[ HARRINGTON ] MINERALOGICAL CHEMISTRY 13 
chemists. The analyses showed cleariy that silicon and aluminium were 
present in the proportion of 1:2, but the percentage of fluorine ranged 
from 16:12 to 18°83. The formula that had generally been accepted for 
the mineral was that suggested by Groth [AI(O.F,)]AISiO,. Penfield 
had, however, previously shown that hydroxyl frequently replaces fluo- 
rine in minerals, andit occurred to him that in the case of topaz the 
varying proportions of fluorine might be due to the presence of varying 
proportions of hydroxyl. Specimens of topaz from a numberof localities 
were examined, and it was found that in every case water was given off 
on ignition, and further that the proportion varied inversely with the 
amount of fluorine. The legitimate conclusion drawn by Penfield was 
that the water was derived from hydroxyl and that in topaz there is an 
isomorphous replacement of fluorine by hydroxyl. In accordance with 
this view the simplest formula that can be assigned to the mineral is 
[AI(F.OH) ]SiO,. Penfield’s admirable paper also brings out in a striking 
manner the close connection subsisting between the chemical composition 
and physical properties of topaz. There is a slight diminution in specific 
gravity corresponding to the increase of hydroxyl, and a corresponding 
increase in the optic axial angle. The indices of refraction also increase 
and the strength of the double refraction decreases in proportion as 
hydroxy! replaces fluorine. 
Look, in the next place, at the case of tale, whose constitution has 
been carefully studied by Clarke and Schneider! The percentage com- 
position of a typical specimen was carefully determined, and it was found 
that scarcely any of the water was driven off at temperatures short of 
redness. On heating the mineral to from 383°—412° in an atmosphere of 
dry hydrochloric acid gas for fifteen hours, scarcely any effect was pro- 
duced. Long continued treatment with aqueous hydrochloric acid produced 
little effect, also pointing to the stability of the compound. The empirical 
formula deducible from the analyses is the ordinary one Mg,H,Si,0,.. 
The simplest interpretation as to constitution is that the compound is an 
acid metasilicate Mg,H.(SiO,),. Groth, however, regarded the mineral 
as a basic salt of the disilicic acid H,Si0,—and gave it the formula 
Mg(MeOH),(5:0,)+ The question therefore which Clarke wished to 
settle was, which formula was most nearly in accord with the facts— 
should tale be regarded as an acid metasilicate or a basic meta-disilicate ? 

The study of a number of other silicates (as already noticed) seemed 
to indicate that the magnesia which is eliminated by the action of dry 
hydrochloric acid gas is present in the mineral in combination with 
hydroxyl (—Mg—OH.) If, therefore, tale were a basic silicate the hydro- 
chlorie acid should have a marked effect in eliminating magnesia ; which 
! Experimental Untersuchungen über de Constitution der natürlichen Silicate, 
von F. W. Clarke and E. H. Schneider in Washington. Zeitschrift ftir Krystallo- 
graphie, etc., xviii., 4, 1890. 

