Conservation is increasingly central to the botanic garden mission. Living plant collections are important components of conservation. Critical evaluation of living conservation collections with population genetic analysis can directly inform ex situ conservation strategy. Here, we quantify the degree of genetic variation captured through a population-based collection protocol, and explore optimal sampling for ex situ conservation. An extensive living collection derived from one population of Leucothrinax morrisii (Arecaceae) provided a model system. We compared 58 specimens from the ex situ collection with 100 individuals from throughout the parent population via 6 ISSR loci. Random bootstrapped resamples of the data were made to model differently structured ex situ collections. Mean diversity (He) differed little between the collection (0.204) and the population (0.216), and genetic distance (D) was very close (0.036). Very few private alleles were found between the collection and the population. Allelic capture, as measured by percent of private alleles, was greater than 94%. Resampled collections of different sizes captured from 48% to 94% of alleles. Pairwise comparison of bootstrapped resamples suggests that increasing the representation of half-sibling groups does not significantly increase allele capture. Increase in allele capture with increasing sample size is greatest at low resample sizes, and showed diminishing returns as resample size increased. No appreciable increase in allele capture was gained through maintaining different half-sibling groups. These data inform sampling for ex situ conservation purposes, and recommend sample sizes of at least 15 individuals, with the upper limit based on resources.