the Distribution of Heat over the Globe. 91 



the Atlantic at, say, 56°*. Then 239 + 56° = 295° represents 

 the number of degrees of rise due to the heat which it receives. 

 In other words, it takes all the heat that the Atlantic receives 

 to maintain its temperature 295° above the temperature of space. 

 Stop the Gulf-stream, and the Atlantic would be deprived of 

 one-fifth of the heat which it possesses. Then, if it takes five parts 

 of heat to maintain a temperature of 295° above that of space, 

 the four parts which would remain after the stream was stopped 

 would only be able to maintain a temperature of four-fifths of 

 295° or 236° above that of space : the stoppage of the Gulf- 

 stream would therefore deprive the Atlantic of an amount of 

 heat which would be sufficient to maintain its temperature 59° 

 above what it would otherwise be, did it depend alone upon the 

 heat received directly from the sun. It does not, of course, follow 

 that the Gulf-stream actually maintains the temperature 59° 

 above what it would otherwise be were there no ocean-currents, 

 because the actual heating-effect of the stream is neutralized to 

 a very considerable extent by cold currents from the arctic regions. 

 But 59° of rise represents its actual power ; consequently 59°, 

 minus the lowering effect of the cold currents, represents the 

 actual rise. What the rise may amount to at any particular 

 place must be determined by other means. 



This method of calculating how much the temperature of the 

 earth's surface would rise or fall from an increase or a decrease 

 in the absolute amount of heat received is that adopted by Sir 

 John Herschel in his ' Outlines of Astronomy/ § 369 a . 



About three years ago, in an article in the ' Reader/ I endea- 

 voured to show that this method is not rigidly correct. It has 

 been shown from the experiments of Dulong and Petit, Dr. Bal- 

 four Stewart, Professor Draper, and others, that the rate at which 

 a body radiates its heat off into space is not directly propor- 

 tionate to its absolute temperature. The rate at which a body 

 loses its heat as its temperature rises increases more rapidly than 

 the temperature. As a body rises in temperature the rate at 

 which it radiates off its heat increases ; but the rate of this in- 

 crease is not uniform, but increases with the temperature. 

 Consequently the temperature is not lowered in proportion to 

 the decrease of the sun's heat. But at the comparatively low 

 temperature with which we have at present to deal, the error 

 resulting from assuming the decrease of temperature to be pro- 

 portionate to the decrease of heat would not be great. 



It may be observed, however, that the experiments referred to 



* The mean temperature of the Atlantic between the tropics and the 

 arctic circle, according to Admiral FitzRoy's chart, is about 60°. But he 

 assigns far too high a temperature for latitudes above 50°. It is probable 

 that 56 c is not far from the truth. 



