288 BERNARD F. DAVIS. 



"A small percentage of iron is present as ferric oxide. 

 The BeO includes any alumina which may be present, 

 although I found none in the qualitative analysis. This 

 result seems very high, and I think the high total is prob- 

 ably due to this. 54*16 represents the percentage of the 

 oxides together, obtained by igniting the oxalates with 

 addition of a drop of nitric acid at the end. The per- 

 centage of yttrium is probably slightly lower and the lan- 

 thanum and didymium correspondingly higher. 



"Dr. Norman Collie very kindly examined the mineral for 

 helium and other gases. He says that 10 grams on heating 

 gave about 10 cc. of 0O 2 and 10 cc. of hydrogen, a little 

 nitrogen and about one bubble of helium. No argon. In 

 his first experiments he thought he detected argon, but it 

 must have been an impurity from the air. 



"The method of analysis was as follows : — Decomposition 

 of the mineral with hydrochloric acid and estimation of the 

 silica in the usual way. Filtrate made up to known volume 

 and measured quantities taken for different estimations. 

 In the slightly acid solution the cerium* and yttrium earths 

 were precipitated with excess of oxalic acid while boiling 

 and allowed to stand usually for twelve hours, sometimes 

 two or three days. The filtrate containing iron and 

 glucinum was evaporated with sulphuric acid and after 

 taking up with hydrochloric acid, ammonia added and the 

 iron and glucinum oxides weighed together. The iron was 

 estimated volumetrically. The oxalates of the cerium and 

 yttrium elements were converted into sulphates and 

 separated by standing two or three days with saturated 

 solution of pot. sulphate. The cerium was estimated both 

 by fusion of the nitrates with nitre and also by Mosander's 

 method. I did not think it necessary to go over the 

 analysis again to correct these errors or verify the result 

 as given. For our present purposes it is sufficient, and 



