Handling by avian frugivores affects diaspore secondary removal

Abstract

The balance between the costs and benefits of fleshy fruit production depends on the feeding behavior of their seed dispersers, which might effectively disperse seeds to farther areas or drop beneath parent plants some diaspores they handle during frugivory bouts. Nevertheless, the consequences of variation in fruit handling by primary seed dispersers on the secondary removal of diaspores remains poorly understood. We conducted a field study to determine how variation in fruit handling by avian frugivores affects short-term secondary removal of Miconia irwinii (Melastomataceae) diaspores by the ground-dwelling fauna in campo rupestre vegetation, southeastern Brazil. We conducted factorial experiments manipulating: (1) different outcomes of primary fruit/seed removal by birds, (2) distances of diaspore deposition from conspecifics, and (3) the access of ants and vertebrates to diaspores. We showed that secondary removal of diaspores was highly variable at the population scale, with an overall low removal rate by the ground-dwelling fauna (13% seeds, 19% fruits). However, we found that gut-passed seeds embedded in bird feces were less removed than seeds expelled from fruits. Gut-passed seeds were more likely to be removed by ant species acting as secondary dispersers, whereas pulp-free seeds dropped by birds were likely to interact with potential seed predators, including ants and rodents. We found no clear effect of dispersal from parent plant vicinity on seed removal, but fruit removal was significantly higher near parent plants. Partially defleshed fruits were more removed than intact fruits. The removal of fruits by ant and vertebrate rescuers, including lizards and birds, might reduce the costs of interactions with less effective dispersers that drop partially defleshed fruits under parent plants. Our study highlights that variation in fruit handling by primary avian seed dispersers mediate subsequent interactions among discarded diaspores and ground-dwelling animals, potentially affecting final seed fates. Moreover, we argue that escape-related benefits of dispersal can be contingent on how primary dispersers handle and discard seeds.

Fig 1. Schematic representation of the factorial experimental design employed to access the effects of diaspore handling, distance of deposition and vertebrate exclosure on removal of Miconia irwinii seeds and fruits by the ground-dwelling fauna in campo rupestre vegetation.

Distribution of diaspore piles across eight treatment combinations, including two treatments simulating outcomes of diaspore handling by birds paired on each tracking station. We made diaspore piles in one tracking station accessible to ants plus vertebrates, whereas the nearby tracking station with diaspore piles were excluded from vertebrates but accessible to ants. We placed one sampling block near and the other 25 meters far from a fruiting plant.

Fig 2. Handling, discard and dispersal of Miconia irwinii diaspores by avian frugivores in campo rupestre vegetation.

(A) Relative contribution of bird species to seed dispersal through endozoochory (fruit ingestion) or diaspore discard near parent plants (fruit pecking or drop). (B) Estimated flight distance of post-feeding displacements made by potential seed dispersers. Elaenia obscura is a gulper, but it was not represented in the figure because it contributed to less than 1% of fruit removal. * Gulpers, ** Mashers.

Fig 3. Variation in number of seeds falling under fruiting Miconia irwinii treelets in campo rupestre vegetation.

Fig 4. Miconia irwinii diaspore removal by the ground-dwelling fauna in campo rupestre vegetation.

(A) Effects of gut passage (gut-passed vs. pulp-free seeds), distance from conspecific (near vs. far) and vertebrate exclosure (open vs. excluded) on the proportion of Miconia irwinii seeds removed. (B) Effects of fruit handling (partially defleshed vs. intact fruits), distance from conspecific and vertebrate exclosure on the proportion of Miconia irwinii fruits removed. Squares represent medians, boxes represent 25–75% percentiles, whiskers represent non-outlier ranges and circles show the outliers. Distinct letters denote significant differences among treatments.

Fig 5. Ants interacting with Miconia irwinii diaspores on the ground of campo rupestre vegetation.

(A) Latency time (time elapsed between placing diaspores on the ground and the first interaction with ants) across different experimental treatments. (B) Number of ant species interacting with diaspore piles. Squares represent mean, boxes standard error and whiskers standard deviation), asterisk denotes significant differences (Paired t test—p < 0.001).