Technical Abstract: The goal of germplasm collections is to obtain representative samples of genetic variation and maintain them as living specimens that can be used in a variety of research applications. A key challenge in this objective is maintaining both genetic diversity and viability. The process of ex situ conservation involves a series of sampling events. Whether sampling directly from natural variation (in situ) or sampling during a seed increase, genetic drift can result in changes in diversity. We developed a computer simulation model to examine how the sampling process influences long term maintenance of genetic variation. In particular, we examine how two common regeneration strategies (parallel regeneration using the most original seed source or serial regeneration using the most recent regeneration) influence the retention of neutral allelic variation over time. In addition we contrast highly heterogeneous accessions (allele frequencies of 0.72, 0.2, 0.04, 0.02, 0.01, 0.005, and 0.005) and highly uniform accessions (allele frequencies of 0.95 and 0.05) that are either highly selfing or highly outcrossing. Lastly we examine how changes in seed longevity influence drift indirectly through changes in viability. Low frequency alleles were quickly removed from wild populations when 50 individuals were selected to regenerate the accession. Allele losses are minimized by performing regenerations in parallel, by improving seed longevity, by performing regeneration events before reductions in seed viability occur, and by increasing the number of individuals in regeneration events.