310 PROCEEDINGS OF THE AMERICAN ACADEMY. 



differentiates of a magma are richest in those compounds which are 

 the first formed secretions from the magma, and observe that the 

 soda-potash-feldspar is clearly the first mineral to crystallize in this 

 porphyry, we may infer that the feldspar was the first mineral to 

 crystallize, and that the possible appearance of the quartz first, may 

 be due to a re\'ersal of the normal order through supercooling, a 

 condition that did not exist to the same extent at the deeper levels. 



In the extreme contact phases at higher levels we find pseudomorphs 

 of what were pretty certainly pyroxene crystals: we have also noted 

 the occurrence of a few phenocrysts of a calcic-iron pyroxene in the 

 granite-porphyry, and that of a more highly sodic variety of pyroxene 

 in the form of small crystals included in the outer parts of the feldspar 

 phenocrysts, as well as in the form of separate crystals. In the rhom- 

 benporphyry a calcic-iron pyroxene began to crystallize before the 

 phenocrystalline feldspar had ceased its growth. From these facts 

 we infer that the less sodic pyroxenes began to crystallize shortly 

 after the feldspar. 



After a considerable amount of quartz and feldspar with small 

 amounts of pyroxene (possibly aenigmatite) had grown as phenocrysts, 

 there occurred a sharp and clearly marked pause in the crystallization 

 of these constituents. The cause of such a pause is undoubtedly to be 

 found in the supersaturation of the residual liquor, a phenomenon 

 which appears to the writer to be an inevitable incident in the crystal- 

 lization of a quickly chilled rock magma, such as that of the granite- 

 porphyry unquestionably was. It is clear that before and perhaps 

 during this pause the magma was in movement, for we find fragments 

 of feldspar phenocrysts as well as the broken and irregular ends of the 

 crystals covered by the later growth, itself unbroken. Furthermore, 

 the sealed cracks crossing so many of the phenocrysts extend only to 

 the edge of the original crystal, but not across the later rim. During 

 this pause it is possible that the resorption of the quartz phenocrysts 

 took place. It is also possible that the pyroxene (of more sodic type) 

 continued its groMth, and the massive parts of the hornblende began 

 their growth, although their crystallization appears to belong to the 

 second period of crystallizating activity, for they never come normally 

 in contact with the original phenocrystalline feldspar but always lie 

 outside of the later groundmass addition to the feldspars. As we 

 should expect in a supercooled liquid, the crystallization when it was 

 resumed took place about vastly more numerous centers, and so far 

 as can be told there was a practically similtaneous growth of the vari- 

 ous minerals. A portion of the feldspars attached themselves as a 



