PROFESSOR KOPP ON THE SPECIFIC HEAT OF SOLID BODIES. 
185 
The preceding contains a generalization of Neumann’s law ; but as certainly as this law 
is recognized in the preceding in a more general manner than was formerly assumed, as 
little is it universally applicable. 
Regnault’s investigations have shown that Neumann’s law is not rigidly valid. Even 
for those compounds which contain the same element as electronegative constituent, 
and have similar atomic constitution, he found the atomic heats as much as to 9- dif- 
ferent from each other*. The reason of this he seeks in the same circumstances, which 
in his view prevent a closer agreement in the atomic weights of the elements (com- 
pare § 92). 
Differences of this kind, and even still more considerable, occur in the atomic heats 
of compounds for which greater agreement in these numbers might be expected — of 
such compounds, "that is, as contain elements of the same, or almost the same atomic heat 
combined with the same other element in the same atomic proportion. To this belongs 
the fact that the atomic heat has been found so different (§ 85) for the isomorphous com- 
pounds, magnetic iron ore (37*7), chrome iron ore (31 ’2), and spinelle (27*7), and for 
alumina (21*3) and for sesquioxide of iron (26‘8). In the atomic heats of such analogous 
compounds there are differences for which, or rather for the magnitude of which, as 
furnished by our present observations, I know at present no adequate explanation. 
But there is another kind of difference in the atomic heats of analogous compounds, 
which exhibits a regularity, and for which an explanation can be given. Certain 
elements impress on all their compounds the common characteristic, that their atomic 
heat is much smaller than that of most analogous compounds. The atomic heat of 
boracic acid, B 2 0 3 , is only 16-6, while that of most other compounds, R 2 0 3 and R 2 0 3 , is 
between 25 and 28 (§ 85). The atomic heat of the borates, R B 0 2 , is (§ 87) only 16-8,. 
while that of R 2 0 2 , as the mean of the determinations in § 85, is 22*2. The atomic 
heat of Rb B 2 0 4 is (§ 87) only 26'5, while that of Ee 3 0 4 (§ 85) in the mean is 37 - 7. 
Similar results have been obtained for compounds of certain other elements, of carbon 
and of silicium for instance, that is, of those elements which in the free state have a 
smaller atomic heat than that of most other elements. 
This observation leads to the question whether the elements enter into compounds 
with the atomic heats which they have in the free state, and in connexion with this, 
how far is it permissible to make an indirect determination of the atomic heat of the 
elements (in their solid state) from the atomic heats of their (solid) compounds. 
96. The assumption that elements enter into compounds with the atomic heats they 
have in the free state would be inadmissible, if not only the atomic structure as ex- 
pressed by the empirical formula, but also the grouping of the elements to proximate 
constituents, as is endeavoured to be expressed by the rational formula, influenced 
the atomic heat of the compounds. That the latter is not the case is very probable 
from the comparisons made in § 94, where approximately equal atomic heats were 
obtained for compounds of analogous empirical formulae, even with the greatest dissi- 
* Ann. de Chim. et de Phys. [3] vol. i. p. 196. 
2 c 2 
