ndicated by mine samples, and the actual value as shown
by yield plus the residues. At Broken Hill, on three lead mines,
the yield is about 12% less than sampling would indicate. This
constancy of error in one direction has not been so generally
acknowledged as would be desirable, and it must be allowed for
in calculating final results. The causes of the exaggeration seem
to be:--

_First_, inability to stope a mine to such fine limitations of
width, or exclusion of unpayable patches, as would appear practicable
when sampling, that is by the inclusion when mining of a certain
amount of barren rock. Even in deposits of about normal stoping
width, it is impossible to prevent the breaking of a certain amount
of waste, even if the ore occurrence is regularly confined by walls.

If the mine be of the impregnation type, such as those at Goldfield,
or Kalgoorlie, with values like plums in a pudding, and the stopes
themselves directed more by assays than by any physical differences
in the ore, the discrepancy becomes very much increased. In mines
where the range of values is narrower than the normal stoping width,
some wall rock must be broken. Although it is customary to allow for
this in calculating the average value from samples, the allowance
seldom seems enough. In mines where the ore is broken on to the
top of stopes filled with waste, there is some loss underground
through mixture with the filling.

_Second_, the metal content of ores, especially when in the form of
sulphides, is usually more friable than the matrix, and in actual
breaking of samples an undue proportion of friable material usually
creeps in. This is true more in lead, copper, and zinc, than in
gold ores. On several gold mines, however, tests on accumulated
samples for their sulphide percentage showed a distinctly greater
ratio than the tenor of the ore itself in the mill. As the gold is
usually associated with the sulphides, the samples showed higher
values than the mill.

In general, some considerable factor of safety must be allowed
after arriving at calculated average of samples,--how much it is
difficult to say, but, in any event, not less than 10%.

CHAPTER II.

Mine Valuation (_Continued_).

CALCULATION OF QUANTITIES OF ORE, AND CLASSIFICATION OF ORE IN SIGHT.

As mines are opened by levels, rises, etc., through the ore, an
extension of these workings has the effect of dividing it into
"blocks." The obvious procedure in determining