Hello Paul,

Posted under "Reviewed papers" on my website (see Welcome) is "Sampling in Mineral Processing" in which I give the variance of Gy's sampling constant, and explain why interleaving sampling protocols give reliable information at affordable cost (the cost of preparing and assaying one more sample).

For heterogeneous sampling units at mineral processing plants, Gy's sampling constant cannot possibly give an unbiased estimate for the distribution component of the variance of the primary sample selection stage but only for its composition component (Gy's fundamental error is, of course, a variance).

A single pair of A- and B-samples (ISO/TC183 jargon for a pair of interleaving samples) gives only one degree of freedom but for monthly production 27-30 degrees of freedom are obtained. The above paper also shows how to calculate confidence limits for metal contents and grades of mill feed, concentrate and tailing. Shortly, I shall post "Process Simulation with Spreadsheet Software".

In fairness to visitors to my Forum, I decided not to respond by email. At a later stage, however, I am prepared to transmit to you by email Excel spreadsheet templates to show how to calculate confidence limits for metal contents and grades.

Kind regards,

Jan W Merks ■