Diet assessment of herbivores provides insight into trophic relationships, the potential for competition, and the influences herbivores may have on an ecosystem (Bookhout 1996). Thus, the determination of their food requirements is imperative prior to the implementation of any management decisions, which must be based on reliable data (Bookhout 1996). Direct observations and fecal analysis are 2 commonly employed techniques for assessing the diet of wild herbivores (Van Aarde and Skinner 1975, Field and Ross 1976, McInnes et al. 1983, Landman and Kerley 2001, Parker et al. 2003). However, both methods are biased in the way in which they quantify herbivore diets (for review see: Dearden et al. 1975, McInnes et al. 1983, Norbury and Sanson 1992, Bookhout 1996).
Traditionally, the diet of megaherbivores (herbivores that exceed 1,000 kg; Owen-Smith 1992) was quantified using direct observations because megaherbivores are conspicuous and easy to observe compared to smaller, more cryptic ungulates. The giraffe (Giraffa camelopardalis) is no exception (Leuthold and Leuthold 1972, Van Aarde and Skinner 1975, Parker et al. 2003). They are large, browsing ruminants (Owen-Smith 1992) that favor open or broken savanna habitats (Skinner and Smithers 1990) and browse exclusively on woody vegetation (Owen-Smith 1992). Giraffes are highly selective compared to other megaherbivores due to their unique foraging technique, whereby they strip the leaves off the terminal shoots of trees with their long tongue (Cooper and Owen-Smith 1986, Owen-Smith 1992). Typically, trees from the genus Acacia are the mainstay of giraffe diet (Du Toit et al. 1990). However, most Acacia species are deciduous and lose their leaves during the dry or cold season (Parker et al. 2003). Consequently, giraffes include a greater proportion of evergreen or semi-deciduous vegetation in their diet during this period (Sauer 1983, Parker et al. 2003). Apart from males having higher forage intake rates than females (Ginnett and Demment 1997), no sex differences in diet composition have been reported in the literature.
We compared direct observations and fecal analysis to determine the diet of a megaherbivore on a regional scale using the giraffe as a model. We tested the ability of the 2 methods to detect the most important plant species and the seasonal changes in the diet. There is no a priori reason to suggest that one method will perform better than the other, and we predicted that it is most likely that the 2 methods together will give the most robust results.