Reject oocytes from in vitro-fertilization patients are currently the only practical source of human oocyte material available for meiotic studies in women. Two hundred clearly analyzable second meiotic (MII) metaphase oocytes from 116 patients were examined for evidence of first meiotic (MI) division errors. The chromosome results, in which 67% of oocytes had a normal 23,X chromosome complement but none had an extra whole chromosome, cast doubt on the relevance, to human oocytes, of those theories of nondisjunction that propose that both chromosomes of the bivalent fail to disjoin at MI so that both move to one pole and result in an additional whole chromosome at MII metaphase. The only class of abnormality found in the MII oocytes had single chromatids (half-chromosomes) replacing whole chromosomes. Analysis of the chromosomally abnormal oocytes revealed an extremely close correlation with data on trisomies in spontaneous abortions, with respect to chromosome distribution, frequency, and maternal age, and indicated the likelihood of the chromatid abnormalities being the MI-division nondisjunction products that lead to trisomy formation after fertilization. The most likely derivation of the abnormalities is through a from of misdivision process usually associated with univalents, in which the centromeres divide precociously at MI, instead of MII, division. In the light of recent data that show that altered recombination patterns of the affected chromosomes are a key feature of most MI-division trisomies, the oocyte data imply that the vulnerable meiotic configurations arising from altered recombination patterns are processed as functional univalents in older women. Preliminary evidence from MI-metaphase oocytes supports this view.

Cytogenetic preparations from oocytes remaining unfertilised after in vitro fertilisation revealed single chromatids (as opposed to whole chromosomes) in 4 out of 38 meiosis II metaphases. In one oocyte, a single chromatid was present in addition to the normal 23,X complement, and in three oocytes, two identical but separate chromatids replaced one whole chromosome within the complement. The data indicate that the chromatids have arisen as a result of premature division of the centromeres at meiosis I ("predivision"). None of the oocytes were hyperhaploid with an extra whole chromosome. These findings are at variance with the general belief that trisomy in man is largely a consequence of nondisjunction of whole bivalents at meiosis I and they suggest that predivision resulting in separate chromatids may be a significant mechanism for human trisomy.

A series of 26 triploid foetuses was ascertained in a survey of spontaneous abortuses and the origin of the additional haploid complement determined in 21 of them by the study of foetal and parental heteromorphisms. In 17 the additional haploid set was paternal in origin, in 3 it was maternal in origin and in one the parental origin could not be determined. The best fit for the data using a maximum-likelihood method was that 66.4% of the triploids were the result of dispermy, 23.6% the result of fertilization of a haploid ovum by a diploid sperm formed by failure of the first meiotic division in the male and 10% the result of a diploid egg formed by failure of the first maternal meiotic division. The possible sources of error inherent in the technique are reviewed and our results compared with previously published data.

Trisomy is the single most frequent type of chromosome abnormality in humans and has considerable impact on many aspects of human pathology. It arises most commonly through “nondisjunction” at maternal meiosis I, but the underlying mechanism of formation remains obscure. Analysis of 100 haploid oocytes at second meiotic metaphase shows that the only type of chromosome abnormality compatible with trisomy formation after fertilisation is the presence of single chromatids in addition to, or replacing, whole chromosomes. The mechanism resulting in the presence of single chromatids is considered to be precocious division of univalents or dyads at first meiotic anaphase.

The mechanism of trisomy formation and its relationship to increased maternal age is not understood. Molecular analysis of the pattern of inheritance of DNA markers in trisomy families shows trisomies can be grouped according to whether the affected chromosomes inherited from their mothers are heterozygous or homozygous with respect to the centromeres. Furthermore, molecular analysis reveals that those that are heterozygous have fewer chiasmata, which are located more distally, while those that are homozygous have more chiasmata proximally located. Cytogenetic analysis of human oocytes shows that the kind of imbalance predicted by the classic hypothesis of nondisjunction, i.e. extra whole chromosomes at the second metaphase, is rarely found, whereas the common expression of imbalance is seen as single chromatids. We hypothesise that one mechanism links these data: the mechanism depends on the prediction from the cytogenetic data that cohesion within the bivalent complex is severely weakened during the extended dictyate stage in older women. Consequently, when meiosis resumes, at the time of ovulation, the bivalent emerges as four chromatids held together only by its chiasmata. In accordance with the rules of orientation on the spindle, the final balanced shape of the configuration achieved at metaphase I, in this case determined by the position of the chiasmata, will dictate whether the subsequent segregation of the chromatids will result in their heterozygosity or homozygosity. It follows that the concept of "first division" and "second division" errors, i.e. of nondisjunction originating at first or second meiotic division as defined by centromeric hetero- or homozygosity, may be erroneous.