Major problems remain in the evaluation of important parameters of steroid contraceptive usage. The questions of contraceptive effectiveness, of “minor” side-effects such as changes in menstrual bleeding pattern or nausea, and of subjective symptoms such as mood and libido changes are not reviewed at this time. Attention is centered on the so-called major side-effects, particularly thromboembolic phenomena, carcinogenesis, metabolic alterations, and certain endocrine effects. Survey-type data have failed to yield any evidence for an increase in thromboembolic disease in users of oral contraceptives. Evaluation of this and related information is complicated by the fact that there has been a rapid rise in idiopathic thromboembolic disease, unrelated to use of oral contraceptives, in several developed countries. Retrospective case-control studies both in England and the U.S.A. point to an increased risk of thromboembolic disease, about sixfold, in oral contraceptive users. There are significant differences between the two sets of studies with respect to the age-factor, and the risk of cerebral thrombosis. These studies are critically examined in terms of the problems inherent in this type of statistical approach, and they are felt to be less than conclusive. The fact that these studies yield similar conclusions may be related to the fact that they are of the same basic design and carried out at a time when unavoidable biases of the same sort may affect them all. It is pointed out that risk of death from thromboembolic disease, at the maximum estimate derived from these studies, is still less than the risk from the use of other contraceptive modalities, when the maternal mortality from the use of these less effective methods is included. Certain basic data relative to carcinogenesis in experimental animals are reviewed. It is pointed out that marked species and strain differences exist, and that primates appear to be unusually resistant to tumor induction by estrogens, as indicated also by the failure of any experiments to date to produce mammary tumors in this species. The importance of continuous (and usually high) dosage in the induction of animal tumors is emphasized, as is the difference between these regimens and those employed in human therapy. The regression of most estrogen- induced animal tumors upon removal of the estrogenic stimulus, as well as the tumors' total dependence on the presence of the hypophysis is a major consideration. The possibility exists that estrogens stimulate some pituitary factor in rodents, and that this pituitary factor is tumorigenic, while the estrogen is involved only indirectly. However, some evidence exists for breast neoplasms induced in human males by prolonged, continuous, intense estrogen exposure. Progestagens have also produced mammary neoplasms in dogs; these have not been malignant, and comparative studies with progesterone itself are not available. Except for the isolated reports in men, incidence studies and retrospective surveys have failed to show an increase in breast cancer in menopausal women using estrogens. There are data which suggest, however, that the hormonal environment at certain periods of premenopausal life may affect the future incidence of breast carcinoma, hence the data from estrogen use by menopausal women may not be directly relatable to the problem of oral contraceptive usage. Numerous metabolic effects, most of them readily predictable from earlier animal experimentation, are manifested in some women using synthetic estrogens, progestagens, or combinations of the two. Simultaneous use of two compounds, as in combination-type contraceptives, confuses evaluation. Such metabolic changes are related in rare instances to clinical complications such as hypertension, etc. The clinical and prognostic significance of chemically demonstrable alterations in carbohydrate metabolism, serum lipids, etc. are felt to be purely speculative at the present time. Endocrine effects persisting after discontinuation of oral contraceptives are rare; apparently both types of steroidal agents play some part. The prognosis for hypothalamo-pituitary dysfunction (not existing prior to the use of these agents) is excellent. Evaluation of published data in this field is difficult because polemical as well as other irrelevant factors have crept into the picture. This feature, as well as notable geographic and population differences, suggest the use of great caution in formulating fixed opinions at the present time.