Washington — Submarine Emotions of 1831 and 1891. 145 



SiO„..-. _.. 44-83 



A1 2 3 



Fel0 3 



FeO .. 



MgO 



CaO 



Na„0 



H 2 + 



H o 0- 



k;o 



Ti0 2 



PA--- 



MnO 



I 



II 



III 



la 



44-83 



44-64 



46-40 



747 



11-73 



12-74 



21'84 



115 



1-35 



4-21 



9-53 



009 



11-79 



11-17 



2-04 



164 



5-50 



5-82 



5*37 



•138 



9'63 



10-12 



10-33 



•172 



3-34 



4-31 



3-27 



•054 



0-81 ) 

 0-10 j 



0-51 



n.d. 











1-40 



1-41 



1-69 



•015 



6-88 



5-86 



n.d. 



•086 



214 



n.d. 



n.d. 



•015 



0-20 



0-20 



n.d. 



•003 



99-70 



100-99 



100-47 





Analyses of basalt scoria from Foerstner Volcano (1891) : 



I. "Analysis by H. S. Washington. 

 II. Analysis by H. Foerstner, Tsch. Min. Petr. Mitth., xii, 

 p. 512, 1891. 



III. Analysis by G. H. Perry, Nature, vol. xlv, p. 252, 1891. 

 la. Molecular ratios of I. 



The amount of nickel was not determined, but it is probably 

 about the same as that found in my analysis of the Graham 

 Island basalt (about 0'10 per cent), as shown by the depth of 

 the greenish tint of the filtrate from the ammonia precipitate. 



Foerstner' s analysis (II) agrees very closely with mine in 

 most respects, and especially shows about the same amount of 

 Ti0 2 and a similar very high ratio of ferrous to ferric oxide, 

 though his figure for the latter is decidedly higher than in I. 

 His alumina must be corrected for one or two per cent of P 2 5 . 



The alumina of Perry's analysis (III) must be corrected for 

 about 9 per cent of Ti0 2 and P 2 5 , the subtraction of which 

 would bring the figure for alumina close to those of the other 

 analyses. The figures for the iron oxides are reversed as com- 

 pared with those of Foerstner and myself, and must be 

 regarded as incorrect, especially as Perry himself expresses 

 doubt as to their accuracy.* The somewhat higher silica may 

 be due to its not having been corrected for impurities by 

 evaporation with HF. The other figures are closely concord- 

 ant with those of the other two analyses. 



Classification. — The norm of I is calculated to be as 

 follows : 



* He remarks, "As the powder was magnetic, Fe 2 3 was probably com- 

 bined with FeO to form Fe 3 4 . This would give FeO 7*55, Fe 2 3 " 4'18." 

 The basis of calculation for these fig-ares is not clear ; as they yield a molecu- 

 lar ratio of FeO : Fe 2 3 =4:1. 



