Genth.] 120 (Aug. 17, 
pure scales (containing only 0.85 % of quartz, gold, &c.) to make one ana- 
lysis, which, as it was made with the greatest care, must be a very close 
approximation of the truth. The material of the other analyses was far . 
more contaminated, and the results were given merely for comparison and 
to show the influence of the admixtures upon the analyses. 
From Prof. Roscoe’s analyses it does not appear that he attempted to 
separate the impurities by chemical means, and thus he gives the compo- 
sition of the whole mixture. 
He assumes the vanadium to be present as pentoxide, the iron as ferric 
oxide, the manganese as manganic oxide, the two latter as replacing alu- 
mina ; and magnesia, lime and soda as replacing potassium oxide. 
As I have made a direct determination of the state of oxydation of the 
vanadium, I can say positively that, if any. only the smaller portion of the 
vanadium is pentoxide. I found the composition of the vanadium oxide 
to be V,O,, = 2 V,0,, VO; ; but as it was obtained after allowing for the 
oxydation of ferrous into ferric oxide, and as the quantities of ferrous ox- 
ide have been found to vary from 1.67 to 3.30 %, it is not impossible that 
an insufficient quantity of oxygen has been deducted, and that the whole 
of the vanadium is present as V,Os. 
Pure roscoelite contains no manganese ; in Prof. Roscoe’s analyses 0.85 
— 1.45 % of manganic oxide have been found, which confirms my opinion 
that his material was not pure ; but what is most astonishing to me is the 
very low per centage of silica which he finds. 
From his analyses he calculates a formula, and from this the per centage 
composition, which, however, is far from corresponding with his analyses, 
as for instance : 
Silica found = 41.25, calculated = 41.18 
Potassium oxide found = 8.56, ef = 14.24 
I had not calculated any formula from my analyses, being in hope that 
I may yet be able to procure this interesting mineral in a still purer state 
for further investigation. For comparison I will insert my analysis (@) 
which certainly closely represents the true composition of Roscoelite, and 
will add the formula corresponding with the same, with this alteration 
however, that I consider all the vanadium as V,O;. It contains, after 
deducting 0.85 % of quartz, gold, &e. : 
ji 
/ 
Found. Calculated. 
SiO, = 47.69 = 49.33 
Al,O, = 14.10 a 14.09 
V.0; = 20.56 al 20.62 
FeO = 1.67 — 1.64 
MgO — 2.00 — 1.83 
Li,O —— | races —_ nase 
Na,O = 0.19 = 
K,O = 7.59 = 7.55 
Tgnition = 
4,96 == 4.94 
98.76 100.00 
