To understand the ways in which condensed tannins (CT) affect primate diet selection and nutritional status, correct measurements are essential. In the majority of studies of the CT contents of primate foods, a tannin source such as "quebracho" is used to standardize CT assays, but the CT in quebracho tannin may not be similar to those in the plants of interest. We investigated how the choice of standard to calibrate CT assays affects the estimation of CT in the diets of mountain gorillas (Gorilla beringei). We purified the CT from gorilla foods and compared the actual amounts of CT in the foods with estimates produced by using the quebracho tannin. When quebracho was used, the estimates of CT contents of gorilla foods were, on average, 3.6 times the actual content of CT so that the amounts in frequently eaten gorilla foods were substantially overestimated. The overestimation for a given plant could not be predicted reliably and the ranking of plants by tannin content differed according to the standard used. Our results demonstrate that accurate measurements of CT necessitate the use of tannins purified from the plant species of interest. A reevaluation of primatology studies using interspecific comparisons of tannin content will provide new insights into primate food selection and nutritional ecology.