138 Proceedings of the Royal Physical Society. 



tives of tropical conditions on another. This must always 

 have been the case, at every period of the Earth's history 

 after our atmosphere had passed into its present state. What 

 is called a glacial period is not necessarily a period of extreme 

 cold. It is really nothing more than the result of a combina- 

 tion of geographical conditions, through which large quantities 

 of aqueous vapour are distilled by the Sun's heat at one part, 

 transported chiefly by the winds to another region, where the 

 vapour is congealed into snow, and where it remains in 

 quantities which are in excess of those removed by melting. 

 In other words, glacial conditions may be said to be those 

 which occur where, each year, more snow falls on the low- 

 lands than is removed by melting. In this sense there is 

 still a glacial period in the Antarctic regions, and also in 

 Greenland and the parts near; but there is not a glacial 

 period in many other parts where the temperature is much 

 lower, because the air in these latter districts is drier, and 

 snow does not fall to an extent sufficient to give rise to 

 glaciers. 



The conditions that give rise to perennial snow on the 

 lowlands are very complicated, and would need considerable 

 space for their full discussion. But the inter-relationships of 

 such of the more important factors as this paper is concerned 

 with may be briefly noticed with advantage. First of all, it 

 is considered that, in the absence of any heat from external 

 sources, the atmosphere would have a temperature of -490° 

 F., and it would help us to gain a clearer view of what 

 climates are if we were to keep this in mind. It is, of 

 course, mainly due to the action of the Sun that the temper- 

 atures of those parts of the atmospheric envelope which are 

 next the Earth range a little more or a little less than 550° 

 F. above that frightfully-low point. It is not difficult to 

 compute what share of the total sun-heat that reaches the 

 Earth should be received per annum within any given lati- 

 tudes ; and, indeed, maps have been constructed which show 

 the " climates " that, in the absence of any modifying causes, 

 would characterise the latitudinal zones. But, as a matter 

 of fact, the actual distribution of heat by no means accords 

 with these " theoretical " or " astronomical " climates so 



