246 SCIENCE PROGRESS 



seems to be explicable by the specific properties of the bodies 

 concerned. The silicates do not in general begin to crystallise 

 at the proper temperatures indicated on a freezing-point diagram, 

 but only at lower temperatures, when a considerable degree of 

 supersaturation has been attained. Crystallisation once initiated 

 proceeds, however, in the case of olivine or anorthite with great 

 readiness, and large crystals may be built up very rapidly. 

 In this way the whole of the olivine in the magma may be 

 exhausted before the crystallisation of the felspar begins, 

 or conversely. With minerals of less rapid crystallisation a 

 different behaviour is indicated. Orthoclase and quartz are such 

 a pair, and rocks composed essentially of these two minerals 

 present a close analogy with the crystallisation of salt 

 and water ; the eutectic mixture of the two crystallised con- 

 stituents being characterised very often by some special 

 kind of intergrowth, clearly demonstrating their simultaneous 

 crystallisation. 



That the interstitial micropegmatite of acid rocks is of the 

 nature of a eutectic aggregate was first suggested by Teall in 

 1888, and in his Presidential Address to the Geological Society 

 in 1901 he returned to the subject, discussing from this point of 

 view graphic intergrowths in general and certain types of 

 spherulitic structures. The analogy of such intergrowths of 

 felspar and quartz with the eutectic aggregates of Guthrie is 

 indeed very close. Their structure is clear evidence of simul- 

 taneous crystallisation of the two minerals involved, and, as 

 Vogt has since proved, the relative proportions of the two 

 constituents are very closely constant, being about 74 parts of 

 orthoclase to 26 of quartz. Further, the manner of occurrence 

 of these intergrowths shows that they are a final residual 

 product of crystallisation. They often make up a ground-mass, 

 enclosing relatively large crystals either of felspar or of quartz, 

 which we must suppose to represent the original excess of one 

 or other constituent over eutectic proportions. With con- 

 siderable probability we may extend a like interpretation to the 

 less frequent graphic intergrowths of oligoclase with quartz or 

 augite or hornblende, of nepheline with augite, etc. Again, we 

 have intimate associations of a somewhat different kind between 

 two minerals which are closely related chemically and crystal- 

 lographically, though not isomorphous, e.g. orthoclase or 

 microcline and albite, or augite and enstatite. Here the two 



