BACKGROUND: To establish zebrafish as a developmental toxicity model, we used 7 well-characterized compounds to examine several parameters of neurotoxicity during development. METHODS: Embryos were exposed by semistatic immersion from 6 hrs postfertilization (hpf). Teratogenicity was assessed using a modified method previously developed by Phylonix. Dying cells in the brain were assessed by acridine orange staining (these cells are likely to be apoptotic). Motor neurons were assessed by antiacetylated tubulin staining and catecholaminergic neurons were visualized by antityrosine hydroxylase staining. RESULTS: Atrazine, dichlorodiphenyltrichloroethane (DDT), and 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) were primarily teratogenic and not specifically neurotoxic. 2,4-dichlorophenoxyacetic acid (2,4-D), dieldrin, and nonylphenol showed specific neurotoxicity; dieldrin and nonylphenol were specifically toxic to catecholaminergic neurons. Malathion, although not teratogenic, showed some nonspecific toxicity. CONCLUSIONS: Teratogenicity measured in 96-hpf zebrafish is predictive of mammalian teratogenicity and is useful in determining whether a compound causes specific neurotoxicity or general developmental toxicity. Induction of apoptosis or necrosis is an indicator of neurotoxicity. An effect on motor neurons in the caudal third of the embryo correlates with expected defects in motility. Overall, our results showed a strong correlation with mammalian data and suggest that zebrafish is a predictive animal model for neurotoxicity screening. Birth Defects Research (Part A) 76:553–567, 2006. Published 2006 Wiley-Liss, Inc.