112 ROYAL SOCIETY OF CANADA 
acids, but we may, as in the case of the feldspar group, have molecular 
compounds in which molecules of different silicates replace one another 
isomorphously. Clark has shown that the variation in the com- 
position of many silicates may be explained by regarding the different 
forms as derived from normal salts, certain constituents of which may 
be replaced by other atoms or molecular groups. The micas, for in- 
stance, he regards as derived from the normal salts Al,(Si0,), and 
Al,(Si,O,),. Rammelsberg also explains the constitution of many com- 
plex silicates, which cannot be referred to any of the simple types enu- 
merated, by regarding them as compounds in varying proportions of 
orthosilicates and metasilicates, sometimes with disilicates, ete. 
It is, of course, possible to assume the existence of a number of 
acids which have not yet been isolated, but the present tendency is 
rather to reduce the number of types. Groth indeed regards ortho- 
trisilicic acid as composed of metasilicic and meta-disilicic acids, thus : 
EMSRO PME SO, Hei 
ae 
On 
He has tried, too, wherever possible, to refer silicates to the two simplest 
acids—orthosilicic and metasilicic. Take, for instance, the case of andalu- 
site, which was often described as a salt of hypothetical parasilicie acid 
(H,SiO, or SiO, 3H,0), in accordance with which the formula would 
be ALSiO. Groth regards it as a basic orthosilicate containing the 
univalent group AIO, and writes the formula (AIO) AlSiO,. Cyanite, 
again, which has precisely the same empirical formula as andalusite, he 
writes as a basic metasilicate (A10}, SiO,, on the ground that cyanite is 
less liable to decomposition than andalusite and that metasilicates gener- 
ally are more stable than orthosilicates. 
Some chemists look with suspicion upon the attempts which are 
being made to assign constitutional formule to minerals, regarding the 
formule as based upon hypotheses and having no foundation of fact. 
This must arise largely from lack of knowledge as to what has been ac- 
complished, and also from a certain unwillingness to apply the same prin- 
ciples to mineral chemistry as are applied to other departments of the 
science. Besides there is no science which has not been helped by work- 
ing hypotheses and theories or by argument from analogy ; and without 
its numerous hypotheses, organic chemistry could never have made the 
marvellous strides that it has during the past forty or fifty years. 
As an illustration of the thorough manner in which minerals are 
being investigated at the present day, let me here cite Professor 8. L, 
Penfield’s' recent study of the mineral topaz. This species had often 
been studied before, and had been analysed and re-analysed by many 

1 On the Chemical Composition and Related Physical Properties of Topaz, by S. 
L. Penfield and J. C. Minor, jr., Am. J. Sc., May, 1894, p. 387. 
