24 
ON PYKTTE AND MARCASITE. 
[BULL. 186. 
containing sulphuric acid and filled with carbon dioxide and titrated. 
After reducing by hydrogen sulphide, filtering off the sulphur, and 
expelling the hydrogen sulphide by boiling in a carbon dioxide cur- 
rent, the total iron is titrated. A portion of the remaining solution 
should always be tested with sulphocyanate to ascertain the extent of 
the reduction. This generally gives a faint reaction for ferric iron. 
If it is decidedly red, either the reduction is incomplete or the mineral 
contains notable amounts of copper. 1 
As small errors in titration produce considerable deviation, an accu- 
rately calibrated burette should be used, so narrow that the bottom 
of the meniscus can be distinctly seen against a light. Those burettes 
which have the graduation carried halfway around the tube are par- 
ticularly good for accurate reading. The permanganate should have 
a strength of about 1.5 grams per liter, and should be run in only to 
the veiy faintest change of color, as compared with a ferric alum solu- 
tion of approximately the same strength and acidity. 
When less accurate results, are required the mineral, after extrac- 
tion with acid, may be simply collected on a hardened filter, washed 
with acid and water, and dried in the vacuum; special purification 
of the carbon dioxide, filtering in a carbon dioxide atmosphere, filtra- 
tion of the reduced solution from sulphur, and other exceptional pre- 
cautions maybe omitted. Results made in duplicate in this way may 
differ 3 or 4 per cent. 
ACCURACY OF THE METHOD. 
If 100 cm. 3 solution and a permanganate of 1.5 grams per liter be 
used in titrating, an excess of 0.1 cm. 3 permanganate, or an oxidation 
of 0.0003 gram of the material, gives the following errors in the value 
of p: 
a. 
/>. 
c. 
Oxidation 
of 0.0003 gram 
sulphide. 
Pyrite 
Marcasite 
+ 1.00 
+0.20 
+0. 12 
+o.o.-) 
-1.00 
-0.20 
-1.00 
-0.15 
Special care is therefore necessary in standardizing the original 
solution and in the final titration. Practically, duplicate determina- 
tions with pyrite tend to agree more closely than those with marca- 
site, as seen in the table below. This is perhaps due to the greater 
tendency of marcasite to flocculate, as indicated above (p. 23), and to 
its greater tendency to oxidize, although the influence of the latter is 
less than in the case of pyrite. The influence of various impurities 
is considered below (p. 31). 
For the explanation of this see p. 45. 
