Maternal serum screening for alpha-fetoprotein (AFP) and human chorionic gonadotropin (hCG) increases the detection rate of Down's syndrome (DS) pregnancies. To estimate the risk of a DS pregnancy for a particular woman, Wald et al. combine a trivariate function of AFP, hCG, and unconjugated oestriol with age-specific risk. Calculation of independent likelihood ratios (LRs) for AFP and hCG has allowed us to examine the predictive value of each test alone and the combination. AFP and hCG were measured in stored serum samples from 672 normal, 8 trisomy 21 (DS), 9 trisomy 18, and 2 trisomy 13 pregnancies. AFP and hCG multiples of the median (MOM) were calculated for each sample. The LRs for AFP MOM and hCG MOM were calculated and combined with age-specific risk. Of eight DS pregnancies, six had increased risk based on age and AFP. Addition of hCG detected two additional DS pregnancies. Of nine trisomy 18 pregnancies, four (44 per cent) had hCG MOM under 0.25. Three out of nine would have been classified as high risk by AFP, but none by combined AFP and hCG. Amniocentesis would have been recommended in 74 per cent of aneuploid pregnancies if both age and serum screening were used. Abandonment of amniocentesis based on age alone would have excluded two abnormal pregnancies from detection. Screening programmes should note that combined risk figures are specific for DS and do not include other trisomies. Detection of other trisomies requires inclusion of low hCG level as a discriminator and continuation of age-based testing.