Viviparity has evolved almost 100 times among lizards and snakes. Most scientific interest in this phenomenon has focussed on environmental variables that may have stimulated the evolution of viviparity via intermediate stages of prolonged oviductal retention of eggs (egg retention, or ER). Little attention has been paid to more proximate questions (e.g. what endocrine mechanisms control the duration of ER?) and their evolutionary consequences (e.g. is the degree of ER in a given female fixed or labile? Are there specific features of endocrine control systems for ER which may preadapt a taxon to the evolution of viviparity?). We develop a recently-proposed physiological model for prolonged uterine retention of eggs (Guillette, 1985), and investigate its implications for the evolution of reptilian viviparity. Our model is unusual in combining proximate mechanisms with ultimate (evolutionary) processes. We suggest that the duration of ER is controlled by circulating levels of progesterone, and that progesterone secretion by the adrenals in response to environmental cues may prolong ER. Hence, there may be a large phenotypic component to variance in ER. This mechanism may facilitate the evolution of viviparity in some taxa and in some habitats, by increasing variance in ER and hence (1) expediting the rate of change in ER inducible by selection, and (2) overcoming the necessity for very brief ER to confer a selective advantage. In situations in which prolonged ER confers a higher fitness to the reproducing organism, this character may evolve through genetic assimilation ("fixing" of a phenotypic response, from facultative to obligate). Our model generates several testable predictions, in terms of both endocrine mechanisms and evolutionary pathways.