Seven samples of trimethyltin obtained from three commercial sources were evaluated for neurotoxic potency in the rat. Hippocampus weight, histology and assays of the astrocyte protein, glial fibrillary acidic protein, were used as indices of neurotoxicity. A single administration (8.0 mg/kg, i.v.) of different samples of TMT resulted in markedly different degrees of neurotoxicity as assessed by all three criteria. Subsequent analysis of each sample for trace metal and speciated organotin content revealed that sample-to-sample differences in neurotoxic potency could be attributed to the presence of several impurities. Indeed, in several samples, sodium was present at levels high enough to affect neurotoxic potency simply by diluting the trimethyltin content. A number of samples also showed contamination with the non-neurotoxic organotin, dimethyltin. The data underscore the need for standardized sources of toxic compounds to insure the reliability of experiment-to-experiment comparisons.