Pirsson — Corundum-bearing Rock from Montana. 423 



should now no longer be urged. This idea was that such 

 corundums had been torn loose from some place below where 

 they had previously existed, and being infusible had spread 

 themselves through the magma. Others again recognized in 

 these corundums infusible but recrystallized portions of rock 

 fragments enclosed in the magma, other portions being con- 

 verted into spinel, cordierite, etc. Lagorio points out, how- 

 ever, that this could not be the case, as corundum dissolves 

 in molten glasses, and he calls attention to the confusion which 

 has existed between fusibility of compounds in molten masses 

 and their solubility in the same, the two being quite distinct. 

 The characteristic form of corundum occurring with igneous 

 rocks is the thin, flat, hexagonal table with low rhombohedron, 

 described in the following paper. 



This occurrence at Yogo Creek is an important addition to 

 the list of pyrogenetic corundum. The clear-cut form of the 

 crystals and their general distribution shows that they have 

 crystallized out of the magma with as much certainty as the 

 well formed phenocrysts of feldspar in a porphyry betray 

 their origin. 



The general character of the rock, however, and its close 

 relationship to the minettes and shonkinite of the region 

 shows that it could not originally have been sufficiently rich 

 in alumina to have allowed a general separation out of corun- 

 dum. The condition of it, as mentioned above, shows that the 

 magma took up great quantities of inclusions from the sedi- 

 ments through which it passed. Among these sediments must 

 have been a great, though unknown, thickness of the Belt for- 

 mation, consisting of clay shales. This formation lies between 

 the Archaean gneisses and the lowest beds of recognized 

 Cambrian. The liability of the beds to be shattered by igne- 

 ous rocks ascending through it and included as fragments, has 

 already been shown elsewhere.* 



Such included fragments of shale, if the magma maintained 

 its heat sufficiently, as confined in dike form it naturally would 

 do, would eventually be dissolved, as the experiments described 

 show. There would thus be formed local areas in the magma 

 very rich in alumina, which, on cooling, would allow crystals 

 of corundum to separate out. This explanation seems to us 

 most in accord both with the facts observed in the field and 

 those obtained by experiment in the laboratory. The form of 

 the crystals is also in accord with that of the pyrogenetic 

 cerundums. 



This occurrence then agrees well with the experiments and 

 views of Lagorio and is indeed an important confirmation of 

 them. 



Mineralogical Petrographical Laboratory, Sheffield Scientific School, Tale Uni- 

 versity, New Haven, June, 1897. 



* Geology of Castle Mt., Bull. 139, U. S. Geol. Survey, p. 72. 



