ANALYSIS OF TERRESTRIAL RECORDS 115 



periodogram. In striking contrast, there is a near absence of random effects 

 in the Yellowstone analyses. Frequency periodograms are given in figure 55. 



ANALYSES OF OTHER GEOLOGICAL MATERIAL 



The very remarkable records revealed by Professor Gerard DeGeer in 

 Sweden and Dr. E. Antevs in North America well deserve careful analysis 

 for cycles. Antevs's measurements of the clay layers in the Connecticut 

 Valley, left in the retreat of the New England ice sheet, received a brief 

 preliminary analysis soon after they were published and a recent, fairly 

 complete review. His sequence, examined for cycles, covers about 4000 

 years with a break in it. Careful cross-dating has no doubt given the separate 

 parts a fine continuity. Cycles between 7 and 8 years in length occur fre- 

 quently; at 8.8 and at 10 years there are strong cycles; the 11-year cycle is 

 weak or absent; only two good examples of it appear in the 4000 years. 

 A cycle shows at 12.5 years, a group at 13 and another at 14.5, some weak 

 cycles at 16 and 17, and strong cycles close to 20 years. This result is essen- 

 tially repeated in the 736 years of Hudson River varves, recently published 

 by Dr. Chester A. Reeds, New York. 



In 1912 a large number of fragments of swamp cypress, Taxodium dis- 

 tichum, described to me as Pleistocene, were secured in the peat beds north 

 of Dresden, Germany. The total rings measured were 1260 in number, 

 distributed in four specimens giving long ring sequences nearly free from 

 errors. One 500-year record gave a strong cycle at 10.9 years; another gave 

 a good cycle at 10.0 years. Others were not particularly satisfactory, owing 

 to the complacency of the growth. These values show resemblance to the 

 varve cycles just described. 



Several years ago the late Professor Udden of the University of Texas 

 discovered lamination in anhydrite material forming a drill core brought up 

 from some 1500 feet in depth in Culbertson County, Texas. He became 

 satisfied of the annual character of the layers and with aid of a grant from 

 the Carnegie Institution he had them measured. He thought there might 

 be an error of the order of 2 per cent in layers lost. The longest good se- 

 quence of layers studied has 1443 years. The measures were turned over to 

 Dr. E. L. Dodd, whose analyses for cycles showed the presence of cycles of 

 10, 11, 19, and 33 years. I was fortunate enough to secure a copy of this 

 series of measures and have checked the analysis and confirm Dr. Dodd's 

 results within small differences. I find cycles at 7 to 7.5 years, 8.8, a very 

 prominent Hellmann cycle at 11.4, a group between 14 and 15 years, a faint 

 cycle at 17.5 years, and a little better one at 19 years. I found only a weak 

 cycle at 10 years. 



