618 JOHN JOHNSTON AND PAUL NIGGLI 
melt, shows that at a pressure of about 100 atm. a considerable 
quantity of SO, is condensed or dissolved in a melt of HgBr, at 
230°, in spite of the fact that the critical temperature of pure SO, 
IS 57456 
In precisely the same way we can have at high pressures water, 
or any of the other gases, present in the magmatic melt, and at 
the same time lowering its melting-point considerably below what 
it would be were no volatile components present.* Initially the 
amount of volatile component might be small, but its proportion in 
the liquid would increase (if the pressure remained constant, i.e., 
unless gas can escape) as crystallization progressed, unless the gas 
itself entered largely into the solid minerals. 
On the other hand, gaseous material at high pressures and at 
temperatures above the critical can exert some degree of solvent 
action, as is evident from the following reasoning. The volatile 
component A as liquid can dissolve the non-volatile component B. 
Now unless the solubility of B in liquid A diminishes very rapidly 
near the critical point, there will be at that point a definite amount 
of B in solution. But since at the critica! point the liquid and 
gaseous phases are by definition identical, there must then be the 
same proportion of B in the gaseous phase; so that above the critical 
temperature we can have a gaseous, or better a fluid, solution of a 
non-volatile substance. In general there is a very marked falling- 
off in solubility just below the critical point, but this in itself is 
no reason why at still higher temperatures and pressures the solu- 
bility of B in A should not again increase. : 
For instance, in fluid SO, at 159°4 and at a pressure of about 
80 atm. the solubilities of HgBr, and HglI, are respectively about 
1.5 per cent and 0.7 per cent by weight. At 230° in the same 
volume (hence at correspondingly higher pressures) the amount 
of HgBr, dissolved is about 8.5 per cent, while at 254° about 6.2 
per cent HglI, dissolves in the fluid phase. This example illustrates 
only the possibility of the phenomenon; the quantitative results 
cannot of course be transferred to magmatic systems. But the 
two cases are entirely analogous in principle, and the residual 
phase from solidification of the magma, even though it be gaseous, 
is to be looked upon as a solution. 
« Direct evidence bearing on this point will be presented in the near future. 
