A reanalysis of the data presented in a recent large twin study suggests that opposite-sexed (OS) pairs may be not as exhaustively tested as same-sexed (SS) pairs on genetic markers. This is contrary to an assumption I made in estimating that there are about 8 SS dizygotic pairs to every 7 OS pairs (thus impugning Weinberg's differential rule). If this assumption is false also in regard to the samples I discussed, then that estimate is unsound and Weinberg's rule is unscathed by empirical data. However, regardless of such considerations, there are strong theoretical reasons for questioning the status of Weinberg's rule. It is based on two assumptions, namely that p (the probability that a dizygotic twin zygote is male) is equal and independent for all dizygotic twin zygotes. Data are adduced here to suggest that both assumptions are false. The upshot is that, at present, without testing, we cannot know, of any given population, whether the rule holds or not. Otherwise, though Weinberg's rule may be useful (like Hellin's law) as a rule-of-thumb, it cannot be assumed as a basis for serious scientific argument.

Based on Dutch twin incidence figures since the beginning of the current century, evidence is provided in support of the idea that the DZ unlike/like sexed ratio has gradually shifted (since 1900) from unity to less than unity. Opposing conclusions with regard to the justification of the use of Weingberg's differential rule are very probably correct in themselves but could depend on country and period of birth of the twin sample used. Furthermore, the fast drop and subsequent rise in DZ twinning rate between about 1963 and 1990 can very likely for the greater part be ascribed to a parallel shift in maternal age.

The usual assumption that monochorionic twins are monozygotic has recently been questioned, based on blood group discordance in 3 of 12 monochorionic pairs. Therefore, this study evaluates the validity of zygosity diagnosis based on examination of placental membranes, and at the same time evaluates Weinberg's differential rule in a Danish consecutive twin series. All twin pairs, 265 like-sexed and 87 unlike-sexed pairs, born at Odense University Hospital, Denmark, from 1 January 1980 through 31 August 1988, were ascertained. In 229 like-sexed pairs, zygosity was determined by serological methods, and in 190 of these, the fetal membranes could be reliably classified. Fiftynine monochorionic twin pairs were identified and none of those pairs were discordant for any genetic markers. Therefore, we conclude that monochorionic twins can be considered monozygotic. The distribution of monozygotic and dizygotic pairs was in accordance with Weinberg's rule.

Most population studies of twins estimate the number of monozygotic (MZ) and dizygotic (DZ) pairs by Weinberg's differential rule. This rule assumes that within the DZ twins the numbers of unlike-sexed (U) and like-sexed (L) twins are equal. The literature on the validity of Weinberg's rule is still controversial. In this prospective population-based study (EFPTS) of 2,589 twin pairs, of whom 2,577 were of known zygosity and placentation, the estimates of Weinberg's rule agree well with the results of direct zygosity determination.

Weinberg's difference method, applied to twin birth statistics, usually shows a dependence of the MZ rate on maternal age, like a thin shadow of the DZ rate. Some of this MZ variation could be explained away by James' finding of more same-sex (SS) than opposite-sex (OS) DZ twins, the excess being mistakenly classified as MZ by Weinberg's assumption of equal numbers. By several methods one can extract a Constant value for the MZ rate and a Constant or nearly Constant value for the DZ SS/OS ratio, but these “constants” are actually arbitrary and they vary between populations.