Methods A longitudinal cohort study beginning in early pregnancy was conducted in rural Bangladesh, where arsenic concentrations in well water vary considerably. We assessed the effects of pre- and postnatal arsenic exposure on development of 2112 children at 18 months of age with Bayley Scales of Infant Development-II (mental and psychomotor development indices), Wolke's Behavior Rating Scale and maternal report of language. We related the measures of child development to arsenic concentrations in maternal urine in gestational weeks 9 and 30 and child's urinary arsenic at 18 months of age. Details of socio-economic background, home stimulation and anthropometric measurements of mothers and children were also available.Results Median maternal urinary arsenic concentration averaged over early and late gestation was 96 mu g/l, whereas children's urine contained 35 mu g/l of arsenic. There was no significant effect of any of the arsenic exposure measures on any of the child development measures after controlling for social and economic confounders, child's age and sex.Conclusion Contrary to expectations, we found no indications of adverse effects of pre- or postnatal arsenic exposure on child development at 18 months. It remains possible that duration of exposure is critical and that effects will become apparent later in childhood.