1G CONTRIBUTIONS TO CHEMISTRY AND MINERALOGY. [bull. 167. 
repeated treatments with ammonium chloride a complete deeomposi, 
tion might be effected, but this question is one upon which it seemed 
unnecessary to spend further time. 
Pyrophyllite. 
The empirical formula for pyrophyllite, AlHSi 2 6 is apparently that 
of an acid metasilicate, and the mineral is therefore peculiarly available 
for fractional analysis. The compact variety from Deep Error, North 
Carolina, was taken for examination, and a uniform sample was pre- 
pared. Analysis gave the following results : 
If, now, pyrophyllite is an acid metasilicate it should break up on 
ignition in accordance with the equation 
2AlHSi 2 0,. = AliSi.,0, + Si0 3 + HO. 
That is, one-fourth of the silica, or 10.18 per °^!^ * ^H 
The mineral itself is very slightly attacked by bmlmg with the sod, n 
carbonate solution, and in an experiment of this kind only 0.72 pe. 
cent of silica was dissolved. Upon ignition under varying circnm 
stances the following data were obtained: 
Ignited ten minutes over a Bunsen burner, and then extracted with sodium ear 
bonate solution, 1.51 per cent of SiO, dissolved. t , „ „„i„w„ 
Ignited fifteen minutes over a Bunsen burner, 1.89 per cent became soluble 
Ignited ten minutes over a Bunsen burner and then fifteen minutes over the blast 
2 84 per cent of silica was liberated. 
These results are of a different order from those given by pectolito 
and talc, and raise the question whether pyrophyllite, despite its ratios 
is a metasilicate at all. So far as the evidence goes, it may with pro 
priety be regarded as a basic salt of the acid H 2 Si 2 5 , and its format 
then becomes 
Si,0 5 = Al — OH. 
This formula is at least as probable as the metasilicate expressior 
which latter rests upon assumption alone. Still other formula-, but 
greater complexity, are possible; but until we know more of the gene, 
and chemical relationships of pyrophyllite, speculate concerning the. 
would be unprofitable. 
