Data on pregnancies with and without Down's syndrome between 10 and 14 weeks of pregnancy were used to determine the performance of combined ultrasound and biochemical markers in prenatal screening for Down's syndrome. We used three datasets: one published by Pandya et al. (1995) on nuchal translucency measurement in 86 Down's syndrome pregnancies; one published by Wald et al. (1996a) on free beta human chorionic gonadotrophin (hCG) and pregnancy-associated plasma protein A (PAPP-A) in 77 affected pregnancies and 385 unaffected pregnancies; and the third from Schuchter et al. on 561 unaffected pregnancies with nuchal translucency measurement. Combining the data from these three datasets showed that screening between 10 and 14 weeks by combining the serum markers with nuchal translucency measurement had a detection rate of 80 per cent for a 5 per cent false-positive rate, better than maternal age with two serum markers (62 per cent for 5 per cent) or maternal age with nuchal translucency measurement (63 per cent for 5 per cent). At this time in pregnancy, it appears that screening using the combined test is better than second-trimester serum screening (76 per cent for 5 per cent), though these estimates do not allow for any association between the markers and spontaneous fetal loss, an issue that needs to be clarified by further research. Meanwhile, these results provide a reasonable working estimate of screening performance using different combinations of these markers.