Analysis of confined placental mosaicism (CPM) for trisomies 2, 3, 7, 8, 9, 16, and 22, in diagnostic chorionic villus sampling procedures, demonstrates apparent incidences of CPM for individual trisomies of between 9 and 91 cases per 100,000 pregnancies, with trisomy 7 being the most common. More detailed analysis of the percentage of aneuploid cells present, and the distribution of abnormality between the cytotrophoblast and extra-embryonic mesoderm cell lineages, shows a highly specific pattern for each chromosome. Theoretical considerations, in conjunction with direct observations, indicate that the overriding influence on the patterns of cell distribution seen in CPM is the distribution of aneuploid cells laid down during blastogenesis. This in turn reflects closely the origin of mosaicism from either correction of a trisomic conception or post-fertilization somatic error. The pattern of aneuploid cells for each trisomy, as seen at the end of the first trimester and later in pregnancy, can therefore be used to predict the relative contribution of meiotic and mitotic errors to CPM, and hence the likely incidences of uniparental disomy from this source, upd(16)mat being the most common (1 in 10,000 continuing pregnancies). In addition, CPM for trisomies 2, 3, and 8 shows strong evidence of a non-random distribution of aneuploid cells between the different extra-embryonic cell lineages. Analysis of comparable data from spontaneous abortion material repeats this non-random pattern for trisomies 2 and 3, and suggests that a similar phenomenon may also be occurring for trisomy 22. A non-random distribution could be attributable to selection for or against, or intolerance of, particular trisomic cells in certain lineages, but is more probably a result of either cell lineage-specific non-disjunction or consistent uneven compartmentalization of aneuploid cells during blastocyst development.