Vertebrate viviparity (live-bearing reproduction), placentation, and placentotrophy are widely assumed to have evolved as three successive, gradualistic transformations. From empirical data and predictive tests on lizards and snakes, this paper indicates that placentae and a degree of placentotrophy have evolved repeatedly as necessary correlates of viviparity, not as subsequent modifications. In addition, information derived from studies of anatomy, physiology, biogeography and systematics is used to evaluate new saltationist and punctuated equilibrium models for the evolution of viviparity. Phylogenetic reconstruction reveals that more than 100 squamate clades have made the transition to viviparity and placentation. However, various phenotypic intermediates postulated by the gradualistic model are either scarce or unrepresented among known forms, including those in which viviparity has evolved at specific and subspecific levels. Evolution in squamates seems to have produced a dichotomy between two evolutionarily stable patterns: (i) retention of weakly shelled or shell-free eggs to term (viviparity), with development of fully functional placentae; and (ii) deposition of shelled eggs at or near the limb bud stage of development (typical oviparity). Conflicting functional demands placed on eggshell morphology may constrain establishment of prolonged, oviparous egg-retention as a viable, historically stable pattern. Alternatively, the costs of prolonged egg-retention associated with decreased female mobility or decreased fecundity may exceed the benefits in oviparous forms.