THE SUN AND THE ATMOSPHERE STETSON 169 



growth of trees, as measured by the widths of their rings, follow 

 very closely the sequences in the sunspot cycle. Since variations in 

 tree growth suggest variations in precipitation, he has accumulated a 

 vast amount of evidence for alternations of wet and dry periods vari- 

 able with the sunspot cycle, carrying records backward for some 

 3,000 years. His studies appear to indicate that at least for selected 

 regions, trees have shown most growth when sunspots were most 

 numerous. It does not appear improbable, however, that the growth 

 of trees integrates all favorable conditions and that temperature, the 

 quality of sunlight, and particularly the percentage of ultraviolet 

 enter into the growth rate of trees as well as does rainfall. 



Sunspot periods have also been traced with minor discrepancies in 

 the flow of rivers and the level of lakes, some regions responding 

 much more clearly than others to the sunspot cycle. 



Altogether we see there are many indications that the earth responds 

 to the changing state of the sun over an interval of a little over 11 

 years and often by double this period or approximately 23 years. 

 Whether all of the effects produced in the earth and its atmosphere 

 that are noticed at sunspot maxima are the result of the sunspots 

 themselves or whether the state of the sun and its whole surroundings 

 are so activated as to change materially the cosmic environment of 

 the earth is a question still unanswered. Enough has been said to 

 indicate that it is eminently desirable to be able to predict sunspots 

 with all the accuracy that our somewhat scanty knowledge will permit. 



Many attempts have been made to analyze the curve of sunspot 

 activity from the time of reliable records, which began in the middle 

 of the eighteenth century, up to the present time. Investigators 

 differ as to the number of significant periods which may enter into 

 the solar graph. In addition to the 11- and 23-year cycles, others of 

 37, 68, 77, 83, 252, 300, and even possibly as long as 1,400 years have 

 been assigned. Authorities are not in agreement as to the reality of 

 some of these intervals. Obviously we have hardly enough sunspot 

 records to be sure of any intervals of 100 years or more. 



Somewhat over 2 years ago, on the basis of the occurrences of the 

 past cycles and the present rate in the rise in sunspots, I ventured to 

 predict that the next sunspot maximum would occur in the early part 

 of 1938, or more than a year before what would have been expected 

 by adding 11 years to the time of the last maximum. If subsequent 

 observations should prove my predictions correct, it would be one of 

 the pleasantest and most unexpected turns of fate for which a scientist 

 could hope. We have not yet an adequate basis for any degree of 

 precision in the sunspot prediction. 



The sunspot records obtained during the last year already have out- 

 topped the records of 1917 and indeed all previous maxima subsequent 

 to that of 1870. It will be observed from the sunspot curve that in 



