For this reason Cl is pronounced to be trivalent. But in C,H,Cl, 
where the rule does hold good, Cl is suddenly again monovalent. In 
C,H,S S would be heaavulent. Besides: several substances which 
more or less deviate, are declared to be associated or slightly asso- 
ciated (at 7). 
For the compound esters one of the O-atoms in COOX is declared 
tetravalent (except for methyl-isobutyrate and methyl propionate, where 
it is unnecessary), the other O atom bivalent: 
p. For the alcohols and the acetic acid M finds the oxygen in 
OH tetravalent, the other atom O in COOH is /ivalent. Compare 
these assertions with ours in § + under ¢). We saw viz. that the 
eause must not be found in the O-atoms, but that the C in COOH 
has the normal fundamental value against 0 in other compounds 
and that also H and OH have the full value, viz. 3,2, against 1,6 
for the other H-atoms (bound to (©). 
For the amines NH,CH, and NH,C,H, deviate, which accordingly 
are declared to be associated with N pentavalent, but for the other 
amines, which do agree, association is unnecessary, and therefore 
N may remain trivalent. 
After these remarks, to which we shall not add anything (in my 
book to be published later [ subject also some very singular the- 
oretical considerations from his latest paper to criticism) the reader 
himself can form an idea of the value which he will have to attach 
to this remarkable “rule” of Mathews. 
In § 4 we have sufficiently shown, that not only the values of 
bz, but also those of Va, can be built up perfectly additively from 
a few fundamental values. With regard to these fundamental values 
themselves, we found with respect to hj exceedingly remarkable 
regularities in connection with the periodic system, and also the 
fundamental values of Va, — which are almost equal for every 
horizontal series — certainly open up important perspectives in 
connection with THomson’s and Nicworson’s theory about the atomic 
structure. 
But to this we shall revert later on. 
Clarens, November 1915, 
