Tropical forest fragmentation threatens biodiversity, yet basic information on population responses for major groups such as plants is lacking. Hypervariable genetic markers were used to reconstruct a population-level pedigree in fragmented tropical forest for the tree Symphonia globulifera. Though seedlings occurred only in remnant forest, the pedigree showed that most seedlings had been produced by sequentially fewer adults in pasture, creating a genetic bottleneck. The pedigree also implicated shifts in the foraging of animals that disperse pollen and seed in a secondary constriction of the bottleneck. These results suggest that tropical conservation strategies should anticipate complex, cryptic responses to fragmentation.

We developed genetic markers for three microsatellite loci in the tropical tree Symphonia globulifera and used them to examine the demographic genetic consequences of forest fragmentation. High levels of genetic variation were revealed in samples of adults, saplings, and seedlings. The more-variable loci exhibited less stability in allelic composition across sites and stages. The number of alleles per hectare (ha) of forest was similar when continuous forest plots were compared to plots from fragmented forest for all three stages. This pattern also held for the number of unique multilocus adult and sapling genotypes, but the number of unique seedling genotypes per ha of fragmented forest greatly exceeded expectations based on continuous forest data, probably due to the concentration of seeds into remnant forest patches by foraging bats. Significant inbreeding and genetic differentiation were most often associated with the fragmented forest and the seedlings. Finally, principal component analysis reaffirmed that a bottleneck, acting in concert with pre-existing genetic structure in the adults, had led to enhanced and rapid divergence in the seedlings following deforestation, a result that is of central interest for landscape management.

No abstract

We provide primer sequences for 14 (GA)n microsatellite loci developed from northern red oak, an important timber species. We screened loci using two sets of samples. A parent–offspring set included DNA from seven acorns collected from one mother tree along with maternal DNA, to determine that all progeny carried a maternal allele at each locus. The other set was comprised of 10 adult trees sampled from Indiana old‐growth forest, providing a measure of diversity revealed by each locus.

The red oaks (Quercus section Lobatae) include important timber species, but we know little about their gene pools. Red oak species can be difficult to identify, possibly because of extensive interspecific hybridization, although most evidence of this is morphological. We used 15 microsatellite loci to examine the genetic composition of a red oak community in 20.6 ha of an Indiana old-growth forest. The community included northern red oak (Quercus rubra L.), Shumard oak (Quercus shumardii Buckl.), and pin oak (Quercus palustris Muenchh.). Species were identified using whole-tree silvic characters, the approach most often implemented by foresters. We found high genetic diversity within species but limited genetic differences between species. Phenetic clustering showed that Q. rubra and Q. shumardii were more genetically similar than either was to Q. palustris, but a neighbor-joining tree revealed that individuals of the different species did not resolve into single-species clusters. We identified four mixed-species subpopulations using Structure, a computer program based on Monte Carlo simulation. The three largest groups are consistent with the following biological interpretations: (i) pure Q. rubra, (ii) Q. rubra, Q. shumardii, and their hybrids, and (iii) Q. rubra, Q. shumardii, Q. palustris, and their hybrids. We discuss the implications of these findings for the whole-tree silvic approach to selection and for management of the red oak gene pool.

Fifty-six accessions of cultivated and wild sorghum were surveyed for genetic diversity using 50 low-copy-number nuclear DNA sequence probes to detect restriction fragment length polymorphisms (RFLPs). These probes revealed greater genetic diversity in wild sorghum than in cultivated sorghum, including a larger number of alleles per locus and a greater portion of polymorphic loci in wild sorghum. In comparison to previously published isozyme analyses of the same accessions, RFLP analysis reveals a greater number of alleles per locus. Furthermore, many RFLP alleles have frequencies between 0.25-0.75, while the vast majority of isozyme alleles are either rare (< 0.25) or near fixation (> 0.75). Correlations between genetic and geographic distances among the accessions were stronger when calculated with RFLP than with isozyme data. Systematic relationships revealed by nuclear and chloroplast restriction site analysis indicate that cultivated sorghum is derived from the wild ssp. arundinaceum. The portion of the wild gene pool most genetically similar to the cultivars is from central-northeastern Africa. Previous published data also suggested that this is most likely the principal area of domestication of sorghum. Introgression between wild and cultivated sorghum was inferred from disconcordant relationships shown by nuclear and chloroplast DNA markers. Introgression apparently occurs infrequently enough that the crop and its wild relatives maintain distinct genetic constitutions.

Polymorphic allozyme loci were used to estimate outcrossing rates for three tree species from a disturbed dry forest in southern Costa Rica. Estimates of the multilocus outcrossing rates of Cedrela odorata and Jacaranda copaia were 0.969 and 0.982, respectively, and suggest that these species may be self‐incompatible. The subcanopy tree Stemmadenia donnell‐smithii also demonstrated little self‐fertilization based on an estimated outcrossing rate of 0.896. Significant heterogeneity in pollen allele frequencies among maternal trees was detected for at least two enzyme loci for each species. A test of correlated mating between progeny of S. donnell‐smithii revealed that all seeds within a fruit were singly sired. In addition, the low estimates of biparental inbreeding and significant differences in pollen and ovule allele frequencies for this species suggest that gene flow into the sampled forest fragment may occur. The implications of deforestation on the mating systems of these tropical tree taxa are discussed.

Patterns of allozyme variation were surveyed in collections of cultivated and wild sorghum from Africa, the Middle East, and Asia. Data for 30 isozyme loci from a total of 2067 plants representing 429 accessions were analyzed. Regional levels of genetic diversity in the cultivars are greater in northern and central Africa compared to southern Africa, the Middle East, or Asia. The spatial distribution of individual alleles at the most variable loci was studied by plotting allele frequencies on geographic maps covering the distribution of sorghum. Generally, many of the alleles with frequencies below 0.25 are localized in specific portions of the range and are commonly present in more than one race in that region. Several alleles occur in both wild and cultivated sorghum of one region and are absent from sorghum elsewhere, suggesting local introgression between the wild and cultivated forms. Although the same most common allele was found in the wild and cultivated gene pools at 29 of the 30 loci, phenetic analyses separated the majority of wild collections from the cultivars, indicating that the two gene pools are distinct. Wild sorghum from northeast and central Africa exhibits greater genetic similarities to the cultivars compared to wild sorghum of northwest or southern Africa. This is consistent with the theory that wild sorghum of northeast-central Africa is ancestral to domesticated sorghum. Wild sorghums of race arundinaceum of northwest Africa and race virgatum from Egypt are shown to be genetically distinct from both other forms of wild sorghum and from the cultivars. Suggestions for genetic conservation are presented in light of these data.