It is well established that the dizygotic (DZ) twinning rate increases with maternal age until about age 38 and thereafter declines, and that within each maternal age category, it increases (apparently indefinitely) with parity [5]. Interpretation of this increase with parity is not immediately clear. One may formulate two hypotheses, one or both of which may be true.Hypothesis 1: Twin-prone women are more fertile than other women and therefore they form a progressively increasing proportion of the mothers at each birth rank and/or Hypothesis 2: Each birth somehow adds an increment to a woman's probability of bearing a DZ twin pair at her next pregnancy.Allen [1] tested the latter hypothesis in the following way. For a group of sibships all of the same size, and all of which contained a pair of DZ twins, he considered only those pregnancies occurring in a given limited age range, e.g., 25-29. The test consisted in observing whether the twin maternities were then related to birth order within this age range. He found that they were. Accordingly, he drew the inference that the second hypothesis was true.This led him [2] to question the evidence which had seemed to support hypothesis 1. Among the sorts of evidence he considered were the prompt postmarital conceptions [4,9] and prompt postwar conceptions [3,8] of DZ twins. He found that within the reproductive histories of women bearing DZ twins, the conception delays leading to first-born DZ twin pairs were less than those leading to first-born singletons. Though he admits that the difference was not significant, he writes: "It probably means that prompt conception is not a constant characteristic of parents of twins; their chance of early conception need not be any greater than that of other couples." It seems to me that (even if the difference were significant) the argument is misleading. The contrast that would be interesting is between the conception delays preceding first-born singletons in DZ twin sibships and the first conception delays in other sibships. If there were no difference between them, then Allen's suggestion would be vindicated.In support of his case, Allen writes [2]: "It seems as if the marriage event itself promotes twinning; and so does return of husbands from war, or an illicit love affair." 0001-5660/82/3101-02-0119S01.00