A new method of estimating fetal exposure is used in a dose-response analysis of data from the 1971 outbreak of methyl mercury poisoning in rural Iraq. An X-ray fluorescence instrument for the measurement of single strands of human hair was employed to obtain longitudinal profiles recapitulating fetal exposure. Logit and hockey-stick models as well as nonparametric smoothing are used to describe data on delayed development and central nervous system abnormality.