An in vitro system has been set up to study the possible genetic effects of diagnostic ultrasound (US) on the developing fetus. Amniotic cells were cultured by the in situ technique. Well established clones were exposed to US for various lengths of time using a linear array transducer within a sound transparent apparatus. Cells were then grown in the presence of 5-bromo-2'-deoxyuridine and processed for scoring the number of sister chromatid exchanges (SCE) as the cytogenetic endpoint. There was no linear relationship between the SCE frequency and the duration of US exposure. Variance analysis showed that only interindividual variability was a significant component of total variation. Neither the main effect of treatment nor the interaction effect were statistically significant. The data suggest that US delivered from a diagnostic unit to actively growing cultured fetal cells in a system closely mimicking the conditions of US exposure during amniocentesis does not induce SCE.