Mating system and genetic diversity of progenies before and after logging: a case study of Bagassa guianensis (Moraceae), a low-density dioecious tree of the Amazonian forest

The logging of large trees in tropical forests causes a decrease in the density of reproductive individuals, which likely affects the pattern of pollen dispersal and the mating system of the remaining trees in the population. Here, we investigate the impact of logging on mating system and genetic diversity of the low-density, thrip-pollinated, dioecious tree Bagassa guianensis within a 500-ha plot at Tapajós National Forest, Pará State, Brazil. Mating system parameters of the logged population were estimated using mixed-mating model. Six microsatellite loci were used to genotype 232 seeds from ten remnant female trees over three seasons (2006, 2007, and 2008). The data were compared with the mating system of the unlogged population of B. guianensis (18 female trees, 488 seeds) in the same plot. The overall number of alleles found in the open-pollinated progenies decreased after logging (K = 71 before logging and K = 57 after logging, considering the three after logging seasons pooled), as well as the average number of alleles per locus (A = 11.8 and 9.5, respectively). Similarly, the number of private alleles, which is defined in the context of this study as the number of alleles observed in the population exclusively before or after logging, also decreased after logging (15 and 1, respectively). However, the average number of alleles per locus and the observed and expected heterozygosities were not significantly higher before logging than after logging. Logging also did not affect the inbreeding of progenies in the population. Somewhat unexpectedly, the effective number of pollen donors was consistently higher after (Nₑₚ= 14.5 in 2006, 8.1 in 2007, and 6.3 in 2008) than before logging (Nₑₚ= 3.4), suggesting that the loss of alleles in the population was compensated by the higher heterogeneity in the pollen pool after the removal of the largest trees by logging. Potential causes explaining the patterns found here include the maintenance of large patches of forest around the logged plot and the species’ ability to perform long-distance pollination by airborne thrips. Maintenance of landscape integrity as logged and unlogged forests around the exploited areas is recommended to enhance pollen migration and to avoid long-term losses in genetic diversity.