HISTORICAL REVIEW Attempts at refining maternal perception of fetal activity by objective observation focused initially on the recording of cardiac activity. In the decades following Cremer’s first successful fetal electrocardiogram ( 1906) the technique was used essentially for the diagnosis of fetal life (Strassmann and Mussey, 1938; Southern, 1954). Later Southern ( 1957) and Hess and Hon ( 1960) attempted to recognise patterns of fetal heart rate recordings, obtained by electrodes placed on the mother’s abdomen and thigh, during labour which correlated with degrees of fetal distress. The use of Doppler ultrasound transducers on the mother’s abdomen to obtain prolonged recordings of the fetal heart rate pattern antepartum (Callagan et al., 1964; Bishop, 1966) allowed this assessment (and the possibility of appropriate intervention) to be brought into the antenatal period (Visser and Huisjes, 1977; Paul and Miller, 1978). Fetal body movements in utero were first recorded objectively with a kymograph by Ahlfeld ( 1905) using the displacement of an inverted glass funnel placed on the mother’s abdomen produced by the distortions of the maternal abdominal wall resulting from movement of the underlying fetus. Subsequently, various methods have been applied to record fetal activity using strain gauges (Sterman, 1967; Timor-Tritsch et al., 1976; Wood et al., 1977), electromagnetic devices (Sadovsky et al., 1973; Adler and Mahler, 1980) and piezoelectric sensors (Sadovsky et al., 1977). Ultrasonic recordings initially used A-mode signals, with reflections of the single ultrasound beam at different distances displayed as spikes on an oscilloscope. Fetal breathing movements could be effectively detected (Boddy and Robinson, 197 1) but more randomly moving fetal parts were difficult to identify with this method. The use of Bmode linear scans to produce a two-dimensional image of cross sections through the uterus was a major advance in demonstrating fetal anatomy (Kratochwil, 1967; Hofmann et al., 1967). A rapid succession of linear scans reproduced on a video screen in real-time allows comprehensive observations of fetal movement, amongst the earliest reports of which were by Hofmann and Hollinder ( 1968) and Reinold (1976). More recent advances in ultrasonic instrumentation include a B-scan phase-locked tracking system, which provides a fine resolution paper record of rhythmic movements of any portion of the fetus selected from a simultaneous conventional real-time picture (Adamson et al., 1980). The clinical significance of ultrasonic observation of musculoskeletal movements of the fetus has been appreciated in terms of establishing viability in early pregnancy (Hollander, 1979), of correlating them with our knowledge of the embryological development of the nervous system (Birnholz et al., 1978) and of using them as a parameter of fetal well being in later pregnancy (Manning et al., 1979).