Large areas of moist upland grassland in KwaZulu-Natal are severely fragmented due to large scale farming of exotic trees. The aims of this thesis were to determine whether habitat fragmentation of these grasslands has a detrimental effect on plant-pollinator interactions and hence the reproductive output of the wildflower species occurring there and whether the magnitude of this effect can be predicted by breeding and pollination system characteristics. The 24 wildflower species included in this study appear to support a rich and diverse pollinator community, including long-tongued solitary bees, long-tongued flies, hawkmoths and sunbirds. Two thirds of the wildflower species appear relatively specialised in terms of pollination, with six species entirely dependent either on a single species or a specific functional type of pollinator for pollination. The majority of wildflower species (90%) were found to be incapable of autonomous self-pollination and thus dependent on pollinators for fruit and seed set. At least six species are obligately xenogamous. Little evidence was found for pollen limitation in undisturbed moist upland grassland, suggesting that these grasslands are characterised by high levels of pollinator activity. Greater levels of pollen limitation of reproductive output in habitat fragments was evident in two species, suggesting that depressed reproductive output in habitat fragments may be the consequence of a decrease in the quantity and/or quality of pollinator services. Significant detrimental effects of habitat fragmentation on reproductive output were evident in two thirds of the wildflower species, with 94% of the species exhibiting overall declines in seed set per flower from the continuous grassland site to the habitat fragments. The median decline in seed set per flower for the wildflower species was found to be 33.0%. Significantly more species experienced overall declines in reproductive output than would have been expected by chance alone. Only specificity of the pollination interaction in terms of number of pollinator taxa involved was found to be significantly related to percentage change in seed set from continuous to fragment habitats. This effect was diminished when other factors were included in a multiple regression. Results support Bond's (1994) hypothesis that degree of specificity in pollination systems is important in determining extinction risk of a given plant species. Declines in reproductive output of a range of wildflower species in grassland habitat fragments may affect the local persistence of these populations, particularly if recruitment is seed-limited. / Thesis (M.Sc.)-University of Natal, Pietermaritzburg, 2002.

The influence of herbivory, grass competition and grazing system on emergence, growth and survival of Acacia seedlings in burnt and unburnt areas was investigated in their first growing season, from September 1997 to May 1998. The study was aimed at determining possible reasons for the increase in woody plant density in semi-arid savannas, specifically the effect of excluding small-mouthed herbivores from domestic and livestock systems, and switching from continuous grazing to rotational grazing systems. The effect of fire intensity and maximum fire temperature on seed viability, germination and mortality, specifically with regard to back versus head fires and seed size , were investigated. Variability among species in the number of seedlings emerging generally resembled differences in viability. Emergence of Acacia karroo and Acacia tortilis was poorer than that of Acacia nilotica under grass competition. Emergence was lower in burnt and open areas, the latter depending on Acacia species. Herbivory did not affect seedling emergence, in both domestic and wildlife systems. Seedling survival and growth was adversely affected by small-mouthed herbivores in both domestic and livestock systems, whilst large-mouthed herbivores exerted no direct effects on woody seedlings, except to a small degree by trampling. Seedlings showed better growth and survival under low grass competition, which also resulted in greater leaf-to-height ratios , indicating that grass interference with irradiance affects woody seedlings. This effect appeared to be greater for cattle and rotational grazing, and for burning, in domestic and wildlife systems respectively. Woody seedling establishment was also better in burnt areas. Whilst seedling growth was better under rotational than continuous grazing, survival was not significantly different. Grass competition influenced seedling establishment to a greater extent than herbivory, whilst burning made a greater impact than grass competition. Seed mortality in response to maximum fire temperature was inversely related to seed size, and trends in seed germination and mortality, although very variable, appeared to be influenced by threshold fire intensities and maximum temperatures. Back fires had hotter maximum temperatures and fire intensities at ground level than head fires, which result in greater seed mortality following fire. Based on the current study it is likely that the removal of grass competition, burning, and a change from continuous to rotational grazing systems, and small-mouthed to large-mouthed herbivores, will result in an increase in woody seedling establishment. / Thesis (M.Sc.)-University of Natal, Pietermaritzburg, 2001.

Giraffe (Giraffa camelopardalis) are extralimital (non-native) to the Eastern Cape Province, South Africa yet they have and continue to be introduced to the region. Financial gain has arguably been the driving force behind these introductions as foreign tourists associate giraffe with Africa and the African wildlife experience. This raises a number of ethical, ecological and philosophical questions especially when it is considered that the impact of these browsers on the indigenous vegetation has remained largely unquantified. In this study I assessed the diet and potential impact of three populations of giraffe in the Eastern Cape Province between January 2002 and October 2003. The diet was assessed by both direct observations and faecal analysis. There was no significant difference (P > 0.05) between the results of the two methods of analysis, although direct observations appeared to be a superior method for assessing the diet of giraffe. The diet of giraffe in the Eastern Cape Province was similar to that within their native range with a deciduous species from the genus Acacia (Acacia karroo) being the most important component of the diet. However, giraffe in the Eastern Cape Province consumed more evergreen plant species than those within their native range. The relative lack of deciduous species in the Eastern Cape Province provides a likely explanation for such a result. Seasonal variation in the consumption of the most important species in the diet was evident and this was attributed to the deciduous nature of A. karroo and the seasonal growth of new shoots which were more palatable. The vegetation of the areas most commonly utilised by giraffe at each site was sampled using the point-centred-quarter method and the results related to the frequency of each species in the diet to calculate preference indices. Giraffe preference was strongest for A. karroo and this was attributed to the highly favourable chemical composition of the species. The browse utilisation of giraffe at each site was determined using the twig-length method and intake rates for the three most important species in the diet calculated using a pre-existing regression equation. Male giraffe fed at a higher rate than females. This was probably due to males adopting a “time-minimising” strategy to their feeding in order to allow more time for reproductive pursuits. Giraffe browse utilisation was highest where giraffe density was highest. However, several species were more heavily browsed than others even when giraffe density was low, suggesting that giraffe are capable of negatively affecting the indigenous flora of the province. I conclude that giraffe numbers should be reduced relative to property size in the Eastern Cape Province and that research into the impact of not only giraffe but the combined effects of giraffe and other extralimital herbivores on the indigenous flora and fauna be continued.

Thesis (MSc)--Stellenbosch University, 2011. / ENGLISH ABSTRACT: With growing concerns about current environmental issues, such as climate change, that affect ecosystems around the world, understanding ecosystem function is becoming increasingly important. In this study, I investigate the plant – seed disperser mutualism between an endemic scatter-hoarding mouse Acomys subspinosus and its Proteaceae plant counterpart Leucadendron sessile in the biodiversity hotspot of the Cape Floristic Region, South Africa. The main objective of this thesis is to investigate the seed selection and caching behaviour of A. subspinosus. First, I determined the seed selection strategy for dispersal and burial by A. subspinosus. Acomys subspinosus may exert stabilizing selection pressure onto L. sessile seeds by dispersing and burying medium seeds with medium hull-thicknesses. Small seeds were eaten in situ and large seeds left at depots. I concluded that the buried L. sessile seeds may have a competitive advantage when it comes to seedling establishment in a post-fire environment, since seeds dispersed by rodents in the fynbos, such as L. sessile, are much larger in size and therefore have more stored nutrients and rapid growth capabilities than seeds dispersed by other vectors. Secondly, I documented rodent dispersal behaviour over a full years’ time. Acomys subspinosus dispersal behaviour changed significantly over the seasons, which I attributed to a change in food availability as the year progressed. Acomys subspinosus buried seeds in autumn after mass seed drop but began to recover caches and consume seeds during winter and spring. The rodent switched to an insectivorous diet in spring. I propose that the A. subspinosus – L. sessile relationship is mutualistic during the year, but the relationship does shift in the favour of the rodent during winter and spring. Lastly, I address the scatter-hoarding behaviour of A. subspinosus and cache recovery ability of its assumed closest food competitor Rhabdomys pumilio. I found that cache size has a profound influence on pilferage rates of L. sessile seeds. Acomys subspinosus scatter-hoarded the majority of seeds singly in the field and R. pumilio had difficulties finding those single-cached seeds in dry substrate under controlled experimental conditions, serving as evidence that scatter-hoarding is an effective method of pilferage mitigation by A. subspinosus during the dry summer months. Relatively little was known about this plant – disperser mutualism and how it functions before this thesis were conducted. I have provided insights into the influence of rodent disperser behaviour on seed morphology development, seed fate and seed persistence in the field and suspect that the plant –disperser relationship may have a larger influence on ecosystem dynamics than previously anticipated. Further research on this system is of importance, especially with today’s emerging environmental instability and human interference that threaten the robustness of highly interconnected ecosystems like the fynbos. / AFRIKAANSE OPSOMMING: Met die huidige omgewingskwessies, soos die klimaatsverandering, wat ekosisteme wêreldwyd affekteer, word die begrip van ekosisteemfunksionering toenemend belangrik. In hierdie studie ondersoek ek die dier – saadverspreidingsmutualisme tussen die endemiese verstrooiings-storing muis Acomys subspinosus en sy Proteaceae plant eweknie Leucadendron sessile in die biodiversiteit 'hotspot’ van die Kaapse Floristiese Ryk, Suid-Afrika. Die hoof doelwit van die tesis is om die saadseleksie en storingsgedrag van A. subspinosus te ondersoek. Eerstens het ek die saadseleksie strategie vir die verspreiding en begrawing deur A. subspinosus bepaal. Acomys subspinosus het direksionele druk uitgeoefen op L. sessile sade deur mediumgrootte sade met medium saadhuiddiktes te versprei en te begrawe. Klein sade was in situ geëet en groot sade was gelaat by afgesette plekke. Ek het die gevolgtrekking gemaak dat die L. sessile sade wat begrawe is ‘n kompeterende voordeel mag hê wanneer dit kom by die vestiging van saailinge in ‘n afgebrande omgewing, aangesien sade wat in die fynbos deur knaagdiere versprei word, soos L. sessile, baie groter is en dus meer gestoorde voedingstowwe en spoedige groeivermoëns het, as sade wat deur ander vektore versprei word. Tweedens het ek die knaagdier verspreidingsgedrag oor die tydperk van ‘n jaar gedokumenteer. Acomys subspinosus se verspreidingsgedrag het beduidend verander deur die verloop van die jaar, wat ek toegeskryf het aan die verandering in voedselbesbikbaarheid soos wat die jaar gevorder het. Acomys subspinosus het sade begrawe in die herfs na grootskaalse vrylating en val van die sade, maar het gestoorde sade begin terug kry en sade begin eet gedurende die winter en lente. Die knaagdier het na ‘n insekvretende dieët omgeskakel in die lente. Ek stel voor dat die A. Supspinosus – L. sessile verhouding nie die hele jaar mutualisties is nie, maar eerder antagonisties, in die knaagdier se guns, gedurende die winter en lente. In die laaste hoofstuk spreek ek die verstrooiings-storingsgedrag van A. subspinosus en storingsverkrygingvermoeë van sy naaste voedselmededinger en deponeringsdief Rhabdomys pumilio aan. Ek het gevind dat die storingsgrootte ‘n beduidende invloed het op die koers van diefstal van L. sessile sade. Acomys subspinosus het die meerderheid van die sade gestoor in die veld en R. pumilio het die enkel-gestoorde sade in droeë substraat onder gekontroleerde eksperimentele kondisies moeiliker gevind. Dit is ondersteunende bewyse dat verstrooings-storingsgedrag ‘n effektiewe metode is om diefstal te verminder in die droë somer in die fynbos. Relatief min was bekend oor hierdie dier – saad verspreidingsmutualisme en hoe dit funksioneer voordat die studie uitgevoer was. Ek het insig verskaf oor die invloed van knaagdier verspreidingsgedrag op saadmorfologie ontwikkeling, die lot van sade en die tydperk wat dit begrawe is in die veld. Ek vermoed dat die mutualisme ‘n hoeksteenproses is in die fynbos en die invloed daarvan op ekosisteemdinamieka mag dalk groter wees as wat voorheen verwag was. Verdere navorsing oor hierdie sisteem is belangrik, veral met vandag se opkomende omgewingsonstabiliteit en menslike inmenging wat die robuustheid van hoogs verbonde-netwerk ekosisteme soos die fynbos bedreig.

A research report submitted to the Faculty of Science, University of the Witwatersrand, Johannesburg, in partial fulfilment of the requirements for the degree of Master of Science. Johannesburg, 2017. / The distribution of grazing herbivores is influenced by several factors, including spatial and temporal availability of resources. The Telperion and Ezemvelo Nature Reserves (TENR), located on the border between Gauteng and Mpumalanga Provinces, experienced a declining hartebeest (Alcelaphus buselaphus caama) distribution over the last years. Simultaneously, an increase in the density of plains zebra (Equus quagga) occurred in the reserves. In this study, I investigated the changes in the distribution of four herbivore species, namely red hartebeest, plains zebra, blue wildebeest and black wildebeest in TENR, and the possible influence that biotic and abiotic factors had on the distribution of hartebeest between 2010 and 2016. Data of aerial surveys conducted in the wet season of each year were used to assess the variation in herbivores distribution along the period. Furthermore, the influence of variables known to affect herbivore distribution (e.g. distance to water, slope gradient, spatial distribution of other herbivore species and vegetation greenness) was tested on hartebeest distribution in TENR. The results indicate that blue wildebeest (Connochaetes taurinus) had the highest distribution variation, while plains zebra presented the most spread distribution among the four species, and black wildebeest (Connochaetes gnou) concentrated their distribution almost entirely at the Ezemvelo nature reserve. Hartebeest avoided areas with the highest concentration of plains zebra, but did not show the same avoidance for areas with prevalence of wildebeest species. Further, high quality forage resources influenced the hartebeest distribution in TENR. Thus, competition for forage resources with bulk feeders may have affected the population and distribution of hartebeest between 2010-2016. / LG2018

Corals, the foundation species of tropical reefs, are in rapid global decline as a result of anthropogenic disturbance. On many reefs, losses of coral have coincided with the over-harvesting of reef herbivores, resulting in ecosystem phase-shifts from coral to macroalgal dominance. It is hypothesized that abundant macroalgae inhibit coral recovery and recruitment, thereby generating ecological feedback processes that reinforce phase-shifts to macroalgae and further diminish reef function. Notwithstanding, the extent to which macroalgae directly outcompete coral, the mechanisms involved, and the species-specificity of algal-coral competition remains debated. Moreover the capacity for herbivores to prevent vs. reverse ecosystem phase-shifts to macroalgae and the roles of herbivore diversity in such phenomena remain poorly understood. Here I demonstrate with a series of field experiments in the tropical Pacific and Caribbean Sea that multiple macroalgae common to degraded reefs directly outcompete coral using chemical warfare, that these interactions are mediated by hydrophobic secondary metabolites transferred from algal to coral surfaces by direct contact, and that the outcomes of these allelopathic interactions are highly species-specific. Using field observations and experiments in the tropical Pacific, I also demonstrate that the process of herbivory attenuates the competitive effects of allelopathic algae on corals by controlling succession of algal communities, and that the herbivore species responsible for macroalgal removal possess complementary tolerances to the diversity of chemical defenses deployed among algae, creating an essential role for herbivore diversity in reversing ecosystem phase-shifts to macroalgae. Lastly, I demonstrate with field experiments in the tropical Pacific that algal-coral competition simultaneously induces allelochemicals and suppresses anti-herbivore deterrents in some algae, likely due to trade-offs in the productions of defense metabolites with differing ecological functions. Together, these studies provide strong evidence that chemically mediated competitive and consumer-prey interactions play principal roles in coral reef degradation and recovery, and should provide resource managers with vital information needed for effective management of these ecologically and economically important but threatened ecosystems.

Degradation of forest environments in South Africa have greatly affected both the plants and animals that interact with these plants. It is therefore important to understand plant-animal interactions in forest ecosystems in an effort to develop strategies for forest conservation. Animal dispersal can assist in the restoration of forest plant diversity and consequently animal diversity. Wahlberg’s epauletted fruit bat (Epomophorus wahlbergi) occurs in the eastern part of South Africa and is known to feed on fruit, nectar, pollen and flowers. It is an important disperser of various indigenous and alien invasive forest fruit species. Little is known about the benefits these bats get from feeding on fruit. Furthermore, the buccal cavity and lingual adaptations of fruit bats in Southern Africa are unknown. This study investigated morphological and physiological adaptations for frugivory in E. wahlbergi. In particular, the palatal and lingual morphological structures of E. wahlbergi were examined by light microscopy and scanning electron microscopy. The results of this study showed that the morphological structures of tongue papillae in E. wahlbergi were similar to that of other fruit and nectar feeding bats. The structural arrangement of tongue papillae in E. wahlbergi probably allows food to move from the anterior part of the tongue and collect at the median line of the posterior part of the tongue, and thereafter move down to the pharynx. These bats had an elongated tongue, wide, flattened molar teeth, and a hard papillae structure observed on the upper palate at the posterior end of the buccal cavity. This papillae structure has not been described previously; and it appears that it is with this structure, together with the palatal ridges, and teeth that E. wahlbergi crushes fleshy fruit to extract the juices when feeding. Consequently palatal and lingual structures of E. wahlbergi showed morphological adaptations for efficiently feeding on fleshy fruit and nectar. South African indigenous and exotic fleshy fruits available to E. wahlbergi are found to be generally hexose dominant. These bats are suggested to have high glucose intake irrespective of sugar concentration to power their high energy demands due to flight. High glucose intake could result in increased blood plasma glucose levels which are detrimental to mammals of small body size. This study investigated the diel variations in blood plasma glucose concentrations of E. wahlbergi. Epomophorus wahlbergi’s blood plasma glucose concentration was lower (5.24 ± 0.38 mm ol/l) at 18h00 before feeding and increased during/after feeding (8.19 ± 1.24 mm ol/l), however bats appeared to regulate it within limits. Their range in concentrations was higher than the normal mammalian blood plasma glucose concentrations range. Consequently these bats appear to regulate their blood plasma glucose concentration although at a range higher than normal mammalian levels and thus reduce the negative consequences associated with hyperglycemia. The data obtained provide a baseline for comparison with free-ranging E. wahlbergi. Proteins are important in a diet to provide the required nitrogen and amino acids necessary for maintaining body tissues. Fruits, however, appear to have energy-rich but protein-limited foods. Frugivores that feed exclusively on fruit may therefore have difficulties in maintaining their protein requirements since fruits are generally high in sugar content but low in protein content. The importance of protein in the diet of E. wahlbergi was determined by measuring diet intake at varying levels of protein in the laboratory. Epomophorus wahlbergi were offered equicaloric 15 % glucose solutions with varying protein concentrations (2.58, 5.68, 7.23 g soy protein/kg H2O) and a solution with no protein. This was repeated using 15 % sucrose instead of glucose solutions. Epomophorus wahlbergi’s volumetric intake of the respective glucose and sucrose solutions varied among individual bats, with total volumetric intake highest for the solution with no protein (control) and lowest at 2.58 g/kg soy protein concentration solution for glucose and 5.68 and 7.23 g/kg soy protein concentration for sucrose. These bats appeared to prefer sugar solutions without or low protein, and their daily protein intake was relatively low. This suggests they have low-protein requirements, and this relates to their characteristic low-protein fruit available in the wild. Feeding and digestive efficiency has been widely studied in frugivorous and/or nectarivorous birds but relatively few studies have been done on bats, particularly African bats. Feeding on a liquid nectar diet and fruit juices could cause physiological challenges for nectarivores and/or frugivores as they have to balance water and energy intake from this liquid diet. Mammalian kidneys have to eliminate salts and nitrogenous wastes, conserve water during water restriction as well as excrete it when ingested in excess. Morphological renal characters are known to be reliable indicators of urinary concentrating abilities in mammals. Short digesta residence time is a digestive trait that is known to be associated with frugivory and this may be a problem since digestive efficiency is a function of the length of digesta retention time. Histological sections of E. wahlbergi kidneys and small intestines were examined under microscopy to determine water regulation and specialization for sugar absorption respectively. Cortex and medulla length measurements were taken to calculate the medulla: cortex ratio (M/C) and the percent medullary thickness (PMT). The observed M/C ratio and PMT of E. wahlbergi is typical of a mesic species. Epomophorus wahlbergi feeds on a watery diet and does not need to concentrate urine; therefore the medulla and cortex were more or less the same size. Analysis of the histological sections showed that the kidney structure of these bats allows efficient water regulation in a mesic environment. The microvilli of E. wahlbergi intestine were relatively long with a large surface area thus serves the purpose for efficient digestion and absorption of sugars in these bats. Future studies need to be done to determine the mechanisms by which these bats regulate their blood plasma glucose levels, and also determine intake on a selection of fruit in order to broaden research to adaptation of these bats on low-protein diets. This study contributes to a greater understanding of the physiological and morphological mechanisms that may affect fruit intake and consequently dispersal. It will contribute to a greater understanding of plant-animal interactions in southern Africa. / Thesis (M.Sc.)-University of KwaZulu-Natal, Piertermaritzburg, 2012.

This study presents the results of a study on introduced giraffe diet and feeding effects within the Eastern Cape Province of South Africa. Five sites were selected over a moisture gradient from the mesic east to the arid west of the Eastern Cape to describe giraffe diet and provide baseline data for feeding effect experiments. At each site faecal samples were collected seasonally and diet composition determined by microhistological analysis. Plant species availability was measured using the point intersect method in order to compare plant availability to plant consumption to determine diet preferences. Giraffe diet across the sites showed a unimodal response to rainfall by utilising fewer species in drier areas where plant availability is low and increasing the number of species consumed as the rainfall increases. In higher rainfall areas, giraffe reduced the number of plant species fed upon because of the high availability of important species that reduces the need for a varied diet. Across all sites, plant availability was dominated by woody plant species (more than 90 percent at each site). A total of 57 plants species were recorded as being eaten across the five sites. Several PDI (Principle Diet Items) plant species (Acacia karroo, Schotia afra, Pappea capensis and Euclea undulata) were eaten across sites. Over all the sites, significantly preferred species were Asparagus striatus, Schotia latifolia, Asparagus suaveolens, Commelina benghalensis, Viscum rotundifolium, and Acacia cyclops. Acacia karroo, Schotia afra, Pappea capensis, Rhus crenata, A. tetracantha, and Grewia robusta were utilised in proportion to their availability. Euclea undulata, Rhus longispina and Putterlickia pyracantha were avoided. Feeding effects were tested by erecting exclosures around trees, covering one half of each tree. Ten trees were selected for these manipulative experiments which ran from September 2003 to February 2005. The number of branches, number of leaves, branch orders, branch diameter, branch length, leaf length and leaf width were measured for ten samples per side (enclosed vs. exposed) and analysed using a Wilcoxon matched pairs test and tested for effect size using Cohen’s d. Although no significant differences were detected between the sides measured, the effect size indicated differences in all measurements between sides ranging from low to moderate. Leaf length was greater on the outside of exclosures and showed a moderate difference between the sides in terms of the d-value with the p-value (0.059) tending towards significance. Although leaf width was higher on the inside of exclosures, it only showed a moderate difference for d with no statistical significance. Leaf area was higher on the inside of exclosures and tended to significance (p = 0.059) and similarly the d value indicated moderate differences between the sides in terms of effect size. Similarly branch length was higher on the inside of exclosures although not significantly, the p-value (0.07) approached significance. The effect size for branch length also showed a moderate difference between the sides. The number of branches and the number of leaves showed no statistical differences between the sides although the p-value (0.059) for the number of branches, as well as number of leaves (p = 0.07) approached significance. For both these variables, effect size showed a moderate difference. Giraffe showed diversity in their diet selection across sites but mainly fed upon available species. Some preferred species were however less available and this preference could possibly result in increased pressure on the selected plant species Exclosure experiments showed no significant differences in leaf and shoot characteristics although all measurements showed differences in terms of effect size meriting further investigation. It is concluded that giraffe impact on sites may be reduced if populations are properly managed however, high densities of giraffe could lead to similar vegetation impacts as observed in other areas where giraffe have been introduced and potentially change vegetation structure and plant community composition.

The nyala on Buffelsdrift Private Game Reserve were found to be mainly browsers and made little use of grass during the study period. Browse contributed 87.8 percent to the annual diet, while grass contributed 12.2 percent. Only six species formed > 5 percent of the annual diet. Of these six only three species formed > 10 percent. They are the shrubs Acacia karroo, Pappea capensis and Grewia robusta. These were followed by the shrub Portulacaria afra, the dwarf shrub Zygophyllum lichtensteinianum and the grass Cynodon dactylon all contributing between 5 - 10 percent. There were definite seasonal dietary shifts with preference to different plant species. The shift takes place between the wet seasons (winter and spring) and the dry season (summer). Herbaceous species such as Cynodon dactylon and Erharta calycina were favoured in the winter/wet season. Shrubs such as P. capensis, P. afra and dwarf shrubs such as Limeum aethiopicum and Z. lichtensteinianum were favoured during the spring/wet and summer/dry seasons. Nyala habitat use in BPGR differed significantly from expected use when taking the relative area of habitats into consideration. However nyala did not display signififcantly different habitat preferences between seasons. Male and female nyala displayed differences in foraging behaviour at the plant form, plant species and habitat scale. Males preferred a greater proportion of shrubs and females a greater proportion of dwarf shrubs. Males preferred the shrubs Pappea capensis and Portulacaria afra the most and females preferred the dwarf shrubs Limeum aethiopicum and Zygophyllum lichtensteinianum the most. Plant species site availability varied significantly between the sexes, indicating that males and females selected different feeding paths. The difference in their feeding strategy appears to be related to the selection of plants that best satisfy the nutrient requirements of each sex. These findings are in agreement with the results from earlier studies on nyala sexual dimorphism. At the habitat scale females selected more for open dwarf shrubland and males more for densely vegetated habitats. This appears to be as a result of males attempting to maximize their foraging opportunities by selecting areas that offer food in greater quantities. Nyala in BPGR appear to co-exist with kudu and eland through spatial resource partitioning. This is suggested to be due to the influence of body size-gut size relationships on forage selectivity.

The giraffe at Nyaru were found to be browsers who made little use of graze during the study period. They utilized a diversity of 20 browse species. Two species, Acacia karroo and Acacia cyclops, formed the bulk of the giraffe diet throughout the year. A definite seasonal dietary shift was evident. A. karroo was favoured in summer and autumn and formed the main food species in spring, summer and autumn. A. cyclops was favoured throughout the study, but its contribution to the diet increased during winter when less A. karroo was consumed. This seasonal shift is related to the deciduous nature of Acacia karroo. Although a seasonal shift in species contribution to giraffe diet has been observed in many other giraffe feeding studies, no studies on giraffe feeding have been done in the Mosaic Thicket of the southern Cape. The ecological browsing capacity for giraffe in thicket was estimated to be between 0.020 BU/ha and 0.095 BU/ha. The browsing capacity for giraffe at Nyaru, based on the available phytomass 2–5 m above the ground, was estimated using those species that formed the bulk of the giraffe diet, and amounted to 0.063 BU/ha. A maximum of three giraffe could thus be stocked on the 157 ha of suitable giraffe habitat on Nyaru. This stocking rate recommendation lies within the range commonly recommended by local consultants for giraffe introductions into the southern Cape. Their recommendations are, however, not based on quantitative assessments such as performed in this study. The recommendation of this study should not be applied as a fixed ecological capacity for giraffe in thicket, but should be seen as a starting point in the adaptive management cycle. Ongoing monitoring of parameters, such as herbaceous composition and phytomass; as well as the condition of key browse species, is strongly advocated. A. karroo was browsed significantly more and carried significantly fewer pods per tree at a heavily used site compared to a lightly used site. Heavy browsing thus appears to affect the reproductive success of A. karroo significantly. Fewer pods are likely to lead to lower regeneration and thus reduced density of A. karroo. Whether A. karroo will maintain its dominance within the thicket vii community in the long run will be related to how individual plants survive and reproduce and if some can escape from herbivory. Fewer G. occidentalis were clumped with other species at the heavily used site compared to the lightly used site. This could possibly be attributed to the fact that intense browsing pressure at the heavily used site caused protective clumps to be eaten away, thus exposing G. occidentalis to higher ungulate browsing. Previous studies have found that nurse shrubs protect G. occidentalis against ungulate browsing. G. occidentalis was browsed significantly more at the heavily used site compared to the lightly used site. There was a general trend of fewer fruits at the heavily used site compared to the lightly used site, while fruits were absent on G. occidentalis growing alone at the heavily used site. This suggests a negative effect of heavy browsing on plant reproductive success and emphasizes the importance of nurse plants for the successful recruitment and hence long term prevalence of G. occidentalis in Mosaic Thicket.