mineral species independent of Phenomena of Isomorphism. 199 
to answer this question ; for the materials at our hands are suffi- 
cient to give us a satisfactory reply. 
There is a compound of antimony and silver called discrasite, 
which occurs in many localities crystallized in trimetric prisms 
homeeomorphous with Sb Zn*. ‘'he formula of the mineral is 
therefore probably Sb Ag*, which would require 71°5 per cent. 
of silver; but the per cent as given by analysis varies between 
7525 and 78 per cent, and one analysis gives the per cent as 
high as 85. Further analyses of this mineral are required in 
order to determine its constitution, but there can be no doubt 
that it varies in composition like Sb Zn?. | 
Silver-glance is another highly crystalline mineral. Theoreti- 
cally it should contain 87-1 per cent of silver and 12:9 per cent 
of sulphur; but in a specimen analysed by Klaproth, the pro- 
portions were 85 and 15. 
Again, the analyses of pyrrhotine (magnetic pyrites) give re- 
sults varying between 38°78 per cent sulphur, 60°52 per cent 
Iron (variety from Bodenmais), and 43°63 sulphur, 56°37 iron 
(variety from Baréges). The constitution of the mineral is still 
uncertain ; but its true formula is probably FeS, which would 
mquire 86°4 per cent sulphur and 63°6 per cent iron. Lastly, 
the analyses of antimony-glance give results varying between 
Antimony 74-06, Antimony 73°65, 
Sulphur” 25-94 994 > Sutphur ” 965 
The true formula of this mineral is undoubtedly Sb S?, which 
would require on] y 72°88 per cent of antimony. 
Ged ar examples might be greatly multiplied. Those just 
(Sra ote Selected at random from the first few pages of Dana's 
ystem of Mineralogy.’ They are all examples of binary com- 
Pounds which occur almost chemically pure in nature; so that 
!e phenomena in question are not complicated by those of 
somorphism, ; 
hen we pass to minerals of more complex constitution, the 
very wlnerals as mica, hornblende, garnet, and tourmaline differ 
es Ria 8teatly from each other,—a difference, moreover, which no 
pg of analysis will explain, and which must therefore be 
this a 0 an actual variation in composition. In the silicates 
wha: aration in composition is made evident by the variation of 
main, 18 termed the “oxygen ratios;” and it is well known to 
“ralogists that in many species this variation is very large, 
