Variation in the nutritional value of primate foods: Among trees, time periods, and areas

The study of nutritional ecology has proven to be useful for understanding many aspects of primate behavior and ecology and is a valuable tool in primate conservation. However, to date this approach has had limited application since chemical analyses of food items is very time-consuming and collections of perishable food material are often made in remote field locations. Such logistic difficulties have led to plant material being collected in a variety of fashions, and it is not known how variation in collection method might influence our understanding of the chemical basis of dietary selection. A standardization of collection methods is greatly needed to allow for direct comparison among studies. To develop an appropriate standardized method and to evaluate past research, it is necessary to understand along what dimensions plant chemistry varies. We evaluated variation in nutritional value--protein, fiber, digestibility, alkaloids, saponins, cyanogenic glycocides, and minerals--of leaf material from species eaten by red colobus (Piliocolobus tephrosceles) and black-and-white colobus (Colobus guereza) of Kibale National Park, Uganda. We consider variation at 3-levels: among trees, time periods, and areas. While there was considerable variation among species with respect to protein, digestibility, and saponins, there was also variation among individuals of the same species; in fact, individuals may vary by as much as 20%. The average coefficients of variation (CV) among individuals of the same species are 13.4 for protein, 12 for digestibility, and 43 for saponins, while the average CV among species are 35, 31.3, and 82.4, respectively. No species showed a variable response with respect to testing for the presence or absence of cyanogenic glycocides, while 2 of 11 species tested for alkaloids showed a variable response. Over 2 years there was evidence of variation among time periods in the chemical composition of the same food items. The protein-to-fiber ratio of mature leaves of the same species collected from 4 sites separated by 12 km within Kibale was also variable and in some cases the variation among sites was greater than the differences among species. For example, while Funtumia latifolia had little variation in protein-to-fiber ratio at 3 sites (0.44 at all sites), the remaining site was 28% greater. Because temporal variation is less than variation among individuals, it is likely more important to sample from multiple trees at a single point in ti