Sexual Selection with a Negative Effect on Fitness: Can it be Reversed?

A Scientific Grant Proposal, subtitled


This experiment will deal with a population of males expressing a sexually selected trait with a physiological (Kotiaho, et all 1998) fitness cost that results in an increased mortality rate. The trait is handicapping and so indicates that males who express it carry good genes. Males who express the trait are referred to as attractive, because in the experimental species, females select mates with good genes based on their expression of the handicapping trait. And females who mate with attractive males will have more attractive sons, and as a result, through a positive feedback loop of coevolution, more grandchildren (Brooks 2000). However, because of their significant handicap, the male offspring of females that mate with attractive males are at a higher mortality risk than those of females who mate with unattractive males. In the experimental species, there is a pre-existing survival differential favoring the offspring of those females who do not select attractive mates (Møller and De Lope, 1994).

Houle and Kondrashov (2002) have created a model of costly good genes display which predicts eventual extinction for a population in which good genes are constantly sexually selected for, even when their expression is costly to the point of certain fatality. However, their model did not incorporate the heritability of female mate choice behavior, and the selective pressure favoring those females who produce the most offspring. Wiens (2001) has used phylogenetic techniques to examine species showing a probable historic loss of a handicapping secondary sexual trait and found tentative evidence that a reversal in female selective behavior may have led to handicapping trait loss in species of swordtail fish, ducks, and flycatchers. This experiment will attempt to drive female mate selection patterns away from a reliable preference for handicapped males, using a number of natural selection regimes approximating the fitness cost of that handicap.


Hirundo rustica, or the barn swallow, is ideal experimental species for this experiment because
  1. It exhibits an easily quantifiable trait that is sexually selected for as well as naturally selected against, and
  2. The rate at which this trait is naturally selected against has already been empirically determined (Møller and De Lope, 1994).
350 barn swallow eggs will be collected from wild populations in central New Your State. In order not to have too large an impact on any one population, these will be taken from at least 5 different colonies of birds. No more than one egg will be taken from a nest that contains four, and no more than two eggs will be taken from a nest that contains five. Eggs will be taken instead of adult animals because barn swallows are monogamous, and if animals were collected that had previously mated, the disruption in their life cycle might skew their mate choice in the laboratory, i.e. females might prefer to mate with individuals resembling their previous mates. In addition, the experiment seeks to observe changes in heritable patterns of mate choice, and removing precognitive eggs from their parents’ nests will eliminate the possibility that females choose mates that resemble their fathers. Only 320 individuals are needed, but the initial population needs to consist of a equal number of males and females, and the excess number of eggs captured will make sure that the necessary number of individuals of either sex are obtained.

The eggs will be hatched and reared using the technique described in the Procedure section. As nestlings they will be housed in incubators, and then transferred as adults to 8 aviaries. The exact feed ratio for nestlings and juveniles are yet to be determined but it will consist of a mixture of ground insects. Adults will be fed live flying insects, which will be introduced into the aviaries in great excess, so that there is no selection for between individuals for optimal aerodynamics or hunting skill.


The initial population of eggs will be raised as described below, and the nestlings will be sexed using the molecular, DNA marker technique described by Griffiths and Tiwari (1993). From these 400 initial nestlings, 4 experimental groups will be generated, numbered A through D. Each group will consist of 40 randomly allocated males and 40 randomly allocated females. The remaining 30 individuals will be released.

Each group will be bred for 5 generations, using the following experimental manipulations of the swallow life cycle:


The eggs will be incubated and hatched, and the young will be fed via a number of robot decoy adult heads that will be programmed to release a small bolus of pulverized insect matter into the mouth of a begging bird. (The viability of this set up will be tested with a small number of young prior to the primary experiment. If the young will not feed from the decoy heads, then a new experimental design will be utilized, with the nestlings being raised by their parents, and the heritable effects of female mate choice not being tested directly.)

Nestlings will be sexed, and juvenile male and female birds will be separated and house apart from each other. Food will be provided ad libidum. Upon reaching adulthood all of the swallows will be fitted with numbered tags. The males’ tail lengths will be measured and recorded. For the first generation of males, a median tail length will be calculated, and the males will be divided into those with tail length above the median (designated “long”) and those below it (“short”). This median length L, from the first generation, is a figure conserved through all experimental generations. Individuals in generations 2 through 5 will be designated long or shot according to their relation to L and not according to their own median length.


Natural selection will be imposed by randomly culling the following fractions of males in each group. Every group will undergo the same overall rate of selection, with 57.8% of all males culled. However, the groups will differ in the selective differential between tail lengths. In the control group, long- and short-tailed males will undergo the same negative selective pressure. At any point that animals are said to be “culled” they will merely be removed from the breeding population and released into the wild.
  • Group A will undergo differential selection favoring short tail length at the rate observed by Møller and Lope: 71.1% of long-tailed males and 50.6% of short-tailed males will be culled.
  • Group B will undergo selection at the same overall rate, but with a stronger skew towards culling long-tailed males. 78.1% of long-tailed and 43.6 of short-tailed males will be culled.
  • Group C will undergo selection at the same overall rate, but with a weaker skew towards culling long-tailed males. 64.1% of long-tailed males and 57.6 of short-tailed males will be culled.
  • Group D will undergo selection at the same overall rate, but with no skew towards culling long-tailed males. An identical proportion, 61.85%, of both long-tailed and short-tailed males culled.

The animals remaining in the mating pools of each group of males will then be presented to females taken from their generation and group- “A” males to “A” females, “B” males to “B” females, etc. However, at this step each group contains far fewer males than females. In order for meaningful mate choice to occur, the number of females in the mating pool will be reduced to 2/3 the number of males. The remaining females will be released. Those females that do mate will be introduced to the entire group of males one at a time so that each female’s mate choice behavior can be recorded. The ID numbers of the males that a given female had to chose from, as well as the ID of the male that she chose, will be recorded. After each female has had the opportunity to chose a mate, those males that are not chosen to be mated with will be released. Male and female pairs will then be allowed to lay and hatch their eggs, but prior to hatching these eggs will be removed and the adult birds released. The eggs will then be incubated, hatched and raised, and the offspring will be taken through the experimental life cycle already described.

Because barn swallows produce an average of 4.5 eggs yearly, generation 1, which initially will contain 80 members and after culling and mate selection is reduced to 16.8 breeding pairs, can be expected to produce 75.6 offspring. However, if sexual selection is strong enough in any of the populations to drive up average tail length, then the number of males culled from each generation will increase and offspring production will fall. In all cases the population of females will not be tampered with, and at each generation the females (as well as the males) will be released after being mated one time. The eggs from the fifth generation will be hatched, reared, culled, and mated, at which point female mate choice will be observed a final time and the experiment will end.

Hypothesized Results

In this experimental set up, females with the stronger selective behavior for long-tailed males will produce fewer offspring: between 61.85% and 78.1% of their male offspring are culled, as opposed to between 43.6% and 61.85% of the male offspring of females who select for short-tailed males. Thus, if there is any variation in the mating behavior of the females of the experimental populations, then over five generations of selection the observed female preference for long-tailed males should decrease, at least in the population in which long-tailed males experience the most stringent negative natural selection.

It is possible, however, that there is no such pre-existing variability in female reproductive behavior. In this case the experimental population will be seen to decline: As the larger proportion of the male offspring in each generation will have long tails, a larger proportion of the male offspring in each generation will be culled. A stable population could only result from this model by the understanding that a population can afford to lose up to 78.1% of its male offspring, due to the large numbers of offspring produced. Group population figures and tail lengths are also of interest and should be recorded, for each generation, before and after culling.

  • Brooks, R. 2000. Negative genetic correlation between male sexual attractiveness and survival. Nature. Vol. 406, pp. 67-70.
  • Griffiths, R. and B. Tiwari. 1993. The isolation of molecular genetic markers for the identification of sex. Proceedings of the National Academy of Sciences of the United States of America. 90 (18) : 8324-8326.
  • Houle, D. and A.S. Kondrashov. 2002. Coevolution of costly mate choice and condition-dependent display of good genes. Proceedings: Biological Sciences. Vol. 269, pp. 97-104.
  • Kotiaho, J., R.V. Alatalo, J. Mappes, S. Parri, and A. Rivero. 1998. Male mating success and risk of predation in a wolf spider: a balance between sexual and natural selection? Journal of Animal Ecology, Vol. 67, pp. 287-291.
  • Møller, A.P. and F. De Lope. 1994. Differential costs of secondary sexual character: An experimental test of the handicap principal. Evolution, Vol. 48, pp. 1676-1683.
  • Wiens, J.J. 2001. Widespread loss of sexually selected traits: how the peacock lost its spots. TRENDS in Ecology and Evolution. Vol 16., pp. 517-523.

Log in or register to write something here or to contact authors.