Acute toxicity and genotoxicity of the 54.8% 2,4-D-based commercial herbicide DMA® were assayed on Cnesterodon decemmaculatus (Pisces, Poeciliidae). Whereas lethal effect was used as the end point for mortality, frequency of micronuclei (MNs), other nuclear abnormalities and primary DNA damage evaluated by the single cell gel electrophoresis (SCGE) assay were employed as end points for genotoxicity. Mortality studies demonstrated an LC50 96 h value of 1008 mg/L (range, 929-1070) of 2,4-D. Behavioral changes, e.g., gathering at the bottom of the aquarium, slowness in motion, slow reaction and abnormal swimming were observed. Exposure to 2,4-D within the 252-756 mg/L range increased the frequency of MNs in fish exposed for both 48 and 96 h. Whereas blebbed nuclei were induced in treatments lasting for 48 and 96 h, notched nuclei were only induced in fish exposed for 96 h. Regardless of both concentration and exposure time, 2,4-D did not induce lobed nuclei and binucleated erythrocytes. In addition, we found that exposure to 2,4-D within the 252-756 mg/L range increased the genetic damage index in treatments lasting for either 48 and 96 h. The results represent the first experimental evidence of the lethal and several sublethal effects, including behavioral alterations and two genotoxic properties namely the induction of MNs and primary DNA strand breaks, exerted by 2,4-D on an endemic organism as C. decemmaculatus.