FORM 01' SOLAR HYPOTHESIS 521 



however, that winter storms increase. N'ot only do the trees suggest that 

 the winters have relatively few storms, but Knllmer's figures demonstrate 

 it. If the average number of storms from December to March during the 

 years 1877-1879 and 1888-1890, when the sun-spots were at a minimum, 

 be taken as 100, the number during the intervening period of maximum 

 spots, 1882-1884, is 81.8. If the summer storms from April to November 

 are taken in the same way, the figures become respectively 100 and 160. 

 In other words, at this particular time of maximum spots winter storms 

 decreased somewhat in northern Germany, while summer storms in- 

 creased greatly. The long curve of tree growth in figure 2 suggests that 

 this happens regularly during each repetition of the sun-spot cycle. *If , 

 this is true, it has an important bearing on what we shall soon say as to 

 changes of climate in earlier times. 



The ''cyclonic" versus the ''caloric' for in of the solar hypothesis.— This 

 brings us to the end of our consideration of modern climatic variations 

 and their causes. In summing up the matter, we see first that the unas- 

 sailable evidence of Newcomb, Koppen, Hann, and others proves that 

 there is a close relationship between changes of temperature in tropical 

 regions and changes in the sun-spot cycle. The work of Arctowski shows 

 that in tropical regions and in the extratropical areas where the climate 

 is under direct solar control, there are synchronous departures of the 

 temperature from its mean value, and that these must be due to some 

 widely acting cause, probably the sun. An amplification of Arctowski's 

 work indicates that where we find discrepancies and apparent contradic- 

 tions between the variations of temperature in different regions, they 

 appear to be explicable as the result of the transportation of heat by cur- 

 rents of air, and to a much greater degree by currents of water. Other 

 discrepancies appear to be due to the dust of volcanic eruptions, and 

 probably there are still other causes which we have not yet discovered. 

 Next we have seen that in spitfe of our conclusions as to the influence of 

 the sun, the measured variations in the solar constant do not lead to the 

 belief that such variations are the main cause of differences in terrestrial 

 temperature. This leads to the consideration of other climatic phenom- 

 ena which may be related to solar changes. We find that in tropical 

 regions there is strong evidence that the number of cyclonic storms varies 

 in direct harmony with the sun-spots. In northern Germany the growth 

 of trees shows that when sun-spots are at a maximum the climate tends 

 to become continental, with wet, stormy summers, relatively dry winters, 

 and early springs. When sun-spots are few, on the contrary, oceanic con- 

 ditions prevail and the contrast between summer and winter diminishes. 

 The actual figures as to the number of storms support this conclusion. 



