While in college I had always believed that if I were somehow unable to pursue an academic career in theoretical physics, my next choice would have been evolutionary biology. As an indication of that interest, I briefly studied under Prof. E.O. Wilson at Harvard, and then a few years ago finally managed to write up and publish in more polished form one of the the papers I had written for him back in 1983 on the Social Darwinist implications of the traditional Chinese rural political economy.
I recently happened to be digging through my old files and came across another college paper I’d written in his seminar that might perhaps be worth republishing in its original form. This analysis suggested an obvious relationship between the typical litter size of a given species and the strength of its innate incest barrier, based upon the intuitive notion of an Effective Reproductive Value (ERV). At the time, Prof. Wilson thought my hypothesis seemed quite novel and something he hadn’t seen discussed elsewhere, and although I’d assume that the subsequent 38 years of academic scholarship in evolutionary theory has filled that gap, I’d certainly be interested if someone can provide some such source references.
Incest is an emotional and important topic among human beings. To a sociobiologist, this fact is not surprising: incest, being a subcategory of reproductive behavior, has a potent impact upon genetic fitness, hence the act and the emotions relating to it are shaped by powerful selective pressures.
Incest is also an important topic among sociobiologists. Incestuous matings greatly increase the probability that offspring will be homozygous in malfunctioning genes and therefore less likely to survive and reproduce.For a discussion of the harms of inbreeding and a survey of the literature, cf. Dobzhansky (1962) 146-150; Wilson (1975) 78-80. A sociobiologist would predict the evolution of human behavioral mechanisms to suppress incest; and case studies of Israeli kibbutz children and Taiwanese child-brides have confirmed the existence of a strong innate (as opposed to cultural) aversion to sexual relations between children raised in the same family. This fact constitutes one of the strongest empirical validations of human sociobiological analysis, and serves to explain the strong incest taboo found in nearly all human societies as a cultural amplification of an innate biological mechanism.For a discussion of this phenomenon, cf. Wilson and Lumsden (1981) 138-141.
Yet incest barriers remain a curiously ambiguous support for sociobiology. Incest is obviously genetically harmful to all organisms, not just humans; but, as any pet owner can testify, domestic dogs and cats have little if any aversion to mating with their parents, siblings, or offspring. The standard explanation presented by sociobiologists is that domesticity constitutes an unnatural environment in which behavior patterns sufficient to prevent inbreeding in the wild (such as the expulsion of adolescents from the family group) are unable to function.
This argument is weak on two grounds. First, it ignores the fact that canines and felines have been subject to domesticity for several thousand generations, long enough for new behavioral structures to have evolved. Second, it does not explain why incest barriers in humans should be so much stronger and more direct than those in dogs and cats. In the primitive state, young human males may have been subject to a certain amount of wanderlust, while young females may have been traded between bands; such mechanisms would prevent much inbreeding, and the need for an extremely powerful “hard-wired” barrier seems unclear. It is the purpose of this paper to suggest a more logical and more convincing explanation of this anomaly using the notions of effective reproductive value and litter size.
Effective Reproductive Value (ERV) and Litter Size
Having one child is good; having twenty is not twenty times as good. On this point, both human women and sociobiologists are in agreement. Past a certain point, each additional offspring means (for species with high post-natal parental investment, especially birds and mammals) just another helpless infant to feed and protect. Doubling the size of a litter may mean doubling the parental investment but actually decreasing the reproductive return if spreading too little food among too many hungry mouths results in general malnutrition, with few of the offspring surviving into reproductive maturity.This whole issue is thoroughly discussed in Lack (1954); Daly and Wilson (1978) 26-29 contains a good summary of the argument and a survey of the literature.
This simple intuitive fact may be described more precisely by defining an “average net effective reproductive value (ERV)” function V as a function of litter size N. V(N) would be composed of a positive term proportional to the average number of offspring in the litter which reach sexual maturity and successfully reproduce, and a negative term which includes the cost to the parent of producing and caring for the offspring. A typical k-selected (high parental investment) net effective reproductive value curve might look like:
We would expect an organism to evolve towards producing a litter of size N.opt, the optimal value point or the point of diminishing returns. This hypothesis has been confirmed by a wide range of empirical studies.Cf. Lack (1954); Daly and Wilson (1978) 26-29; Fuchs (1982).
One feature which we might expect to find in many such net ERV curves would be a near plateau in the vicinity of the central maximum due to the buffering effect of increased parental investment. For example, if a typical (hence optimal) feline litter contains five kittens of which usually only two survive to age six months, a four kitten litter might very likely also yield two survivors on average because of the increased food and care available to each kitten. Four kittens might have (say) 95% of the ERV of five kittens rather than the 80% a strictly linear value function would predict.
Obviously the above analysis is purely speculative and theoretical. The exact nature of a given species’ effective reproductive value function is generated by a complex interaction of environmental and biological factors, and can only be determined empirically. However, the logical case for a buffered plateau around the central maximum seems quite plausible.
More decisively, the existence of such a plateau is generally supported by the little empirical evidence available. Visible ERV plateaus seem to be present in the plotted data for the Boat-Tailed GrackleDaly and Wilson (1978) 28. and the Great Tit,Daly and Wilson (1978) 29. two bird species whose breeding success/clutch size characteristics have been numerically analyzed. Among other bird species such as hawks, owls, storks, crows, buzzards, and common swifts, it is the overwhelming rule for one or more nestlings (the “runts of the litter”) to die of starvation during most years (during years in which food is especially plentiful, they are raised to maturity); these appear to constitute a “reserve army” of offspring, to be supported during good years and cast-off to starve during average or bad. The presence of “throw-away” offspring implies the existence of an ERV plateau.
Among mammalian species, the evidence for an ERV plateau is also common. A study of the litters of laboratory Guinea Pigs shows that the average number of offspring weaned per litter of 3 is 80% of that per litter of 5 (the optimum), 36% higher than a simple linear ERV would predict. Furthermore, if we assume that average weight at weaning is proportional to future survival/reproductive success, then the ERV becomes virtually constant around the central plateau:Lack (1954) 46.
at Weaning (gm)
(% of Maximum)
That such a smooth ERV plateau occurs under laboratory conditions is very remarkable. In the natural state, we would expect limited parental resources of food and security against predation to generate a much smoother ERV plateau among most species.
Controlled litter size experiments performed on house mice seem to yield similar results. Survivorship, weight at weaning, and future reproductive fitness of offspring all fall with increased litter size. Furthermore, in especially large litters, the rate of maternal infanticide and cannibalism of offspring rises dramatically; such behavior very likely serves to conserve maternal resources (and augment them by “recycling” surplus young).Fuchs (1982) 40-43. All of these results are completely consistent with the existence of a broad ERV plateau, and indeed would probably be enhanced in the natural state.
From the above theoretical and empirical evidence, it seems likely that an ERV plateau is found in a wide range of k-selected species. Below, we shall see that such species would be considerably buffered against the deleterious effects of incestuous matings.
The Impact of Incest
The harms of incest are two-fold. There is some loss of genetic variation, which may make it less likely for future offspring to possess the genes necessary for success in a changed environment. But far more immediate and significant are the harms directly associated with homozygous gene pairs, sometimes relative to a superior heterozygous state (as in the case of sickle-cell genes)In tropical regions, the heterozygous sickle-cell state is substantially superior to the normal state because it provides a degree of protection from malaria. Cf. Dobzhansky (1962) 152. but more often absolute (as in the case of individuals homozygous in a mutated and non-functional gene important to survival). Among humans, these direct harms may include increased likelihood of stillbirth, blindness, dwarfism, and severe deformities;Cf. Wilson (1975) 78. and this is probably the same for most organisms. In the natural condition, an infant of any species born with such handicaps would have virtually no chance of surviving and reproducing.
To a species with a very small litter-size—notably homo sapiens with an average litter of one infant—incest is a true disaster. A study of the results of r=1/2 (parent-child or sibling-sibling) incest among human women showed that almost half of the infants born either died within the first year of life or exhibited crippling defects; children born to these same women through non-incestuous relations exhibited a (normal) 10% rate of early death or disability.Cf. Wilson (1975) 79. Human incest therefore means a 50-50 chance that the infant produced will be defective and an entire reproductive cycle will have been wasted.
On the other hand, the negative consequences of incest will be substantially buffered in the case of a species with a large average litter size and a plateau in its effective reproductive value function. Losing several offspring very soon after birth would decrease the total ERV of the litter only slightly. This would be expecially true in those species (such as mammals) in which the bulk of the parental investment is post-natal and pro-rated: among house mice, the energetic cost of producing milk to feed a pup is over three times as great as the energetic cost of the producing the pup itself, hence losing a pup or two because of incest would mean only a very small lost investment.Fuchs (1982) 49. Whether this argument applies to domestic canines and felines and explains their very weak incest barriers can only be determined by controlled breeding experiments (see following section); but it seems a plausible hypothesis.
Experimental Suggestions and Predictions
This section is perhaps the most important one. Far too much of modern sociobiology has been concerned with highly abstract (and suspect) theoretical reasoning unbuttressed by empirical evidence. Fortunately, the general predictions made in this paper should be fairly easy to confirm or contradict experimentally.
(1) Among most higher organisms there should be a loose inverse correlation between litter size and inbreeding aversion. In particular, very strong incest barriers should exist in species possessing the following characteristics: (a) an average litter size of one; (b) intelligence high enough for individual recognition to be possible; and (c) small family-groupings in which siblings remain together for at least several years and must be prevented from developing future sexual attachments. (Animals living in large herds or with very little sibling contact should exhibit weaker incest aversion, since the randomizing mating pattern would largely negate the risk of incest.) Prime candidates for high incest barriers would include elephants, bears, higher primates, and man.
(2) The net reproductive value of domestic feline and canine litters resulting from incest should be close to the net reproductive value of normal litters. This could easily be determined by very simple controlled breeding experiments.
(3) In general, it might be very enlightening to attempt to determine the nature of the effective reproductive value functions for canine, feline, and other species by removing selected numbers of offspring from normal litters and then recording the average number of remaining offspring that successfully reach maturity. It may be possible to establish some relation between these species-specific functions and general ecological or taxonomical characteristics.
The prediction of an inverse relationship between incest barrier strength and litter size, if confirmed by experimental data, should further establish the usefulness of the sociobiological analysis of animal behavior. Furthermore, the determination of species-specific effective reproductive value (ERV) functions might open new and interesting lines of sociobiological research.
M. Daly and M. Wilson, Sex, Evolution, and Behavior, 1978.
Dobzhansky, Mankind Evolving, 1962.
S. Fuchs, “Optimality of Parental Investment,” in Behavioral Ecology and Sociobiology, pp. 39-51, 1982.
D. Lack, The Natural Regulation of Animal Numbers, 1954.
E.O. Wilson, Sociolobiology, 1975.
E.O. Wilson and C.J. Lumsden, Genes, Mind, and Culture, 1981.
 For a discussion of the harms of inbreeding and a survey of the literature, cf. Dobzhansky (1962) 146-150; Wilson (1975) 78-80.
 For a discussion of this phenomenon, cf. Wilson and Lumsden (1981) 138-141.
 This whole issue is thoroughly discussed in Lack (1954); Daly and Wilson (1978) 26-29 contains a good summary of the argument and a survey of the literature.
 Cf. Lack (1954); Daly and Wilson (1978) 26-29; Fuchs (1982).
 Daly and Wilson (1978) 28.
 Daly and Wilson (1978) 29.
 Lack (1954) 46.
 Fuchs (1982) 40-43.
 In tropical regions, the heterozygous sickle-cell state is substantially superior to the normal state because it provides a degree of protection from malaria. Cf. Dobzhansky (1962) 152.
 Cf. Wilson (1975) 78.
 Cf. Wilson (1975) 79.
 Fuchs (1982) 49.