Convergent evolution in feeding types: Salivary gland mass differences in wild ruminant species

In the ongoing debate about divergent evolutionary morphophysiological adaptations of grazing and browsing ruminants, the size of the salivary glands has received special attention. Here, we report the most comprehensive dataset on ruminant salivary glands so far, with data on the Glandula parotis (n = 62 species), Gl. mandibularis (n = 61), Gl. buccalis ventralis (n = 44), and Gl. sublingualis (n = 30). All four salivary gland complexes showed allometric scaling with body mass (BM); in all cases, the 95% confidence interval for the allometric exponent included 0.75 but did not include 1.0 (linearity); therefore, like other parameters linked to the process of food intake, salivary gland mass appears to be correlated to metabolic body weight (BM0.75), and comparisons of relative salivary gland mass between species should rather be made on the basis of BM0.75 than as a percentage of BM. In the subsequent analyses, the percentage of grass (%grass) in the natural diet was used to characterize the feeding type; the phylogenetic tree used for a controlled statistical evaluation was entirely based on mitochondrial DNA information. Regardless of phylogenetic control in the statistical treatment, there was, for all four gland complexes, a significant positive correlation of BM and gland mass, and a significant negative correlation between %grass in the natural diet and gland mass. If the Gl. parotis was analyzed either for cervid or for bovid species only, the negative correlation of gland mass and %grass was still significant in either case; an inspection of certain ruminant subfamilies, however, suggested that a convergent evolutionary adaptation can only be demonstrated if a sufficient variety of ruminant subfamilies are included in a dataset. The results support the concept that ruminant species that ingest more grass have smaller salivary glands, possibly indicating a reduced requirement for the production of salivary tannin-binding proteins.