Haplodloritlc and Related Magmas. 183 



from Ab,An s to Ab 2 Aii, and, if it had happened that the liquid 

 was separated from the crystals before the later zones of plagio- 

 clase were formed, this liquid would, of course, have crystal- 

 lized to a body of haplodiorite. Again, this separated haplo- 

 diorite magma might have crystallized under conditions which 

 produced a zoning of its plagioclases, ranging, say, from A^Anj 

 to Ab 4 An„ and if the liquid were separated at a late stage it 

 could give a body of haplosyenite. 



If sinking of crystals took place as they grew in the liquid it 

 is plain that the effect on the uppermost liquid would be of the 

 same kind, indeed only stirring of considerable vigor could 

 prevent the formation of haplosyenite as the upper differentiate 

 of a large, slowly-cooled mass of haplobasaltic magma. The 

 only limit beyond which this differentiation could not pass is 

 a mixture of 97 per cent albite and 3 per cent diopside. 



Another feature of these differentiates is worthy of note, 

 viz., the fact that as the plagioclase becomes more alkalic the 

 percentage of diopside (colored constituent)* decreases rapidly. 

 Thus in the original haplogabbro there was 50 per cent diop- 

 side, in haplodiorite derived in the manner outlined about 30 

 per cent, in haplosyenite 10-15 per cent. On the other hand, 

 if it is imagined that these latter types were not derived by 

 this process of development from more basic types but were, 

 say, specially created, then there would be no necessary rela- 

 tion between the alkalinity of the plagioclase and the propor- 

 tion of diopside. 



The crystallization of the natural analogues of these melts, 

 gabbro, diorite, etc., is beyond doubt a considerably more com- 

 plex process, nevertheless the plagioclase mix-crystal series 

 must exert a similar influence, and when we turn to the rocks 

 this influence is plain in what is commonly termed the sub- 

 alkaline series. The increasing alkalinity of the feldspar and 

 the accompanying decreasing importance of the colored con- 

 stituents in the later members (members of latest consolidation) 

 of this series is a well-attested fact. One important difference 

 between the artificial melts here described and the natural 

 series is the prevalence of free silica as quartz in the later 

 members. In another paper this matter will be discussed in 

 some detail and it will be shown that the formation of olivine 

 at an early stage and the later formation of hornblende and 

 especially of biotite in tbe presence of water account satisfac- 

 torily for the development of quartz. 



The natural series presents a further important difference 

 from the ' haplo ' series in that there is no special mixture, 

 analogous to that containing 97 per cent albite and 3 per cent 

 diopside, which can be referred to as a necessary end, unless 



*I. e., corresponding to a colored, constituent as these terms are ordinarily 

 used though the pure, artificial diopside of these mixtures is not colored. 



