The teratogenic potential of commercial formulations of atrazine (40.8%) and 2,4-D was evaluated using FETAX (frog embryo teratogenic assay--Xenopus). Because these herbicides have been detected in ground and surface water, this study was designed to determine the adverse effects in buffer and natural water for both herbicides. All treatments showed a significant concentration-response effect on exposed embryos, except for the 2,4-D natural water sample. Atrazine (solubility of the commercial formula used 70 mg/L at 20 degrees C), compared to 2,4-D (solubility = 311 mg/L at pH = 1 and 25 degrees C), had a significantly greater teratogenic effect in both the buffer (atrazine EC50 = 33 mg/L, LC50 = 100 mg/L, TI = 3.03; 2,4-D EC50 = 245 mg/L, LC50 = 254 mg/L, TI = 1.04) and natural water samples (atrazine EC50 < 8 mg/L, LC50 = 126 mg/L; 2,4-D EC50 and LC50 > 270 mg/L). The 2,4-D EC50 and LC50 values for the buffer were similar at 245 mg/L and 254 mg/L. These similar values and the teratogenic index (TI) of 1.04 suggested that 2,4-D was more embryotoxic than teratogenic to frog embryos at high concentrations. Atrazine in natural water demonstrated a significantly greater EC50 (100% abnormality at 8 mg/L, the lowest test concentration) to frog embryos than the buffer experiment (EC50 = 33 mg/L). The extrapolated lowest observable adverse effect concentration (LOAEC) for the natural water experiment was 1.1 mg/L. These results suggest that atrazine toxicity is enhanced by the synergistic or additive effects of some component of the water or atrazine was already present in the sample. In contrast to atrazine, 2,4-D was less toxic in natural water than buffer. These results suggest that both atrazine and 2,4-D pose little threat, since their embryotoxicity and teratogenicity to frog embryos occur at high concentrations approaching their maximum solubility levels in water.