The aim of this study was to determine the efficacy of cell separation with single density and triple density-gradient techniques in the yield of foetal erythroblasts isolated from maternal blood. Maternal blood was obtained from 20 singleton pregnancies at 11-14 weeks of gestation immediately before foetal karyotyping by chorionic villus sampling. In each woman, the blood sample was divided into two portions; one portion was used for single density-gradient separation and the other, for triple density-gradient separation. Magnetic cell sorting (MACS) was subsequently performed with anti-CD71/antiglycophorin-A. The enriched erythroblasts were stained with Kleihauer-Giemsa and with fluorescent antibodies for the gamma, epsilon and zeta globin chains. The percentage of foetal cells positive for each stain was calculated. Fluorescence in situ hybridization (FISH) for X- and Y-chromosomes was also performed. Comparison was made in the proportion of enriched foetal cells between the two separation methods for each CD71 and glycophorin-A (GPA) antibody. The percentage of erythroblasts enriched from maternal blood that stained positive for gamma, epsilon and zeta globin chains and with Kleihauer-Giemsa was significantly higher in the triple density-gradient separation fractions compared with the single density-gradient fractions with both anti-CD71 and GPA MACS. FISH analysis for the Y-chromosome confirmed the increase in foetal cell proportion in the triple density-gradient samples. Isolation of foetal erythroblasts from maternal blood using triple density-gradient separation and MACS is more effective with regard to foetal cell yield and purity than single density-gradient separation and MACS.