SILICON. 18 ( J 



This almost coincides with the analysis of Do- 

 bereiner. 



But waving this probable explanation of the 

 excess, let us calculate the atomic weight of si- 

 lica, on the supposition that the mineral contains 

 only simple silicates, and that the analysis is quite 

 correct. The excess of the silica must be divid- 

 ed between the two silicates in the ratio of 14*22 

 : 15*55 ; or we must add 1-42 of it to the silica 

 united to the manganese, and 1*3 to the silica 

 united to the iron. This gives us silicate of iron 

 composed as follows : 



Gives 

 Silica . . 15-52 or 2-1827 2-1827. 



Protoxide of iron 32 - 4-5 



Let us now take a mean of the weight of an 

 atom of silica, deduced from the analyses of the 

 8 minerals which we have just examined. 



1 Table spar gives the atom of silica 1-96800 



2 Bucholzite . 1-997595 



3 Nepheline .... 1 94582 



4 Hedenbergite .... 1-90350 



5 Silicate of zinc . . . 1-99800 



6 Dioptase 2-00000 



7 Bisilicate of manganese . . 2-02399 



8 Knebelite .... 2-18270 



Mean 

 Mean . 2-00245 2-00245. 



This weight exceeds 2 only by T oVo th part. The 

 atomic weight given by Berzelius, from his own 



