144 Proceedings of Royal Society of Edinburgh. [mar. 21 , 
brine as it might now be called, containing 1 per cent, of chlorine, 
and remaining liquid at temperatures above - 1°*0 C. When 90 
more parts of ice had been deposited, we should have 10 parts of 
concentrated brine containing 10 per cent, chlorine and remaining 
liquid as low as - 13° C. In the case imagined, we assume the saline 
contents to consist of NaCI only, and with further concentration the 
cryohydrate would no doubt separate out and the mass become 
really solid. On reversing the operations, that is, warming the ice just 
formed, we should, when the temperature has risen to about — 13° C., 
have 999,990 parts ice and 10 brine containing 10 per cent, chlorine. 
Now, owing to the remarkable fact that pure ice, in contact with a 
saline solution, melts at a temperature which depends on the nature 
and the amount of the salt in the solution, and is identical with the 
temperature at which ice separates from a solution of the same 
composition on cooling, the brine liquefies more and more ice at 
progressively rising temperatures, until, as before, when the temper- 
ature of the mass has risen to -0°T C., it consists of 999,000 
parts of ice and 1000 parts of liquid water, containing 1 part of 
chlorine. The remainder of the ice will melt at a temperature 
gradually rising from - 0 o, l to 0° C. 
The consideration of this example furnishes an easy explanation 
of the anomalous behaviour of ice formed from anything but the 
very purest distilled water, in the neighbourhood of its melting- 
point. This subject has been studied with great care and thorough- 
ness by Pettersson. The apparent expansion of all but the very 
purest ice, when cooled below 0° C., is ascribed by him in part to 
solid saline contents of the ice which exercise a disturbing and 
unexplained influence on its physical properties. Viewed in the 
light of the fact that the presence of even the smallest quantity of 
saline matter in solution prevents the formation of ice at 0° C., and 
promotes its liquefaction at temperatures below 0° C., we see that 
this apparent expansion of the ice on cooling is probably due to 
the fact that we are dealing, not with homogeneous solid ice, but 
with a mixture of ice and saline solution. As the temperature 
falls this solution deposits more and more ice, and its volume 
increases. But the increase of volume is due to the formation of 
ice out of water, and not to the expansion of a crystalline solid 
already formed. 
