Changes in fetal motor activity may reflect changes in central nervous function and in the health of the fetus. Recording of fetal movement (FM) has therefore been suggested as a method of assessing fetal well-being. To establish the normal range of FMs in the third trimester of pregnancy, FMs were studied cross-sectionally in 180 and longitudinally in 6 healthy women with normal pregnancies. FMs were recorded simultaneously by the mother and by a FM detector using four piezo-electric crystals attached to the maternal abdomen. The recordings lasted for 45 min (cross-sectional study) or 30 min (longitudinal study). The median numbers of FMs recorded during 45 min by the FM detector and by the mother were 85 (2.5th percentile, 14; 97.5th percentile, 232) and 41 (2.5th percentile, 10; 97.5th percentile, 135), respectively. The median 45-min incidences of FMs recorded by the FM detector and by the mother were 8.1% (2.5th percentile, 1.3; 97.5th percentile, 30.2) and 3.7% (2.5th percentile, 0.9; 97.5th percentile, 15.6), respectively. Neither the number nor the incidence of FMs changed appreciably as gestational age advanced. The interindividual variance in the number and incidence of FMs was two to three times greater than the intra-individual variance (longitudinal study). The large variation in the quantity of FMs recorded during 30 and 45 min limits the value of using quantitative FM recording for antepartum fetal monitoring. Individualized normal limits of the quantity of FMs might be preferable to general limits.