Industrial uses of lignocellulosic biomass are sensitive to changes in feedstock compositions. Therefore, obtaining representative feedstock samples is important for optimizing overall process efficiencies. Some challenges faced in sampling lignocellulosic biomass along feedstock supply chains are a lack of flowability, heterogeneous feedstocks, potentially rapid biomass deterioration, large particle size distributions, time constraints, and scaling issues.  Further, bulk lignocellulosic biomass quickly sorts, with larger particles migrating to the top and the fines to the bottom. These different particle sizes exhibit unique compositions as well as reactive properties.   We have measured significant differences in composition and processing responses due to particle sizes and anatomical fractions of the biomass. Typically used methods of sampling such as riffle splitting, grab-sampling, or coning and quartering were difficult to apply in several circumstances, such as obtaining biomass subsamples from a combine. We suggest using an application of Pierre Gy’s two-dimensional incremental delimitation method to provide an accurate and cost effective sampling method.