<h2>Summary</h2> In this paper we have reviewed major findings from several studies designed to test the hypothesis that anoxia is a significant perinatal experience. Several design factors which may help to account for the discordant conclusions arrived at by various investigators were pointed out. These were: approach to the problem (prospective or retrospective), criteria of later impairment, use of control groups, and criterion of early anoxia. It is the factor of definition of early anoxia with which we have been especially concerned in this and earlier studies of neonates. On the basis of these studies, several conclusions and suggestions emerged. First, blood oxygen saturation appears to reflect very transitory levels of oxygenation. Since it is not necessarily closely related to even the immediate neonatal condition, it would not be surprising if there were no relationship with more remote events. It seems necessary, consequently, to employ a less exact but more meaningful clinical criterion of anoxia. Second, even though a clinical criterion is used, it should permit some ordering as to degree in addition to satisfying the usual scientific demands for objectivity and reliability. Third, there is considerable variability in the postnatal rate of recovery of anoxic infants. Measurement of these differences provides a way of identifying those infants who may be most likely to show residual effects. Fourth, events believed to lead to fetal anoxia were associated, in our subjects, with relatively greater impairment in performance on Newborn Behavior Tests than either postnatal anoxia or nonmechanical delivery complications. This finding highlights the necessity of controlling for fetal factors in any research designed to assay the effects of anoxia.