40 JOURNAL OF GEOLOGY— SUPPLEMENT 



slowly and continual sinking of crystals occurred, the inevitable 

 result would be a body consisting of calcic plagioclase, olivine, and 

 magnesian pyroxene in its lower parts (i.e., of a gabbroidal nature) 

 and of sodic plagioclase approaching albite, diopsidic pyroxene, and 

 free silica in its upper parts (i.e., of a granitic nature), with various 

 intermediate types in the intermediate layers. If the freedom of 

 sinking of crystals were somewhat restricted, one of these inter- 

 mediate types, say a granodiorite or a diorite, would occur as the 

 uppermost differentiate, the limit of the process under these less 

 favorable conditions. The composition of the residual Hquid might, 

 moreover, have been similarly affected by zoning of crystals even 

 if there were no opportunity for the sinking of crystals, and in this 

 case the interstitial material of late crystallization would be the 

 same salic material as that found in the upper layers when sinking 

 of crystals took place. If a certain amount of both zoning and 

 sinking of crystals took place, a body would result showing the salic 

 differentiate both as interstitial material and as a separate upper 

 layer. The same possibilities of the formation of peridotite and 

 spinel-bearing peridotite obtain in this case with the same conditions 

 favoring their formation as those discussed under the anorthite, 

 forsterite, silica system. 



It has been possible, then, to deduce from facts ascertained 

 experimentally the crystallization with quick and slow cooling of 

 mixtures which give results closely analogous to the occurrence 

 observed in nature of diabase in small dykes and small sills (quickly 

 cooled) and of diabase with micropegmatite interstices or a granitic 

 or granodioritic differentiate in larger bodies (slowly cooled). 

 There are many differences and complications in the natural 

 magma in the matter of details, but it is clear that the broad scheme 

 is well understood and that crystallization is the sole control. 

 There is no necessity for assuming that assimilation of siliceous 

 material is essential to the formation of the salic differentiate, nor 

 that its separation is accomphshed by the process of liquid immisci- 

 bihty. The Palisade diabase sill with its "ledge" rich in sunken 

 oHvine crystals near the base and the micropegmatite interstices 

 in the main mass of the diabase is a case in which the working 

 of the process is clearly exhibited.^ 



^ J. V. Lewis, Ann. Rept. State Geologist, New Jersey, 1907, p. 125. 



