of Chondrodite, Humite and Clinohumite. 197 





I. 



II. 



III. 



IV. 



V. 



VI. 



VII. 



VIII. 



Sp. gr. 



3-190 



3-125 



.... . 















3-22 



Si0 2 



33-26 



33-92 



33-63 



33-60 



33-66 



33-13 



34-01 



34-08 



MgO 



57-92 



59-33 



58-13 



53-05 



53-26 



54-71 



5497 



53-44 



FeO 



2-30 



1-11 



3-88 



8-67 



6-82 



4-95 



4-62 



7-22 



F 



5-04 



2-59 



5-20 



5-58 



4-24 



4-99 



4-56 



4-01 



H 2 









1-37 



1-31 







•55 



•61 







A1 2 S 



1-06 



■97 











•67 



•68 



•71 



•44 



CaO 



•74 



---- 



Fe 2 3 -12 



.... 



MnO '75 



= F 



100-32 



2-12 



98-52 

 109 



102-21 102-33 

 2-19 2-35 



98-65 

 1-78 



99-76 

 2-L0 



Cal.HsO 



98-20 

 2-94 



97-43 



4-19 



100-02 9998 



1-58 1-24 



9687 



3-12 



97-66 



2-29 



100-29 99-19 

 1-92 1-69 



9837 97-50 



2-45 3-27 



101-14 101-62 101-60 101-22 99'99 99-95 10082 100'77 



With two exceptions, analyses III and IV, it will be ob- 

 served that after deducting the oxygen equivalent to fluorine 

 the analyses add up to only about 96 to 98 per cent, and it is be- 

 lieved that this loss must be due to water which was not deter- 

 mined. Taking the protoxides, MgO and FeO, as a basis of 

 calculation, we have determined the amounts of water which 

 should be added to the analyses in order to yield the ratio 

 RO : (F-f-OH) = 5 : 2 established by us. These quantities 

 {Cal . R„G) when added give very satisfactory results. In all 

 of the analyses except those of Wingard, III and IV, the 

 summation is nearer 100 per cent than before and this is espe- 

 cially shown in the last four. This method of treating the 

 water is probably fairer than to derive it from the deficiency 

 in the analyses, as by the latter treatment, all of the errors of 

 analysis would be thrown upon this constituent, which, owing 

 to its low molecular weight, needs to be very accurately deter- 

 mined to yield a sharp and correct ratio. The percentages of 

 Si0 2 vary from 33-13 to 34*08 while in our analyses they run 

 from 33*33 to 33*87. In the following table, where the ratios 

 of Si0 2 : RO are given, it may be observed that they approxi- 

 mate closely to 2 : 5 as required by our formula. 





Si0 2 



RO 







Si0 2 



•RO 





I. 



■554 



1-489 = 1-88 



5 



V. 



•561 



: 1-426 = 1-96 



. 5 



II. 



•565 



1-507= 1-88 



5 



VI. 



•552 



1-447 = 1-91 



5 



III. 



•561 



1-507= 1-86 



5 



VII. 



•567: 



1-448= 1-96 



5 



IV. 



•560: 



1-446 = 1-94 



5 



VIII. 



•568 



1-436= 1-98 



5 



It is very gratifying to be able to show that the formula 

 established by us is thus substantiated by the analyses of 

 previous investigators. 



