A a joe tore ss rata Pl tio aw eo ok ate i ne at a Santen tn terest ter ata tac crnnte saree setetor aetna a “s - ey ce A One Oe LAD ee nee al ~ een eer or ine eras = a nent Se nn ee eh ane manera etna RN ene ee tasenentt pena SS ere serene pane ba eens ee ete ee a renee tte ap naanchrnen geo sneOr
I ea ors apt AP EPSP A F Cent Pet a ah ove ’ ta maaan as ae erate a a nS ee ee ee Sar cage aa en ngs eon “ - 0 at te eT TP OTT he petite art hae OT S Cr atiner os Ole e=O thea namaste dens ota ten tT en asa ae ree MPD DBT MDITOE DE eH
ae aor we ae oon? appear er Tee mene = a me ad nigel ee thn IN npn ere ~e8to= at te ee ni tel ty OTOP Tet eg ~~
srapaltrtes ee etn ee ee ree TR Sem
ee oe -
ee ee AM a ee Ol On |
‘'D
be thts Seine
A) of x fl a é BAI ae 1 4 | hae . | 4 fl : Oe ee arn Dani & , 7, eer ant uA a - tL, \,
ae i
mt Sas a
SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOL: 101
tte
“EVERY MAN IS A VALUABLE MEMBER OF SOCIETY WHO, BY HIS OBSERVATIONS, RESEARCHES, AND EXPERIMENTS, PROCURES KNOWLEDGE FOR MEN ’’—SMITHSON
(PUBLICATION 3695)
CITY OF WASHINGTON PUBLISHED BY THE SMITHSONIAN INSTITUTION 1942
The Lord Baltimore Press
BALTIMORE, MD., U. 8 As
ADVERTISEMENT
The Smithsonian Miscellaneous Collections series contains, since the suspension in 1916 of the Smithsonian Contributions to Knowledge, all the publications of the Institution except the Annual Report, the annual volume describing the Institution’s field work, and occasional publications of a special nature. As the name of the series implies, its scope is not limited, and the volumes thus far issued relate to nearly every branch of science. Papers in the fields of biology, geology, anthropology, and astrophysics have predominated.
C. G. Aszort, Secretary of the Smithsonian Institution.
(iil)
i 7
ai \\:
. iV
on
N
10.
te
nee
CONTENTS
Asgot, C. G. An important weather element hitherto generally disregarded. 34 pp., 11 figs., May 27, 1941. (Publ. 3637.) MiTTLEMAN, M. B., and Jopson, Harry G. M. A new sala- mander of the genus Gyrinophilus from the southern Appa-
lachians. 5 pp., I pl., July 14, 1941. (Publ. 3638.)
WeEDEL, WaLpo R. Environment and native subsistence econo- mies in the central Great Plains. 29 pp., 5 pls., 1 fig., Aug. 20, 1941. (Publ. 36309.)
HrpiicKa, ALES. Diseases of and artifacts on skulls and bones from Kodiak Island. 14 pp., 11 pls., Sept. 25, 1941. (Publ. 3640. )
ArcTOWSKI, HENRYK. On solar constant and atmospheric tem- perature changes. 62 pp., 33 figs., Nov. 7, 1941. (Publ. 3641.)
DoszHANSky, TH. Beetles of the genus Hyperaspis inhabiting the United States. 94 pp., 6 pls., Dec. 31, 1941. (Publ. 3642.)
WeEDEL, WaAtpo R. Archeological remains in central Kansas and their possible bearing on the location of Quivira. 24 pp., roipls;, 1 fis., Jans 15, 1942. (Publ. 3647.)
CocKERELL, T. D. A. Bees of the family Hylaeidae from the Ethiopian region. 15 pp., Feb. 19, 1942. (Publ. 3649.)
SHOEMAKER, CLARENCE R. Notes on some American fresh- water amphipod crustaceans and descriptions of a new genus and two new species. 31 pp., 12 figs., Feb. 16, 1942. (Publ. 36735.)
REsSER, CHARLES E. Faunal content of the Maryville formation. Sopp, Feb: 12,1942. (Publ. 3676:)
SHOEMAKER, CLARENCE R. Amphipod crustaceans collected on the Presidential Cruise of 1938. 52 pp., 17 figs., Mar. 5, 1942. (Publ. 3677.)
Assot, C. G. The quantity of vaporous water in the atmophere. 7 pp., Mar. 23, 1942. (Publ. 3678.)
Gazin, C. Lewis, and Suttivan, J. MacRuper. A new titano- there from the Eocene of Mississippi with notes on the cor- relation between the marine Eocene of the Gulf Coastal Plain and continental Eocene of the Rocky Mountain region. 13 pp., a pls; 1 fig.. Apr: 23, 1942, (Publ: 3679.)
(v)
14.
18.
WeEtTMorE, ALEXANDER. Two new fossil birds from the Oligo- cene of South Dakota. 6 pp., 13 figs., May 11, 1942. (Publ. 3080. ) :
RessEr, CHARLES FE. Fifth contribution to nomenclature of Cambrian fossils. 58 pp., May 22, 1942. (Publ. 3682.)
Kot, ErzsEBeT. The snow and ice algae of Alaska. 36 pp., Gipls.,5 fles.j) Sepe) 19, 1042, (Publ: 526832)
WEINTRAUB, RoBERT L., and McAListTEr, Epwarp D. Develop- mental physiology of the grass seedling. 1. Inhibition of the mesocotyl of Avena sativa by continuous exposure to light of low intensities. 10 pp., 1 pl., 4 figs., June 24, 1942. (Publ. 3685.)
Scumitt, WaLpo L. A new species of sand bug, Blepharipoda doelloi, from Argentina. 10 pp., I pl., Aug. 10, 1942. (Publ. 3687.)
(vi)
i cee al: ral ae We eC ts Pa Ay ae jl iy Pa | h i }
Ay ania
NA ul
Tn
Lie ; \ \
tt i ls
An
nk aM, iV fe
Me fit ny i i]
iy
np, a FM 1e is ms
. \
o]
es AWW) i i, i
+
ah
i,
le hile | ay
i ru, i
i i Wi
ry
- SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 101, NUMBER 1
“AN. ‘IMPORTANT WEATHER ELEMENT
Mf 5a - fF \
he \ We } | 4s ion qt A, . 3 nay ; er S eas 4 & | oa , yy A an ‘ 4 ’
HITHERTO GENERALLY DISREGARDED
BY) G. G. ABBOT
Secretary, Smithsonian’ Institution —
(PUBLICATION 3637)
ony OF WASHIN GTON
pate PUBLISHED BY THE SMITHSONIAN INSTITUTION
MAY 27, 1941
ey
SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 101, NUMBER 1
AN IMPORTANT WEATHER ELEMENT HITHERTO GENERALLY DISREGARDED
BY CG. G. ABBOT
Secretary, Smithsonian Institution
(PUBLICATION 3637)
CITY OF WASHINGTON PUBLISHED BY THE SMITHSONIAN INSTITUTION MAY 27, 1941
THe Lord Baltimore Press
BALTIMORE, MD., U. 8 A.
AN IMPORTANT WEATHER ELEMENT HITHERTO GENERALLY DISREGARDED
By C. G. ABBOT
Secretary, Smithsonian Institution
I wish to present evidence that, though small in percentage, the variation of the sun’s output of radiation is an effective weather element.
THE VARIATION OF THE SUN
In other papers I have described the means used since 1902 by Smithsonian observers to measure the energy contained in solar radia- tion, to estimate the losses which the solar beam suffers in traversing the atmosphere, and to evaluate the “solar constant of radiation.” The solar constant may be defined as the average intensity of solar radia- tion in free space at mean distance of the earth from the sun. Ex- pressed in heat units its value is about 1.94 calories per square centimeter per minute.
By many thousands of determinations, Smithsonian observers have found that the solar constant changes from day to day and from month to month around its mean value. The extreme range of these excursions thus far observed is 54 percent. Monthly mean values since 1920 have an extreme range from 1.91 to 1.96 calories or 24 percent. Superposed on these fluctuations of monthly means are de- partures of a few days in length dependent to a great extent on the rotation of the sun. These short-interval changes, superposed on the changes indicated by monthly mean values, widen the extreme limits of solar variation to those given above as 54 percent.
The Smithsonian Institution operates several desert mountain solar- observing stations. Two of these, long occupied, are Table Mountain, Calif. (prior to 1926 at Harqua Hala, Ariz.), and Montezuma, Chile. These stations are at elevations of 7,500 and 9,000 feet, respectively. About 80 percent of all the days of the year are nearly cloudless at these stations, though not always of first-rate quality for solar-constant work.
Figure I gives the march of the monthly means of the solar constants from (A) combined results of Harqua Hala and Table Mountain, and (B) from Montezuma. The curves cover the years 1920, to 1939.
SMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 101, No. 1
2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
The preparation of figure I was delayed in order to use later results of the revision of solar-constant values. Individual daily values, how- ever, will still be under consideration and revision for several more months, so that even the curves shown in figure 1 are not final. Yet they are improved and differing somewhat from those shown in figure 8, which was prepared at an earlier stage of the revision.
Figure I employs revised Montezuma data, beginning with Sep- tember 1923, and revised Table Mountain data, beginning with December 1925. Montezuma data from August 1920 to August 1923
1927 1928
Fic. 1—Solar constant monthly means. Dotted curve North America, full curve South America.
are taken from table 31 of volume 5 of the Annals of the Astro- physical Observatory.
The Harqua Hala data employed in figure 1 are not fully satis- factory. They are derived as follows. Table 29-a of volume 5 of the Annals gives daily ‘“‘long-method” Harqua Hala solar constants, from which monthly mean values were computed. The resulting monthly means, when compared with simultaneous Montezuma “‘short-method” data given in table 31 of volume 5 of the Annals, indicate an average deficiency of 0.004 calorie in scale for the Harqua Hala “long-method”’ data. Referring to table 1 of my paper “Provisional Solar Constant
NO. I AN IMPORTANT WEATHER ELEMENT—ABBOT 3 Values,” * the monthly mean “short-method” Harqua Hala values were recomputed from October 1920 to November 1924. Daily values marked ‘‘U” were omitted from the means. Upon comparing these monthly “short-method” means with the simultaneous monthly “long- method” means from table 29-a of volume 5 of the Annals, there seemed to be no marked difference in scale up to December 1922, but thereafter there was found to be an average deficiency in the “short- method” monthly means of 0.008 calorie up to January 1924, and of 0.023 calorie, February 1924 to October 1925. It is not possible now, without a very tedious and thorough examination of the original records to detect the cause of this change of scale, but an allowance was made, as follows.
The “short-method” daily values are much more numerous than the “long-method” ones, and are also regarded as less subject to accidental errors. Hence, I preferred to use them for figure 1. Taking account of the discrepancies of scale just mentioned, the Harqua Hala results plotted in figure 1 are based on the “short-method” daily values, but are reduced to the Montezuma scale by adding 0.004 calorie for the interval October 1920 to December 1922, and 0.004+0.008=0.012 calorie from January 1923 to January 1924. From February 1924 to October 1925 there was added 0.004 + 0.023 =0.027 calorie.
The interval December 1924 to December 1925, at which latter month the revised Table Mountain values begin, is not covered in “Provisional Solar Constant Values.’ * This interval has been nearly closed by using hitherto unpublished Harqua Hala “short-method” monthly means. Comparison with “long methods” determines their scale corrections in the same manner explained above, and I increased these “‘short-method” monthly means, as stated above, by 0.027 calorie. As the Harqua Hala data are obviously less eligible, I give the average monthly mean differences from Montezuma values for the interval 1920 to 1925, inclusive, separately from those for 1926 to 1940. But all are used in combination in the grand mean for the interval 1920 to 1940.
Mean (Montezuma—Harqua Hala) 1920 to 1925 = +0.32 percent Mean (Montezuma—Table Mt.) 1926 to 1940 =+0. i Grand mean, 1920 to 1940 = 20)
Notwithstanding differences of 1,500 feet in altitude, 4,600 miles in latitude, summer conditions at the one station simultaneous with winter conditions at the other, 34° of north latitude against 22° of
* Smithsonian Misc. Coll., vol. 77, No. 3, 1925.
4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
south latitude, these independently derived monthly means, which rest on determinations of approximately 5,000 separate days at each station, agree to within an average monthly difference of 0.24 percent. Fluc- tuations are shown of 24 percent in the solar constant which are common to both stations.
As computed from the average monthly differences, the probable error of a monthly mean value derived from the work of two stations is
0.84 X 0.24 7 ; ——_——=0.14 percent. Thus the range of variation found is 18
Wee times the probable error of the determination. In many cases the two stations unite to show continued trends of increasing or of decreasing solar-constant values during many months. In such cases, successive values support each other, so that the probability with which such trends may be ascribed to observational error is further reduced. These facts support the validity of solar variation.
I conceive, however, that evidences of the reality of solar variation, stronger even than this agreement between results of high accuracy from two far-separated stations, are to be found in correlations of these results with other phenomena, both solar and terrestrial. I shall discuss these below.
But before presenting the most striking of these correlations I will refer to many supporting evidences of solar variability heretofore published. Though these evidences individually may not all be weighty, yet taken altogether they are highly convincing, on the principle of the well-known fable of the bundle of fagots. I present them in figure 2 and give appropriate references to the literature, in what follows, covering these and other cases of similar bearing.
Figure 2 has reference letters, A, B, C, D, E, F, G, for convenience in citing evidence of different kinds.
A, B. SOLAR CONSTANT AND SOLAR CONTRAST
Graphs A and B relate to the contrast of brightness along the radius of the solar disk, We were accustomed at Mount Wilson to allow a large solar image to drift centrally across the slit of the spectro- bolometer. In this way on each day of observation of the solar con- stant, we recorded the relative brightness of the sun’s disk along its E-W diameter for a number of wave-length regions of the spectrum. For each chosen wave length, results of many days in the year 1913 gave the average march of brightness from center to limb of the sun’s disk. The observed march on each individual day was then compared
NO. I AN IMPORTANT WEATHER ELEMENT—ABBOT 5
with this average. In figure 2 we see that on September 22, 1913, a day of high solar-constant value, the contrast of brightness, limb to center, was greater than normal, while on October 20, a day of low
3 2 Fee g42e2838ee88
eS ESS Ere ees ab Bre Ced CN Sh a a eee |
————
cer Oe @ PS Py
:
ear ea Piles eee scar = n sp Bet | pen hie aa a 6
ec lnce = aa Waa
I
ar.
20
Wilson 11905 -! Chile Statians 1916 +1924
eS tions
Mt h
Fic. 2.—Various evidences of solar variation and its effects.
‘or’ eee
> Mt. » Montezuma
een ee ew
PNG KES ;
Pea ES ec
Si
‘ola:
solar-constant value, the contrast was less than normal. The effect was far greater at wave length 3737 than at wave length 5955. This is as was to be expected, whether we regard the change of the solar constant as having been caused by change of the effective temperature
6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOE; LOX
of the sun’s radiating layers, or by a change in the absorptive qualities of the sun’s outer layers. Either change should affect shorter wave lengths most. This observation naturally leads us to graph C. But first let it be pointed out that from all wave lengths and all days of observation in the year 1913 there resulted a correlation coefficient of 0.601 + 0.067 between solar-constant and solar-contrast changes. ( See Annals, vol. 2, pp. 214-217, 1908; vol. 3, pp. 153-165, 1913; vol. 4, pp. 183, 184, 217-257, 1922; also Smithsonian Misc. Coll., vol. 78, No. 5, 1926.)
C. SPECTRAL DISTRIBUTION OF SOLAR-CONSTANT CHANGES
As foreshadowed by a comparison of graphs A and B, we are to expect that with increased solar-constant values the shorter wave lengths of the spectrum will show most increase. Graph C gives com- parisons of this kind culled from results of Mount Wilson and early Montezuma work. Although the day-to-day solar-constant changes that figure in the comparison average little more than 0.5 percent, and though accidental errors make the graphs rather rough, both stations yield the same general indication, namely, that solar-constant increase is almost wholly confined to the visible and ultraviolet spectrum, and may be 10 or more times as great in percentage for wave lengths less than 3500 as for the total radiation. An unpublished investigation recently made with the latest revised Montezuma results confirms this conclusion. (See also Annals, vol. 2, pp. 105, 106, 1908; vol. 3, pp.
121-133, 1913; vol. 4, pp. 204-207, 1922); vol. 5, <p: 29, 1932))
D. SUNSPOTS AND THE SOLAR CONSTANT
The graph correlates the sunspot numbers of Wolf and his suc- cessors with over 2,000 observations of the solar constant of radiation made at Mount Wilson and in Chile between the years 1905 and 1924. The tendency is toward the conclusion that higher sunspot numbers are associated with higher solar-constant values. But, as graph E clearly indicates, on March 23, 1920, the central passage of an enor- mous sunspot acted strongly to diminish the solar constant. (These ' contrary tendencies are discussed by A. Angstrom, Astrophys. Journ.,
vol. 55, pp. 24-29, 1922.) E. OTHER SOLAR-SURFACE PHENOMENA AND THE SOLAR CONSTANT
In this graph we compare the solar-constant values of successive days, March 1 to May 11, 1920, observed at Calama, Chile, with the
NO. I AN IMPORTANT WEATHER ELEMENT—ABBOT We
areas of calcium flocculi within 15° of the central meridian of the solar disk as measured at the observatory of Ebro in Spain. The dates given are for the solar-constant values. The dates of the calcium flocculi are displaced forward by one day in order to make best agree~ ment. (See also Smithsonian Misc. Coll., vol. 77, No. 5, p. 22, fig. 14, and p. 23, fig. 16, 1925; vol. 77, No. 6, pp. 42-54 and 59-63, 1925; also Proc. Nat. Acad. Sci., vol. 26, No. 6, pp. 406-411, 1940).
F. THE SOLAR-ROTATION PERIOD AND THE SOLAR CONSTANT
This graph shows the march of correlation coefficients computed between the zero day and the days 1-37 thereafter, for all solar- constant values observed at Mount Wilson in the year 1915. The solar-rotation period is very plainly associated with a range of correla- tion coefficients from —0.40 to +0.40, or 0.80 altogether. (See Smithsonian Misc. Coll., vol. 69, No. 6, 1918, where the same phe- nomenon appears in the years 1910 and 1916, though less strongly marked ; see also Smithsonian Misc. Coll., vol. 71, No. 3, p. 21, fig. 5,
and pp. 41, 42, 43, 1920.)
G. ATMOSPHERIC PRESSURE AND THE SOLAR CONSTANT
This graph shows for three meteorological stations that the baro- metric pressure follows opposite courses after high and low solar- constant values, respectively. At Winnipeg the greatest opposition occurs on the zero day, at Chicago on the second, and at New York on the third day after the solar-constant influence. It is confirmatory of the value of this evidence that the greatest oppositions of pressure effects are found attending the widest departures of the solar constant, and that the oppositions of pressure effects consistently diminish as the solar-constant departures become less. (See also Smithsonian Misc. Coll., vol. 89, No. 15, pp. 13-35, 1934; also Bull. Amer. Meteorol. mac. vol. 21, No, 6; p. 257, et. seq., 1940. See also figs. 1, 2, 11, 12, 16 of Smithsonian Misc. Coll., vol. 77, No. 5, 1925.)
A summary of many evidences of solar variation is given by H. H. Clayton (Smithsonian Misc. Coll., vol. 78, No. 4, pp. 50-62, 1926.)
A physicist to whom I showed figure 2 informed me that he con- sidered graph F, which connects solar-constant variation with the sun’s rotation period, to be the most convincing of all evidences of solar variation. Feeling that others may have the same view, I have obtained new evidence of this kind as given in figure 3. The curve F of figure 2 derives from Mount Wilson observations of 1915. In figure 3 I employ the revised solar-constant data, as yet unpublished.
8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
They are derived by combining results of daily observations at both Montezuma and Table Mountain, in the years 1929, 1930, 1931 as
ae of ae aaar thy ae od 7A Wey (eae ir 19
2) [tT acids of Sold pct | TT ug! te
Sle Sales vr 80 we ces i a AL th ae Pol EF ee Ae a is La aa a Ee SE a Se ea a 53 G2 = 8, i A A Di PE cai Bio7N NIA oo Fan A Pie 5 A De | A es ai Sho Abie Vale ee S 2 EI Ee Oe a Nee if
WS AS AO OD
a he
: ae nee
ool a
Fic. 3.—Solar rotation and solar variation. Note the rise of radiation at the 19th day, curves A-E.
revised by my colleagues, Messrs. Aldrich and Hoover, and soon to be published in Volume 6 of our Annals.
The data used to obtain figure 3 also yield by the method of correla- tion coefficients the fact that the solar constant has a 27-day periodicity
NO. I AN IMPORTANT WEATHER ELEMENT—ABBOT 9
of variation. The correlation coefficient between day 1 and day 27 from November 17, 1929, to October 5, 1930, is 45.5+7.5 percent. By plotting the marches of means, each of four consecutive periods of 27 days, as in figure 3, they give the amplitudes of variation. The maximum amplitude at this time was about 0.5 percent.
My friend, the physicist, asked me to interpret for him the physical explanation of this 27-day period in solar variation. I replied that I conceive there is an emanation from the sun, especially active in sunspot regions, and emitted approximately normally to the solar surface. This emanation produces either scattering or absorbing, or both, for solar rays. Its columns are shot outward from the sun for hundreds of millions of miles, and take roughly conical shapes, with vertices at the sun’s surface. As the sun rotates, these columns sweep through space, and when issuing from highly active regions of emission
central to the solar disk, as occurred about March 22, 1920, the columns _ reach the earth and the solar constant is diminished by several percent.
The atmospheric transmission at such times is also materially re- duced. The curves F and G of figure 3 show the position of the immense sunspots which prevailed on March 21, 1920, the depression of the solar constant on that day, and the simultaneous decrease of atmospheric transmission for homogeneous spectral rays at that time. The reader will, of course, understand that if the apparent decrease of atmospheric transmission were spurious, the effect of such an error would have been to raise, not ower, the solar-constant value computed.
The graphs in figure 3 should be compared with graph E of figure 2, that the reader may appreciate how frequently phenomena observed photographically near the center of the sun’s disk are attended almost simultaneously by depressions of the solar constant. The sun’s rotation brings these occurrences along every few days, and is the most prolific and readily understood cause of the fluctuation of the solar constant. The periodic long-range solar changes, which we shall describe further in what follows, are by no means so easy to understand. They seem to present features which are contradictory to expectation. These unexpected and contradictory phenomena will be discussed in the forthcoming Volume 6 of our Annals.
Solar rotation may also bring increases of solar radiation of 27-day periodicity, as shown beginning at the 18th day in curves A to E of figure 3. These increases occur (a) when the sun’s rotation brings over the limb a sunspot group previously invisible, or (b) when new sunspots form or increase in size on the visible solar disk. Mount Wilson Observatory photographs, kindly loaned me by Dr. W. S. Adams, show that one or more of these phenomena occurred after
IO SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
the 18th day of each of the rotation periods dealt with in curves A to E of figure 3. Shorter lived cases also occur. See days 9, 15, B.
PERIODS IN SOLAR VARIATION
In order to introduce other solar and meteorological correlations above referred to, I now remind the reader that I published several years ago my discovery of a number of regular periods in the observed solar variation.” Several of these periods have been confirmed to within the error of determination by T. E. Sterne,’ using other methods than mine.
I called attention, in my paper just cited, to the fact that the periods in solar variation, as I there determined them, had approximately a least common multiple of 23 years. This is approximately two sunspot periods of 11+ years. I mentioned also that the sunspot numbers themselves indicate a master cycle of two sunspot periods. I now give - further evidence thereon, derived from summations I have made of monthly sunspot numbers published recently by W. Brunner, of the Observatory of Zurich.*
Beginning with the sunspot minimum in the year 1810, the com- parative intensities of the 12 succeeding 11-year sunspot periods, as measured by the areas included under sunspot-number curves, are as follows. The 12th period being still incomplete, its area as given below is too small. The areas are given in arbitrary units.
Number of period..... i 3 &§ 7 9 Ir Total areas WATE AE aires sieve crave Sinvelscor ts 2354. 6501 5451. 3804 3730 (4018 25907 Number of period..... 2 4 6 8 10 12 ‘Total areas MASE Eatin a crstiatrieecelercest cele 3879 60908 6222 4651 4412 5068 31239
Each odd-numbered area is smaller than the next succeeding even- numbered area, and the totals representing the relative average intensi- ties of odd- and even-numbered sunspot periods from 1810 to 1940 are approximately as 5 to 6. This shows that for the last 130 years there has been a cycle in sunspot intensities comprising two 11-year periods. The average length of these periods since 1810 is 11.3 years.
The second correlation above referred to relates to G. EX. Hale’s observations of magnetism in sunspots. Hale found that sunspots prevailingly appear in pairs, of which the two members are of opposite magnetic polarity. During each 11-year sunspot period polarities per-
? Smithsonian Misc. Coll., vol. 94, No. 10, 1935. * Proc. Nat. Acad. Sci., vol. 25, No. 11, 1939, and vol. 26, No. 6, 1940. *See Terrestrial Magnetism and Atmospheric Electricity, September 10939.
NO. I AN IMPORTANT WEATHER ELEMENT—ABBOT II
sist unchanged with reference to advancing and following positions in solar rotation. But the polarities reverse at the beginning of each suc- cessive II-year period. Hence, the double sunspot period of approxi- mately 271 months is also the complete cycle in sunspot magnetism.
Various weather phenomena are also correlated with this complete sunspot cycle. Numerous meteorologists have noted a 23-year period in weather phenomena.’ I give in figure 4 an example of 23-year periodicity in tree-ring widths. These curves I have computed by combining results from five localities in Southern California and Ne- vada including about 40 trees in all. The results are from data pub- lished by A. E. Douglass. I reproduce here also in figure 5 my illustration given as figure 33 in my paper “Solar Radiation and Weather Studies,” already cited.
According to my former Smithsonian researches, the well-verified solar cycle of about 23 years which is accompanied by terrestrial effects, is the approximate least common multiple of 12 regular periodic varia- tions of solar radiation.
In his first paper on these periodicities, Dr. Sterne found high probabilities in favor of my periods of 92, 114, 21, 25, 394, 46, and 68 months. The data were insufficient to enable him to test periods of g2 and 276 months. He found little evidence favorable to my periods of 7, 8, or 34 months. I am disposed to agree with him as to the elimination of periods of 7 and 34 months, but shall submit further evidence below favorable to a period of about 8 months.
However, the interval of 20 years, during which high-grade solar- constant observations have been carried on, is too short to fix accu- rately the lengths of the solar periods, or to indicate whether they continue indefinitely without shifting of phases, or changing of ampli- tudes. I desired to use long meteorological records to throw light on the first two of these interesting questions, but was at first balked, as meteorologists have frequently been, by changes of phase in the ter- restrial periodicities which are supposedly associated with those in solar radiation.
CAUSE OF SOME PHASE CHANGES IN TERRESTRIAL PERIODS
It occurred to me that if, as is indicated by the 20-year analysis of solar variation which I shall present below, the solar periods follow on without changes of phase, then such a period as 8 months, if it has a maximum in January 1900, must have others in September 1900
° See Wild, Die Temperatur Verhaltnisse des Russische Reiches, p. 279, 1881; also Quart. Journ. Roy. Meteorol. Soc., vol. 62, p. 481, 1036; also Douglass, Climatic cycles and tree growth, vol. 2, pp. 131, 132, 1928.
I2
YEARS
SMITHSONIAN MISCELLANEOUS COLLECTIONS
23-YEAR PERIOD IN TREE RINGS FROM 5 GROUPS, SO. CALIFORNIA 15
5 10 20
Fic. 4.
VOL.
IOI
13
IMPORTANT WEATHER ELEMENT—ABBOT
AN
I
NO.
*sia}}0] Aq pazyeoipur soinjeay Surpuodsaii0oy ‘[I[J ‘eoag ye uolepdioaid ul ajoAd reak-€z ay [.—S ‘ory
a a eS Pe eS es (are Sse ew EN intl ba eee oe
|
es fan pe elle AE API ce
86st
szgl
ool Ny LZ A A at Paes Ee SAW cA a ee Ba EA Ae oe Ee =e es a ee ee es | ee ee eee es ee ee ee ee ee Se
WWHON LN39DUd3d
14 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
and in April 1901. It might very well be that the changed conditions, as regards snow covering and atmospheric circulation, in these months of three different seasons of the year, would tend to produce changes of phase in the terrestrial response of the weather elements to the regularly recurring solar cause.
I tested this hypothesis for several periods and for a variety of weather stations as indicated in figure 6. In all the graphs of figure 6 the abscissae are the months of the years in which the solar cause re- curs, and the ordinates are the months of the period in which the terres- trial response is observed. The latter data were obtained in each case by plotting the successive months of departures from normal weather since 1920 in separate graphs, each of the length to include only one recurrence of the period in question. In each such plot the maximum point was located, and serves to establish one point among the ordinates in figure 6.
Of course the influences of any terrestrial modifying causes, and the influences of other solar periods, would often tend to displace maxima in these working graphs. Hence it is not surprising that the points plotted in figure 6 do not fall exactly upon the straight lines drawn. But it is plain that these straight lines tend strongly to repre- sent the mean indications of the groups of points. I conclude, there- fore, that for various solar periods, and for many terrestrial stations, changes of phase in the weather periods, associated with the periodic solar variations, are due to seasonal terrestrial influences, and not to’ changes of solar phases.
This conclusion reached, phase changes may be eliminated from computations. Consider, for instance, the period of 21 months. What- ever solar phase for this period was operative in January 1900, that same solar phase will have been operative in January of each 7th year, counting from 1900, both before and after, provided successive oc- currences of the solar periods do not themselves shift in their phase relations. Similar considerations enable us to compute the years when recurrences of each of the other subordinate solar periods had approxi- mately the same relations to the months of the year. These computa- tions having been made, we were able to select all of the approximately comparable recurrences of each of the solar periods, from the year 1800 to the year 1940.
15
AN IMPORTANT WEATHER ELEMENT—ABBOT
I
NO.
“INDO Sasned JelOS WoYM Ieok oy} JO uoseosS Aq peuroA0s $I [eII}Sotto} JO Ssoseyud— 9) ‘OI
UwWFA SO SHLNOW Be GS Se OR MIA A IT 2€ zx wr Mm a ZT Lf FOE OLE FIN A pg 4 ie 246 Ta A IT E
Ea
1 Owe HIIMNITSO
; : ae
0 0 ioeemesl {~ ° o
On 26 HOIMNIZYD
COlYFd SO SHLNOW
16 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
METEOROLOGICAL EVIDENCE ON THE REALITY OF PERIODIC SOLAR VARTATIONS, AND) ON THEIR EXACL LENGE
In order to illustrate and develop the evidence, let us make four assumptions, subject to verification.
1. There is a periodic solar variation of about 8 months.
2. Its phases are unchangeable.
3. It produces appreciable changes in temperature and precipitation.
4. Its effects on weather are of changeable phase, depending on the relation of the unchangeable solar phases to the season of the year.
If the period is about 8 months, it will recur at the same season of the year after about 24 months. Hence, we shall compare together weather intervals, each 8 months long, which recur at intervals of 2 years. Proceeding in this way, let us employ the departures from normal temperature at Copenhagen, Denmark, beginning with 1800 and smoothed by 5-month consecutive means. These departures have been derived with reference to mean values computed for the interval 1800 to 1932. Thus is formed the following table in which maxima are indicated by bold-faced type, minima by italics.
To test the preferred period we adopt January 1900 as a base date, and working both ways from it, we compute a series of dates 8% months apart, extending backward to 1798 and forward to 1937. These dates fall in every month of the year. To study them with regard to the seasons and to the passage of the century, the departures for 8-month intervals, beginning with each of these dates, have been arranged in 24 groups. January and February dates taken together are segregated into four groups, covering the years 1800 to 1833, 1834 to 1867, 1868 to 1901, and 1902 to 1935. A similar grouping has been made for each succeeding pair of the months of the year.
The following table gives the data for November and December by individual group means, as well as by the grand mean, and gives the grand means for all other pairs of months. It is seen that there is no progressive side shift of the individual means for November and December. The same is true for the other pairs of months. There is, however, a marked shifting of terrestrial phases governed by the different seasons at which identical solar phases recur. This is indi- cated in the table by positions of the bold-face and italic numbers corresponding, respectively, to maxima and minima.
The numerical results just set forth appear to confirm the reality of all four of the assumptions made above. The several grand mean values indicate 8!-month periodicities with ranges of from 0°47 to 1°44 Centigrade throughout an interval of 139 years. The range depends on the time of the year at which identical solar phases recur.
17
IMPORTANT WEATHER ELEMENT—ABBOT
AN
NO.
‘unuliulm 94} Ag ‘y}UOUI — =
Ayjoexa youurd 9uo s}jIyS 2S1e, yons YUMA
1
tz
of—
gi
6F+
or
zs—
‘syjuow fg pottod 9y} Jajatd 0} pue ‘yyUOUT
SI ZIXOOI, I gXxI1
6G Rae 90) SECT. = gor ZI I— I cz oS fe le— 88 tr SII Wi" (eh iy GI =ig eee SOT 9 — zl— ¥S— 6€— goI— to—
suvsy jo 21qR
wnuiixem 34} Ag
‘steah OIL Ul
vl
tcl
Zel
oS
:a}BoIpul JoABMOY S}JIYS Soy]
ofol
0081
6h
Qe
II
9 GI
“sny
§ ig
UOT}OIIIOD 9Y} UO XY O} SN pos] Set} JIYIAN | “yqyuouw
cI—
Li
FI
cf
OI
1
Aint
el—
ci—
II
ci—
Ei—
aunf
SV ODTGOD) OOl/ ly eSUeOlLus JON ea) Ob Toop NOLL Mo) Pp
poisagd uquou-9 fo jsat, “payjoowms ‘sainjavgap aanqosaguiay uabpyuagdoy— 1
Ajuo ysnsny 0} Arenuef jo sanjeA
rs o£ OOL O ZI yn S == ig Of kg OS AP == of— + oz Vt eh Kg ONS 6 {Ei (it 8 i 17 — II Qi Se= Ol zz 6z EE Ly— Ae cady "ae :yup aay |
vo— I
cI
‘WP
6 oe es
ZI eee ewes
z1XOll gx 61
‘[e11} 9uo0 YIM portied 9y} 0} UOTZDaI109 s9do1d 9Y4} sUTUIIO}ap
WYySI1 9y} 0} syjJuoUI 61 uNUTUI oy} pue syjUOU I: 4y1YyS 0} Sivadde umuwixem oy Z
uedyq
Or
IOI
VOL.
SMITHSONIAN MISCELLANEOUS COLLECTIONS
18
*1oqjoue auo 0} yoodso1 yYM soseyd jo jusWeoR[dsIp Moys Op 1eaA ay} JO SuOseasS JUdIDYIp }e Sutuurseq sdnois yng “Joquisdaq—1equUIsAONY JOJ a[qGQr} Papua}xe 94} Aq UMOYS SI ‘iva 94} JO UOseas jURJsUOD & je Suluulseq Sdno1s jo suvow jo saseyd ay} Jo yuatooeldsip Ie[noes oarIssa1s01d ou o19y SI a194} JEU,
0 zy ov gI G 9 cs ze WG 8229 yee 12S LE61 ‘sny 6z— SI— ¢ IP S60) Oz= = ee. g/d on ic— 8 gl Gl = 72 RS Kg — BOO OOK. HYG 61 aah LE61 ounf} ., || o— S— b— 62 pf Gg of gI—''**'4 9 Gp ieee ai—s> v Bee = See Sapgaae ere pue 3, | S061 ACWW £061 TOV |e |e. —"—ee— Lhe SE Cpe ee Sg zoe or SI 62 — 08 —— 00 — OL LO OC OG na pue -S QogI N pl AN QOgI Ger 60 — S10 — 2S = Ol a a ee ey 9 = it == (45 ve 2 Cee ee es pue efgr uel f£g1 BG 9— gI t= =p 2 Ve = VE Glee ee 0} 6z— SI— € Iv S9 62 Ze (eee nL
“AON ssuluuiseq Jaqui909(7—19quIaAON SIvI_ ynoysno1sy} *D sdIsep OOI/I : UA A[UO UdAIS sueaTy ‘“poAojduia syyuoU [Te Jo sanjeA
poisag yyuou-29 fo jsa fT, “payjoows ‘sasnjavdap aanyosaduiay uaboyuagoj—z ATav J,
NO. I AN IMPORTANT WEATHER ELEMENT—ABBOT 19
This difference in amplitude is due not so much, I think, to a difference in effectiveness per se of the solar cause, as to a fitfulness of the terrestrial dealings with it. From this latter cause individual periods which are combined together in the seasonal means clash together in phase, and so reduce the ranges of the mean.
WEATHER RESPONSES TO THE SUBORDINATE SOLAR PERIODS
We have made use of monthly temperature records * from Copen- hagen, Vienna, and New Haven, beginning with the year 1800, to investigate the weather responses to 7 of the 10 periodic solar varia- tions. We have first computed the years when the temperature responses to the assumed solar periods must be in approximately cor- responding phases, provided that the solar periods recur without displacements of phase and that they have the lengths we had attrib- uted to them.
Before particularly discussing these investigations, I shall make a few remarks on the nature of the periods I have in mind. Many investigators show strong preference for the use of sine and cosine forms in studying periodicities. While it was shown long ago that any curve, however irregular, may be represented by a summation of a series of sine and cosine terms, this is an entirely forced and arbitrary mathematical device. It is so elaborate, in the case of re- curring periodic curves of irregular form, that if adhered to with the use of series having sufficient terms to give a true representation, one could not, with ordinary means, find time for such a research as I contemplate here.
It has seemed to me not only simpler, but truer to the actual facts, to test the validity of periodicities, and to express the forms of them, by tabulations of actual ordinates and abscissae, determined in records of the phenomena. One uses enough repetitions of the proposed period in each table to yield a fairly representative mean form, if indeed there results a plausible form, indicative of a real period. Then by comparing the mean forms determined by successive tabulations, one sees whether the proposed period persists throughout the entire interval investigated. By this means one learns whether to reject the period as nonexistent. But one also often detects lateral shiftings in the same sense, among the curves determined by successive tabulations. These shiftings enable a correction to the length of the assumed period to be computed. Then a new tabulation may be made with the improved
*World Weather Records, Smithsonian Institution, 1927 and 1934.
20 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
period, and, if necessary, a further correction to its length may be made and subjected to a new test by tabulation.
When a very long interval is under consideration, in the examina- tion of relatively short periods, there will be so many repetitions that very exact lengths of the periods must be found, or else in the mean of all recurrences the amplitudes of the periods will disappear. As a corollary it will be plain that of two periods slightly different in length, that is the better for which the average amplitude over a very long interval is the greater.
When using odd lengths of periods, such as 9.79 or 11.29 months, we have first to prepare a table showing the exact day of ending of each recurrence. Thus in our tabulation we must occasionally repeat a month, or omit a month, so as to keep always within 5 month of the true times when the periodicity recurred.
TABLE 3.—Average march of two assumed periodicities, Copenhagen and New Haven, 1800-1932
Assumed period Station Months I II Ill IV Vv VI Vil VIII Range
Copen- hagen..8.00 0°10 0°41 —o°14 —o°12 —o?08 —o0°25 —o°02 —0°33 0°74C. 8.12 —o°14 0°29 0°55 0°79 0°33 —0°33 —0°50 —o?22 1°29C. New Haven. .8.00 0°18 0°20 —o0°13 —o0°04 0°25 0°08 —o°02 0°04 o°38F. 8.12. | 0°69 0°45. 0°52 0246 0°17 —o%08 —o°62 0713 I9gnEe
To illustrate the necessity of accuracy in the lengths of periods when treating of periodicities in so long an interval as 132 years, I give the results of two sets of analyses. In the first the period of 8.00 months is assumed to exist in the departures from normal tempera- tures computed for the interval 1800 to 1932 from records of Copen- hagen and New Haven. In the second the period is taken as 8.12 months. To avoid seasonal influences, the comparison includes only periods when the solar pulse was in identical phases and running from January to August. Equal numbers of such periods were used in computing each of the results shown in table 3.
It will be seen that in each of the analyses, based on an assumed period of 8.12 months, there results a fairly regularly defined periodic curve of considerable amplitude. The other analyses based on 8.00 months are much less satisfactory. I conceive that the excellent show- ing of the period of 8.12 months for each of two far-separated stations, and throughout an interval of 132 years, is satisfactory proof of the validity of this period in weather.
NO. I AN IMPORTANT WEATHER ELEMENT—ABBOT 21
To further emphasize and explain the preceding discussion :
With an 8-month period there are 66 even-numbered years from 1800 to 1932 when the terrestrial response should be in the same phase in January. Six tables may be prepared from monthly means of tem- perature departures at Copenhagen, each covering IT recurrences. In each table there will be in each line departures for the 8 successive months, January to August, making 8 columns, each column 11 lines long. Taking the mean values of the 8 columns of departures, we obtain the average march of the 8-month period in Copenhagen tem- perature departures for an interval of 22 years. On comparing the
Rear cba of dy | | | PENS, he
CoPENHAGEN VIENNA NEW Haven
Fic. 7—Greater amplitudes of periodic curves when precise period is found.
6 mean curves thus obtained we found a slight displacement of maxima and minima from epoch to epoch indicating that not 8.00 but approxi- mately 8 months is the true period. We have repeated the computa- tion at Copenhagen, at Vienna, and at New Haven with the period 8! months and employing only such epochs (37 in number) as bring corresponding solar phase dates within less than +1 month of Jan- uary 1. Thus we find the curves given in figure 7 representing the interval 1800 to 1932. The period 8$ months suits all three stations. But compare the ranges of these 3 curves of 84 months period with the ranges of the other 3 curves of 8.00 months period also given in figure 7. Evidently this slight difference in length of period is highly important for so long an interval as 132 years.
I shall postpone a more detailed report of this investigation which concerns the determination of the exact lengths of the periods. It will
22 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
probably be published in Volume 6 of the Annals of the Smithsonian Astrophysical Observatory, now in preparation. However, the com- putations from the three stations, Copenhagen, Vienna, and New Haven, agree very well as to the best lengths of the 10 solar periods.
.
CONSTANCY OF PHASE IN SOLAR PERIODIC VARIATION
Having observed the solar constant daily for only 20 years, it would be uncertain, if it were not for the meteorological evidence just described, whether the solar periodic variations continued for long intervals, such as a century, without displacements of phase. But if there were frequent displacements of phases in the solar periods then the numerous repetitions of the responding periods in the weather during 132 years would bring out the discrepancy. They would not in that case yield periodic mean weather curves of considerable ampli- tudes, such as would, if synthesized, reproduce to a considerable extent
TABLE 4.—Possible range of solar temperature effects
Period, months
Station Sir2) ) or70) § h-2o)n 2r-0oN | 255 3904 454 Sums ee Amplitude in degrees a a Copenhagen ....... C. 41229)-0°79' 1°40 4047. 1200) 1276 0863) 6274s AeA alot ontecre. Cy ears “1820 1387) Zeon ocGomte2s WOO mNO 4b ee News Hlavenweren ac. Ey) 1330) 1200) 3230 10200), 1270) 0858) 10200) 1O.Souue
the range of actual weather fluctuations. So many repetitions, if the periodic causes had different phases, would tend to flatten the general mean curves toward zero amplitude.
Accordingly it is of much interest to assemble the mean periodic weather curves of Copenhagen, Vienna, and New Haven representing the interval of 132 years and see how wide ranges they could produce from extreme crest to extreme trough if made to recur till all came simultaneously to maxima at one epoch, and to minima at another. This assembly is given in table 4. It shows, when we consider only 7 of the 10 known periods, that they are of sufficient combined range to account for the full ranges of smoothed monthly mean temperature departures from the normal at Copenhagen, Vienna, and New Haven. I suppose that the observed ranges of temperature departures from the normal at these stations are actually less than might result from solar variation alone for the following reason. Since there are terres- trial phase changes as explained above, and since the periods are not exactly commensurable, the 10 periodic solar causes, in fact, are never in complete harmony of phase as to maxima and minima of their
NO. I AN IMPORTANT WEATHER ELEMENT—ABBOT 23
temperature effects. Weather changes produced by solar variation, in other words, can never exactly repeat, because of terrestrial phase changes, and of the noncommensurability of the solar periods.
ANALYSIS AND SYNTHESIS OF SOLAR VARIATION
These meteorological researches having, as I conceive, fixed the lengths of the solar periods with considerable accuracy, I now present the analysis of the curve of solar variation for the years 1920 to 1930. It is derived from Montezuma observations alone after the year 1923 and from combined results of Montezuma and Harqua Hala from 1920 to 1923. I used these data from our best station of the years 1923 to 1939 for this purpose before the completion of revisions of data from other stations. As shown in figure 1, the agreement of the monthly means from different stations is so good that I have no reason to think the results would be much changed had the final data com- bined from all stations been available.
In figure 8, curve A comprises these monthly mean observations of the solar constant for 20 years. The analysis of the curve was made as described in my paper “Solar Radiation and Weather Studies,” * and in my paper “The Variations of the Solar Constant and Their Relations to Weather.”’* Seven of the periods used were those given above, and, in addition 68, 91, and 276 months. The results of the analyses are printed in the legend for figure 8. In making these analyses, we verified the constancy of the phases by dividing the 20 years of data into several parts for the shorter periods.
Curve B is computed by synthesizing the 10 periodic curves indi- cated by numbers below it. It is obviously very like curve A. The average monthly deviation for 240 months between curves A and B is 0.0026 calorie, or 0.13 percent.
AMPLITUDES OF THE SOLAR PERIODS
Having satisfied ourselves, as explained above, that there has been no notable change in the orderly succession of phases in the subordinate solar periods for 140 years, it is of interest to know if the amplitudes of these periods also remain unchanged. Here we have to rely on the solar observations of the past 20 years alone. So far as we see, the meteorological records cannot help to solve this question.
It would be expected, by analogy with the sunspot numbers, that these amplitudes will change. Our meager data of only 20 years dura-
“Smithsonian Misc. Coll., vol. 94, No. 10, 1935. * Quart. Journ. Roy. Meteorol. Soc., vol. 65, No. 280, 1930.
IOI
VOL.
SMITHSONIAN MISCELLANEOUS COLLECTIONS
24
"gI— oz— ‘vi— ‘ZL ‘6z ‘hE ‘6E ‘Ib ‘OE ‘EE ‘ot ‘g— ‘h— ‘£1 ‘Oz ‘be ‘z— ‘gi— ‘S— ‘E— ‘gf— ‘ob— ‘he— :1¥eak-Ez AG fo cc ‘c= — ‘o— ‘0 Or ‘z Of) ‘h Oy ‘bp “p Lf td ts! Lis) EC *¢ 4 ‘o ce — ‘2 .— OE Ge eC) a A ee Oe CN A ee OC NP Th te ‘g— ‘g— ‘g— ‘g— ‘L— ‘g— ‘g— ‘g— ‘S— ‘S— ‘o ‘or ‘br ‘Zr ‘gr ‘gr ‘gr ‘or ‘g *z1— ‘g— ‘S— ‘oS ‘4 ‘or “by “Si ‘Sr “Zr “gt “oz ‘ee “be “ye ‘zz ‘gt ‘or +00°16 ai? rake tah oye) 54, Giese Cre Ge Uae Oe Asie Ohne Cie Girne Gian Uipie Cepin Gini ys O78 OS eg Gre Oo IS A eS eh Ue a ‘LS— ‘pS— ‘€S— ‘6b— ‘Eb— ‘PE— ‘of— ‘Ez— ‘61 "SI p=, ri Yo Cee Gea sie ee More Ce U4 oy O74: Mey Cigid Gopi. Wier Cureton Mone Moje Stara Ue Ute Sere 8 9 8 2 ee) 9 89 ho) YE Vane ae epee Oe Oey Tsay Otsisn “Sr ‘hr ‘hr ‘hr ‘Ex ‘zr ‘11 ‘11 ‘or ‘6 ‘g ‘LZ ‘g ‘SG ‘E ‘0 fo fo ‘z— ‘S— ‘g— ‘o1— ‘S1— ‘gi— ‘ez— ‘Ez— “bz— ‘Se— ‘Sz— ‘Sze— ‘bz— ‘oz— ‘S1— ‘o1— :¥SP “bh— ‘Sb— ‘ob— *SE— ‘Lz-— ‘41 — ‘S— ‘tr ‘1p ‘1G ‘19 ‘ES ‘Sh ‘1b SLE ‘SE “HE ‘ih ‘Ib ‘zh ‘Of ‘zz SE ‘1 ‘I— ‘e— ‘e— ‘E— ‘E— ‘h— “hp— ‘L— ‘11— “ES— ‘gq— ‘gq— ‘ZS— ‘Sh— ‘ob— :6E "gt ‘zr ‘b ‘g ‘Sz— ‘b— ‘S1— ‘g— ‘Sr ‘11 ‘oz ‘€— ‘o1— ‘z1— ‘gz— ‘zz— ‘Sz— ‘11— ‘ze ‘1— “gi ‘of *(£€z) ‘gi ‘11 ‘o Sz "g— ‘ze ‘hz ‘g— ‘gi ‘Ez ‘6 ‘g— ‘SG ‘S ‘6— ‘Ex ‘g ‘bh ‘gr1— ‘11— ‘g— ‘L— “S1— 00°12 *9— ‘o1— ‘61 ‘11 ‘Zr ‘61 6g “bi— ‘gi— ‘zI— ‘11— 62°11 py ‘LZ ‘€— €— “ SL— ‘11— ‘€— ‘E ‘LZ 36276 *6— ‘€ ‘gI Csi day ees ‘g— as ‘9— 789 syjuour
000'01 ‘gc61 Arenuef ulsaq Mojeq saseyd ¢ at10]e9 ae Qu, ‘oz61 Arenuef sutuursaq
‘soIqIotporiod O1 SUIMOTIO} dy} JO stsayyUAs “gq BAIND ‘ aTIYD “euINZoJUOTY ‘san[eA JuL}sUOD-Ie[OS UROL AjJyyuow ‘vy sAIND— s “DIY
noe Te peeled Ns ae
A Aerts De sitet ees ta este Sem
er SAY oe
CC ALA
in
FS a al
pam ede athe)
J See
NO. I AN IMPORTANT WEATHER ELEMENT—ABBOT 25
tion cannot give indications as to the longer periods. But we have explored the 8.12-, 9.89-, and 11.29-month periods as to their ampli- tudes since 1920. For this purpose we prepared tables of Io lines for the 8.12-month period, of 8 lines for the 9.79-month period, and of 7 lines for the 11.29-month period. Taking the mean values for these several arrangements, we can compare average amplitudes at three epochs during the 20 years of observation. Of course, the mean values for tables of so few lines are not very reliable, considering that influ- ences of accidental error and of interference by others of the 10 periodicities are simultaneously operative. But for such worth as they may have the comparative results are as follows:
Relative amplitudes of solar periodicities in three epochs
Periodicity, months
Epoch 8.12 9-79 11.29 ani LOZO=Iune mm 1OZOseyes eerie: 49 83 105 alyiero26—D ec TOs 2"-nee seston 26 39 30 Jan!) 1933—June T0380... os6n6 600600 36 38 54
All three periodicities seem to be stronger in the first epoch, 1920- 1926, than later. But the data are not sufficient to be decisive.
SYNTHETIC WEATHER PREDICTIONS
It appears that 10 periodic variations of the sun have persisted in regular phase relations for 20 years, and probably for 140 years. They are each apparently associated with periodic departures from normal temperatures, and also (although to save space we have not illustrated it) with periodic departures from normal precipitation. For the periods of shorter lengths these departures from normal weather conditions exhibit phase changes. But these phase-changes in weather seem to depend on the season of the year when the solar cause operates, and can be allowed for on that basis. The solar periodicities may not be of uniform amplitude as they recur, but this matter is uncertain as yet.
These facts led me to consider if weather might be predicted by a synthesis of the average effects of the periodic solar causes, just as the solar variation is itself predicted for the last years shown in figure 8.
With the assistance of Miss N. M. McCandlish in the computing, we have tried 5-year predictions for various stations, both for pre- cipitation and temperature. Using only the recorded weather data up to December 1934 as a basis for the forecasts, we have made synthetic predictions for the later years, 1935 to 1939. These predictions were then compared with the observed weather records of 1935 to 1939,
26 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
but these latter records had no part in influencing the computations of synthetic weather indications.
In illustration of the method I give in table 5 the components of a synthetization of 1 year in the precipitation of Peoria, Ill. In figure 9 I give predicted and observed weather for 5-year periods in several stations. As these long periodicities cannot have to do with the vicissi- tudes occurring within individual months, I use 5-month running means of departures in weather data in these illustrations.
Considerable similarity appears between predictions and events. On the average about two-thirds of the months of a 5-year prediction seem to show fairly good correspondence. Some stations turn out
TABLE 5.—Synthesis of precipitation, Peoria, Ill., for the year 1938*
Percentages of normal. 5-month running means
Period, Forecast Ob-
months 8 92> 11}? 21 25 394 46 68 (Product)* served Janse aeaerts 97 82 100 OS l22) 123 e101 95 112 134 Heb. apo 98 86 100 O30) 2r) 120 97. ~=«+100 110 147 Maia Mame 98 go Io! O32 eTOSH) H122 7 l22 96 128 156 NBS tctie 98) 495) ' 09. 103, OF © 130] |..1377 87s 14g is IMD, Loo ode TOO}! 1T03 "IO PTI Se i ey 84 138 156 June see LOZ) 103) e1LGON IT 76m IZ Silay 80 140 143 Nitalyameertocte 103. +IOI Ove ean Sh arr6) 188 88 128 131 ENE. no dai 103 05 93 ~=2+108 88 106 105 04 90 109 Septiiic.e' 102 99 63) LOL 07 77 07 98 67 86 Octsprvccuce LOL LOS 96 92. I01 69 89 ~=104 59 66 INOW Sonct 97. =—-103 97 95 104 7, 102) « sf02 79 85 Deer tics 98 106 98 93 Os) loz 99 96 87 103
* The periods used differ slightly from those now preferred. b Where dashes are inserted in these columns, changed forms of periods begin, according with the control of periods by season of the year, above explained. ¢ This forecast would be improved if it lay 10 percent higher. Its range of 84 percent compares with the observed range of 90 percent favorably, and its phases are quite correct.
better than others. For another somewhat better synthetic prediction of the precipitation at Peoria, Ill., not here illustrated, I have worked out the correlation coefficient between prediction and event over a 60-month interval. It results 70+5 percent.
What seems to me particularly significant in these comparisons is this: The range of departures forecasted is nearly as great as the range observed. If the solar variations were negligible in their effects, or even nonexistent, then the average weather effects corresponding to them over a period of 20 to 50 years, which are the basic elements in our forecast, would also tend to be negligible. Moreover, if they had no common thread of causation connecting the several solar periods with the weather, then the synthesis of these supposed spurious or negligible average weather periodicities would tend to be zero. On
INO. LE AN IMPORTANT WEATHER ELEMENT—ABBOT 27),
the contrary we find their synthetization gives values of weather departures which show nearly the same ranges and phases as the actual weather records. To me this indicates that the variation of the sun is actually a principal cause of weather changes.
This method of forecasting by synthesis of periodic effects, like that about to be mentioned, while hopefully successful in many cases, shows bad timing or dissimilarity with events in others which destroys its
131
31235. 1936 1937 1938 1939 ine | oa iL JeL Ee. |e
APT ae eH Ae a SSRN eaenaes ee Ae INS PAM Lal Et VS IT NAT a | ECHR a
ie isi SECC ECCEC ACE Ere
Fic. 9.—Predictions of weather aa on the previous effects of 10 periodic changes in solar radiation.
usefulness. There appears to be some other variable, as yet obscure, which must be discovered and applied before valuable forecasts can be made consistently hy this method.
WEATHER FORECASTS. BASED ON 23-YEAR CYCLES
With further regard to long-range forecasts depending on solar variation, as | remarked above, the solar periodicities approach rather closely to a least common multiple in 23 years. But this relationship is not exact. Moreover, the seasonal phase change of terrestrial response to corresponding solar causes of weather effects complicates
28 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
the matter. Hence the phases and amplitudes of weather features should not exactly recur at intervals of 23 years, even if there were no other terrestrial complications. Yet the tendency to a repetition of weather features each 23 years, such as is shown in figure 5, itself affords a method of long-range forecasting which gives some promise of usefulness. It may be compared in its results with the synthetic method explained above.
In applying this method, one must first determine with what dis- placements in phase the weather features happening recently occur with respect to those of 23, 46, and 69 years ago. Also one must be guided as regards amplitudes by those former appearances. One pieces his prediction onto the curve of departures as it stood at the time the forecast began, and thus accepts the prevailing trend, whether positive or negative, as a point of departure.
I now give in figure Io several examples of forecasts and events with which I have tested this method in the past 6 years. Here, as in figure Q, the data used are 5-month running mean values of departures from the normal.
The forecasts shown in figure 10 were all made previous to the events. While records of 23 or 46 years preceding are shown in the figure by way of illustration, other preceding data were also considered in making the predictions.
It will be seen that the forecast for 1934, 1935, and 1936 of pre- cipitation at Burlington, Vt., is very good. Other forecasts have lesser accordance with events. As a rule, precipitation seems more amenable to this kind of forecasting than temperature. This is also the case with synthetic forecasts previously explained. However, in almost all cases of failure, the trouble seems to be, not that coming weather features were unperceived, but that differences of timing of as much as 3, or even 4 months occur, as between the prediction and the event. This of course destroys the usefulness of the method.
It appears that other factors, still obscure, must be discovered, whereby these errors of timing may be anticipated and corrected, before this method of forecasting can be very useful. Nevertheless there seems to me to be so much evidence thai the periodic solar varia- tions, and their cycles of approximately 23 years and multiples thereof, are governing influences over weather, that I cannot but hope that experts in meteorology will take up this clue, and go forward to greater success, by combining with it their own knowledge of atmos- pheric processes.
Much more might be said of these long-range solar periods, and their application to weather forecasts. Indeed, without additional de-
NO. I AN IMPORTANT WEATHER ELEMENT—ABBOT 29
tails, the statements made here may be thought by some readers to rest too much on the good faith of the narrator. But this paper has already exceeded its intended limits, and as I wish still to bring up the subject
TEMPERATURE ——PREDICTED
5-MQNTH RUNNING MEANS.
OBSERVED ---
eke (| A
Sy,
Fic. 10.—Predictions of weather based on the 23-year cycle.
~-PRECIPITA TION-——PREDICTED
5-MONTH RUNNING NEANS
O®SERVED--
ae ces = 2 fs a SF is aS pe Pita Te Pe
of short-time an fi and their weather effects, I will say no more of very long-range weather prediction.
120,54
SHORT-INTERVAL SOLAR CHANGES
The sun is not equally bright at all points on its surface. This is visually obvious with regard to sunspots and faculae. My colleagues,
30 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. LOE
Fowle and Arctowski, as well as H. H. Clayton, have demonstrated correlation between solar appearances and variation of observed solar radiation.’ As the sun rotates upon its axis in approximately 27 days, a succession of unequally radiant solar surfaces is presented toward the earth. This leads to irregular variations of the solar constant of short intervals. There is a tendency, also, to create families of solar changes of about 27 days’ period. However, the changes in the sun’s surface are so rapid that generally new distributions of brightness take place before more than one or two solar rotations are completed. The local variations of solar brightness usually produce very small per- centage changes in the sun’s total radiation. On these accounts it is only rarely that the 27-day period can be distinctly shown by our still unsatisfactorily accurate solar-constant values. However, some good cases are on record.”
In two papers I have discussed the effects of short-interval solar variation on the temperatures of Washington, St. Louis, and Helena in the United States, and Potsdam in Germany.” Now that longer series of daily solar-constant values are available, I have recently repeated in part the computations described in these papers, using 20 years instead of 12 years of observations. The results of these recent studies do not alter the conclusions that opposite trends of tempera- ture follow, respectively, rising and falling sequences of solar change, and that these temperature changes are of major significance. I have treated the subject as well as I can in my paper in reply to Paranjpe and Brunt,” and will now quote therefrom:
Referring to figure 1 of my just cited paper, here reproduced,” the reader will find there two curves for each month of the year showing departures from normal temperature at Washington, D. C. In each month the curves show a well- marked opposition like the right and left hands. The separations of the curves in the months January, February, March, April, May, June, August, October, November, and December range from 14° to 24° F., and evidently constitute major departures from normal temperatures. Similarly, results showing in almost all cases opposition like the right and left hands, but differing widely in actual march of the pairs of curves, are shown for St. Louis, Helena, and Potsdam in other illustrations in the cited publications. The curves are com- puted for all these cities starting from identical dates, 320 in number, scattered over I2 years. Some 10 to 20 cases are combined in each curve shown. The
® See Smithsonian Misc. Coll., vol. 77, No. 5, pp. 21-23, 1925; also Proc. Nat. Acad. Sci., vol. 26, No. 6, pp. 406-411, 1940.
* See, for instance, Smithsonian Misc. Coll., vol. 69, No. 6, pp. 7-8, Sec. (3), and the curve for 1915, fig. 1, 1918. See also my paper in Science, April 11, 1941.
*% Smithsonian Misc. Coll., vol. 95, No. 12 and No. 15, 1936.
* Quart. Journ. Roy. Meteorol. Soc., vol. 65, No. 280, 1939.
* Here again reproduced as figure 11 of this paper.
INO: I AN IMPORTANT WEATHER ELEMENT—ABBOT 31
data of temperature departures in each case cover 16 days following the starting date selected.
How were these 320 dates selected? They are chosen as dates when solar variations commenced. As shown in figure 1 and table 1 of “The Dependence of Terrestrial Temperatures on the Variations of the Sun’s Radiation,” they
MEAN RESULTS 1924-1935 TEMPERATURE CHANGES FOLLOWING RISING SOLAR RADIATION SHOW LL CURVES, FALLING BY DOTTED CURVES. 4.460 18) 10) 112) 141116
Cee et RSCTA CEPT CPs COTE Ca in FICE
Re OliCe ora aie Ta
al = al Ka zal i = I a a ia ga pT e me i aa >| Ea
Inds ai a
ee apy
_— bul m o 2 € > 2
=a i a i a i
if
je isc
es ees Bsa t+ Be == Ses ele my aa (res ieee] ea aa aN Eas = ae
-—
ba SH
ISS eS ae ales | aS
sea meee as = Etealea Hes ial 4 p— = =: =
TEMPERATURE SCALE FAHR.
AS ENS ie ele = Pe a eS
a = pal
Biz) oases
== iz (Ee Ss - = i
SGA eae ae SCUGBE SRI THHAGU AIIMRABIIICE a 4 oy
DAUvaAGRERand
FE
£.
D> | CES = Se ere Sa
a Ea
=e
fel ss ates 2g,
Be
SOLAR VARIATION AND TEMPERATURE DEPARTURES AT WASHINGTON
Fic. 11.—Oppositeness of temperature departures at Washington which follow average rising and falling sequences of solar variation.
comprise all the dates during 12 years when good consecutive solar-constant observations, made mostly at Montezuma, Chile, began to indicate rising or falling sequences of the sun’s output of radiation. The range of these sequences is only about 0.7 percent of the solar constant. Owing to the interference caused by changes in atmospheric transparency, superposed on the inevitable accidental errors of measurement, it is highly probable that some of these 320 cases are
32 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
a |
spurious. If the spurious cases could be eliminated the average temperature departures would doubtless be increased above their already large magnitudes.
The plain inference from these data is that short-interval solar changes are major causes, controlling weather for many days. It is so unexpected that I made other checks, described as follows:
. as I showed in my paper, “The Dependence of Terrestrial Temperatures on the Variations of the Sun’s Radiation,’ above cited, not only do several cities show large opposing trends of temperature following rising and falling sequences of observed solar variation, but 46 cases of unusually great observed solar changes were followed on the average by 1.95 times as large temperature changes in the same phases as the mean of 150 cases of all amplitudes. Again, the average trends of temperature following solar changes, as observed in the years 1924 to 1930, were nearly identical in phase, magnitude, and form with those observed in the years 1931 to 1935.
But now I offer a new evidence which I think is even more convincing. If, in reality, the observed variations of the sun were real, and influenced tem- peratures greatly for 16 days after their incidence, there still seems no reason to think there should have been any unusual temperature effects immediately before their occurrence. I have therefore computed for each of the 320 dates the march of temperature departures from normal for 16 days preceding the dates in question. I have then computed correlation coefficients for Washington as between the average marches of temperature attending rising and falling solar sequences, both after and before the beginnings of the sequences of solar change.
To fix ideas, I recall that in each division of this test there are 24 lines comprising 17 values each, two lines for each month of the year, selected from the 12 years 1924 to 1935. These pairs of 24 lines of the divisions are separated into two types, one type containing 17 values for days following, and 17 for days preceding the beginning of sequences of observed rising solar radiation. The other type comprises 17 values for days following and 17 values for days preceding the beginning of sequences of observed falling solar radiation. Two correlation coefficients are to be computed, one including the 204 values of the two contrasted types following the supposed critical dates, the other for the 204 values of the two types preceding them.
In order to avoid diluting the correlations by including extraneous influences due to previous conditions, each line was first reduced to the level of zero temperature departure, by adding to all 17 values in that line a constant quantity such as to make the average temperature departure for that line zero.
Having thus arranged the values, correlation coefficients were computed between the two types for the two divisions. They resulted as follows:
After appearance of solar change, r=— 54.3 4.9 percent, which is significant.
Before appearance of solar change, r=-+11.1+6.0 percent, which is meaningless.
The inference is obvious that the 320 dates, above described, were dates of real significance, since no other consideration was used in selecting them, and it is difficult to avoid the conclusion that they were dates when real solar changes began.
INO 1 AN IMPORTANT WEATHER ELEMENT—ABBOT 33
Are such small solar changes adequate to produce the apparent effects?
Sir George Simpson in his classic paper, “Further Studies in Terrestrial Radiation” (Simpson, 1928), concludes that 1 percent change of solar radiation might make up to 2° C. average change in the radiative temperature of the earth’s surface, but he says that this change would not be distributed uniformly. As a matter of fact, the temperature at any single station is strongly associated with the direction of the wind. The direction of the wind depends on the location of the station with respect to cyclonic centers. Any cause which alters the paths of cyclonic wind movements, alters wind directions and temperatures greatly. Clayton, from statistical studies, finds that variations of the sun, whether indicated by sunspot numbers or by solar-constant observations, are associated with large geographic changes of the centers of the barometric lows and highs. Hence, although a 0.7 percent change of solar radiation cannot produce all over the earth’s surface coincidentally a change in the same direction of 5° or 10° F. in temperature, it may very well produce on a given day a rise of 5° or 10° in one place, accompanied by a fall of 5° or 10° in another place. And this indeed is quite in agreement with the comparative results given in figures 1 and 6 of my paper, “The Dependence of Terrestrial Temperatures on the Variations of the Sun’s Radiation.”
SUMMARY
The preceding paper covers the following points:
1. Evidence that solar variation has been observed by independent stations in the northern and southern hemispheres with amplitudes as much as 18 times the probable error of the mean of the observations as deduced from their average differences.
2. The reality of solar variation is confirmed by many other corre- lations, including studies of the distribution of brightness across the sun’s disk, the spectral distribution of apparent solar variation, its relation to sunspot numbers, and to faculae and flocculae, the 27-day period in solar variation, and relations to atmospheric temperature and pressure.
3. Ten long periods in solar variation are found, ranging from 8 to 273 months. These have nearly a common multiple in 273 months.
4. These solar periods are traceable in terrestrial temperature and precipitation, but phases of terrestrial responses vary with seasons.
5. By analysis of temperature and precipitation records of Copen- hagen, Vienna, and New Haven, it is shown that the solar periodicities continued with unaltered phase for at least 140 years, and produced temperature responses apparently competent in their aggregate to account for the total range of departures from normal temperature.
6. Attempts to forecast weather conditions for 5 years in advance by evaluating the solar influences from past weather records and synthesizing them, proved fairly successful. ‘
34 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
7. Attempts to forecast weather conditions for 3 years in advance by using the master cycle of 23 years proved fairly successful. 8. Day-to-day changes in solar radiation, largely influenced by the
sun’s rotation, appear to be major factors in controlling weather for 2 weeks in advance.
APPALACH IANS
(With ONE von)
BY -M. B MITTLEMAN- pL Ohio bis eae Pe he Sco A ND “HARRY G. M. JOPSON ” Bridgewater College”
(PUBLICATION 3638)
-~
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOLUME 101, NUMBER 2
A NEW SALAMANDER OF THE GENUS GYRINOPHILUS FROM THE SOUTHERN APPALACHIANS
(WITH ONE PLATE)
BY M. B. MITTLEMAN Ohio University AND HARRY G. M. JOPSON
Bridgewater College
(PUBLICATION 3638)
GITY OF WASHINGTON PUBLISHED BY THE SMITHSONIAN INSTITUTION JULY 14, 1941
~ iss The Lord Baltimore Press BALTIMORE, MD., U. S. A. — ; | \
\ rae eh
we ae if : | | NMaVna th |
} tid): ay A Wee COs CIR Ree bys
A NEW SALAMANDER OF THE GENUS GYRINOPHILUS FROM THE SOUTHERN APPALACHIANS
By M. B. MITTLEMAN, Ohio University * AND HARRY G. M. JOPSON, Bridgewater College?
(WitH ONE PLATE)
Some few years ago the junior author chanced upon a single salamander of the genus Gyrinophilus while collecting in a small stream near Caesars Head, S. C., which could not be assigned to any known species in the genus. Discussion with Dr. E. R. Dunn, and an examination of the specimen by him, disclosed that it might possibly represent an undescribed form. However, no further speci- ‘mens could be obtained until E. B. Chamberlin of the Charleston Museum kindly lent the junior author a single adult Gyrinophilus which he had collected at Rocky Bottom, Pickens County, S. C., and which is plainly referable to the same form as the Caesars Head specimen. In view of the paucity of material, it was deemed best to withold description of the animal until such a time as further specimens might appear.
In the course of a revision of the genus Gyrinophilus, the senior author accidentally chanced upon the same form, represented in the collection of the United States National Museum by the Caesars Head specimen. Since additional material has become available, and with it a more accurate picture of the variation and distribution of the animal, the following description and notes are offered pending the publication by the senior author of a paper dealing with the entire genus.
*Contribution No. 23, from the Department of Zoology, Ohio University, Athens, Ohio.
* Contribution from the Department of Biology, Bridgewater College, Bridge- water, Va.
SMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 101, No. 2
2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
GYRINOPHILUS DUNNI, new species
PLATE I
Gyrinophilus porphyriticus dunni STEJNEGER, Rep. U. S. Nat. Mus. for 1937, p. 30 (nomen nudem).
Type.—vU.S.N.M. No. 113230 (Mittleman Coll. No. 382), female, collected on the campus of Clemson College, Clemson, Pickens County, S. C., 700 feet altitude, by Arnold Grobman, April 8, 1941.
Paratypes—U.S.N.M. No. 102440, Caesars Head, Greenville County, S. C., 3,000 feet; U.S.N.M. No. 102441, larva, Rabun Gap, Rabun County, Ga., 2,300 feet; U.S.N.M. No. 68168, Jefferson City, Jefferson County, Tenn., 1,250 feet; U.S.N.M. No. 68820, Indian Cave, near Jefferson City, Jefferson County, Tenn., 1,200 feet ; Three Springs, near Russellville, Hawkins County, Tenn., 1,350 feet; Charleston Museum No. 35.141.12, Rocky Bottom, Pickens County, S. C., 2,000 feet ; Charleston Museum No. 28.144.4, larva, Greenville, Greenville County, S. C., 1,000 feet; Sherman C. Bishop collection, one specimen from Sunburst, Haywood County, N. C., 3,400 feet; North Carolina State Museum No. 4905, Cane River, Yancey County, N. C., 3,000 feet; North Carolina State Museum No. 7594, Clemson College, Pickens County, S. C., 700 feet; North Carolina State Museum No. 7198, Cowee Mountains, between Jackson and Macon Counties, N. C., 3,000 feet; M. B. Mittleman collection Nos. 383-4, Clemson College, Pickens County, S. C., 700 feet ; Clemson College Division Ent. and Zool. Nos. 36 and 58, Clemson College, Pickens County, S. C., 700 feet; Clemson College Division Ent. and Zool. No. 105, Walhalla, Oconee County, S. C., 1,000 feet.
Diagnosis -——A Gyrinophilus which is red in life, with numerous tiny black dorsal flecks on head, limbs, body, and tail; throat pale, without dark reticulations; venter usually immaculate, or else with a few very small, dark flecks; canthus rostralis with a white orbito- labial line which is heavily bordered with black beneath, and imper- fectly bordered with black above; vomerines meeting the para- sphenoids at an acute angle; maximum size less than 160 mm. (average total length of type series, 125 mm.).
Description of type——Costal grooves 18 (including one axillary and two inguinal branches) ; 7 coastal grooves between appressed toes ; head width 13.62 percent of length from snout to vent; head length 14.20 percent of length from snout to vent; eye slightly shorter than the distance from its anterior angle to the nostril; snout swollen, a small tubercle at lower end of nasolabial groove; canthus rostralis prominent ; outline of jaw sinuous in its lateral aspect; angle of jaw
NO. 2 A NEW SALAMANDER—MITTLEMAN AND JOPSON 3
posterior to the posterior angle of eye; both eyelids fitting under a fold of skin at their posterior angle; a groove extending posteriorly from the eye almost to the gular fold, but intercepted by a small intervening fold; limbs well developed; fingers 3-2-4-1, in order of length, barely webbed at base; toes 3-4-2-5-1, in order of length, first toe half webbed, other toes very slightly webbed at base; vent grooved ; tail shorter than head and body, much compressed, keeled prominently except on the basal portion ; vomerine teeth 8-8 in series, commencing about half their length beyond the outer border of the inner nares, and curved anteriorly to become confluent with the parasphenoids at an acute angle; parasphenoids extending posteriorly beyond the rictus of the jaws for a distance equal to half their length, and separated their entire length by a distance approximately equal to half the length of a vomerine series; coloration in life, reddish, with thickly scattered minute brown flecks; coloration in alcohol, yellowish tan above, with brown flecks ; the narrow, white orbitolabial line of canthus rostralis bordered heavily beneath with blackish, imperfectly lined above with same; dorsal flecks of body assuming the vague outlines of chevrons; labial region with numerous heavy, black bars, whfch tend to form reticulations ; venter of head, limbs, body, and tail immaculate, except for a very few tiny flecks; heavy flecking of dorsum abruptly diminishing on the lateral areas, to be- come almost completely absent on the ventrolateral surfaces. Mea- surements: Total length, 136 mm.; snout to gular fold, 17 mm.; head width, 12.5 mm.; snout to anus, 82.5 mm.; tail, 53.5 mm.
Distribution—The southern Appalachian uplift below 3,500 feet, in North Carolina, South Carolina, Tennessee, and Georgia.
Remarks.—The range of G. dunni may possibly extend as far south as northern Jackson County, Ala.; Dunn (1926, p. 266) in- cludes a record for “Gyrinophilus porphyriticus’” from Sand Moun- tain, recorded by Holt, which is within the expected range of the new species.
The new form is considered a species rather than a subspecies because of the lack of what may be considered intermediate material between dunni and danielsi. Although this intergradation will prob- ably be shown to exist some day, we prefer the binomial in view of the absence of the critical specimens. Certain specimens are extant from higher localities (above 3,500 feet) in North Carolina that are not referable to dunni nor to danielsi; these specimens may ulti- mately be shown to be either an undescribed form or intergrades between dunni and danielsi. Similar-appearing specimens from
4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
Tennessee have been referred to by King (1939, p. 554) as danielsi x duryi. We do not agree with these conclusions, and a future paper by the senior author will deal more thoroughly with these forms. Undoubtedly, however, dunni is commoner in collections than present records would indicate; numerous specimens are undoubtedly extant under the name Gyrinophilus p. daniels.
The principal variation observed in available specimens is in the distribution of the dorsal flecks. These vary from very tiny, close-set flecks, which grossly present an irorated appearance, to somewhat larger spots. In no case, however, do these marks approach the size of those found in danielsi, nor is there ever the coalescence of markings in dunni that so frequently appears in danielsi. Occasional examples of dunni tend to have the dorsal markings arranged so as to present the appearance of short series of chevrons. Typically, there is a heavily barred or reticulated labial region, but a few specimens show only a lightly spotted or flecked surface. There are often a few very tiny spots or flecks widely distributed on the venter, but in no case is the ventral surface so heavily suffused with markings as it is in danielsi; indeed, it is generally immaculate. Costal grooves between appressed toes vary from 4 to 8, average 6.2. The angle of the vomerines with the parasphenoids is occasionally as acute as 45°, but never as obtuse as 90°. Available specimens tend to show that dunni is a considerably smaller animal than danielsi; as stated previously, the type series averages 125 mm. in total length of body plus tail. Only two specimens are larger than the type, both of these being 160 mm. in total length. Other than these, the type series, which is composed of sexually mature individuals, varies from 100 to 136 mm.
Dunn (1926, p. 271) has postulated that danielsi is probably more primitive than porphyriticus, and to this we add the hypothesis that dunni is mostly likely parental to danielsi. It would seem that danielsi has arisen from the more widespread, generalized dunni stock, and has become differentiated in its ecological niche, which, judging by the large size and robust form attained, has proved to be a favorable one.
ACKNOWLEDGMENTS
The senior author is indebted for advice and the loan of specimens to Dr. Sherman C. Bishop, Dr. Doris M. Cochran, Prof. Franklin Sherman, C. S. Brimley, and Dr. Willis King. A special word of thanks is due to Dr. Herschel T. Gier, of Ohio University, for his
NOt 2 A NEW SALAMANDER—MITTLEMAN AND JOPSON 5
many fruitful criticisms and unfailing interest in the present work.
The junior author would acknowledge especially the kindnesses and encouragement shown to him by Drs. E. R. Dunn and A. H. Wright. E. B. Chamberlin, of the Charleston Museum, has gener- ously loaned specimens in his care.
LITERATURE, CITED
Dunn, Emmett REID 1926. The salamanders of the family Plethodontidae. Smith. Coll. Publ., pp. viii 1-441, figs. 1-85, 2 pls. Kino, WILLIS 1939. A survey of the herpetology of Great Smoky Mountains National Park. Amer. Midl. Nat., vol. 21, No. 3, pp. 531-582, figs. 1-8.
rs i i Mal vey fey al Ti ee j nye bes 4 gue i } i ine t + 1 i H er f Uh hl } j i i ; { SOL i ike vie Ct a Cie i " Vol Hy i I a 5 | j rey i t ete 2 tas ‘ i fy ify) a eee Le | on : i hh ae vt ; Ay ea OM i 4 lle tora { if i Me Wee! Bi Hae Pe aA ® 1 aL ih { } i, } 4 M , } Ts hank (ua Pesce nt j ! aan } he 4 i Ms f i 4 j has 14 yi att? ve Awl AGL ae (hc te ; FS, f } nate i f i y j ¥ es abe | ; ei I Te a | ta yO as teh { { i ne « } ae } 1 \ 4 i 4 han : Wa ( PtAY j < ue y hi i i} oy i $ + i | kia tk m i i \ | ; i , { | 1 ly { ia ‘ ' { Pa yi th 4 Ay i 1 ‘ 7) i 1 } i] i fi 1 j i vA i i i [ ie i b | | i i ‘ mn } { fl i i thi { i i ALO seen 1 i i We fi uM | } ' i Ke tet heck Wahi ie t ‘ ba ! He ee | Team ij aly Hy 7 Thy wis pune Lee | Ui j } | { , j i i i it ha ; , i ary) i j l ya a he] } 1 fh A pa i TL tee,
| i i | ) u dak aa i } } : af ie PR CAOe emer Et Zo RETR Uae ls Mae dak Ooh ela IP Lhe j j ri 1 fi f ma Lhe it We yl iy / ‘ mit Mi
at i a AN ie Vii, ri
a. Yael t ove Lira ey, Wis
ree: 7
re
SMITHSONIAN MISCELLANEOUS COLLECTIONS WAOIES AO INOS 245, TEES
B
GYRINOPHILUS DUNNI, SP. NOV.
Type —U.S.N.M. No. 113230 (Mittleman Coll. No. 382), female, Clemson College, Clemson, Pickens County, S. C., 700 feet. A, dorsal view; B, lateral view; C, ventral view. Actual length, snout to anus, 82.5 mm.
tae wi)
; Wi i f | ; f, 1 j \ } { yn ' f j f i | ! f ie 5) i} j i jl a | \ he | i ie) if : I i it nul ° ve il ‘et al iy j I i i + ; VY ol " i! i v i i } | 4 if : i 7 } t \ 7 i] 9 ti Wh , } { ; 5 4 rei f ! , j 4, i f ti % Ta ; | in | \ 4 04 Pau i ingle ye { ty } iy t ; ik | | re} i Hy f i piuy ; : id if hy : i ! j i t 1 ’ ‘ i { } ha : i
i) i i Aah Ly ‘ Hat tly i i i} u uf ven 4 | ie Dee Kee nae ‘yield ; | ai hi | Mae iN . FER vied ies y j f ' tf At ra i ut i a : . ; i yh me | i ‘mae , a)
ni vie,
SMITHSONIAN, MISCELLANEOUS COLLECTIONS peat OLE: 101, ‘NUMBER 3
"GREAT PLAINS
/ 7}
: “ewirs . ie PLATES)
“WALDO R. WEDEL
"Assistant Curator, Division of Archeology Wee eens National “ak gikocrs
ee corey, oF WASHINGTON Nip UBLISHED BY THE SMITHSONIAN INSTITUTION
t
é AUGUST. 20, 1941
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOLUME 101, NUMBER 3
ENVIRONMENT AND NATIVE SUBSISTENCE ECONOMIES IN THE CENTRAL GREAT PLAINS
CWiTH FIVE Piates)
BY WALDO R. WEDEL
Assistant Curator, Division of Archeology U. S. National Museum
yy
<< NY “Paso DENTON
(PUBLICATION 3639)
GITY OF WASHINGTON PUBLISHED BY THE SMITHSONIAN INSTITUTION AUGUST 20, 1941
f ' os) ; ep) . ’ " , \ tu hes . ra ts Dar aN Al it Ph it fi mn) ‘ y, ‘ ‘ ia &. \ ) i" + The Lord Baltimore Press BALTIMORE, MD., U. & A. . .
: fl y
i fhe
y i
} 4 Al A
: ae ¥ WAN ih: ‘a RRL Pie VY er er 4
ENVIRONMENT AND NATIVE SUBSISTENCE ECONOMIES IN THE CENTRAL GREAT PLAINS
By WALDO R. WEDEL Assistant Curator, Division of Archeology, U. S. National Musewm
(WitH Five PLATEs)
During the past 10 years an increasing interest has been manifested in the relations of man to environment in the Great Plains. Wide- spread droughts, spectacular dust storms, and recurrent crop failures are driving home again a fact which had been largely forgotten during the preceding prosperous decades—namely, that the climatic fluctua- tions to which the region is subject can be of sufficient magnitude to render man’s occupation precarious. Numerous farms have been abandoned, and there is a rather general belief that much of the land is wholly unsuited to agriculture. Students of ecology and geography, recalling similar happenings in the past, have been insisting again that a long-range program of land utilization in place of the present hap- hazard methods would make possible the recovery of much of the supposedly worthless area.
It is not my intention here to suggest a cure for the economic problems arising from the conditions just noted, but rather to examine certain pertinent facts brought out by recent archeological investiga- tions. We know now that long before white explorers ventured into the Great Plains, the region had been exploited in different ways by various native peoples. There is a growing belief that some of these aboriginal groups may have had to cope with adverse climatic condi- tions similar to those faced by man here today. The evidence is still fragmentary and scattered, because the area involved is enormous and the workers are few. Still, it may be worth while to indicate the directions in which the available data appear to lead.
For present purposes the central Great Plains comprise the area included in the States of Kansas and Nebraska. We shall review briefly the environmental setting as a background for an outline of the historic and prehistoric native subsistence economies. This will be
SMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 101, No. 3
2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
followed by a consideration of the possible significance of certain inferred climatic variations upon past human populations.’
ENVIRONMENTAL FACTORS
To the casual observer traveling across Nebraska and Kansas along the main thoroughfares, the region exhibits a rather wearisome uni- formity. In broadest outline it is a land of low relief, few trees, and little rainfall—of sun and wind and grass. A closer regard for the details of topography, native flora and fauna, and other aspects of the environment will show, however, that there are a number of natural variations (see Shelford, 1926, and Fenneman, 1928) which bear on man’s utilization of the land.
The western portion of the area is a part of the High Plains proy- ince (fig. 1)—the remnant of a great outwash plain which in Tertiary times reached from the mountains into eastern Nebraska and Kansas. The High Plains are characterized by broad, monotonously flat upland areas which, toward the north, tend to become uneven or gently rolling. Rivers heading in the Rocky Mountains, such as the Arkansas and Platte, flow eastward across this belt in wide, flat-floored valleys. Lesser streams rising within the High Plains occupy shallow, open valleys which in places give way to picturesque rock-walled canyons. Where the secondary valleys have been cut through the unconsoli- dated Tertiary silts, sands, and gravels into the impervious underlying formations permanent springs occur. These give rise to perennial creeks flowing in verdant valleys which contrast strikingly with the surrounding uplands and which, since time immemorial, have pro- vided ideal camp and village locations for primitive man (pl. 3). | Where not under cultivation, the uplands are dominated by buffalo and grama grasses, with yucca, cactus, and sagebrush locally abun- dant. In the valleys there are groves of hackberry, cottonwood, and willow, with some elm and ash. Thickets of wild plum, elderberry, and other edible native plants are scattered along the ravines and stream valleys. Juniper grows along the valley rims, and in parts of Nebraska there are stands of western yellow pine. Native fauna included notably such animals as the bison, antelope, mule deer,
*For helpful information and stimulating suggestions as this study developed, I am particularly indebted to Dr. C. E. Leighty, J. S. Cole, and O. R. Mathews, agronomists at the Division of Dry Land Agriculture, Bureau of Plant Industry, U. S. Department of Agriculture; and to Harry E. Weakly, junior agronomist at the North Platte Experimental Substation of the College of Agriculture, University of Nebraska.
NO. 3 CENTRAL GREAT PLAINS—-WEDEL 3
\
Priore au
| | | |
SOUTH DAKOTA Ae
_——
100 MILES 102
Fic. 1—Map of the central Great Plains and upper Missouri Valley, show-
ing physiographic divisions, tribal groups, and archeological sites considered in the present paper.
4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
prairie dog, coyote, black-footed ferret, jack rabbit, badger, and smaller forms, and game birds such as prairie chicken and grouse.
East of the High Plains there is a marked change in the terrain. North of the Platte River in Nebraska are the Sandhills, a rough, hilly area dotted with ponds, lakes, marshes, and hay flats. The hills are in reality sand dunes which are held fast by a cover of bunchgrass interspersed with yucca and cactus. Trees are scarce except in the vicinity of ranch buildings. There are few streams, but those which head in the region, notably the several forks and upper tributaries of the Loup, carry an abundance of excellent water throughout the year. They head. in shallow, grassy swales which soon give way to deep canyons lined with deciduous timber, plum thickets, and other vegeta- tion.” Wild rice grew in some of the shallow lakes and was an impor- tant native food item (Gilmore, 1919, p. 67). Beaver, muskrat, deer, and smaller mammals inhabit the valleys; coyotes and jack rabbits abound; and great numbers of waterfowl still have their nesting grounds about the lake margins. Like the High Plains, the Sandhills are a region of low rainfall. Since the soil is loose and sandy and tends to blow readily when the sod is removed, large-scale agriculture is impracticable and cattle raising is the principal activity today.
South of the Republican River, in north-central Kansas, the eastern front of the High Plains has been dissected by stream erosion to produce the so-called Plains Border. This is a belt of high plateaus and prominent east-facing escarpments cut through by the deep, broad valleys of the Smoky Hill, Saline, Solomon, and their tributaries. Buttes and ridges are a conspicuous feature of the landscape. The fertile river valleys are fringed with forests of oak, elm, ash, walnut, cottonwood, and other hardwoods, and bluestem and bunchgrasses formerly clothed the upland areas. Native fauna included all the larger forms found on the High Plains together with numerous other species more typical of areas to the east.
*In 1895 Bessey (1896) observed that relict stands of western yellow pine were to be found in the canyons along the Niobrara River as far east as Holt County, along the North Platte and Lodgepole Rivers to Lincoln County, in widely isolated canyons in Valley, Custer, Greeley, and other Counties in central Nebraska, and along the Republican to the vicinity of Franklin. These localities lie for the most part west of the 99th meridian. Hussong (1896) also calls attention to the fact that “yellow pine grew formerly in and along the canyons south of the Republican River in Franklin County ...” Early settlers cut the larger trees for fuel and posts, and transplanted the smaller ones to gardens and lawns. Bessey held that these distributions were best explained “by supposing that the central region was once wholly or in part covered with forests.”
NO. 3 CENTRAL GREAT PLAINS—WEDEL 5
East of the Sandhills and Plains Border provinces the Great Plains give way to what was formerly a bluestem and prairie-grass region. In eastern Nebraska the Loess Plains present a flat to rolling or hilly terrain. A thick mantle of loess extends from the Missouri Valley westward to the Sandhills and, as a narrower band, between the Platte and Republican Valleys to the High Plains in southwestern Nebraska (pl. 4, fig. 1). To the southeast it merges into the unglaciated gently rolling Osage Plains lying south of Kansas River. Throughout all this region, the streams run in broad, bluff-lined, often terraced, val- leys. Fine stands of burr oak, elm, walnut, hickory, sycamore, and other hardwoods skirt the streams. Edible plants include the wild grape, plum, chokecherry, mulberry, and a number of tuber-bearing forms (see Gilmore, 1919). Other items which unquestionably figured in aboriginal economy were the Osage orange or bois d’arc, papaw, pecan, and persimmon, all found in eastern and southeastern Kansas. Bison, elk, antelope, and deer found unlimited-pasturage in the valleys and on the deeply grassed uplands. Wolves, coyotes, badgers, and rabbits were plentiful; along the streams were beaver, bear, otter, raccoon, cougar, opossum, wildcat, fox squirrels, and a host of lesser forms. Most of these followed the wooded valleys westward far into the plains. Permanent streams are, or were, the rule, and fine springs formerly abounded on the larger as well as on many of the lesser watercourses. Precipitation, except in the west, is ample for agricultural needs, and the soils everywhere are deep, rich, and easily worked. Today this is the most densely populated and the finest agricultural land in the central Great Plains.
The Loess Plains are flanked by a narrow strip of rugged hills on the east where short, deeply incised creeks and narrow, heavily timbered ravines empty directly into the Missouri River. Broadleaf forests and dense thickets were formerly dominant, and the district abounded with game and other wild-food resources. Historical ac- counts indicate that the lofty bluffs fronting on the Missouri were bare of trees during the early nineteenth century, but the valleys must always have been well wooded.
From the standpoint of agriculture, whether modern or aboriginal, probably the most important factor in utilization of the Great Plains by man is climate (Kincer, 1923, and Climatic summary of the United States, sections 38-41). The region as a whole is characterized by warm summers with abundant sunshine; by winters that are cold and dry; and by considerable windiness throughout the year. In the central portion, as we have defined it, there is a frost-free growing
6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
season of 150-200 days, sufficient to mature most cereals and vegetable crops. Precipitation decreases markedly from an annual total of about 40 inches in southeastern Kansas to 15 or 18 inches in western Kansas and Nebraska. Because of the dry winters, the moisture stored in the soil is generally scanty; hence crops depend chiefly on the rainfall during the growing season. In this respect, the average annual precipi- tation shows a favorable distribution, since about 70 percent or more falls between April and September.
As Smith (1925, p. 413) has pointed out, “averages do not tell the whole story. Averages rarely happen. The freaks of the season decide man’s chances .. .” In the Great Plains, the minimum precipitation required for successful agriculture by ordinary methods is somewhere between 15 and 20 inches annually. This means that over much of the region, particularly west of the 98th or goth meridian, the yearly average is very close to the minimum required for successful crop growth, and hence that the year to year variation is of very great importance. A decrease in rainfall of only a few inches, particularly if continued for several years, may result in a major disaster for the grain farmer.’ Furthermore, long-time weather records show that in the western Great Plains “there is less than the normal amount of rainfall in more than half the years.” Farming except by specialized methods is rendered still more precarious by the fact that there may be several seasons of deficient moisture. These fluctuations are of vari- able duration, do not come in regular succession, and cannot be fore- cast with any accuracy.
The character of the summer rains must also be considered. In large part, they come as thunderstorms, often of great violence and short duration. Downpours of 3 to 6 inches within a 24-hour period have been recorded at many points, this sometimes exceeding in amount the normal precipitation for the month in which it occurs. The rains may come with such force that the ground surface is puddled, so that most of the water runs off before it can be absorbed
* The delicate balance between yearly rainfall and crop yields in this western area has been strikingly demonstrated by Cole (1938). From weather and crop records for 14 stations in western Nebraska, Wyoming, North and South Dakota, and Montana, he has determined the mean precipitation and mean average yield of spring wheat over periods varying in length from 10 to 28 years. With an average precipitation of just under 15 inches, the average wheat yield was slightly over 15 bushels per acre. When average precipitation fell 20 percent to 12 inches, wheat yield dropped nearly 50 percent to a trifle over 8 bushels. A further lowering of precipitation to 50 percent of normal thus resulted not in a half crop but in no crop at all. No comparable statistics are available on rainfall and corn growth in this region, but a similar correlation may be suspected.
NO. 3 CENTRAL GREAT PLAINS—WEDEL 4
by the soil. These storms, furthermore, are mostly local, with short, erratic courses that may leave one small area drenched while sur- rounding sections receive little or no moisture. Similar local dif- ferences are reflected in the annual precipitation records from time to time. Thus, in 1875 and again in 1901-3 inclusive at Dodge City, Kans., annual precipitation was from 50 to go percent of normal (average, 19.9 inches) ; but during the same years at Hays, 85 miles to the north, there was an excess of moisture ranging from 5 to nearly 50 percent (average, 21.28 inches).
Droughty conditions in summer are often attended by prolonged periods of high temperatures. Southerly winds predominate; being warm, they have a drying effect, and their high velocity favors rapid evaporation. Particularly destructive are the so-called “hot winds” which may accompany shade temperatures of 100° to 110°. They have been likened to a blast from a hot furnace, and frequently cause much damage to crops and serious discomfort to animal life. Immense havoc may be wrought in a few hours if these winds occur at critical stages of crop development, and when they continue for several days man and beast alike suffer intensely and widespread crop failures result. Many farmers insist that these hot winds can kill the corn crop even when through subirrigation or otherwise the soil is moist enough to meet the normal requirements of the growing plants. I am not certain, however, that this view has general acceptance among impartial observers.
HISTORIC SUBSISTENCE ECONOMIES
The ggth meridian, lying somewhat east of the line of 20-inch annual precipitation, may be regarded as the approximate dividing line be- tween the Great Plains on the west and the true prairies or prairie plains on the east. It will serve also as the line of demarkation between two strikingly divergent native subsistence economies in the central Great Plains during the nineteenth century. To the east, where soil and especially climatic conditions are today recognized as most favor- able for farming, the native economy was based on horticulture with hunting secondary. West of this line, where bitter experience has since shown the white settler that agriculture is likely to be a highly uncertain venture, hunting was of primary importance.
The principal natural game resources of the High Plains have already been noted. What they may have lacked in variety was more than offset by the abundance of certain species. During the nineteenth century the High Plains and the Plains Border immediately to the
8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
east were preeminently the range of the great bison herds. Parasitic on these were several Siouan, Algonquian, and Shoshonean Indian tribes whom we may term the migratory bison hunters (fig. 1). North of the Platte were the Dakota bands ; farther south was the habitat of the Cheyenne, Arapaho, Comanche, and Kiowa. Throughout the spring, summer, and early fall organized bands of these Indians hung about the flanks of the herds, subsisting chiefly on the flesh of the bison and drying large quantities of meat for winter use. In winter the roving village units usually returned to certain sheltered spots which because of water, wood, and forage for horses were used year after year. Large herds of horses were to be seen at every camp. The portable skin tipi was universally used (pl. 1). Skin working was highly developed; otherwise, implements, utensils, and industries were limited to essentials which could be moved easily and conveniently on horseback from camp to camp. Surplus foods were stored in skin containers. Agriculture was nonexistent, as was the potter’s art. The gathering of quantities of wild fruits, nuts, berries, and starchy roots and tubers supplemented the chase (Carlson and Jones, 1940) ; maize was obtained by trade or theft from settled horticultural tribes. An elaborate militaristic system had been built up, and much of the time not occupied in food getting was given over to warfare and horse stealing. These peoples recognized no definite tribal boundaries, and the distances traveled by them during their annual hunting trips and in raiding forays frequently totaled many hundreds of miles.
East of the 99th meridian, since the coming of the white man, have dwelt chiefly Siouan- and Caddoan-speaking tribes. Eastern Kansas was held by the Osage, northeastern Kansas by the Kansa, eastern Nebraska by the Oto, Missouri, Omaha, and Ponca. Farther to the west, on the Loup, Platte, and Republican Rivers, stood the villages of the Pawnee, a confederacy of Caddoan tribes whose nearest kindred linguistically were the Arikara in South Dakota. At the dawn of the contact period, the Wichita are believed to have had a group of settle- ments in central Kansas in the vicinity of the great bend of the Arkansas. These were abandoned during the eighteenth century for other lands farther south. The rest of the tribes enumerated remained in their respective locations until reservation days a century or so later.
All these tribes dwelt in large fixed villages situated near streams where wood, permanent water, and arable ground were to be had. Habitations were circular earth-covered or grass-thatched lodges (pl. 2). Subsistence was based primarily on the cultivation of maize, beans, and squash, to which were added a long list of wild berries,
NO. 3 CENTRAL GREAT PLAINS—-WEDEL 9
fruits, and tubers (Gilmore, 1913, and 1919). The latter included wild plum, hackberry, chokecherry, sand cherry, wild potato (Ipomoea pandurata), ground bean (Apios tuberosa), the pomme blanche (Psoralea esculenta), and others. Of considerable importance, too, especially after acquisition of the horse about 1700, were the products of the chase. Bison were the principal game animal, and to obtain them one or two well-organized hunting trips were made annually into the western plains. The Omaha and Ponca hunted north of the Platte into the Sandhills; the Pawnee went either up the Platte and Republican Valleys, or else shared with the Kansa and Osages the Plains Border and adjacent regions in central Kansas. At such times the entire population of the villages excepting the very young, the senile, and the decrepit moved en masse, dwelling in portable skin tipis and hauling their.impedimenta on horseback or by travois, and living in general like the migratory bison hunters. At the temporarily deserted villages, the possessions which could not be carried along were concealed in underground pits. Under aboriginal conditions all these groups made pottery, and possessed in addition well-developed industries in stone, bone, horn, shell, and other materials. Whereas the temporary campsites of the migratory bison hunters today show little evidence of occupancy beyond hearth areas and possibly a few stone implements and animal bones, the abandoned house sites, cache pits, and accumulated refuse deposits of the village dwellers usually yield a rich harvest for the archeologist.
Fundamentally, the native agriculture of the Great Plains was of southeastern type, with tillage mostly or entirely by the hoe. The old type of hoe consisting of a bison shoulder blade lashed to a bent or forked stick survived until very late times, being used side by side with iron tools supplied by the traders. The fields—more accurately de- scribed as gardens—were small, ranging in size from 4 to 3 or 4 acres. No attempt was made to break out the tough sod of the uplands. In the valley bottoms, the plantings were confined to little patches of loose alluvial soil scattered along the creek banks or lying at the mouth of a ravine. Because such spots were usually limited in number, the women often found it necessary to travel from 5 to 10 miles to and from their gardens. Corn, beans, and squash were the principal crops, but sunflowers, tobacco, and watermelons were also grown (Gilmore, 1913, p. 322). Women did all the planting and cultivating. Fertilizers were unknown and there is no evidence that irrigation was attempted. It is not definitely known whether the Indians of the central Great Plains had developed special deep-rooted, early-maturing, or drought-
15@) SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
resistant varieties of corn, as had the Mandans of the upper Missouri (cf. Will, 1922), but it is quite possible that in the course of years some selection of this sort had taken place.
The fields were ordinarily hoed only once or twice. After the second hoeing, in June, the entire population of the village set out on the summer hunt, returning in September to harvest the crops. Surplus corn was boiled, cut from the cob, and dried, and then stored in underground caches. These caches, when emptied of foodstuffs or when rendered unfit for further use by spoilage of their contents, were abandoned, often to be refilled with refuse. Such pits are a common feature of every village site so far excavated in the central Great Plains where horticulture is evidenced. In the early historic villages of the Pawnee along the Loup and at the contemporary Wichita (?) sites in central Kansas, these caches are often 6 feet deep, and they have been known to attain a depth and a diameter of 10 feet or even more. At later sites, particularly in those dating after 1800, caches seldom reach these dimensions. It is not yet clear whether this decrease in size reflects smaller crops due to a slackening interest in farming and a correspondingly greater reliance on hunting or is attributable to some other factor.
The Pawnee and Omaha used an upright wooden mortar with pestle for grinding corn; their protohistoric contemporaries in central Kansas used the flat or hollowed stone mealing slab with muller.
ARCHEOLOGICAL CONSIDERATIONS
As systematic archeology adds perspective to our picture of native life in the central Great Plains, it becomes increasingly clear that the two fundamental economic patterns outlined above, or variants thereof, have long been present locally, but that man’s emphasis has shifted back and forth from one to the other. As Kroeber (1939, pp. 76-79) has indicated, ““. . . the historic Plains [horse] culture was a late high- pressure center of culture in a region which previously had been rather conspicuously low-pressure.” He is undoubtedly correct, too, when he observes further that in the prehistoric period, prior to the sixteenth century, the plains were a cultural margin. By comparison with the eastern Woodlands and the Southeast, where advanced mound-building civilizations once flourished, the region west of the Missouri is characterized by antiquities of quite unspectacular nature. Temple mounds, for example, are nonexistent; and the practice of raising tumuli over the dead, which serves to emphasize the highly elaborated burial cults of the east, extends only a little way into the
NOS 3 CENTRAL GREAT PLAINS—WEDEL - II
prairie plains of eastern Kansas and Nebraska and is not indicated at all for the plains proper.
The admittedly marginal nature of the semisedentary village cultures in historic times, together with the nomadic mode of life followed by the “typical” plains tribes farther west, has tended to obscure certain facts relating to the earlier agrarian peoples in the area. Kroeber (op. cit.) states that “It is scarcely contendable that the western plains were wholly uninhabited before the horse was available. Agricultural groups from east and west probably strayed in now and then and tried to farm. Small groups could make a living by com- bining bison and river bottom hunting with berry and root-gather- ing...” Archeology shows that primitive maize growers from the east had indeed penetrated far beyond the western margin of the prairie plains and had established themselves along many of the stream valleys in the High Plains. Their settlements were much smaller—and far more numerous—than those of such historic village tribes as the Pawnee. It is hardly accurate to speak of these pre- historic groups as mere “strays,” for they came and spread in suf- ficiently leisurely fashion to scatter their remains along almost every arable stream valley with reasonably sure water as far west as the Colorado line. The diffuse nature of this early occupancy is in striking contrast to that of historic times when such tribes as the Pawnee dwelt in a very few large compactly built towns within a few miles of one another.*
Strong (1935) was the first to point out clearly that in the light of archeology the limitations of environment in the Great Plains were not so severe as many have been led to believe. In the long-range view, the droughts, excessive temperatures, and searing winds, which have played havoc with the present-day farmer and his commercial ventures, are comparatively transient if recurrent phenomena. Under normal climatic conditions the region is less hostile. Some degree of success, at least, must have attended the efforts of the native peoples who, content with a subsistence agriculture, ventured to try their hand at wresting a living from the soil of the short-grass plains.
As an example we may cite the recent discovery of pottery-bearing sites yielding definite proof of native farming activities in Chase
*The Pawnee after 1800 were variously credited with 5,000 to 12,000 persons. I suspect that if these were redistributed among the older sites, many of the latter would be decidedly underpopulated. Even granting that not all the small pre- historic sites were inhabited synchronously, I am of the opinion that there may at times have been about as many Indian farmers in prehistoric Nebraska as there were during the nineteenth century. A very considerable proportion of these, moreover, lived in the High Plains area.
12 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
County, Nebr. (see fig. 1, Ch: 1), and in Scott County, Kans. (fig. 1, Sc:1).° These sites lie near the 101st meridian, 300 miles or more west of the Missouri, and well within the dry High Plains province (pl. 3). Both have yielded charred corn, together with bone hoes made from the scapula of the bison. At the Scott County site (pl. 3, fig. 1), where certain puebloan influences suggest the possibility of irrigation, the remains of squashes or gourds were also found. The great abun- dance of animal bones, as well as the very limited positive evidence of fixed habitations, leads to the belief that hunting probably ranked first in the food economy, with horticulture perhaps a side line. There is no reason to suppose that the inhabitants of these sites were related to the Pawnee or to any of the other Caddoan or Siouan village tribes of the eastern plains, or that they were directly ancestral to the Dakota, Cheyenne, and Arapaho who roamed the same area in the late eighteenth and nineteenth centuries. Scant amounts of iron trade materials indicate an early historic or protohistoric dating. The sites have been assigned to the Dismal River culture ; it is possible that they will eventually be attributable to some of the semihorticultural Apache communities which according to seventeenth- and eighteenth-century Spanish documents formerly lived in the bison plains. At present they represent the westernmost points at which maize specimens have been reported archeologically in the central Great Plains.
In an earlier period, before any European influences had yet reached the area, horticultural peoples of another sort dominated the central Great Plains. Most widely distributed and best known are those who left the remains comprising the Upper Republican culture (Strong, 1935, pp. 69-124, 245-250, 275-278; Wedel, 1935, and 1940b, pp. 310- Bui Champe, 1930). These remains occur throughout the Loess Plains to the edge of the Sandhills and south to the Smoky Hill- Kansas River drainage, with a westward extension far up the Repub- lican (pl. 4, fig. 1) and Platte basins.” They consist of innumerable small village sites situated near former springs or other permanent water on the flood-free terraces which characteristically border many
> Wedel, 1940a; unpublished field notes by A. T. Hill, Lincoln, Nebr.
®It is interesting to note that in terms of present-day agricultural regions, the village sites of these prehistoric agrarian peoples are most plentiful in those por- tions of Nebraska and northern Kansas which are assigned to the Corn Belt (see O. E. Baker’s map in Atlas of the Historical Geography of the United States, pl. 142A). Their range also includes adjacent districts in northern Kansas now given over to hard winter wheat. From surface finds of pottery, it appears that they may have inhabited the present corn-wheat zone in northeastern Colorado.
NO. 3 CENTRAL GREAT PLAINS—-WEDEL 13
of the creek valleys. The villages included from a half dozen to two or three dozen rectangular pit houses scattered over several acres of ground. Unlike the historic Pawnee villages, these earlier communi- ties were apparently unfortified. Cache pits are present in all these sites ; they seldom exceed 3 or 4 feet in greatest depth and diameter, and thus average much smaller than those of the early historic Pawnee. In them are found charred maize, beans, animal bones, pottery, stone and bone artifacts, and other evidences of human industry. Universally present is the bone hoe. Charred corn, cobs, and beans have been found in these sites as far west as Medicine Creek in Frontier County, Nebr. Typical pottery, together with evidence of earth-lodge remains, occurs still farther west in Chase County, and excavation will prob- ably show that the occupants of these villages also practiced farming. Inferentially, this must have been on a small scale and by methods closely similar to those described above for the Pawnee and their neighbors. The hoe, and also the crops evidenced through archeology, all have historic counterparts. Since the prehistoric villages were much smaller, it can be assumed that sufficient arable ground was available in the bottom lands close at hand. The horse was unknown, but since the villages were scattered widely over much of the choicest bison range it is probable that ample supplies of meat could be obtained by foot hunters whenever desired. At any rate, the bone refuse in their sites shows that the Upper Republican peoples relied probably more on the bison than on any other single species for meat.
A somewhat variant contemporaneous manifestation along the Missouri River bluffs has been termed the Nebraska culture. Here, too, are found rectangular earth lodges, cache pits, pottery, and other evidences of a settled horticultural mode of life. Animal bones indicate that white-tailed deer, elk, and smaller mammals figured much more heavily in the native diet than did the presumably more distant and less easily obtainable bison. In this respect the nonequestrian pre- historic inhabitants of the immediate valley of the Missouri differed from the historic Siouan tribes who, mounted, could and did travel hundreds of miles westward in quest of the bison.
Underlying the widespread Upper Republican and contemporaneous manifestations are other pottery-bearing horizons. Remains designated as Woodland are found in small obscurely situated sites which have so far received scant notice from archeologists. In the ravines of eastern Nebraska they occur as artifact-bearing occupational strata exposed in newly cut banks, with an overburden that varies from 6 to 25 feet or more (pl. 5, fig. 1). Farther west, as in the High Plains
14 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
of Lane County, Kans., and elsewhere, similar strata are often over- lain by wind-blown soils. It is still uncertain whether these overlying materials represent short cycles of abnormally heavy precipitation (in the east) or of excessively dry, windy conditions (in the west), or are due to slower but long-continued erosional processes. The extent to which the Woodland groups depended on cultivated crops is also a question since the sole evidences yet reported of horticulture—at the Walker Gilmore Site in Cass County, Nebr.—involve only the squash and gourd (Strong, 1935, p. 193).
Apparently contemporaneous with the Woodland sites are others along the Missouri in northeastern Kansas and nearby Missouri, which show a close similarity in several respects to the Hopewellian remains of the Mississippi-Ohio drainage. Evidence of horticulture is generally wanting throughout the Hopewellian horizon in the eastern United States, but it has been postulated that such an economic basis would have been necessary to so highly developed a civilization (Setz- ler, 1940, p. 262). At any rate, the only Hopewellian-like village site so far investigated seriously on the Missouri yielded direct proof of horticulture in the form of charred maize and beans (Wedel, 1938, p. 101). The bone hoe is absent, probably having been replaced by implements of stone or other material. To what extent the local economy was based on gardening is not clear, but cache pits possibly for storage of domestic crops were relatively plentiful. These yielded considerable quantities of deer and raccoon bones, but almost none of the bison. Since this complex does not appear to have penetrated very far west beyond the Missouri it is of minor interest so far as primitive horticulture in the Great Plains is concerned.
The prehistoric potters and farmers of the central Great Plains are thought to have moved into the region from a general easterly direc- tion. If the Woodland peoples practiced horticulture, as we know the later Upper Republican groups did, then at least two principal waves of immigration by native farming economies are indicated. How long the interval separating these two is we do not know; there is no clear proof that the Upper Republican developed directly out of the Wood- land, though some contact between the two is indicated (cf. Wedel, 1940b, p. 346). Both groups spread westward into the High Plains, nearly or quite to the present Colorado line. There is no way of telling whether either found the western plains uninhabited. It is abundantly clear that hunting economies had occupied much of the region at a far earlier time, as shown by the presence of Folsom, Yuma, and other ancient nonagricultural remains. Since hunting tribes again
NO. 3 CENTRAL GREAT PLAINS—WEDEL 15
controlled the area in historic times, it can be inferred that occupation by native subsistence economies based on maize constituted a rela- tively brief interlude which was preceded and followed by very much longer periods of occupancy by nomadic or seminomadic bison hunters.
The first Spanish explorers to visit the western plains in the six- teenth century found them occupied by nomadic Indians who had “no other settlement or location than comes from traveling around with the cows.” In terms of modern linguistic groups, these are thought to have been Athabascans, probably Apache or Lipanan (Harrington, 1940, p. 510). Just when they arrived is not clear, but by 1541 the little farming communities over most of the Upper Republican area had evidently been given up. At any rate, there is no mention in the narratives of the Coronado or subsequent expeditions of anything corresponding te the Upper Republican village sites as these have been defined by archeology. The first permanent settlements seen by the sixteenth-century Spaniards in what is now central Kansas were large, some of them estimated to number 200 houses (Winship, 1896). The houses were of straw, and the natives are described as having corn, beans, and melons. All of this is reminiscent of the large proto- historic villages found in Rice and McPherson Counties, Kans. Insofar as it concerns fixed villages of horticultural peoples it also calls to mind the great fortified towns of the protohistoric Pawnee on the Loup River in Nebraska—possibly the Harahey of Coronado’s chronicles (see Lesser and Weltfish, 1932, p. 12). Farther west, according to these explorers, were only migratory hunters whose mode of life was essentially the same as that of the historic hunters except for innovations taken over by the latter from white men. The cause or causes for the observed abandonment of the western plains by native farming peoples in late prehistoric times is one of the problems now confronting plains archeologists.
DROUGHTS AND’ PREHISTORY
Van Royen (1937, p. 637) has remarked that “there is little doubt in the minds of students of weather and climate that wide borderland areas between humid and arid regions will always be subject to recur- rent droughts of varying duration and intensity, such as those experi- enced in historical times. Also, before the dawn of recorded history droughts occurred, some of which were brief, others evidently very long.” The Great Plains constitute just such a borderland zone, with arid regions on the west and southwest and humid regions to the east. We may turn, therefore, to a consideration of certain phenomena
2
16 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
which suggest drought conditions during the prehistoric occupancy of the central Great Plains, keeping in mind particularly their possible effects on the native horticultural peoples.
Two sites in Nebraska are of especial interest in this respect. They are Signal Butte in the North Platte Valley near the Wyoming line, and the Lynch Site on lower Ponca Creek in Boyd County about 12 miles from the South Dakota border. Archeological remains at Signal Butte and their geologic context have already been detailed elsewhere (Strong, 1935, p. 224-239). Briefly, they consist of three prehis- toric levels of human occupancy separated from one another by layers of wind-deposited loess from 18 to 24 inches thick. The topmost cul- tural horizon includes pottery of Upper Republican and Dismal River types, and may be no more than 300 to 500 years old. The middle and lower strata, yielding no pottery, are believed to represent much older hunting cultures. The occupational strata consist in part of humus, and presumably indicate periods of increased humidity. The intervening sterile layers are attributed to dry, windy periods. Van Royen suggests that “one or both of the sterile strata on Signal Butte may correspond to a long dry period from three to four thousand years ago or that they may be even older.”’ Strong’s esti- mate is more generous, with 8,000 to 10,000 years given as the pos- sible time which has elapsed since level I (the lowest) was inhabited.
At Lynch (Van Royen, 1937, p. 638), the archeological remains cover a considerable area on the bluffs just north of Ponca Creek. As revealed in excavations by the University of Nebraska, their most striking feature is the presence of “a thick dark stratum, which near the ends of the [| University test] trenches was found to lie about a foot beneath the surface and which was covered near the center by eight feet of sand... .’ These sands, according to Van Royen, are wind-deposited, and were derived from the water-laid Pleistocene sands on the high terrace north and northwest of the areas of accumulation. The topsoil to a depth of 12 to 18 inches has been colored gray by plant matter and still supports a fair stand of grass. Wind activity here today is not great enough to produce dunes or sand drifts except where the grass cover has been killed off by cultiva- tion or by overgrazing. The gray topsoil is too deep to be accounted for by the few decades which have elapsed since introduction of the plow in the district, and “since the culture stratum does not show any influence of the white man” it is suggested that the period of pronounced sand movement antedates the coming of Europeans. The observed conditions would imply a prolonged period of lower
NO. 3 CENTRAL GREAT PLAINS—-WEDEL 7
rainfall and destruction of the grasses, followed by increased wind action on the denuded ground surface.
The question of dating even approximately the Indian occupation at Lynch, and through this the drought which must have followed it closely, hinges very largely on accurate identification of the archeo- logical materials in the dark stratum. A detailed report on these has not yet appeared, but certain generalizations can be ventured on the basis of sherds collected on the surface during several visits I made before and at the time of the excavations (Wedel, 1940b, p. 317; see also Van Royen, 1937, p. 647). The sandy overburden has blown extensively wherever modern cultivation is under way, this being especially true on fields a few hundred yards east of the diggings. On the denuded village surface there were abundant remains, and hearths could be found only a few inches below the plowed topsoil. A collection of several hundred potsherds including numerous rim pieces, as well as many end scrapers, projectile points, and other chipped forms indicated an interesting mixture of types. One group of sherds exhibited features characteristic of the Upper Republican horizon ; others, including a few rim pieces with handles, were remi- niscent of Nebraska culture remains. A third group, in which shell tempering was noted, included incised or trailed decoration, rims, and handles resembling in most particulars the Oneota wares of the upper Mississippi and Missouri Valleys. A few sherds bore parallel ridges on their exterior surfaces, apparently produced by the same paddling technique used so widely by the Pawnee, Mandan, Arikara, and other tribes in protohistoric and historic times.
The Upper Republican and Nebraska culture manifestations throughout the central Great Plains have been extensively worked, and in no case has iron, glass, or other evidence of contact with white men been noted. On the other hand, Oneota village sites in Iowa, Missouri, and Kansas have yielded small amounts of such material. All the available evidence indicates that in the Missouri Valley and westward, the Oneota remains are late. Some of the sites may ante- date slightly the arrival locally of white men, but the strikingly uniform character of the remains over most of the area occupied would indicate that they were not spread over a very long period of time.
It is not clear whether all the several pottery types noted at Lynch occur together or whether there was a stratified succession of wares. It is possible that a late phase of the Upper Republican survived here for a time alongside an unclassified peripheral variant of the Oneota
18 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
manifestation. In any case, the fact that pottery with definite Oneota affinities underlies the sand deposits is strong evidence that the latter were laid down within the last three to five centuries.
Much less striking than the sand accumulations at the Lynch site or the stratified remains at Signal Butte is the occurrence at many sites in the central Great Plains of a soil cover equally suggestive of dry, windy conditions. In a great many localities soil profiles have been partially obliterated by modern agricultural activities, but here and there in stream terraces can still be seen evidences of a dark humus zone buried under to to 30 inches of fine light-gray loess, without doubt wind-laid. That this old humus zone was at one time an inhabited surface is indicated by archeological observations. In the Republican drainage of southern Nebraska, as for example, on Medicine and Lost Creeks, Upper Republican pit-house sites have been found excavated into the humus line, with potsherds, bones, and similar village refuse littering the same level (Wedel, 1934, pp. 149, 152, 154; Strong, 1935, p. 76). That these houses were not dug from the present surface is shown by the presence of culturally sterile wind-blown materials which extend evenly and uninterruptedly across the old house basins and over the adjoining detritus-strewn humus zone. The present surface of this soil cover is usually flat with no suggestion of dunes or drifts. The material itself is finer than that at Lynch, and superficially resembles rather closely the sterile layers separating culture horizons at Signal Butte.
Upper Republican village sites blanketed in this fashion occur throughout the Republican River basin in southern Nebraska from Frontier, or possibly Hayes, County eastward at least to Webster County. I have observed a similar though thinner unbroken cover- ing on Upper Republican sites in the lower valley of the North Loup River (pl. 4, fig. 2). Since widely scattered sites are thus involved the factors responsible must have been of more than local magnitude. At the same time, it must be pointed out that other Upper Republican sites in the Loup drainage are marked by shallow surface depressions, indicating either that less soil was deposited over them or else that the houses and caches were dug through the cover- ing material and are thus later. The villages occupied by the Pawnee on the lower Loup and Platte Rivers seem never to have been thus buried, and the old lodge circles were always clearly visible before their obliteration by the plow. Still farther east, along the Missouri, the Nebraska culture sites which are believed to have been occupied synchronously with some of the Upper Republican villages, charac-
NO. 3 CENTRAL GREAT PLAINS—WEDEL Ig
teristically show deep, well-marked house pits where they are not under cultivation.
It has already been noted that Woodland sherds and stone artifacts have been found in western Kansas and in Nebraska in similar buried humus strata. Wherever Upper Republican and Woodland remains occur on the same location, the latter are always at the bottom. For example, near Healy, in Lane County, Kans., Upper Republican village remains (pl. 5, fig. 2, stratum A) occur just below the wind-eroded surface of several small terraces. Below, and separated by a few inches to nearly 2 feet of sterile gray soil (pl. 5, fig. 2, stratum D), is a dark-gray stratum yielding hearths and Woodland artifacts (pl. 5, fig. 2, stratum B)." The number of known similar occurrences in the Republican, Loup, and other more northerly river valleys is increasing.
Insofar as they relate to prehistoric man, the dust deposits of the Republican Valley and adjacent areas, as just described, have not been closely studied by physiographers or geologists. It should be noted that the repeated dust storms of the past decade accompanying droughts which wiped out the corn crop in many localities have not produced comparable formations in the Republican and Loup Valleys, though elsewhere deep drifts and dunes have been formed where fences, hedges, and other obstructions tend to break the wind. It is not certain that these deposits result from a single short, intense drought such as that evidenced at Lynch. It has been suggested to me that they can more reasonably be interpreted as a gradual accumulation over a period of many years. At the same time, the fact that the dust covers a dark humus stratum which often contains archeological remains would seem to indicate that a period of fairly rapid deposition followed a more humid interval which lasted long enough to produce a vegetative cover and to become the home of sedentary farming peoples. In other words, I see no reason why the different strata cannot be viewed as evidence of climatic fluctuations analogous to those inferred from the findings at Signal Butte.
Carefully controlled studies in the past decade have made it possible to arrange the major archeological horizons of the central Great Plains in sequential order (see summary in Wedel, 1940b). Where evi- dences of severe drought, or of prolonged periods of subnormal pre- cipitation and consequent increased soil deposition, are definitely linked with these horizons it may be possible to determine the ap-
“What is possibly a third occupation zone here is indicated as stratum C in plate 5, figure 2; from it came only broken animal bones and fire-cracked stones, hence the horizon remains unidentified.
20 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
proximate time involved. At Signal Butte, as already indicated, the dry periods have been tentatively assigned an antiquity of several millenia. At Lynch, on the other hand, the archeological evidence points toward a very much more recent drought—one that might have occurred as late as the sixteenth or seventeenth century.” The dust blanket on the Upper Republican sites in southern Nebraska must have begun to accumulate at least 400 years ago, since there is evidence that the Pawnee were living in large villages in eastern Nebraska in Coronado’s time (1541) but none whatever that contemporary horti- cultural earth-lodge-using peoples lived in the western plains.
The estimates of age here given are inferential, and rest on archeo- logical and historical data. At present they cannot be checked by pre- cise methods such as dendrochronology offers in the Southwest. How- ever, as already indicated, red cedar (Juniperus virginiana) is widely distributed throughout the High Plains, and there is good evidence that western yellow pine grew scatteringly eastward to the goth meridian as late as the nineteenth century. Moreover, it should be borne in mind that charred pine and juniper have been found in
°M. E. Kirby, of the U. S. Engineers office at Omaha, Nebr., has called my attention to relevant data by Upham (1895, p. 504) concerning fluctuations in the surface levels of certain glacial lakes in North Dakota. Devils Lake and nearby Stump Lake in the northeastern part of the State present highly irregular out- lines and are believed to occupy the valley of a preglacial river which has been elsewhere buried with drift. In historic times Devils Lake reached its highest level about 1830. At about the same time occurred record high-water levels in Red River and in the Great Lakes. Thus, 1830 would appear to represent the high stage in a period of heavy precipitation. The shore line of Devils Lake at that time is indicated by a line of heavy timber. Between this and the recent shore line Upham noted a stand of smaller trees which in 1889 showed a maxi- mum of 57 annual growth rings. During the 1830 high stage the waters flowed into Stump Lake a few miles to the southeast, but the latter never attained the same level as Devils Lake owing to evaporation. The early postglacial outlet of Stump Lake into the Sheyenne River was dry during the 1830 high water, as shown by a stand of large timber growing across the channel. This timber corre- sponds to that marking the 1830 water line in Devils Lake.
At present, in the bed of Stump Lake, North and South Washington Lakes, and Lake Coe, all situated near Devils Lake, there are old stumps of trees which grew when the lakes were dry some time before the 1830 rise. Many of these have been uprooted and used for fuel. Some of the stumps on the lake bed showed as many as II5 annual rings, indicating well over a century of deficient rainfall. It is not known just when the protracted period of desiccation here indicated came to an end, but it must have been long before 1830, since many years of heavy rainfall would be required to refill the lakes. Upham suggests that the drought represented here may have coincided with the arid conditions in the Great Basin which are supposed to have dried up Pyramid, Winnemucca, and other lakes in Nevada about 300 years ago.
NO. 3 CENTRAL GREAT PLAINS—WEDEL 21
Upper Republican pit houses (Wedel, 1935, p. 170) ; that the former distribution of pine and the present occurrence of juniper overlap the known range of the prehistoric Upper Republican horizon; and that much of the area of overlap in the western plains has a low rainfall which is apparently directly reflected in tree growth. Work- ing on wood specimens from historic log structures and from old stumps in gully fills in Lincoln County, Nebr., Weakly (1940) reports a continuous tree-ring sequence reaching back to about A. D. 1480. The buried material is said to have given “very readable ring se- quences,’ but absolute dates have not yet been assigned. All this raises the hope that continued research, combining dendrochronology with archeology, will soon produce exact tree-ring datings for some of the late prehistoric and protohistoric culture horizons in the western plains, and also establish the time of some of the associated drought evidences in the same region.
In another paper (Wedel, 1940b, p. 329), I have suggested the possibility that abandonment of the western plains by sedentary horti- cultural peoples in late prehistoric times may have been due in part to inability to cope with drought conditions. This view has been questioned by ecologists with whom I have discussed the point. Their contention is that the small gardens of the Indians, unlike the present- day farms, would have been situated only in sheltered bottom-land pockets where there was maximum protection against hot winds and where natural drainage conditions would have provided subirrigation. Moreover, the practice of storing one to several years’ supply of corn against the contingency of crop failure would have carried them through droughts such as those of the historic period. In this con- nection a perusal of the reports of the various Indian agents in the Kansas and Nebraska territories is instructive.
The effects of the droughts of 1860, 1870, and 1893-96, in terms of large-scale populational movements out of the plains by white set- tlers, have been frequently recounted.’ Their effects on the native and transplanted Indian populations, however, seem to have gotten little attention. According to the report of the Commissioner of In- dian Affairs for 1860, the Pawnee (pp. 94-95) ‘“‘had about 800 acres of corn, pumpkins, beans, etc.; but owing to the extreme drought in this section of the country, and improvident farming, their crop is very light.” Among the Oto and Missouri (ibid., pp. 96, 97) a highly favorable spring was followed by three rainless months, with “a con- stant burning sun and scorching wind, the result of which is the
® Summarized in Clements and Chaney, 1937, p. 41.
22 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI entire loss of their crops... .” Even the grass was burned up, so that neither the Indians nor the agency were able to lay up their usual reserve of hay for the winter. Along the Nemaha, the Sac and Fox (ibid., p. 99) fared a little better, with the prospect of half a corn crop and considerable hay. The Kickapoo (ibid., p. 101) in northeastern Kansas who had just been presented with oxen, plows, and other farm machinery, were giving up their farms consisting of “8 to 10 acres of soft turfless ground among the girdled trees skirting the narrow borders of the creeks’ and were venturing out into the prairies. For them, 1860 brought the worst drought of their 30-year sojourn west of the Missouri, and a complete crop failure. The Potta- wotamie (ibid., p. 41), the Sac, Fox, and Ottawa (ibid., p. 111), the Kansa (ibid., p. 113) together with the Kaskaskia, Peoria, and other tribes on the Osage River in eastern Kansas were equally hard hit. Total or partial crop failures are also reported for the Choctaw, Cherokee, Chickasaw, Quapaw, and Wichita (ibid., pp. 114-123) far to the south in the Indian Territory. According to the agents, the alternatives confronting the natives everywhere throughout this vast region were starvation, government aid, or recourse to the plunder- ing of such of their white neighbors as were receiving relief from the government or from friends in the east.
A decade later, in 1870, drought struck again, but this time it af- fected the Indians over a much smaller area. At the Whetstone Agency, Dakota Territory, the crops were destroyed by a 2-month drought in June and July. At the Santee Agency, in northeastern Nebraska, 370 acres of corn were a total loss. The agent for the Ponca wrote that “Had it not been for the very long and severe drought, they would have realized a yield of at least 14,000 bushels. On the first of July their crops looked well, and I was of the opinion that the bottom lands would not be affected by the drought; on the contrary, the whole was an entire failure. . . . This tribe is now bor- dering on starvation.” In southern Nebraska, of the Oto it is said that “On the bottom lands a few will probably succeed in harvesting light crops of corn; but I fear those ... on the upland... will experience an almost entire failure of their crops... .” Through- out northeastern Kansas the tribes generally seem to have fared very well. Weather records show that Leavenworth, Kans., enjoyed an excess precipitation of nearly 25 percent in 1870, which suggests that the reservation Indians hereabouts may have been out of the main drought area.”
1 References to insect pests are rare in these reports. In 1870, however, the agent for the Yancton Sioux suggested (p. 212) the advisability of removing
NO. 3 CENTRAL GREAT PLAINS—WEDEL 23
Prior to 1860, data on drought conditions in the central Great Plains are scarce. We may note, however, that in 1848 the Council Bluffs Agency reported that “The government has purchased the past season for the Pawnees between 1700 and 1800 bushels of corn, to keep them from suffering. . . . They are still in a miserable condi- tion; their crops this season have almost been an entire failure, ow- ing to the drought. Their corn in the Platte bottoms was literally burned up; . . . they will suffer, unless they make an unusual good hunt this winter.”
The distressing conditions set forth in these reports are a sig- nificant commentary on the reactions of Plains Indian farming com- munities to drought. It can be objected that at this late date the old cultural patterns of such tribes as the Pawnee, Omaha, Oto, and Ponca were swiftly falling apart. Their horticultural econ- omy was only a part-time interest, and probably but a dim reflec- tion of what it evidently had been a hundred or two hundred years before. Moreover, tribes like the Kickapoo, Sac, Fox, Ottawa, and others, were originally residents of more easterly drought-free areas who had been moved onto reservations west of the Missouri so re- cently that there was insufficient time or incentive to make the neces- sary adjustments to a changed environment. All this is true. But it is also true that the area inhabited at this time represents the best corn-growing section in the central Great Plains, climatically and otherwise. It must be remembered, too, that there was no con- temporary Indian agriculture beyond the g9th meridian in the High Plains, and the above observations contain about the only extant data concerning such activity under pronounced drought conditions beyond the Missouri.
The recorded droughts of the nineteenth century, despite their oc- casionally calamitous effects, were relatively short-lived affairs. Most of them meant only a year or two of subnormal precipitation ; in no case has there been, since the coming of white settlers, as much as a decade of continuous drought and unbroken crop failure. For this reason, it is impossible to say how long or how intense a period of drought would be needed to kill off the sod cover to the point where large-scale soil movements could take place. Clements (1938, p. 202) states that “even a thin cover of vegetation controls the wind so
these Indians from “a climate where crops are so uncertain, owing to the scarcity of rain and the ravages of the grasshopper. . . .” He says further that “ ... in five years of the last ten the crops were totally destroyed by the drought and grasshopper, and in one year of the ten there was about half a crop...”
» 24 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
effectively that soil-drift and deposition during the past must have been limited to bare areas such as ocean-strands and river-banks. During the historical period, dust storms have come only from soils exposed by man in the course of settlement.” The second statement, at least, may need revision, since terrific sand and dust storms ap- parently rivaling in intensity those of recent years were experienced by McCoy and his surveying party in northern Kansas 200 miles west of the Missouri as early as 1830 (McCoy, 1840, pp. 408-409). The material which overlies Upper Republican sites can hardly have been due to man’s work, since it unquestionably antedates the era of modern agriculture and could never have been an aftermath of the small-scale horticulture practiced by the people whose remains it covers.
In his discussion of the Lynch site, Van Royen (1937, pp. 645, 648) expresses doubt that 20 to 25 years of drought would have been sufficient to destroy the local sod cover and inaugurate wind erosion leading to the topographic changes there noted. On the other hand, many farmers in the western plains maintain that 5 years may be enough to kill the grass and start serious soil movements, even where no overgrazing has taken place. On this point, agricultural experts with whom I have talked corroborate the observations of the farmers. In light of these latter statements it would be interesting to know just how long the droughts suggested by the archeological record may have lasted. A possible clue lies in the tree-ring studies already mentioned.
Weakly (1940) has called attention to the presence of aeolian de- posits in canyons in the vicinity of North Platte, Nebr., beneath which red-cedar stumps are buried. Annual rings on these stumps “indicate a period of over 30 years with deficient moisture. Appar- ently this drought period contributed very largely to the death of these trees.’’ Presumably, the fill overlying these stumps is a result of wind action on the surrounding uplands after these had been par- tially or largely denuded of their sod cover. The drought conditions manifested here have not yet been dated. They have significant implications, however, for they suggest that prolonged droughts com- parable to those held responsible for extensive ethnic disturbances in the Southwest in 1276-1299, and again in 1573-1593 (Douglass, 1935, p. 48), might well have occurred in the Great Plains, though not necessarily concurrent.
In addition to this as yet undated major drought, Weakly believes his tree-ring data prove the occurrence of a number of shorter periods of deficient rainfall. Some of these “would have been very
NO. 3 CENTRAL GREAT PLAINS—WEDEL 25
severe even on the present population of this section and . . . would have been a major catastrophe to a population of aboriginal farmers. Several of the drouths were of sufficient severity to very largely de- populate the plains even now. .. .” (Letter of March 7, 1941.) In this same letter, Weakly informs me that the major dated drought periods previous to 1700 are as follows:” 1439-54 (I5 years), 1459-68 (9 years), 1539-64 (26 years), 1587-1605 (except for 1594- 96). I am inclined to suspect that if Weakly’s chronology has been extended beyond question to the early part of the fifteenth century, he may be on the threshold of an exact dating for some of the Upper Republican village sites in western Nebraska.
We have already indicated that the western portions of the Upper Republican habitat have a low irregular rainfall which borders on the minimum required for successful farming. It can be assumed that these peoples farmed intensively and that they were far-sighted enough to lay by seed corn and food against a year or two of crop shortage. At the same time it may be doubted that their harvests were ever on the scale of those normally enjoyed by the later Pawnee and other horticultural peoples farther east. In light of the diff- culties experienced by Indians in the eastern plains during the brief droughts of 1860 and 1870, I am led to believe that when hot, sear- ing winds and droughty summers visited the aboriginal farmers throughout the western Great Plains, crop failures and some measure of destitution were their lot in prehistoric days as well. Dry years may also have given rise to insect plagues, as they do today, further complicating the native economy. The delicate balance between annual precipitation and crop yield has been noted elsewhere in this paper. If drought conditions recurred for several successive years, or if there was a drop of several inches in the average annual pre- cipitation over a period of Io or 20 years or more, perhaps with springs and watercourses drying up, there would have been no choice for the natives other than that of abandoning their villages and re- moving eastward to better-watered and more dependable regions. A sedentary mode of life such as that indicated for the Upper Re- publican peoples, involving relatively permanent earth-lodge villages, would have been impossible when climatic or other factors prevented the cultivation of maize. All this leads me to suggest again that the early horticultural peoples of the Upper Republican communities may have been forced out of their habitat in western Kansas and Nebraska, probably toward the east, in large part by a long-continued
* A manuscript detailing the results of Mr. Weakly’s tree-ring studies is awaiting publication by the U. S. Department of Agriculture.
20 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
period of subnormal rainfall, the outstanding physiographic after- math of which is seen in the dust which today covers many of their ancient living sites.”
“The relation between the prehistoric Upper Republican peoples and the historic Pawnee is still a puzzle. There is nothing in Pawnee traditions to sug- gest that this group was a late arrival west of the Missouri, and it is generally believed that they were firmly established in or very near their historic locale at the time of Coronado’s march into the plains. Strong (1935, p. 277) has suggested that the Pawnee may have been the lineal descendants of the Upper Republican peoples. This is based on a number of resemblances involving basic house types, pottery, and the general semisedentary horticultural mode of life. Persistent search has so far failed to produce a single site in Nebraska which could be regarded as intermediate between the Upper Republican horizon and the earliest protohistoric village remains attributable to the Pawnee. On the contrary, the small, open, undefended villages, prevailingly rectangular pit houses, cord-roughened pottery, and communal ossuary burials of the Upper Republican peoples are consistently in contrast to the large, defensively situated, fortified towns, invariably circular earth lodges, corrugated paddle pottery, and individual flesh interments of the Pawnee. The much sought “clear unbroken line of ceramic and other development” foreseen by Strong is not yet at hand, nor can we say how long the gap in the proposed sequence is. The data on physical anthro- pology, which may well be crucial in this connection, are either nonexistent or unpublished.
Recent excavations (Wedel, 1941) in Rice County, Kans., have shown that small but consistent amounts of Upper Republican-like cord-roughened pottery occur here, along with puebloan sherds of ca. 1525-1650, at village sites tenta- tively ascribed to the Wichita. There is also some evidence of ossuary burial. Strong (1940, p. 382) has recently proposed that the Arzberger site near Pierre, S. Dak., represents “a late prehistoric horizon, basically Upper Republican, but in process of development into the more specialized and later protohistoric Pawnee (to the south) and Arikara (in the north).”’ As I have indicated in this paper, there is also a strong possibility that a late phase of the Upper Republican culture survived into virtually protohistoric times in northeastern Nebraska, as shown by the pottery remains at Lynch. In other words, traits which may be regarded as of Upper Republican derivation appear to have survived later in the eastern plains and in South Dakota and Kansas than in the Upper Republican- Pawnee region in Nebraska. One wonders, therefore, whether the postulated development from a prehistoric into a historic entity, or entities, may not have taken place outside the Nebraska area, with the Pawnee on the Loup and Platte Rivers representing a backwash. According to their traditions, the Pawnee entered Nebraska from the east and south rather than from the north, which might explain certain ceramic (as for example, the cloistered rims) and other elements in their material culture which have an easterly rather than Upper Republican flavor. This is admittedly a thorny problem and one which cannot be certainly answered with the information now at hand. As a working hypothesis, however, it may be well to bear in mind the possibility that the threads running from the Upper Republican to the Pawnee, if they exist, may have to be traced out of Nebraska and then back at a somewhat later period, which, in any case, should precede 1541.
NO. 3 CENTRAL GREAT PLAINS—WEDEL 27
As has been pointed out elsewhere, the occurrence of dust-covered humus strata containing human debris is more frequent in the central Great Plains than has commonly been supposed. Moreover, to the best of my knowledge this does not involve the remains of such his- toric tribes as the Pawnee and their Siouan contemporaries whose villages and campsites seem always to have been situated on the present ground surface. If the similarity between Signal Butte and other later stratified or “buried” sites can be taken as evidence of like climatic fluctuations, then it would appear that the record of man’s activities in the western plains is linked with the periodic develop- ment of stable humus zones between which there are dry-weather dust deposits. Thus, in the archeological record we have Signal Butte I and Signal Butte II, each succeeded by periods of unde- termined duration during which dust was being laid down. Later came pottery-making groups: First the Woodland, then the patently horticultural Upper Republican, each followed by periods of in- creased deposition. There is at present no way of determining accurately the length of time required for the development of these now-buried humus zones on which prehistoric farming peoples once carried on their everyday activities over much of the central and western Great Plains. It seems to me, however, that these in- tervals must have been of some length since the human occupancy spread westward about as far as climatic factors would permit corn growing. In any event, if the repeated interludes of deposition were indeed the aftermath of decreasing rainfall and increasing wind activity, as contrasted to the more humid periods which produced the humus layers, we may visualize the farming Indians as having ventured far out into the Great Plains during favorable times only to withdraw when droughts set in.
CONCLUSION
It need not be assumed from the findings of archeology that any major climatic change or permanent desiccation has taken place in the central Great Plains within the span of time represented by the various aboriginal pottery-making groups formerly resident there— or, as a guess, within the past 8 or Io centuries. In all likelihood, however, shorter or longer periods of deficient rainfall have oc- curred repeatedly in prehistoric as in historic times. Some of these periods were probably of sufficient duration or intensity to depopu- late the western plains for a time. The occupation of the region by migratory hunters since at least 1541 may well be due to the acci- dents of history as much as to any unfavorable climatic trends. That
28 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
is to say, it is conceivable that had the hunters first seen by Coro- nado’s party not received horses and firearms from the whites, maize- growing peoples might again have extended their occupation west- ward beyond the g9th meridian. In view of the still sketchy nature of our information on plains prehistory, it is impossible at this time to evaluate fully the respective importance of environment and of historical accident on the alignment of native economies in the region. Nevertheless, it begins to look as though alternate settlement and abandonment was true of primitive man’s occupation of the western plains just as it has characterized the subsequent white man’s tenure where large-scale government aid was not forthcoming during periods of adverse climatic conditions.
LITERATURE CITED Bessey, C. E. 1896. Were the Sandhills of Nebraska formerly covered with forests? Publ. Nebraska Acad. Sci. V, Proc. 1894-1895, p. 7. Carson, G. G., and Jonss, V. H. 1940. Some notes on uses of plants by the Comanche Indians. Pap. Michigan Acad. Sci., Arts, and Letters, vol. 25, pp. 517-542. CHAMPE, JOHN L. 1936. The Sweetwater culture complex. Chapters in Nebraska Archeol., vol. 1, No. 3, pp. 249-299. Lincoln. CLEMENTS, FREDERIC E. 1938. Climatic cycles and human populations in the Great Plains. Sci. Monthly, vol. 47, No. 3, pp. 193-210. CLEMENTS, F. E., and CHaney, R. W. 1937. Environment and life in the Great Plains. Carnegie Inst. Washington, Suppl. Publ. No. 24 (revised ed.). Comes: 1938. Correlations between annual precipitation and the yield of spring wheat in the Great Plains. Techn. Bull. No. 636, U. S. Dep. Agr. Douetass, A. E. 1935. Dating Pueblo Bonito and other ruins of the Southwest. Contr. Techn. Pap., Pueblo Bonito Series No. 1, Nat. Geogr. Soc. FENNEMAN, N. M. 1928. Physiographic divisions of the United States. Ann. Assoc. Amer. Geogr., vol. 18, No. 4, pp. 261-353. GitmoreE, M. R. 1913. The aboriginal geography of the Nebraska country. Proc. Mississippi Valley Hist. Assoc., vol. 6, pp. 317-331. 1919. Uses of plants by the Indians of the Missouri River region. 33rd Ann. Rep. Bur. Amer. Ethnol. Harrincton, J. P. 1940. Southern peripheral Athapaskawan origins, divisions, and migrations. Smithsonian Misc. Coll., vol. 100, pp. 503-532. Hussone, E. M. 1896. The yellow pine in the Republican Valley. Publ. Nebraska Acad. Sci. V, Proc. 1894-1895, p. 7.
NO. 3 CENTRAL GREAT PLAINS—WEDEL 29
Kincer, J. B. 1923. The climate of the Great Plains as a factor in their utilization. Ann. Assoc. Amer. Geogr., vol. 13, No. 2. Kroeser, A. L. 1939. Cultural and natural areas of native North America. Univ. California Publ. Amer. Archeol. and Ethnol., vol. 38. Lesser, A., and WELTFISH, G. 1932. Composition of the Caddoan linguistic stock. Smithsonian Misc. Coll., vol. 87, No. 6. McCoy, Isaac. 1840. History of Baptist Indian Missions. New York. Serzier, F. M. 1940. Archeological perspectives in the northern Mississippi Valley. Smith- sonian Misc. Coll., vol. 100, pp. 253-290. SHELFORD, V. E. 1926. Naturalists’ guide to the Americas. Baltimore. SmirH, J. RUSSEL. 1925. North America. Harcourt, Brace and Co., New York. Strone, W. D. 1935. An introduction to Nebraska archeology. Smithsonian Misc. Coll., vol. 93, No. 10. 1940. From history to prehistory in the northern Great Plains. Smithsonian Misc. Coll., vol. 100, pp. 353-394. UpHam, WARREN. 1895. The glacial Lake Agassiz. U. S. Geol. Surv. Monogr., vol. 25. Van Royen, W. 1937. Prehistoric droughts in the central Great Plains. Geogr. Rev., vol. 27, No. 4, pp. 637-650. WEAELY, H. E. 1940. Tree-rings as a record of precipitation in western Nebraska. Tree- ring Bull., vol. 6, No. 3, pp. 18-19, Tucson, January. WEDEL, W. R. 1934. Preliminary report on the archeology of Medicine Valley in south- western Nebraska. Nebraska Hist. Mag., vol. 14, No. 3, pp. 144-166. 1935. Contributions to the archeology of the Upper Republican Valley. Nebraska Hist. Mag., vol. 15, No. 3. 1938. Hopewellian remains near Kansas City, Missouri. Proc. U. S. Nat. Mus., vol. 86. 1940a. Archeological explorations in western Kansas. Expl. and Field-work Smithsonian Inst. in 1930, pp. 83-86. 1940b. Culture sequence in the central Great Plains. Smithsonian Misc. Coll., vol. 100, pp. 291-352. 1941. In search of Coronado’s “Province of Quivira.” Expl. and Field-work Smithsonian Inst. in 1940, pp. 71-74. Marr, G: F, 1922. Indian agriculture at its northern limits in the Great Plains region of North America. 20th Int. Congr. Americanists, Rio de Janeiro. Winsuip, G. P. 1896. The Coronado expedition. 14th Ann. Rep. Bur. Amer. Ethnol., pt. 1
hy ye iy ih
by Ae
| LF hive pit a 7 , ee tag t f i Seq an it, Viel 1 he ee LL ey eae Hah Ne
| t cy
——
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOERNOMny INOZ cose blest
TYPICAL SCENE IN A PLAINS INDIAN HUNTING CAMP OF THE NINETEENTH CENTURY
Photograph by S. J. Morrow, probably made in the upper Missouri Valley about 1870.
“1Zg1 ‘uosyoerf “HM Aq Yydeisojoyg ‘MYSGSN ‘VON3SD YVAN ‘YSAIY dnNOoT SHL NO ‘SNIV1d NYSLSV¥9 SHL 4O FEIN L IWYNLINDILYOHINAS IWOIdAL V ‘SSENMWd SHL AO SOVTNNA 39007-HLaVa
Galdns) cON) sO 110A SNOILOS11IOO SNOANVIISOSIN NVINOSHLIWS
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOLE. 10:1), NO. 3, PL. 3
1. BEAVER CREEK VALLEY IN SCOTT COUNTY, KANS.
The even skyline, scattered trees, and short grass are characteristic of the High Plains. A Pueblo ruin and other protohistoric Indian remains lie at the center of the view.
2, VIEW IN THE VALLEY OF STINKING WATER CREEK, CHASE COUNTY, NEBR.
The remains of a protohistoric village of hunting and farming Indians occupy the flat immediately across the stream.
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 101, NO. 3, PL. 4
1. VIEW IN THE REPUBLICAN VALLEY, FRANKLIN COUNTY. NEBR.
The Dooley site, type station for the prehistoric Upper Republican culture, lies on Lost Creek, just to the left of the loess bluffs.
2. STRATIFIED SITE ON DAVIS CREEK, NEAR COTESFIELD, NEBR.
Twelve-inch dust mantle (A) overlying dark gray humus stratum (B) contain- ing potsherds and other remains of Upper Republican type.
‘uolzeur[dx9a 10] 61 os6ed 9a¢G
"SNVH “ALNNOD ANY] °MH3SRHD LIVS NO ALIS G3alsILvVYyLS “2
‘YSE4N ‘ALNNOD SSVD (GNVIGOOM) 3ALIS AYOWNTIYD Yay 1IVM LV SHLYVSAH GalenNg AWds30
SNOII9D3A1109 SNOANVTIZOSIN NVINOSHLINS
el!
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOLUME 101, NUMBER 4
(Wirt 11 Ait iil
BY
ALES HRDLIGKA
Curator, Division of Physical Anthropology U. S. National Museum
(PUBLICATION 3640) :
. CITY OF WASHINGTON _ PUBLISHED BY THE SMITHSONIAN, INSTITUTION SEPTEMBER 25, 1941
SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 101, NUMBER 4
DISEASES OF AND ARTIFACTS ON SKULLS AND BONES FROM KODIAK ISLAND
(WiTH 11 PLATES)
BY ALES HRDLIGKA
Curator, Division of Physical Anthropology U. S. National Museum
(PUBLICATION 3640)
CITY OF WASHINGTON PUBLISHED BY THE SMITHSONIAN INSTITUTION SEPTEMBER 25, 1941
The Lord Gattimore Press
BALTIMORE, MD., U. S. A.
DISEASES OF AND ARTIFACTS ON SKULLS AND BONES FROM KODIAK ISLAND
By ALES HRDLICKA Curator, Division of Physical Anthropology U. S. National Museum
(WitH 11 PLaTEs)
During the field seasons of 1931, 1932, and 1934-1937 extensive excavations were carried on by the writer and his various field parties, under the auspices of the Smithsonian Institution, in the pre-Russian sites of Kodiak Island and particularly in a remarkable old site on what became known as “Our Point,’ Uyak Bay. The chief object of these excavations was to recover the numerous skeletal remains buried in the deposits and thus obtain light on the nature of the old inhabitants of this anthropologically important region. As a result there were gathered several hundred skeletons or parts of skeletons which showed that long before the island was inhabited by the strain of people found there by the Russians and called by them the “Koniags,” there was an extensive older and also culturally different, pre-Koniag, population. The skulls and bones of this hitherto un- known population showed a number of special features which deserve to be reported apart from the general study of the specimens.
The Pre-Koniags were an oblong-headed, moderate-sized people of much artistic ability, with some Eskimoid features, but related es- sentially to the oblong-headed American Indian. They were the first permanent inhabitants of the island, having settled there soon after the passing of glacial conditions, which, however, was fairly late— probably at about the beginning of the Christian era. They left deep, condensed deposits, in part already concreted, and larded with their own remains and those of animals. They apparently developed—or possibly brought in—cannibalism, and during the latter part of their period of occupancy, the evidence showed, did various curious things with their skeletal remains and especially with the skulls. It was very common to find a skeleton without the skull or with whole limbs miss- ing, individual or even small groups of skulls apart, and spare bones disseminated throughout the accumulations. Some of the skulls and bones, moreover, showed conditions or artifacts which will be described in this paper.
SMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 101, No. 4
2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
DISEASE
If there be excluded one skeleton, in all probability from Russian times, that showed multiple and very marked tuberculous lesions, and cases of arthritis, a rare fracture, or a dental abscess, the numerous skulls and bones from Our Point, as well as those from the pre- White deposits elsewhere on Kodiak Island, are wholly free from disease. This is very remarkable, for the material is both extensive and exhaustive. There is no trace of any of the constitutional dis- eases or distrophies that leave their marks on the skeleton, and there are no tumors. Even fractures are much less common than with us. These findings apply to both the Pre-Koniag and the Koniag peoples.
The arthritic lesions in the elderly were common. They show the usual variety of alterations in the spine and the joints, but “mush- roomhead” femora, relatively frequent in old Peru, do not occur.
The rarity of fractures, even in the ribs, points on one hand to the resistance of the bones, and on the other hand to rarity of such violence as would lead to fractures. There were, however, one or two skeletons in which several of the parts had been badly broken before death—possibly through encounters with bears.
As elsewhere in pre-White Alaska, there were no dental caries; but almost from the beginning of adult life there was progressive wear of the teeth, until eventually in instances the pulp cavity was exposed, infection followed, an apical abscess developed, and the tooth would be lost or rendered useless.
ANOMALIES
Cranial and skeletal anomalies, too, were rare in both the Pre- Koniags and the Koniags. As all through western Alaska, there was in both the Kodiak peoples occasional narrowing of the nasal bones at the root, and in rare instances a metopic suture or some form of an “Inca” bone. There are of course, as in all crania, numerous minor exceptional features, but they are of little importance.
DEFORMATION
A good many of the adult Koniag skulls, particularly males, show a slight asymmetry, probably caused by the subjects in infancy lying habitually on the back more to one side than to the other. Such asymmetries are common in brachycephalic crania everywhere and can hardly be called deformations. They may have had a slight effect on the cranial index in the males, but this could not be material. The Koniags had cradleboards, but used no bandaging of the heads of the infants.
NO. 4 SKULLS AND BONES FROM KODIAK ISLAND—HRDLICKA 3
The case was somewhat different among the Pre-Koniags. We do not know whether they had any form of a cradleboard ; if they had, it would have been made of perishable material which would long since have disappeared. What is certain is that some deformation begins with this people. It was an occasional characteristic slight to moderate occipital compression that raised the parietal part of the vault, leaving the coronal region flat or even with a slight postcoronal depression. The deformation in only a very few cases was enough to necessitate the elimination of the specimen from the measuring ; and fortunately such cases were so infrequent that they did not affect the high value of the collection.
ARTIFACTS
No artifacts were encountered on any of the Koniag skulls or bones, but some were present on those of the Pre-Koniags. They consisted in a few cases of drilled holes, in the skull, lower jaw, a scapula, or a pelvic bone, for the passing of a cord by which the bone or part was suspended; in one case of the insertion of artificial (ivory) eyes into the sockets of a skull; and in one or two cases of trephining. Brief descriptions of the individual cases, supplemented with photographs, follow.
DRILLED SKULLS AND BONES
Drilled skull—U.S.N.M. No. 377738; from the intermediary pre- Koniag deposits, Our Point, Uyak Bay, Kodiak Island; a broken-out portion of an adult, probably male, skull that may have been trephined in the postcoronal region and was bored through at bregma. (PI. 1, fig. 1, lower.)
This specimen, as far as the drilling at bregma is concerned, is an exact counterpart of some of those from Michigan reported in 1875- 1877 by Gillman.” It shows at bregma a clean somewhat funnel- shaped perforation, 7 mm. in diameter dorsally and 4 mm. ventrally. A shelf deep inside the opening indicates that it was made first by a
* Gillman, H., Certain characteristics pertaining to ancient man in Michigan. Ann. Rep. Smithsonian Inst. 1875, pp. 234-245, 1876; La perforation cranienne du Michigan. Bull. Soc. Anthrop. Paris, vol. 11, pp. 435-436, 1876; Cranes perforés du Michigan, ibid., vol. 12, p. 82, 1877; Additional facts concerning artificial perforation of the cranium in ancient mounds in Michigan. Proc. Amer. Assoc. Adv. Sci., vol. 26, pp. 335-339, 1877. See also Fletcher, Robert, On prehistoric trephining and cranial amulets. Contr. North Amer. Ethnol., vol. 5, 1882; Cranial amulets and prehistoric trephining. Trans. Anthrop. Soc. Washington, vol. 1, pp. 47-51, 1882.
4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
smaller and then by a larger drill; and 14 mm. posteriorly to it the bone had been cut across and its upper edge beveled, all the other edges being left raw and irregular. The whole piece may have been a crude disk, worn as a breastplate during some observances, or as an amulet on some occasions.
Plate 1, figure 1, shows the specimen, together with a previously unreported fragment from a mound in Michigan with three similar perforations.
A very good late report on artifacts of this nature was published in 1936 by Hinsdale and Greenman.’ They refer to other post- mortem drilled or cut skulls, summarizing the subject thus (p. 12):
Crania with a single perforation at the vertex, drilled after death, have been found at four sites in southeastern Michigan and at three sites in western Ontario; crania with more than one perforation, or with an unknown number, have been found at one site in Michigan, at seven in Ontario, and at two in Ohio. Circular disks cut from human crania and perforated with two or more holes, are found in Ohio, Ontario, and New York.
In addition these authors report having found artificial perfora- tions in the femora and tibiae of a skeleton. The case is unique and so curious that the details deserve to be quoted; they read (pp. 4-5):
There are also perforations in long bones from the Farmington site. Two femurs and two tibiae are perforated near the ends. Although dissociated from any other parts of a skeleton they apparently belonged to the same individual and were lying in normal articulation. A hole 5 mm. in diameter is in the distal end of the shaft of the left femur, anteroposterior. A hole in similar position on the anterior surface of the right femur was carried only to a depth of about I cm. The drill was started in the opposite surface of the femur to meet this hole but was only carried to a depth of about 1 mm. Another hole had been started in the anterior surface at the other end of this femur, at the distal end of the neck. It is 8 mm. deep. Both femurs have the heads cut so that each head is a disk about 23 mm. wide. At the other ends both condylar surfaces are missing, apparently they were severed (Pl. IV, Fig. 1).
The perforations in the two tibiae are at the proximal ends. Both articular surfaces have been cut off, and both anterior borders have been shaved down I to 2 cm. distal to the tuberosities. The distal ends have also been cut, but the greater part of each articular surface remains. There are two holes through the proximal end of each tibia, one anteroposterior and the other lateral, so that originally they may have intersected in the interior of the shafts, where cancellous tissue has not fallen away. The holes are from 6 to 7 mm. in diameter. The hole in the external surface of the left tibia could not have been more than 7 mm. deep.
? Hinsdale, W. B., and Greenman, Emerson F., Perforated Indian crania in Michigan. Occas. Contr. Mus. Anthrop., Univ. Michigan, No. 5, pp. I-15, 5 pl., 2 figs., 1936.
NO. 4 SKULLS AND BONES FROM KODIAK ISLAND—HRDLICKA 5
The perforations are pictured in plate IV of the above publication. They are small and served, it would seem to the present writer, for fastening together the respective femora and tibiae.
Drilled skull —U.S.N.M. No. 374674 ; a normal Pre-Koniag middle- aged male; marks of knife across the forehead suggest that scalp had been removed artificially. (Pl. 1, fig. 2.)
The greater wing of the right sphenoid shows two smaller drill holes (4 mm. wide), that of the left sphenoid, one larger one (6 mm. wide), doubtless for thongs which attached the lower jaw to the skull, or by which the skull was suspended. The skull was one of several isolated crania laid down in such a way that they formed a right angle in the deposits. It is well preserved and has evidently been handled but little. There is no indication as to why it had received the special attention.
Drilled mandible —U.S.N.M. No. 379244; from Our Point, Uyak Bay ; the normal, well-preserved lower jaw of a Pre-Koniag boy of epoutsta. (Pl. 2, fie. 1.)
In each ascending portion below its middle is a drill hole 6 mm. wide; and a similar although incomplete perforation 5 mm. wide with concave base is seen anteriorly just below the root of the lateral right incisor.
Drilled scapula—U.S.N.M. No. 379249; from Our Point, Uyak Bay; Pre-Koniag adult, sex uncertain; considerably damaged and acromion process evidently hacked off, but nothing pathological or wndsual. (Pl.'2, fig. 2.)
A drill hole 5 mm. wide passes through the body of the spine about 3 cm. from its vertebral end; another similar but incomplete drill hole, with nicely concave base, is seen on the dorsal surface of the bone 3 cm. below the lower border of the glenoid cavity and 9 mm. from the axillary border of the bone.
Drilled sphenoid—U.S.N.M. No. 379250; a portion of the great wing of a normal sphenoid, with a perforation 5 mm. wide near its pteric end; bone marked “black” * (Koniag), but may be displaced wred o( Pre-Koniap)) | \(Pl.3, fig. 1, right.)
The fragment, though found as such, was plainly broken off secondarily. It was the skull when whole that was drilled for sus- pension, as in No. 374674, and not the fragment.
*The term “black” was used throughout the excavations for the Koniag de-
posits, with “red” for the later and “blue” for the older pre-Koniag accumula- tions ; and specimens from the different layers were marked accordingly.
6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
Drilled pelvic bone of a.child—U.S.N.M. No. 379251; right ilium of a small child; bone is marked “black’’ (Koniag), but is probably a displaced “red’’ (Pre-Koniag). (PI. 3, fig. 1, left.)
Near the anterosuperior spine is a perforation 6 mm. wide made from the back forward and very similar to those on the other drilled ilia. There may possibly have been another hole farther on, but the middle portion of the upper part of the ilium has been broken off. The bone is strong and rather flat but shows nothing pathological.
Drilled pelvic bone —U.S.N.M. No. 379248; from Our Point, Uyak Bay ; normal right Pre-Koniag pelvic bone. (PI. 3, fig. 2, right.)
The ilium near its upper border shows two drilled holes, each 5 mm. in diameter. The bone is that of a young adult, rather small but not weak, sex uncertain. It shows no injury and no marked handling. Posteriorly, a short distance below the anterior perforation is seen the beginning of another, consisting of a sharp rim with a smooth, large dotlike hollow in the middle.
Drilled pelvic bone—U.S.N.M. No. 377701; from Our Point, Uyak Bay; right pelvic bone of a Pre-Koniag adult male. (PI. 3, fig) 2: left:)
The ilium shows four borings, three of 8 mm. diameter and one of 6 mm., made without much care from the back forward. Their loca- tion and disposal are shown in plate 3, figure 2, left. They were doubtless made for a vertical suspension of the bone, or possibly even of the pelvis as a whole; but the bone was found alone. The pubic part of the specimen had been broken off, but that might have been incidental. The bone itself is normal and indicates a small presenile male individual.
TREPANATION
In the course of the excavations at Our Point, Uyak Bay, Kodiak Island, the “red” or later pre-Koniag deposits yielded a human skull that shows a form of trepanation.
Such operations have not as yet been recorded from anywhere else in Alaska, but there is an example of the practice in the United States National Museum collections from Kagamil Island, and another from Prince William Sound. Several cases also have been found among remains of related type in the oldest deposits at the mouth of the Fraser River * and in those of the nearby Boundary Bay. Gill-
*Smith, Harlan I. (quoting Hrdlicka), Trephined aboriginal skulls from British Columbia and Washington. Amer. Journ. Phys. Anthrop., vol. 7, pp. 447- 452. 1924. One or two additional specimens recovered since the publication.
NO. 4 SKULLS AND BONES FROM KODIAK ISLAND—HRDLICKA 7
man * mentions in addition a skull reported to him from Santa Bar-
bara, Calif., which showed an artificial perforation in the vault that
may possibly have been made in life. Several cases of skull trephining
are on record or await publication from different parts of the United
States and northern as well as southern Mexico (Zapotec tombs,
Caso; unpublished), besides the many from Peru and Bolivia. The description of the Kodiak specimen follows:
Trephined skull—U.S.N.M. No. 379252; from later pre-Koniag deposits, Our Point, Uyak Bay; skull of an elderly woman, damaged before coming into deposits, breaks in basal parts, lower face miss- ing, otherwise normal. (PI. 4, fig. 1.)
Along the middle of the upper third of the frontal and adjoining portion of the parietals is a smooth elliptical depression 6 cm. long, at maximum 2.1 cm. wide, and up to 4 mm. deep, the result of an ancient operation—incomplete trepanation. The hollow in all prob- ability was made by scraping, long before the death of the woman. There is no indication as to why it was made, and there was a perfect healing.
Trephined skull—Reported with the preceding may be the case of skull U.S.N.M. No. 262170, from Knights Island, Prince William Sound; gift of Dr. F. M. Boyle. (PI. 4, fig. 2.)
The skull is that of a young adult female, probably not very ancient. It shows some disturbance of development—the occiput is bulging and somewhat asymmetrical, with a rather marked supralambdoid “set-back” of the parietals ; the left parietal shows over the eminence a fairly large, old, well-healed lesion, which, however, did not affect the inner wall and so may not have been the cause of the operation ; yet the thickening of the external plate about the lesion is somewhat pronounced especially anteriorly, where it extends nearly to the rear margin of the lower perforation. Aside from these features the specimen is normal.
The postcoronal upper two-thirds of the left parietal of this skull shows two good-sized openings; one of these is surely due to an
° Gillman, H., Certain characteristics pertaining to ancient man in Michigan. Ann. Rep. Smithsonian Inst. 1875, p. 242, 1876.
*See Lumholtz, Carl, and Hrdlicka, Ales, Trephining in Mexico. Amer. Anthrop., vol. 10, No. 12, pp. 389-396, 1897; Cosgrove, C. B., A note on a trephined Indian skull from Georgia. Amer. Journ. Phys. Anthrop., vol. 13, No. 2, pp. 353-357, 1929; Shapiro, H. L., Primitive surgery: First evidence of trephining in the Southwest. Nat. Hist., vol. 27, No. 3, pp. 266-269, 1927; Hrdli¢éka, A., Trepanation among prehistoric people, especially in America. Ciba Symposia, vol. 1, No. 6, bibliography, 1930.
8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
operation, and the other probably is. The upper hole is circular, 2 cm. in maximum diameter, 1 cm. in the rear of bregma and involving a portion of the sagittal suture. The edges of the opening are fairly sharp, its external border is beveled, the internal border but little affected. The opening bears all the marks of an old, well-healed trepanation. The bone behind it externally is somewhat thickened but otherwise looks normal.
The lower lesion is not easy to describe. From above downward it consists of a nearly circular portion, 7 mm. in greatest width, and apparently cut out in that form. Below this the outer wall of the skull has been scraped and in it there has been made an oblong, somewhat obliquely quadrilateral opening 22 mm. long by 13 mm. wide. This now has thin and very sharp borders, communicates with the small circular aperture above, and reaches antero-inferiorly, by what looks like a broad crack, a small focus of perforative osteitis. The inner table about this whole second opening is somewhat uneven, but there is no breaking down of the bone, and the whole wound, less the small osteitic lesion, has long since been well healed.
The face and base of the skull are normal.
Trephined skull—Still another specimen that may well be reported here is U.S.N.M. No. 243974, a normal skull of a Pre-Aleut elderly female from Nazan Bay, Atka Island; collected in the early seventies by William H. Dall. (PI. 5, fig. 1.)
In the anterior third of the right parietal, 3.5 cm. from the coronal and 1.7 cm. from the sagittal suture, there is an oval opening 12 mm. long by 7 mm. wide, with beveled edge all around and some brown discoloration outside of this. The inner surface of the skull about the opening is entirely normal. The bone about the opening shows healing, but little if any restitution.
It seems impossible to diagnose the lesion as anything else but a trepanation, during life, by cutting. But there is no indication as to why the operation may have been performed; the vault of the skull is wholly normal and shows only one possible injury—a moderate- sized, irregular-edged opening in the unusually thin right postero- inferior portion of the frontal, 3 mm. above the sphenofrontal suture, which appears to be accidental—its edges are rough, there is no beveling, and no trace of cutting (Pl. 5, fig. 2). There are many post-mortem defects in the orbits, the maxillae, and neighboring parts, none of which, however, can have any connection with the trepanation.
Peculiar lesions in skull bones of an infant—U.S.N.M. No. 372883 ; from Our Point, Uyak Bay ; two portions of the skull of a Pre-Koniag intant.(CP1+6,)
NO. 4 SKULLS AND BONES FROM KODIAK ISLAND—HRDLICKA 9
Present, a large portion of the right and a piece of the left parietal. Each shows curious lesions that, especially on the left fragment, appear to have been artificially produced, although they may possibly have been due to some natural cause.
The right bone shows a streaked outer surface, due to corrosion by roots. In its lower third is a large, nearly circular lesion that looks as though a disk of the bone had been cut out. The edges of the opening are sharp, and its inner border is naturally beveled.
The part of the left bone presents two lesions, one nearly circular, the other nearly right-angular, with the angle itself rounded. The bone is thin, apparently normal, and without any traces of disease or inflammation. The openings look exactly as though someone had cut out two pieces of the bone, but the edges of the openings are almost knife-sharp, and ventrally there is a distinct beveling about each of the defects.
The specimens are the only ones of that nature found in the Far North. They cannot, it would seem, well be trepanations, but what else they could be seems impossible to decide. They might possibly be defects caused by absorption from within, but it is hard to con- ceive what could cause such odd absorptions.
BREAKING OUT, CUTTING
A broken-out portion of face—U.S.N.M. No. 379245; from Our Point, Uyak Bay; a normal specimen, Pre-Koniag, strong adolescent male. (PI. 7.)
The joint maxillary portion of the face has intentionally been broken out at the level of the maxillozygomatic sutures, and the nasal septum, the turbinated bones, and the nasal wall of each antrum were removed; but the good denture, not fully complete, and the whole front of the piece, as well as the palate, were fully preserved. It is difficult to understand the motives behind the making of this piece, or for what purpose it could have served; but it is surely not accidental.
Trepanation on the humerus—uU.S.N.M. No. 332608(6) ; from Point Hope (old burials), Alaska; normally developed medium- strong right male humerus, 32.3 cm. long; the lower end shows marked arthritic changes, with upper end unaffected. (PI. 8, fig. 1.)
Eight cm. from the upper end of the bone, externally, at the upper limit of the deltoid ridge, there is a clean-cut somewhat lozenge- shaped opening, 12 mm. long by 8 mm. wide. The operation was done in life, was uneventful, and was followed by perfect healing of the edges. There is not the slightest fracture nor any sign of inflamma-
Io SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
tion about the locality ; yet the shape, size, and perfect edges of the opening can leave no doubt as to its representing a planned and successful operation.
This is the only known example of a trepanation of a long bone from the Far North, or from anywhere else in the Americas.
Cutting away of the head of a femur—vU.S.N.M. No. 372822; from Our Point, Uyak Bay ; normal right moderate-sized femur from a Pre-Koniag pre-middle-aged female. (PI. 8, fig. 2.)
The head of the bone has been cleanly cut away post mortem with a sharp stone knife. The cut is so neat that it must have been made while the bone was still fresh. This is the only specimen of this nature found in the Kodiak excavations, or so far in Alaska, and there is nothing to indicate the object of the cutting. The same de- posits, however, yielded several cups each made out of the head of the humerus of a bear, and just as the people made “baby” stone lamps, they may also have made a “baby” bone cup. It may, of course, have been used as a handle to some tool, or for still another purpose.
SKULL WITH IVORY EYES
One of the first outstanding specimens recovered from the deposits at Our Point, Uyak Bay, was the skull, less lower jaw, of a middle- aged Pre-Koniag. The orbits of this skull had originally been filled with some organic mass, possibly gum, which eventually left but a dark, loose residue ; and in this mass in each orbit was fixed an ivory eye, with pupil of a black stone. (PI. 11, fig. 1.)
The specimen was found in the very first week of the work on Kodiak. It lay with its face down in what were designated the ‘“‘red” or later pre-Koniag layers. It was alone, without the lower jaw or any other part of the skeleton. Without anticipating anything unusual I carefully took it out, saw the orbits filled with what looked like nearly black “dirt” and began to shake this out, when there dropped into my hand an ivory eye; and at the same time Dr. Rich, who was - looking on, called my attention to something white that dropped out of the other orbit to the ground, and which proved to be the mate of the first eye.
Plate 11, figure 1, shows the two eyes in their natural size. Nothing of such nature has ever come out of the deposits before or since, and nothing similar has ever been known from any part of Alaska before; recently (1940), however, some related finds of this nature were made by Rainey in an interesting new site at Point Hope.’
* Rainey, Froelich, A new form of culture on the Arctic coast. Proc. Nat. Acad. Sci., vol. 27, No. 3, pp. 141-144, 1041.
NO. 4 SKULLS AND BONES FROM KODIAK ISLAND—HRDLICKA EE
The Uyak skull was a normal specimen. It had evidently been handled, so that some of the teeth were lost. It may have been the head of some outstanding hunter or chief or a distinguished enemy ; in the first case it would have been a memorial, in the latter a prized trophy.
SKULL CUPS, BOWLS, AND DIPPERS
The Pre-Koniags of Uyak Bay, Kodiak Island, used a part of the human skull occasionally as a cup, bowl, or dipper. They broke out a suitable piece and trimmed it more or less to suit their purposes.
All these specimens came from the “red” or later pre-Koniag layers of the deposits. At this period the people were certainly not shy of human bones; they practiced cannibalism and must have done all sorts of things with skulls and bones as shown by the skeletal remains in the deposits.
Skull bowls were used, it is known, in Tibet, and may have had a wider distribution in eastern Asia. The practice, in fact, was once widely distributed over the world, even among peoples of the white race, and was very ancient, dating back to at least the Upper Paleo- lithic (Magdalenian, Solutrean) .*
In America, in the words of Friederici,’
the skull cup may be followed over the whole continent. It is found in the region of the great Canadian lakes, in Massachusetts and probably in Carolina, in Sinaloa and Michoacan, in Darien and in the Antilles. In South America it is met with in Venezuela, in the realm of the Incas, among the Maynas, Mochos and in the Amazon region. Among the Chaco tribes it was especially reported in the Abipones, Tobas, and Macobies, while the Matacos drank from cups made of the scalps. Especially good information in this respect exists about the Araucano, among whom the cups from the skulls of the Governors Valdivia and Loyala were their most prized trophies.”
®* See Breuil, H., and Obermaier, H., Cranes paléolithiques fagonnés en coupes. L’Anthrop., vol. 20, pp. 523-530, 1909. See also Kuhn, A., Namen von Ge- fassen, Namentlich von Kochgefassen. Zeitschr. Ethnol., vol. 9, p. 480, 1877.
® Friederici, G., Skalpieren und ahnliche Kriegsgebrauche in Amerika. Inaug. Diss., p. 96, Braunschweig, 1906.
*Tdem, bibliography. There is here, regrettably, some confusion between cups made of human skulls and cups or vessels of clay in the form of skulls. Thus Chase, Henry E., Notes on the Wampanoag Indians, Ann. Rep. Smithsonian Inst. 1883, p. 904, says the Wampanoags “have large drinking-cups made like skulls”; and Holmes, William H., Ancient pottery of the Mississippi Valley, 4th Ann. Rep. Bur. Ethnol., p. 407, referring to the prehistoric industry of Arkansas, speaks of “vessels imitating the human head.”
I2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
In South ‘America, particularly in Brazil and Chile, cups or bowls made of human skulls were reported, according to Vignati,” from the Guaycurie, Abipon, Chiriguano, Toba, and Araucano-Mapuche Indians. The making and the use of such cups was generally associ- ated with war; the skulls used were those of trophy heads, and the drinking from such cups was more or less of a ceremonial nature.
The incidence of cups or bowls made from human skulls in North America is as yet not sufficiently clear. Specimens of this nature could readily be passed over by the finder, or be mentioned where the physical anthropologist would not be likely to see it. An example is the brief report by E. R. Quick on “A Prehistoric Cup Made from a Human Cranium,” from Cedar Grove, Ind., published in the Janu- ary 1881 number of the Journal of the Cincinnati Society of Natural History. The “cup or bowl was made of a child’s skull,’ and accom- panied an old man’s skeleton. In this case,
the base of the skull has been roughly cut away and scraped smooth, leaving an irregular margin or rim to the vessel. Both the inner and outer surface has been scraped with some rough-edged tool, leaving numerous scratches. Two holes were drilled through the side, near the upper part of the cup, for the purpose of mending a crack by tying the fractured parts together.
References to similar practices may doubtless be found in the vast ethnographic and archeological literature on the North American ab- origines. Dr. J. R. Swanton, of the Bureau of American Ethnology, has given me the following information:
In my own notes I find but one reference and that merely states that as a war trophy besides the scalp they sometimes took the upper part of the skull. This is in Lawson’s History of Carolina, Raleigh edition, page 321. It does not state that the part so taken was used definitely as a cup. And, in Pierre Margry’s great work, “Découvertes et éstablissements des Frangais dans |’ouest et dans le sud de l’Amérique septentrionale,” volume 5, page 96, some of the northern Indians threaten to drink out of the skulls of their captives.
The specimens in the United States National Museum are as follows:
Skull bowl—U.S.N.M. No. 379243; from Our Point, Uyak Bay; the rear half of a normal skull of a “blue” (deep) Pre-Koniag adult female, shaped artificially though crudely into what evidently served as a cup or a bowl. (PI. 9, lower.)
The specimen now has a hole near the center caused possibly by one of my boys’ picks, though it may have been made by the people
“ Vignati, M. A., Los craneos trofeo. Arch. Mus. Etnografico, No. 1, pp. 93, 118-120, 153, 1930.
NO. 4 SKULLS AND BONES FROM KODIAK ISLAND—HRDLICKA 13
“killing” the cup after the decease of the owner, as they habitually did with the stone lamps and other articles. The blow was delivered from the outside, causing about the wound characteristic peelings of the inner table.
Skull bowl—U.S.N.M. No. 377738; from Our Point, Uyak Bay ; skull of a Pre-Koniag female. (PI. 10, lower.)
The rear third of the parietals and the occipital squama were cut off and rounded to a moderate-sized bowl 10 cm. wide by 12 cm. long. The ventral ridges of the occipital were leveled to make the interior of the bowl somewhat more even and, incidentally, to give a better thumb-spot for handling the vessel. It would have been very useful to drink from, or to contain more or less liquid food.
Skull dipper —U.S.N.M. No. 379247 ; from Our Point, Uyak Bay ; occipital bone of a Pre-Koniag, probably an old male. (PI. 9, upper.)
The borders of the bone were trimmed all around, and the right lower end was left protruding somewhat to form a convenient handle. The specimen could have been used as a shallow cup from which to drink water or as a ladle for dipping out more or less liquid food. It shows no special marks, but considerable handling.
Skull dipper —vU.S.N.M. No. 379246; from Our Point, Uyak Bay; left parietal of a Pre-Koniag adult. (PI. 10, upper.)
The specimen had been trimmed all around to form a dipper, the lower left corner being left somewhat protruding for a handle. There is nothing extraordinary about the bone itself, and it bears no marks ; but it had been broken (“killed” ?) and some pieces were lost before it came into the deposits.
Skull dipper from Aleutian Islands —vU.S.N.M. No. 379253; rear portion of the skull of a sub-middle-aged male, probably Pre-Aleut, from one of the mummy caves on Kagamil Island, cut into the form of a large dipper or a bowl. (PI. 11, fig. 2.)
The piece consists of the occipital, a small part of the right and large part of the left parietal, and the left temporal, with a portion of the base which served for a handle. The parietals and temporal were cleanly cut by a very sharp knife. The whole formed a spacious bowl, and the cut edges are worn, showing considerable use. The dish had evidently been used only for liquids, for its surfaces show no abrasion. The rostrum of the stella turcica had been removed, prob- ably to facilitate holding, and a part of the squama of the cut left temporal has been broken off, either during or after the deposition of the specimen in the cave.
I4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IOI
SUMMARY AND CONCLUSIONS
This paper calls attention to the limited pathology and grosser anomalies, as well as the various artifacts, shown by the skeletal remains of the pre-Russian and particularly the Pre-Koniag people of Kodiak Island, and reports a few related specimens from southwest Alaska.
The Kodiak remains show absence of all constitutional diseases that leave their marks on the skeleton, with the exception of senile arthritis.
The artifacts, limited largely to the later period of the Pre-Koniag people, include some slight grades of head deformation ; drilling of skulls or bones for suspension as charms or trophies; trepanation ; the cutting of bones or breaking out of parts for some utilitarian or other purpose ; and the making of cups or bowls from human skulls.
An apparently significant fact is that most if not all the drilled bones are male, and the three trephined crania all female.
There are here presented thus, from the Far Northwest, a whole series of observances with the human skull and bones previously un- known from that area. As excavations in those regions proceed, this field will doubtless be further enriched. It shows the existence among the peoples of pre-White western Alaska of considerable ceremo- nialism which, together with mummification, sacrifice of slaves, and other practices, establishes on the one hand further cultural links with the rest of the American continent, and on the other hand suggests strongly an ancient pre-American origin of all these usages. It in- creases greatly the need of the extension of explorations beyond the Bering Strait and Sea into the vast and archeologically still largely unknown territories of eastern Siberia.
‘WNS'f ‘uoisuadsns 104
pe[[ip OM} YUM ‘purysy
INCN'S'() ‘eUseaq ye sjoy pel[lap
‘(UMOYS JOU) IJ9] 94} UL auUO puke UOIsI1 a[dutl9a} JYSII 9Y4} UT saToYy
uiog INQ Wojy [[Nys a[eul seiuoy-a1g VY *
yeipoy ‘Aeq yxeAp
"RELLLE > SUIMOYS “purys] 1} [NYS sviu0y
Avg yeky, uliog ing u
jo uolji0od ev ‘tamoTqy ‘“z47ZF1 “ON “JAN'S (1) ay} jo doy 94} Ul SSUIT[I1p sulMmoys “uBS “IYOIJ, WO, [[NAS uvlpuy uv jo yuoursery ctrvddy “1
bt “Id ‘tp “ON ‘LOL “10A
SNOILO31100 SNOANVIISOSIN NVINOSHLIWS
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 101, NO. 4, PL.
1. Lower jaw of a Pre-Koniag boy of about 14, from Our Point, Uyak Bay, Kodiak Island, drilled for suspension through each ascending portion above the angle, and also incompletely below right lateral incisor. U.S.N.M. No. 379244.
2. Pre-Koniag right scapula, from Our Point, Uyak Bay, Kodiak Island, show- ing one complete and one incomplete drill hoie, for suspension. U.S.N.M. No. 379249.
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOLES 101), INOS 4, RE. 3
1. Pre-Koniag specimens from Our Point, Uyak Bay, Kodiak Island. Left. child’s pelvic bone, drilled for suspension. U.S.N.M. No. 379251. Right, a portion of a sphenoid of an adult skull, drilled for suspension. U.S.N.M. No. 379250.
2. Pre-Koniag pelvic bones from two separate subjects, drilled for suspension; from Our Point, Uyak Bay, Kodiak Island. Left, U.S.N.M. No. 377701; right, U.S.N.M. No. 379248.
‘oL1zc92 ‘ON “NN SA “SULUTYE } SuULMOYS “punog UTeTT ITAA 9OULT | “puelsy] syystuy WOT} [[NYS Vs[euloy “Z INQ) Wott A
{ 4PA “yuUl0d Yootg Ajtepja ue jo [pnyS ‘1
“puej[sy xt I MOUS “HPUIOM
SidienvasON| sO); lO” SNOILD3S11090 SNOANVIISOSIN NVINOSHLINS
‘punOM 9Y} 0} A[[euTsi10 parjdde [ersrazeu ofues10Utr
aWIOS 0} anp aq Ajqissod Avur Sutuado ayy ynoqe woeAO[OISIP IY J, ‘“I9pso0q st} ‘PLOfte ‘ON “JWN'S’Q ‘uoljeurdsat, surmoys ‘uleyd JO Surjaaaq Surmoys ‘1 sinsSy Ul pejzesjsni[i wonrurdsty Jo MaTA posiejug “Z uelyns[y ‘purlsy ByPyY Wo1y ‘uULMIOM yna]Y-a41d Ajtopja ue JO [[NHS “1
G ‘Id ‘bt “ON ‘LOL “10A SNOILO31100 SNOANVIISOSIW NVINOSHLINS
REG
NO. 4,
01,
VOL. 1
SMITHSONIAN MISCELLANEOUS COLLECTIONS
inings.
g lesions resembling treph
OW 1
372883.
No.
Uyak Bay, Kodiak Island, sh«
U.S.N.M.
rom Our Point,
¢ 1
i929 infant skull,
Parts of a Pre-Koni
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 101, NO. 4, PL.
Front and rear view of a purposely broken out Pre-Koniag maxilla, from Our Point, Uyak Bay, Kodiak Island. U.S.N.M. No. 379245.
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOLE. 10d); INO: 4, (PE. 8
sis a Y
1. Right male humerus, from 2. A_Pre-Koniag_ female femur, from Our Point, Point Hope, Alaska, showing Uyak Bay, Kodiak Island, the head of which has been trepanation. U.S.N.M. No. neatly cut off. U.S.N.M. No. 372822.
332608 (6).
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOES Oni NOS 4s) PEa9
Pre-Koniag dipper and bowl made from human _ skulls; from, Our Point, Uyak Bay, Kodiak Island. Upper. U.S.N.M. No. 379247; lower, U.S.N.M. No. 379243.
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 101, NO. 4, PL. 10
Pre-Koniag dipper and bowl made from human skulls; from Our Point, Uyak Bay, Kodiak Island. Upper, U.S.N.M. No. 379246; lower, U.S.N.M. No. 377738.
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOUS On), NO 4.) PE tt
1. Ivory eyes, from a Pre-Koniag male skull. Our Point, Uyak Bay, Kodiak Island. Natural size.
2. A dipper made from a human skull, from Kagamil Island, Aleutian Chain. U.S.N-M. No: 379253:
ie Di it A ne
RoR ey rece
* ni 1
oes von f
‘be i
a ‘ _ SMITHSONIAN MISCELLANEOUS COLLECTIONS VOLUME 101, NUMBER Ss
. BY | HENRYK ARCTOWSKI
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOLUME 101, NUMBER 5
Roebling JFund
ON SOLAR-CONSTANT AND ATMOSPHERIC TEMPERATURE CHANGES
BY HENRYK ARCTOWSKI
nes
ph
(PUBLICATION 3641)
GITY OF WASHINGTON PUBLISHED BY THE SMITHSONIAN INSTITUTION NOVEMBER 7, 1941
The Lord Baltimore Press
BALTIMORE, MD., U. 8 As
PREFACE
In recent time, correspondence reaching me from various sources convinces me that the view is growing among experts that the vari- ation of solar radiation is a weather element of first-rate importance. This would not have occurred had it not been for the generous, enthusi- astic, and discriminating support afforded by Mr. John A. Roebling to the Smithsonian solar investigations throughout the past 20 years.
When the lamentable war in Europe stranded Dr. Arctowski in this country, it seemed to me that it would be very helpful if an unbiased research meteorologist of Dr. Arctowski’s long experience and stand- ing would make a study of the supposed solar variation with reference to world weather conditions. I proposed this project to Mr. Roebling. He enthusiastically embraced it, and generously provided means for the investigation. The present paper represents the first fruits of this work, and it is published under a grant from Mr. Roebling.
Dr. Arctowski has confessed to me that he undertook the assign- ment unbelieving in Smithsonian claims regarding the importance of solar variability. He said that he had intended to work faithfully on the problem for 2 or 3 months, and then, finding nothing, to tear up his papers and resign, as he could not accept compensation under false premises. But to his surprise he became fully convinced within a few weeks that solar variation is a main cause of weather changes.
But obvious as he finds the influence of solar variability in weather to be, the complexity of its effects is baffling. He regards his present paper, which is the fruit of 2 years of the most devoted and un- remitting work, as my colleagues of the Institution will testify, to be merely a preliminary survey of some of the factors which must be considered.
Dr. Arctowski asked me to read his paper critically in manuscript. I approached the matter, not as a qualified meteorologist, but as one experienced in the measurement of solar radiation. It appeared to -me that my comments from this viewpoint might add to the interest of readers of Dr. Arctowski’s paper ; hence this introduction.
I think no one who has not struggled for many years, as my col- leagues and I have done, with the actual field work and computations in solar-constant measurements can have a vivid appreciation either of the strength or of the weakness of our results.
The ocean of atmosphere, even above our best desert mountain stations, is constantly changing its transparency. We have not the
iii
iv PREFACE
advantage of variable-star observers, who can compare their chosen star object with dozens of other stars nearby, the rays of all of which experience nearly identical percentage losses within our atmosphere. The observer of solar variability can have no comparison stars. He must compare absolute measurements of today with those of tomorrow and succeeding days, and evaluate the sun’s small percentage variation by differences of total radiation. This makes it well-nigh impossible to obtain the necessary degree of accuracy to disclose the solar variation.
The reader will find that almost none of the cases of solar variation which Dr. Arctowski quotes in his paper, and which he correlates with atmospheric and weather changes, reach the magnitude of one-half of I percent. To my colleagues and myself, who know so well the multitudinous sources of error in solar-constant determinations, it seems only by chance that any single isolated instance of solar vari- ation of such small magnitude as one-half of 1 percent should deserve individual consideration. It is true, however, that Dr. Arctowski forti- fies his data by using only cases where there is a regular approach and recession accompanying lows and highs of the solar constant. Yet in my own studies of correlation between solar variation and weather records, I have almost always thought it necessary to use mean values of many cases of solar variation in order to reduce accidental errors.
But Dr. Arctowski, as a meteorologist rather than a solar-constant observer, is more impressed by the changeability of a multiplicity of terrestrial circumstances than by the fallibility of individual deter- minations of solar-constant fluctuations. Hence he feels the necessity of dealing with individual cases, having no confidence that they are sufficiently comparable among themselves to give sound mean values. I greatly regret