66 THE CONSTITUTION OF THE SILICATES. [bull. 125. 
Woerthite is an altered sillinianite, and westanite is perhaps a similar 
derivative of andalusite. The newtonite compound has already appeared 
in tbe mica series among the conrponents of cookeite and rumpfite. 
The best that can be said for these .formulae is that they are sug- 
gestive. In one respect they are highly questionable, for the reason 
that the group — Si0 4 =H 3 is indicative of loosely combined water, 
while in these particular clays the water is quite firmly retained. On 
this point much fuller information is needed, and future evidence may 
prove that the serial relation indicated is apparent only. 
The composition of rectorite may be Al :! (Si0 4 ) 3 H 3 +2H 2 0, or that of a 
hydrous kryptotile. Halloysite has the composition of kaolin plus one 
molecule of water, the latter being removable at or about 100°. Halloy- 
site, however, differs from kaolin in being decomposable by hydrochloric 
acid, and hence it is unlikely that the two species have similar struct- 
ure. Allophane is perhaps ( A10 2 H 2 ) 2 H 2 Si0 4 +3H 2 0, or it may be written 
analogously to andalusite, 
. Al 3 (Si0 4 ) 3 (A10 2 H 2 ) 3 .12H 2 0. 
Neither formula is sustained by any good evidence. Other clays are 
possibly as follows : 
Samoite Al 4 (Si0 4 ) 3 . ,0H 2 O 
Cimolite Al 4 (Si 3 8 ) 3 .6H 2 
Oollyrite (A10) 4 Si0 4 .6H 2 
Schrotterite (A10 2 H 2 ) 4 Si0 4 .6H 2 
None of these formulae can be construed as anything more than a 
temporary suggestion, which may help research. The hydrous ferric 
silicates are, if anything, less satisfactory than the aluminum salts. 
Anthosiderite is representable by the formula Fe 4 (Si 3 8 ) 3 .2H 2 0; and 
chloropal by the expression Fe 2 (Si0 4 ) 3 H 6 . Hisingerite seems to range 
from* a ferric kaolin to a ferric halloysite, and farther than this it is not 
worth while to go. The remaining iron clays which have received 
specific names are altogether doubtful. The chromium clays, wolchon- 
skoite, etc., are also of uncertain character. 
To cerite, which is of doubtful composition, the provisional formula 
^SiO.^OeO.H, 
Ce— Si0 4 =OeO.H 2 
\siO 4 =0eO.H 2 
may be assigned. Other earth metals — lanthanum, the two didymi- 
ums, etc. — replace a considerable part of the cerium. With cerium 
only the formula requires 
SiO, 20. 22 
Ce,0 3 73.71 
H 2 6. 07 
100. 00 
Cenosite, H 4 Oa 2 Y 2 OSi 4 Oi 7 , and ardennite, H 5 Mn 4 Al 4 VSi 4 23 , can each 
be written structurally in more than one way, with no determining 
evidence upon which to base a decision. They may be orthosilicates or 
