| Abstract: | Reproductive success is widely used as a measure of fitness. However, offspring quantity may not reflect the genetic contribution to subsequent generations if there is nonrandom variation in offspring quality. Offspring quality is likely to be an important component of human fitness, and tradeoffs between offspring quantity and quality have been reported. As such, studies using offspring quantity as a proxy for fitness may yield erroneous projections of evolutionary change, for example if there is little or no genetic variance in number of grandoffspring or if its genetic variance is to some extent independent of the genetic variance in number of offspring. To address this, we performed a quantitative genetic analysis on the reproductive history of 16,268 Swedish twins born between 1915 and 1929 and their offspring. There was significant sex limitation in the sources of familial variation, but the magnitudes of the genetic and environmental effects were the same in males and females. We found significant genetic variation in number of offspring and grandoffspring (heritability = 24% and 16%, respectively), and genetic variation in the two variables completely overlapped—i.e., there was a perfect genetic correlation between number of offspring and grandoffspring. Shared environment played a smaller but significant role in number of offspring and grandoffspring; again, there was a perfect shared environmental correlation between the two variables. These findings support the use of lifetime reproductive success as a proxy for fitness in populations like the one used here, but we caution against generalizing this conclusion to other kinds of human societies.Measuring selection and projecting evolutionary change, including in contemporary human populations (1), relies on validly measuring fitness (i.e., the genetic contribution to future generations). Fitness is usually measured by a metric of reproductive success, i.e., offspring quantity (1, 2). However, offspring quantity may be a poor proxy for fitness when there is nonrandom variation in the reproductive quality of offspring (ref. 3; e.g., due to differences in offsprings’ viability, attractiveness to mates, or intrasexual competitive ability). For example, a female might have few offspring but increase their reproductive quality (and the female’s own fitness) by investing parental care and resources in the offspring, by choosing a mate who invests in the offspring (4), and/or by choosing a mate whose superior (5) or more compatible (6) genetic makeup improves the genetic quality of the offspring. A second female might have more offspring but fewer grandoffspring (and so lower fitness) if she and her mate(s) confer lesser material or genetic benefits to her offspring. The same of course applies to males.Given humans’ exceptionally slow life history (∼15 y to sexual maturity) and high degree of biparental investment in offspring, the quality of those offspring is likely to be an important component of fitness in humans (7, 8), and extended parental investment improves quality of offspring in terms of their reproductive success (9). Research from preindustrial societies provides evidence for a tradeoff between offspring quantity and reproductive quality (e.g., refs. 2, 8, and 10–12), and there is evidence in postindustrial societies that offspring quantity is associated with lower parental investment in each offspring (13) and with detriments in offspring quality measures such as intelligence (14) and childhood growth (15) (see ref. 16 for a review of quantity–quality tradeoffs in humans). Such tradeoffs could mean that number of offspring might be a misleading indicator of longer-range (i.e., better) measures of fitness, e.g., number of grandoffspring.Evolutionary change in a trait (i.e., the shift in population mean over generations) due to selection depends on the trait’s genetic covariation with fitness (17–19). In this way (i.e., using the Robertson–Price identity), recent high-profile studies have projected evolutionary change in human traits (20, 21). However, because they used number of offspring to measure fitness, the projected magnitude or direction of evolutionary change could be wrong. For example, although previous research has revealed genetic variation (39% of the total variation) in number of offspring (22), there might be little or no genetic variation in number of grandoffspring, which would yield little or no long-term evolutionary change. Alternatively, if there is genetic variation in number of grandoffspring, it might not be captured by the genetic variation in number of offspring (e.g., because of the genetic variation in traits relating to maternal investment, mate choice, or mate retention), which would affect the magnitude or direction of the genetic covariation with the trait. However, it could be that the genetic variation in number of grandoffspring completely overlaps (i.e., rg = 1.0) with the genetic variation in number of offspring, which would validate using number of offspring as a measure of fitness.The classical twin design uses the greater genetic similarity of identical twins (100%) compared with nonidentical twins (50%) to partition traits’ variance and covariance into genetic and environmental sources. Here we examine Swedish twins born between 1915 and 1929 (n = 16,268) and their number of offspring and grandoffspring born, which, for the vast majority of the sample, reflect lifetime reproductive fitness in both generations (Methods). We estimate the genetic variation in these variables and assess whether there are genetic influences on number of grandoffspring that are independent of the genetic influences on number of offspring. |
