26 THE UNIVERSITY SCIENCE BULLETIN. 



satisfying the necessary conditions and having smaller variations, 

 unless possibly by introducing more points of maxima and minima 

 upon it. Such a complication would be much less probable than the 

 variations shown by the present one, all of which are less than the 

 variations from the mean value of complete cycles of approximately 

 11 years have been in the rather recent past, as shown by table 1. 



THE RAINFALL DATA EXAMINED. 



I have examined the rainfall averages of each of the forty-two 

 sections in which the United States has been divided by the Weather 

 Bureau, of a number of stations in Central Siberia, of the Punjab in 

 India, of a few towns in Chile, of complete records of Denmark and 

 Sweden and stations in Holland and England, of South Australia, 

 of Jamaica, and of Tananarive, Madagascar. I had a small amount 

 of data from the Soudan and Abyssinia and scattered small amounts 

 from other countries, but none of these enough to examine with any 

 weight. There were also data such as received from Canada, where 

 the proximity of countries for which I had data made it seem un- 

 wise to take the great amount of time necessary to average the in- 

 dividual stations, and where, unlike Madagascar, thousands of miles 

 from the nearest data used, it seemed useless to obtain results with 

 the little weight that would be attached to one station. 



The results from each of the sections named above are discussed 

 here, the tables are given from which these results are deduced, the 

 values are given for each individual cycle, and the means of the 

 halves or thirds are given and plotted, as also the curves from the 

 whole data. The sections are grouped in three main divisions: 



(A) Interiors and eastern coasts of large continents. There are 

 three such -sections: Eastern United States, Central Siberia, and the 

 Punjab. 



(B) Western coasts of continents. This group includes the Pa- 

 cific coast of the United States, the group of countries from the 

 northwest European coast, and a very small amount of data from 

 Chile. 



(C) Other sections. This includes South Australia, Jamaica and 

 Tananarive, Madagascar. 



The last sun-spot maximum occurred in 1917, and all data since 

 then are thus unavailable for use in examining the existence of the 

 period. This would not be a serious handicap for predicting, if the 

 period should be proved to exist, since the course of the maxima and 

 minima could be followed from cycle to cycle by using means from 



