F. W. Clarke on the Atomic Volume of Compounds. 175 
a minute error in the determination of the first value must, in 
for 
stance, with Melene, CaH¢, an error of pee 0 4 j in the specific 
gravity will alter the atomic volume about 24:00. And it is very 
common to find the determinations of specific gravity for a single 
er oe differing more than 0°04. And, as I have — 
tal error. For this purpose, instead of directly comparing the ac- 
tual atomic volumes with the results of theory, I have preferred 
to compare the specific gravities calculated from the latter with 
those really determined by experiment. Even here much care 
is needful, since errors are more likely to occur with some com- 
pounds than with others. Thus, there is much more danger 
of error in determining the specific gravity of anhydrous mag- 
nesic chlorid, than in taking that of baric sulphate. 
Furthermore, many different modes of taking specific gravity 
their melting point, an others distant therefrom. Neverthe- 
less, in spite of all these pombleiet of error and chances of 
irregularity, one curious series of relations between atomic 
volumes are easily demonstrated; which, taken together, hint 
strongly at a es a more general theory of atomic volumes 
oe pe hitherto been enunciated. 
to begin with, we ca a brief resumé .: seen points 
mits ‘have been demonstrated by other observ 
First comes the axiomatic statement of Schrider that the 
atomic volume of a compound must equal the sum of the ato- 
the sums of those of the free elements composing 
them, many other compounds have values lower than such sums, 
thereby indicating condensation. And some iodids have ato. 
mic volumes greater than the sums of those of the eae and 
the iodine. So here arises the question,—when a compound 
possesses a 1 greater or less volume than the Fest contained 
in it doin the free state, do those elements condense or expand 
in Pheer or in different ratios? It is really upon this ares that 
tudy of atomic volumes hinges; so that, stripped of all 
Bilichyaical notions of atoms and spheres of heat, the subject 
