The human fetus is exposed to a variety of environmental agents and drugs which cross the placenta and can induce DNA damage. Micronucleus (MN) determination is a suitable and sensitive method for measuring DNA damage and since umbilical cord blood is obtained without any risk for the newborn, we measured the frequency of MN in cells from cord blood in four groups of healthy newborns (NB): 35 NB whose mothers lived in two urban cities (groups I and II); 16 NB from an agricultural area (group III); and 15 NB of mothers with high-risk pregnancy (group IV). MN were also evaluated in the mothers of NB from group I (n=17) and group III (n=14). Acetylcholinesterase (AChE) concentration was measured in groups I and III. The average frequency of binucleated cells with MN was 3.7+/-1.4 in 1000 cells in mothers and 1+/-0.9 in 1000 cells in NB from urban areas; and 4.5+/-2.4 in 1000 cells in mothers and 2+/-1.5 in 1000 cells in NB from the agricultural area. The correlation between the frequency of MN in mothers and NB was significant (r=0.61, p<0.01). AChE levels of samples obtained both from group III mothers and from newborns were similar to those of group I. The Wilcoxon's rank-sum test was applied to measure differences in MN frequency; NB of group I were used as control group. A significant (p<0.01) higher frequency of MN (4+/-2) was found only in lymphocytes from NB from high-risk pregnancies. Data indicate that MN evaluation in umbilical cord samples might be useful in the identification of transplacental mutagens.