30 ON PYRITE AND MARCASITE. [bull. 186. 
DISCUSSION OF THE CURVE. 
The curve in PL I is based on the mean values of p thus obtained, 
with the exception of pure marcasite, where the more probable value 
18 is used (p. 27). 
Influence of the minimum: indirect determination of small amounts 
of pyrite in marcasite. — The most obvious features of this curve 
are the presence of a minimum point when the amount of pyrite 
reaches about 10 per cent and the increasing steepness as 100 per 
cent is approached. The minimum implies that a mixture of marca- 
site with a little pyrite has a less percentage of sulphur oxidized 
than has pure marcasite. This unexpected result, which is abun- 
dantly confirmed by numerous other preliminary data not published, 
may be partly due to the fact that pyrite uses up more ferric salt than 
does the equivalent amount of marcasite; the latter is therefore acted 
on by a solution weaker in Fe'" than if it alone were present, the 
effect of which is to lower the oxidation coefficient. Possibly there 
may also be an electrochemical action between the two. A minimum 
also occurs in the curve for pyrite-galena, while pyrite-chaleopyrite 
shows a maximum. The existence of these points shows that it is 
impossible to predict, from a knowledge of the values of p for any 
two sulphides, whether the result of the contamination of one with a 
small quantity of the other will be to raise or to lower the oxidation 
coefficient. The variation of p from 16.3 to 18 for nearly pure marca- 
sites is presumably due to small amounts of impurities other than 
pyrite. Practically these have but little effect on the determination 
of pyrite and marcasite, for, as seen in the table, the three samples 
of marcasite, though varying from 1G.5 to 18, give essentially the same 
figures when mixed with even as little as 5 per cent of pyrite. 
As a result of the minimum, any specimen in which p is about 18 
or less may have one of two possible compositions. Which of these 
is actually present may be tested by adding one-ninth its weight of 
pyrite and determining the value of p for the new mixture. If 
depression is produced, the true percentage of pyrite lies to the left 
of the minimum; if elevation, to the right. In fact, the freedom of a 
marcasite from pyrite can be regarded as absolutely proved only 
when this precaution is taken. The determination of p in the origi- 
nal specimen is not necessary; the marcasite in question may at once 
be mixed with a known amount of pyrite and the original contents of 
pyrite thus determined. If x represent the percentage of pyrite in 
the original sample, then a mixture of 00 parts of this with 10 parts 
pyrite will give 
(X9 x +- 10— per cent pyrite in mixture. 
The latter is found by one determination and the composition of the 
original sample deduced. An application of this method may be seen 
below (pp. 36-38) in the case of the marcasites from Garfield tunnel 
