No abstract

The nature and the strength of plant–frugivore interactions often vary along an antagonism–mutualism continuum and are highly influenced by the local ecological context (e.g. level of environmental disturbances). However, little is known concerning how the local ecological setting where plant–frugivore interactions take place affects the seed dispersal effectiveness (SDE) and, eventually, plant recruitment. This knowledge gap relates to the scarcity of empirical investigations on individual‐based plant–frugivore networks. We assessed whether the assembly of interaction modes (antagonist and mutualist) of the dwarf palm Chamaerops humilis (Arecaeae) affect the individual and the population level SDE in two Mediterranean sites differing in perturbation levels. We analysed the frequency distribution of interaction typologies and test its relationship with variation among individual palms in SDE. Additionally, we document how variation in interaction motif frequency (e.g. overrepresented interaction typologies) relates to changes in the SDE landscape of both disturbed study sites. We found that the interaction typologies of individual palms and its frugivores did not occur randomly. In a more complex landscape, interactions were more diverse and less dominated by simpler typologies, which may increase the stability of the plant–disperser interaction over the long term. We demonstrate that plants that interact with a more complex assemblage of frugivores, including both mutualistic and antagonistic partners, had the highest probability of recruitment. We found that the highly variable SDE among individual palms resulted from a few interaction modes occurring at higher than expected frequencies. We suggest this may be a reasonably frequent situation in diversified frugivore assemblages with a high heterogeneity of functional effects among mutualistic and antagonistic partners.

It has been suggested that there is a geographic dichotomy in the pollination systems of chiropterophilous columnar cacti: in intra-tropical areas they are pollinated almost exclusively by bats, whereas in extratropical areas they are pollinated by bats, birds and bees. However, currently the studies are clumped both taxonomically (mainly Pachycereeae species) and geographically (mainly in the Tehuacan Valley and the Sonoran Desert). This clumping limits the possibility of generalising the pattern to other regions or cactus tribes. Only four of the 36 chiropterophilous cacti in Pilosocereus have been studied. Despite the tropical distribution of two Pilosocereus species, bees account for 40-100% of their fruit set. We examined how specialised is the pollination system of P. leucocephalus in eastern Mexico. As we studied tropical populations, we expected a bat-specialised pollination system. However, previous studies of Pilosocereus suggest that a generalised pollination system is also possible. We found that this cactus is mainly bat-pollinated (bats account for 33-65% of fruit set); although to a lesser degree, diurnal visitors also caused some fruit set (7-15%). Diurnal visitors were more effective in populations containing honeybee hives. P. leucocephalus is partially self-compatible (14-18% of fructification) but unable to set fruit without visitors. Despite the variation in pollination system, P. leucocephalus shows more affinity with other columnar cacti from tropical regions than with those from extratropical regions. Although we report here that a new species of tropical Pilosocereus is relatively bat-specialised, this Cereeae genus is more flexible in its pollination system than the Pachycereeae genera.