538 Mr. W. Sutherland 



on a 



If we now compare our expressions (11) and (12) for the 

 bulk-modulus and latent heat with those which before gave 

 results respectively 65 and 4' 75 times too large, and neglect 

 for a moment the small terms, we see that in a general way 

 the theoretical bulk-modulus (11) and the experimental 

 would agree if b' were put equal to lb as 8 2 = 64, and the 

 theoretical latent heat (12) would agree with the experi- 

 mental if V =3'75&. But if in (11) and (12) we take account 

 of the small terms and solve for b\ we find that on the aver- 

 age the relation b' = 6b will bring the theoretical values of 

 both the bulk-modulus and latent heat into accord with the 

 experimental. The best way of showing this will be to 

 tabulate the values calculated from (11) and (12) by means 

 of the relation b' = 6b, and compare them with the experi- 

 mental. For the bulk-modulus the mean of k l9 k 2 , k 3 given 

 before is taken as the experimental value. 







io- 6 



times the Bulk-modulus 











Cu. 



Ag. 



Au. 



Mg. 



Zn. 



Al. 



Pb. 



Fe. 



Pt. 



Quasi -Exp.. 



. 1550 



810 



1370 



325 



625 



565 



174 



1660 



790 



Theory 



. 1710 



920 



1790 



310 



410 



560 



216 



3740 



5310 



Ratio 



. 11 



11 



1-3 



•96 



■66 



1-0 



1-2 



22 



6-7 



On the whole the values of the ratio show that b' = 6b 

 brings experiment and theory into good accord as regards 

 bulk-modulus. 







Latent Heat. 











Ag. 



Zn. 



Cd. 



Hg. 



Su. 



Pb. 



Exp. ... 



16-2 



24-5 



12-5 



2-8 



12-8 



4-8 



Theory... 



17 2 



23-7 



10-4 



1-9 



8-0 



4-6 



Ratio . . . 



1-06 



•97 



•83 



•67 



•63 



•9 



If, for the reasons given before, we attach less weight to 

 the values for mercury and tin, the value of the ratio for the 

 other four metals is satisfactorily near to 1. 



4. Further evidence of the Alteration of Molecules ivith Change 

 of Temperature. — It is important now to see what further 

 justification there is for the idea that molecules shrink with 

 rising temperature. The small amount by which metals ex- 

 pand on being heated from absolute zero to their melting- 

 points, namely about 2 per cent, of their linear dimensions, 

 has made it difficult to form a mechanical conception of mole- 

 cular behaviour during melting, because if the molecules are 

 reallv at rest against one another at absolute zero it is hard 



