Washington — Submarine Eruptions of 1831 and 1891. 137 



about Ab,Au 9 or Ab 1 An 3 . These feldspars are highly euhe- 

 dral and the laths show well-formed terminal planes. The 

 tables sometimes surround portions of olivine and augite phe- 

 uocrysts, but carry few inclusions, mostly of black dust. Oli- 

 vine is rather abundant, in subhedral, and often highly euhedral, 

 equant individuals, up to 0*5 mtn in diameter. It is quite fresh 

 and colorless, and carries few inclusions of ores and dust. 

 Augite is rather less abundant than the olivine, in anhedral to 

 subhedral, equant individuals, up to 05 mm , or rather larger. 

 It is almost colorless, but slightly yellowish or greenish, and 

 contains few inclusions. Opaque grains of magnetite, presum- 

 ably titaniferous, are not uncommon, though small in size. 



Under low powers the groundmass is black and quite opaque, 

 but high powers resolve it into a colorless, isotropic glass, 

 thickly sprinkled with very minute black dusty grains. In 

 general characters my specimen resembles that of Foerstner 

 which he obtained from Gemmellaro, though the glass in his 

 was coffee-brown. 



Chemical composition. — The results of my analysis are shown 

 in column I below. The analyses of Foerstner and Abich are 

 given also in II and III. 



Analysis I is, in general terms, that of an ordinary feldspar- 

 basalt, and is remarkable in itself chiefly for the rather high 

 Ti0 2 and the presence of considerable nickel. The high 

 amount of FeO as compared with the Fe 2 3 is also a note- 

 worthy feature, which will be referred to again. 



Foerstner's analysis (II) corresponds in general very well 

 with mine, showing but slightly more Si0 2 and Fe a 8 and a 

 little less Na 2 and K 2 0, while the differences in the figures 

 for FeO, MgO, and CaO are somewhat larger, but sufficiently 

 close to indicate the practical chemical identity of the two 

 specimens.. His much higher figure for Al 2 3 is, of course, to 

 be ascribed to the non-determination of Ti0 2 and P 2 5 , both of 

 which would be weighed with the alumina. In a later paper,* 

 discussing the lava of the submarine eruption of 1891 in the 

 analysis of which he determined Ti0 2 , he considers that in his 

 analyses of the Graham Island and Pantelleria basalts the Ti0 2 

 was weighed with the silica, which should be corrected for 

 about 5 per cent of titanium dioxide. That this supposition is 

 incorrect is shown by the figures in analyses I and II above, as 

 by the facts of analysis. The residue left on correction of the 

 silica for impurities by evaporation with sulphuric and hydro- 

 fluoric acids consists largely of alumina and ferric oxide, and 

 does not contain all the titanium. f It may be noted in this 

 connection that this residue in my analysis amounted to only 

 1*73 per cent. 



* H. Foerstner, Tsch. Min. Petr. Mitth., vol. xii, p. 520, 1891. 



f Cf. W. F. Hillebrand, Bull. U. S. Geol. Surv. ; No. 305, p. 80, 1907. 



