Dioecious species require a conspecific for sexual reproduction. As such, they demand a high level of flowering synchronization. In this scenario, phenology can deepen the effect of density, sex ratio, and distance over fruit production. Also, pollinator reliability can affect a species reproductive success (RS). Pollinator specialist depend on their few pollinators have a high population size; in contrast, generalist (pollinated by many taxa) can surpass this because of their condition. In

more »

... paper, we analyzed the effect of phenology, density, distance and pollinator abundance on the female RS of the dioecious palm Chamaedorea pinnatifrons (Palmae) at a cloud forest in Costa Rica. This palm is pollinated by wind and a single species of thrips (Thysanoptera: Thripidae) in a mixed pollination system. The pollination system seems to be highly specialized, as the thrips species has been recorded only in flowers of Chamaedorea Methods: In the 2012 season we monitored the phenology of a population (n = 88 males, 94 females) in a 40x40 plot. All individuals were sexed, mapped, had their height measured and number of leaves counted. RS was measured as the number of fruits/inflorescence and as fruit set. Generalized Linear Mixed Effects Model (GLMM) were used to test the effect of flowering time, distance to nearest male, density of males, number of flowers, height, number of leaves and number of flowers over the two response variables. We also estimated thrips population by sampling insects directly from staminate inflorescences in anthesis, from early March to mid-May (11 weeks, n = 33 inflorescences). Results: Flowering of C. pinnatifrons lasted five months: it began by early February, had a peak by early April, and ended by July. Mean fruit production was 99.7 (± 12.1 SE, min = 0, max = 735). The amount of fruits produced was significantly higher in inflorescences from individuals with shorter stems, had more flowers, a lower flowering synchrony with males and flowered late. Mean fruit set was 0.144 ± 0.016%. Fruit set was also significantly higher in inflorescences from lower stems, with lower male synchrony and flowered late. The mean amount of thrips sampled in staminate inflorescences increased 13-fold between early March (20.8 ± 10.6 thrips/rachillae) and mid-May (282.0 ± 74.4). Discussion: The higher RS of late flowering plants can be partially explained by the progressive increase in the population of pollinators. Our results in C. pinnatifrons seem to oppose recent analyses in that early flowering plants should be positively selected. A deeper understanding of a species pollination system is necessary to understand the strength of the agents of selection.