308 
She. Saye haeveet Se et ee been eae ee age eee 
Alvin; ©) .%52 0 0: ss te Os ene ee 
Protoxide ofiron,. . . . - .-.- + 883 
Maenesia, ee ee oe es a ee 
Mae tc ee ea ss eS ee 
Oxide of chrome,.° . 9. . . . . . . a trace 
100.46 
This analysis was twice repeated by Mr. Alex. MacDon- 
nell, who obtained numbers almost identical with those above 
given, and which it is, therefore, unnecessary to adduce. 
These numbers, it may be easily shown, correspond very ex- 
‘actly with the empirical formula,— 
Si O,+3 Al, O,;+29RO+21 HO; 
and the atoms may be so arranged as to give the rational for- 
mula,— 
3 (Al, O,, Si O;) + 8 (3 RO, SiO., 2HO) +5 (RO, HO). 
Pennine, it was subsequently found, had been previously 
twice analyzed, first by Schweitzer, whose results do not dif- 
fer very widely from those just given, and afterwards by Ma- 
rignac and Descloiseaux, who are generally considered to have 
accurately fixed its constitution. They give the following as 
its empirical formula, viz. :— 
5 SiO, + 2 Al, O;+12RO+10 HO, 
which, as is obvious, is utterly irreconcileable with the analy- 
sis just brought under the notice of the Academy. As respects 
the cause of these discrepancies, the conjecture may be ha- 
zarded, that they are due to the presence of intermixed por- 
tions of other minerals. The specimens, for example, of 
Pennine, brought to Dublin by Professor Jellett, are (some of 
them) intersected by threads of tale, and incrusted with mi- 
nute grossular garnets in perfect dodecahedral crystals. 
The crystalline system of Pennine is usually set down as 
the third, the most common crystal being an acute rhombo- 
hedron, whose apices are deeply truncated. The specimen, 
