On some Points in Climatology. 143 



by Mr. Croll at all. Experiments on radiation, commenced 

 by Dulong and Petit, tend to show that Newton's theory of 

 the proportionality between temperature and radiation is not 

 well founded, and that, as temperature rises, radiation increases 

 in a much higher ratio. To speak more exactly, if we take a 

 series of temperatures in arithmetical progression, the corre- 

 sponding rates of radiation of heat will not be in arithmetical 

 progression, but in a series of which the differences continually 

 increase. An immediate inference from this general law is 

 that, if an isolated body receive a given amount of radiant 

 energy per annum, its mean annual temperature will be a 

 maximum when this radiation is uniform, and will be lower 

 the more irregular the reception of heat. 



Now it is well known that the total amount of heat received, 

 not only by the earth as a whole, but by each hemisphere, is 

 constant, notwithstanding the change in the earth's eccen- 

 tricity; but in virtue of the law just stated, any portion of the 

 earth's surface on which a large portion of the annual supply 

 of heat is delivered during a short summer will have a lower 

 mean temperature than the hemisphere on which the heat is 

 distributed more uniformly. But Mr. Croll does not, so far as 

 I have ever noticed, adduce this law at all. On the contrary, 

 he assumes Newton's law of radiation proportional to tempera- 

 ture, under which the cause would not act in the way suggested. 



One great source of inconclusiveness in Mr. Croll's results 

 seems to me to be a lack of quantitative precision in his 

 language. Though he may use numbers wherever it seems to 

 him they are applicable, one can hardly fail to notice that the 

 quantitative terms he most uses are such as "great," "very 

 great," " small," "comparatively small/' and these without 

 any statement of the units of comparison relatively to which 

 the expressions are used. Now I deem it not improbable that 

 the difference between a cold and a hot epoch may be due to 

 the very small preponderance of one or the other of several 

 antagonistic causes ; and, if so, quantitative precision is neces- 

 sary to lead to any reliable conclusion. 



1 shall now enter into some details. Mr. Croll suggests that 

 I may have forgotten the researches of Pouillet and Herschel 

 into the temperature of space. I reply that I regard the con- 

 clusion that the temperature of space is —239° as having no 

 sound basis. To speak with greater quantitative exactness, it 

 has precisely the same value as a photometric estimate of the 

 intensity of starlight, founded on observations of the sky 

 made in full day, with an attempt to eliminate the light 

 reflected by the sky so as to find what residue comes from the 

 stars. The fact is, that no observations of radiant heat from 



