Page 231 - Introduction to Mineral Exploration

214 M.K.G. WHATELEY & B. SCOTT

FIG. 10.8 A Jones sample splitter.
For explanation see text.

using Gy’s formula. This is then compared crushing–grinding equipment does not have to
with the previously estimated total error, on be enlarged, as in the first option.
the following basis: The total sampling error of any sample
1 When the observed total error is smaller than reduction system should be checked at re-
or of the same order of magnitude as the esti- gular intervals, say every 6 months depending
mated total error nothing should be changed upon throughput and use, to ensure that it is
in the reduction system. This provides a safety working satisfactorily.
margin against the future wear of the equip-
ment which will deliver progressively coarser
material for which the sampling variance 10.1.5 Preparation of samples containing
increases. native gold and platinum
2 When the observed total error is larger With native precious metals a problem arises
than the estimated total error the former has from the significant difference between the
to be reduced and particularly at the first (or density of an associated gangue (about 2.6) and
coarsest) reduction stage where the variance that of the metals (110.0 to 21.5). Gold may
2
is likely to be the major component of S (FE). be diluted with silver and copper to a density
This can be achieved by retaining the same in the order of 15.0 but even at this level its
number of sampling stages but increasing the density is twice that of galena, the densest of all
weight (i.e. the mass) treated at each separate common sulfide minerals, and six times that
stage and increasing the number of stages but of quartz. This contrast means that when
with the same sample weight at each stage. such valuable constituents are liberated from
Usually the latter choice is preferable as the their gangue, flakes and nuggets will segregate