To compare the indicators and levels of mercury (Hg) exposure in the mother with those in the fetal compartments, and determine its effects on the newborn. Hg levels using atomic absorption spectrophotometry were determined in maternal blood, breast milk, cord blood, infants' hair, and meconium of 78 consecutive mother-infant pairs in a community with high Hg pollution. The prevalence and levels of Hg both in meconium and in cord blood were correlated with maternal and infant risk factors. The prevalence of Hg in the fetal compartments was higher than in the maternal fluid compartments. Hg was present in 6.4% of maternal blood and 6.4% of breast milk, as compared with 16.7% of cord blood, 31.6% of infants' hair, and 46.1% of meconium. Forty-six percent of infants with Hg in cord blood had none in meconium, whereas 80.6% with Hg in meconium had none in cord blood. Hg was not present in the maternal blood of all infants (n = 36) with Hg in their meconium. Among those with detectable Hg, the mean levels were: mothers' blood 24 parts per billion +/- 5.47, cord blood 53.3 parts per billion +/- 37.49, and meconium 48.6 +/- 43.48. Quantitative measurement in hair was not done because of insufficient sample. Paired comparisons were all significant between Hg levels in the mothers' blood and meconium, mothers' blood and cord blood, and cord blood and meconium. Regression analysis showed Hg levels in meconium to be correlated with prevalence of Hg in infants' hair, length of stay in Tagum, and meconium-stained amniotic fluid. Fisher's Exact probability test showed that the prevalence of Hg in meconium was significantly related to the prevalence of Hg in the mothers' blood and length of stay in Tagum. The prevalence of Hg in cord blood was significantly related to the prevalence in the mothers' blood. Regression analysis of levels of Hg in cord blood showed a significant relation to levels in mothers' blood (.0001), prevalence in infants' hair (.0126), gestational age (GA) (.0091), and head circumference (HC) (.0469). By quadrant analysis of weight against HC in 66 full-term infants all of 4 infants weighing an average of >3000 g at birth and with HCs lower than the fifth percentile had Hg in meconium. The higher prevalence and levels of Hg in the fetal compartments reflect the ease of placental transfer with fetal trapping. Hg determinations in the mothers' blood underestimate the degree and extent of fetal exposure. There is a significant difference in each compartment's ability to reflect Hg exposure of the fetus. A small HC may be associated with the presence of Hg in meconium. Hg in meconium should be measured in addition to cord blood to determine the load of fetal Hg.