A comparison and critique of different scat-analysis methods for determining carnivore diet

Abstract

For terrestrial carnivores, scat analysis is the technique most often used to deter- mine diets. Various methods of interpreting scat-analysis data exist; however, little is known about how the choice of method affects the results. 2. We reviewed 50 scat-analysis papers to assess the range of methods currently used. Furthermore, we used a large data set from cape fox Vulpes chama and black-backed jackal Canis mesomelas scats to compare 11 scat-analysis methods. Techniques tested included five biomass calculation methods, four frequency of occurrence methods, one method that estimated volume in scats, and another that estimated mass of food items in scats. 3. Frequency of occurrence methods were used in 94% of reviewed papers, and in 50% of papers they were the sole methods used. However, we conclude that fre- quency of occurrence has the least ecological significance and results can be mis- leading. Although biomass calculations probably provide the best approximation to true diets, only 23% of reviewed papers used suitable biomass calculation methods when models were available for the study species. 4. Analysis of fox and jackal scats showed that there were significant differences among methods when calculating percent diet composition and niche breadth. Additionally, dietary overlap between species differed considerably among the methods (range of R0 = 0.29-0.79). We conclude that the choice of method can have a significant impact on the results of dietary analysis, and can lead to very different conclusions about a species' ecology. 5. The best approximation of the true diet can be obtained by using a biomass calculation model that was developed for the same species, or for a closely related species with a similar food spectrum. When no such model is available, either the volume or mass of diet components in the scats should be used. To document rare food items, frequency of occurrence data could also be given.