224 J. P. Cooke, Jr., on Cryophyliite. 
tainty than was possible with the minute scales from which the 
species was first established. 
Analysis of Rockport Lepidomelane.—My analyses of this mica 
are given in the following table. In the analyses 2, 3, and 4, 
the same methods were followed which have already been de- 
scribed in connection with cryophyllite. In analysis 5, I em- 
ne be with some slight modification, the method first proposed 
y H. Saint Claire Deville,* and the close agreement of the fig- 
ures obtained by such very different methods serves to confirm 
the general accuracy of the result. The determination of the 
water was the chief difficulty met with in the analyses. The 
water enters so intimately into the composition of the mineral 
that it is not expelled below a low red heat, and then only quite 
slowly. Moreover at this temperature the mineral is rapidly 
oxydized. But by igniting the mineral in a current of carbonic 
acid gas the oxydation was prevented and constant results were 
obtained. With this precaution portions of the same powder 
were ignited, first without and then with oxyd of lead with the 
following results; 
1 2. s 
Loss on ignition without PbO, 1-54 1-52 1-52 153 
“ “ with “ 
The difference is within the limit of probable error, and the 
results prove that the fluorine-is not expelled at a low red heat. 
Moreover the mineral after ignition gave as strong a reaction for 
fluorine as before, and when heated to the highest temperature 
which could be obtained with a gas blast lamp, underwent no 
farther loss of weight. This is in accordance with what we 
found to be true of cryophyllite; but not in accordance with 
the comportment of hydrous minerals into which fluorine enters 
as an essential constituent. Compare analyses of Cookeite, this 
Journal, [2], xli, 246. These facts, taken in connection with the 
circumstance that fluorine was not found by Soltmann in the lepr 
domelane of Wermland, lead to the inference that the fluorine 
in the Rockport mineral is an accidental constituent, deriv 
from the admixture of material foreign to the species itself. Th 
amount of fluorid of silicon in analysis 2 was determined by the 
method of Berzelius; but the limit of error in this process 18 8° 
wide that the result can only be regarded as approximate. The 
analyses which follow were all made on material dried at 100°, 
and selected to represent the different varieties of the mineral 
found at Rockport. Analyses 2 and 3 were made from portions” 
of a mass of comparatively compact material. No.4 from 
perfect crystal imbedded in the massive feldspar of the veln 
already referred to, and the silica determination. No. 1 from 
scales picked out from the granite of Boylston Hall in Cam- 
bridge, which is built of stone from the Rockport quarries. 
* Annales de Chimie et de Physique, [3], xxxvili, 5. 
