112 



ANNALS NEW YORK ACADEMY OF SCIENCES 



Therefore, since, as has been demonstrated in this memoir, the Are- 

 quipa variation is not exclusively an e<]uatorial phenomenon, hut appears, 

 more or less modiiied, in North America, P]urope and even in the arctic 

 regions, the question is whether the years of conjoined pleions do not 

 correspond to crests of the Arequipa curve and whether the depressions 

 of this curve do not correspond to years^of isolated pleions in a net of 

 conjoined antipleions. In fact, the maps 1"or 1900-1909 show that the 

 years 1904 and 1907, closely following the Arcfiuipa depressions, are years 

 of conjoint antipleions, and that the years 1900, 1908 and to a certain 

 extent 1905, are pleionian years with isolated antipleions. 



Pig. 61. — Pleionian connections 



The existence of macropleionian variations, the close correlation of the 

 pleionian phenomenon with the Arequipa variation, the compensating 

 antipleions, and, finally, the dynamic character of these climatic changes, 

 eliminate, it seems to me, the hypothesis attributing such changes exclu- 

 sively to the presence of variable quantities of volcanic dust in the higher 

 layers of our atmosphere. 



Variations of the solar radiation must be the real and most important 

 cause producing the changes of our climates and keeping them in a 

 dynamic state. 



'^I'he elaborate investigations pursued at the Smithsonian Astrophysical 

 Observatory, and the Mount Wilson Observations in particular, give 

 striking support to this conclusion. 



In fact, considering the means of the solar constants, observed at Mount 

 Wilson during the summer months of 1905, 1906, 1905, 1909 and 1910, 

 and comparing the differences of these mean values with the correspond- 



