CONSTITUTION AND TEMPERATURE ON MAGNETIC SUSCEPTIBILITY. 261 



H be zero or not. It is really due to the spontaneous local intensity of magneti- 

 zation per unit volume and corresponds to the similar energy term, NP (Wfiiss), 

 due to the spontaneous magnetization in iron. Using the values which have already 

 been assigned to a! c and I,* viz., 



a' c = 2'5xl0 4 t ) 



1 = 400 /'. (4) 



we see at once that this term is large in comparison with \ . k t . H 2 and therefore the 

 change of volume to which the potential energy term \.a! e . P gives rise will be large 

 in comparison with that which we can produce artificially in a liquid by applying the 

 largest field available in the laboratory. 



When a diamagnetic substance crystallizes, the alteration of internal pressure will 

 be \a! c . P and therefore the accompanying change of volume in cubic centimetres per 



cubic centimetre will be 



,W = i . c . ' c . P, 



and substituting from (4) we find 



$V = %. 0'8x lO-^'x IGx 10 4 x2T)x 10 4 = O'lG c.c./c.c. 



to the appropriate order. 



Some values of <W in cubic centimetres per cubic centimetre are : 



Substance. 8V. 



Benzene O'lO 



Naphthalene OM4 



Benzophenone O'lO 



Di-phenylamine O'lO . , 



Formic acid O'lO 



Sodium 0-03 



Potassium 0'03 



Mercury 0'036 



Lead 0'03 



Tin . . ' 0-03 



Bismuth 0'03 j" 8 



Iron (at A 8 point) . . 0'003j 



The values calculated agree as well as could be expected with the experimental deter- 

 minations, since we know the orders of magnitude only of a' e and I, for these are 



* See p. 257 supra. 



t A physical explanation of this large value of a' e and of the corresponding constant N in ferro- 

 magnetism is given on p. 267 infra. 



| G. TAMMANN, ' Kristallisieren u. Schmelzen,' Leipzig, 1903, pp. 204 et seq. 

 DESCH, ' Metallography,' p. 242. 



