groups Results-Cases (%) Jackknifed Results-Cases (%) RPl RP2 RPl RP2 RPl RP2 Total % Correct 16(100) 0 0 15(100) 100 15(94) 0 1(6) 15(100) 97 Reproductive Success Model Canonical Variable 1: 0.386*pH - 1.235*soilmoisture - 0.651*organicmatter + 0.937*phosphorous - 0.438*potas- sium + 0. 106*magnesium + 0.422*totalplantcover + 0.756*ammonium_log + 1 .406*nitrates + nitritesJog Canonical Variable 2: 0.482*pH + 0.771*soilmoisture + 0.258*organicmatter — 1.265*phosphorous + 0.421*potas- sium + 0.786*magnesium + 0.253*totalplantcover + 0.754*ammoniumJog + 0.497*nitrates + nitritesJog Classification results: Actual groups Results — Cases (%) Jackknifed Results — Cases (%) RSI (%) RS2 (%) RS3 (%) RSI (%) RS2 (%) RS3 (%) RSI RS2 RS3 Total % Correct Variables measured 10(100) 0 0 0 11(100) 0 0 0 10(100) 100 10(100) 0 0 0 11(100) 0 Group Means for Models Reproductive Potential Reproductive Success RPl RP2 RSI RS2 0 0 10(100) 100 RS3 pH 7.59 7.83 7.43 7.78 7.90 % Soil moisture 29.6 40.3 25.9 37.2 41 % Organic matter 2.62 3.65 1.84 3.39 4.11 Phosphorous (ppm) 46.1 52.7 39.90 53.41 54.05 Potassium (ppm) 414.6 346.6 425.60 347.86 375.10 Magnesium (ppm) 1036.6 1104.2 1024.60 1030.82 1156.35 % Total plant cover 76.3 64 89.8 65.9 55.7 Log-Ammonium (ppm) 1.15 1.11 1.09 1.09 1.22 Log-Nitrates + Nitrites (ppm) 1.38 1.06 1.68 0.96 1.06 % Non-native plant cover 27 8.5 25.0NI 19.2NI 9.9NI Log-Calcium (ppm) 3.42 3.76 3.35NI 3.65NI 3.75NI Log-CEC (meq/lOOg) 1.75 2.26 1.83NI 2.12NI 2.04NI Soil salinity (ppt) 58.7NI 33. INI NA 38.6NI 42.4NI NA-Not available; NLNot included in models (see Methods and/or Results for explanation.) variables that would allow us to discriminate be- tween these groups? And, if so, what combination of variables would allow us to best predict the group to which the sampling location belonged? Two models were used. One model separated sam- pling locations into two groups (RPl, RP2) based on differences in sample means for reproductive potential variables, and another separated sampling locations into three groups (RSI, RS2, and RS3) based on differences in sample means for repro- ductive success variables. The groupings were es- sentially the same as described previously, except for the absence of San Dieguito Lagoon and San Elijo Lagoon 1999: no biotic and abiotic data were collected in 1999. The 13 biotic and abiotic vari- ables used in the models were: pH, salinity, soil moisture, organic matter, ammonium, nitrates + ni- trites, phosphorous, potassium, magnesium, calci- um, cation exchange capacity, total vegetation cov- er, and non-native plant species cover. While soil salinity appeared to be higher for RPl than RP2 (Table 2), it did not have a strong loading in pre- liminary analyses for either model or correlation with other variables and was therefore not incor- porated into final analyses. According to the reproductive potential analysis, the biotic and abiotic variables used discriminated well between the groups suggested by multiple comparison results (F = 8.34, n = 31, P < 0.0001). Canonical scores of group means were 2.21 for RPl and -2.36 for RP2. The canonical discrimi- nant function accounted for approximately 100 per- cent of the total dispersion in the data. Based on the standardized functions, most of the group sep- aration came from soil moisture, phosphorous, cal- cium, and cation exchange capacity. The canonical variable and a list of group means is provided in Table 2. Using the canonical variable, the model PARSONS AND WHELCHEL: CLIMATIC VARIABILITY AND RARE PLANT 10.0 183 MARSH/YEAR GROUPS RS1 RS2 I RS3 -7.0 -3.6 "0.2 3.2 6.6 CANONICAL VARIABLE 1 10,0 Fig. 3. Canonical scores plot for the reproductive success model of the discriminant function analysis. Soil moisture and nitrate + nitrite concentrations appeared to provide most of the separation between RSI and RS2/RS3. Abiotic factors were also able to separate RS2 and RS3, although the separation appeared weaker. was able to correctly predict or classify sampling locations 100 percent of the time (jackknifed clas- sification matrix = 97 percent). The reproductive success analysis was also suc- cessful at using nine of the biotic and abiotic vari- ables to discriminate between the three groups sug- gested by multiple comparison results (F = 18.04, n = 31, P < 0.0001; Fig. 3). Canonical scores of group means were (8.09, 0.03) for RSI, (-3.72, -1.29) for RS2, and (-4.00, 1.38) for RS3. The canonical discriminant function, which has two variables, accounted for 100 percent of the total dispersion in the data. The first canonical variable accounted for approximately 96 percent of the data dispersion. The canonical variables and a list of group means are provided in Table 2. As Figure 3 illustrates, most of the group separation comes from the first canonical variable, which appears to split RSI (San Dieguito 1996) from the other marsh/ years. The first canonical variable had strong load- ings for nitrates + nitrites, soil moisture, phospho- rous, ammonium, and organic matter. The second canonical variable involved separation of RS2 from RS3, although the degree of separation relative to RSI appeared weaker. At least three of the sam- pling locations from RS2 appeared to associate more strongly with RS3, while one of the RS3 lo- cations was grouped with RS2 (Fig. 3). The second canonical variable had strong loadings for phos- phorous, magnesium, soil moisture, ammonium, and nitrates + nitrites. Using the canonical vari- ables, the model was able to correctly classify sam- pling locations 100 percent of the time (jackknifed classification matrix = 100 percent). The strong separation of marsh/year groups ef- fected by soil moisture, phosphorous, cation ex- change capacity, and calcium supports the premise that resources are the primary limiting factors of reproductive potential. Standardized canonical co- efficients and group means for soil moisture, cation exchange capacity, and calcium concentrations point to a positive, perhaps even linear, relationship between resource factor and plant variable (Table 2 — canonical variable). The relationship between phosphorous concen- trations and L. glabrata subsp. coulteri yield ap- pears somewhat more complicated than that for soil moisture, cation exchange capacity, and calcium concentrations. Standardized coefficients suggest that elevated phosphorous concentrations may ac- tually drive the canonical score toward RPl, which had a low yield. In contrast, group means show highest phosphorous concentrations in marsh/years with the highest reproductive potential. This dis- parity between group means and standardized co- efficients for phosphorous also occurred in the re- productive success model. As shown in Figure 3, resources also appear to play a role, if perhaps a more limited one, in re- productive success. As hypothesized earlier, the suite of factors influencing vegetative yield ap- peared to be slightly different from that affecting seed production, which may account for the differ- ence in reproductive success observed between marsh/years with similarly sized individuals (San Dieguito Lagoon 1996 and San Dieguito Lagoon 1997). While soil moisture and phosphorous con- centrations featured prominently in both analyses. 184 MADRONO [Vol. 47 nitrogen concentrations (ammonium and nitrates + nitrites) appeared to have a larger effect on repro- ductive success than on reproductive potential. As was the case with phosphorous, standardized coef- ficients for nitrates + nitrites and ammonium sug- gest that elevated concentrations of inorganic nitro- gen may actually drive canonical scores toward RSI and a reduced reproductive yield, although group means for ammonium were marginally high- er in RS3 than in either RSI or RS2. Based on standardized coefficients and group means, only or- ganic nitrogen sources such as organic matter ap- peared to contribute directly to enhanced seed pro- duction. The evidence for resource limitation of seed pro- duction is somewhat weaker for the RS2 and RS3 groups. The reproductive success analysis did pro- vide at least enough separation between RS2 and RS3 using canonical variable 2 to enable successful group differentiation and prediction. As with ca- nonical variable 1 (Table 2), standardized coeffi- cients for phosphorous in canonical variable 2 again appear to drive canonical scoring toward re- duced yield, despite the fact that phosphorous con- centrations were slightly higher in the group pro- ducing the most seeds (RS3). However, in contrast to canonical variable 1, inorganic nitrogen, along with soil moisture and magnesium, appeared to play a positive role in influencing reproductive yield. While the biotic and abiotic factors included in the analysis do enable successful separation be- tween RS2 and RS3, the slight to moderate overlap between groups displayed graphically in Figure 3 suggest that, at some resource level, the number of seeds produced may be driven by other factors not included in this analysis, the most probable of which is pollen supply. Discussion Soil moisture would seem an unlikely constraint in a salt marsh, but the high marsh represents a distinct ecotone in an aquatic environment. In gen- eral, high marsh species must contend with a com- plex series of hydrologic cycles: days or even weeks of flooding in the winter may be followed by months where the high marsh or marsh periph- ery is only inundated or saturated from subsurface flow during the highest high tides. The hydrologic complexity is compounded in managed lagoons, where the lagoon may be flooded deliberately to attract waterfowl or the tidal inlet may remain closed for most of the year even after winter storms elevate internal water levels. In several instances, water or moisture stress has been singled out as a primary factor limiting growth of species in the up- per marsh zones (Boorman 1971; De Leeuw et al. 1990). Conversely, too much water or waterlogging can negatively affect species adapted to the typi- cally well-drained soils of the high marsh or marsh periphery (Phleger 1971; Nestler 1977; Parrondo et al. 1978; Cooper 1982; Seliskar 1985; Adams and I Bate 1994). | Waterlogging may account for the anomalous re- 1 suits recorded in 1998, when survival (L. Parsons personal observation) and reproductive yield of the San Elijo Lagoon population plummeted and was significantly less than that of San Dieguito Lagoon. Nineteen ninety-eight was the one year during our study when rainfall was above average (178 percent of average during the months October- April; Table 1) . During that year, back-up of run-off and creek flows kept water levels within the eastern area of the lagoon substantially elevated for weeks. In gen- eral, reproductive yield of this population was ac- tually highest in the two years where rainfall was slightly below average — 1997 (73 percent of aver- age) and 1999 (67 percent of average). In below- average years, the current hydrologic management regime, in which the sluice gates are closed for wa- terfowl enhancement and outflow is provided through dips in a dike system, may actually en- hance the population by artificially maintaining sat- urated soil conditions within the eastern portion of the lagoon. Conversely, the response of the San Dieguito Lagoon population to rainfall is more con- sistent with plants being limited by lack of water. Above-average rainfall during 1998 was directly associated with dramatic increases in vegetative and reproductive yield. The positive association be- tween rainfall and yield, combined with the strong evidence of resource limitation in discriminant function analyses, suggests that, at San Dieguito Lagoon, rainfall both directly and indirectly boosts input and cycling of resources such as water and nutrients. The importance of nutrient limitation in coastal salt marsh plant communities has been well docu- mented (Tyler 1967; Pomeroy et al. 1969; Valiela and Teal 1974; DeLaune et al. 1979; Smart 1982; Long and Mason 1983; Mitsch and Gosselink 1986; Covin and Zedler 1988; Langis et al. 1991; Parsons and Zedler 1997; Boyer and Zedler 1988 and 1999). Our results generally show that higher yields are linked to higher nutrient concentrations. The seemingly negative relationship between phospho- rous and inorganic nitrogen concentrations and plant yield observed in analyses could have resulted from some indirect effect of nutrient influx, such as greater competition with more abundant species for light, moisture, or nutrients (BoUens et al. 1998). However, neither total plant cover or cover of non- native species factored strongly into the discrimi- nant function analyses. Based on group means showing elevated levels of phosphorous and, to some extent, ammonium in high yield plots (Table 2) , it is more probable that these nutrients must in- teract with other resource variables in such a way that yield is maximized in areas with moderate con- centrations of phosphorous and inorganic nitrogen. If such an interaction exists, our analyses were not sensitive enough to detect it, as no strong correla- 2000] PARSONS AND WHELCHEL: CLIMATIC VARIABILITY AND RARE PLANT 185 tion was evident between biotic and abiotic depen- dent or predictor variables (correlation <71 per- cent). In general, plasticity in growth or reproduction in relation to rainfall and changes in soil moisture and nutrient input should be expected in annual plant species within Mediterranean climates, even in aquatic systems such as salt marshes. These op- portunistic life forms must rely almost entirely on nature's largesse to propagate, survive, and succeed as they have none of the mechanisms (e.g., deep taproots, strongly developed mycorrhizal associa- tions, waxy cuticle layer on leaves, etc.) that enable perennial plants to cope with drought and other cli- matic challenges. Several studies on salt marsh an- nuals, including an occurrence of L. glabrata subsp. coulteri at Carpinteria Marsh near Santa Barbara, California, have linked above average rainfall to increases in relative abundance (Allison 1992; Par- sons and Zedler 1997) and density, distribution, and biomass (Callaway and Sabraw 1994). Terrestrial species are also strongly influenced by soil moisture (Reynolds et al. 1997; Center for Conservation Bi- ology 1994), with yield for grassland members of Lasthenia such as L. californica optimized both during wet years and when growing in wet micro- sites (Hobbs and Mooney 1991, 1995). For peren- nial species, the effect of below average rainfall may be more subtle than for annuals, though no less significant, resulting in substantial reductions in seed set (Morgan 1999) and ultimately recruitment and population growth rates (Maschinski et al. 1997). Plasticity in reproduction can be exacerbated by inter-annual variability in other types of "re- sources" such as pollen. While no information ex- ists on the mating system of southern California coastal populations, in general, L. glabrata subsp. coulteri has been categorized as one of the 14 of 17 Lasthenia species that is self-incompatible (Orn- duff 1966). Several species of insects such as bees, beeflies, flies, and beetles were observed visiting flowers, although what role these species have in effecting pollination of L. glabrata subsp. coulteri is unknown. Based on the species' presumed status as an entomophilous outcrosser, reproductive yield must depend to some degree on pollination success. As with their host species, pollinators, some of which are believed to nest in marshes or adjacent upland areas, can be affected by climatic variations and watershed disturbances, including flooding (Stephen et al. 1969). Given the myriad of ecological interactions in- volved, it is not surprising that the factors govern- ing reproductive potential and success of L. glabra- ta subsp. coulteri may prove complex both in terms of time and scale. A number of recent studies have supported the potential for spatial or temporal het- erogeneity in resource and pollen limitations (McCall and Primack 1987; Zimmerman and Aide 1989; Campbell and Halama 1993; Lawrence 1993; Parsons 1994; Parsons and Zedler 1997). Our study supports not only inter-annual heterogeneity in re- source limitations, but possibly intra-annual hetero- geneity, as well. For example, while reproductive potential and rainfall totals between October- March were similar for the 1996 and 1999 San Die- guito Lagoon populations, seed set was higher in 1999 than in 1996. A series of storms in early April 1999 may have eased resource constraints during the seed set period, allowing the sparse population of small plants to produce comparatively larger numbers of seed. In general, however, the complex hydrology of urbanized watersheds with dams, year-round urban run-off, and mouth closures would seemingly argue against a tight linkage be- tween rainfall patterns and resource inputs and cy- cling. Conclusions As we originally surmised, the L. glabrata subsp. coulteri population at San Dieguito Lagoon is not only smaller than the one at San Elijo Lagoon, but, based on trends in seed production, less stable, as well. For the most part, plants at San Dieguito La- goon were smaller and produced less flowers and capitulescences and seed than those at San Elijo Lagoon. However, the nature of this relationship appears to be temporally variable and highly de- pendent on climatic conditions such as rainfall to- tals and distribution. In a year with above-average rainfall, yield of the San Dieguito Lagoon popula- tion was similar to and, in some ways, greater than that of the more stable one at San Elijo Lagoon. As rainfall is often linked directly and indirectly to in- puts and cycling of resources such as water and nutrients, the strong association found between re- sources and reproductive potential and, to some ex- tent, reproductive success is certainly not surpris- ing, although the relationship was not always either simple or linear. Too much water actually appeared to decrease survival and reproductive yield of the 1998 San Elijo Lagoon population by inducing "waterlogging." In addition, some nutrients such as inorganic nitrogen and phosphorous may require higher levels of other resources such as soil mois- ture before exerting a positive effect on growth or reproduction of L. glabrata subsp. coulteri. When resources are sufficient, seed production appears to be limited by other "resources," the most probable of which is pollen supply. The importance of the relationship between cli- matic conditions and population productivity as- sumes a deeper significance when considering the long-term viability of the small San Dieguito La- goon population. Obviously, less seed will be pro- duced in years when few plants are present or plant vigor is reduced. Still, even when the San Dieguito Lagoon population was relatively large and pro- duced more seed per plant than the San Elijo La- goon one, productivity of the San Dieguito Lagoon 186 MADRONO [Vol. 47 population as a whole was still comparatively low- er, because of the difference between marshes in population size. To some extent, the impact of con- sistently producing small numbers of seed could be offset if seed banks are long-lived and/or seed vi- ability and germination rates are high. No research has been specifically conducted on seed bank lon- gevity of L. glabrata subsp. coulteri, but studies on various Lasthenia species have documented long- lived seed banks (10 y; Vivrette 1999) and high germination rates in the field (25 to 69 percent; Thorp 1976) and laboratory (34 to 90 percent; Kingsbury et al. 1976; Callaway et al. 1990; Ra- jakaruna and Bohm 1999; Michael Wall personal communication, March 1999; Doug Gibson unpub- lished data). However, there are indications that germination or emergence from the seed bank for some Lasthenia species may be tightly regulated by the same climatic conditions (Vivrette 1999) that appear to negatively affect yield of L. glabrata subsp. coulteri, at least at San Dieguito Lagoon. Some evidence for this could be seen in the low number of plants present at San Dieguito Lagoon in 2000 (six plants; A. Thorpe personal communi- cation), when rainfall during the primary germina- tion period (October — January) totaled only 10.4 percent of average (San Diego NWS-Lindbergh; California Department of Water Resources, Cali- fornia Data Exchange Center). In drought years, then, both recruitment and individual yield could be reduced, thereby further diminishing productiv- ity of the population as a whole. Poor recruitment and yield in all but above-av- erage rainfall years is of concern for populations in a region such as southern California, where above- average rainfall years are sporadic and often inter- spersed between lengthy periods of drought or be- low-average rainfall. In San Diego County, below- average rainfall occurs 60 percent of the time, while above-average rainfall occurs about 40 percent of the time (Elwany et al. 1998). There are sugges- tions that variability of this already extremely vari- able climate may be increasing due to global warm- ing. Chronically low numbers of plants in average to below-average rainfall years can increase popu- lations' susceptibility to genetic bottlenecks or ex- tinction due to stochastic or disturbance-related events. Long-term viability of small populations such as San Dieguito Lagoon will probably depend on whether the species can germinate and repro- duce successfully under average, as well as above- average, rainfall and climatic conditions. Future monitoring efforts should focus on assessing repro- ductive potential and success of this population un- der a variety of climatic and hydrologic conditions, as well as better defining pollinator relationships, breeding system, survivorship, seed bank dynam- ics, and field germination rates of L. glabrata subsp. coulteri. Implications for Management and Restoration The information from this study will provide both preserve and restoration managers with some guidelines for future efforts to enhance or even re- introduce L. glabrata subsp. coulteri into salt marshes. Based on our results, L. glabrata subsp. coulteri grows best in marshes with moist, but not waterlogged, soils with low to moderate salinity, high cation exchange capacity, high percentage of organic matter, and moderate concentrations of phosphorous, calcium, and possibly ammonium. To ensure a high potential for project success, man- agers interested in conducting enhancement or re- introduction projects should carefully evaluate site conditions and hydrologic management regimes. While the goal of restoration and enhancement con- tinues to revolve around creation of self-sustaining ecosystems, the reality is that many of our wetland ecosystems are now highly managed through tide or sluice gates, dikes, culverts, mechanical mouth breaching, and even deliberate floodings to attract waterfowl. If management cannot be avoided, it can perhaps be manipulated to provide benefits to spe- cies other than waterfowl. Indeed, the high yield recorded at San Elijo Lagoon in years with below- average rainfall may result in part from artificially elevated soil moisture conditions created by back- up of run-off and creek flows when sluice gates are closed during the winter. While, as a science, restoration ecology has moved away from a single-species management ap- proach, there is still a strong need for single-spe- cies-focused research. Without carefully under- standing the biotic and abiotic relationships that drive individual species within an ecosystem, we might be tempted to make gross generalizations about the habitat linkages without ever really grasp- ing the framework of those linkages. For example, what functions of the high marsh are particularly important for L. glabrata subsp. coulteri, and how do these needs complement or detract from those of other species inhabiting this fragile ecotone, such as Cordylanthus maritimus Benth. subsp. maritimus or Panoquina errans (wandering skipper butterfly)? Directed research on each of these species provides the pieces for the larger ecosystem puzzle. It is up to restoration and preserve managers to put the puz- zle together in a manner that will maximize benefits for as many species as possible, as well as the eco- system as a whole. Acknowledgments The authors would like to thank Doug Gibson, John Boland, Troy Kelly, and Roger Briggs for their assistance with this project. We also thank Bonnie Peterson for her contribution to data collection. The comments of Kathy Boyer, Jessica Martini-Lamb, and two reviewers improved this manuscript. 2000] PARSONS AND WHELCHEL: CLIMATIC VARIABILITY AND RARE PLANT 187 Literature Cited Adams, J. B. and G. C. Bate. 1994. The effect of salinity and inundation on the estuarine macrophyte Sarco- cornia pererinis (Mill.) A. J. Scott. Aquatic Botany 47:341-348. Allison, S. K. 1992. The influence of rainfall variability on the species composition of a northern California salt marsh plant assemblage. Vegetatio 101:145-160. Barrett, S. C. H. and J. R. Kohn. 1991. Genetic and evolutionary consequences of small population size in plants: implications for conservation. Pp. 3-30. in D. A. Falk, and K. E. Holsinger (eds.). Genetics and conservation of rare plants. Oxford University Press, New York, N.Y. BoLLENS, U., S. GusEWELL AND F. Klotzli. 1998. On the relative importance of nutrient availability and water level for the biodiversity of wet meadows. Bulletin of the Geobotanical Institute ETH 64(0):91-101. Boorman, L. a. 1971. Studies in salt marsh ecology with special reference to the genus Limonium. Journal of Ecology 59:103-120. BoYER, K. E. AND J. B. Zedler. 1998. Effects of nitrogen additions on the vertical structure of a constructed cordgrass marsh. Ecological Applications 8(3):692- 705. Boyer, K. E. and J. B. Zedler. 1999. Nitrogen addition could shift plant community composition in a restored California salt marsh. Restoration Ecology 7(1):74- 85. California Wetlands Information System. 1996. Cah- fornia Environmental Research Exchange System (CERES). California Resources Agency, Sacramento, CA. Electronic communication (web site). Callaway, R. M., S. Jones, W. R. Ferren, Jr. and A. Parikh. 1990. Ecology of a mediterranean-climate es- tuarine wetland at Carpinteria, California: plant dis- tributions and soil salinity in the upper marsh. Ca- nadian Journal of Botany 68: 1 139-1 146. Callaway, R. M. and C. S. Sabraw. 1994. Effects of variable precipitation on the structure and diversity of a California salt marsh community. Journal of Vege- tation Science 5:433-438. Campbell, D. R. and K. J. Halama. 1993. Resource and pollen limitations in seed production in a natural plant population. Ecology 74(4): 1043-105 1 . Center for Conservation Biology. 1994. Conservation of the palmate-bracted bird's beak, Cordylanthus pal- matiis. Stanford University, Stanford, CA. Charlesworth, D. and B. Charlesworth. 1987. In- breeding depression and its evolutionary conse- quences. Annual Review of Ecology and Systematics 18:237-268. Cooper, A. 1982. The effects of salinity and waterlogging on the growth and cation uptake of salt marsh plants. New Phytologist 90:263-275. Covin, B. and J. B. Zedler. 1988. Nitrogen effects on Spartina foliosa and Salicoruia virginica in the salt marsh at Tijuana Estuary, Calif. Wetlands 8:51-65. Davy, A. J. and R. L. Jefferies. 1981. Approaches to the monitoring of rare plant populations. Pp. 219-232. in H. Synge, (ed.). The biological aspects of rare plant conservation. John Wiley & Sons, London, United Kingdom. DeLaune, R. D., R. J. BuRESH and W. H. Patrick, Jr. 1979. Relationship of soil properties to standing crop biomass of Spartina alterniflora in a Louisiana marsh. Estuarine Coastal Marine Science 8:477-487. De Leeuw, J., H. Olff and J. P. Barker. 1990. Year-to- year variation in peak above-ground biomass of six salt marsh angiosperm communities as related to rain- fall deficit and inundation frequency. Aquatic Botany 36:139-151. Elwany, M. H. S., R. E. Flick and S. Aijaz. 1998. Open- ing and closure of a marginal southern California la- goon inlet. Estuaries 2 1 (2):246-254. Elwany, M. H. S. 1999. Professor. Center for Coastal Studies, Scripps Institution of Oceanography, La Jol- la, CA. Unpublished data. Gardner, W H. 1986. Water content in Methods of Soil Analysis, part 1: physical and mineralogical methods, 2nd ed. American Society of Agronomy — Soil Sci- ence Society of America. Madison, WI. HaMRICK, J. L., Y. B. LiNHART AND J. B. MiTTON. 1979. Relationships between life history characteristics and electrophoretically detectable genetic variation in plants. Annual Review of Ecology and Systematics 10:173-200. Harper, J. L. 1977. Population biology of plants. Aca- demic Press. London, United Kingdom. Hedrick, p. W. 1983. Genetics of populations. Science Books International. Boston, MA. Hickman, J. C. 1993. The Jepson manual: higher plants of California. University of California Press, Berke- ley. Kingsbury, R. W, A. Radlow, P. J. Mudie, J. Ruther- ford AND R. Radlow. 1976. Salt stress responses in Lasthenia glabrata, a winter annual composite en- demic to saline soils. Canadian Journal of Botany 54: 1377-1385. Langis, R., M. Zalejko and J. B. Zedler. 1991. Nitrogen assessments in a constructed and natural salt marsh of San Diego Bay. Ecological Applications 1:40-51. Lawrence, W. S. 1993. Resource and pollen limitation: plant size-dependent reproductive patterns in Physalis longifolia. American Naturahst 141(2):296-313. Ledig, F. T. 1986. Heterozygosity, heterosis, and fitness in outbreeding plants, p. 77-104. in M. E. Soule, (ed.). Conservation biology: the science of scarcity and di- versity. Sinauer Press, Sunderland, MA. Long, S. P. and C. F. Mason. 1983. Saltmarsh ecology. Cambridge University Press, Cambridge, United Kingdom. Maschinski, J., R. Frye and S. Rutman. 1997. Demog- raphy and population viability of an endangered plant species before and after protection from trampling. Conservation Biology 1 1(4):990-999. McCall, C. and R. B. Primack. 1987. Resources limit the fecundity of three woodland herbs. Oecologia 71: 431-435. MEC Analytical Systems, Inc. 1993. San Dieguito La- goon restoration project biological baseline study, March 1992-May 1993, Vol. I, 2nd ed. Draft tech- nical memorandum. Prepared for Southern California Edison. Rosemead, CA. Menges, E. S. 1986. Predicting the future of rare plant populations: demographic monitoring and modeling. Natural Areas Journal 6(3): 13-25. Menges, E. S. and D. R. Gordon. 1996. Three levels of monitoring intensity for rare plant species. Natural Areas Journal 16(3):227-237. MiTSCH, W. J. AND J. G. Gosselink. 1986. Wedands. Van Norstrand Reinhold Company, New York, NY. Morgan, J. W. 1999. Effects of population size on seed production and germinability in an endangered, frag- 188 MADRONO [Vol. 47 merited grassland plant. Conservation Biology 13(2): 266-273. Natural Diversity Data Base (NDDB). 1998. Califor- nia Department of Fish and Game. Sacramento, CA. Nei, M. Molecular population genetics and evolution. El- sevier, New York, N.Y. Nestler, J. 1977. Interstitial salinity as a cause of eco- phenic variation in Spartina alterniflora. Estuarine and Coastal Marine Science 5:707-714. Ornduff, R. 1966. A biosystematic survey of the gold- field genus Lasthenia (Compositae:Helenieae). Uni- versity of California Publications in Botany 40:1-92. Parrondo, R. T, J. G. Gosselink and C. S. Hopkinson. 1978. Effects of salinity and drainage on the growth of three salt marsh grasses. Botanical Gazette 139(1): 102-107. Parsons, L. S. 1994. Re-establishment of salt marsh bird's beak at Sweetwater Marsh: factors affecting repro- ductive success. Master's Thesis, San Diego State University, San Diego, CA. Parsons, L. S. and J. B. Zedler. 1997. Factors affecting reestablishment of an endangered annual plant at a California salt marsh. Ecological Applications 7(1): 253-267. Pavlik, B. M. 1994. Demographic monitoring and the re- covery of endangered plants. Pp. 322-350. in M. L. Bowles, and C. J. Whelan (eds.). Restoration of en- dangered species: conceptual issues, planning and im- plementation. Cambridge University Press, Cam- bridge, United Kingdom. Phleger, C. F. 1 97 1 . Effect of salinity on growth of salt marsh grass. Ecology 52:908-911. Pomeroy, L. R., R. E. Johannes, E. P. Odum and B. Roff- man. 1969. The phosphorous and zinc cycles and pro- ductivity of a salt marsh. Pp. 412-419. in D. J. Nel- son, and F. C. Evans (eds.). Proc. 2nd Natural Sym- posium on Radioecology. Powell, A. H. and G. V. N. Powell. 1987. Population dynamics of male euglossine bees in Amazonian for- est fragments. Biotropica 19:176-179. Purer, E. A. 1942. Plant ecology of the coastal salt marshlands of San Diego County, California. Ecolog- ical Monographs 12(1):83-1 1 1. Rajakaruna, N. and B. A. Bohm. 1999. The edaphic fac- tor and patterns of variation in Lasthenia californica (Asteraceae). American Journal of Botany 86(11): 1576-1596. Reynolds, H. L., B. A. Hungate, F. S. Chapin, III and C. M. D' Antonio. 1997. Soil heterogeneity and plant competition in an annual grassland. Ecology 78(7): 2076-2090. San Diego County Department of Parks, and Recre- ation. Unpublished data. Recording of date of mouth opening, duration of opening, and type of opening (artificial vs. natural). San Diego, CA. ScHEMSKE, D. W., B. C. Husband, M. H. Ruckelshaus, C. Goodwillie, I. M. Parker, and J. G. Bishop. 1994. Evaluating approaches to the conservation of rare and endangered plants. Ecology 75(3):584-606. Sea Science Services, and Pacific Southwest Biologi- cal Services, Inc. 1980. Report on environmental baseline studies for the San Dieguito Lagoon En- hancement Plan, Del Mar, California. Prepared for the Department of Planning and Community Develop- ment, City of Del Mar and the State of California Coastal Conservancy. Seliskar, D. M. 1985. Morphometric variations in five tidal marsh halophytes along environmental gradi- ents. American Journal of Botany 72(9): 1340- 1352. Shaffer, M. L. 1981. Minimum population sizes for spe- cies conservation. Bioscience 31:131-134. Skinner, M. W. and B. M. Pavlik. 1994. Inventory of rare and endangered vascular plants of California. California Native Plant Society, Sacramento, CA. Smart, R. M. 1982. Distribution and environmental con- trol of productivity and growth form of Spartina al- terniflora. in D. N. Sen, and K. S. Rajpurokit (eds.). Contributions to the ecology of halophytes. Dr. Junk, The Hague, The Netherlands. SOKAL, R. R. AND F J. ROHLF. 1981. Biometry. W. H. Freeman and Company, New York, N.Y. Stephen, W. P, G. E. Bohart and R F Torchio. 1969. The biology and external morphology of bees. Agri- cultural Experiment Station, Oregon State University, Corvallis, OR. Thorp, R. W. 1 976. Insect pollination of vernal pool flow- ers. Pp. 36-40. in S. Jain (ed.). Vernal pools, their ecology, and conservation. Institute of Ecology Pub- lication No. 9. University of California-Davis, Davis, CA. Tyler, G. 1967. On the effect of phosphorous and nitro- gen, supplied to Baltic shore-meadow vegetation. Bo- taniska Notiser 120:433-447. Valiela, I. AND J. M. Teal. 1974. Nutrient limitation in salt marsh vegetation. Pp. 547-563. in R. J. Reinhold and W. H. Queen (eds.). Ecology of halophytes. Ac- ademic Press, New York, NY. Vivrette, N. 1999. Coastal bluff vegetation change over 25 years on Santa Cruz Island in Fifth California Is- lands Symposium: schedule of symposium proceed- ings and events and presentation and poster abstracts. March 29-April 1, 1999. Wall, M. 1999. Seed curator. Rancho Santa Ana Botanic Garden at Claremont, Claremont, CA. Personal com- munication. Zar, J. H. 1984. Biostatistical analysis. Prentice-Hall, En- glewood Cliffs, NJ. Zimmerman, J. K. and T. M. Aide. 1989. Patterns of fruit production in a neotropical orchid: pollinators vs. re- source limitation. American Journal of Botany 76(1): 67-73. Madrono, Vol. 47, No. 3, pp. 189-194, 2000 PLEISTOCENE MACROFOSSIL RECORDS OF FOUR-NEEDLED PINYON OR JUNIPER ENCINAL IN THE NORTHERN VIZCAINO DESERT, BAJA CALIFORNIA DEL NORTE Philip V. Wells Department of Ecology and Evolutionary Biology, University of Kansas, Lawrence, KS 66045 Abstract Two late-Pleistocene Neotoma (wood rat) middens have been dated by four radiocarbon analyses at 10,000-10,200 and 17,470 radiocarbon years. Both deposits document by numerous macrofossils the abundance of Jimiperus californica Carriere, but in the older deposit dominance is shared with Pinus quadrifolia Pari. Both deposits contain lesser quantities of the principal dominant shrub of the California chaparral, the chamise {Adenostoma fasciciilcitwn Hook. & Arn.) together with the shrubby oak, Quercus turhinella E. Greene, and other chaparral genera at about 30°N in the northern part of the Vizcaino desert. The existing desert vegetation at both sites is dominated by giant columnar xerophytes and several species of low, desert shrubs, no trace of which has been detected in either of the dated middens. Although abundance of macrofossils of woodland trees with lesser amounts of chaparral shrubs, in conjunction with absence of any species of desert shrub, document a modest displacement of desert vegetation at moderate elevations (550-594 m) in the northernmost Vizcaino desert, this evidence cannot be extrapolated to include the entire peninsula of Baja California. Very substantial biogeographic and ecologically pertinent physiographic evidence suggest a major desert barrier in the central part of the peninsula that also may explain the high degree of endemism in the desert flora. Resumen En Desierto Vizcaino, Baja California del Norte, Mexico, en 550-594 m, dos depositos de epoca tardo- Pleistocene con datos 10,000-10,200 y 17,470 anos, documentan con macrofosiles numerosos la presencia de Pinus quadrifolia y/o Juniperus californica, con arbustos de encinal (Quercus turbinella, Adenostoma fasciculatum). Sin embargo, arbustos de desierto moderno ausente en los dos depositos. Some of the most spectacular desert vegetation in North America occupies a relatively restricted sector of northern Baja California, north of the Viz- caino Plain in the central part of the peninsula. Shreve (1951), who was very familiar with all the North American deserts, referred to it as a desert wonderland (on p. 108). This unique assemblage of bizarre xerophytes of gigantic stature extends from just south of the lofty Sierra San Pedro Martir east of El Rosario on the Pacific Coast, south to about San Borja, a distance of about 250 km. The most spectacular scenery is more localized, reaching a peak in areas of weathered granitic rocks, such as the extensive tract north of Santa Ines or farther south near Portezuelo. The unique plant is the cirio tree {Idria colum- naris Kell.), a remote relative of the common oco- tillo {Fouquieria splendens Engelm. also present) of the same family, but the cirio attains a height of more than 16 m and has a single main trunk with spongy wood that stores water It has been likened to an inverted, spiny carrot with innumerable short branches arranged in dense spiral phyllotaxy. Al- though it resembles a giant columnar cactus, it has ordinary C3 photosynthesis, unlike most cacti which have crassulacean acid metabolism (CAM). The cirio is practically endemic to this sector of Baja California, reaching its southern limit on the high Tres Virgenes volcanoes at over 1600 m, east of San Ignacio. The equally impressive cardon cac- tus (Pachycereus pringlei, (S. Wats.) Britt. & Rose) has about the same northern limit on the Pacific slope, but extends south to the Cape in Baja Cali- fornia del Sur where it is the ubiquitous dominant giant of desert elevations. Another pachycaul with a swollen trunk, the el- ephant tree {Pachyconmis discolor (Benth.) Covi- lle) is prominent in the cirio area but extends south with the cardon, though not nearly to the Cape. Closer to the cardon in distribution is another dis- tinctive quasi-endemic tree, Viscainoa genicidata (Kell.) Greene, which associates with cirio in its more limited range. Sprawling beneath the vertical cardons is another large cactus, but of a horizontal mode of growth, the endemic Mcichaerocereus gummosus, (Engelm.) Britt. & Rose, with a wind- swept appearance and tart fruits ("pitahaya agria"). The lesser cacti include various species of Opuntia, especially chollas {O. cholla Weber, O. molesta Brandegee). A lesser succulent is the euphorbia- ceous Pedilanthus macrocarpus Benth. The domi- nant understory shrubs are mainly dull gray like Ambrosia chenopodii folia (Benth.) Payne, enliv- ened by weedy patches of Encelia californica Nutt. with its masses of sunflowers, and punctuated by the scarlet flowers of Beloperone shrubs (seen after the winter rains). 190 MADRONO [Vol. 47 Table 1 . Plant Composition of Neotoma Middens from the Cirio-Cardon Desert of Baja California del Norte, j Locations shown on Fig. 1: site 1 = northwest of Mision San Fernando at ca. 30°N; site 2 = northwest of Ranchoi Santa Ines at 29°46'N. Relative abundances: + + + = major constituent; ++ = lesser; + = trace. 1 SITES: San Fernando, at 594 m Santa Ines, at 550 m AGES: 10,000-10,200 ± 135 17,470 ± 200 (UCLA: 1365, 1366, (Beta 9372) 1367) MACROFOSSILS CONIFERS Pinus quadrifolia (intact leaf fascicles) Juniperiis californica (leafy twigs, seeds) CHAPARRAL Adenostoma fasciculatum van obtusifolium (leaves) Querciis turbinella (acorns, leaves) Primus lyonii (leaves, endocarps) Arctostaphylos glandulosa (nutlets of drupe) Eriodictyon angustifolium (leaf) absent + + + + + + + It was in this most spectacular part of Baja Cal- ifornia that I first sought for evidence of Pleistocene vegetation early in 1968. The Neotoma method (Wells 1976) had already proven fruitful in the Mo- have, western Sonoran ("Colorado" desert, Cali- fornia), and Chihuahuan deserts (Wells 1966; Wells and Berger 1967). Equipped with a % ton 4X4 pickup and camper, my lifetime friend. Jack Yri- zarry, and I crawled down the thousand-mile desert track to the Cape. "The Baja Run" was then little more than a very rough, rocky trail with incredibly steep grades in places and was strictly one lane in the best stretches, where a speed of 15 mph might be briefly reached! The average speed with a big, rocking camper was about 5 mph. That there was any "road" at all was due to a sparse procession of big-wheeled trucks driven by native truckers, who provided "servicio particular" to isolated local rancheros. Methods and Materials The Neotoma macrofossil method (first mono- graphed in Wells 1976) is deceptively simple, but requires thorough training in taxonomic and mor- phological botany and proficiency with existing flo- ras over wide areas of North America. Furthermore, it is essential that proficiency extends to minute de- tails exhibited by numerous tiny macrofossils. In- terpretation of the results requires training in ecol- ogy of the vegetation of North America and de- tailed knowledge of its physiography, geology, and climatology. Skilled sampling of the deposits is even more essential. The key to critical stratigraphic analysis is a focus on friable, macrofossil-rich layers. The friable stratigraphic layers are split off as separate units that disintegrate in the dry state into their con- stituent macrofossils. Thus, bulk-processing of mid- dens in homogenizing water baths is avoided. In- stead, the method of dry-processing provides a co- pious yield of macrofossils from discrete strati- graphic units that are carefully sorted on a multiple-mesh seive set. The macrofossils are sort- ed as to species, weighed, and reported as percent biomass, or with simpler assemblages assigned rel- ative abundances. Most species are present as mere traces. There is often an overwhelming dominance of woody plants that fully justifies the vernacular name "wood rat," coined by mammalogists for the genus Neotoma. Aliquots of the same friable strati- graphic layers burned for carbon dating are curated separately as vouchers for documentation and fur- ther study. Results Less than a week of crawling south of Ensenada got us well within the northern part of the great cirio desert, northwest of Mision San Fernando. Ex- ploring all of the rock shelters we could spot with 10 X binoculars en route, we finally found an old wood rat {Neotoma) midden in a cavelet located in volcanic rocks. The midden was about one meter thick and well within a secure shelter, large and dark enough to harbor a few bats. I knew it was old because it contained none of the desert plants mentioned above, but rather was composed of woodland or chaparral trees and shrubs, predomi- nantly Juniperus californica, accompanied by leaves and acorns of a shrubby live-oak {Quercus turbinella) and tiny but numerous leaves of cha- mise {Adenostoma fasciculatum), a principal dom- inant of the California chaparral (Table 1). Also present were a few nutlets of Arctostaphylos glan- dulosa Eastw. More remarkable were the large leaves and endocarps (cherry pits) of Prunus lyonii Eastw., the Catalina cherry. A thorough reconnaisance of the area showed a desert vegetation characterized by dominance of all of the giant columnar xerophytes and a profusion of lesser cacti and Agavaceae, notably Agave shawii En- gelm., A. deserti Engelm., and Yucca whipplei Torrey. There were no chaparral shrubs except for one indi- vidual of the xerophytic monotype, Xylococcus. A fly in the ointment was the presence of a few, live Ju- 2000] WELLS: PLEISTOCENE ENCINAL IN CIRIO DESERT 191 niperus califomica Carriere about 200 m from the \ midden site at an elevation of 610 m (2000 ft). This somewhat vitiated the significance of the juniper re- cord, but the chaparral sclerophylls justified the three radiocarbon dates kindly provided by Rainer Berger: 10,000 to 10,200 radiocarbon years before present from top to bottom of the midden. This was cetainly not a pleniglacial date, but rather late-glacial to Ho- locene transition (Table 1). Proceeding slowly south for another few days, we entered the extensive granitic area north of San- ta Ines, where we explored many Neotoma middens in cavities within the small exfoliation domes of the granitoid rocks. Unfortunately, all we saw were re- cords of desert plants. This set a pattern of more desert records all the way to La Paz. Needless to say, I was discouraged with the prospects, but later switched to mainland Mexico, where I have more unpublished records. Still later, I decided to con- centrate on the Great Basin (Wells 1983). Shortly after the 1983 paper appeared, William H. Clark of Albertson College, Caldwell, Idaho, kindly sent me samples of an obviously ancient wood rat midden from the Santa Ines granitic area. It contained Pinus quadrifolia, a four-needled pinyon pine species previously unrecorded in the Neotoma fossil record anywhere, so I had the stratum dated: 17,470 ± 200 radiocarbon years before present (Beta 9322). In 1988, 1 responded to Bill's generous invitation to visit the site to collect more material. Instead of a week's drive below Tijuana, I made it in the afternoon of the same day, blacktop all the way! Additional material from the dated sector of the midden provided much more Pinus quadrifolia and Juniperus califomica and a similar assemblage of chaparral sclerophylls previ- ously recorded in the 10,000 yr old midden, plus Er- iodictyon angustifolium Nutt. Consistently absent from both the 10,000 and 17,400 y-old deposits were any xerophytes of the existing desert vegetation (Ta- ble 1). The combination of chaparral shrubs, junipers, and pinyons was called encinal by Forrest Shreve, using a Spanish word for an oak community. The west slope of the Sierra San Pedro Martir at middle elevations has chaparral, partly dominated by Quer- cus turbinella and Adenostoma, associated with Pi- nus quadrifolia and Juniperus califomica but mi- nus any desert xerophytes, aside from the chaparral Yucca, Y. whipplei. Thus, the encinal recorded in late-glacial and pleniglacial middens in what is now cirio desert has a nearly exact analog in the San Pedro Martir mountains to the north. Discussion The contrast between the spectacular modern de- sert vegetation dominated by giant xerophytes like cirios and cardons at both Neotoma sites and the midden macrofossil evidence for a Pleistocene en- cinal, lacking desert xerophytes, staggers the imag- ination (Table 1). The recorded displacement was complete at this latitude, ca. 30°N. Pinus quadrifolia, the four-needled pinyon, is one of the more mesophytic pinyon pines. At pre- sent it is restricted to moderately high elevations (ca. 1100 to 2100 m, or 3500 to 7000 ft), mainly on the western or Pacific slopes of the Peninsular Range as far south as the San Pedro Martir. In con- trast, the one-needled P. californiarum D.K. Bailey (1987) occurs on the rain-shadowed eastern slope of the Peninsular Range (Bailey, personal commu- nication 1975-1990; Wells 1995). Where the two distinct species occur on the same mountain, as on Mt. San Jacinto, California, P. quadrifolia forms a zone well above P. californiarum. The latter, one- needled, pinyon alone has a far southern disjunction on Cerro San Luis (to 1550 m) in the Sierra de Calamajue at about 29°N, an isolated peak not far north of the high San Borja Mountains (to 1700 m + ), where no pinyon pines have ever been recorded (Fig. 1). The nominate subspecies of Pinus califor- niarum also extends disjunctly far to the north of P. quadrifolia on low, isolated mountains in arid parts of the Mohave Desert, e.g., the Coxcombs, Eagle, Old Woman, and northeast in the Provi- dence; it dominates the pinyon-juniper zone in all of these ranges (Wells 1995). From the distribution maps of Critchfield and Little (1966), one would have hypothesized P. ca- liforniarum (then under the Great Basin species P. monophylla Torrey & Fremont) to have extended farther south on the peninsula than P. quadrifolia during the Ice Ages, because they show the south- erly outlier of the former on Cerro San Luis. This may well have been true on the east or Gulf of California slope, where we have no Neotoma re- cords. The macrofossil record of P. quadrifolia we do have is from the Pacific side of the peninsular divide in the cirio zone at the substantial elevation of 550 m (1800 ft), less than 500 m below its mod- ern lower Hmit (Wells 1986). The assemblages of evergreen sclerophylls {Ade- nostoma fasciculatum, or chamise, a principal dom- inant of California chaparral, Prunus lyonii, Arc- tostaphylos glandulosa or Eastwood manzanita, etc.) are consonant with the presence of Juniperus califomica or Pinus quadrifolia in the same strata of the Neotoma middens. Today, the juniper and four-needled pinyon both associate with a broad zone of chaparral below montane forest of Pinus jejfreyi Grev. & Balf. that is mainly above 2500 m; the chaparral belt (largely dominated above by Arc- tostaphylos peninsularis Wells) extends down to <1000 m. Chaparral extends southward from California in the Peninsular Range to Cerro Matomi (to 1 370 m), the southern extremity of the San Pedro Martir. Aside from small populations of manzanitas (A. pe- ninsularis) on isolated peaks: Cerro San Juan de Dios (to 1300 m at 30°N) and Cerro San Luis (1550 m: the southern limit of Pinus californiarum at 29°19'N) there is a major disjunction of chaparral species to the high San Borja Mountains (to 1700 Fig. 1. Map of Baja California, showing principal mountains mentioned in text and numbered wood rat (Neotoma) midden sites of Pleistocene age in white circles: 1 = northwest of Mision San Fernando, 30°N, at 594 m (1950 ft). 2 = northwest of Rancho Santa Ines, 29°46'N, at 550 m (1800 ft). m) rising above the Gulf of California at 28°47'N (Fig. 1). The San Borjas are the southern limit of Juniperus californica and Arctostaphylos peninsu- laris, but lack pinyon pines (Reid Moran, personal communication 1969; Wells 1972). Still farther south at 27°30'N there is a southern outpost of chaparral on the lofty Tres Virgenes volcanoes (to 2000 m) above Santa Rosalia on the Gulf; the Tres Virgenes are also a southern outpost for cirio trees. Possibly the most unexpected macrofossil occur- 2000] WELLS: PLEISTOCENE ENCINAL IN CIRIO DESERT 193 rence in the northern cirio desert at both Neotoma sites (Table 1) is the Catalina cherry tree {Prunus lyonii). This large-leaved cherry occurs in Califor- nia on the larger Channel Islands, but not on the mainland. In Baja California, however, this ever- green cherry has widely disjunct stations in deep canyons of the isolated and inaccessible Sierra de San Francisco (or Francisquito) of northern Baja California del Sur at ca. 27°30'N (Fig. 1). The two Pleistocene records of Prunus lyonii in the northern part of the cirio zone suggest (but do not prove) a former continuity of range. A major biogeographic anomaly that might shed light on the Pleistocene location of the Sonoran De- sert with its rich array of endemic plants (about 30% of the 2500+ species are endemic: cf. Wells 1970), is posed by the distribution of pinyon pines (cf. Wells 1986). As discussed above, Pinus qua- drifolia is presently restricted to the Sierra San Pe- dro Martir south to about 30°20'N, and the Neoto- ma record at Santa Ines extends that to 29°46'N, a scant 80 km farther south. The one-needled Pinus californiarum has an isolated southern outpost on Cerro San Luis (to 1550 m) at 29°19'N. Neither of these pinyon pines is known from the higher Sierra San Borja (to 1700 m +), where Juniperus califor- nica reaches its southern limit. None of these three conifers occurs in Baja California del Sur, the southern half of the peninsula, which has suitably high mountains such as Tres Virgenes (to 2000 m). Sierra de Santa Lucia (to 2000 m). Sierra de la Gi- ganta (to 1770 m), and Sierra de la Victoria (to 2070 m) = Laguna Mountains (Fig. 1). Instead, the Laguna Mountains support an extensive zone of the three-needled Mexican pinyon, Pinus cembroides Zucc. (subsp. Lagunae Bailey), which has very dis- tinctive pink "endosperm" (gametophytic tissue). In all other species of pinyon pines the food re- serves of the seed are white. Pinus cembroides has a very wide distribution on the mainland of Mexi- co. There are also some live-oaks in the Lagunas identical to mainland species (Que reus reticulata H. & B., Q. tuberculata Liebmann). Other broad sclerophylls in the isolated Lagunas include the toyon, Heteromeles arbutifolia (Lind- ley) Roemer, Arbutus peninsularis Rose & Gold- man, and Garrya salicifolia Eastw. Most remark- ably, no taxa of Arctostaphylos, Ceanothus, or Ade- nostoma are known from any mountains of Baja California del Sur (S. de la Giganta, S. de Laguna). Absence of these three most characteristic genera of the California chaparral, including the two most speciose genera, Arctostaphylos and Ceanothus, is strong evidence of a major isolating barrier in the central sector of the peninsula. The southernmost known occurrence of Arctostaphylos peninsularis is in the high Sierra de San Borja at 28°45'N (Wells 2000). The isolated Sierra de la Laguna pinyon-oak woodland is mainly above tropical deciduous for- est, which occupies the lower slopes and surround- ing foothills of the Lagunas that are on the Tropic of Cancer (23 y2°N). The absence of Juniperus ca- lifornica or either of the two northern pinyons may mean that these conifers never migrated this far south. Almost certainly, had Juniperus californica colonized any of the high mountains of Baja Cali- fornia del Sur, possibly even the summer-rainy La- gunas, it may have survived to the present, inas- much as its niche is vacant, there being no other species of Juniperus in Baja California. Although the Neotoma macrofossil evidence from the northern cirio zone documents a modest displacement of desert vegetation by a mesophytic pinyon pine and evergreen chaparral, there is as yet no evidence as to how far south this Pleistocene climatic effect extended. Even if all the elevated areas north of the Vizcaino plain in the central sec- tor of Baja California were affected, there would be ample room for desert vegetation farther south (Fig. 1). The peninsula is immensely long, extend- ing far into subtropical latitudes. The apparent fail- ure of even the relatively xerophytic Juniperus ca- lifornica to colonize any of the mountains of the southern half of Baja California (it stopped far short in the San Borjas at 28°45'N) suggests a major de- sert barrier in the central sector of the peninsula, where temperature-sensitive giant cacti, and quasi- endemics like Idria (cirio), Pachycormus, Viscai- noa and others may have survived the long Pleis- tocene periods of climatic displacement unscathed. Acknowledgments Research supported by grants from the National Science Foundation: GB-5002, DEB 78-1 11 87. Special thanks are extended to William H. Clark and his long-term research team at Santa Ines, chiefly from Caldwell, Idaho, who generously shared their discovery of the ancient pinyon midden near their annual camp at Santa Ines. Literature Cited Bailey, D. K. 1975. Personal communication 1975-1990. . 1987. A study of Pinus, subsection Cembroides: the single-needed pinyons of the Californias and the Great Basin. Notes Royal Botanic Garden, Edinburg 44:275-310. Critchfield, W. B. and E. L. Little. 1966. Geographic distribution of the pines of the world. U.S.D.A. Forest Service Misc. Publ. 991. Moran, R. 1969. In letter. Shreve, E 1 95 1 . Vegetation of the Sonoran Desert. Carnegie Institute of Washington Publication 591, Vol. 1. Wells, P. V. 1966. Late Pleistocene vegetation and degree of pluvial climatic change in the Chihuahuan Desert. Science 153:970-975. . 1970. Historical factors controlling vegetation patterns and floristic distributions in the Central Plains region of North America. Department of Ge- ology, University of Kansas, Special Publ. 3. . 1972. The manzanitas of Baja California, includ- ing a new species of Arctostaphvlos. Madrono 21: 268-273. . 1976. Macrofossil analysis of wood rat (Neoto- ma) middens as a key to the Quaternary vegetational 94 MADRONO [Vol. 47 history of arid America. Quaternary Research 6:223- 248. . 1983. Paleobiogeography of montane islands in the Great Basin since the last glaciopluvial. Ecol. Mo- nogr. 53:341-382. . 1986. Systematics and distribution of pinyons in the late Quaternary, Pp. 104-108. in R. L. Everett compiler. Proceedings: Pinyon-Juniper Conference. . 1995. Recognizing the new single-leaf pinyon pine {Pinus californiarum Bailey) of southern Cali- fornia. The Four Seasons 10:53-58. . 2000. The Manzanitas of California, Mexico and the World. Published by the author; 151 p., 150 figs. , AND R. Berger. 1967. Late Pleistocene history of coniferous woodland in the Mohave Desert. Sci- ence 155:1640-1647. Madrono, Vol. 47, No. 3, pp. 195-203, 2000 SEED BANKS OF LONG-UNBURNED STANDS OF MARITIME CHAPARRAL: COMPOSITION, GERMINATION BEHAVIOR, AND SURVIVAL WITH FIRE Dennis C. Odion' Department of Geography, University of California, Santa Barbara, CA 93106 Abstract Seed germination requirements in the California chaparral have been described mainly from freshly collected seed. However, uncertainties remain because the behavior of seeds in the soil can differ. I studied germination of the seed bank in long-unburned stands of maritime chaparral in central coastal California. I quantified seedlings emerging from soil samples provided with appropriate temperature and moisture conditions following I) no other treatment, 2) a heat treatment to optimize germination of heat-stimulated species, 3) the same heat with the addition of charred wood, and 4) the burning of chaparral stands prior to collection of samples. I compared germination in these treatments with seedling emergence in the field following fire. I also collected and divided samples into 0-2.5 and 2.5-7.5 cm depth fractions to evaluate abundance of seed at the surface and depth before and after fire. Seed of one annual had reduced germination following the heat treatment. Seeds of all other species common enough to evaluate statistically were heat tolerant. However, because seeds were found to be mostly near the surface, there was considerable mortality with fire. Moreover, seedling populations in the field only accounted for a fraction of the seed bank that survived fire, and seventeen species that ger- minated in samples did not germinate and/or emerge in the field. Most species' germination and emergence was influenced in some way by heat and/or charate. Germination of two Ceanothus was dependent on heat. Adeuostoma fascicidatum Hook. & Arn., Arctostaphylos piirissima P. Wells, and two annuals had germination that was enhanced by heat and enhanced further when charate was added. Despite the im- portance of fire effects, there were no short-lived species having entirely fire-dependent germination. Germination and/or emergence of 3 species was negatively affected by charate. These germinated spar- ingly or not at all after fire. One of the most prominent evolutionary special- izations to fire exists in the germination ecology of seeds from plants found in Mediterranean shrub- lands, particularly those of Australia, South Africa, and California (Bond and Van Wilgen 1996). This subject has received considerable attention (Review by Keeley 1991), revealing a fascinating complex- ity of features that insure germination will coincide with the anomalously favorable conditions for seed- ling establishment that exist after fire. There are physical features such as bradyspory (or serotiny) where seeds stored in fruits and cones are released when heated by fire (Whelan 1995), and impervi- ous seed coats that open with the heat of fire (Swee- ney 1956; Quick and Quick 1961; Auld and O'Connell 1991). Physiologically dormant seed may be induced to grow following fire by chemi- cals washed from charred wood (Wicklow 1977; Keeley 1984, 1987; Keeley et al. 1985; Keeley and Pizzorno 1986), water soluble nitrogenous com- pounds (Thanos and Rundel 1996) and smoke (Keith 1997; Keeley and Fotheringham 1997, 1998). For each fire-related germination cue, there are multiple dormancy-releasing mechanisms that have evolved convergently among disparate floras (Baskin and Baskin 1998). In chaparral, germina- ' Present address: Marine Science Institute, University of California, Santa Barbara, CA 93106. tion without fire may also be inhibited by allelo- pathic chemicals washed from foliage or litter (Muller et al. 1968; McPherson and Muller 1969), and/or phytotoxins produced by soil microbes (Ka- minsky 1981). Fire eliminates these compounds. Seeds of chaparral plants range from readily ger- minable at the time of dispersal (non-refractory) to deeply dormant (refractory) as a result of multiple barriers to germination (Keeley and Fotheringham 1998). Some species produce a portion of seed that is refractory and a portion that is not (Emery 1988; Parker and Kelly 1989). Generalizations about the type(s) of dormancy species exhibit derive mainly from tests on freshly collected and stored seed. Germination of seeds residing in the soil may differ significantly, as has been documented for Adeuos- toma fascicidatum Hook. & Arn. dind Arctostaphy- los canescens Eastw. among others (Stone and Juhren 1951; Parker 1987; Keeley and Fotheringh- am 1998). Seeds exposed to allelopathic chemicals and phytotoxins found in chaparral soils may ex- hibit enforced dormancy (Muller et al. 1968; Mc- Pherson and Muller 1969; Keeley 1991). Therefore, it is imperative to study the soil seed bank to un- derstand how chaparral germination is controlled in nature. The potential for germination in the chaparral seed bank without fire is thought to be low for most species because seedlings are rarely apparent under 196 MADRONO [Vol. 47 the shrub canopy. However, Christensen and Muller (1975), Tyler (1995), and Swank and Oechel (1991) reported considerable seedling growth under Ad- enostoma in plots protected from herbivory, sug- gesting more germination can occur without fire than is evident. In addition, Zammit and Zedler (1988, 1994) and Holl et al. (2000) found that many species germinated readily from chaparral soil seed bank samples. To test how much germination can occur in species' seed banks without fire and how much requires heat and/or chemicals produced by fire, I compared emergence from controls and uni- form fire treatments that are known to maximize germination of refractory seed without inducing mortality. I then analyzed how much of the in situ seed bank was eliminated above and below 2.5 cm in the soil by fire in the chaparral stands. Finally, I enumerated seedling emergence in the field to com- pare germination in nature vs. in collected samples. The chaparral I studied is geographically-isolated and its environment differs in many respects from that found in the Transverse and Peninsular ranges inland. With Santa Ana winds absent, and a lower frequency of ignitions, coastal environments have likely supported less frequent and dynamic fire, at least prior to human dominance of the fire regime (Odion et al. 1992; Odion et al. 1993). The average lifespans of the Ceanothus spp. in maritime chap- arral are particularly short (Davis et al. 1988) com- pared to those inland (Keeley 1975, 1992). Death of the non-sprouters opens space for recruitment by numerous herbs and subshrubs (Odion and Davis 2000). These and other factors such as soil and cli- mate may help explain differences in post-fire re- generation in maritime vs. nearby inland chaparral reported by Tyler (1995); they may also have con- tributed to evolutionary divergence in maritime chaparral taxa that has produced endemic Arctos- taphylos and Ceanothus (Griffin 1978). I have eval- uated my germination data for any evidence that the environment and insular nature of the study area manifested variation in germination ecology. Study Area Samples were collected from Adenostoma fascic- ulatum (hereafter Adenostoma) chaparral located near sea-level, within Vandenberg Air Force Base in central, coastal California as described in D' Antonio et al. (1993), and Odion and Davis (2000). Substratum here is Pleistocene eolian sand (Dibblee 1950). Climate in the area is strongly in- fluenced by the prevailing onshore winds and cool ocean, and the temperature regime is mild, es- pecially for a chaparral environment. Maritime chaparral of the area has been described in detail by Davis et al. (1988). The average annual precip- itation is 35.3 cm. I counted annual rings from the obligate seeders, Arctostaphylos purissima P. Wells, and Ceanothus cuneatus (Hook.) Nutt. var. fasci- cularis (McMinn) Hoover (Keeley 1993) to estab- lish that the chaparral had not burned for 75-80 y at the site where more intensive sampling was un- dertaken (site 1). Samples were also collected from a second, nearby site which had not previously burned for about 50 y. Both sites were dominated by Adenostoma. Methods Transects consisting of 47 contiguous 1 m^ plots were established in dense chaparral dominated by Adenostoma. Nine 5 cm diam, 7.5 cm deep cores of soil were obtained per plot at site 1 in the fall of 1988. Five cores were collected in fall, 1989 at site 2. Chaparral at both sites was burned soon thereafter, with low fuel moisture contributing to relatively intense fires (Odion and Davis 2000). I collected 5 cores per plot the day after each fire. Seed bank cores for each plot were composited, and 350 cc subsamples were removed from each ho- mogenized sample. Pre-burn samples from site 1 were given three treatments: 1) heat, 2) heat and charate (charred, pencil-sized Adenostoma stems collected after the fire and ground up) and 3) control (no heat or char- ate). Only the second of these treatments was used on pre-burn samples from site 2. Samples to be heated were spread to a depth of 2-3 mm on alu- minum cooking trays. Based on studies by Wright (1931), Sampson (1944), Sweeney (1956), Keeley and co-workers (several publications, see Keeley (1991), heating at 100°C for ~5 min typically pro- duces the greatest germination response among fire- recruiters, and is well within their heat tolerance. Given the slight insulation the soil would provide, I decided to use a 7 min duration. Heat-induced seed mortality is controlled predominantly by max- imum temperature, as opposed to duration (Borch- ert and Odion 1995), so it is unlikely that this change effected mortality. Heating was done in a forced-air oven. The subsamples were spread on sterile sand in 20 cm plastic pots. The amount of charate added was 2 rounded tablespoons (21.3 ± 0.94 g, n = 11). The pots were covered with clear plastic, pro- tected from herbivory, and kept moist out-of-doors under 50 percent shade cloth at Cal-Orchid Nursery in Santa Barbara, where temperature fluctuations were analogous to the field. Potting was complete in late November, at which time all samples were given their first watering. All the samples were ex- posed to outdoor temperatures from the time of col- lection through the subsequent growing season to provide natural temperature stratification. Seed bank sampling was also undertaken at ran- domly located --1.5 m diam canopy gap areas ad- jacent to the site 1 transect. I took samples from the center of the gap as well as the edge and un- derstory of the adjacent Adenostoma canopy. These cores were separated into 0-2.5 cm and 2.5-7.5 cm 2000] ODION: GERMINATION OF MARITIME CHAPARRAL SEED BANK 197 depth fractions and given the heat and charate treat- ment. Due to smaller amounts available, 175 cc subsamples were spread over sand in 16 cm diam plastic pots placed with the others. All germinants were identified and removed from pots through the growing season. Nomenclature I follows Hickman (1993). Specimens whose identity ' was uncertain were grown until it was determined. I Five of the pre-burn samples treated to heat and charate were removed from pots after emergence stopped. I repotted these the following autumn. No further germination occurred in these. Results General patterns. Seed from 72 species germi- nated and emerged from samples collected along transects (Table 1). More than half (48) were an- nuals. Site 2 had greater diversity (60 vs. 48 spe- cies). Many of the same species emerged abun- dantly from samples from both sites (e.g., Aden- ostoma, Helianthemum scopariiim Nutt., Crassula connata (Ruiz Lopez & Pavon) A. Berger, Cen- taurium davyi (Jepson) Abrams, and Navarretia atractyloides (Benth) Hook & Arn.). Despite this, there were only 33 species in common. In addition, two subshrubs, Miniidus aiirantiacus Curtis and Lo- tus scoparius (Nutt.) Ottley, were abundant in site 2 samples and absent in those from site 1. Several annuals were common in samples from one site but not the other. Nine species were non-native. All 9 are widespread weeds. Post-burn samples contained 44 species ( 1 non- native) and substantially reduced numbers of ger- minants. The reduction varied with species depend- ing on the proportion of seed present at depth, as described below. Reduced germination was also strongly correlated with the amount of soil heating that occurred where samples were located (Odion and Davis 2000). Thus, the number of post-burn germinants was much greater in samples from gaps vs. under the shrub canopy. Horizontal patterns of seed abundance are analyzed in Odion and Davis (2000). There were 20 and 3 1 species that germinated in the field respectively at the two sites (Table 1 ). Twelve were at both. Four species germinated in the field plots but not seed bank samples. There was only one individual of each. Seventeen species ger- minated in samples but did not germinate and/or emerge in the field, including species whose seed was among the most abundant (e.g., Centaiiriiim davyi, Mimulus floribiindiis Lindley). Another, Crassula connata, was virtually absent in the field in the burn areas, although it was common in ad- jacent unburned chaparral. Field populations for most species were much smaller than post-burn seed bank populations — between ~5 and 14 times smaller for shrubs, and generally even smaller for other species. Germination treatments. Only two species, both perennial Gnaphalinms, had germination that was not affected by the heat treatment (Table 1). One of these, G. microcephalum Nutt. was significantly negatively affected by charate. Among heat-affect- ed species, the two obligate-seeding species of Ce- anothus, the subshrub Helianthemum scoparium, and the annual Thfolium microcephalum Pursh had significantly greater germination with heat alone, while the opposite occurred for the annual Calan- drinia ciliata (Ruiz Lopez & Pavon) DC. (Table 1). Other important species that had a positive response to heat were also affected by charate. Adenostoma, Arctostaphylos purissima, and Lotus strigosus had significantly greater germination with heat and charate compared to heat alone. Germination of Centaurium davyii and Crassula connata with heat and charate was not only signif- icantly lower than with heat alone, but also lower than with no treatment. Crassula was rare in the burn areas, and Centaurium did not occur there un- til the third year after fire. Both species were fairly common in the surrounding unburned chaparral. Mimulus floribundus had much lower germination with heat and charate than with heat alone, but heat and charate germinants outnumbered those in con- trol samples ( P > 0.05, NS). This species, though abundant in several samples, was absent from most. It was never observed in the field, including in un- burned chaparral. It is typically found in seasonal wetlands like two other species that were found in samples, but not in the field, Crassula aquatica (L.) Schonl. and Centunculus minimus L. With a relatively high proportion of seed at depth (76 percent below 2.5 cm in gap, edge, and under- story samples combined. Table 2), Arctostaphylos purissima had better survival (post-burn/pre-burn = 17 percent) than Adenostoma (site 1 = 2 percent, site 3 = 3 percent) which only had 22 percent of its seed below 2.5 cm. These survival percentages are from transect samples. The 2.5-7.5 cm depth samples had relatively little emergence of species with charate-enhanced germination. Seeds of other shrubs were not abundant enough to evaluate depth distribution. The high survival of Ceanothus cu- neatus at site 2 (33 percent) as well as results from a fuel translocation experiment (Odion and Davis 2000) suggest this obligate-seeder had a high pro- portion of seed at depth. Helianthemum scoparium was particularly abun- dant and not affected by charate, so the effect of the depth distribution of its seed is relatively clear. The subshrub had 53 and 75 percent of its seed bank below 2.5 cm in gap and understory samples respectively. Despite the greater proportion of seed at depth at understory plots, survival was similar there (22 percent) compared to gaps (24 percent). Survival along the site 1 transect was 9 percent (post-burn/pre-burn heat). After fire, 94 percent of Helianthemum scoparium seeds were below 2.5 cm in the soil in understory plots, only half were at depth in gap plots. 198 MADRONO [Vol. 47 tu c C/5 C3 1^ m V „1 ^ - S 5 O C3 < '^H ^ ^ s s < 2 3 P3 cd 'o cd i-L- a - zee c/) ocN r~- ^r-H in On "n (N (N M3(N(N '^^^ °9 ^00 oi^oo-H'Kr^O'^r-^o^''^oooo6o^rjo ^ (N in in in ^ on (N r-H in m dooooooodoooooodoo So i: cx, >- li 8 g I § 1-2 l>5 2 s CO t~ S ^ to Co ^ 5 ^ cs^ ^ --^ ^ cio s- ^ ^ So Q :::: 200 MADRONO [Vol. 47 Table 2. Numbers of Germinants, Expressed as Average Number per m% From 0-2.5 cm and 2.5-7.5 cm Depth Fractions Before and After Fire, and in the Field From the Same Plots in Which the Separate Depth Samples Were Taken. Values are the averages from 30 samples expressed as the density of seed per m-. Pre-burn samples were treated with heat and charate. pre-burn post-burn field 0-2.5 2.5-7.5 0-2.5 2.5-7.5 Shrubs Adenostoma fasciciilatiun Z4U. 1 jZ.Vz Q O O.O 5.4 A rctostaphylos purissima 44.1 141.1 36.8 26.5 10.6 Subshrubs Helianthemiim scoparium 41 1.6 582.6 88.2 145.6 47.0 Perennial herbs Mimulus florihundus 1519.0 538.0 39.2 26.5 , 0 Annual herbs Apiastrum angustifoliiim 245.0 26.5 0 17.6 0.8 Centaurium davyi 1396.5 608.6 558.6 185.2 0 Crassula connata 6056.4 1525.9 1 166.2 493.9 ' 1.6 Cryptantha clevelandii 78.4 17.6 0 0 0.1 Lotus strigosus 83.3 88.2 19.6 79.4 11.0 Na va rettia atractyloides 3013.5 299.9 274.4 61.7 29.7 Survival percentages of the seed bank for species whose germination was negatively affected by charate are equivocal because post-burn samples presumably contained the inhibitors(s). Among the remaining annuals, high mortality was common. In fact, Apiastrum angustifoliwn Nutt. though fairly common in pre-burn samples, was not detected in post-burn transect samples, and was rare in the field. Seed of this diminutive plant was predomi- nantly near the surface (Table 2). The second most abundant species in the field after fire was the an- nual Navarretia atractyloides. Combining data from gap and understory plots in Table 2, while better illustrating survival at depth, obscured other patterns. Where seed of this species was concen- trated, in canopy gap areas, only 6 percent of its seed was below 2.5 cm in the soil, explaining why only 10 percent survived there despite relatively low soil heating with fire. Conversely, in the un- derstory, 16 percent of seed was in the deeper frac- tion, explaining the relatively high survival (22 per- cent) along the site 1 transect (predominantly un- derstory). Seed of Lotus strigosus (Nutt.) E. Greene was equally abundant in deep and shallow samples overall (Table 2). Survival in surface samples was a relatively high 24 percent, and at depth 90 per- cent. Discussion My procedures indicated that seed mortality with fire was substantial, and greater in the older stand. Previous studies have also found that a significant number of seeds do not survive fire in chaparral (Keeley 1977; Davey 1982; Bullock 1982; Zammit and Zedler 1988; Davis et al. 1989). For species to ensure successful recruitment after fire, their seeds must accumulate at depths in the soil where they will be safe. There must be strong resistance to ger- mination in the absence of fire for this to occur. Consistent with this, I found that most of the seed bank for many species needed fire to germinate (Ta- ble 1 ). Two shrubs, both species of Ceanothus, had germination that was entirely fire-dependent. How- ever, I also found that a small but distinct portion of the seed bank for all other fire-recruiters ger- minated with simply moisture and natural temper- ature fluctuations. In addition, there were no short- lived species detected that had entirely fire-depen- dent germination {i.e., there were no specialized fire annuals), which is not typical for chaparral. Seed- lings of these are usually found only the first year after fire (Keeley et al. 1981). Short-lived species in my study all produced seedlings after the first post-fire growing season (Odion 1995, unpublished data). Thus, fire-recruiters in this study, other than Ceanothus, produce seed that is both refractory and not. Based on my germination results and those by Davis et al. (1989) and Holl et al. (2000), as well as extensive field observation (Davis et al. 1988; Odion et al. 1992; Odion et al. 1993) non-refractory seed may be somewhat more important in maritime vs. inland chaparral, at least among short-lived spe- cies. Conversely, for Adenostoma, the proportion of non-refractory seed (19%, Table 1) is in agreement with what has been found at more inland locations (Stone and Juhren 1953; Zammit and Zedler 1988). How might having seed that is both refractory and not be a selective advantage in chaparral? By producing seed that germinates readily in the field, short-lived species may grow and reproduce throughout the fire cycle, which may be critical for them to sustain seed populations from one fire to the next (Zammit and Zedler 1988). As fire interval increases, the capability to augment the seed bank between fires will be of increasing importance to short-lived species because their seed banks will i 2000] ODION: GERMINATION OF MA] I otherwise diminish due to mortality and predation. Therefore, considering the past likelihood of rela- tively long fire intervals in maritime chaparral, it is not surprising that I found non-refractory seed to be so prevalent, even among fire-recruiters. It is possible that some non-refractory seed remains dor- mant under the chaparral canopy due to inhibitors. However, short-lived species such as Navarettia atractyloides, Helianthemum scoparium, Chorizan- the spp. and Camissonia micrantha (Sprengel) Ra- ven are common in old age class maritime chap- arral (Davis et al. 1988; Holl et al. 2000), particu- larly in the canopy gaps that typify this vegetation. Their seeds are concentrated in gaps, an advantage because survival with fire is much greater there (Odion and Davis 2000). The germination ecology of these and other short-lived species allows them to exploit gaps when they appear in maritime chap- arral, resulting in more abundant post-fire recruit- ment than would otherwise occur. Such opportunis- tic germination has been documented in other chap- arral (Zammit and Zedler 1988), and linked to can- opy gaps (Shmida and Whittaker 1981), but its relative importance undoubtedly varies with chap- arral canopy dynamics. The dominant shrubs in the study, Adenostoma, Arctostaphylos, and the two species of Ceanothus, are fire-recruiters. Nearly all chaparral areas are dominated by some combination of these genera and there has been much interest in their germina- tion ecology. For Adenostoma, the most widespread and abundant chaparral shrub, a question that has persisted had been, what is the role of heat in ger- mination of refractory seed? Adenostoma seed banks have been studied previously by Christensen and MuUer (1975) who heated soil under shrubs in situ Zammit and Zedler (1988, 1994) who burned straw over soil placed in flats, and Parker (1987) who oven-heated soil and supplied charred wood extract. The first two procedures enhanced germi- nation of Adenostoma, but it is unclear whether this was a direct or indirect effect of heat. Parker (1987) found that charred wood extract, not heat, enhanced germination. It is possible that the heat he supplied (100°C for 1 h) was in excess of what the seeds in the samples could tolerate since it resulted in a de- crease in germination. On seeds collected from shrubs, oven-heating stimulated germination (Wright 1931; Sampson 1944; Stone and Juhren 1953). In addition, Keeley (1987) found that heat alone increased germination compared to controls in 6 of 6 different temperature treatments, but there was not a statistically significant effect. However, Keeley did find significantly enhanced Adenostoma germination with charate, and that there was a syn- ergistic effect with heat and charate. I also found that heat and charate produced a synergistic effect, but unlike Keeley, that germination was signifi- cantly enhanced with heat alone. It is possible that stimulatory substances were formed and/or inhibi- tors destroyed when I heated soils. Chemical stim- HME CHAPARRAL SEED BANK 201 ulants can be produced when soil or wood are heat- ed to 175°C for 10-30 min (Keeley and Nitzberg 1984; Keeley and Pizzorno 1986). However, these stimulants are effective in very low concentrations (Keeley and Pizzorno 1986) and my heat treatment did not produce a germination effect comparable to that found with heat and charate. In contrast to Adenostoma, the Arctostaphylos and Ceanothus'' in this study are narrowly distrib- uted taxa whose germination ecology has not been previously studied. However, congeneric ecological analogs can be compared. I found that heat was effective in inducing Arctostaphylos purissima seeds to germinate, and again that there was a syn- ergistic effect with both heat and charate. This ef- fect has been found in one oihQV Arctostaphylos that coincidentally is also a narrow endemic from mar- itime chaparral, A. morroensis Wiesl. & B. Schrei- ber (Tyler et al. 1998; Tyler et al. 2000). Their methods avoided potential influence of soil-derived stimulants because seeds were extracted from the soil prior to heating. Germination doubled with heat and charate, but, there was no effect with heat alone, or charate alone. Conversely, Parker (1987) found that dormancy of A. canescens seeds extract- ed from the soil was overcome by charred wood extract alone. Freshly collected seed remained dor- mant with the same treatment. Other research using freshly collected seed has found that just heat (Sampson 1944; Berg 1974) or charate (Keeley 1987) can be effective in breaking seed dormancy of Arctostaphylos. Further research on Arctostaphy- los spp. seed banks is needed to determine how variation in their germination may be correlated with fire regime or other environmental variables. Ceanothus spp. have a hard seed coat that can be cracked by heat (Quick 1935; Quick and Quick 1961). There can be some germination in the ab- sence of fire if the impermeability of the seed coat deteriorates over time (Quick and Quick 1961), e.g., C. greggi A. Gray, (Moreno and Oechel 1991; Zammit and Zedler 1994). In addition, Keeley (1991) reports that it is typical for a few percent of the seeds of Ceanothus to be non-refractory. In the stands I studied, seeds of C. cuneatus van fascicu- laris and C. impressus had resided in the soil for a considerable length of time. Neither species was in the pre-burn vegetation in the older stand; both drop out as stands of maritime chaparral age-after only —20 y in the case of C. impressus (Davis et al. 1988). Nonetheless, I found no germination of this species or C. cuneatus var. fascicularis without heat. Because input into the seed bank for these species will cease in long-unburned stands, seeds must survive and remain dormant for their seed banks to persist. These two species may have seed coats that are especially resistant to deterioration, perhaps because they are relatively thick. Thickness of seed coats is correlated with heat endurance (Wright 1931), and I found these Ceanothus spp. were more capable of germinating in areas of great- 202 MADRONO [Vol. 47 er soil heating than any other species at the two burn sites (Odion and Davis 2000). Another hard-seeded species with heat-stimulat- ed germination was the annual Lotus strigosus (Ta- ble 1). It is curious that this species, unlike the Ce- anothus spp., had germination further enhanced by heat and charate. Lotus strigosus was common the second spring after fire. Second year plants may have emerged from seed that did not germinate the first year, or from seed produced the first year. Charate-induced germination could allow newly produced seed to germinate the following year without heat if water soluble byproducts of wood combustion are still present in the burn area. An- other hard-seeded annual legume Trifolium micro- cephalum did not have charate-enhanced germina- tion, and the phenomenon has not been reported among other hard-seeded species (Baskin and Bas- kin 1998; Table 10.6). Baskin and Baskin (1998; Tables 10.4 and 10.7) list a few chaparral species that may have germi- nation reduced by heat or charred wood extracts. I found germination that was suppressed by heat (one species), and by heat and charate (three species; Table 1). Curiously, two of the species suppressed by heat and charate (Crassula connata and Cen- taurium davyi) had germination that was enhanced by heat alone. Both were uncommon in the field after fire despite having abundant seed in the soil. In fact, it was not until three to four years after each burn that Centaurium seedlings appeared, so the same mechanism that inhibited germination in samples apparently operated in the field. The results for Crassula may be at odds with what occurs else- where. This species is often apparent after fire in chaparral, however, this may be due to increased biomass of individuals, not increased densities. In conclusion, seed banks in the maritime chap- arral I studied may differ from most inland coun- terparts in the following ways: 1) greater impor- tance of non-refractory seed, 2) lack of entirely fire- dependent germination in short-lived species, 3) germination among Arctostaphylos stimulated by heat and especially heat and charred wood extracts together, 4) more strongly enforced dormancy among Ceanothus spp. and 5) greater importance of fire-suppressed germination. Conversely, germi- nation of the Adenostoma seed bank appears con- sistent with what occurs elsewhere. Further study of soil seed banks will be required to determine whether there has in fact been divergence in the germination ecology of maritime chaparral. In par- ticular, it would be illuminating to directly compare seed banks of species that occur in both inland and maritime chaparral. Acknowledgments This research was supported by NSF Grant No. SES- 8721494. James and Lauds Rose provided outdoor grow- ing space at their nursery. Frank Davis, Claudia Tyler, Mark Borchert, Bruce Mahall, Carla D' Antonio, Bob Haller, and Nancy Vivrette provided insightful feedback L over the course of the study. Pat Shafroth, Joe Mailander, and Kelly Cayocca assisted with the collection and treat- ment of seed bank samples. Clifton Smith identified a number of species. I express my sincere thanks to each for these contributions. Literature Cited AuLD, T. D. AND M. A. O'CoNNELL. 1991. Prcdicfing pat- terns of post-fire germination in 35 eastern Australian Fabaceae. Australian Journal of Ecology 16:53-70. i Baskin, C. C. and J. M. Baskin. 1998. Seeds: ecology, I biogeography, and evolution of dormancy and ger- mination. Academic Press, San Diego, CA. I Berg, A. R. 1974. Arctostaphylos Adans. manzanita. j Seeds of woody plants in the United States, USDA ! Forest Service. Bond, W. J. and B. W. van Wilgen. 1996. Fire and plants. Chapman and Hall, New York, New York. Borchert, M. I. and D. C. Odion. 1995. Fire intensity and vegetation recovery in chaparral: a review, p 91- 100. in J. E. Keeley, and T. Scott (eds.), Brushfires in i California: ecology and resource management. Inter- j national Association of Wildland Fire, Fairfield, WA. Bullock, S. H. 1982. Reproducfive Ecology of Ceanothus \ cordiilatiis. M. A. Thesis, California State University, ! Fresno. \ Christensen, N. L. and C. H. Muller. 1975. Effects of fire on factors controlling plant growth in Adenosto- ma chaparral. Ecological Monographs 45:29-55. D'Antonio, C. M., D. C. Odion and C. M. Tyler. 1993. Invasion of maritime chaparral by the alien succulent Carpobrotiis edulis: the roles of fire and herbivory. Oecologia 95:14-21. Davey, J. R. 1982. Stand Replacement in Ceanothus cras- sifolius. M.S. Thesis, California State Polytechnic University, Pomona. Davis, F W., D. E. Hickson and D. C. Odion. 1988. Com- position of maritime chaparral related to fire history } and soil. Burton Mesa, California. Madrono 35:169- i 195. , M. I. Borchert and D. C. Odion. 1989. Estab- lishment of microscale vegetation pattern in maritime chaparral after fire. Vegetatio 84:53-67 Dibblee, T W. 1950. Geology of southwestern Santa Bar- bara County. Bulletin 150, California Division of Mines, Sacramento. Emery, D. 1988. Seed Propagafion of Native Cafifornia Plants. Santa Barbara Botanic Garden, Santa Barbara, California. Griffin, J. R. 1978. Maritime chaparral and endemic shrubs of the Monterey Region, California. Madrono 25:65-81. Hickman, J. C. (ed.). 1993. The Jepson manual: higher plants of California. University of California Press, Berkeley. Holl, K. D., H. N. Steele, M. H. Fusari and L. R. Fox. Seed banks of maritime chaparral and abandoned roads: potential for vegetation recovery. Journal of the Torrey Botanical Society 127:207-220. Kaminsky, R. 1981. The microbial origin of the allelo- pathic potential of Adenostoma fasciculatum H. & A. Ecological Monographs 51:365-382. Keeley, J. E. 1975. Longevity of nonsprouting Ceano- thus. American Midland Naturalist 93:505-507. . 1977. Seed production, seed populations in soil, and seedling producfion after fire for two congeneric 2000] ODION: GERMINATION OF MARITIME CHAPARRAL SEED BANK 203 pairs of sprouting and nonsprouting chaparral shrubs. Ecology 58:820-829. ■ . 1984. Factors affecting germination of chaparral seeds. Bulletin of the Southern California Academy of Sciences 83:1 13-120. . 1987. Role of fire in seed germination of woody taxa in California chaparral. Ecology 68:434-444. . 1991. Seed germination and life history syn- dromes in the California chaparral. The Botanical Re- view 57:81-1 16. . 1992. Demographic structure of California chap- arral in the long-term absence of fire. Journal of Veg- etation Science 3:79-90. . 1993. Utility of growth rings in the age deter- mination of chaparral shrubs. Madroiio 40:1-14. , AND M. NiTZBERG. 1984. The role of charred wood in the germination of the chaparral herbs E)n- menanthe penduliflora (Hydrophyllaceae) and Erio- phvilum confertiflorum { Asteraceae). Madroiio, 208- 218. , B. A. Morton, A. Pedrosa and P. Trotter. 1985. The role of allelopathy, heat and charred wood in the germination of chaparral herbs and suffrutes- cents. Journal of Ecology 73:445-458. , and C. J. Fotheringham. 1997. Trace gas emis- sions and smoke-induced seed germination. Science 276:1248-1250. , and . 1998. Smoke-induced seed germi- nation in California chaparral. Ecology 79:2320- 2336. Keeley, S. C, J. E. Keeley, S. M. Hutchinson and A. W. Johnson. 1981. Postfire succession of the herba- ceous flora in southern California chaparral. Ecology 62:1608-1621. , and M. Pizzorno. 1986. Charred wood stimulat- ed germination of two-fire following herbs of the Cal- ifornia chaparral and the role of hemicellulose. Amer- ican Journal of Botany 73:1289-1297. Keith, D. A. 1987. Combined effects of heat shock, smoke and darkness on germination of Epacris stuar- tii Stapf., an endangered fire-prone Australian shrub. Oecologia 112:340-344. Mcpherson, J. K. and C. H. Muller. 1969. Allelopathic effects of Adenostomci fasciciilatum, 'chamise\ in the California chaparral. Ecological Monographs 39:177- 198. Moreno, J. M. and W. C. Oechel. 1991. Fire intensity effects on germination of shrubs and herbs in south- ern California chaparral. Ecology 72:1993-2004. MuLLER, C. H., R. B. Hanawalt and J. K. Mcpherson. 1968. Allelopathic control of herb growth in the fire cycle of California chaparral. Bulletin of the Torrey Botanical Club 95:225-231. Odion, D. C. 1995. Effects of variation in soil heating during fire on patterns of plant establishment and re- growth in maritime chaparral. Dissertation, Univer- sity of California, Santa Barbara. , AND F. W. Davis. 2000. Fire, soil heating, and the formation of vegetation patterns in chaparral. Ecolog- ical Monographs 70:149-169. , D. E. Hickson, and C. M. D' Antonio. 1992. Central coast maritime chaparral on Vandenberg air force base: an inventory and analysis of management needs for a threatened vegetation association. Report prepared for The Nature Conservancy. , J. Storrer and V. Semonsen. 1993. Biological resources assessment. Burton Mesa Project Area, Santa Barbara County, California. Report prepared for Santa Barbara County, Resource Management De- partment. Parker, V. T. 1987. Effects of wet-season management burns on chaparral vegetation: implications for rare species. Pp. 233-237. in T. S. Elias and J. Nelson (eds.). Conservation and management of rare and en- dangered plants. California Native Plant Society, Sac- ramento. , and V. R. Kelly. 1989. Seed banks in California chaparral and other Mediterranean climate shrub- lands. Pp. 231-255. in M. A. Leek, V. T Parker and R. L. Simpson (eds.). Ecology of soil seed banks. Academic Press Inc., San Diego, CA. Quick, C. R. 1935. Notes on the germination of Ceano- thus seeds. Madrofio 3:135-140. , AND A. S. Quick. 1961. Germination of Ceano- thus seeds. Madrofio 16:23-30. Sampson, A. W. 1944. Plant succession on burned chap- arral lands in northern California. University of Cal- ifornia Agricultural Experiment Station Bulletin 685. Shmida, a. and R. H. Whittaker. 1981. Pattern and bi- ological microsite effects in two shrub communities. Southern California. Ecology 62:234-251. Stone, E. C. and G. Juhren. 1951. The effects of fire on the germination of the seed of Rhus ovata Wats. American Journal of Botany 38:368-372. , AND . 1953. Fire stimulated germination. California Agriculture 7:13-14. Sweeney, J. R. 1956. Responses of vegetation to fire. Uni- versity of California Publications in Botany 28:143- 250. Swank, S. E. and W. C. Oechel. 1991. Interactions among the effects of herbivory, competition, and re- source limitation on chaparral herbs. Ecology 72: 104-115. Thanos, C. a. AND P. W. RuNDEL. 1995. Fire-followers in chaparral: nitrogenous compounds trigger seed ger- mination. Journal of Ecology 83:207-216. Tyler, C. M. 1995. Factors contributing to postfire seed- ling establishment in chaparral: direct and indirect ef- fects of fire. Journal of Ecology 83:1009-1020. , D. C. Odion and D. Meade. 1998. Ecological studies of Morro Manzanita {Arctostaphylos mor- roensis), seed ecology and reproductive biology. Re- port prepared for California Department of Fish and Game, Species Conservation and Recovery Program. , , AND M. A. Moritz. 2000. Fac- tors affecting regeneration of Morro Manzanita (Arc- tostaphylos morroensis): reproductive biology and re- sponse to prescribed burning. Report prepared for the California Department of Fish and Game, Species Conservation and Recovery Program. Wells, P. V. 1969. The relation between mode of repro- duction and extent of speciation in the woody genera of the California chaparral. Evolution 23:264-267. Wicklow, D. T. 1977. Germination response in Emmen- anthe penduliflora (Hydrophyllaceae). Ecology 58: 201-205. Whelan, R. J. 1995. The ecology of fire. Cambridge Uni- versity Press, Cambridge, Great Britain. Wright, E. 1931. The effect of high temperature on seed germination. Journal of Forestry 29:679-687. Zammit, C. a. and R H. Zedler. 1988. The influence of dominant shrubs, fire and time since fire on soil seed banks in mixed chaparral. Vegetatio 75:175-187. , and p. H. Zedler. 1994. Organisation of the soil seed bank in mixed chaparral. Vegetatio 1 1 1:1-16. Madrono, Vol. 47, No. 3, pp. 204-205, 2000 HEDOSYNE (COMPOSITAE, AMBROSIINAE), A NEW GENUS FOR IVA AMBROSIIFOLIA John L. Strother 1001 Valley Life Sciences Building #2465, University Herbarium, University of California, Berkeley, California 94720-2465 Abstract Hedosyne Strother is a new genus based on Iva ambrosiifolia (A. Gray) A. Gray {=Euphrosyne ambrosiifolia A. Gray = Hedosyne ambrosiifolia (A. Gray) Strother]. Plants of Hedosyne differ from those of Iva s.s. in having leaves mostly alternate, leaf blades 1-3 times pinnately divided or lobed, and capitulescences paniculiform and either ebracteate or with 3-6+ heads per bract. In morphology-based cladistic analyses of Iva L. and other genera of Ambrosiinae, Bolick (1983) placed Iva ambrosiifolia (A. Gray) A. Gray sister to Xanthiiim L. and Karis (1995) placed /. ambro- siifolia sister to Euphrosyne parthenifolia DC. (the type and only species of Euphrosyne). Miao et al. (1995a, b) reviewed relationships of ivas and other Ambrosiinae with respect to variations in chloro- plast DNA and nuclear rDNA; they concluded that /. ambrosiifolia did not result from hybridization and they placed /. ambrosiifolia sister to Dicoria canescens A. Gray. Although they differ in their placements of some species, Bolick, Karis, and Miao et al. all consid- ered Iva S.I., i.e., Iva sensu Jackson (1960), to in- clude species that have closer relationships outside Iva s.l. than within. I agree that five species (con- stituting Iva sect. Cyclachaena (Fresenius) A. Gray, sensu R. C. Jackson) should be withdrawn from Iva s.l. and I treat them as monotypic genera: Chorisiva Rydberg [C. nevadensis (M. E. Jones) Rydberg = Iva nevadensis M. E. Jones], Cyclachaena Fresen- ius [C. xanthifolia (Nuttall) Fresenius = Iva xan- thifolia Nuttall], Leuciva Rydberg [L. dealbata (A. Gray) Rydberg = Iva dealbata A. Gray], Oxytenia Nuttall [O. acerosa Nuttall = Iva acerosa (Nuttall) R. C. Jackson], and a new genus, Hedosyne [see following]. Iva s.s. and the other genera may be distin- guished as indicated in the following key: 1 . Capitulescences racemiform or spiciform, bracteate with 1-2 heads per bract Iva s.s. 1 . Capitulescences paniculiform, ± ebracteate or with 3-6+ heads per bract, or heads ± scattered. 2. Leaves all or mostly opposite, blades rarely lobed or divided, mostly deltate, triplinerved, and ± toothed Cyclachaena 2. Leaves all or mostly alternate, some or all blades ± pinnately laciniate-lobed or 1-3 times pinnately divided. 3. Plants suffrutescent or shrubby; phyllaries, paleae, and cypselae ± villous Oxytenia 3. Plants mostly herbs, rarely woody at base; phyllar- ies, paleae, and cypselae glabrous or strigillose and/ or hispidulous. 4. Leaf blades laciniately lobed, the lobes mostly 3- 12+ mm wide, abaxial faces ± lanate, the adaxial ± tomemtose Leuciva 4. Leaf blades mostly 1-3 times pinnately divided, the lobes 1-3 mm wide, abaxial and adaxial faces ± scabrellous and/or hispidulous. 5. Heads ± scattered; herbaceous phyllaries usually 3, usually longer than the florets; lobes of corollas of functionally staminate florets erect Chorisiva 5. Heads in paniculiform arrays; herbaceous phyllaries usually 5, ± equalling the florets; lobes of corollas of functionally staminate florets reflexed . . . Hedosyne Hedosyne Strother, gen. nov. A Iva s.s. foliis pro parte maxima alternatis 1-3- pinnatis, capitulescentiis laxe paniculiformibus ± ebracteatis vel capitulis 3—6+ ad quoque bracteam, et corollis florum pistillatorum nullis differt. Type: Euphrosyne ambrosiifolia A. Gray = Iva ambrosiifolia (A. Gray) A. Gray = Hedosyne am- brosiifolia (A. Gray) Strother The name Hedosyne comes from the Greek word hedosyne, meaning delight (see Brown 1956), and is, I believe, a suitable name for a sister or step- sister to Euphrosyne, one of the three Graces. As here circumscribed, Hedosyne includes a single species: Hedosyne ambrosiifolia (A. Gray) Strother, comb, nov. Basionym: Euphrosyne ambrosiifolia A. Gray, PI. Wright. 1:102. 1852, as ambrosiaefolia. = Iva ambrosiifolia (A. Gray) A. Gray in A. Gray et al., Syn. Fl. N. Amer. 1(2): 246. 1884. = Cyclachaena ambrosiifolia (A. Gray) Rydberg N. L. Britton et al., N. Amer. Fl. 33:10. 1922. — Type: western Texas or adjacent New Mexico, IMay-Oct. 1849, C. Wright "310" (GH; isotypes: UC! US). Cyclachaena lobata Rydberg in N. L. Britton et al. N. Amer Fl. 33:10. 1922. = Iva ambrosiifolia (A. Gray) A. Gray subsp. lobata (Rydberg) R. C. Jackson, Univ. Kansas Sci. Bull. 41:838. 1960. — Type: IVlexico, Nuevo Leon, Monterrey, Aug 1911, Alban and Arsene 208 (US; isotype: MO). 2000] STROTHER: HEDOSYNE 205 Habit annual. Stems erect, l-5(-10) dm. Leaves mostly alternate, petioles 5-12(-45) mm long, blades deltate or ovate to lanceolate in outline, mostly 3-5(-9) cm long, 4-5(-8) cm wide, 1-3 times pinnately divided, ultimate lobes oblong to lance-linear, 1-3 mm wide, faces scabrellous and/ or hispid, usually gland-dotted. Capitulescences loosely paniculiform, ± ebracteate or heads 3-6+ along an axis from the axil of each bract; peduncles 3-12+ mm long. Involucres ± hemispheric, 2-3 + mm high. Phyllaries 10-12+ in 2+ series, free, the outer 5 ± herbaceous, about equalling the florets, the inner phyllaries scarious to membranous, equalling or shorter than the outer. Pistillate florets 5-10, corollas none. Functionally staminate flo- rets 5-10(-20+), corollas funnelform, 1.5-2 mm long, the lobes soon reflexed. Receptacles hemi- spheric; paleae spatulate to linear, membranous. Cypselae pyriform, ± obcompressed, 1.4-1.7 mm long, finely striate, glabrous (said to become mu- ricate in age); pappus none, jc = 18. Plants of Hedosyne ambrosiifolia usually grow in sandy, sometimes gypseous or calcareous soils, often in disturbed places (roadsides, washes, etc.) in southwestern United States (Arizona, New Mex- ico, Texas) and northwestern Mexico (Chihuahua, Coahuila, Durango, Nuevo Leon, San Luis Potosi, Sonora, Zacatecas). Acknowledgments I thank B. Baldwin, R. C. Jackson, G. Nesom, and A. R. Smith for help in various ways. Literature Cited BoLiCK, M. R. 1983. A cladistic analysis of the Ambro- siinae Less, and Engelmanniinae Stuessy, Pp. 125- 141. in N. 1. Platnick and V. A. Funk (eds.). Advances in cladistics. Vol. 2. Brown, R. W. 1956. Composition of scientific words, re- vised edition. Published by the author, Baltimore, Md. Jackson, R. C. 1960. A revision of the genus Iva. Uni- versity of Kansas Science Bulletin 41:793-876. Karis, p. O. 1995. Cladistics of the subtribe Ambrosiinae (Asteraceae: Heliantheae). Systematic Botany 20:40- 54. MiAO, B., B. L. Turner, and T. J. Mabry. 1995a. Chlo- roplast DNA variations in sect. Cyclachaena of Iva (Asteraceae). American Journal of Botany 82:919- 923. , , and . 1995b. Systematic impli- cations of chloroplast DNA variation in the subtribe Ambrosiinae (Asteraceae: Heliantheae). American Journal of Botany 82:924-932. Madrono, Vol. 47, No. 3, pp. 206-208, 2000 BOOK REVIEWS 2'"' Interface Between Ecology and Land Develop- ment in California. Edited by J. E. Keeley, M. Baer-Keeley, and C. J. Fotheringham. 2000. U.S. Geological Survey Open-File Report 00-62. Sac- ramento. This book brings together a set of papers deliv- ered at the 2"^^ Interface Between Ecology and Land Development Conference held in 1997. I had the pleasure to review the proceedings from the first conference, held in 1992 (Keeley 1993), and it is interesting to observe how the issues have both de- veloped and remained the same in the intervening time. The issues and problems of rampaging land de- velopment and how these impact the natural envi- ronment are important virtually everywhere in the world. They are particularly acute in California, where urban development, in particular, appears to the outsider to be virtually out of control. These pressures, together with the state's biotic diversity place California as one of the biodiversity "hot- spots" of the world (Myers et al. 2000) and one of the regions most likely to undergo massive biotic change (Sala et al. 2000). California is one of the few recognized hotspots in the northern hemi- sphere, and is unique in its position in the world economy. In few parts of the world is such excep- tional affluence and quality of life set against a rich and varied natural environment and biota. Califor- nia thus presents an interesting litmus test for whether we can successfully develop methods and approaches to integrating development and conser- vation. If California, with its affluence and rela- tively educated population, cannot tackle the prob- lems effectively, what hope is there for the rest of the world, where public and private funds are scarc- er and conservation ranks much lower on the list of important issues? So this collection of papers is exceptionally in- teresting from an outsider's perspective. The issues discussed here, while focusing on the Californian situation, are relevant in most parts of the world. The book starts with a paper from Mike Soule, which is the transcript from his keynote address at the conference, and as such is very conversational and discursive. Soule's topic is the Wildlands Pro- ject and the need to be bold when tackling the creeping development crisis. While I'm sympathet- ic to his argument, I'm not sure that it does much to help the main issue of the conference. In the face of rampaging land development, it's not enough any more to say "Stop it" and seek to put as much land as possible into conservation networks. Over much of coastal California, it's too late for a Wild- lands Project approach, but there is nevertheless much of conservation value that needs to be pro- tected and managed, and much that can be done to direct and control development so that it is more ecologically acceptable. The rest of the book tackles some key areas around this problem. The first section looks at the thorny issue of fire management along the wild- land/urban interface, and the papers reflect a range of opinions and approaches. A key message in many of the papers is that prescribed fire is likely to be a key management tool, but that there are many questions and issues still to be resolved. The next section deals with habitat fragmentation and its impacts on biota, and the next with NCCP and Land Planning. A final, long section discusses various aspects of restoration ecology, or how to repair damaged eco- systems and return species to areas of their former range. Some may view this increasing emphasis on restoration as defeatist, maintaining that the main game should still be the preservation of undam- aged, pristine habitat. The unfortunate fact is that the amount of undamaged habitat remaining is de- clining all the time and the maintenance of some ecosystems and species demands that we take re- medial action and repair some of the past damage. This series of papers represents a fine collection of current ideas and approaches in this area. I cannot agree with Peter Bowler's plea that everyone, in- cluding individuals and agencies, has to concentrate on getting to know small local areas intimately and restoring a few acres. This is fine in theory and will work in a few cases — however, it belies the im- mensity of the task ahead. If we are serious about broad-scale management and restoration, we will have to deal with vast areas, and this demands that we have effective, replicable treatments which can be applied over large areas. Of course, small-scale restoration is also essential as an ideal way to in- volve our increasingly urbanized population in con- tact with nature — recreating bits of nature in cities probably represents one of the best hopes for turn- ing around our current crazy development path. This needs more emphasis on the value of city wildlife and more ecologists prepared to spend time in the cities instead of seeking out nice wild areas far away! Overall, this collection of papers presents an in- teresting cross-section of current issues and ap- proaches which should be of value to anyone with an interest in ecology and conservation in Califor- nia. The coverage is inevitably variable in depth and comprehensiveness, but there is still a wealth of information here. The index is a bit if a token effort and not really very helpful, and there are a few more typos than you would expect. Perhaps the 2000] REVIEWS 207 most serious omission from the book is a synthetic chapter at the end. This would have increased the value of the book immensely. If we are to make progress in dealing with issues such as the interface between ecology and land development, we as sci- entists need to make sure that we ourselves inter- face effectively. Only by communicating clearly and concisely about these issues with the people who are making and regulating development deci- sions will we have any hope of changing the way things happen. Most developers and regulators will not read the papers in this book, but they are the ones that we most need to engage with. Maybe if there is to be a third conference in this series, that is the issue on which it should concentrate. — Richard J. Hobbs, School of Environmental Science, Murdoch University, Murdoch, WA 6150, Australia rhobbs@essun 1 .murdoch.edu.au Literature Cited Keeley, J. E. (ed.). 1993. Interface between ecology and land development in Calfirornia. Southern California Academy of Sciences, Los Angeles. Myers, N., R. A. Mittermeier, C. G. Mittermeier, G. A. B. DA Fonseca and J. Kent. 2000. Biodiversity hot- spots for conservation priorities. Nature 403:853- 858. Sala, O. E., F. S. I. Chapin, J. J. Armesto, E. Berlow, J. Bloomfield, R. Dirzo, E. Huber-Sanwald, L. F. Huenneke, R. B. Jackson, A. Kjnzig, R. Leemans, D. M. Lodge, H. A. Mooney, M. Oesterheld, N. L. Pope, M. T. Sykes, B. H. Walker, M. Walker and D. H. Wall. 2000. Global biodiversity scenarios for the year 2100. Science 287:1770-1174. Synthesis of the North American Flora. Version 1 .0. By John T. Kartesz and Christopher A. Meacham. 1999. North Carolina Botanical Garden, University of North Carolina at Chapel Hill. In cooperation with The Nature Conservancy, NRCS, USDA, USFWS, USDI. $495.00. ISBN 1-889065-05-6. Minimums requirements, a Pentium 90 MHz class processor, 32 MB RAM, 25 MB free hard drive space, SVGA display (minimum resolution, 800 by 600, 1280 by 1024 recommended) with 16 colors, Microsoft Windows 3.1, 95, 98, NT, or 2000 operating system, CD-ROM drive for installation. This is a most impressive work. An update of, and an expansion on Kartesz 1994, A Synonymized Checklist of the Vascular Flora of the United States, Canada, and Greenland, Second edition. Volume 1, Checklist and 2, Thesaurus, Timber press, Portland. Unlike its predecessor, near-instant answers, com- parisons, and analyses can be obtained to a multi- tude of questions within and beyond the scope of the printed work. It contains a comprehensive da- tabase with a high level of accuracy on the taxon- omy, nomenclature, phytogeography, and biologi- cal attributes of the North American vascular flora (by Kartesz) combined with highly functional soft- ware for accessing the database (by Meacham). Thus, a slick and versatile product. The cover in the jewel case is a six page insert. Included is "No- menclatural Innovations" with 41 new combination (see International Code of Botanical Nomenclature (Saint Louis Code), 2000; Recommendation 30A.1. Ex. 1). Installation of the product is simple. The "Overview of Basic Functions" in the help menu can be printed for immediate reference or accessed as needed. As the title indicates, this work covers North America north of Mexico. Treated are all continen- tal states and the District of Columbia for the U.S.A., all provinces of Canada with Newfound- land displayed separate from Labrador and the Northwest Territories by administrative district (Keewatin, Mackenzie, and Franklin), the islands of St. Pierre and Miquelon, and Greenland. Further- more, Puerto Rico, the U.S. Virgin Islands, and Ha- waii are also included. The primary screen contains three nomenclature windows on the left with lists by family, genus, infrageneric name (specific, subspecific, varietal ep- ithets), respectively. A box above each list allows one to type the first few letters of a name and then click on it, or one can scroll and click on a name (options, with common or contrived names, au- thors, hybrids, synonyms, and either in checklist or thesaurus fo