Arsenic is a known teratogen and developmental toxicant in many animal models. The aim of the present study was to determine the influence of arsenic oxidation state, concentration, duration of exposure, and embryonic gestational age on arsenic-induced developmental toxicity. For these studies whole embryo culture was used since this experimental model allows an assessment of the direct effect of the toxicant on the embryo and precise control of the variables of interest. ICR and CD1 mouse embryos were prepared for whole embryo culture and exposed to concentrations of trivalent (1, 2, 5, 7.5, 10, 20, and 30 microM of sodium arsenite) and pentavalent arsenic (5, 10, 20, 50, and 100 microM of sodium arsenate) at different developmental stages (3, 4-6, 8-10, or 20-23 pairs of somites) and for different exposure periods (1, 4, 6, or 24 hr). Embryonic growth, development, malformation rates, and viability were evaluated. A comparison of the ED50s of the two oxidation states showed that arsenite was about three times more potent than arsenate with respect to both malformations and lethality. The pattern of malformations was similar for both arsenite and arsenate and involved nonclosure of the cranial neural tube, prosencephalic hypoplasia, dysmorphogenesis of the optic and otic anlagen, and pharyngeal arch defects. ICR conceptuses were more sensitive than CD1s with regard to perturbation in embryonic growth by both forms of arsenic. ICRs were also more sensitive to otic, pharyngeal arch, and somite dysmorphogenesis induced by arsenite. With increasing gestational age there was an increasing resistance to arsenic-induced effects. In comparison to the 4-6 somite stage, the ED50 for induction of dysmorphogenesis was increased about twice at the 8-10 somite and over three times in 20-23 somite stage embryos. Exposure to arsenite/arsenate for a 1-hr period was sufficient to induce maldevelopment. A 6-hr exposure induced prosencephalic, otic, and optic abnormal development at a rate similar to that produced by a 24-hr exposure. The malformation pattern produced by exposure to arsenite/arsenate in vitro closely corresponds to that produced by maternal administration at the same gestational stage. This indicates that the arsenic embryopathy may be the result of a direct impact of the agent on the conceptus.