An historical cohort study of twins, aged 6 to 15 years, found reduced cognitive performance related to subclinical exposure to lead (Pb) much earlier in life. Pairs of twins discordant for blood Pb (low-Pb twins ranged from 30-50 micrograms/dl and high-Pb twins ranged from 43-80 micrograms/dl) exhibited reduced learning of a computer-administered visual discrimination and reversal by the twin having the higher exposure. There was no evidence that sensory or motor impairment contributed to the cognitive deficit. Performance of a reference group indicated that test's validity for age-related cognitive development and the method's suitability for children of varying socio-economic or racial status. Because the reversal learning approach required only one 20-min test session and was powerful enough to document Pb-related deficits in a small sample (n = 8), it may indicate a practical method for obtaining early evidence of environmentally induced developmental delays. An advantage of the test is that it can be used with both humans and animals. Tests capable of direct comparison of behavioral data from humans and animals can guide the search for behavioral and biological mechanisms in experiments that can be done only with animals. They also can augment the clinical procedures currently used.