38 
THE CONSTITUTION OF THE SILICATES. 
[BULL. 125. 
with tlie foregoing conjecture. Taking all the data into account, chaba- 
zite may be formulated as a mixture of the compound 
,Si0 4 =Al 
Al-Si0 4 =H 2 .AlH 2 2 
\siO 4 =0a 
I 
Ca +14H 2 
I 
,Si 3 8 =Ca 
Al-Si 3 8 =H 2 .AlH 2 2 
\si 3 8 3EAl 
with the salts 
,Si0 4 =]s T a 3 
A1-SK) 4 =H 2 .A1H 2 2 . + 7H 2 
\si0 4 =Al 
ySi 3 8 =Na 3 
Al-Si 3 8 =H 2 .AlH 2 Q 2 . + 7H 2 
\S1,0 8 =A1 
the latter being isomeric with the gmelinite molecules. A sodium 
chabazite agrees in empirical composition with gmelinite, and the cal- 
cium salt by itself has the following composition, as compared with 
chabazite from Aussig, as analyzed by Lemberg:* 
Found. ' Calculated. 
47.50 
47.43 
20.00 
20.15 
10.20 
11.07 
.23 
1.24 
21.40 
21.35 
100. 57 
100. 00 
It is perfectly evident that the calcium chabazite may also be written 
as an equimolecular mixture derived from anorthite and its trisilicate 
equivalent ; and in levynite, the last member of the rhombohedral group, 
we have a mineral of the same type, although differing in hydration. 
Instead of 18 molecules of water, as in chabazite, levynite contains 
only 15 ; but in both cases 4 molecules are to be regarded as consti- 
tutional, in accordance with the observations made by Dainour. 
Computing from Hillebrand's analyses,! levynite contains the ortho- 
and trisilicate molecules in the ratio of 3:2, giving the following empir- 
ical composition to the species : 
3Al 6 (Si0 4 ) 6 Ca 3 . 15H 2 () + 2Al 6 (Si 3 Q 8 ) 6 Ca 3 . 15H 2 0, 
* Analysis No. 4 in Dana's table, Syst. Min., p. 591. 
t Bull. XL S. Geol. Surv. No. 20, p. 38. 
