Energy digestibility in dog and cat food can be predicted by the linear regression of the fiber content on a dry matter (DM) basis. Fiber is the independent variable and in vivo measured energy digestibility the dependent variable. This has been demonstrated for crude fiber (CF) in both species and in dogs for total fiber as analyzed by the Englyst and Cummings method (1,2). To date, however, few observations have been published measuring both total dietary fiber (TDF, according to the method of Prosky et al. [3]) and energy digestibility in dogs and cats. From the existing data points, it is thus not possible to 1) calculate a regression equation as for crude fiber or total fiber that is likely to predict digestibility more accurately in pet foods and 2) compare the accuracy of prediction of energy digestibility by CF and TDF methods.