354 



REPORT — 1886. 



moment, first, acids and basic hydrates ; second, salts decomposed by water, double 

 salts, &c. ; third, neutral salta which remain abnormal even in very dilute solution ; 

 and consider only the salts I have called nonnaV 



The class of bodies most favourable to ]M. Kohlrausch's results are those relating 

 to anhydrous and nearly ' normal ' salts. Here are his numbers for the following ions. 

 The numbers for Am, K, Cy, 01, Br, I, and NOj differ very little from a mean value 50. 

 But for Ag, SO4, CIO3, the atomic conductivity is 40, and for CO3 it is 36. 



' To compare these numbers with my own, this is what I have done. Consider, 

 for example, K^SO^ ; its molecular resistance according to Kohlrausch is 



2 1 __88_ 

 48 "*" 40 ~ lyiiO' 

 That of KCl, which I have used as a standard, is 



1 1 97 



48 "^49' 



iJ35:^ 



The ratio of these is 



i-ni4.i 



I have measured this same ratio at different concentrations, viz. '—, ^\^, y~, -^^-^ , 

 and find numbers which, divided by the ratio of equivalents, give respectively 



1-288 1-144 1-074 1-011. 



These evidently tend towards the limit 1, and not to the number 1-1114 proposed 

 by M. Kohlrausch.' 



Similarly with K3CO3 and AgNO,. 



' For sulphate of silver compared with ECl, Kohlrausch's table would give 1*21 , 

 which I find by experiment -981. 



' These divergencies are not enormous, but they get much bigger if one proceeds 

 to study normal h3'drated salts whose conductivity varies so rapidly with dilution. 

 ' Kohlrausch's numbers for the following metals are very different from 50 : — 



Ca 26 

 Ba 29 



Cu 29 and 12 

 Mg 23 and 14 

 Zn — and 12 



* When two coefficients are given, the former belongs to ordinary salts, the other 

 to sulphates, which otherwise M. Kohlrausch cannot get to obey his law.' 



Comparing these with KCl in the manner already explained, M. Bouty gets the 

 ioUowing table, in which Kohlrausch's numbers are also inserted for comparison. 



' This is a mistake. The velocity numbers themselves must be added, not their 

 reciprocals ; so for K2SO4 48 + 40 = 88 is the conductivity, and for KCl it is 97 ; hence 

 the ratio is 1-10. The difference being so small I have left the figures of M. Bouty 

 unchanged, especially as the ground of his arguments is only partially affected, not 

 removed. In a note to the September 1886 Journal de Physiqne, p. 428, M. Bouty 

 admits his slip. 



