168 Woeil-qf — CrolVs IIyj)otheses of Geological Climates. 



even the warmest and coldest months (which are generally not 

 January and July) do not differ even by 3°, and nowhere as 

 much as by 5°. 



Dr. Croll has answered the Rev. O. Fisher,* and this answer 

 is a curious example of the difficulties in which he has involved 

 himself. The question as to why the annual range on the equator 

 is so small is a very simple one. The quantity of heat received 

 from the sun does not vary more than in the ratio of 100 : 115 

 and, besides, the observations on which depends our knowl- 

 edge of the temperature of the equator, were mostly made on 

 the sea-coast, two good reasons indeed for a small annual range. 

 On the 50° N. the quantity received on the days of the winter 

 and summer solstices vary in the ratio of 100: 562, and yet in 

 some places of England the yearly range is not above 20° F., 

 and nowhere above 27° F. Thus it is the small annual range 

 in many regions of the middle and high latitudes which much 

 more needs explanation. Dr. Croll, in his reply, expresses the 

 opinion that the Northern hemisphere is the dominant one, 

 and, as the whole earth has a higher temperature in July than 

 in January, so by the operation of this cause, the normal excess 

 of the temperature of the equator in January is weakened and 

 even abolished. I replied to this, that the temperature on the 

 equator was nearly entirely influenced by cold winds, from 

 cold ocean-currents, in some regions (the West coasts of Africa 

 and America and the adjoining parts of the Pacific and Atlan- 

 tic), and in these the equator was considerably colder in July 

 than in January. For example, at the island of St. Thomas, 

 West Africa, by 2*7. The cold winds come from the South, 

 while winds from the North seldom reach the equator, and can 

 never have a depressing influence on the temperature. In most 

 places the temperatures on the equator were influenced by the 

 rainy season, so that when it was at its height in January, this 

 month was cooler than July, (this is the case even in Batavia, 

 7° S.), while when January was dry and July rainy, the former 

 month was warmer, not only on the equator, but even to some 

 degrees north of it. So for example, it is warmer by 7'2 at Lado, 

 Upper Nile 5° N., and hy 3*1 at Freetown, Sierra Leone, 

 8|° N.f 



All this is, I think, conclusive enough, and proves that Dr. 

 CroWs system of estimating temperatures breaks down when tested 

 seriously. Small errors would be quite natural in a question of 

 that kind, but I have shown that the errors are enormous, amount- 

 ing to 100° F., and more ; that is, they are greater than the dif- 

 ference of annual temperature between the equator and the North 

 pole. 



* ' Nature,' vol. xx, p. 577. 



f ' Nature,' vol. xxi. p. 249; reprinted also in this Journal. 1880. 



