The Mauna Kea silversword, Argyroxiphium sandwicense ssp. sandwicense, has experienced both a severe population crash associated with an increase in alien ungulate populations on Mauna Kea, and a population bottleneck associated with reintroduction. In this paper, we address the genetic consequences of both demographic events using eight microsatellite loci. The population crash was not accompanied by a significant reduction in number of alleles or heterozygosity. However, the population bottleneck was accompanied by significant reductions in observed number of alleles, effective number of alleles, and expected heterozygosity, though not in observed heterozygosity. The effective size of the population bottleneck was calculated using both observed heterozygosities and allele frequency variances. Both methods corroborated the historical census size of the population bottleneck of at most three individuals. The results suggest that: (i) small populations, even those that result from severe reductions in historical population size and extent, are not necessarily genetically depauperate; and (ii) species reintroduction plans need to be conceived and implemented carefully, with due consideration to the genetic impact of sampling for reintroduction.

Reintroduction of populations of endangered species is a challenging task, involving a number of environmental, demographic and genetic factors. Genetic parameters of interest include historical patterns of genetic structure and gene flow. Care must be taken during reintroduction to balance the contrasting risks of inbreeding and outbreeding depression. The Mauna Loa silversword, Argyroxiphium kauense, has experienced a severe decline in population size and distribution in the recent past. Currently, three populations with a total of fewer than 1000 individuals remain. We measured genetic variation within and among the remnant populations using seven microsatellite loci. We found significant genetic variation remaining within all populations, probably related to the recent nature of the population impact, the longevity of the plants, and their apparent self-incompatibility. We also found significant genetic differentiation among the populations, reinforcing previous observations of ecological and morphological differentiation. With respect to reintroduction, the results suggest that, in the absence of additional data to the contrary, inbreeding depression may not be a substantial risk as long as propagules for the founding of new populations are adequately sampled from within each source population before additional inbreeding takes place. The results further suggest that if mixing of propagules from different source populations is not required to increase within-population genetic variation in the reintroduced populations, it may best be avoided.