32 On Residual Phenomena . [January, 
sumed law by any known expedient. Even if we were to 
fix the atomic weight of carbon at 24, the results, 4*82 for 
graphite and 3*92 for diamond, would still remain distinctly 
anomalous. The only method of saving the law is the 
assumption that we have here before us a residual pheno- 
menon, some other law exercising a modifying result, either 
on these three elements alone or at least to a greater extent 
than upon other simple bodies. It must be remembered 
that the atomic weights of the three elementary bodies con- 
cerned are low, and that of carbon, at least, maybe regarded 
as most satisfactorily ascertained, so that the solution of the 
difficultv cannot lie in that direction. 
• 
It is a somewhat curious circumstance that though the 
law of Dulong and Petit, as it is called, has been now for a 
long time before the scientific world, and though the ano- 
malies to which we have drawn attention have been recog- 
nised for more than a quarter of a century, the explanation 
is still wanting. Ivopp, as the result of his investigations, 
concludes that the law does not hold good for all the com- 
monly reputed elements in their solid state ; but if so we 
naturally inquire into the reason of its limited scope, and 
without some reply we feel inclined to doubt whether its 
approximate applicability in other cases may be anything 
more than a casual coincidence. In the meantime the law 
of Dulong and Petit should be regarded as on its trial, and 
the weight allowed it should be limited indeed. It would, 
for instance, be scarcely justifiable to rejeCt any new hypo- 
thesis because it should happen to lead to results incom- 
patible with the law in question. 
In the instances we have been considering, and in all 
others of an analogous nature, there can be no doubt 
whether, or in how far, the law is in harmony with the 
phenomena which it embraces. But where the results are 
incapable of being stated numerically a serious difficulty 
arises. Issue may be joined on the fundamental question 
whether the theory agrees with the faCts or not, and a 
thoroughly indisputable decision is not easily arrived at. 
Hence the recognition of residual phenomena — not to speak 
of their explanation — is far less easy in biology or geology 
than in astronomy, physics, or chemistry. One and the 
same theory may be by different authorities accepted as 
satisfactory and complete, rejected as altogether illusory, 
and again admitted as a partial view of the truth, which, 
however, leaves many points still in the dark. To us it 
seems that the origin of species supplies an admirable 
