Effectiveness of the non-native Japanese white-eye as a novel pollinator of endemic Hawaiian plants

Background/Question/Methods

Current global extinction rates, arising from rapid environmental change, generate broadscale disruption in interspecific interactions. Losses of mutualists can erode ecosystem functions, provoke novel numerical and behavioral responses among remnant native species, and further threaten biodiversity. In some settings, non-native species appear to carry out the functions that had been provided by extinct native species. Hypothetically, such functional replacement could maintain native species dynamics and biodiversity. However, non-native mutualists likely differ behaviorally and morphologically from native mutualists, potentially leading to underperformance. Due to a combination of habitat loss, predator introductions, and disease, the Hawaiian Islands have lost the large majority of their native avian pollinators. We examined three endemic, congeneric Hawaiian plant species, contrasting in flower size, to determine whether and under what circumstances non-native birds are capable of effectively pollinating these native plants. Our approach coupled 306 hours of pollinator visitation observations with seed set data obtained from selfed, hand-supplemented, unmanipulated, and bird-visited flowers. Visitation rates by bird and plant species were calculated across focal study stands and compared with receptive flower availability. Treatment effects on seed set were examined using a linear mixed effects model with study stand and focal tree as blocking variables.

Results/Conclusions

Non-native Japanese white-eyes (Zosterops japonicus) performed the large majority of visitation to small-flowered Clermontia parviflora and midsized-flowered C. montis-loa (83.9% and 96.5%, respectively). Daily visitation rates to these species exceeded flower availability in all study stands (two-tailed Student’s t-test; p < 0.05). By contrast, the large-flowered C. hawaiiensis received only rare visitation from native birds (Hemignathus virens) and no visitation from Z. japonicus. Seed set following hand-supplementation, compared with unmanipulated flowers, was significantly enhanced across all three focal plant species (linear mixed model; p < 0.05). Unmanipulated flowers set significantly more seed than selfed flowers for C. parviflora and C. montis-loa (linear mixed model; p < 0.05), but not C. hawaiiensis (p = 0.72). These results suggest that successful pollination of C. parviflora and C. montis-loa is largely attributable to the non-native Z. japonicus. However, most current seed set by C. hawaiiensis is likely due to selfing. Furthermore, direct morphological comparisons between Z. japonicus and C. hawaiiensis demonstrate that the bird’s bill is too short to easily access the flower’s nectar pool. As a novel pollinator in Hawaii, the effectiveness of Z. japonicus appears to vary by target plant species and may be dictated by flower size and morphology.