THE LIFE OF THE SUN 13 



150 miles he imagines that there has been almost no 

 cooling. If, however, we take one of the cases put 

 by Professor Perry, and assume that below a depth of 

 four miles there is ten times the conductivity, we find 

 that after a period of 10,000,000,000 years the gradient 

 at the surface is still i F. for every fifty feet; but that 

 we have to descend to a depth of 1,500 miles before we 

 find the initial temperature of 7,000 F. undiminished 

 by cooling. In fact the earth, as a whole, has cooled 

 far more quickly than under Lord Kelvin's conditions, 

 the greater conductivity enabling a far larger amount 

 of the internal heat to escape; but in escaping it has 

 kept up the temperature gradient at the surface. 



Lord Kelvin, replying to Professor Perry's criticisms, 

 quite admits that the age at which he had arrived by 

 the use of this argument may be insufficient. Thus, he 

 says, in his letter : l 'I thought my range from 20 

 millions to 400 millions was probably wide enough, but 

 it is quite possible that I should have put the superior 

 limit a good deal higher, perhaps 4,000 instead of 400.' 



The third argument was suggested by Helmholtz, and 

 depends on the life of the sun. If the energy of the sun 

 is due only to the mutual gravitation of its parts, and if 

 the sun is now of uniform density, ' the amount of heat 

 generated by his contraction to his present volume would 

 have been sufficient to last eighteen million years at his 

 present rate of radiation.' 3 Lord Kelvin rejects the 

 assumption of uniform density, and is, in consequence of 

 this change, able to offer a much higher upward limit 

 of 500 million years. 



This argument also implies the strictest uniformitarian- 

 ism as regards the sun. We know that other suns may 

 suddenly gain a great accession of energy, so that their 

 radiation is immensely increased. We only detect such 

 changes when they are large and sudden, but they prepare 

 us to believe that smaller accessions may be much more 

 frequent, and perhaps a normal occurrence in the evolution 



1 Nature, January 3, 1895. 



2 Newcomb's Popular Astronomy, p. 523. 



