398 



P. EE TIE Y OX TEE MICEOSCOPIC CHAEACTEES 



crack, may hare been formed subsequently to the crystal which is 

 traversed by that crack. What are we to infer from these crack- 

 begirt crystals ? Surely we have here something like the contraction 

 experimentally determined by Bischof in rocks of precisely similar 

 constitution. The crystal is a constituent of trachyte ; it may be 

 regarded even as a small piece of trachyte developed within obsidian. 

 The strain upon the surrounding glass, engendered by the formation 

 of this crystal, is, I have no doubt, the cause of the rupture in the 

 surrounding glass ; and the gape of the fissure is the measure of the 

 contraction for any one plane. The polarization-figures seen within 

 the perlitic areas surely, then, bespeak the incipient strain which 

 heralds the production of the crystal ; and I believe that the 

 rounded boundaries and the curved cracks, so frequently seen within 

 the angles of olivine and other crystals developed in vitreous rocks, 

 are but another expression of the laws which crystallization enforces 

 in a surrounding amorphous mass. In studying these questions we 

 seem almost to stand upon the threshold of crystallography ; and I 

 may well close this paper with the hope that the subject may be 

 taken up by more able hands : for it requires further investigation, 

 since there are periites in which no apparent relation exists between 

 the perlitic structure and the porphyritic crystals — as in some of the 

 periites of Chemmitz, where even a score of small perlitoids may be 

 seen to abut against the margin of a single crystal of sanidine or 

 of magnesian mica. In these rocks we have, indeed, always assumed 

 that the development of perlitic structure was subsequent to the 

 formation of the porphyritic crystals, and not approximately syn- 

 chronous with it, as I think has been the case in the materials of 

 the obsidian tuff just described. In an inquiry of this kind thick 

 sections, as well as thin ones, should be prepared and examined, 

 and conclusion after conclusion discarded if needful, until the truth 

 is reached. Indeed, I am not unwilling to believe that some of the 

 phenomena which we have just discussed may be due to different 

 operations of one force, or of different forces ; and, just as the same 

 mineral may sometimes be formed by a wet or by a dry process, so 

 it is possible that similar structures may not always be due to the 

 same cause. 



Appendix. — Microscopic Characters of Volcanic- Rocks of Montana. 



2s o. 9 (the rock from Mount Washburn) is an andesite which seems to 

 be intermediate in character between the augite- and the hornblende- 

 andesites, and approximates in mineral constitution to a basalt. 



The constituents are chiefly triclinic felspar, augite, hornblende, 

 magnetite, and vitreous matter. 



A little sanidine also appears to be present. The augite and 

 hornblende crystals generally show a dark border when examined 

 under the microscope with substage-illumination. By reflected light 

 these borders appear of a bright rust-red, and are doubtless due to 

 marginal decomposition of the hornblende and augite. These stains 

 and granules of peroxide of iron impart a reddish brown- colour to 

 the rock. The vitreous matter constitutes a considerable proportion 



