272 SCIENCE PROGRESS 



name of 'Kyschtymite, shows a silica percentage as low as i6*8o, 

 and yet it is a persalane. Indeed, rocks consisting of practically 

 pure corundum, such as have been described from India, 1 would 

 still be persalanes, though they are almost free from silica. The 

 range of the alumina might theoretically be as great as that of 

 the silica. Apparently Kyschtymite has about 70 per cent., and 

 the Indian rocks nearly 100, while intrusive quartz has none. 

 The alkalies, too, may exceed 21 per cent., as in the soda of 

 a rock such as is known to exist composed almost entirely of 

 nepheline, and may fall to zero. 



The rocks of this class have therefore nothing in common 

 except the fact that they contain less than 12*5 per cent, of 

 ferromagnesian silicates and other minerals classed as femic. 



In the other classes there is a similar wide range of silica 

 percentage: in the dosalanes from 75*40 to 38*11; in the 

 salfemanes from 73*60 to 36*36; in the dofemanes from 53*05 to 

 31*59 and even 9*79 in the case of certain iron ores which are 

 believed to be of igneous origin ; and in the perfemanes from 

 55*52 in a lherzolite to 0*76 in an iron ore. 2 The limits between 

 which the silica might possibly vary would, I need not say, be 

 still wider. The other oxides have ranges almost as great. 



Even the total amount of iron, magnesium, and calcium 

 oxides, which is the most constant character of the classes, shows 

 wide variations. The persalane analyses range from 1*05 to 

 22*07 P er cent., and might theoretically reach 30 per cent. In 

 the dosalanes the percentage varies from 2*44 to 30*03, and 

 might be as high as 50. In the salfemanes it ranges from 

 13*21 to 39*87 per cent., with a higher limit of 70; in the 

 dofemanes, from 33*32 to 65*90, with a possible percentage of 90; 

 while in the perfemanes it varies between 41*16 and 72*55, and 

 might rise to 100 per cent. 3 These figures will not be materially 



1 See F. R. Mallet, Records Geol. Surv. India (1872), vol. v. p. 20 and vol. vi. 

 (1873) p. 43; and The Manual of the Geological Survey of India, Mineralogy, 

 (Calcutta, 1887) p. 48 ; and Prof. J. W. Judd, " On some Simple Massive Minerals 

 (Crystalline Rocks) from India and Australia," Alineralogical Magazine, vol. xi. 

 (1895) p. 56. At the Imperial Institute there are specimens of microcrystalline 

 corundum from the Malay Peninsula which might also be classed as a rock. It 

 is not clear, however, whether these rocks are of igneous origin. 



2 See Washington, op. cit., pp. 218, 306, 310, 350, 354, 364 ; Iddings, Chem. 

 Comp., supra cit., pp. 59, 62. 



3 See Washington, op. cit., pp. 142, 206, 218, 292, 312, 338, 354, 364, 366, and 

 368. 



