The Seychelles Child Development Study (SCDS) is testing the hypothesis that prenatal exposure to low concentrations of methylmercury from a maternal diet high in fish is related to the child's developmental outcome. In this report, 217 children from a pilot cohort were reevaluated at 66 months of age. The evaluation included the McCarthy Scales of Children's Abilities, the Preschool Language Scale, and age-appropriate sub-tests from the Woodcock-Johnson Tests of Achievement. Maternal hair total mercury, measured by cold vapor atomic absorption in a maternal hair segment corresponding to pregnancy, revealed a median exposure of 7.1 ppm. The association between maternal hair mercury levels and neurodevelopmental outcomes at 66 months of age was examined by multiple linear regression analysis with adjustment for important confounding variables. The results indicated that mercury exposure was negatively associated with four endpoints (the McCarthy General Cognitive Index and Perceptual Performance subscale and The Preschool Language Scale Total Language and Auditory Comprehension subscale). After normalizing the data by removal of a small number of outliers or highly influential scores, the mercury effects were no longer significant except for auditory comprehension. These results should be viewed as preliminary and interpreted with caution, since the SCDS main study 66-month evaluations, which are better controlled with more detailed endpoints are being analyzed. This study highlights the difficulties in interpreting epidemiologic studies of this type and the degree to which overall results in multivariate analyses can be influenced by a very small number of cases.