Beautiful but buzz-less: Urban bushland fragments harbor few native bees to the detriment of an endemic buzz-pollinated plant

Background/Question/Methods

Urban remnants of native vegetation can act as rich repositories of biodiversity, but their conservation value depends on whether they also conserve ecological processes such as pollination. We expect that larger and more recently isolated remnants have a greater likelihood of containing higher biodiversity and maintaining ecological function. Native pollinators, particularly bees, are in decline in much of the world due to habitat loss and competition from non-native pollinators. Changes in their abundance may especially affect the reproduction of plants that are pollen-limited and rely on specialized pollinators. We investigated whether native bee abundance in urban remnants is predicted by remnant size or time since isolation. We sampled bees with three methods (window traps, colored pan traps, and netting) at 8 urban Banksia woodland remnants (2 small young, 2 small old, 2 large young, 2 large old) near Perth, Western Australia once per month from spring to early summer 2010 (September to December). We also recorded numbers and identities of flowering plants in randomly selected quadrats at the sites. In 2011, at three of the sites we conducted a pollen supplementation and pollinator exclusion experiment on an endemic plant (Conostephium pendulum(Ericaceae)) that releases its pollen only when it is sonicated (buzz pollinated) by certain native bee species.

Results/Conclusions

From all methods, sampling dates, and sites combined, we captured a total of 316 bees representing 10 genera, of which 75 (24%) were native buzz-pollinating bees, 23 (7%) were non-buzz pollinating native bees, and 218 (69%) were the introduced European honeybee, Apis mellifera. We found no significant relationship between the number of native or honey bees captured and size or age of the remnant, or the number of flowers or flowering plants or diversity of flowers. Pollen supplemented C. pendulum had 45% fruit-set compared to 0-5% fruit-set for flowers left open for natural pollination (p <0.05).  Only two native bees were observed to visit C. pendulum in 18 hours of observations. We conclude that size and time since isolation are not good predictors of native or honey bee abundance in Banksia woodland remnants, possibly because native bees sensitive to habitat changes or the presence of the honey bee may have already disappeared, and honey bees have longer flight ranges and may be drawn from managed hives. Plants that rely on native bees for pollination, such as buzz-pollinated plants, may be at risk of eventual extinction.