Page:The Journal of geology (IA journalofgeology11893univers).pdf/353

difficultly soluble compounds diffuse towards the plane of cooling"^[1] are misconceptions. It is the proportion between the solvent and the dissolved substance which is changed and this is all—so far as we know at present. Consequently, in order that one may use "Soret's principle" for the purposes of theoretical petrography it is quite necessary to have the question settled: what is "the solvent" and what "the thing dissolved?"

Vogt^[2] avoids this difficulty in the following way. He says: "Owing to chemical action certain 'liquid-molecules' are individualized, which are preliminarily kept dissolved in the resting magma, and which only by a subsequent lowering of temperature, or pressure, are separated in the solid condition. The minerals which crystallize first at every stage may consequently be considered originally 'dissolved' in the remaining 'mother-liquor.'" Here we find at first the supposition, that certain compounds are "individualized"^[3] in preference to others, and consequently the latter as not "individualized" form a sort of chaos. But this remainder must certainly consist also of chemical compounds. The author has perhaps thought that they should be dissociated, but it must be remembered that the free ions cannot diffuse independently of one another.

In the latter part of the quotation it is stated, that the substance which crystallizes out first when temperature sinks is to be considered as dissolved in the solvent, which crystallizes at a still lower temperature. But, in general, it is the solvent which crystallizes out first when the temperature falls, and this crystallization goes on until the "cutectic proportion" (Guthrie) is reached, when both the substance dissolved and the solvent crystallize simultaneously until the whole is solidified. If Vogt's reasoning is correct, the more a dilute solution of nitre is diluted with water, so much the more should the water be regarded as the substance dissolved.