Hyndman and Bonne// — Analysis of Spherulites. 367 



surrounding the better-formed spherulites. This material possessed 

 the sea-green colour mentioned above, and when finely powdered was 

 white, while the spherulites gave a reddish powder. It was found 

 that quite a minute portion of a partly-formed spherulite produced 

 a distinct tinge. The results were as follows : — 



RO : E2O3 : Si0 2 ... 2-74 : 6"73 : 39-05 2-345 : 4'16 : 41-81 



Including H 2 O 3-45 3"68 



The method of analysis employed for all except the alkali metals 

 was thorough fusion in a platinum crucible with mixed alkaline 

 carbonates. The spherulite gave a light greenish mass, and the 

 matrix a still lighter one ; they both dissolved easily in dilute 

 hydrochloric acid, and were evaporated to dryness twice with this 

 acid. The sulphuretted hydrogen precipitate was brownish black 

 in both cases. After the separation of the aluminum and iron 

 hydroxides, the nickel was precipitated by ammonium sulphide, and 

 the amount being small it was weighed as sulphide. The double 

 determinations of the silica were obtained by treating the preliminary 

 qualitative examination as quantitative, until the silica had been 

 separated. 



In each case the powdered rock, spread out on a watch-glass, was 

 heated for some hours, to about 110° F. ; the loss of weight was small, 

 only amounting to -36 per cent, in the matrix. The powder was 



1 Mean of two. 2 All the iron is calculated as Fe O. 



3 Another calculated as Na 2 O = 7-48. 



. _, , . . , , c ! oxide in rock x % oxygen in oxide 



4 I he oxygen ratio is equal to — ^qq 



