Offspring sex ratio and age of parents.

Abstract

with the hypothesis that parents of intermediate ages produce a greater proportion of female offspring than do younger or older parents. Lowe (1969) and Meadows (1969) implied that the differences Kamaljan reported were statistically significant. Further, Lowe (1969) examined data from red deer, Cervus elaphus, for such an effect, found it, and constructed fecundity tables on the assumption that females between 5 and 8 years of age produce more female than male offspring. If this effect is real and general, most current models of mammalian population dynamics will require modification. Before such a large step is taken, it might be as well to re-examine the evidence on which the hypothesis is based. Kamaljan (1962) presented sex ratios of offspring classified by age of both parents. Parents were divided into 'young', 'intermediate' and 'old', the interval of years these classes represent being given for each species in Table 1. Table 2 shows the 2 values of association between sex ratio of the offspring and the age of the parents. Each results from a 2 x 3 contingency test, the three age-classes of parents being those in Table 1. Since each test has two degrees of freedom, a 2 of at least 5-99 is required before an association is acceptable at the 5% level of significance. Of the ten examples, only one comparison (age of male parent against sex ratio of first-born offspring in man) reached the required level. In each of Kamaljan's examples, parents of intermediate ages produced offspring whose sex ratio was biassed further towards females than was the sex ratio of offspring produced by younger or older parents. This provides strong evidence in favour of the hypothesis, regardless of the general lack ofsignificance shown by individual examples. There is, however, an alternative explanation. When sampling variation is high consequent on low numbers, it is possible unconsciously to choose a set of contiguous ages ofparents whose offspring have a sex ratio deviating from the mean, as may have happened here. Table 1 shows inconsistency in Kamaljan's choice of an intermediate-age set. That for male cattle is narrower than for females, in pigs it is broader for males, and in horses the intermediate age interval is the same for both parental sexes. Even so, the subjectivity evident in the choice of age intervals should have little effect if the number of offspring is large. But, in the present case, subjectivity of selection cannot be ignored as a possible influence on the results: the ranks of