A comparison of direct and indirect measures of gene flow in the bat-pollinated tree Hymenaea courbaril in the dry forest life zone of southwestern Puerto Rico

Gene flow was estimated in the tropical tree Hymenaea courbaril using indirect measures based on the proportion of total genetic diversity from differences among populations (F-ST) and using direct paternity analyses. Outcrossing estimates (t(m) = 1.07) are consistent with indications that H. courbaril is self-incompatible. Direct estimates of pollen gene flow into two small (n < 6) stands of trees indicate that 42% and 62% of the pollen came from outside the stand. Despite this level of external gene flow, genetic drift is expected to cause divergence of populations over time (migration rate N(e)m = 0.40 and 0.63), mainly because of the small number of adult trees within each population. Indirect estimates of gene flow based on nine larger populations of 21-48 trees gave estimates of 2.91 migrants per generation, but the absence of migration-drift equilibrium renders this estimate unreliable. Analysis of pollen movement among 21 adult trees indicated that the effective number of fathers was only 6.9, with the majority of pollinations coming from just three trees. Asynchronous flowering, and perhaps restricted use of flowering resources by bats, the primary pollinators of H. courbaril, is the likely cause of this low number. Genetic diversity values based on 31 allozyme loci were lower than those reported for other outcrossing, gravity-dispersed tree species. Few effective fathers, and a relatively low population density, are the likely reasons.