Document Type



We studied the time interval to the first birth and to the twin birth using statistical and mathematical models in two groups of mothers, those with twins and those with singletons, from the same population. We made use of a pair-matched case-control design. We treated the maternal birth cohort and parity as confounders and thus as controlled. We also investigated the sex of twin pairs as an interactive variable, employing such methods as survival curve testing and using geometric, gamma, and exponential distributions where appropriate. The expectations derived from the mathematical models yield numerical estimates of fertility components. The results suggest that unlike-sex twin-prone mothers have higher fecundity than controls when they conceive singletons. Further, fecundity appears high and unimpaired before the birth of twins. Mothers of like-sex twins experience somewhat shorter and but more variable birth intervals than corresponding controls before the birth of twins, suggesting within-group heterogeneity. Specifically, the birth of like-sex twins is preceded by low fecundity and a short period of postpartum amenorrhea. Biologically, like-sex (presumably monozygotic) twin-prone mothers have a hormonal defect related eventually to menopausal status that interferes with ovulation and perhaps with lactation. As for unlike-sex twin-bearing mothers, they probably experience a displacement of their maximum fertility potential toward early reproductive life and an extension of their menstrual life. From a methodologic standpoint, the study of the fertility of twin-prone mothers cannot proceed without estimates of the fertility components of birth intervals, as these intervals do not lend themselves to straightforward analytical interpretations by statistical analyses.