26 THE CONSTITUTION OF THE SILICATES. [hull. 125. 
of genesis; that vesuvianite alters into garnet, and also, like garnet, 
into micas and chlorites; and that both species, as shown by Doelter,* 
yield essentially the same products upon fusion. The analogies con 
necting them are remarkably suggestive and complete, and the formulas 
here proposed render those analogies intelligible. In the Swedish 
"mangau-idocrase" a salt occurs which is doubtless the vesuvianite 
equivalent of spessartite, but the compound in a pure state is unknown. 
Two additional species, meliliteand gehlenite, which, like vesuvianite, 
are also tetragonal, may perhaps be best considered at this stage of 
the discussion. For melilite, the analyses are somewhat discordant; 
but on the whole they agree best with Groth's formula (CaMgN"a 2 ) 6 
{AlFe) 2 Si 5 19 . This may be written so as to resemble the formula of 
vesuvianite, but with the linking Si0 4 group replaced by the bivalent 
Si0 3 . For gehlenite, the empirical formula Ca 3 Al 2 Si 2 O 10 is commonly 
assumed, but a careful comparison of the published analyses has con- 
vinced me that it is not entirely satisfactory. Most of the analyses 
show appreciable amounts of water, but this may be due to alteration ; 
for the artificial gehlenite, prepared by the igneous method, is non- 
hydrated. A different formula, which relates gehlenite to melilite 
and vesuvianite, seems to be preferable; and such a formula is empiric- 
ally Ca 7 MgAl4Si 5 2 4. In this the univalent group — Al\ /E" is to 
be assumed, which occurs also among the clintonite micas, and which, 
in natural gehlenite, may be replaced in part by — Al = (OH) 2 . On 
this basis the three species are comparable as follows : 
(lehtenite. 
/Si0 4 = ) ■ 
R ;/ 3 / > R 7/ 3 
Al-Si0 4 ^ ) 
\Sib 4 ==(A10 8 B")2 
Al-Si0 4 = ) 
R ;/ 3 \ > R"-. 
These symbols are not absolutely unobjectionable, but they are useful 
for purposes of correlation. Gehlenite alters into garnet, melilite some- 
times occurs among the products of fusion of garnet, and the formulae 
suggest such relations. Furthermore, Lembergt has shown that geh- 
lenite, heated to 200° with a solution of potassium carbonate, gives cal- 
cium carbonate and a product having the composition of a potash mica; 
while similar treatment with sodium carbonate converts the mineral 
into cancrinite. Gehlenite, garnet, cancrinite, and muscovite are there- 
fore related to one another, and this fact is expressed by the formulas 
proposed. In melilite the calcium is partly replaced by sodium, and 
Vesuvianite. 
Melilite. 
/Si0 4 = ) 
/ R" 3 
Al-Si0 4 ~ ) 
,Si0 4 = 
Al-Si0 4 = 
\si0 4 =Al-OH 
Al-Si0 4 — ) 
\Si0 4 = > 
)>Si0 3 
Al-Si0 4 = 
^Si0 4 = 
*Allgem. Cheui. Hineralogie, p. 183. 1 Zeitsch. Deutsch. Geolog. Gesell., 1892, p. 231 
