With the current global concerns about pollinators, relationships between species interactions and diversity are pivotal. If pollinator communities depend strongly on the diversity of flowering plants and vice versa, anthropogenic influences—whether positive or negative—on one partner will cause changes in the other. Here we ask whether nectarivorous bird communities are structured by resource abundance (Proteaceae nectar) or Proteaceae diversity at different spatial scales in the Cape fynbos of South Africa. On a small spatial scale, we sampled 34 one‐hectare plots across the Cape Floristic Region (CFR) for flowering Proteaceae species, number of inflorescences, nectar volume, vegetation age, nectar‐feeding bird abundance and species richness. At small scale, nectar—rather than vegetation structure or plant community composition—was the most strongly correlated to nectar‐feeding bird diversity and abundance. On a landscape scale we investigated the spatio‐temporal flowering patterns of ornithophilous Proteaceae throughout the CFR. Similar flowering patterns—with a winter floral abundance peak—were found throughout the region, but Protea, Leucospermum and Mimetes showed complementary flowering periods. At large spatial scales ornithophilous Proteaceae species richness is strongly correlated—more so than plant or floral abundance—to the nectar‐feeding bird community. At large spatial scales resource diversity—and at a smaller scale resource abundance, shapes nectar‐feeding bird communities. Providing high volumes of nectar sugar throughout the year is key to restore the nectar‐feeding bird communities in small conservation areas.

Biological communities are increasingly faced with novel urban habitats and their response may depend on a combination of biological and habitat traits. The response of pollinator species to urban habitats are of particular importance because all species involved in the pollination mutualism may be affected. Nectarivorous bird communities worldwide show varying tolerances to urban areas, but studies from Africa are lacking. We investigated nectarivorous bird communities in a medium‐sized South African city and asked which biological and garden traits best predict the community assembly of specialist and opportunistic nectarivorous birds. Information was collected on garden traits and the frequency of nine nectarivorous bird species for 193 gardens by means of a questionnaire. Information on biological traits of birds was obtained from published literature. Habitat generalism and tree nesting were identified as the most important biological traits influencing bird occurrence in gardens. A greater diversity of indigenous bird‐pollinated plants and the presence of sugar water feeders increased the numbers of nectar specialist birds and species richness of nectarivorous birds. While bird baths increased the species richness of nectar specialist birds, opportunistic birds’ urban adjustment was further facilitated by large vegetated areas in gardens and limited by the distance to the nearest natural habitat. In conclusion, though some biological traits and dispersal barriers seem to limit urban adjustment, a combination of natural and artificial nectar resource provisioning could facilitate this adjustment.

1. Flower colour differs dramatically between populations for some plant species; yet, we know little about what drives this variation. Such polymorphisms can be influenced by plant-pollinator interactions, but whether they are also influenced by pollinator-mediated plant-plant interactions is unknown.2. We test whether flower colour polymorphisms can arise through convergence (facilitation) or divergence (competition) of flower phenotypes resulting from plantplant interactions mediated by the shared, and only, pollinator (orange-breasted sunbird) of 10 Erica communities in South Africa.3. Sunbird visitation rates to the less-preferred Erica species in communities increased with colour similarity to the most-visited species, suggesting that polymorphisms can be maintained by selection for colour convergence within communities, which promotes shared signals through rewarding mimicry (i.e. facilitation). 4. Colour similarity was lowest when risk of reproductive interference was greatest: that is, when reproductive morphology (stigma-anther distance) and flowering phenology was most similar, and when floral density differed greatly between species. This implies that polymorphisms can also be maintained by selection for colour divergence, since this promotes assortative pollinator foraging (i.e. competition).5. Different populations of a species may experience opposite selective pressures, depending on their community context. This is the first evidence that a geographic mosaic of plant-plant pollination interactions could maintain colour polymorphisms in congeneric species sharing a single pollinator.

No abstract