We shall now shortly pass in review the different groups of 
substances, in which the little serupulous manipulations will be 
manifest with which M. gets rid of exceedingly troublesome 
deviations of up to 3800°/,. 
a. For the valenceless noble gases Marrrws’ rule is, of course, not 
valid at all. For if » =O, then Wa ==0 would also be = 0. But this 
diffieulty the author entirely ignores. He simply assumes the formula 
to be valid, and now simply calculates the valencies of He from 
it ete. Thus he finds for He »v = 0,04 a 0,07, for Neon » = 0,32, 
for Argon v—=1,12 a 1,35, for v—= 1,23, and finally for Xenon 
p= 1,80 à 1,95 (loc. cit p. 339). 
Hence He '/,,-valent, Ne '/,-valent, Ar. and Kr. 1'/,-valent and X 
very nearly bivalent ! 
And in order to justify these singular broken valencies (among 
which < 1), M assumes that these inactive gases are, indeed, 
valenceless, but that they yet possess two ‘extra’’-valencies, which 
are, however only partially active. One of these two valencies in 
extraordinary service would be positive, the other negative, but only 
for a part of the molecules present those valencies — which are 
neutralized in ordinary circumstances, — would be “closed”, as 
M expresses himself free and “open”. And only these “open” 
valencies are revealed by M’s rule. Thus 90°/ 
open for X, about 65°/, for Kr., 60°/, for Argon, 16°/, for Ne, and 
at last about 5°/,, for He (loc. cit. p. 341). 
of the valencies are 
0 
After this ingenious explanation — we see our way to explain 
away all discrepancies in the whole of Nature in this way — there 
is no need of anything further. 
B The other elements. Where with us only hydrogen departs from 
the rule with regard to Va, not a single element is found to conform 
to his “rule” in Marnews. In order to make this rule valid also now, 
however, he declares the factor *’/,, inaccurate by 60°/, for e.g. H, 
(though it happens to be almost perfectly accurate for H,, being 
4= 0,999); further N and O to be monovalent in N, and O,, and 
chlorine trivalent in Cl. 
y. The anorganic compounds. In our table all the normal com- 
pounds (also CO and NO after correction of py — see hj. in § 3) 
are in good agreement. With Matnews, however, 12 of the 14 
substances of our table do not agree at all. Only H,Se and SO, 
to 92/,! For ordinary substances this singular formula gives already errors of 
15%, for the factor 27/,4, and the errors can rise to more than 200°/) when we 
approach ideal substances. But his purpose is attained: the values of dj now 
coincide with those of a) (Gf. the last of the cited papers). 
SO” 
