Outbreaks of methylmercury poisoning in Japan and Iraq have demonstrated the sensitivity of the fetus to neurotoxic effects. Based on toxicokinetics and considerations of practicability, the optimal biomarker of methylmercury exposure is the hair concentration, but whole-blood measurements of mercury are also useful. Dose-response relations are still incompletely known, especially concentrating developmental neurotoxicity under conditions of chronic exposure. Available evidence indicates that neurobehavioral dysfunction in children may occur if the maternal mercury concentration in hair is > 6 micrograms/g (30 nmol/g). This value corresponds to a blood mercury concentration of approximately 24 micrograms/L (120 nmol/L). The period of maximum sensitivity of the nervous system to methylmercury toxicity is unknown, but the transfer of mercury to the newborn through human milk may represent an additional risk. In view of the wide occurrence of mercury contamination in developing countries, increased use of the exposure biomarkers is encouraged.