A group of 137 patients, 113 of whom had obstetric complications, recorded fetal movements by a subjective method. Low fetal movement counts were associated with maternal cigarette smoking and prolongation of pregnancy and high counts with maternal ingestion of sedatives or tranquillizers, abnormal fetal heart rate in labour and a true knot in the umbilical cord. The absence of fetal movement during four consecutive 20-minute counting-sessions was not associated with poor fetal outcome. Further investigations to assess the usefulness of fetal movement counting are recommended. Such investigations should take into account the limitations imposed by the methods and the obstetric population for study.