Human-induced land use change is a primary driver of species loss within ecological communities.   Recent experimental studies have documented the importance of abundance and richness of species within communities for ecological functioning; however, we have limited understanding of how community disassembly resulting from landscape change affects ecosystem functioning in natural systems.  We measured differences in pollinator communities and pollination of two spring ephemeral wildflower species along a gradient of increasing suburban development and decreasing woodland to explore whether landscape change drives disassembly of the native pollinator community and associated pollination function.  Specifically, we ask (1) does loss of natural habitat and associated increase in developed land lead to decreases in pollinator abundance and species richness; and (2) do these differences in pollinator communities affect the level of pollination to native plants in remnant habitats.  We established standard sized arrays of potted Claytonia virginica and Polemonium reptans at 20 woodland sites in the Delaware Valley region within and surrounding Philadelphia, USA.    All sites were in patches of mature deciduous woodland.  The surrounding landscape ranged from 12%-98% suburban/urban landcover within 1 km of the sites. To measure pollinator communities we recorded flower visitation by all insect taxa for two hours at each site during the main bloom period of each plant species.  On the same days that visitation was recorded we also subjected a sample of flowers of each species to either open pollination or open-plus-supplemented pollination treatments.  We recorded fruit and seed set for all treatment flowers.  Comparison of open versus supplemented treatments (pollen limitation) provided a measure of differences in pollination functioning among sites.  

Results/Conclusions

Pollen limitation depended on the rate of visitation for both C. virginica (r² = 0.49, F_1,18 =  5.30, P= 0.03) and P. reptans (r² = 0.25, F_1,24 = 8.09, P = 0.01).  Despite the dramatic landscape gradient, visitation, and visitor species richness did not depend on the proportion of developed or woodland habitat surrounding the site (C. virginica visits: F_1,15 = 0.35, P = 0.58, richness: F_1,15 = 0.12, P =0.74;  P. reptans visits:  F_1,20 = 1.92, P =0.18, richness: F_1,20 =1.07, P = 0.31 ).  As a result pollination was not affected by differences in surrounding landscape.  These results suggest that pollinator loss could affect plant reproduction of remnant native plant populations; however, urban development and the associated loss of natural vegetation does not drive pollinator loss in this system.