How important is the relationship between Protea humiflora (Proteaceae) and its non-flying mammal pollinators?

The interaction between Protea humiflora and its non-flying mammal pollinators was investigated with the aim of quantifying the relationship for both the plants and mammals involved. We manipulated flower density by inflorescence removal and determined the effect upon mammal captures, and excluded mammals to determine their effect upon seed set. Of three mammal species captured frequently, the smallest, Acomys subspinosus (18±4 SD g, Rodentia: Muridae), demonstrated the strongest relationship with P. humiflora. Breeding coincided with flowering, and pollen constituted 33.5±27.5% by volume (max 80%, n=19) of winter scats. Acomys captures on control grids were twice those on inflorescence-removal grids during flowering. Aethomys namaquensis (51±14 g, Rodentia: Muridae) included flower products in its diet (pollen comprising 3.2±4.9% of winter scats, n=73), but captures were not affected by inflorescence removal. Aethomys is a seasonal breeder and juveniles entered the population 2 months after flowering. Elephantulus edwardii, although an insectivore (48±7 g, Macroscelidea: Macroscelididae), carried more pollen grains on its nose than the rodents, acquired while foraging in inflorescences for insects. Pollen comprised 3.1±5.4% (n=23) of winter scats. From the plant perspective, seed set was significantly reduced in caged inflorescences and in those >20 cm above the ground, presumably due to limited accessibility by terrestrial animals. We conclude that small mammals are responsible for around half (56%) of the effective seed set in P. humiflora.