The Snap Lake Phase I project initiated a subsampling campaign on the material (kimberlite/metavolcanic/granite) mined and treated for diamonds. The campaign was aimed at determining the optimum subsampling program to subsample the feed to the bulk sample plant (BSP) for the duration of a sampling campaign._x000D_
This paper describes the application of Gy's formula for determining the subsample size and other good sampling practices to determine this subsampling campaign._x000D_
Three main issues are discussed: the use of a hammer sampler to sample off a moving conveyor and the practical experiment to determine the optimum subsampling program; bias introduced by laboratory equipment including two sizes rotary splitters; and variability of the feed material to the BSP over time.