The changes in plant quality and availability in space and time present a substantial problem to mammalian herbivores. As a result, these herbivores need to alter their foraging behaviour to maximize their energy gain at both small (plant level) and large (landscape level) scales. A megaherbivore, the African elephant (Loxodonta africana), has been shown to be selective in its foraging choices at both of these scales. Furthermore, the ratio of palatability:defences (e.g. fibre and total polyphenols) has been highlighted as an important determinant of habitat selection in elephants. The elephants in Ithala Game Reserve (IGR) frequently leave IGR during the wet season and forage outside the reserve. However, they predominantly feed on the low-nutrient granite soils of the reserve and return to a high-nutrient area with dolerite soils during the dry season. In an attempt to understand these seasonal movements, I focused on how the small-scale foraging decisions of the elephants lead to large-scale seasonal movements in IGR, KwaZulu-Natal. Plant availability was determined seasonally for seven target species across four areas in the reserve. Crude protein, fibre, energy and total polyphenols as well as the ratios of palatability:digestion-reducing substances were analysed in the wet and dry seasons. All factors and their interactions were significant in a MANOVA. Consequently, I employed a dimension-reducing Principal Components Analysis (PCA) to better understand the factors of greatest importance. The PCA highlighted four of the six most important factors to be the ratios of palatability:digestion-reducing substances. The other two important variables were total polyphenols (negative effect) and crude protein (positive effect). At small spatial scales, I found that the elephants were selective in their decisions, especially during the dry season. For example, the increased inclusion of the principal tree species Acacia nilotica from 2.9% in the wet season to 39.3% during the dry season appears to be a result of a decline in total polyphenols and fibre during the dry season. At large spatial scales, the elephants moved back into IGR from the low-nutrient granite soils in the east in response to an increase in forage quality in the west as the quality declined in the east at the same time. However, it is unclear as to why the elephants are leaving the reserve during the wet season. Some possible explanations for this are discussed.Key-words: acid detergent fibre, crude protein, Loxodonta africana, neutral detergent fibre, total polyphenols, plant part quality. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2013.

Diet selection and its variation due to difference in forages nutritive value are central processes in plant-herbivore interactions. An incorporating of these parameters may allow predicting diet selection patterns which are a basic underpin for sustainable vegetation management and profitable animal production. To achieve these issues many studies were done two of which were run to study the browse-browser interactions in a sub-humid subtropical savannah, a herd of goats was used as a model browser in natural pasture. Other studies evaluated the nutritive value of five plant species which were the most selected by goats. The objectives of this study were to (i) determine the seasonal patterns of diet selection of goats on woody species and their nutritive value and (ii) to predict the diet selection. The secondary objectivrs were to (i) investigate how plant characteristics such as plant morphology (spinescence, and broad vs. fine leaves), phenology (evergreen vs. deciduous species) and plant chemistry (nitrogen, neutral detergent fibre, acid detergent fibre, acid detergent lignin, condensed tannin, cellulose and hemicellulose) are relate to the diet selection patterns and to (ii) evaluate the nutritive value of the most plant species selected by goats during the different seasons. Two experiments were conducted at Owen Sitole College of Agriculture, KwaZulu-Natal, to determine diet selection of goats grazing/browsing in Zululand Thornveld. The first experiment was done in the early wet season (November 2007) and late wet season (March/April 2008) while the second experiment was conducted in the dry season (June/July 2008), early wet season (November/December 2008) and late wet season (February/March 2009). The relative abundance of browse species in the veld was determined and used to estimate the selection index. In the first part, diet selection was estimated using (i) direct observation of animals (account bouts) by observing every one minute two goats while they were foraging with others for 3 hours in the morning and 1.5 hours in the afternoon on alternate days for a total of six days and (ii) an indirect plant-based method by observing at regular intervals 40 marked branches on ten plants of each browse species every two days for a total of six days. The browse species observed were: Gymnosporia senegalensis, Acacia nilotica, Acacia karroo (Acacia natalitia), Scutia myrtina and Dichrostachys cinerea. In both seasons, the three most selected species according to the plant-based observations were S. myrtina, A. karroo and D. cinerea. Consistently, A. nilotica experienced moderate defoliation and G. senegalensis the least. However, on the basis of the selection index, the species followed the order: A. nilotica > D. cinerea > A. karroo > S. myrtina > G. senegalensis in the early wet season and A. nilotica > S. myrtina > D. cinerea > A. karroo > G. senegalensis in the late wet season. Both methods did not rank species in the same order. The selection index was negatively (P<0.05) correlated to neutral detergent fibre (NDF), acid detergent fibre (ADF) and acid detergent lignin (ADL). In the second part, observations were taken on feeding time of two goats, randomly selected per day for 7-8 days. Observations were made during active foraging periods for 2 hrs in the morning and 1.5 hrs in the afternoon. The duration of each feeding bout and the species of woody plant from which bites were cropped at each feeding station were recorded. Season and plant species affected diet selection and preference. The five main species selected (utilized) by goats in decreasing order were S. myrtina, A. nilotica, D. cinerea, Acacia natalitia and Chromolaena odorata. Scutia myrtina was the most preferred (highest utilisation relative to availability) in the dry season while D. cinerea comprised the greatest proportion in the early and late wet seasons. Scutia myrtina was most preferred in the dry and early wet seasons while A. nilotica was most preferred in the late wet season. Spinescent species were generally selected more than non-spinescent species in all seasons, while fine-leaf and deciduous species were selected more than broad-leaf and evergreen ones in the wet seasons. However, preference for broad-leaf and evergreen species increased in the early wet season. Although plant chemistry varied across seasons, it did not explain the preference of goats for various plant species in this study. Instead, effects of chemistry were species- specific. The nutritive value of the five main species selected by goats was evaluated by chemical composition, in vitro gas production, in vitro degradability and in sacco degradability. Two experiments were conducted for the in vitro studies, the first experiment was to test the effect of season and species, while the second experiment tested the biological effect of tannins using polyethylene glycol 4000 (PEG). The PEG treatment was applied to samples of the early and late wet seasons only. The parameters were maximum gas production (GP), degradation rate (C), lag time (lt), maximum rate of GP at the point of inflection (μ), half time to the maximum gas volume(T1/2), gas produced from fermentation of soluble and slowly degradable fractions (A and B, respectively), their degradation rates (c1 and c2, respectively), apparent (ApDeg) and true degradability (TrDeg), microbial yield (MY), partitioning factor (PF), degradation efficiency factor (DEF) and short chain fatty acids (SCFA). There were wide variations among seasons and species in crude protein (CP), NDF, ADF and condensed tannins. Season and species affected kinetics of gas production, GP, TrDeg, MY, PF and SCFA. Chromolaena odorata had the highest CP (185.8-226.4 g kgˉ¹), GP (87.3-104.1 ml gˉ¹ DM), gas produce from soluble (47.6-50.9 ml gˉ¹ DM) and insoluble (39.8-50.9 ml gˉ¹ DM) fractions during the three seasons compared to other browse species. The TrDeg ranged from 634 to 856 g kgˉ¹ DM. The total SCFA varied between seasons and among species. The addition of PEG decreased TrDeg and PF, and increased GP and total SCFA. The GP, its degradation rate (C) and gas from the soluble fraction were positively correlated to CP without and with PEG. Gas from the soluble fraction was negatively correlated to NDF, ADL and CT; and GP to CT without PEG. In the in sacco study, the parameters of dry matter (DM) and nitrogen (N) degradability were soluble (a) and slowly (b) degradable degradability, the degradation rate (c), potential degradable (PD), effective degradability (ED) and lag time (lt). Season affected PD of DM degradability and ED of both DM and CP. Browse species and its interaction with season affected all parameters except b fraction of CP and interaction did not affected PD of CP. Chromolaena odorata had the highest estimated parameters of degradation during the three seasons compared to other browse species. Based on PD and ED, the plant species followed this decreasing order: C. odorata, A. nilotica, A. natalitia, S. myrtina and D. cinerea. All parameters were positively correlated to CP except the a and b fractions of DM and CP, respectively, and lt of both DM and CP. All parameters were negatively correlated to NDF, ADF and ADL except the b fraction and rate of degradation. Condensed tannins (CT) were negatively correlated with all parameters except the b fraction and PD of DM and CP, and lt of DM degradation. Cellulose was negatively correlated with a of DM and lt of CP degradation. The broad objective of this study was to assess whether plant characteristics, in sacco degradability, in vitro gas production, digestibility or chemical composition (NDF, ADF, ADL, CT and CP) could predict the selection of browse species. The parameters were selection index, plant characteristics (spinescence, or leaves phenology), chemical composition, in vitro gas production, in vitro degradability and in sacco degradability parameters. Most of these parameters were poor for predicting selection by goats of browse species. Spinescence (spn), leaves phenology (phen), NDF, CP and CT accounted for 86% of the variation in selection index (y) = -5.91 - 0.01CT - 0.002NDF + 0.02CP + 6.18spn + 2.43phen; (R2=0.86; n=20; RMSE=0.406; P<0.001; for phen, 1 = evergreen, 0 = deciduous; for spn, 1 = spinescent, 0 = spineless). It was concluded that in vitro gas production, in vitro degradability and in sacco degradability were poor predictors of selection index of browse species used by goats. Spinescence, leaf phenology, condensed tannins, CP and NDF were predictors of diet selection index and suggested that these plant species have potential to be used as feed supplements. Chromolaena odorata has the highest potential as feed protein source in goats. The addition of PEG (tannin binding agent) emphasizes that the inhibitory effect of tannins on rumen microbes was great in tannin-rich feeds. / Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2012.

Diet selection and foraging efficiency are primary driving behavioural processes in mammalian herbivore plant interaction. They determine the quality of production and also impact on vegetation, necessitating managerial inputs that will sustain productivity and conserve or improve on its plant species composition. In view of a better understanding of browse/browser interaction in subtropical arid savannah, Nguni goats were adopted as our model browser considering their socioeconomic importance to local, regional and global economy. The primary objective of this research was to determine the seasonal and plant species trend of diet selection indexed by intake and foraging efficiency indexed by intake rate. The secondary objective was to investigate how plant morphological variables such as spinescence, shoot morphology such as long shoots as opposed to short shoots, and broad as against fine leaves, phenological variables such as evergreen versus deciduous species, and plant chemical components including: crude protein, neutral detergent fibre (NDF), acid detergent fibre (ADF), hemicellulose, cellulose, acid detergent lignin and condensed tannins relate and possibly explain the observed patterns. Six browse plants, Acacia nilotica, Acacia robusta, Ehretia rigida, Euclea crispa, Rhus lancea and Ziziphus mucronata were used in the first experiment to determine short term intake during the dry and rainy seasons. In the second experiment, 10 browse plants: Acacia nilotica, Acacia robusta, Combretum epiculatum, Dicrostachys cinerea, Ehretia rigida, Euclea crispa, Gymnosporia buxifolia, Rhus lancea and Ziziphus mucronata were used to determine short term intake rate in the dry, early rainy and late rainy seasons. Diet selection was significantly affected by species (F5,60= 31.87 ; p< 0.05) and the interaction between seasons and species (F,60= 2.52; p= 0.039). Intake was not significantly higher in the rainy relative to the dry season, 3.94 ± 0.42a gDM compared to 3.57 ± 0.42a gDM. Rhus lancea and A. robusta were the most preferred species in both seasons, while Z. mucronata was more preferred in the dry season than E. crispa but in the rainy season it was the opposite. Dichrostachys cinerea and A. nilotica were the least preferred in both seasons, but with improved intake of D. cinerea in the rainy season, whereas that of A. nilotica declined significantly. Intake of plant species with long shoots, which were also broad leafed, was significantly higher than that of old short shoot species (t= -3.99; df= 65.80; p= 0.0002). Similarly, intake of evergreen species was significantly higher than that of deciduous species (t= 5.22; df= 34.02; p< 0.0001). Cellulose in the rainy season positively correlated with intake (r= 0.85; n= 6; p= 0.031), bite number cropped (r= 0.98; n= 6; p= 0.001) and browsing time (r= 0.93; n= 6; p= 0.007). In the dry season there was no correlation of any plant chemical components with intake except cellulose that correlated with the intake variable of bite size (r= 0.87; n= 6; p= 0.023). The following regression equation was generated from the correlation to model intake: Intake= 0.68 + BS (3.07), (R2= 0.71; p< 0.05), where BS= bite size. Intake rate varied amongst species (p< 0.05), as well as seasons (p< 0.05). Instantaneous intake rate in the dry season were highest for A. caffra and G. buxifolia, followed by C. apiculatum, R. lancea, E. crispa, A. robusta and Z. mucronata. The bottom tier comprised of E. rigida, D. cinerea and A. nilotica. At the onset of the rainy season, A. robusta, A. caffra and R. lancea had the highest instantaneous intake rate, followed by C. apiculatum, E. crispa and Z. mucronata. The lowest instantaneous intake rates were derived from D. cinerea, E. rigida, G. buxifolia and A. nilotica. At late rainy season, A. robusta, A. caffra, C. apiculatum, E. crispa and R. lancea yielded the highest intake rate, followed by D. cinerea, G. buxifolia and Z. mucronata in the middle tier, and the least were derived from E. rigida and A. nilotica. Acacia caffra, A. robusta, R. lancea, C. apiculatum and E. crispa featured amongst species with the higest intake rate in all three seasons, whereas G. buxifolia from which one of the highest intake rates was derived in the dry season declined to amongst those with the least in both early and late rainy seasons. Ehretia rigida, D. cinerae and A. nilotica were consistent in all three seasons as the least efficiently foraged Browses with long shoots were more efficiently foraged upon than those with short shoots (t= -3.2; df= 284.64; p< 0.002). Evergreen plant species yielded higher instantaneous intake rate relative to deciduous species (t= 2.02; df= 278.01; p< 0.044). Intake rate of spinescent relative to spineless species was not significantly different but more bites were cropped from spinescent relative to spineless species (t= 2.80; df= 256.77; p= 0.006). Bite sizes were smaller for spinescent relative to spineless species (t= -3.30; df= 227.4; p= 0.001). There were significant positive correlations of cellulose with intake rate (r= 0.7; n= 10; p= 0.023), and with condensed tannins (r= 0.7; n= 10; p= 0.022) in the dry season. Additionally, cellulose also had a positive correlation with bite size in the same season (r= 0.7; n= 10; p= 0.016). At early rainy season, there was no correlation of intake with any of the plant chemical components, whereas, NDF (r= 0.6; n= 10; p= 0.044), hemicellulose (r= 0.8; n= 10; p= 0.012) and condensed tannins (r= 0.7; n= 10; p= 0.032) all correlated positively with browsing time at early rainy season. The following regression equations were derived to model intake rate and some intake rate related variables: (1) intake rate= 0.433 + BSDM (0.081), R2= 0.68, p< 0.05 and BSDM= bite size dry matter; (2) Bite rate= 0.503 –BSDM (0.147), R2= 0.597, p< 0.05 and BSDM= bite size dry matter; (3) Bite size dry matter= 1.713 – BR (4.210), R2= 0.597 ,p< 0.05 and BR= bite rate, and (4) Intake dry matter= 3.804 + BSDM (0.865), R2= 0.607, p< 0.05 and BSDM= bite size dry matter. Diet selection and foraging efficiency were both influenced more by shoot and leaf morphology than by spinescence. Plant chemistry on the other hand had diverse and inconsistent effect on diet selection and foraging efficiency trends. The results therefore have important and far reaching implications for Nguni goat and goat production as a whole in the bush veld of Gauteng and related ecological zones. / Thesis (M.Sc.Agric.)-University of KwaZulu-Natal, Pietermaritzburg, 2012.