250 SCIENCE PROGRESS 



advanced on this question by Zirkel, Cross, Pirsson, and others 

 it is not necessary to enter. The laws of crystallisation of 

 solutions lead at once to a general interpretation of the structure, 

 which is independent of any discontinuous change of con- 

 ditions. The porphyritic elements are to be regarded as 

 representing approximately the excess of particular constituents 

 over the eutectic proportions, and the ground-mass the eutectic 

 mixture itself. This view, clearly set forth by Becker, has been 

 discussed by Vogt for the particular case of the felspar-quartz- 

 rocks. The data are provided by a large series of comparative 

 analyses (of total rock, porphyritic crystals, and ground-mass) 

 published twenty years ago by Lagorio. Many of these repre- 

 sent rocks composed essentially of felspars and quartz, with 

 only a small proportion of ferro-magnesian minerals. They differ 

 rather widely in bulk-analysis ; and still greater difference is 

 seen in the porphyritic elements, which are sometimes felspars, 

 sometimes quartz, sometimes both felspars and quartz. The 

 analyses of ground-mass or glassy base show a notably smaller 

 range of difference. For rocks with about 74 per cent, of silica 

 the ground-mass has approximately the same composition as 

 the whole ; for more acid examples the ground-mass is less 

 acid than the whole rock, owing to abstraction of quartz in the 

 form of porphyritic crystals ; for more basic examples the 

 ground-mass is more acid than the rock, owing to the abstraction 

 of felspar. There is thus in the ground-mass of such rocks a 

 tendency towards a certain standard composition, with about 

 74 per cent, of silica, agreeing so far with the eutectic mixture of 

 felspar and quartz as determined from graphic intergrowths. 



It is important to remark, however, that, while this tendency 

 is unmistakably indicated, the ideal eutectic composition is in 

 most cases realised only with a rude approximation. Analytical 

 errors and the disturbing influence of the ferro-magnesian con- 

 stituents present will account for some part of the discrepancy, 

 but there are more important reasons. The rocks analysed by 

 Lagorio are almost all volcanic rocks, in which therefore the 

 progress of crystallisation has generally been interrupted by an 

 abrupt change of conditions ; and this break may often have 

 occurred before the excess of one or other constituent had com- 

 pletely crystallised out. Especially, we have to reckon with the 

 consequences of supersaturation in the magma. Vogt has shown 

 that this consideration not only goes far to account for the 



