ANALYSIS OF SEX DISTRIBUTION WITHIN FAMILIES IN A LARGE LATIN-AMERICAN SAMPLE

A sample of 59,496 Latin American individuals belonging to 28,545 sibs hips was taken from a large case-control Collaborative Study on Congen ital Malformations (ECLAMC). Newborn babies with congenital malformati ons, and their respective controls, were excluded from the sample in o rder to avoid biases associated with malformation. On the assumption t hat Poisson and Markovian effects are not important in determing the o verall population distribution of sibships, some statistical models we re applied. The general model was a double binomial with one-tail exce ss. This model can be simplified to a double binomial, a simple binomi al with one-tail excess or a simple binomial. The first three distribu tions fitted the data well (chi(15)(2) = 16.46, P = 0.35; chi(16)(2) = 20.92, P = 0.18; chi(17)(2) = 19.88, P = 0.28), while the simple bino mial did not (chi(18)(2) = 35.01, P = 0.009). The same models were als o applied to several studies carried out in different countries at dif ferent times, with similar results. The most parsimonious model that f its most data sets is the double binomial, the results from which sugg est that 9% of couples segregate X-linked recessive lethals, since the two segregation proportions of the model are closely in line with the genetic prediction.