Comparison of the reproducibility of fetal crown rump length measurements (CRL) obtained in a resting state by sonar with a phase focused linear array real time scanner, a real time mechanical sector scanner, and a static B-scanner did not show any significant differences between machines. Maternal movement was found to produce a change in fetal intrauterine position and to induce fetal movement. Examination following maternal movement was felt to be more representative of clinical conditions and produced greater variability of CRL measurement, shown by all machines to a differing extent, although the differences between machines were not substantial. There is some evidence that the discernment of fetal movement and the quality of fetal echoes obtained with different machines affects the reproducibility of CRL measurement. Thus variability of CRL measurement with every machine is small in terms of gestational age and justifies the use of real time machines to establish gestational age.THE measurement by sonar of fetal crown rump length (CRL) in the first trimester of pregnancy is a recognised method of establishing gestational age (Robinson, 1973;Drumm et al, 1976). Real time ultrasound systems, because of their portability (Balfour, 1978) and their ease of use, especially by inexperienced operators (Docker and Settatree, 1977), are particularly suitable for this role.Adam et a1 (1979) have shown a good correlation between CRL measurements made with one type of real time machine and a static 734