Search

Newly germinated seedlings are vulnerable to biomass removal but usually have at least 6 months to grow before they are exposed to dry‐season fires, a major disturbance in savannas. In contrast, plants are exposed to browsers from the time they germinate, making browsing potentially a very powerful bottleneck for establishing seedlings.Here we assess the resilience of seedlings of 10 savanna tree species to topkill during the first 6 months of growth. Newly germinated seeds from four dominant African genera from across the rainfall gradient were planted in a common garden experiment at the Wits Rural Facility and clipped at 1 cm when they were ~2, 3, 4 and 5 months old. Survival, growth and key plant traits were monitored for the following 2.5 years.Seedlings from environments with high herbivory pressure survived topkill at a younger age than those from low‐herbivore environments, and more palatable genera had higher herbivore tolerance. Most individuals that survived were able to recover lost biomass within 12 months, but the clipping treatment affected root mass fraction and branching patterns.Synthesis. The impact of early browsing as a demographic bottleneck can be predicted by integrating information on the probability of being browsed and the probability of surviving a browse event. Establishment limitation through early browsing is an under‐recognized constraint on savanna tree species distributions. [ABSTRACT FROM AUTHOR]

In savannas, ruminant herbivores can have divergent impacts on tree recruitment and subsequent woody cover. Whereas heavy grazing by cattle results in woody thickening, intensive grazing by wildlife instead tends to be associated with lower woody cover.To disentangle why woody cover is low in areas heavily grazed by wildlife, we investigated (a) whether grazing lawns attract indigenous mammalian browsers, and if a preference for short‐grass habitat decreases with browser body mass as predator susceptibility decreases; and (b) whether browser attraction to grazing lawns translates into the suppression of woody plants, including seedlings and saplings, thus maintaining the openness of heavily grazed short‐grass areas.In Kruger National Park, South Africa, we contrasted browser abundance (using dung counts) on grazing lawns and on low‐herbivory sites characterised by tall grass. Additionally, a herbivore exclosure experiment was set up to investigate the combined impact of browser removal and grass height habitat type on seedling survival and sapling growth of a dominant woody plant species. Finally, in Hluhluwe‐iMfolozi Park (HiP), we examined the long‐term (10 years) impact of browser removal on the growth rates of a range of woody species, monitored across ten sites along a gradient of herbivory ranging from grazing lawn to tall grass.Steenbok and impala selected short‐ over tall grass as preferred browsing sites, while elephant preferred tall grass. Browser abundance on short grass decreased with browser body mass, indicating that predator avoidance might be a key factor driving mesoherbivores to utilise grazing lawns.Seedling survival was lowest on grazing lawns when browsers were present, with mortality occurring in two out of every three seedlings. Similarly, sapling growth was lowest on grazing lawns, although browser removal had no significant effect. Evidence for increased browser impact on grazing lawns was clearest from our long‐term herbivore exclosure experiment in HiP, which demonstrated that browsers strongly modify the growth rates of woody plants in short‐grass habitats.Synthesis. These results provide support for the hypothesis that browsers, particularly browsing mesoherbivores and mixed feeders, are attracted to short‐grass habitats, and that they help maintain grazing lawn openness by suppressing seedling survival and woody plant growth where grass is kept short by grazers. [ABSTRACT FROM AUTHOR]