Four environmental agents have been tested for activity in a recently developed in vitro teratogen assay system. All four agents inhibited attachment. The agents were 40-fold concentrated drinking water (ID50 = 0.45 ml/ml), whole cigarette smoke condensate (ID50 = 85 micrograms/ml), kerosene soot (ID50 = 90 micrograms/ml), and commercial formulations of the pesticide carbaryl (ID50 approximately 150 micrograms/ml). On the basis of these examples appropriate criteria for the validation of in vitro teratogen assay systems are discussed. It is concluded that criteria are critically dependent on the specific applications of the assay system. For example, the false-positive rate must be minimized to make a wide-ranging screen of water samples useful. On the other hand, an investigation of impurities in commercial compounds requires low false-negative rates. In every case a quantitative measure of the potential teratogenic potency, in vivo, is desirable.