The Society for Integrative and Comparative Biology
Chance, Time Allocation, and The Evolution of Adaptively Flexible Sex Role Behavior1
1 Institute of Ecology, University of Georgia, Athens, Georgia 30602
2 Department of Plant Biology, University of Georgia, Athens, Georgia 30602
 |
SYNOPSIS |
|---|
TOP
 SYNOPSIS INTRODUCTIONCLASSIC IDEAS IN SEXUAL...CHANCE EFFECTS ON TIME...CHANCE, TIME, AND AN...DYNAMATE AND ADAPTIVE...TESTING THE MODEL OF...POTENTIAL REPRODUCTIVE RATES AND...CAUSE OR CORRELATION IN...COERCION AND MANIPULATION VIA...WHY ARE THE SEXES...References |
|---|
An alternative to classic sexual selection hypotheses for sex differentiated pre-mating behavior is that time available for mating—as individuals experience it—along with fitness differences among alternative potential mates, induces choosy versus indiscriminate mating behavior. This alternative hypothesis says that selection has acted so that all individuals flexibly express fitness-enhancing choosy, indiscriminate, and competitive mating behavior, induced by time-varying life histories, environmental and social cues. Key predictions of DYNAMATE, the formal model of adaptively flexible sex role behavior of individuals of both sexes within dynamically changing populations, include: (1) All individuals regardless of sex assess likely fitness outcomes from mating with alternative potential mates before expressing choosy or indiscriminate behavior. (2) Males and females express adaptively flexible, choosy and indiscriminate behavior so that individuals may change their behavior—from moment to moment—to fit dynamically changing circumstances. (3) Indiscriminate behavior of males and (4) choosy behavior of females would often be maladaptive even in species with greater female than male parental investment, when females have longer latencies to receptivity to re-mating than males, and when the relative reproductive rate of males is greater than in females. (5) Whether or not females show choosy behavior will not affect whether or not males exhibit choosy or indiscriminate behavior, and vice versa. (6) When other model parameters are equal, the proportion of individuals of a given sex expressing choosy or indiscriminate mating behavior is a function of the distribution of fitness ratios (a distribution of all fitness differences that would be conferred on an individual by mating with any two sequentially or simultaneously encountered alternative potential mates). (7) Whether same-sex individuals behaviorally compete is a function of the fitness that would be conferred if the strategist won access to a potential mate, but not a function of relative reproductive rate or its proxy, the operational sex ratio. We call for re-evaluation of sex differences in choosy, indiscriminate, and competitive behavior under strong experimental controls that level the ecological playing fields of males and females, i.e., under experimental conditions informing the mechanisms of phenotypic expression. We end with comments on the classic question of questions: why are the sexes as they are?
|
INTRODUCTION |
|---|
|
TOP
SYNOPSISINTRODUCTIONCLASSIC IDEAS IN SEXUAL...CHANCE EFFECTS ON TIME...CHANCE, TIME, AND AN...DYNAMATE AND ADAPTIVE...TESTING THE MODEL OF...POTENTIAL REPRODUCTIVE RATES AND...CAUSE OR CORRELATION IN...COERCION AND MANIPULATION VIA...WHY ARE THE SEXES...References |
|---|
Hypotheses explaining how sex roles (sex differences in choosy, indiscriminate, and competitive behavior) are determined continue to be debated. Do genes alone determine sex differences ("nature") in behavior? Do culture, learning, and socialization experienced during development induce sometimes-arbitrary and/or maladaptive sex role differences ("nurture") in behavior? This paper is about DYNAMATE, a quantitative model of a third alternative, that nature is inextricably intertwined with nurture in the production of behavioral phenotypes (West-Eberhard, 2003
). DYNAMATE assumes that individuals are fixed (genetically invariant) so that they can express adaptively flexible behavior induced by environmental and social differences individuals experience. DYNAMATE simulates individual behavior, mating success, reproductive success, and survival of individuals in populations. Its predictions are about flexible behavior and its real-time fitness effects in demographic and life history contexts. Its most important rules differ from those of classic sexual selection, and allow comparison of DYNAMATE's predictions to those predicted by classic sexual selection.
To explain the new rules of DYNAMATE, we begin the paper with a review of how sexual selection operates through mating success variances. Second, to introduce the idea that the time available to an individual for mating (Fig. 1) is a critical determinant of mating success—an essential of DYNAMATE—we briefly review two largely overlooked models of chance effects on within-sex mating success variances (MSV) (Sutherland, 1985a) and lifetime mating success variances (LMSV) (Hubbell and Johnson, 1987). These models showed that to estimate the opportunity for sexual selection non-heritable components of MSVs and LMSVs should be subtracted off of estimates of total MSV and LMSV (Fig. 2). Readers familiar with classical sexual selection and time allocation effects on MSV and LMSV might skip the first two sections. Third, we describe Hubbell and Johnson's (1987) ESS model of choosy and indiscriminate behavior, which said that mating success variances set the stage for the expression of choosy and indiscriminate mating, a claim contrasting those of classical sexual selection (Fig. 3). DYNAMATE evolved from this ESS model, and it uses the logic of the "switch point" theorem from the ESS model. Fourth, we describe DYNAMATE, a formal quantitative model of the hypothesis (Fig. 4) that selection has operated such that ecological, social, and demographic conditions induce individuals, regardless of their sex, to flexibly and adaptively express choosy or indiscriminate and competitive behavior. DYNAMATE's stochastic simulations attempt to capture the real-life contingencies, including non-heritable (random or fixed) and heritable factors that individuals experience when making reproductive decisions (Fig. 5). Should individuals wait—be choosy, or mate-as-they-encounter potential mates—be indiscriminate (Figs. 6–8)? Should individuals compete over access to a potential mate (Figs. 9, 10)? What are the fitness payouts of these options for individuals? We discuss the quantitative predictions of DYNAMATE, contrast quantitative predictions with other conceptual models that make similar, but qualitative predictions, and discuss some tests and implications of DYNAMATE. To increase the transparency of the paper, readers may want to review the Figures first.
|
|
|
|
|
|
|
|
CLASSIC IDEAS IN SEXUAL SELECTION |
|---|
|
TOP
SYNOPSISINTRODUCTIONCLASSIC IDEAS IN SEXUAL...CHANCE EFFECTS ON TIME...CHANCE, TIME, AND AN...DYNAMATE AND ADAPTIVE...TESTING THE MODEL OF...POTENTIAL REPRODUCTIVE RATES AND...CAUSE OR CORRELATION IN...COERCION AND MANIPULATION VIA...WHY ARE THE SEXES...References |
|---|
Ever since Bateman (1948)
, sexual selection, i.e., selection via reproductive competition—Darwin's (1871) male–male behavioral competition and female choice—has been associated with within-sex variance in mating success, where "mating success" is defined as the number of mates an individual has during a reproductive bout or over its lifetime. In this section we stress the importance of within sex mating success variance (MSV) to sexual selection.
Darwin's challenge
Early critics of natural selection challenged Darwin (1859), when they noted that the evolution of many bizarre and elaborate traits, usually in males, decreased their bearers' survival probabilities and therefore natural selection (survival selection) could not explain their evolution. In response, Darwin (1871) said that sexual selection, a type of natural selection having only to do with reproductive competition, acting through male–male behavioral contests and female choice, could account for the evolution of such traits because these mechanisms would favor males possessing traits that facilitated winning male–male behavioral contests and being chosen as mates by females. Darwin argued that the behavior of choosy, but passive females and competitive and ardent males parsed mating success among males so that some males had much higher reproductive success than others, a selection differential that would favor the evolution of male weapons and traits attractive to females. What Darwin did not explain was why female mate choice seemed more prevalent than male mate choice and why male–male behavioral contests seemed more prevalent than female–female behavioral contests. Williams (1966), Parker (1972), and Trivers (1972) picked up that thread. They based their arguments on Bateman's (1948) results.
Bateman's experiment
Bateman (1948) reported an experiment in which he allowed small numbers of male and female Drosophila melanogaster to cohabit for three or four days and to mate. Using individuals of different phenotypes, influenced by dominantly inherited alleles, he was able to count the number of sires in a given female's brood to get an estimate of variance among females in mating success. By counting the number of times that a given male appeared in the sire list for different females, he was able to estimate the number of mates each individual had and the variance in male and female mating success. His data showed that mating success variance was greater among males than females—at least under the conditions of his tests, just what one would expect if sexual selection among males was at work. Bateman concluded that the high mating success variances in males were due to female choice and male–male competition over access to females. Ever since, between-sex differences in mating success variances have been considered an indicator of sexual selection.
Genes for choosy females and indiscriminate males
Williams (1966), Parker et al. (1972), and Trivers (1972) used Bateman's results as the starting place to explain why females so often appear to be the choosy sex and males so often the indiscriminate sex (see Gowaty [2004] for a further review). They argued that the fixed life-history differences between the sexes result in differential costs of gamete production (Parker et al., 1972) and of post-gametic parental investment, which together favored the evolution of genes in females for choosy behavior and genes in males for indiscriminate and competitive behavior (Fig. 3, top panel). They thus provided an adaptive explanation for choosy and indiscriminate behavior that Darwin left out. Bateman's paper, which reported sex differences in mating success variance showed for the first time, that mating success variances were larger in males, just what one would expect if females choose and males are competitive and indiscriminate. Bateman's paper is one of the most cited papers in the sexual selection literature.
Mating success variances and selection
Even in the face of modern criticisms of Bateman's paper (e.g., Tang Martinez, 2005), the idea remains key, because within-sex variances in mating success go to the heart of selection theory that holds that the larger the within-sex variance in fitness associated with heritable traits, the stronger the opportunity for sexual selection. And, his data demonstrated exactly the result one would expect if male–male combat over access to mates and female choice of mates operated so that some males were more successful than others in mating. Bateman did not observe behavior or, if he did, he did not report his observations. Thus, he was unable to definitively associate the observed within-sex variances with particular heritable traits. This task, left to later observers, has been a central focus of modern studies of female mate choice (Andersson, 1994; Eberhard, 1996), and male–male behavioral contests. Very few people, until recently, have questioned the significance of Bateman's results. In the next section, we briefly review the work of those critics, who initially championed questions about Bateman's results in relation to the study of sexual selection. Because DYNAMATE is a model of time allocation, the arguments in the next section are critical to an understanding of DYNAMATE.
|
CHANCE EFFECTS ON TIME ALLOCATION, MSV, AND LMSV |
|---|
|
TOP
SYNOPSISINTRODUCTIONCLASSIC IDEAS IN SEXUAL...CHANCE EFFECTS ON TIME...CHANCE, TIME, AND AN...DYNAMATE AND ADAPTIVE...TESTING THE MODEL OF...POTENTIAL REPRODUCTIVE RATES AND...CAUSE OR CORRELATION IN...COERCION AND MANIPULATION VIA...WHY ARE THE SEXES...References |
|---|
Mating success variances in the absence of choice and competition
Sutherland (1985a
) confronted Bateman's conclusion by showing that the "handling time" associated with a mating along with random variation in numbers of encounters with potential mates, in the absence of mate choice and male–male contests, could account for greater MSV of one sex compared to the other. Sutherland tested his neutral model's predictions against Bateman's data, and found that his model of random mating predicted MSVs not statistically different from Bateman's observations.
Sutherland's model is about individual time allocation (Fig. 1A). It is similar to Holling's (1959) handling time model of foraging. It assumes that over their lifespan, individuals have a certain amount of time, T, to devote to mating. T is affected by encounter rates (higher encounters mean more opportunities to mate, fewer encounters mean fewer opportunities to mate), and T includes the time individuals have to deal with the activities consequent to mating, something Sutherland called the handling time, H, when an individual is unable to seek a mate. H includes non-heritable and/ or fixed aspects of copulation, parental care, and replacing gametes. Thus, the time available for searching for mates is T minus H. The important point is that while an individual is in state H, it is losing opportunities to mate with others. Sutherland calculated the number of mates over T under the assumptions that individuals mate as they randomly encounter potential mates at rate µ. While mating with one, an individual is losing opportunities to mate with another. The relative probability of having b mates over T followed a Poisson distribution modified to include H. Holding encounter rates equal, longer H reduces the opportunities to mate compared to shorter H, so longer H necessarily yields lower MSV than shorter H (Fig. 1A).
Lifetime mating success variances in clones without reproductive competition and mate choice
Hubbell and Johnson (1987) extended Sutherland's time allocation logic to show effects on lifetime mating success variance, LMSV. To Sutherland's model they added a third parameter, s, individual survival probability per unit time (Fig. 1B). Adding survival sets a realistic constraint that individual reproductive lifespans be finite (everyone eventually dies). Also they imposed the necessary constraint that average mating success must be equal for mating females and mating males. Just as the length of search time varies because of higher or lower encounter rates, variation in lifespan can result in more or fewer opportunities to mate. Higher survival probabilities give individuals more time to mate; lower survival probabilities less opportunity to mate (Fig. 1B). The two constraints, lifespan is finite and average mating success of mating males and mating females are equal, are just facts of life that make the results of models conform more closely to reality. Thus, the variable life-time mating success among individuals in the model comes from random contingencies that individuals experience.
Hubbell and Johnson (1987) used an absorbing finite Markov chain model to derive analytical expressions for the expected lifetime mating success (LMS) of clones of males and females under chance variation in parameters affecting time available for mating. LMS is finite in the model, because each individual eventually dies and thus stops reproducing (goes to the absorbing state of death). LMS of each individual is a stochastic variable: its mean and variance depend upon several conditional transition probabilities of the individual's life history and its social environment, including encounters with potential mates, latency from the onset of one copulation to receptivity to the next (similar to Sutherland's H), and survival probabilities. Thus, within-sex lifetime mating success variance (LMSV) resulted from fixed (and therefore, non-heritable) latencies to re-mating receptivity (fixed n), a stochastic distribution of lifespans (probability s) and random encounters with potential mates (probability a). All individuals of a given sex had identical values of parameter n and probabilities s and a. Thus, the LMS model intentionally left out competitive differences among males and among females—in searching ability, in combat, in individuals' abilities to sequester or monopolize individuals of the opposite sex, in their attractiveness to the opposite sex, in the genetic composition of spatially explicit populations, etc. Like Sutherland's results, Hubbell and Johnson's depended only on fixed life history effects and chance: Heritable variation among individuals was not part of the model. Neither of these models addressed what may give rise to fixed sex differences in life histories.
The time allocation model of LMSVs worked as follows: Hubbell and Johnson assumed, as Sutherland did, that individuals at any given moment will be in different states of their reproductive cycle. They distinguished four states in the reproductive cycle of an individual. An individual can be (1) receptive to re-mating and searching for a mate, (2) encountering a potential mate, (3) actually mating, or (4) in a reproductive "time out" period, during which the individual is processing the current reproductive bout and is non-receptive to mating. Hubbell and Johnson refer to all but the encounter period as a composite, the "post-mating latency" period; for operational clarity, we note that it is the "latency from the beginning of one copulation to receptivity to re-mating" and call it "latency to receptivity to re-mating," "latency to re-mating receptivity," "latency or n." One can simulate a population of males and females using the transition probabilities to compute the numbers of times that individuals of a given sex mate over their life times. With these data, one can compute within sex LMSVs, and, then, if one wants, compare LMSVs of males and females. One might also simply use the analytical expression in Hubbell and Johnson (1987) to compute sex differences in LMSVs.
Mating success variances by themselves are inadequate for concluding sexual selection
In the context of the model of LMSV and random mating, sex differences in LMSV, when they arise, may be due to random and/or fixed (i.e., non-heritable) sex differences in life history, survival probabilities, or encounters. Such sex differences may affect individual time budgets but not affect within-sex selection (because all individuals of a sex are fixed or the within-sex variation is stochastic). In these time allocation models (Fig. 1A and B) in which mating is random (mate choice and within-sex competition do not exist), MSV results from non-heritable variation between individuals—not from variation in the ability of individuals to find, win, or compete for mates, and not from selection on traits associated with the cost of parental investment. The unavoidable conclusion was that MSVs could be entirely due to random, fixed, and non-heritable factors affecting individual time budgets (Hubbell and Johnson, 1987; Sutherland, 1985), a result with profound implications for the study of sexual selection.
The conclusion that MSVs by themselves are inadequate evidence for the operation of sexual selection was sensible (Sutherland, 1985b; Sutherland, 1987), and suggested a redefinition of the opportunity for sexual selection (Hubbell and Johnson, 1987) as the residual within-sex LMSV not ascribable to random or fixed (non-heritable) variation in encounters, survival, and latencies (Fig. 2). As far as we know, only a few investigators (Finke, 1982; Mackenzie et al., 1995) have attempted to partition MSV and LMSV into components associated with non-heritable and heritable effects. In fact, some think it may not be necessary to partition variances. For example, see discussion (p. 39) in Shuster and Wade (2003) in which they disagree with Sutherland (1985), Hubbell and Johnson (1987) and Clutton-Brock (1988). The resolution of this debate awaits empirical evaluation. However, as we discuss in the next section, partitioning of fitness variances to random, fixed, and heritable effects may have implications far beyond the usual questions about the evolution of showy traits in males.
|
CHANCE, TIME, AND AN ESS FOR CHOOSY AND INDISCRIMINATE BEHAVIOR |
|---|
|
TOP
SYNOPSISINTRODUCTIONCLASSIC IDEAS IN SEXUAL...CHANCE EFFECTS ON TIME...CHANCE, TIME, AND AN...DYNAMATE AND ADAPTIVE...TESTING THE MODEL OF...POTENTIAL REPRODUCTIVE RATES AND...CAUSE OR CORRELATION IN...COERCION AND MANIPULATION VIA...WHY ARE THE SEXES...References |
|---|
Clones and selection favoring within-sex ESSs for choosy and indiscriminate mating behavior
The most novel contribution of Hubbell and Johnson (1987)
was a within-sex ESS (Evolutionary Stable Strategy is one that is non-invasible) model for choosy and indiscriminate behavior of individuals regardless of their sex. They added to their calculations of non-heritable effects on LMSV of clones of males and females, fitness differences that strategists would experience if they mated with alternative potential mates. Their approach suggested that mating could be "non-conceptive," not resulting in viable zygotes or living offspring, because no matter how many successful matings an individual has, if the fitness conferred from the matings is zero, reproductive success would also be zero. Thus, given differences in both mating success and fitness conferred from mating with alternative potential mates, they were able to compute lifetime reproductive success (LRS), not just mating success, for individuals mating as choosy and indiscriminate. The Hubbell and Johnson ESS model was not a game theoretic model, but one in which selection favored within each sex, "choosy" or "indiscriminate" individuals, depending on the environments they experienced and their associated fitness payouts.
The switch point
In the model, indiscriminate strategists were receptive individuals that mated as they encountered potential mates; choosy strategists were receptive individuals that mated only when they encountered high quality mates. The fitness point at which LRS was equal for choosy and indiscriminate individuals defined the within-sex ESS. Note that here fitness variation is more than mating success variation. LRS is the product of mating success and actual conferred fitness from matings, a formulation that took into account the undeniable fact that some matings fail to result in fertilized eggs, zygotes, or offspring, i.e., are non-conceptive, and that, once produced, zygotes and offspring vary in quality. The fitness conferred by alternative potential mates may be represented by number of eggs laid, the number of offspring born, offspring at the age of independence, or even in increases and decreases in the strategists' viability. Under the simple conditions of their model, non-heritable variation in LMSVs and differential fitness from mating with alternative potential mates, they showed that selection would favor—within each sex—an ESS for choosy and indiscriminate behavior. Their results depended only on non-heritable variation in encounter rates (a), survival rates (s), latency to receptivity to re-mating (n), and the ratio of fitnesses (WL/WH) that would be conferred by mating with alternative encountered mates. A crucial prediction was that even in species with strong asymmetries in the energetic costs of parental investment, some individuals in both sexes would express choosy and some indiscriminate behavior. This model of time allocation remains a largely overlooked alternative to classical models of sex associated behavior (Fig. 3).
As far as we know, no one has tested the Hubbell and Johnson model. The switch point ratio predicts the behavior of descendants, whose parents experienced particular suites of a, s, n, and WL/WH. If a particular individual was choosy, but experienced environmental conditions favoring indiscriminate, selection would be against genes for choosy. Those with genes for choosy would mate less frequently and accumulate lower reproductive success than an individual with genes for indiscriminate in that circumstance. Thus, this model assumes that the ESS conditions favored genes for two strategies within each sex. We envision a test that would begin by subjecting an ancestral population to controlled variation in a, s, n, and WL/WH followed by behavioral testing of descendants to evaluate if temporal environments experienced by parents resulted in the predicted within-sex ESS distribution of choosy and indiscriminate among their offspring.
The arrow of causation was turned around
Even without empirical test, the results of the Hubbell and Johnson model were extraordinary, because it turned the arrow of theoretical causation around (Fig. 3). In the classic sexual selection models (Williams, 1966; Parker et al., 1972; Trivers, 1972), fixed life history differences and the differential costs of parental investment favored genes for choosy females and competitive males, and the expression of sex-differentiated behavior resulted in larger MSVs among males than among females. In the Hubbell and Johnson model, life-history differences in latencies to receptivity to remating (n), and chance effects on encounter rates (a) and survival probabilities (s) combined to produce non-heritable variation in MSVs, which then were a part of the environmental variation that selected choosy or indiscriminate behavior. The ESS model raised, for the first time, the question: which came first, the MSVs or the behavior?
|
DYNAMATE AND ADAPTIVE FLEXIBILITY IN SEX ROLE BEHAVIOR |
|---|
|
TOP
SYNOPSISINTRODUCTIONCLASSIC IDEAS IN SEXUAL...CHANCE EFFECTS ON TIME...CHANCE, TIME, AND AN...DYNAMATE AND ADAPTIVE...TESTING THE MODEL OF...POTENTIAL REPRODUCTIVE RATES AND...CAUSE OR CORRELATION IN...COERCION AND MANIPULATION VIA...WHY ARE THE SEXES...References |
|---|
Gowaty (2003
, 2004) extended the ESS model to populations of variable individuals, when she hypothesized that selection favored not sexes, but individuals—regardless of their sex—that respond flexibly and dynamically, moment to moment to changing social and life-history situations, as choosy or indiscriminate to enhance their current contributions to lifetime reproductive success (Fig. 4). If this hypothesis is so, what genes may influence and development mold are individuals' abilities to assess fitness differences of potential mates, sensitivities to the time they have left to reproduce, and their abilities to respond to the time-variable environmental cues, so that adaptive, flexible behavior is induced (West-Eberhard, 2003).
The formal model (Gowaty and Hubbell, unpublished manuscript) of Gowaty's (2003, 2004) flexible sex roles hypothesis is DYNAMATE, a simulation model (Fig. 5) of populations of individuals of both sexes. DYNAMATE calculates individual flexibility in behavior, MSVs, and reproductive success instantaneously using the switch point theorem from Hubbell and Johnson (1987), taking into account dynamically changing encounter rates, life history parameters, survival probabilities, and the fitness differences among potential mates.
The parameters of DYNAMATE
DYNAMATE's parameters include:
- a = encounter probability per unit time with potential mates of the opposite sex
- n = number of time units of latency from beginning of one copulation to onset to receptivity to another;
- s = probability of survival per unit time
- WL = fitness conferred from mating with a self-referentially low quality mate
- WH = fitness conferred from mating with a self-referentially high quality mate.
- n = number of time units of latency from beginning of one copulation to onset to receptivity to another;
DYNAMATE necessarily considers males and females simultaneously, because mating outcomes are constrained by what is happening with other individuals in the population. (For example, if all males die before all females die, the mating success of the remaining females will be affected). DYNAMATE can consider populations of seasonal or continuously breeding individuals. DYNAMATE can assume non-overlapping or overlapping generations.
Definitions of choosy and indiscriminate
DYNAMATE defines choosy and indiscriminate behavior as follows. Indiscriminate receptive individuals mate with all receptive individuals of the opposite sex as they encounter them. Choosy receptive individuals mate only with receptive individuals of the opposite sex that will confer on them relatively high fitness. The point of equal fitness between choosy and indiscriminate behavior is the switch point ratio (SPR) from the theorem in Hubbell and Johnson (1987). The SPR = as(1 – sn+1)/(1 – s + as[1 – sn+1]). Fitness ratios have to be small enough to be below the SPR to favor the expression of choosy behavior by individuals, and fitness ratios have to be large enough to be above the SPR to favor indiscriminate behavior by individuals. Note that on the graphs of results (Figs. 6–8) small differences between encountered potential mates are indicated by high fitness ratios (WL/WH), while large differences are indicated by small fitness ratios.
Relaxed assumptions about mate quality
DYNAMATE relaxes the ESS assumption (Hubbell and Johnson, 1987) of discreet high and low quality mates, so that there is a continuous distribution of fitnesses conferred by mating with alternative potential mates, which are high and low relative to one another and the individual assessing potential fitness consequences (i.e., not absolute fitness differences). We further assume that mate assessment by individuals of either sex is self-referential, perhaps for genetic dissimilarity and complementarity (Wedekind, 1999), so that whomever may be the best mate for one male (or female) may not be best for any other males (or females) in the population. A critically important idea in DYNAMATE is that individuals assess likely fitness rewards or deficits from mating with alternative potential mates before they express choosy or indiscriminate (competitive) behavior.
Definitions of fitness ratios
Fitness ratios, WL/WH, might be thought of as an expression of differences in fitness conferred by mating with alternative potential mates. One could imagine potential fitnesses conferred as absolutes, so that all individuals of one sex would have identical fitness conferred by mating with any given opposite sex individual. Alternatively, one might imagine that fitness conferred is relative to the pair of mating individuals, so that any distribution is relative—self-referential—to the individual strategist making a mating decision, a version of DYNAMATE that particularly interests us. DYNAMATE calculates fitness ratios in both of these ways to test "good genes" versus "complementary genes" mate preferences models. The population of fitnesses conferred can be imagined as a matrix of fitnesses that result when all males are mated to all females and all females mated to all males. These fitnesses can be expressed within a sex as a series of fitness ratios, which we constrain to vary between 0 and 1 (by always putting the smaller fitness in the numerator). The distributions of fitness ratios, whether based on absolute or self-referential rules, change dynamically in DYNAMATE as individuals enter latency or die, as they are no longer available to mate. The distribution of fitness ratios can be estimated as a characteristic of a population, but in DYNAMATE the distributions dynamically change as a function of which strategist is making the assessment.
Assessment rules
DYNAMATE accommodates different assessment rules. DYNAMATE says that all individuals assess fitness outcomes before deciding to be choosy or indiscriminate, but the model is agnostic about exactly how the assessment occurs. We have designed DYNAMATE so that it can be set so that assessment fits any of the assessment models including well known models such as best-of-n. In some versions strategists may keep a running average of all previously encountered potential mates and compare that average to the fitness conferred by mating with the current encountered potential mate. In other versions, strategists may only remember the last encountered potential mate and compare that individual to the currently encountered potential mate. In yet other versions (e.g., Fig. 5), strategists have universal knowledge of the fitness that could be conferred by mating with any member of the opposite sex in the local population. In yet other versions, strategists only assess potential mates when they encounter at least two simultaneously.
Some results of DYNAMATE
The dynamic behavior of individuals under DYNAMATE rules can be visualized in relation to the graph in Figure 4, which shows the isocline of switch points in fitness space for individuals that dynamically experience environments as they change from moment to moment. The line that bisects the graph is the collection of switch points of equal fitness for choosy and indiscriminate behavior. In the graph in Figure 4, we set s = 0.99, and a = 0.2, and studied the effect of variation in the other parameters. Individuals falling in the fitness space above the line will enjoy higher fitness from indiscriminate mating behavior; individuals falling below the line will have higher fitness from choosy mating behavior.
Encounter rates, a
Figure 6 shows the result that, all else equal, as encounter rates increase from 0.01 (Fig. 6a) to 0.2 (Fig. 6c) and higher, a wider range of fitness ratios induce choosy behavior, independent of the strategist's sex. The reason higher encounter rates increase choosy behavior is they increase strategist's opportunities to encounter potential mates with relatively large fitness differences.
Survival probability, s
Figure 7 shows that, all else equal, as survival probabilities decline, a wider range of fitness ratios induce indiscriminate behavior, independent of the strategists' sex. The reason lower survival probabilities increase indiscriminate behavior is they decrease strategist's opportunities to encounter potential mates with relatively large fitness differences. Data on switches from choosy to random mating under predation risk (Dill et al., 1999; Godin and Briggs, 1996; Gong, 1997; Gwynne, 1984, 1985; Hedrick and Dill, 1993) accumulated over the last decade are consistent with the results of the predictions in Figure 7. These studies showed that females switch from choosy to random mating when individuals' s declines, in these cases, under predation risk. All else equal, as survival probability declines (Fig. 7a–c), the fitness payout from indiscriminate mating behavior increases, predicting an increase in indiscriminate behavior. What is also predicted is that for a given distribution of fitness ratios and survival probability (due to variation in predation risk, pathogens, or social risk) the proportion of individuals that express choosy and indiscriminate is also predicted (see below).
|
Latencies to receptivity to re-mating
As the previous Figures (6 and 7) show, all else equal within a sex, as n increases, a wider range of fitness ratios induce choosy behavior. The reason higher n increases choosy behavior is that during relatively longer ns, individuals lose opportunities to mate. All else equal, choosy behavior, following a relatively long n, will attempt to make up relative fitness losses compared to other same sex strategists with shorter n. It is worth emphasizing that in a strict version of DYNAMATE in which only time allocation effects are under consideration, n for virgins is equal to zero and n therefore would have no effect on choosy versus indiscriminate behavior.
Fitness ratios
DYNAMATE predicts from the population distribution of fitness ratios (WL/WH), the proportion of individuals expressing choosy or indiscriminate behavior, all else constant (Fig. 8). Recall that differences in fitnesses conferred by mating with alternative potential mates made possible the calculation of LRS and the ESS for choosy and indiscriminate behavior in Hubbell and Johnson (1987). In the new model of individual flexible behavior, the distribution of fitness ratios allows predictions of the proportion of individuals that will express choosy and indiscriminate behavior. In Figure 8 we illustrate several possible distributions of fitness ratios as probability density functions on the y-axis. In each graph the SPRs were calculated for identical values of s = 0.99 and a = 0.2, over values of n. Fitness ratio distributions might be uniform (top panel); or skewed towards large ratios (second panel), in which the differences in fitness conferred are large; normally distributed (middle panel); or skewed towards small ratios (second from bottom panel), in which the differences in fitness conferred are mostly large; or the distributions could be bi-model (bottom panel). Other distributions are also possible. Suppose we had a known distribution of fitness ratios. Then we could use it to predict the proportion of individuals that will express choosy in any case. To do this pick a point on the line of SPRs. Now draw a line to the y-axis. The distribution of fitness ratios bisects the population into individuals above the line (indiscriminate) and those below (choosy). The fractional area under the fitness ratio distribution above and below the SPR is the expected proportion of individuals expressing indiscriminate or choosy behavior, respectively. For example, when all else is equal and the distributions of fitness ratios are mostly large (most fitness differences are small as in the second panel), most individuals will be indiscriminate; whereas, when distributions of fitness ratios are mostly small, most individuals will be choosy. If individuals encounter potential mates with fitness ratios = SPR, they will be either choosy or indiscriminate. Given values of s, n, and a will describe the SPR for strategists, and the proportion of individuals predicted to be choosy or indiscriminate can be read directly from probability density function of fitness ratios.
|
Almost nothing is known about the distribution of fitness ratios within populations. We have completed one study (P. A. Gowaty, B. T. LeBow, S. P. Hubbell, Y. K. Kim and W. W. Anderson, unpublished manuscript) estimating the distributions of fitness ratios in captive populations of inbred and outbred Drosophila pseudoobscura. Moreover, the distribution of fitness ratios within populations are likely to be dynamic, changing due to disease, resource fluctuations, and other factors. Thus, if the hypothesis is correct, the proportion of individuals expressing adaptive alternatives will also change dynamically and we can expect individuals to change their behavior to enhance their fitness given changes in their circumstances. Also, as individuals in the population enter and exit n, or die, the distribution of fitness ratios will also change, and we can expect the proportion of strategists expressing each alternative to dynamically change with changes in the fitness differences among encountered potential mates.
Behavior results from composite effects of a, s, n, and WL/WH
With Figures 6, 7, and 8, we have discussed results in terms of changes in a single parameter at a time, while all else is constant. However in natural populations these parameters all potentially vary simultaneously, and individuals are induced to express choosy or indiscriminate behavior depending on the current values of encounter rate (a), survival probabilities (s), latencies to re-mating receptivity (n), and the distribution of fitness ratios.
|
TESTING THE MODEL OF ADAPTIVELY, FLEXIBLE SEX ROLE BEHAVIOR |
|---|
|
TOP
SYNOPSISINTRODUCTIONCLASSIC IDEAS IN SEXUAL...CHANCE EFFECTS ON TIME...CHANCE, TIME, AND AN...DYNAMATE AND ADAPTIVE...TESTING THE MODEL OF...POTENTIAL REPRODUCTIVE RATES AND...CAUSE OR CORRELATION IN...COERCION AND MANIPULATION VIA...WHY ARE THE SEXES...References |
|---|
Males, not just females, flexibly adjust choosy and indiscriminate behavior
Both the ESS sex role model and the time allocation hypothesis of flexible behavior and its formal version, DYNAMATE, predict that, regardless of parental investment patterns, not just females, but also males, will be choosy under some circumstances. Altmann (1997)
made a similar prediction. One way to test this prediction is to expose subjects to potential mates representing large differences in fitness conferred. From the perspective that females differ in the fitness they would confer on males with whom they mate, observations of male mate choice with large fitness payoffs are not surprising (Drickamer et al., 2003; Gowaty et al., 2003; Gowaty et al., 2004; Gowaty et al., 2002; Gwynne, 1985; Johnson and Hubbell, 1984; Ryan and Altmann, 2001). Further tests of male mate choice behavior in species with typical female-biased patterns of parental investment would be useful.
Selection will sometimes select against choosy females and indiscriminate males
Both the ESS sex role model and DYNAMATE emphasize that there exist circumstances in which selection is against, even in species with female-biased parental investment, indiscriminate mating by males and choosy mating by females (the prediction from classical sexual selection). Consider a species in which an individual male has a very short n = 1, an a = 0.2, and s = 0.99 (as in Fig. 6, lower panel, for example). If the differences in fitness of encountered potential mates are relatively large, so that the fitness ratios are < 0.2, such individuals will enhance their immediate contributions to life time fitness accumulations by choosy mating. In contrast, if they mate indiscriminately as predicted by parental investment theory, they would decrease their LRS. Now consider a female in the same species with an n =15, a = 0.2, and s = 0.99 (also as in Fig. 6, lower panel). If she encounters individuals with small fitness differences, so that the ratio of the fitnesses conferred is large, she will have higher fitness by indiscriminate mating. One can test these predictions several ways. One way would be to estimate the likely fitness benefits from mating with closely versus more distantly related individuals in experimental populations as Ryan and Altmann (2001) did. One can hold a, s, n constant, and hold known sources of fitness conferred constant, to test the prediction that females will mate indiscriminately. Alternatively, one might vary a, n, and s to predict switches from normal behavior, as experimentalists that vary predation risk have done.
Individuals will dynamically alter their choosy, indiscriminate and competitive behavior
Testing the prediction that particular individuals dynamically change their mate choice behavior is slightly more complex. With reliable estimates of within population variation in distributions of fitness ratios, one can experimentally control encounter rate and survival probability, to predict the proportion of randomly tested individuals in a population that express choosy and indiscriminate behavior. Retests of these same individuals under altered experimental circumstances would then allow direct test of adaptively, flexible real-time changes in the behavior of individuals. We have made the first steps in estimating the distributions of fitness ratios in several populations of D. pseudoobscura (Gowaty et al., unpublished). We have previously described frequencies of receptive individuals that express choosy and indiscriminate, so what remains is to evaluate the distribution of fitness ratios and the proportion of individuals choosy and indiscriminate in the same population.
In the pure time allocation model, n equals zero for virgin females and males
All else equal within a sex, as an individual's latency to receptivity to remating increases, the fitness space under which choosy is favored increases. It is worth considering why this is so in time allocation models. In this model, latency to receptivity to remating for virgins is zero, while for non-virgins of either sex, latency to receptivity to remating may be 
0. This condition emphasizes that the effect of n on choosy and indiscriminate behavior is due to its real-time consequences for fitness accrual. Individuals that have never mated have not yet experienced real-time effects of re-mating latencies on mating success. Thus, virgins, independent of their sexes, experience the same starting conditions in n in terms of real time allocations on their mating decisions. Once mated, however, an individual with a longer n will have fewer subsequent opportunities to mate, so that, depending on the fitness accrued by same-sex competitors, such individuals must use future matings to make up fitness deficits that may have been accrued by other same-sex individuals with shorter n.
One can use DYNAMATE to generate predictions that depend only on time allocation or on past selection favoring, say, choosy females and indiscriminate males associated with energy or time allocations for reproduction. If one wishes to test the fitness payouts from a fixed versus adaptively flexible approach, one can compete two versions of DYNAMATE by calculating the SPR for virgins and already mating individuals in two ways. For the adaptively flexible case, one would set n = 0 for virgins, as we discussed above, and n 0 for already-mated individuals, both assignments being consistent with the time allocation model. The competing version would reflect the force of past selection consistent with the arguments of parental investment theory. Here one would set n = the mean latency to receptivity to remating experienced by individuals of a given sex whether or not they are virgins or already-mated. This option captures in a quantitative way the force of past selection acting through the costs of parental investment. (This is what Clutton-Brock and Parker [1992] did in their model of potential reproductive rates, which we discuss below.) All else equal, these two different approaches for assigning n would yield two different predicted proportions of choosy and indiscriminate individuals within a sex, which experimentalists can use in tests with living organisms.
|
POTENTIAL REPRODUCTIVE RATES AND THE OPERATION OF SEXUAL SELECTION |
|---|
|
TOP
SYNOPSISINTRODUCTIONCLASSIC IDEAS IN SEXUAL...CHANCE EFFECTS ON TIME...CHANCE, TIME, AND AN...DYNAMATE AND ADAPTIVE...TESTING THE MODEL OF...POTENTIAL REPRODUCTIVE RATES AND...CAUSE OR CORRELATION IN...COERCION AND MANIPULATION VIA...WHY ARE THE SEXES...References |
|---|
The OSR does not predict individual behavior
Clutton-Brock and Parker (1992)
also built on the idea that reproductive time outs vary, usually by sex. If one assumes that the adult sex ratio is equal, individuals of the sex with the longest H, to use Sutherland's term, or the longest n, to use our term, would be less and less frequently available for reproduction thereby skewing the operational sex ratio (OSR). Thus H would affect the numbers of males to females available for mating, i.e., the OSR, and the amount of deviation from 1 would predict the likelihood of competition over access to mates among the sex with the most individuals available to mate. This idea has been successful in predicting the average behavior of the sexes, but it has been less successful in predicting individual behavior (Parker and Simmons, 1996; Simmons, 1995). There are at least three reasons for this. The first is that their OSR approach did not incorporate fitness benefits of mating with alternative potential mates, which in the model of flexible sex roles is essential to the prediction of individual choosy and indiscriminate behavior. Fitness benefits and costs of combat over mating with alternative potential mates also predicts the likelihood of combat by individuals (Gowaty and Hubbell, unpublished manuscript). The second reason is the underlying algebra of the potential reproductive rate model. Clutton-Brock and Parker used the length of timeouts after mating, assigning the average value of all timeouts to all individuals within a sex regardless of their mating history. In DYNAMATE, time outs or latencies to receptivity to remating do not enter the predictive equations for individuals until they have mated, so that the values for virgins are n = 0, while for nonvirgins n may be 0; in the flexible model latencies to receptivity to remating are characteristics of individuals. A third reason is that the OSR, the ratio of searching males to searching females may not represent the encounter rate for a given strategist: all strategists may not encounter all searching individuals, so that a perceived sex ratio (PSR) may be a better predictor of what individuals do.
Potential reproductive rate does not predict individual behavior
Because of the general significance of the potential reproductive rate model, which also emphasized time allocation, we studied the relationship of the OSR to predictions of choosy and indiscriminate (competitive) behavior. DYNAMATE calculates both the OSR and the PSR. In DYNAMATE, the PSR is operationally the sex ratio of available mates as perceived by strategists, i.e., the number of males searching divided by the number of females searching. When strategists have perfect knowledge of all searching individuals, the OSR is also the PSR. In DYNAMATE both the OSR and the PSR are based on dynamically changing values of encountered potential mates.
Figure 9 shows a result from DYNAMATE, an analysis of the effects of OSR on choosy and indiscriminate behavior of males and females. In these runs we assumed that all individuals had perfect knowledge of the numbers of searching individuals, so the OSR and the PSR are equal in the populations we studied. The populations were seasonally breeding and consisted of non-overlapping generations. We began with 200 adults, 100 males and 100 females. We ran the model until the last adult died. Because the start point was the same for all individuals, breeding synchrony was strong at the start of the season. We set the initial encounter rate at 0.2, s for males and females = 0.99, and male n = 1 and female n = 10. Because of breeding synchrony, the OSR-PSR oscillated throughout the season, as individuals moved in and out of reproduction and began searching again. Through time, the OSR-PSR exhibited damped oscillations as the degree of reproductive synchrony declined through the breeding season. At the end, they rapidly declined due to drift, when few individuals of either sex were left. The proportion of females expressing choosy behavior was larger than the proportion of males over most of the season, largely because in this example we set n for males = 1, but for females = 10, exaggerating the differences in time-outs between the sexes. As can be seen, the OSR-PSR has little or no value as a predictor of the percent of males or females expressing choosy or indiscriminate (and competitive) behavior (the percent indiscriminate within each sex equals 100 minus the percent choosy for that sex). The decline in choosy behavior (and the increase in indiscriminate behavior) is driven primarily by declining population size (Fig. 9A). As population size declines, indiscriminate behavior of both sexes is favored because of lower encounter rates. Figure 10 shows the lack of predictive power of the OSR-PSR more directly. Individual males and females expressed choosy and indiscriminate behavior at all OSRs-PSRs. The cloud of points of choosy and indiscriminate individuals at OSR-PSR of around 1.5 is due to the long period over which the sex ratio equaled about 1.5 in these populations (Fig. 10). The OSR-PSR did not predict the behavior of individuals in these runs of DYNAMATE, while the composite of a, s, n, and the distribution of fitness ratios did. We are continuing to study the predictive power of OSR-PSRs, as the OSR is a relatively easy variable for empiricists to measure.
|
|
CAUSE OR CORRELATION IN SEX DIFFERENTIATED BEHAVIOR |
|---|
|
TOP
SYNOPSISINTRODUCTIONCLASSIC IDEAS IN SEXUAL...CHANCE EFFECTS ON TIME...CHANCE, TIME, AND AN...DYNAMATE AND ADAPTIVE...TESTING THE MODEL OF...POTENTIAL REPRODUCTIVE RATES AND...CAUSE OR CORRELATION IN...COERCION AND MANIPULATION VIA...WHY ARE THE SEXES...References |
|---|
As we discussed above (Fig. 3), the ESS model reverses the arrow of causation from that described by classic sexual selection ideas, which say that the cost of parental investment favored genes for choosy and indiscriminate behavior, which in turn resulted in higher male than female MSVs. In contrast, the ESS model asserts that non-heritable life history variation generates the MSVs, which come first. Then when these MSVs are coupled with variation in fitness conferred by mating with alternative potential mates, the model predicts when selection favors choosy and indiscriminate behavior. The adaptively flexible sex role hypothesis (Fig. 4) and DYNAMATE (Fig. 5) (Gowaty and Hubbell, unpublished manuscript) build on the ESS sex role model. The flexible sex role hypothesis says that flexibility in the expression of choosy and indiscriminate behavior is very old evolutionarily, and possibly fixed in all sexual organisms. Moreover, selection can start from both non-heritable and heritable sources of variation in a, s and n, and differences in fitness conferred by mating with alternative potential mates.
Support for either the classic or the time allocation hypotheses will come from correlations between choosy/indiscriminate behavior and MSVs. For example, under some conditions both hypotheses predict that choosy behavior may negatively correlate with MSV and indiscriminate behavior may positively correlate with MSV. Therefore, it seems important to test the alternative causal pathways whereby these correlations may arise. This suggests to us that systematic study of the random, fixed, and sexually selected contributions to a, s, n, and WL/WH (Fig. 2) would be a profitable route to further discovery of the force these separate effects have on MSVs and behavior.
It is worth emphasizing, that even when variation in a, s, and n is due entirely to chance factors, selection will favor individuals that flexibly modify or switch their behavior, as long as there remains variation in the fitness conferred by mating with alternative potential mates. If this is so, as we discussed in relation to Figure 4, the scope for further evolution would lie primarily in genetic and developmental influences on sensory and motor systems allowing adaptive flexibility in behavior. Thus, we expect that selection will act on sensory systems providing individuals with information about n, a, and s; on individuals' abilities to assess likely fitness outcomes of mating with alternative potential mates; and on individuals' abilities to respond to environmental, social, and life-history cues.
|
COERCION AND MANIPULATION VIA MODIFICATION OF A, S, AND N |
|---|
|
TOP
SYNOPSISINTRODUCTIONCLASSIC IDEAS IN SEXUAL...CHANCE EFFECTS ON TIME...CHANCE, TIME, AND AN...DYNAMATE AND ADAPTIVE...TESTING THE MODEL OF...POTENTIAL REPRODUCTIVE RATES AND...CAUSE OR CORRELATION IN...COERCION AND MANIPULATION VIA...WHY ARE THE SEXES...References |
|---|
DYNAMATE raises many questions about mechanisms of sensitivities and responses to time left for reproduction. If individual sensory systems have been shaped by the fitness rewards of sensitivity to variation in a, s, and n as the model suggests, the opportunity to manipulate or control the behavior of other individuals, may reside in "exploitation of pre-existing sensory biases." This suggests that conspecifics or others could modify strategists' mating behavior by altering the information available to them about a, s, n, and the fitness differences among alternative potential mates. This logic suggests that an early pathway to classic Darwinian reproductive competition may have included exploitation and modification of cues about time left for mating.
The studies of Chapman and colleagues (Chapman et al., 2000; Chapman and Partridge, 1996) are consistent with the idea that one sex manipulates the behavior of the other sex, and with the idea that females control the outcomes of copulation (Eberhard, 1996). Eberhard's book and Chapman et al. demonstrated that conceptive mating is not simply about the transfer of gametes, as we once thought. In Drosophila melanogaster, and other species, peptides in inseminates act as chemical chastity belts increasing the duration of latency to receptivity to remating in females. This has been considered a form of sexually selected manipulation of female behavior through chemical mate-guarding. There may be another indirect effect of these peptides on females consistent with the manipulation by males of cues that induce females to be choosy or indiscriminate. By increasing females' latencies to remating receptivity, the peptides may also increase the likelihood that these females will express choosy behavior in subsequent matings. If a male that previously mated a female is unlikely to encounter her again, it would be to his advantage to increase the likelihood that she would express choosy behavior on subsequent encounters, because her later choosy behavior would thereby decrease the likelihood of sperm competition for the previously mated male. This means, among other things, that the choosy behavior that Bateman, following Darwin, attributed to the evolved nature of coy and passive females might have been induced by manipulative intervention of males, as Gowaty (1997) suggested in a more general discussion of ways males could manipulate and attempt to control the behavior of females.
The time allocation models of flexible behavior also suggest that the frequently observed indiscriminate mating behavior of males could result when males or other conspecifics keep other males from interacting with females. Keeping males away from females would reduce the information available to them about fitness differences among potential mates and it would reduce their encounters with potential mates, and thus, keeping males away from females also would reduce the access of some males to cues that would induce them to be choosy. As the flexible models of choosy and indiscriminate and DYNAMATE predict, all else equal, the lower the encounter rates of individual males with potentially mating females, the more likely a male is to express indiscriminate behavior. A corollary is that in species with variation among males in the expression of showy traits, which attract females, males with the most exaggerated expression may have higher encounter rates with potentially mating females than males with less exaggerated expression of these traits. All else equal, this predicts that within a species the more highly ornamented or showy males are more likely than less ornamented or showy males to express choosy mating behavior. A similar idea suggests that in species in which males defend resources more or less attractive to females, that males with more attractive resources have more encounters with females and thus they also will more reliably express choosy behavior (Itzkowitz and Haley, 1999).
|
WHY ARE THE SEXES AS THEY ARE? |
|---|
|
TOP
SYNOPSISINTRODUCTIONCLASSIC IDEAS IN SEXUAL...CHANCE EFFECTS ON TIME...CHANCE, TIME, AND AN...DYNAMATE AND ADAPTIVE...TESTING THE MODEL OF...POTENTIAL REPRODUCTIVE RATES AND...CAUSE OR CORRELATION IN...COERCION AND MANIPULATION VIA...WHY ARE THE SEXES...References |
|---|
The discussion above connects with the key question that still incites so much interest in animal behavior, why the sexes are as they are, which now seems a considerably more complex question than it once was. How different are the sexes when it comes to classic "sex typical behavior" of choosy and indiscriminate mating? To answer that question, we still need to know if between-sex variation is really larger than within-sex variation in key behavioral states that are used to characterize the sexes, "sex typical," and sex differentiated behavior. If within-sex variation in sensitivity to key inducing factors is greater than between sex variation, are sex "roles" really roles after all? Are "typical differences" in the sexes due to genetically influenced sex-linked traits or to the real-time ecological forces experienced by individuals? Or to some combinations of both? And, if both, what are the combinations? We will not know the answers to these questions until we level the ecological playing fields of females and males and observe behavior under controlled conditions. Until we do this, we will be unable to clearly attribute sex differences in choosy, indiscriminate and competitive behavior to genetically influenced traits or to environmental cues that induce ecologically relevant and, as we argue here—adaptive, differences. If the ecological theatres in which the sexes play are different and if different ecological theatres account for sex differentiated behavior, the question will become: what makes for ecological and social differences experienced by individual males and females?
|
ACKNOWLEDGMENTS |
|---|
We thank Zuleyma Tang Martinez for inviting our contribution to this symposium. We thank Bill Sutherland for the initial inspiration that started this all off, and for his generous conversations with us over the last year. We thank Geoff Parker for his insightful discussions with us about these ideas and for sharing with us that he had drawn stick models of time allocation in his teaching notes as long ago as the 1960s. We thank John Avise, Celeste Condit, Lee Drickamer, John Harte, Jan Leonard, Judy Stamps, and Mary Jane West-Eberhard for candid comments that helped us to be clearer, we hope. We thank Jerry Downhower for his comments on a very preliminary version, and Wyatt Anderson for his cheerful responses to these ideas. We thank PAG's students Jason Lang, Jill Goldstein, Leslie Ruyle, and particularly Brian Snyder, who took a special interest in these ideas, for helpful discussion. We thank the formal reviewers, Chuck Snowdon and anonymous, for their careful readings and very helpful advice.
|
FOOTNOTES |
|---|
1 From the Symposium Bateman's Principle: Is It Time for a Reevaluation? presented at the Annual Meeting of the Society for the Integrative and Comparative Biology, 5–9 January 2004, at New Orleans, Louisiana.
|
References |
|---|
|
TOP
SYNOPSISINTRODUCTIONCLASSIC IDEAS IN SEXUAL...CHANCE EFFECTS ON TIME...CHANCE, TIME, AND AN...DYNAMATE AND ADAPTIVE...TESTING THE MODEL OF...POTENTIAL REPRODUCTIVE RATES AND...CAUSE OR CORRELATION IN...COERCION AND MANIPULATION VIA...WHY ARE THE SEXES...References |
|---|
Altmann, J. 1997. Mate choice and intrasexual reproductive competition: Contributions that go beyond acquiring more mates. In P. A. Gowaty (ed.), Feminism and evolutionary biology: Boundaries, intersections, and frontiers, pp. 320–333. Chapman and Hall, New York.
Andersson, M. 1994. Sexual selection. Princeton University Press, Princeton.
Bateman, A. J. 1948. Intra-sexual selection in Drosophila. Heredity, 2:349-368.[Web of Science][Medline]
Chapman, T., D. M. Neubaum, M. F. Wolfner, and L. Partridge. 2000. The role of male accessory gland protein Acp36DE in sperm competition in Drosophila melanogaster. Proc. R. Soc. London Ser. B-Biol. Sci, 267:1097-1105.
Chapman, T., and L. Partridge. 1996. Female fitness in Drosophila melanogaster: An interaction between the effect of nutrition and of encounter rate with males. Proc. R. Soc. London Ser. B-Biol. Sci, 263:755-759.
Clutton-Brock, T. H., and G. A. Parker. 1992. Potential reproductive rates and the operation of sexual selection. Quart. Rev. Biol, 67:437-456.[CrossRef]
Darwin, C. 1859. The origin of species by means of natural selection. John Murray, London.
Darwin, C. 1871. The descent of man and selection in relation to sex. John Murray, London.
Dill, L. M., A. V. Hedrick, and A. Fraser. 1999. Male mating strategies under predation risk: Do females call the shots? Behav. Ecol, 10:452-461.
Drickamer, L. C., P. A. Gowaty, and D. M. Wagner. 2003. Free mutual mate preferences in house mice affect reproductive success and offspring performance. Anim. Behav, 65:105-114.[CrossRef]
Eberhard, W. G. 1996. Female control: Sexual selection by cryptic female choice. Princeton University Press, Princeton.
Finke, O. M. 1982. Lifetime mating success in a natural population of the damselfly, Enallagma hageni (Walsh) (Odonata: Coenagrionidae). Behav. Ecol. Sociobiol, 10:293-302.[CrossRef][Web of Science]
Godin, J. G. J., and S. E. Briggs. 1996. Female mate choice under predation risk in the guppy. Anim. Behav, 51:117-130.[CrossRef]
Gong, A. 1997. The effects of predator exposure on the female choice of guppies (Poecilla reticulata) from a high-predation population. Behaviour, 134:373-389.
Gowaty, P. A. 1997. Sexual dialectics, sexual selection, and variation in mating behavior. In P. A. Gowaty (ed.), Feminism and evolutionary biology: boundaries, intersections, and frontiers, pp. 351–384. Chapman Hall, New York.
Gowaty, P. A. 2003. Power asymmetries between the sexes, mate preferences, and components of fitness. In C. Travis (ed.), Women, evolution, and rape, pp. 61–86. MIT Press, Cambridge, Massachusetts.
Gowaty, P. A. 2004. Sex roles, contests for the control of reproduction, and sexual selection. In P. M. Kappeler and C. P. van Schaik (eds.), Sexual selection in primates: New and comparative perspectives, pp. 163–221. Cambridge University Press, Cambridge.
Gowaty, P. A., L. C. Drickamer, and S.-S. Holmes. 2003. Male house mice produce fewer offspring with lower viability and poorer performance when mated with females they do not prefer. Anim. Behav, 65:95-103.[CrossRef]
Gowaty, P. A., R. Steinechen, and W. W. Anderson. 2004. Indiscriminate females and choosy males: Within- and between-species variation in Drosophila. Evolution, 57:2037-2045.
Gowaty, P. A., R. Steinichen, and W. W. Anderson. 2002. Mutual interest between the sexes and reproductive success in Drosphila pseudoobscura. Evolution, 56:2537-2540.[CrossRef][Web of Science][Medline]
Gwynne, D. T. 1984. Sexual selection and sexual differences in mormon crickets (Orthoptera, Tettigoniidae, Anabrus-Simplex). Evolution, 38:1011-1022.[CrossRef][Web of Science]
Gwynne, D. T. 1985. Role reversal in katydids: Habitat influences reproductive behavior (Orthoptera, Tettigoniidae, Metaballus Sp). Behav. Ecol. Sociobiol, 16:355-361.[CrossRef]
Hedrick, A. V., and L. M. Dill. 1993. Mate choice by female crickets is influenced by predation risk. Anim. Behav, 46:193-196.[CrossRef][Web of Science]
Holling, C. S. 1959. Some characteristics of simple types of predation and parasitism. Canad. Entomol, 91:385-398.
Hubbell, S. P., and L. K. Johnson. 1987. Environmental variance in lifetime mating success, mate choice, and sexual selection. Am. Nat, 130:91-112.[CrossRef][Web of Science]
Itzkowitz, M., and M. Haley. 1999. Are males with more attractive resources more selective in their mate preferences? A test in a polygynous species. Behav. Ecol, 10:366-371.
Johnson, L. K., and S. P. Hubbell. 1984. Male choice: Experimental demonstration in a Brentid weevil. Behav. Ecol. Sociobiol, 15:183-188.[CrossRef]
Mackenzie, A., J. D. Reynolds, V. J. Brown, and W. J. Sutherland. 1995. Variation in male mating success on leks. Am. Nat, 145:633-652.[CrossRef][Web of Science]
Parker, G. A., R. R. Baker, and V. G. F. Smith. 1972. The origin and evolution of gamete dimorphism and the male–female phenomenon. J. Theoret. Biol, 36:529-553.[CrossRef][Web of Science][Medline]
Parker, G. A., and L. W. Simmons. 1996. Parental investment and the control of sexual selection: Predicting the direction of sexual competition. Proc. R. Soc. London, B315-321.
Ryan, K. K., and J. Altmann. 2001. Selection for male choice based primarily on mate compatibility in the oldfield mouse, Peromyscus polionotus rhoadsi. Behav. Ecol. Sociobiol, 50:436-440.[CrossRef][Web of Science]
Shuster, S. M., and M. J. Wade. 2003. Mating systems and strategies. Princeton University Press, Princeton, New Jersey.
Simmons, L. W. 1995. Relative parental expenditure, potential reproductive rates, and the control of sexual selection in katydids. Am. Nat, 145:797-808.[CrossRef][Web of Science]
Sutherland, W. J. 1985a. Chance can produce a sex difference in variance in mating success and account for Bateman's data. Anim. Behav, 33:1349-1352.[CrossRef]
Sutherland, W. J. 1985b. Measures of sexual selection. In R. Dawkins and M. Ridley (eds.), Oxfords surveys in evolutionary biology, pp. 90–101. Oxford University Press, Oxford.
Sutherland, W. J. 1987. Random and deterministic componsents of variance in mating success. In J. W. Bradbury and M. B. Andersson (eds.), Sexual selection: Testing the alternatives, pp. 209–219. John Wiley & Sons Limited, New York.
Tang-Martinez, Z., and T. B. Ryder. 2005. The problem with paradigms: Bateman's worldview as a case study. Integr. Comp. Biol, 45:821-830.
Trivers, R. L. 1972. Parental investment and sexual selection. In B. Campbell (ed.), Sexual selection and the descent of man, pp. 136–179. Aldine, Chicago.
Wedekind, C. 1999. Pathogen-driven sexual selection and the evolution of health. In S. C. Stearns (ed.), Evolution in health and disease, pp. 102–107. Oxford University Press, Oxford.
West-Eberhard., 2003. Developmental plasticity and evolution. Oxford University Press, New York, New York.
Williams, G. C. 1966. Adaptation and natural selection. Princeton University Press, Princeton.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. Clutton-Brock Sexual Selection in Males and Females Science, December 21, 2007; 318(5858): 1882 - 1885. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. L. Leonard Sexual selection: lessons from hermaphrodite mating systems Integr. Comp. Biol., August 1, 2006; 46(4): 349 - 367. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. Tang-Martinez and T. B. Ryder The Problem with Paradigms: Bateman's Worldview as a Case Study Integr. Comp. Biol., November 1, 2005; 45(5): 821 - 830. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. L. Leonard Bateman's Principle and Simultaneous Hermaphrodites: A Paradox Integr. Comp. Biol., November 1, 2005; 45(5): 856 - 873. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||