Assessing Food Preferences of Ungulates by Acceptability Indices

Abstract

We assessed the preferences of greater kudus (Tragelaphus strepsiceros) for different woody plant species using acceptance or rejection frequencies, feeding durations, and forage ratios. We recorded the foraging behavior of hand-reared kudus in a 213-ha enclosure containing natural savanna vegetation in the Nylsvley Nature Reserve, South Africa. Plant-based acceptance was based on plants within neck reach of the animal during 1-hour observation sessions. Site-based acceptance was based on the presence of a species within readily visible range of the animal in 30-minute periods during all-day observations. Sitebased acceptances revealed a categorical distinction between favored and neglected plant species. Feeding durations were longer on high-acceptability species than on low-acceptability species. Plant-based acceptances showed a less clearcut pattern, due to statistical problems of interdependence between successive plants of the same species encountered. Forage ratios were subject to sampling errors in estimating the abundance and dietary proportions of less common plant species, especially where these were clumped in distribution. Rank orders of preference had little consistency, as species within the favored and neglected categories did not differ significantly in acceptability. Advantages of the site-based acceptability index include use of simple binomial statistics for confidence ranges and the elimination of time-consuming vegetation surveys to estimate abundance. J. WILDL. MANAGE. 51(2):372-378 To classify plant species in terms of their value as food resources for large herbivores some measure of relative preferences for them is required. The most widely used index is the forage ratio (FR) (Ivlev 1961, Petrides 1975). This is calculated by dividing the relative abundance of a food in the diet by its relative abundance in the environment. Foods that yield a ratio >1 are assumed to be positively selected (preferred or favored). Those that yield a ratio <1 are regarded as selected against (rejected or neglected). The FR has several shortcomings: it varies asymmetrically between 0 and infinity, confidence limits are not simply estimated, and conclusions about whether particular foods are positively or negatively selected depend upon the set of food types that the investigator deems to be available to the animal. Accordingly, various modifications of the simple FR have been proposed. The electivity index of Ivlev (1961) achieved a range of variation between limits of +1 and -1. The alpha index of Chesson (1978, 1983), and Vanderploeg and Scavia's (1979) E*, are uninfluenced by changes in the number of food types considered available, or the abundance of these foods. Lechowicz (1982) reviewed the numerical behavior of these and other selection indices. To circumvent difficulties in deciding what food types were available, Johnson (1980) suggested that the rank orders of usage and availability be compared. Our study assessed the relative preferences of kudus for different woody plant species in a South African savanna. We present the results obtained from 2 acceptability indices that we derived, and compare them with other preference measures, including durations of feeding events and the forage ratio. Our study was carried out with the support of the Found. Res. Dev. of the Counc. Sci. and Ind. Res., and formed part of the S. Afr. Savanna Ecosystem Proj.