Evidence for pollinator dependence and absence of pollen limitation in Crassula (Crassulaceae) species in South Africa’s Drakensberg Mountains

Mountain regions are often characterized by highly variable pollination environments, in which self-compatibility and capacity for autonomous self-fertilization may buffer reproductive success of insect-pollinated plants. Within the biodiverse flora of southern Africa, the Drakensberg Mountain Centre (DMC) is characterized by high levels of floristic endemism, but very little is known about plants’ dependence on pollinators and vulnerability to variation in pollinator activity in this region. We used controlled pollination experiments to investigate reliance on pollinators and the contribution of pollinators to reproductive success in four Crassula (Crassulaceae) species that are indigenous to the montane elevation zone of the DMC. In all three species for which breeding systems were assessed, fecundity of flowers in the autonomous and hand self-pollination treatments was consistently lower than for flowers subjected to hand cross-pollination, natural-pollination, and pollen-supplementation treatments, indicating predominant reliance on receipt of cross-pollen for reproductive success. Indices of self-incompatibility (ISI) indicated strong self-incompatibility for C. natalensis and C. sarcocaulis subsp. rupicola (ISI = 0.84 and 0.94, respectively) and intermediate self-compatibility for C. setulosa var. setulosa (ISI = 0.54), whereas indices of autofertility (IAF) ranged from close to zero to a maximum of 0.2, indicating strong reliance on pollinators. Comparison of reproductive success of flowers which were naturally pollinated with that of flowers from which pollinators were excluded further confirmed the exclusive to near-exclusive dependence of all studied Crassula species on pollinators for fecundity. This also revealed high levels of pollinator-mediated reproductive success under natural conditions: vector-mediated pollination contributes at least 80% to 90% of fecundity in all four studied species. Despite strong pollinator dependence, manual addition of cross-pollen did not significantly increase fecundity in any of the three species, indicating that reproductive success was not limited by pollinator-mediated pollen receipt (pollen limitation index values all < 0.15 for three species). Results showing absence of pollen limitation despite strong pollinator dependence suggest that investigation of pollinators of the studied species may provide insight into pollination systems that can provide reliable pollinator service in high-elevation montane environments.