146 W. JF. Ilillebrand — Composition of Yttrialite. 



into UO a , thus throwing it with the Th0 2 , where it naturally 

 belongs. The agreement is indeed "gar nicht gut." 



Benedicks has made the grave mistake of counting Mackin- 

 tosh's monoxide bases a second time, thus making a basic salt 

 B" W\ Si 4 15 instead of the normal one B //V 4 Si 4 14 , to 

 which Mackintosh's results closely conform. 



Discussion of Benedicks' formula for Thalenite. 



Moreover, in the light of the data furnished by Benedicks 

 himself it cannot be admitted that the formula H 2 Y 4 Si 4 15 „ 

 for thalenite is established. 



Water was determined by him according to Penfield's meth- 

 od,* but without any hint as to the particular modification 

 employed. If, as seems probable, the water expelled from the 

 mineral was caused to recondense in the cooler part of the 

 ignition tube, the latter being then weighed and again after 

 driving the condensed water out, two serious sources of error 

 have to be considered: (1) The C0 2 present in the mineral, 

 which would count in part as water unless a very careful cor- 

 rection was made, as provided for by Penfield. No mention 

 is made by Benedicks of any such correction. (2) Nitrogen 

 and helium are said to comprise 1*4 per cent of the mineral by 

 weight. If so, these would introduce an error in the above 

 water determination of contrary sign to that due to C0 2 , and 

 if the proportion of helium were large this error might be of 

 very considerable magnitude. 



In an appendix to his paper Benedicks gives an analysis of 

 what he considers to be a very pure form of thalenite. He 

 makes no comparison of this with his earlier analysis, nor does 

 he deduce a molecular ratio, which I find to be 1 : 2"6 : 5*15, 

 or 1 : 3*03 : 6*02 if small amounts of lime, magnesia, and soda 

 are neglected, instead of 1 : 2 : 4 as required by his formula. 

 There being no C0 2 in this purer material, the value for water 

 (if determined as above surmised) may be supposed to be 

 affected only by the error due to nitrogen and helium. It will 

 be seen that the neglect to regard lime, magnesia, and soda in 

 his second analysis affects the ratio very seriously. This neg- 

 lect may be justified in figuring on his first analysis because 

 of an approximate balancing by C0 2 , but it would be by no 

 means so in the other in spite of the very satisfactory ratio 

 obtained and leading to the empirical formula B/'K'",. Si 6 22 , 

 which is susceptible of a variety of interpretations. It may 

 represent a basic salt of diorthosilicic acid B" B //; 5 (B //; Oy 

 Si 2 7 ) 3 or of metasilicic acid B" K'", (K^O)', (8iO,)„ or pos- 

 sibly even of other acids. 



* This Journal, xlviii, 31, 1894. 



