GENERAL DESCRIPTION 259 



precise data could not be obtained on the fine particles. Occasional 

 minute, elongated bubbles (up to 0.05 mm. in length) were also observed. 

 The density (referred to water at 4° C. and to vacuum) of the obsidian 

 is 2.387. 



CHEMICAL CHARACTERISTICS 



For the chemical analysis of the black obsidian I am indebted to Mr. 

 J. B. Ferguson, of the Geophysical Laboratory, and express herewith my 

 appreciation of his kindness. The material selected for analysis was part 

 of specimen 88428," a jet black glass free from spherulites and litlio- 

 [)hys8e, but containing many of the minute crystallites noted above. The 

 analysis is that of a fairly normal rhyolitic obsidian. Interesting and 

 important is the presence of CI and SO3 in appreciable amounts. It will 

 be shown later that the release of these volatile components in the magma 

 had much to do with the formation of the lithophysae, while the character 

 of the physico-chemical system thus produced caused the simultaneous 

 formation of crystals of fayalite and of tridymite at relatively low tem- 

 peratures. This mineral association is not that of ordinary igneous 

 I'ocks or lavas, but seems to be characteristic of lithophysae in obsidian, 

 notwithstanding the low content in oxides of iron (from 3 to 4 per cent). 

 'I'll us in 1827 Gustav Rose discovered fayalite and tridymite in the litho- 

 physae of the obsidian from Cerro de las Navajas (analysis VT : FeO -j- 

 Fe2O3 = 2.20 per cent) ; in 1885 Iddings found fayalite under similar 

 conditions in the lithophysa} of Obsidian (Tiff, Yellowstone National 

 Park (analysis IV; FeO + Fe.Og = l.(iH per cent). 



The occurrence of an orthosilicate like olivine with tridymite is rare, 

 if not unknown, in intrusive rocks. It is less rare in ciriisive rocks and 

 indicates that physico-chemical conditions of equilibrium at the time of 

 formation of the crystals may be very different even for magmas of the 

 same general total chemical composition. The computed noi'mative c<mi- 

 position in such cases would be the same, but the actual modal composi- 

 tion may be totally different, thus emphasizing the difficulty of setting 

 ii|) a pioper normative composition which even approximates the actual 

 minerid composition of the i-ock. 



'"The specimens described iu this paper have been deposited in the V. S. National 

 .Muiieiim ; the nuniher of the si)ecinien is, in eacli case, its cataloffiie number In the 

 National Museum. 



