I04 P- LOEWINSON-LESSING 



sink in this syenitic liquid. But we must not forget that this 

 bytownite is, according to Bowen, only a temporary component. 

 In the following stages of crystallization by reaction with the Hquid, 

 the albite is extracted, and gradually the feldspar is transformed 

 into labradorite, representing the final product of crystalHzation. 

 If so, the Hquid will again change from syenitic to leucocratic- 

 gabbroid, and the sunken crystals will once more float. 



The foregoing considerations therefore, lead to two possibilities : 

 (i) In consequence of the sinking and subsequent floating of the 

 crystals while the diopside continues to crystalhze, and in conse- 

 quence of vertical and other currents that must arise under these 

 circumstances, the crystals of diopside and labradorite will be 

 mixed and no anorthosite will be formed. (2) If the crystallization 

 proceeds in the manner assumed in the first case, namely if diopside 

 and labradorite crystallize directly, there will be a tendency toward 

 sinking and floating of the minerals. This tendency can produce a 

 complete separation of the two minerals only in the case of mingling 

 movements (as is illustrated in the separation of minerals in heavy 

 liquids). Such movements being absent and the liquid being quiet, 

 there can arise melanocratic and leucocratic facies with intermediate 

 members, but of course not a syenite nor a granite. 



Summary. — Summarizing all that has been said above, we must 

 concede that the problem of the anorthosites and other monomineral 

 igneous rocks is not yet completely solved. It must be conceded 

 that in certain cases several factors, to which different authors 

 - refer, may possibly or actually take part in the formation of these 

 rocks; but every attempt to put forward only one of these factors 

 to the exclusion of the others, gives proof of exclusiveness and in- 

 sufficiency. Indeed, such rocks as gabbros with anorthositic or 

 pyroxenitic bands cannot be understood without admitting the 

 existence of anorthosite and pyroxenite in the liquid state. But I 

 am far from asserting that the only possible explanation is the 

 hypothesis of liquid immiscibihty, and that this is applicable to 

 all forms of anorthositic rocks. And again, the possibiHty of the 

 separation of crystals of pyroxene and labradorite from a gabbroic 

 magma is sufficiently illustrated by the analogous phenomenon in 

 porphyritic lavas, and is corroborated by their respective specific 



