PeerJ (Sep 2016)
Reproductive ecology and isolation of Psittacanthus calyculatus and P. auriculatus mistletoes (Loranthaceae)
Abstract
Background Relationships between floral biology and pollinator behavior are important to understanding species diversity of hemiparasitic Psittacanthus mistletoes (c. 120 species). We aimed to investigate trait divergence linked to pollinator attraction and reproductive isolation (RI) in two hummingbird-pollinated and bird-dispersed Psittacanthus species with range overlap. Methods We investigated the phylogenetic relationships, floral biology, pollinator assemblages, seed dispersers and host usage, and the breeding system and female reproductive success of two sympatric populations of P. calyculatus and P. auriculatus, and one allopatric population of P. calyculatus. Flowers in sympatry were also reciprocally pollinated to assess a post-mating component of RI. Results Hummingbird assemblages differed between calyculatus populations, while allopatric plants of calyculatus opened more but smaller flowers with longer lifespans and produced less nectar than those in sympatry. Bayesian-based phylogenetic analysis indicated monophyly for calyculatus populations (i.e. both populations belong to the same species). In sympatry, calyculatus plants opened more and larger flowers with longer lifespans and produced same nectar volume than those of auriculatus; populations shared pollinators but seed dispersers and host usage differed between species. Nectar standing crops differed between sympatric populations, with lower visitation in calyculatus. Hand pollination experiments indicated a predominant outcrossing breeding system, with fruit set after interspecific pollination two times higher from calyculatus to auriculatus than in the opposite direction. Conclusions Given the low genetic differentiation between calyculatus populations, observed trait divergence could have resulted from changes regarding the local communities of pollinators and, therefore, expected divergence for peripheral, allopatric populations. Using RI estimates, there were fewer heterospecific matings than expected by chance in P. calyculatus (RI4A = 0.629) as compared to P. auriculatus (RI4A = 0.20). When considering other factors of ecological isolation that affect co-occurrence, the RI4C values indicate that isolation by hummingbird pollinators was less effective (0.20) than isolation by host tree species and seed dispersers (0.80 and 0.60, respectively), suggesting that host usage is the most important ecological isolation factor between the two species. Accordingly, the absolute and relative cumulative strength values indicated that the host tree species’ barrier is currently contributing the most to maintaining these species in sympatry.
Keywords