Any conservation actions that preserve some populations and not others will have genetic consequences. We used empirical data from four rare plant taxa to assess these consequences in terms of how well allele numbers (all alleles and alleles occurring at a frequency of > 0.05 in any population) and expected heterozygosity are represented when different numbers of populations are conserved. We determined sampling distributions for these three measures of genetic diversity using Monte Carlo methods. We assessed the proportion of alleles included in the number of populations considered adequate for conservation, needed to capture all alleles, and needed to meet an accepted standard of genetic-diversity conservation of having a 90–95% probability of including all common alleles. We also assessed the number of populations necessary to obtain values of heterozygosity within $±$10% of the value obtained from all populations. Numbers of alleles were strongly affected by the number of populations sampled. Heterozygosity was only slightly less sensitive to numbers of populations than were alleles. On average, currently advocated conservation intensities represented 67–83% of all alleles and 85–93% of common alleles. The smallest number of populations to include all alleles ranged from 6 to 17 (42–57%), but < 0.2% of 1000 samples of these numbers of populations included them all. It was necessary to conserve 16–29 (53–93%) of the sampled populations to meet the standard for common alleles. Between 20% and 64% of populations were needed to reliably represent species-level heterozygosity. Thus, higher percentages of populations are needed than are currently considered adequate to conserve genetic diversity if populations are selected without genetic data.