Adding landscape genetics and individual traits to the ecosystem function paradigm reveals the importance of species functional breadth

Abstract

Animal pollination mediates both reproduction and gene flow for the majority of plant species across the globe. However, past functional studies have focused largely on seed production; although useful, this focus on seed set does not provide information regarding species-specific contributions to pollen-mediated gene flow. Here we quantify pollen dispersal for individual pollinator species across more than 690 ha of tropical forest. Specifically, we examine visitation, seed production, and pollen-dispersal ability for the entire pollinator community of a common tropical tree using a series of individual-based pollinator-exclusion experiments followed by molecular-based fractional paternity analyses. We investigate the effects of pollinator body size, plant size (as a proxy of floral display), local plant density, and local plant kinship on seed production and pollen-dispersal distance. Our results show that while large-bodied pollinators set more seeds per visit, small-bodied bees visited flowers more frequently and were responsible for more than 49% of all long-distance (beyond 1 km) pollen-dispersal events. Thus, despite their size, small-bodied bees play a critical role in facilitating long-distance pollen-mediated gene flow. We also found that both plant size and local plant kinship negatively impact pollen dispersal and seed production. By incorporating genetic and trait-based data into the quantification of pollination services, we highlight the diversity in ecological function mediated by pollinators, the influential role that plant and population attributes play in driving service provision, and the unexpected importance of small-bodied pollinators in the recruitment of plant genetic diversity.

Keywords: bees; body size; ecological services; pollen-mediated gene flow; pollination.

Fig. 1.

Effect of pollinator body size as measured by ITD (A) and local kinship (B) on seed viability. Seed viability refers to the ratio between viable and aborted seeds. Dots represent the predicted means from the model at the inflorescence level on a logarithmic scale. SEs were calculated using 100 bootstrap replicates.

Fig. 2.

Pollen-dispersal distances recorded for each M. affinis pollinator. (A) Boxplots indicate the median, upper, and lower quartile for each M. affinis pollinator. Increasing numbers in the y axis correspond with the following pollinator species sorted in ascending order regarding their body size (ITD): 1: Halictidae sp 2; 2: Trigonisca buyssoni; 3: Tetragonisca angustula; 4: Halictidae sp 1; 5: Trigona muzoensis; 6: Paratetrapedia lineata; 7: Trigona fuscipennis; 8: Trigona amalthea; 9: Trigona fulviventris; 10: Pseudaugochloropsis sp 1; 11: Melipona fuliginosa; 12: Melipona panamica; 13: Centris dichrootricha; 14: Xylocopa fimbriata. (B–E) Frequency distribution of the pollen-dispersal distances for a subsample of M. affinis pollinators representative of the body-size gradient: (B) Melipona panamica; (C) Trigona fulviventris; (D) Trigona fuscipennis; and (E) Tetragonisca angustula.

Fig. 3.

Effect of DBH_mother on pollen-dispersal distance. (A) Pollen-dispersal distance was square root (sqrt) transformed, and DBH_mother is plotted at its original scale. Dots represent the predicted means from the model at the infructescence level. SEs were calculated using 1,000 bootstrap replicates. (B) Relationship between DBH_mother and total number of inflorescences per tree. Transparency was used to show the level of overlapping among dots.