39 



The constitution of the albite syenite shows how power- 

 fully the presence of mineralizers has affected the order of 

 precipitation of the minerals, holding albite in solution until 

 microcline disappeared and free quartz was almost gone. 



(3) DISCUSSION OF THE VARIATION DIAGRAM. 



The foregoing observations have an important bearing on 

 the construction of the variation diagram (fig. 4) and on the 

 limitations of variation diagrams generally. The use of Si02 

 percentages as abscissae was obviously out of the question in 

 the present instance, and the specific gravity of the rocks was 

 indicated as the most characteristic and consistent reference- 

 variable, as indeed it should be if the heavier minerals 

 crystallized first. The regularity of the curves, particularly 

 those for FeO-f Fe^Og, MgO, and NaoO, justifies the choice of 

 abscissae. 



[Tlie experiment of making specific gravity variation 

 diagrams for a number of recorded occurrences of related 

 plutonic rocks was tried by the ^\Titer with a considerable 

 amount of success, which is after all only natural, seeing that 

 for the most part the order of crystallization of a magma, 

 that is the order in which rocks are formed by fractionation, 

 happens to be one of decreasing specific gravity.] 



As the latest stages of crystallization were entirely con- 

 trolled by mineralizers, and albite as a consequence was not 

 precipitated till the very end, the order of decreasing specific 

 gravity was destroyed and the end product (albite syenite) 

 will not fit along with the other rocks into any diagram that 

 can be constructed. 



The present rock series illustrates the limitations of the 

 usefulness of SiOj percentages as abscissae for a variation 

 diagram. Their employment really presupposes that on the 

 whole the sequence of crystallization under the given circum- 

 stances was that of lower silicates first with progressively 

 higher silicates following, and ending with quartz, if the 

 composition of the magma permitted. The ultimate product 

 of such a sequence would be an aplite or even a pure quartz 

 rock. Now while experience has shown this to be a common 

 case it must be recogiiized that the rule is by no means uni- 

 versal. The order of crystallization of minerals from a magma 

 is essentially one of increasing solubility under the conditions 

 obtaining, and this is not necessarily or always that of decreas- 

 ing SiOo percentage. For example, microcline usually follows 

 quartz in granites, as in the present instance, and here the 

 successive products of fractionation will be richer not merely 

 in quartz, but also, to a greater degree, in microcline. 



