stokes.] DETAILS OF THE METHOD. 23 
of the flask; when this occurs, a ring of basic sulphate tends to form 
at the surface of the liquid. About 250 cm. 3 of the standard solution 
are used. The small loss by wetting of the condenser is compensated 
by having it wet at the outset. 
Blank experiments of this apparatus lasting two hours show no 
deposition of basic salt, no loss by volatilization (if the ingoing gas is 
saturated with moisture), and no reduction of the ferric salt. 
DETAILS OF THE OPERATION. 
As the solution contains enough air seriously to affect the results, 
and as the percentage of sulphur oxidized varies with the tempera- 
ture, it is necessary to introduce the substance into the boiling liquid, 
and not until all the air has been boiled out. For this purpose it is 
rapidly placed in the small bucket, shown on a large scale in fig. 2, A, 
which consists of a small test tube cut off and constricted at the 
bottom and provided with a loosely fitting platinum disk, made tight 
by pouring in a little iron-free asbestos emulsion, such as is used in 
Gooch crucibles. The bucket is suspended by a platinum wire, and 
is dropped into the liquid after this lias boiled ten minutes. It is 
necessary that the powder shall disseminate through the liquid and 
not settle out on the bottom for any considerable time. If this be not 
the case, abnormal results are obtained, for reasons which are not 
clear. The powder, as well as the apparatus, must therefore lie 
strictly free from grease, and, to aid the dissemination, the apparatus 
must be so mounted on a stand as to admit of agitating the liquid 
until the powder remains suspended. Pyrite disseminates readily, 
but marcasite shows a strong tendency to flocculate at first, and the 
two may generally be distinguished by this means. 
During the whole operation a very slow current of carbon dioxide 
is passed through. This must, of course, be absolutely free from 
oxygen, and as some air is present even in the most carefully pre- 
pared carbon dioxide, I always pass it over red-hot copper. 1 It must 
also be free from hydrogen sulphide, which is not always the case, 
and it can easily be freed from this by passing it through a tube filled 
with beads and containing saturated copper sulphate solution. 
The greater part of the ferric sulphate is reduced after a few min- 
utes' boiling, but in order to effect complete reduction the boiling 
should be continued two hours. The liquid is then cooled and filtered 
through dry filters into two accurate 100 cm. 3 flasks, which are filled 
with carbon dioxide and placed in cylinders of the same. In order 
entirely to avoid contact with the air, it is driven over through the 
inlet tube by means of carbon dioxide introduced through the con- 
denser. The flasks arc then brought to the temperature of th per- 
manganate to be used, and the contents are transferred to flasks 
1 When many determinations are to be made, it is convenient to place the copper in an iron 
gas pipe, water jacketed at both ends if necessary. 
