PLAN OF FRUSTRUM DIVIDED.

ELEVATION OF FRUSTRUM DIVIDED.]

If the material to be sampled is already a dry powder, 5 per cent. of it
should be heaped in a cone; each lot being added on the apex of the
cone already formed, so that it may distribute itself by falling evenly
in all directions. When the cone is completed, convert it into a low
frustrum of a cone by drawing stuff uniformly and in a direct line from
the centre to the circumference. Draw two diameters at right angles to
each other, and reserving any two alternate quarters, reject the others.
Mix; and form another cone, and proceed until a sample is got of the
bulk required.

This is the usual plan, and all samples should be treated in this way
when the stuff is fine enough to fall evenly down the sides of a cone.

Samples as they reach the assay office are seldom in a fit state for the
work of the assayer; they are generally too coarse, and ought always to
be more than he wants for any particular determination. The portion he
requires should never be taken at hap-hazard; the sample must be reduced
systematically to the quantity required.

1. _If the sample is a liquid:_ it is sufficient to shake the bottle,
and take out a measured or weighed quantity for the assay.

2. _If a liquid with a solid in suspension:_ measure the whole of it.
Filter. Make up the filtrate with the wash-water or water to the
original bulk. Assay it. Dry and weigh the residue, and make a separate