We review information on the structure of proline-rich proteins (PRPs), their various functions related to oral homeostasis and dietary tannin, and the structural basis of these functions. Consideration of the multifunctional nature of these salivary proteins helps explain both the subtle and large variations found in structure and secretion rates both within individuals and between species. We propose that the ancestral function of PRPs is in maintaining oral homeostasis and that counteracting dietary tannins by binding with them is a derived function. PRPs are effective in oral homeostasis at low secretion levels, whereas counteracting tannin depends on high secretion levels. In the dietary habits ranging from carnivores through omnivores to exclusively planteaters, the dietary nitrogen level is progressively reduced, and plant allelochemical intake, including tannins, increases. We suggest that during this evolution from meat-eater to plant-eater, there was some point in omnivory at which selective pressure from nitrogen limitations, arising from a low nitrogen/high tannin diet, became sufficiently great for the evolution of increased secretion level and diversification of PRPs for dealing with tannin. If this hypothesis is correct, carnivorous mammals should secrete low levels of PRPs for oral homeostasis, but should never secrete high levels, unless they are secondarily carnivorous. Omnivores consuming a diet of very little animal tissue but higher levels of tannin-containing foliage or fruit should generally have the capacity to produce high levels of salivary PRPs. Browsers and frugivores should also produce high levels of PRPs, but grazers may have reduced secretion rates depending on the antiquity of the dietary habit. This hypothesis is consistent with the limited information available on the abundance, type, and distribution of PRPs in mammals. Studies are suggested which would test the functional and evolutionary arguments presented.