6CHAU.ER.] 
DUMORTIERITE. 
117 
the same locality, it seems much preferable to regard the determi- 
nation of the water in Ford's analysis as too high rather than to 
attempt an explanation of it by the assumption of " isomorphous 
replacement, etc." 
Of the third analysis by Ford of the New York mineral, he says: "It 
is fully realized that the results are not to be looked upon as being as 
exact as those of the other analyses." 
His analysis is: 
Si() 2 
AU> : 
Fe 2 : 
BA 
H 2 
Katios. 
6.23 or 6X1. 04 
7.28 8X .90 
1.00 1X1.06 
1.40 1X1.40 
Assuming that his water content is somewhat high, the ratios agree 
well with the new formula. 
This concludes the list of analyses of dumortierite which were made 
on pure material and with due knowledge of what was to be determined. 
A number of other analyses will now be given, which serve in a gen- 
eral way for the determination of the composition of the mineral, but 
can not be relied on for exact results. 
One, which has heretofore not been published, was made by the 
writer on the Washington dumortierite. About 200 grams of the rock 
was powdered, and by repeated separation with heavy solutions about 
4 grams of dumortierite were obtained. The sample was by no means 
pure. It was found impossible to remove all the andalusite, an unknown 
but small amount remaining. It was found during the course of the 
analysis that the samples also contained some titanite (leucoxene) and 
a very small amount of pyrite. The analysis was made with all the 
care possible, and the following results were obtained: 
SiQ 2 . 
A1 2 3 
FeA 
Ti( ), . 
H 2 . 
BA- 
Ca() . 
100.03 
Ratios after deduct- 
iTig titanite. 
28. 51 
5. 97 « 
r 0X0.99 
59. 75 
2. 48 
} 
7. so 
8X .97 
. 95 
2. 12 
1.53 
1X1.53 
5. 54 
L.03 
1X1.03 
. 68 
