HYPOTHESIS OF PRIMARY ORIGIN. • 125 



to adopt Brogger's suggestion that a partial crystallization of the feld- 

 spar set in at the same time that the segregation of the basic ingredients 

 was taking place. Backstrom considers* that the formation of basic in- 

 clusions by diffusion is improbable, since there can be no difference in 

 temperature between these and the surrounding magma. 



The second view (6), that the basic segregations may be due to the re- 

 crystallization of fused portions of the pyroxenite, is not improbable 

 though evidence of this is not at hand. The occurrence of inclosures of 

 the pyroxenite in the syenite-gneiss is frequently observed. These, how- 

 ever, usually retain more or less angular contours and do not appear to 

 have suffered very much from fusion. Professor Lawson has described f 

 inclosures of hornblende schist in the Laurentian gneiss from the Rainy 

 Lake region, which occur in sharply angular, subangular or somewhat 

 rounded blocks, or as more or less attenuated bands drawn out parallel 

 to the foliation of the gneiss and confused with it. He finds evidence of 

 total or partial fusion and recrystallization to such an extent often as to 

 admit of the deformation of the fragments and their being drawn out into 

 lenses. He thinks that where the magma had higher temperatures the 

 inclosures were entirely absorbed, leaving no trace of their existence 

 except a more basic local facies of the gneiss. 



Differentiation of the ellipsoidal rock. — On the assumption of a primary 

 origin it may be held that the distribution of the pyroxene in the ellip- 

 soidal rock represents — 



a. A molecular differentiation of the original magma, or 



b. That it is due to the movement of a partially differentiated magma. 

 In considering the first view (a) two hypotheses present themselves : 

 In the first hypothesis (I), which attempts to explain by Soret's prin- 

 ciple the peculiar distribution of the basic minerals in the ellipsoidal rock, 

 there arises at the outset great difficulty in conceiving conditions that 

 would give a difference of temperature between the feldspathic lumps 

 and the interstitial pyroxene bands. Moreover, the presence of the 

 gneissic structure would seem to render this view improbable. 



In the second hypothesis (II), based on the liquation theory, it may be 

 assumed that there has been a separation of tlie magma into layers of 

 different composition, and that owing to some disturbance they were 

 broken up and subsequently crystallized in the forms observed. This, 

 however, is considered improbable from the relations of the coarse grained 

 and ellipsoidal varieties and the absence in the former of any indications 

 of differentiation in layers such as the theory might lead one to expect. 

 Moreover, the presence of the gneissic structure seems to preclude anv 



* Jour. Geol., vol. 1, p. 777. 



tA. C. Lawson : Annual Report Geol. Survey of Canada, 1887-88, new series, vol. iii, part 1, p.lSOF. 



