Evidence for rodent pollination in Erica hanekomii (Ericaceae)

Pollination by rodents, originally discovered in Cape Proteaceae, is known from a handful of plant lineages and may be more widespread than current data suggest. Here, we demonstrate the occurrence of this mode of pollination in Erica, the largest plant genus in the Cape flora of South Africa. The localized endemic Erica hanekomii appears to be adapted for pollination by rodents, on account of its low, mat‐forming habit with many dull‐coloured, pendulous flower heads, large volumes (up to c. 30 µL per inflorescence) of sucrose‐dominated nectar, hook‐shaped styles and late winter flowering phenology. Rodents trapped in the vicinity of E. hanekomii plants had many thousands of Erica tetrads in their droppings. Nocturnal laboratory observation showed that Acomys subspinosus (Cape spiny mouse) is adept at manipulating E. hanekomii inflorescences and lapping nectar without destroying gynoecia. During terrarium experiments, pollen was deposited mainly on the rostrum, but also on the chest area and forelegs of A. subspinosus. No birds or insects were seen to visit E. hanekomii in the field. Whereas natural fruit set was 73%, inflorescences from which rodents, but not insects, were excluded using mesh cages, had fruit set of just 3.3%. Exclusion of all visitors resulted in zero seed production. Broken anther rings (a sign of flower visitation in Erica) were observed in 90.1% of flowers in the field and 92% of flowers exposed to rodents in the laboratory, but in virtually none of the flowers from which rodents were selectively excluded. This first record of a non‐flying mammal pollination syndrome in Ericaceae raises interesting new questions about the origins of the rodent pollination system and the selective factors that lead to its evolution. © 2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 166 , 163–170.