We report 611 non-mosaic and 91 mosaic findings of trisomies 13, 18 and 21, and numerical sex chromosome abnormalities in a series of 20,527 CVS, in the Association of Clinical Cytogeneticists U.K. Collaborative Study, the majority with analysis of both direct preparations and cultured cells. No false-positive results were encountered among the 611 non-mosaic cases, making these findings a very reliable indicator of the fetal karyotype. One false-negative case was reported. In contrast, the 91 mosaic abnormalities were unreliable predictors of fetal abnormality. Many were associated with normal outcomes, but a significant proportion of cases of each individual aneuploidy proved genuine. Mosaicism for 45,X, and trisomies 13 and 18 was disproportionately common. 17 of the mosaic cases showed complete discordance between the karyotype from direct preparations and that from cultured cells. All would have resulted in either a false-positive or a false-negative finding if only one technique had been used. Based on our experience, and that of others, we believe that the highest level of predictive accuracy using CVS can only be achieved if both direct preparation and cell culture are performed. In addition, we continue to recommend that all pregnancies demonstrating mosaicism for these aneuploidies at CVS undergo amniocentesis or fetal blood sampling to differentiate between confined placental mosaicism and true fetal karyotypic abnormality.