1825.] M. Berzelius's Hypothesis of the Atomic Theory. 353 



" Further to illustrate our analysis, let us insulate each of the 

 species of garnet contained in the mixture. If we employ the 

 trioxide of iron, to make a melanite garnet {en faisaut lui grenat 



melanite) of the formula 2 F tSi + Ca' iSi', we must take a 

 quantity of 



Trioxide of iron, whose oxygen = 2-3 corresponds to /'^O 



Silica 4-(J !J-14 



Lime 2-3 8-19 



24-83 



" There remains a portion of lime whose oxygen is o 84, with 

 which we may make a giossular garnet {dont on peut {aire un 



greiKjf grusf^iihtire) of the formula 2 A Si 4 Ca' Si-, by taking 

 a quantity of 



Silica, whose oxygen = IkG^^ corresponds to 23-22 



Alumina ,r84 12'o0 



Lime 5-84 20-79 



5ti-ol 



" There then remains a quantity of silica whose oxygen is 

 2"33, and to use it up (pvnr Cempluyei), we uray tirst make a 

 garnet of alumina and magnesia [an pent fa ire d'abord un grenat 



d'alitntiiic et magnesie) of the fornmla 2 A bi 4- M' Si-, by tak- 

 ing a quantity of 



Silica, whose oxygen = 0-92 corresponds U) I -83 



Alumina .\ . . . 0-4t> 0-98 



Magnesia 0-4G |-1'J 



4-00 

 " Lastly, we shall make of the remainder {ou j'lra de ro^te) a 



garnet of manganese atid magnesia, of the formuia 2 Ma Si + 



M" Si', by taking the residual 



Silica, whose oxygen = 1-41 corresj)jn(is to 2*80 

 Tiiuxide of manganese (l'70a 2';}S 



Magnesia ()-7U5 1*82 



7 00 

 " All these products subtracted, there only remains 



Trioxide of mnnganesn 2*38 



Ms'-ncsla. 3-47 



0-85 

 -Vcuj Striea, voi,. ix. 2 A 



