14 ROYAL SOCIETY OF CANADA 
effect, as we have seen, is not produced. Further it was argued that if 
Groth’s formula rightly represented the constitution of tale, then the 
mineral on heating would simply lose water and no liberation of silica 
would take place :— | 
Si,0. — Me — OH _ SO, — Mg 
MES 0 Me On 7 MO OM ee 
whereas. if the mineral were a metasilicate, the probability was that not 
only would water be given off, but silica set free, thus : 
H Me, (Si0,), = gSiO, + SiO, + HO. 
On heating the mineral to a white heat for half an hour, and afterwards 
boiling the ignited product with soda solution, 15°36 per cent of silica 
actually went into solution—that is practically one-fourth of the silica in 
the mineral (4 = 15°57 per cent). Tale, on the other hand, which had 
not been previously ignited was found to be scarcely attacked by similar 
treatment with soda. Certainly Clarke seems to have been justified in 
concluding that tale is an acid metasilicate. 
Another interesting example of the investigation of mineral constitu- 
tion is to be found in Mr. Amos P. Brown’s recent study of pyrite and 
marcasite. Hitherto we have had no satisfactory explanation of the 
fact that marcasite undergoes decomposition so much more readily than 
pyrite ; but Mr. Brown explains this by showing that in marcasite all 
the iron is in the ferrous state, while in pyrite four-fifths of the iron exist 
in the ferric condition. The structural formula suggested’ for pyrite 
accordingly, 
2 § 
Fe< 
| 
S 
S — Fe 
ee 
| 
PS 
S — Fe 
S 
| 
VS 
Kec 
bose S 
while that of marcasite may be 
AE LIN 
Fe: | or some polymer of this. 
ASS 
1 A Comparative Study of the Chemical Behaviour of Pyrite and Marcasite, 
Proc. Am. Philosoph. Soc., xxxiii., no. 145. 
TA 


