Full-stream mechanised sampling apparatus, that collect a full cross-sectional increment from the complete screen discharge, will always be the preferred option for sampling screen overflows. However, the reality is that provisions for automated sampling are rarely, if ever, provided at the discharges of individual screens within beneficiation plants.Sampling from the discharge of high capacity vibrating screens, especially multi-slope screens, is subject to sampling bias, due to the stratification of the ore that occurs on the vibrating screen deck. The stratification in the vibrating bed translates to a stratified falling stream at the discharge end of the screen, typically with smaller particles at the back of the falling stream and large particles at the front of the falling stream (Figure 1).The chemical and physical characteristics of heterogeneous ore (mineral or coal) particles vary significantly with particle size. If samples do not correctly reflect the source material size distribution, they will be biased and the resulting analytical/assay data will not reflect the actual physical nor chemical characteristics of the source material.Prior published data for coal preparation research (Australian Coal Association Research Program) has demonstrated that a bias appears to exist for samples collected from high capacity multi-slope screens.A new form of sampling scoop has been developed with the aim of overcoming this issue.Bias was evaluated at a coal preparation plant by comparing the desliming screen overflow size distribution to that for reference' coal preparation plant feed samples collected using the recognised stopped-belt technique. The study procured 20 sets of comparison samples in order to ensure statistical relevance.Two forms of sampling device were employed for the multislope screen discharge.A conventional square aperture scoop was demonstrated to exhibit a significant coarse bias, in that the resulting samples exhibited a significantly coarser size distribution compared to the reference samples. In other words it was evident that the conventional scoop favoured' material from the front of the falling stream.A new collapsible' sample scoop was evaluated in parallel, and its samples were found to exhibit no relevant bias.CITATION:Atkinson, B W and Swanson, A R, 2012. A new method for sampling from high capacity screens, in Proceedings Sampling 2012 , pp 89-94 (The Australasian Institute of Mining and Metallurgy: Melbourne).