382 Scheerer on Isomorphism. 
tion of the earths. I however found that the alumina thus 
precipitated was not free from magnesia. After separating it 
therefore by filtering and dissolving it in muriatic acid, it was 
treated again ina similar manner. But even now the alumina 
thrown down was not entirely free from magnesia, as was subse- 
quently shown when it was treated with caustic potash, by which 
operation there was left undissolved a substance consisting chiefly 
of oxyd of iron, from which minute quantities of alumina and 
magnesia were precipitated. The principal amount of the mag- 
nesia that was held in solution by the bicarbonate of soda, was 
separated with the necessary precautions by boiling and other 
methods, the slight quantity of the oxyd of iron present which 
had been taken up with the magnesia in the large excess of the 
alkaline solution, having been previously thrown down by a few 
drops of sulph-hydrate of ammonia. This sulphuret of iron was 
not altogether free from traces of magnesia. 
From two analyses of the Kragerée cordierite, in which the 
separation of the alumina and magnesia was performed as here 
described, the following results were obtained : 
1: Mean 
Silica, 50°44 50°44 50°44 
Alumina, . 3°22 32-68 32°95 
Magnesia, 12-43 13-08 12°76 
hitless Fy 1-08 17 1+12 
Protoxyd of iron, . 0-79 1-12 0:96 
Protoxyd of manganese, trace. trace. trace. 
Ware * 3: 1-17 0-87 1-02 
99-13 99°36 99°25 
The amount of silica could be determined in only one of these 
experiments, inasmuch as in one analysis the mineral was not 
fused with carbonate of soda but was attacked with fluoric acid.* 
The ratio of the oxygen of the silica to that of alumina and to 
that of the 1: 1 atomic bases is consequently as follows: 
Si Al R 
26°20 : 15°26 : 5-48, 
assuming the small amount of iron in the mineral to be in the 
form of the protoxyd. This, however, can scarcely be the case 
with the cordierite here the subject of analysis, which was almost 
entirely colorless, and did not show the slightest tinge of greed, 
whereas it is well known that, comparatively speaking, very 
ima deieadiain ill 
* Plattner’s apparatus was had recourse to for this fusion. The mass was 4 
completely fused, but was rendered thoroughly attackable by hydrochloric act” 
Frequent experience has convinced me of the great advantages whi I 
tus offers to the mineralogical chemist whose time and trouble are materially 
saved thereby. 
