With biomass densities comparable to large ungulates and megaherbivores, termites play a key functional role in many tropical savanna ecosystems. This study focuses on vegetated termite mounds (termitaria) constructed by the Termitidae species Macrotermes herus. We studied how resource rich termitaria affect graminoid herbs (Poaceae and Cyperaceae), forbs and woody species composition and diversity. The density of termitaria explained 89% of the variation in dense thickets in the area. Fire tolerant Acacia species dominated the open savanna while fire sensitive species like Grewia spp. and the succulent Euphorbia candelabrum were restricted to termite mounds. Termitaria plots had four times the mean number of woody species and supported three times as many forb species as the adjacent savanna. For woody species, both the Shannon-Wiener index and the Shannon evenness index were higher on temitaria than on the savanna. There were no differences for graminoid herbs, except for the Shannon evenness index which was higher on termitaria. Our results indicate that graminoid herb richness peaks at lower productivity levels than trees and forbs in savanna ecosystems, as also recently found in temperate areas.

Both density dependent and density independent processes such as climate affect population dynamics in large herbivores. Understanding herbivore foraging patterns is essential to identify the underlying mechanisms behind variation in vital rates. However, very little is known about how animals vary their selection of habitat temporally, alone or in interaction with density during summer. At the foraging scale, we tested using a fully replicated experiment whether domestic sheep Ovis aries stocked at high (80 per km 2 ) and low (25 per km 2 ) densities (spatial contrasts) varied their habitat selection temporally over a four year period. We predicted reduced selection of high productivity vegetation types with increasing density, and that seasonal and annual variation in climate would affect this density dependent selection pattern by increasing competition for high quality habitats in late grazing season and in years with poor vegetation development and over time related to vegetation responses to grazing. As predicted from the Ideal free distribution model, selection of high productivity habitat decreased at high density. There was also a marked temporal variation in habitat selection. Selection of the most productive vegetation types declined towards the end of each grazing season, but increased over years both at low and high sheep density. There was only weak evidence for interactions, as selection ratio of highly productive habitats tended to increase more over years at low density as compared to high density. Limited interactive effects of density and annual variation on habitat selection during summer may explain why similar interactions in vital rates have rarely been reported for summer seasons. Our results are consistent with the view that variation in habitat selection is a central mechanism for climate and density related variation in vital rates.

In constructing large vegetated mounds, Macrotermes termites play a key functional role in many African savanna systems. This study focuses on ungulate feeding on Macrotermes termitaria vegetation in Lake Mburo National Park, Uganda. With the exception of oribi Ourebia ourebi, all the species studied (i.e. impala Aepyceros melampus, Burchell's zebra Equus burchelli, Defassa waterbuck Kobus ellipsiprymnus, topi Damaliscus lunatus, bushbuck Tragelaphus scriptus and warthog Phacochoerus africanus) grazed preferentially on mound areas compared to the adjacent savanna. Ruminant species grazed closer to termitaria than non-ruminants and female impala both browsed and grazed closer to mounds than males. No sexual difference in grazing distance to mounds was found for waterbuck.

No abstract

Density dependent processes affecting foraging strategies may in turn influence vital rates and population regulation in large herbivores. Increased competition may lower both forage availability and quality, but whether the main activity constraint at high density is increased searching time or increased digestion time is poorly investigated. In a fully replicated landscape‐scale experiment, we used long‐term data (2003–2009) from domestic sheep grazing at high and low density (80 and 25 sheep km–2, respectively) on alpine summer ranges to test density dependence in allocation of time to feeding (moving) vs digestion (resting) activities and how this in turn affected body growth. Sheep at high density spent more time actively feeding than sheep at low density, but sheep moved shorter distances while foraging at high density. Increased activity levels at high density suggest that the main activity constraint at high density was availability of high‐quality food increasing searching time and possibly reducing intake rates. Increased movement distances at low density is consistent with a higher selection for more productive vegetation types since high‐quality patches are dispersed in the landscape. The alternative hypothesis, that food processing time increased at high density was not supported as it would have reduced overall activity levels. Individual activity levels increased body growth, but this was not sufficient to fully compensate for lower habitat quality leading to an overall reduced body growth at high density. Our experiment clearly documents changes in activity budgets and movement distances of a large herbivore at high population density, providing one potential behavioural mechanism of density dependent responses observed in vital rates.