CHASE-AWAY SEXUAL SELECTION: RESISTANCE TO “RESISTANCE”

Abstract

Holland and Rice (1998) propose a model of the evolution of elaborate male traits and female preferences. A series of empirical studies has demonstrated antagonistic coevolution of the sexes in Drosophila melanogaster, in part due to the evolution of sperm toxicity and female to this toxicity (e.g., Rice 1996). Holland and Rice extend this observation to a more general model for the evolution of male and female traits. Under chase-away sexual selection, intersexual conflict over reproduction drives a cyclical process whereby males evolve traits to stimulate females into mating suboptimally (e.g., at a rate so high that it reduces female fitness). Females evolve resistance to the traits, and males then evolve novel or exaggerated traits. The model's two central assumptions are: (1) male traits seduce females into mating in a suboptimal way; and (2) selection thus favors female resistance to these traits. Empirical studies, including the ones cited by Holland and Rice, have generally not corroborated these assumptions. Nevertheless, the model makes predictions that can be readily tested in future studies. We will address the assumptions in turn, including a need for clarification of the term resistance. We then discuss the empirical evidence thus far cited in support of chase-away and propose a program for testing the chase-away model. Holland and Rice do not propose a general mechanism by which male traits can lower female fitness through inducing suboptimal mating. They concentrate primarily on male traits in a promiscuous mating system acting to increase female mating rate or extend a female's temporal window of sexual responsiveness. For either of these mechanisms to operate, successful male traits must be able to increase the number of matings per male in a labile pool of matings. For example, in a lekking species females must mate a variable number of times depending on the traits they perceive. It may be, however, that successful male traits instead act to increase the proportion of matings per male in a fixed pool of available matings. Lekking females may arrive at a lek, mate once, and then depart for that season. For selection on mating rate to favor female resistance, the former type of scenario must apply. Attractive traits must enhance reproductive success by seducing females into mating more frequently or at inopportune times, rather than by simply making their bearers more likely than other males to mate. To our knowledge, there is no evidence that sexual communication signals per se can influence female mating in this manner (excluding interactions between heterospecifics; e.g., Pierotti and Annett 1993). Female reproductive physiology can indeed be influenced by communication signals (e.g., Kroodsma 1976; Brown 1985), but there is no evidence that this is not an optimal response for the females in these specific cases. A near-universal of reproductive physiology is that females divide their time between sexually responsive and refractory periods. Most work on sexually selected traits has focused on their effects on females assumed to be responsive, rather than on their alteration of female motivational state. Many well-studied systems, including most anuran amphibians and lekking birds, appear to involve males that are competing for responsive females rather than competing for the ability to induce responsiveness. The relationship of female motivation to male signals is a topic in dire need of further study. Holland and Rice argue for a model of sexual selection driven by increased female resistance to male traits. They do not provide a clear definition of the term resistance and its relationship to preference. We distinguish between two different components of a female's preference function for a given trait: her threshold of acceptability (the minimum trait value required for mating) and her discrimination among acceptable mates (Fig. 1). The threshold is ordinarily conceived in the context of species recognition, where females only respond to species-typical values of a trait, but there is no reason that it could not lie within the range of nonspecific signals as well (see Ryan and Rand 1993). Resistance, as formulated in the chase-away model, appears to involve an increase in this threshold of acceptability. Resistance and discrimination (among acceptable mates) can change independently (Fig. 2). Increased female resistance, as defined by Holland and Rice, involves moving the threshold to the right such that only males with exaggerated trait values are acceptable. The movement of the x-intercept is effectively independent of the slope of the discrimination portion of the female's response function (Fig. 2). Experimental evidence (Berglund 1993; Hedrick and Dill 1993) suggests that these two quantities can indeed change independently. When the cost of mate choice is high, females continue mating, but change the pattern of discrimination among potential mates. If Holland and Rice meant for increased resistance to