352 ANNUAL EEPOET SMITHSONIAN INSTITUTION, 1908. 



in the obliquity of the ecliptic, or to variations in the carbonic acid 

 in the air. He shows that either would give effects of the magnitude 

 required; but it seems doubtful whether either will agree with the 

 records of historical geology ; for as regards the first cause, the change 

 in the obliquity is, geologically speaking, a short and constant oscilla- 

 tion; and, as to the second, as it rests on the variation of carbonic 

 acid, it is open to the same objections as to those of Arrhenius's 

 theory. 



CHANGES IN ATMOSPHERIC CIRCULATION, 



That the explanation of glacial periods is to be sought in changes 

 in the circulation of the atmosphere resulting from geographical 

 changes has been several times suggested, in accordance with Bu- 

 chan's results.*^ This principle has received its fullest application 

 to a specific case by Harmer ^ to the Pleistocene climate of north- 

 western Europe. And, moreover, Dickson has shown how the distri- 

 bution of the glaciations in that case corresponds with what would 

 be expected if they were due to differences in atmospheric circulation. 

 Such meteorological changes would be quite inadequate to explain 

 the occurrence of a tropical climate in the arctic regions, but they 

 would account for changes of temperature of a few degrees, and for 

 glaciations by local concentrations of the snowfall. The difference 

 between the climates of western Europe and eastern America is ob- 

 viously due to meteorological conditions, resulting from geographical 

 position. The differences on the two coasts of the North Atlantic 

 w^ere naturally first attributed to the influence of ocean currents ; but 

 with our present knowledge as to their feebleness and the bending of 

 the Gulf Stream off Newfoundland, ocean currents may be dismissed 

 as a very subordinate factor. A different distribution in air pressure 

 resulting in a different circulation of the wind would probably be a 

 more effective cause, and appears to me at present to offer the beft 

 prospects of a satisfactory solution to the problem. It is the only 

 explanation that seems to agree with the essential facts, viz, the 

 development of glaciation from scattered centers, and at somewhat 

 different dates, and the apparent independence of the glaciations in 

 distant continents, and their apparent direct dependence on a par- 

 ticular adjustment of meteorological conditions. 



The slow march of glaciation across North America and possibly 

 also across Europe is intelligible on this hypothesis, and there is no 

 reason, on that theory, to expect coincidence of glaciations in the 

 Northern and Southern hemispheres. The former glacial extensions 



° For instance, I endeavored to show in 1894 that the more extensive glacia- 

 tion of Mount Kenya was due to a local difference in the atmospheric pressure 

 due to the former greater height of this denuded volcano. ("The glacial 

 geology of Mount Kenya," Quart. Journ. Geol. Soc, Vol. L, 1894, pp. 527-530.) 



" F. W. Harmer, Quart. Journ. Geol. Soc, Vol. LVII, pp. 405-472. 



