Abstract
Several methods of calculating ore grades by diamond-drill sampling in the Homestake mine, Lead, S. Dak., Have been evaluated by statistical analyses of assay data. The natural distribution of gold particles, partly masked through sampling and assaying procedures, is probably lognormal. As a consequence of the gold particles being clustered rather than randomly distributed within the ore bodies, (1) more mine samples are needed to obtain a specified precision of mean grade than the number predicted by the standard-error-of-the-mean law; (2) the variance of gold values is not inversely proportional to sample volume, and, therefore, small- diameter drill core yields samples nearly as good as larger diameter core; and (3) sampling at 5-foot intervals provides nearly as good results as sampling at shorter intervals.
