NOKMAL SEQUENCES 629 



sequence, in the sense that many occurrences illustrate it. There is not, 

 however, any great preponderance of such well studied series gabbro to 

 granite over series between the other extremes. If the series from basalt 

 to granite is the main result of wholesale differentiation in nature, the 

 process has apparently tended to obscure itself, probably because the 

 large size of the resulting masses makes it difficult to follow the grada- 

 tion to an extreme. The emphasis on this series in many published dis- 

 cussion is based more on the abundance of the extreme products, basalt 

 and granite, than the abundant illustrations of the gradational series. 

 Any series which will explain the relations of our commonest rocks is 

 especially noteworthy. 



The lines on the plots run gabbro — diorite — granodiorite — granite. 

 The granites show a good deal of variation from siliceous to alkalic types, 

 both being well illustrated. Several of the series extend, at the basic 

 end, to peridotite. These make it evident that there is no real division 

 between Eosenbusch's two main series. 



Bowen does not stop his series at granite, but extends it through gran- 

 its to alkaline types. In this he goes farther than the other writers just 

 mentioned in connecting rocks in a supposed genetic series. The series 

 from granite to alkaline types is clear enough, but the argument that the 

 normal descent runs from gabbro to granite to alkaline types is based 

 on the broader argument that practically all the granites are derived 

 from basaltic magma. ISTo single mass of igneous rock seems to show 

 such an extended series. Furthermore, a series of this sort plotted on 

 figure 1 would give such a sharp change in direction as to be very sur- 

 prising. Bowen may have appreciated this, for he suggests that the 

 control of the process is modified at this stage by the increase in the 

 amount of water in the magma. 



The clear suggestions to be derived from the diagrams are, first, that 

 the known series do not ran continuously from gabbro to granite to 

 nephelite syenite; and, second, that if the absence of such a series is due 

 to incomplete exposures or analyses, the process of differentiation changes 

 most remarkably in passing the stage of granite. The curve shows too 

 sharp a turn to be suggestive of a continuous process. 



Bowen's "variant" series 24 only emphasizes the difficulty arising from 

 an attempt to follow his theoretical series through the diagram. He 

 supposes in this case that the magma passes from syenite to granite to 

 alkaline types. If the rock names mentioned in such a series are average 

 types, the sequence when plotted seems wholly improbable. Attention 



24 Op. cit., p. 77. 



