in the Mineral Series. 351 



Silica 29-54 = 29-66 



Sesquioxide of iron . . 23-08 = 23-25 



Protoxide of iron . . . 31-25 = 31'49 



Protoxide of manganese . -1*35 = 1'46 



Lime 14-04 = 14-14 



99-26 100-00 



Now 23-25 parts of Fe 2 O 3 require 10-46 parts of FeO to form 

 magnetic iron ore ( = FeO, Fe 2 O 3 ). This, deducted from 31*49., 

 leaves 21 '03 for the protoxide of iron present (on the above sup- 

 position) in the silicate. The analysis consequently stands as 

 below : — 



Silica 29-66 = 44-75 



Protoxide of iron . . . 21-03 = 31-72 

 Protoxide of manganese . 1'46 = 2*20 

 Lime 1404 = 21-33 



66-29 100-00 



Calculating the oxygen-ratios of the above, we obtain — 



SiO 2 . ....... 23-24=l| = 7=14 



FeO . --. . r\ 



MnO. . . A, . . . 13-59 = 1 =4= 8 



CaO . . . .J 



Here, then, we have 8 atoms of RO to 14 atoms of SiO 2 , in place 

 of 2 to 1 , as required by the chrysolite7ormula. If the oxygen- 

 ratios in any of the above analyses had come out as 9 : 6 : 14, 

 we might have written the formula 



7(RO, SiO 9 ) + 2(FeO, Fe 2 O 3 ) • 



but the analyses do not yield these values. The view, therefore, 

 as suggested above, that the peculiar composition of Lievrite 

 may arise from the presence of magnetic iron ore, cannot be 

 sustained. 



But the formula of Lievrite may be brought to coincide with 

 that of the chrysolite series by assuming the iron to have been 

 originally present in the condition of FeO only. If this be 

 assumed — and the assumption is in part warranted by the well- 

 known fact that Lievrite is especially subject to alteration, the 

 conversion of the FeO into Fe 2 O 3 still going on in many speci- 

 mens — the difficulty is at once removed. Taking the mean 

 composition as given above, and calculating the weight of FeO 

 corresponding to the 23-25 per cent, of Fe 2 O 3 , we obtain 20*92. 

 Adding this to the FeO, and correcting the whole to 100 

 parts, the analysis reads as follows : — 



