Do surface plant and soil seed bank populations differ genetically? A multipopulation study of the desert mustard Lesquerella fendleri (Brassicaceae).

Seed banks are an important component of many plant populations, but few empirical studies have investigated the genetic relationship between soil seeds and surface plants. We compared the genetic structure of soil seeds and surface plants of the desert mustard Lesquerella fendleri within and among five ecologically diverse populations at the Sevilleta National Wildlife Refuge in Central New Mexico. At each site, 40 Lesquerella surface plants and 40 samples of soil seeds were mapped and genetically analyzed using starch gel electrophoresis. Overall allele frequencies of soil seeds and surface plants showed significant differences across the five populations and within three of the five individual populations. Surface plants had significantly greater amounts of single and multilocus heterozygosity, and mean surface plant heterozygosity was also greater at the total population level and in four of the five individual populations. Overall soil seed (bot not surface plant) homozygosity was significantly greater than predicted by Hardy-Weinberg expectations at the total and individual population levels. Although F-alpha estimates revealed similarly small but significant genetic divergence within each life-history stage, estimates of coancestry showed that fine-scale (0.5-2 m) genetic correlations among the surface plant genotypes were roughly twice those of soil seed genotypes. An unweighted pair group method with arithrnetic mean cluster analysis indicated that in the two geographically closest sites, the surface plants were slightly more genetically similar to each other than to their own respective seed banks. We also found weak and/or negative demographic associations between Lesquerella soil seed and surface plant densities within each of the five sites. We discuss the difficulties involved with sampling and genetically comparing these two life-history stages.