uld be free from lime and chlorine, and contain only traces of
arsenic, sulphuric acid, &c.

5. Should not leave more than 0.25 per cent. of inorganic and organic
residue together when evaporated in a platinum dish without ebullition
(about 160 deg. C.) or partial decomposition.

6. Silver test fair.

7. The glycerine, when diluted one-half, should give no deposit or
separation of fatty acids when nitric peroxide gas is passed through it.
(Nos. 1, 2, 3, and 5 are the most essential.)

The white flocculent matter sometimes formed is a very great nuisance, and
any sample of glycerol which gives such a precipitate when tried in the
laboratory should at once be rejected, as it will give no end of trouble
in the separating house, and also in the filter house, and it will be very
difficult indeed to make the nitro-glycerine pass the heat test. The out-
turn of nitro-glycerine also will be very low. The trouble will show
itself chiefly in the separating operation. Very often 2 or 3 inches will
rise to the surface or hang about in the nitro-glycerine, and at the point
of contact between it and the mixed acids, and will afterwards be very
difficult to get rid of by filtration. The material appears to be partly
an emulsion of the glycerine, and partly due to fatty acids, and as there
appears to be no really satisfactory method of preventing its formation,
or of getting rid of it, the better plan is not to use any glycerine for
nitrating that has been found by experiment upon the laboratory scale to
give this objectionable matter. One of the most useful methods of testing
the glycerine, other than nitrating, is to dilute the sample one-half with
water, and then to pass a current of nitric peroxide gas through it, when
a flocculent precipitate of elaidic acid (less soluble in glycerine than
the original oleic acid) will be formed. Nitrogen peroxide, N_{2}O_{4}, is
best obtained by heating dry lead nitrate (see Allen, "Commercial Organic
Analysis," vol. ii., 301).

When a sample of nitro-glycerine is brought to the laboratory from the
filter house, it should first be examined to see that it is not acid.[A] A
weak solution of Congo red or methyl orange may be used. If it appears to
be decidedly alkaline, it should be poured into a separating funnel, and
shaken with a little distilled water. This should be repeated, and the
washings (about 400 c.c.) run into a beaker, a drop of Congo red or methyl
orange added, and a drop or so of