ANALYSIS OF TERRESTRIAL RECORDS 99 



sunspot cycle was unstable? Can we infer a length of the sunspot cycle and 

 dates of maxima or minima from tree records extending back 1900 years in 

 Arizona and 3200 years in California, the California records though longer 

 having a less well-established relation to rainfall? And, more important 

 than the other questions perhaps, can we find evidence of recurrence of the 

 same cycle; is there some plan on which it comes back; and can we infer 

 any probability of its return at a definite time? Such knowledge would 

 compensate for some of the difficulties introduced by the temporary character 

 of climatic cycles. What of the cycles in the much more distant past? 

 Good tree records are available from tertiary sources, say 50,000,000 years 

 ago, perhaps very much more, and sediments carry us to much more distant 

 ages. Do we get any hints from these sources regarding the stability or 

 instability of our sun? 



We have at this time two types of records of past climates: namely, 

 accurately dated records in Arizona and California, abundantly cross- 

 identified in great numbers of trees, and, second, undated geological records 

 in fossil trees largely without cross-identification, and in sediments with 

 cross-identification, in all of which cycles may be studied. 



SEQUOIA CHRONOLOGY AND CYCLE RECURRENCE 



The difficulty with the study of cycles in climate has been the limited 

 lifetime or duration of any one cycle. The favorable possibility of finding 

 a definite recurrence of cycles has been recognized since 1927, and the great 

 sequoias have seemed the most promising location for the search. We have 

 in our laboratory 50 radials of these trees, of which about half show 2000 

 rings or more, and four extend back 3100 years; one reaches to 1305 B.C. 

 They fall into two groups, the Grant Park sequoias and the Springville se- 

 quoias, some 40 miles apart. It was discovered about 1921 that the double 

 sunspot cycle of 22 or 23 years is more common in these trees than the 11- 

 year cycle. The pine tree curve of the Sierra Nevadas resembles the se- 

 quoias in giving prominence to a cycle of about 23 years. A correlation 

 periodogram of this series of data by Dr. Alter, reproduced in figure 53, 

 shows a pronounced crest at about 22^ years and its multiples, thus confirm- 

 ing these results. 



In 1927 a study was made of several different sequoia groups in which 

 the presence of the solar influence had been suspected. Trials of a 23-year 

 cycle give us a series of curves each 23 years long, that are consecutive and 

 continuous through nearly 3000 years and each curve is a running or over- 

 lapping mean of three successive 23-year intervals. 



A large percentage of these 130 curves show clearly a 23-year cycle or 

 some subdivision of it up to five equal parts. The division into 4 parts is 

 illustrated in figure 42. Of course, this work is far from finished, but these 

 results apparently agree in the two separated sequoia groves. Limited 

 checking has also been carried successfully to other sequoia groupings that 

 for various reasons have been segregated. As a matter of safety a couple 



