W. F. Ford — Chemical Composition of Axinite. 197 



Discussion of the Analyses. — In considering the results of 

 the analyses it is first necessary to determine the roles played by 

 the boric oxide and the water in the mineral. The boric oxide 

 in each analysis gives a simple and constant ratio with the Si0 2 

 and other oxides, that of Si0 2 to B 2 3 being 8 : 1, and there 

 seems to be no doubt but that it is always present in definite 

 proportion and not as an isomorphous constituent replacing 

 A1 2 3 . Concerning the water, the percentages in the two 

 analyses are small, and the ratios of Si0 2 : H 2 are 8 : 0*86 and 

 8 : 0*77, respectively, in both cases falling short of 8:1. More- 

 over, the ratios of Si0 2 to the protoxide bases, less the water, 

 are not constant, being in the two analyses 8 : 6*12 and 8 : 6*23 

 respectively. If, however, the water is regarded as basic and 

 isomorphous with the other protoxide bases, there are obtained 

 such clear and exact ratios in both cases as to leave no doubt 

 that this is the correct method of interpreting the results ; 

 moreover, the high temperature necessary to drive off the 

 water and the study of other analyses, which will be quoted 

 later, only serve to strengthen and reinforce this conclusion. 

 Consequently in obtaining the ratios from the analyses and in 

 deducing the formula of the mineral from them, the water has 

 been considered as basic and as isomorphous with the protoxide 

 group. According to this interpretation the two analyses give 

 very satisfactory results, the ratios yielding almost exactly whole 

 numbers. The ratios are as follows : 



Si0 2 



B 2 3 



R 2 3 



RO 



8-00 



0-97 



1-99 



6*99 



8-00 



0-92 



1-99 



7-00 



These ratios give the formula, which may be expressed as 

 an orthosilicate, K" 7 E in 4 B 2 (Si0 4 ) 8 ; K 11 being chiefly Ca, with 

 varying amounts of Mn, Fe, and Mg, and H 2 in small propor- 

 tions. R lu is chiefly Al, but replaced always by a little Fe. 



Discussion of Former Analyses. — The two analyses by 

 Genth of the beautiful yellow axinite from Franklin seem to 

 have only one defect, namely, that the water is given as loss 

 on ignition ; hence the percentage of MnO being high, it would 

 be expected that the partial oxidation of this constituent would 

 cause the results of the water determinations to be low. On 

 studying his analyses they were found to give ratios closely 

 agreeing with those demanded by the new formula with the 

 exception of the protoxides, which were lower than required. 

 It was decided, therefore, to make direct determinations of 

 H 2 on the material in the Brush collection, the results being, 

 as was expected, considerably higher than those given by 

 Genth. The corrected analyses are given below and the ratios, 

 as may be seen, are fully in accord with those given on page 



