574 Mr. w. a. mciilHDsox: a MiOKOiiETEic [vol. Ixxix, 



granite in its mineral content than that of other areas. The 

 outer zone, noteworthy for hiotite and plagioelase, could be ex- 

 plained either by Dr. A. Harker's localization of crystallization, 

 or by Dr. N". L. Bowen's conception of the consolidation of the 

 lirst intruded phase. It differs, however, as do the other districts, 

 by having a central orthoclase-zone with a more quartzose margin : 

 that is, by the outer concentration of a mineral of the latest stage 

 of separation. This feature, as will be seen, also finds explanation 

 on the grounds of a pre -intrusive crystallization. 



This cpiestion of the high negative correlation of quartz and 

 orthoclase, and the marginal position of the more quartzose zones, 

 affects all the areas dealt with in this paper. Sinking of the indi- 

 viduals of earlier separation does not account for the phenomena : 

 for, in that case, plagioelase shotdd be centrally concentrated, 

 especially as plagioelase has a higher specific gravity than orthoclase. 

 On the contrary, there is a very small negative correlation between 

 quartz and plagioelase, and their positive areas overlap, even if 

 they do not quite coincide. Moreover, marked central concen- 

 tration of plagioelase is seen only in the St. Stephen's area. 

 Some other explanation must be sought. Examination of geo- 

 logical sections across batholiths shows the tendencj^ to develop 

 cupolas in the central parts of the roof. This leads to the con- 

 clusion that the foundering of the roof, whatever may be the 

 cause, takes place from the centre of the sides. In these circum- 

 stances, the magmatic stream will flow most rapidly as a central 

 column, spreading sideways to fill the chamber as the outer parts 

 of the roof fall in ; and, where the stream is deflected at the 

 roof, a sudden decrease in velocity results. If such a stream were 

 charged with crystalline elements, including mica-flakes, small 

 plagioclase-crystals, and larger perthite-plates, it is evident that 

 the check in velocity would cause a partial dropping of the load at 

 this point. Despite their higher specific gravity, the thin mica- 

 fiakes would be carried on ; ■ the large perthite-plates would be 

 left ; some of the denser plagioclase-crystals would perhaps be 

 carried forward, and some remain in the central parts. Such 

 crystal-sorting by streaming explains the central disposition of 

 the perthites, but it implies that at the time of intrusion into the 

 present position crystallization had proceeded to some length. The 

 mother-liquor would then be siliceous, and the remaining products 

 of crystallization would be partly orthoclase, but largely quartz. 

 If the central region Avere more completely occupied by large 

 perthite-plates than the margins, it follows that the marginal zone 

 after completed consolidation will contain more quartz, so 

 explaining the negative correlation between the two minerals 

 that predominates in this district. 



The concentration of mineralizers in the western areas has had 

 other effects besides the formation of peculiar minerals. The 

 occurrence of narrow positive zones of pneumatolytic minerals 



