All observed bovid species showed relatively narrow dietary niches: they feed on few species of grasses only. Puku fed mainly on Panicum spp., Brachiaria sp., Sporobolus spp., Hyparrhenia/Andropogon-grasses and Eragrostris ssp. – but also on other grasses that remained not determined. Except for Sporobolus ssp. and Brachiaria sp., these grasses were main components in the diet of reedbuck, sable, hartebeest and buffalo in Kasanka NP: The same applied to impala and waterbuck in Kafue Region. Impala, waterbuck and sable showed a considerable amount of browse in their diet. This group of plants was considered as a whole during data analyses, not split up into genera or even species. This might affect the interpretation of competition between the considered species. Nevertheless, as puku consumed very low to zero amounts of browse, this fact does not concern the competition with puku and hence does not touch this study.
The amounts of these grasses differed in the diet of the bovid species. Above all, reedbuck, sable and hartebeest in Kasanka NP showed a very similar diet. According to the results of the JACOBS Index, these three species showed also similar preferences/avoidance patterns for of grass species. The same applied for the results of the PIANKA index: values between these bovid species indicated a high degree of overlap. This competition happened during the cool season (only for reedbuck and sable) and especially during hot dry season and was supported by the results of the PCA as well. It probably resulted from the food shortage in the dry season. Lack of food and mortality are highest as reported for grazers of the Serengeti (SINCLAIR 1979b). This matches with the observation in the dietary overlap between bovid species at Chikufwe in Kasanka NP.
Food plants of sitatunga in the cool dry season differed from those of other bovid species. Consequently the PIANKA index resulted in low values of overlap for the cool dry season, supported by the results of the PCA. In the hot dry season, overlap between sitatunga and other bovid species was slightly higher. Puku and sitatunga showed the lowest degree of overlap according to the PIANKA index and to PCA. The consumed food plants were more similar to those of other bovids and also puku. In the hot dry season, the samples of sitatunga came from Kasanka River and not from Fibwe plain in the hot dry season. But in contrast to puku at the same site, sitatunga showed a strong avoidance of Brachiaria sp.
In the cool dry season, the PIANKA index for puku-regional samples and other bovids was higher than this was the case for Kasanka River puku. This was especially true for sites near or in more wooded areas as Luwombwa, New Airstrip or Wasa (P-0, P-C and P-F). Here,
the diets of the puku were more similar to reedbuck and hartebeest indicating a possibly higher degree of competition.
The JACOBS index revealed inversed patterns of avoidance and preference comparing puku and the other bovids: Puku slightly preferred Brachiaria sp. and preferred Panicum ssp.
more than the others. Reedbuck, sable, hartebeest and, to a certain degree, buffalo preferred Hyparrhenia/Andropogon grasses and avoided Eragrostis ssp. to a much lesser extent than puku. Nevertheless, all bovids showed a slight preference for Panicum ssp. that probably maintains its good nutritive values throughout the dry season (JUNGIUS 1971). The status of competition for food between the Kasanka River puku and the other bovid species was relatively low through the two surveyed seasons. This low overlap was even decreasing with the progression of the dry season, when comparing the competition between puku and reedbuck and between puku and sable. This light degree of competition between the Kasanka River puku and the other bovid species in Kasanka NP was emphasised or even amplified by the results of the PCA. It has been reported from other protected areas that other antelopes come into the range of reedbuck especially after fires (JUNGIUS 1971). To avoid competition with other ungulate species, reedbuck feed on relatively unpalatable grasses (HUFFMAN 2011). While sable showed strong avoidance for Loudetia sp., this grass was less avoided by reedbuck but also hartebeest. Is it possible that especially sable are attracted by the freshly sprouting other grasses? Hyparrhenia is the characteristic grass of the fire climax vegetation in the miombo ecoregion (KINGDON 1982). It was less avoided or even slightly preferred by reedbuck, sable, hartebeest and buffalo in contrast to the puku, and sitatunga, which avoided it. Irregardless of the moving power, puku mainly occurred at different sites than reedbuck, hartebeest, sable (see chapter 2) and buffalo. This separation on a spatial scale is another step to avoid competition.
In Kafue Region, dietary overlap between puku and impala was assessed. Although other bovid species occurred in the study regions, no dung samples could be gathered. The results from the line transect sampling revealed a high spatial overlap between puku and impala in the cool dry season, during which the PIANKA index values were low between those two in both PPKR and Kafue NP. The scatterplots for the sub-regions showed impala and puku well separated from each other. This was different in the late rainy season where PIANKA index values increase between puku and impala. This degree of higher potential overlap was reflected by the results of the PCA as well. To evaluate the dietary overlap and potential competition between puku and impala, dung was taken from the middens that are used mainly by territorial males (STUART &STUART 2000). These territorial males are reported
to have fewer amounts of dicotyledons in their diet than females or bachelor males (VAN
ROOYEN & SKINNER 1989, SPONHEIMER et al. 2003b). If this is true for Kafue Region the actual dietary overlap between puku and impala might be even less marked. This could be one factor that explains the high numbers of these two antelope species in Kafue Region and their sympatric occurrence. Although waterbuck was not included in the statistical evaluation, the results of its diet suggest a medium overlap in the cool dry season based on the fact that waterbuck include considerable browse in its diet.
Bovid species are characterised by anatomical differences. They are assumed to reflect the proportions of C3 browse and C4 grass in the diet, but are proven to correlate with the degree of feeding selectivity also in regard to short-, medium- and tall grass grazing (CODRON
et al. 2008). It is the part of the plant or the specific growth stage of the plants that might be consumed by each bovid species (MURRAY & BROWN 1993) that leads to resource partitioning.
The coexistence of sables with zebra and buffalo in Kruger NP is enabled by their precise selection for green parts of the plants. Buffalo tolerate grass that is mostly brown (MACANDZA
et al. 2012). The methods used for this study did not allow a differentiation of the growth stage or the parts of the plants taken neither would they allowed a statement on the nutritional value of the food plants actually taken which all probably affect the choice of food plants.
The bovids influence the vegetation by their grazing activity. MURRAY & ILLUS (2000) reported on wildebeest (Connochaetes taurinus) and topi (Damaliscus lunatus), going for bite quantity and bite quality, respectively. They can modify the vegetation in such a way that it is less profitable for the competing species. Thus herbivores can preclude the utilisation of resources by other species (MURRAY & ILLUS 2000).
It has to be mentioned that although bovid species compete with each other, the ecosystem has further species of whom they might be influenced in regard to their feeding.
SINCLAIR (1979a) reports of invertebrates having a much higher impact on the grasslands than large mammals. By consequence, mammals could be affected by the fluctuations of phytophagous insects (SINCLAIR 1979a). From the small to the big member of the ecosystem:
The feeding activity of megaherbivores, especially of elephants (Loxodonta Africana), affects mediums sized mixed feeders (FRITZ et al. 2002). Populations of these mesomixed feeders, e.g. impala, can decline when megaherbivores are abundant (FRITZ et al. 2002). Mesograzers, thus also puku, are supposed to be rather unaffected by the presence of megaherbivores
(FRITZ et al. 2002). In Chobe NP, increasing number of elephants lead to increasing numbers of puku: elephants opened the woodland; more grassland was available (DIPOTSO & SKARPE
2006). Thus here puku were positively affected by the elephants.
Two bovid species may be competitive if they exhibit an important degree of similarity in resource use at least during periods of forage scarcity (BAGCHI et al. 2003). This was the case for the bovid species in Kasanka NP: reedbuck, sable, hartebeest and, to a lesser extent, buffalo. The Kasanka River puku showed little dietary overlap with the species mentioned above as well as with sitatunga especially in the dry season, which is the time of suggested food scarcity. The dietary overlap between puku and impala, which showed the same population densities in Kafue Region, was low in the cool dry season, where each species appears to occupy a segregated food niche. In the late rainy season, the time of food abundance, especially impala changed their diets in a way that these two antelopes showed an increased dietary overlap. Unfortunately apart from puku, dung collection was not possible for the rainy season Kasanka NP.
The bovid species and their status of competition depend on the availability of food (SINCLAIR 1979a, WHITE 1978). In return, plant growth is limited by soil moisture (JARMAN &
SINCLAIR 1979) which is mostly influenced by rainfall. But this rainfall, as stated by MCNAUGHTON (1979) for the Serengeti ecosystem, is extremely variable over space and time.
In consequence, competition varies throughout a year and also between years. Thus the results of this thesis give only a small insight into the dietary competition between bovid species in Zambia.