JOHNS : EXPERIMENTAL PETROLOGY. 
465 
together with more recent field evidence, estabhshes the reciprocal 
solution theory as applicable to petrological problems. The 
case of the minerals which, like magnetite, crystallise at a very 
early stage of consolidation, often requires different treatment. 
Here it is not a question of the substance in excess over the 
eutectic ratio separating out, but rather that limited solubility 
is the determining factor, the solvent being a complex fused 
silicate. 
It might not be out of place to state in more detail what 
is meant by this theory of " reciprocal solution " which is now 
offered as a working hypothesis upon which to base experimental 
research. It assumes — until new data prove that it requires 
modification — that the igneous rocks have consolidated from 
a fused mass, which was originally a more or less homogeneous 
mixture of various substances, the silicates largely predominating, 
dissolved reciprocally in each other, and that in this mixture 
might be dissolved other minerals of limited solubility and 
occurring more sparingly. At a sufficiently high temperature 
these would probably form in any proportions a true solution, 
that is, a homogeneous mixture. The lowering of the tempera- 
ture would at a certain point destroy equilibrium and cause the 
difl'erentiation of the mass into distinct portions that would, if 
the mass were not too viscous, separate according to their specific 
gravities. 
As the temperature fell the mineral constituents would 
separate out, not necessarily in the order of basicity or in the 
order of fusibility, but according to their solubility in the still 
unconsolidated, but progressively changing, mother liquor. If 
at any stage a eutectic mixture results — and this is not necessarily 
the final stage — its structure would depend upon the temperature 
gradient. If that were too rapid, an emulsified structure would 
result. If normal, then the usual eutectic system of parallel 
plates ; or if the gradient were very flat the minerals might have 
time to separate even still more. Then it should be remembered 
that these transformations are reversible. If, say during 
intrusion or extrusion, the pressure sustained by the mass be 
reduced and sensible heat liberated, the temperature might rise 
