130 Mr. W. W. J. Nicol on the Molecular 
The last column (D) contains the difference divided by x, 
and thus gives the increase of volume for each molecule of 
salt present. 
A glance at the table will show that, in every case, the 
increase of the molecular volume by heating from 20° to 40° 
is inversely proportional to the strength of the solution. In 
other -words, the more dilute a solution is the greater the 
increase of its molecular volume by heat per molecule of 
dissolved salt. This is, I believe, inexplicable by the hydrate 
theory of solution. It appears impossible that a solution 
which contains only one molecule of salt to 100 molecules 
of water should experience a greater change of molecular 
volume per salt-molecule than a solution of the same salt 
containing two salt-molecules. This is, however, in ac- 
cordance with the theory * of expansion of salt-solutions 
which I put forward at the end of my paper on the Nature 
of Solution. At that time the only data at my disposal were 
the results obtained by Kremers and others, whose experi- 
ments were made with non-molecular solutions. The above 
table, however, fully bears out my previous statement that the 
expansion of a solution is the result of the action of heat on 
the resultant of the three forces — the attraction of water for 
water, that of water for salt, and that of salt for salt, the 
last two being inversely proportional to one another. 
In the cases considered in the first part of this paper, it is 
clear that a variation in the composition of the salt necessitates 
a corresponding change in the molecular volume of the 
solution containing that salt. This, however, holds true only 
of salts non-isomorphous in the strict sense of the term — that 
is, of heteromorphous and homeomorphous salts. When, how- 
ever, two salts are strictly isomcrphous, that is isotomous, 
they preserve their isotomy in solution. The last six salts 
given in Table I. come under this head ; they are the members 
of the so-called magnesian-sulphate group, which crystallize 
(except CuS0 4 ) with 7H 2 0, and then possess approximately 
identical molecular volumes |. This is also true to some 
extent of their solutions, which have nearly the same molecular 
volumes. What the slight variations may be due to I am not 
in a position at present to prove ; but enough has been said to 
show that the intermolecular spaces in isomorphous salts are 
comparable, and that true isomorphism is conditioned by the 
molecule itself ; while with salts which are isomorphous in the 
wider sense (homeomorphous), such as the chlorides, bromides, 
* Phil. Mag. 1883, xv. p. 99. 
t Thorpe, Journ. of Chem. Soc. 1880, p. 102. 
