A conceptual review of mate choice: stochastic demography, within‐sex phenotypic plasticity, and individual flexibility

Abstract Mate choice hypotheses usually focus on trait variation of chosen individuals. Recently, mate choice studies have increasingly attended to the environmental circumstances affecting variation in choosers' behavior and choosers' traits. We reviewed the literature on phenotypic plasticity in mate choice with the goal of exploring whether phenotypic plasticity can be interpreted as individual flexibility in the context of the switch point theorem, SPT (Gowaty and Hubbell 2009). We found >3000 studies; 198 were empirical studies of within‐sex phenotypic plasticity, and sixteen showed no evidence of mate choice plasticity. Most studies reported changes from choosy to indiscriminate behavior of subjects. Investigators attributed changes to one or more causes including operational sex ratio, adult sex ratio, potential reproductive rate, predation risk, disease risk, chooser's mating experience, chooser's age, chooser's condition, or chooser's resources. The studies together indicate that “choosiness” of potential mates is environmentally and socially labile, that is, induced – not fixed – in “the choosy sex” with results consistent with choosers' intrinsic characteristics or their ecological circumstances mattering more to mate choice than the traits of potential mates. We show that plasticity‐associated variables factor into the simpler SPT variables. We propose that it is time to complete the move from questions about within‐sex plasticity in the choosy sex to between‐ and within‐individual flexibility in reproductive decision‐making of both sexes simultaneously. Currently, unanswered empirical questions are about the force of alternative constraints and opportunities as inducers of individual flexibility in reproductive decision‐making, and the ecological, social, and developmental sources of similarities and differences between individuals. To make progress, we need studies (1) of simultaneous and symmetric attention to individual mate preferences and subsequent behavior in both sexes, (2) controlled for within‐individual variation in choice behavior as demography changes, and which (3) report effects on fitness from movement of individual's switch points.

Most empirical research on mate choice followed publication of William's (1966) cost of reproduction argument, Trivers' (1972) parental investment theory, and Parker's et al. (1972) anisogamy theory. These related ideas provided scenarios for the evolution of fixed, sex-differentiated behavior due to posited ancient selection pressures acting on sex biases in gamete sizes and "parental investment." Traditionally, therefore, investigators have assumed that mate choice is directional and fixed within a species, sex typical, and static within individuals over time and that choosersusually femaleschose mates on the basis of exaggerated, sexually selected traitsusually in males. As the current review shows, observations increasingly demonstrate that there is, in many species, considerable within-sex phenotypic plasticity for choosy versus indiscriminate mating behavior (de Gaudemar 1998;Qvarnstrom et al. 2000;Forsgren et al. 2004;Plaistow et al. 2004;Lynch et al. 2005;Simcox et al. 2005;Lehmann 2007; Chaine and Lyon 2008;Heubel and Schlupp 2008;Ah-King and Nylin 2010).
We wonder how much within-sex phenotypic plasticity is actually among-individual or even within-individual flexibility. It is possible that individual flexibility is expressed independent of an individual's sex Hubbell 2005, 2009), and it is certain that one cannot know whether this is the case without within-species, within-population symmetric tests on individuals of different sexes. It is also possible that the relatively commonly observed within-sex phenotypic plasticity that we catalog here is really individual flexibility. We propose a conceptual transition to empirical studies of the inducers of individual flexibility (and the limits to flexibility) with renewed interest in the real-time fitness effects of any observed flexibility.

Background
Almost twenty-five years ago, Hubbell and Johnson's (1987) discrete time mating theory (hereafter H&J's mating theory) opened the doors to tests of quantitative predictions of ecological and social constraints on individual flexibility in reproductive decisions. Their model provided analytical solutions to the expected mean and variance in lifetime mating success and, for the first time, an alternative to the parental investment hypotheses for choosy and indiscriminate behavior. Their results showed that the evolution of choosy and indiscriminate behavior of individuals depended on (1) probabilistic demography and (2) variation in the quality of mates. H&J's mating theory predicted adaptive phenotypic plasticity (see discussion in Gowaty and Hubbell 2005). Some authors used the concepts of H&J's mating theory to explore variation in mating behavior (Bjorklund 1990;McLain 1991McLain , 1992Michiels and Dhondt 1991;Travers and Sih 1991;Wickman 1992;Berglund 1993), and Crowley et al. (1991) produced a simulation model of individuals in a Figure 1. The evolution of adaptive, fitness enhancing, and flexible individuals (the fourth column above) able to switch their reproductive decisions based on their current demographic situations depends upon probabilistic (stochastic) variation in (first column above) a focal individual's encounter probability with potential mates, e, their survival probability s, the duration of any postmating time-outs that the focal has experienced o, and the number of potential mates in the population n, which together predict an individual's expected mean lifetime number of mates under demographic stochasticity. The second column above indicates the SPT's explicit dependence upon the within-population random distribution of fitness that would be conferred. The third column above indicates that the SPT assumes that selection occurred so that what evolved was (1) individual sensitivities to probabilities of encounter of potential mates e, probability of survival s, the duration of postmating time-outs o, and the number of potential mates in the population n and the w-distribution and in (2) abilities to assess the fitness that would be conferred by any potential mate. The SPT proved mathematically (the fourth column above) that individuals fixed in their reproductive behavior would be selected against relative to flexible individuals able to make real-time mating decisions fit to their current ecological and social situations, as though decision-makers are Bayesians able to update their priors to better fit their actions to the demographic and social circumstances they are in (Gowaty and Hubbell 2013). seasonal population based on the parameters associated with waiting to mate (being "choosy"), further inspiring research about environmental sources of variation in reproductive decisions. A later idea related to H&J's mating theorybut with important differences in assumptionscaught on and spread, inspiring many more empirical studies: Clutton-Brock and Parker (1992) said that the potential reproductive rate (PRR) of the sexes determined the operational sex ratio (OSR) and, in turn, determined the opportunities for within-sex competition over mates. Under the influence of PRR theory, investigators found compelling cases of "reversed sex roles" in choosy and indiscriminate behavior (Clutton-Brock and Vincent 1991;Berglund 1994;Kvarnemo and Simmons 1998;Jirotkul 1999;Forsgren et al. 2004;Klug et al. 2008), so that females were called "the competitive sex" when males were rare and males "the competitive sex" when females were rare. Empirical discoveries stimulated additional continuous time models, which remained consistent with the cost of reproduction expectations and most often sought solutions to sex-differentiated equilibrium conditions to predict mating rates as a function of costs of reproduction, population density, OSR, etc. The factors were hypothesized to affect the "direction of sexual selection" because each affected the relative rarity of one or the other sex. Today, there are dozens of papers reporting indiscriminate behavior in the "choosy sex" (Bjorklund 1990;Berglund and Rosenqvist 1993) and investigators and theorists have produced a large number of conceptual and theoretical explanations for observations of switches in which sex is "choosy." Yet, few have concluded the obvious: Within-sex phenotypic plasticity is inconsistent with the predictions from the cost of reproduction arguments of fixed sex differences in reproductive decision-making.
A discrete time, analytical model, the switch point theorem (SPT) (Gowaty and Hubbell 2009) says that flexible individuals trade-off time available for mating with fitness that would be conferred from mating with this or that potential mate, and it proved theoretically that individual flexibility in accepting potential mates on encounter ("indiscriminate" behavior) or rejecting potential mates and waiting for a better option ("choosy" behavior) is adaptive when demographic situations fluctuate or change. The SPT proved that adaptive flexibility increases an individual's expected lifetime reproductive success, irrespective of the sex of the individual, and thereby also proved that fixed choosy or indiscriminate mating behavior would be maladaptive and likely selected against. The parameters of the SPT are individual survival probability per unit time s, the probability of encountering potential mates per unit time during periods of receptivity e, the duration of any postmating time-out or latency before reentering receptivity o, the number of potential mates n, and the distribution of fitness that would be conferred, the w-distribution. The "switch point" is the point along an axis of ranked potential mates that indicates the fitness that would be conferred on a focal individual if they mated with a given potential mate. The ranks that a focal individual self-referentially assigns (Box 1) to potential mates do not change: What does change is the demographic circumstances of the focal individual. For example, under variation in a focal individual's survival probability, his or her switch points between acceptable and unacceptable ranked potential mates may change: If the focal individual's survival probability increases, the switch point may move to better ranks so that the focal individual deems fewer potential mates acceptable and more unacceptable; if the focal individual's survival probability decreases, the switch point may move to potential mates with worse fitness ranks, so that the focal individual deems more potential mates acceptable, fewer unacceptable.
The SPT changed the subject from sex-specific behavior to individual-specific behavior. It also changes the subject from the traits of the chosen sex to the social, ecological, and trait variation in the individuals doing the choosing. As we show here and as Hubbell (2005, 2009) argued, the scenario ( Fig. 1) from the SPT for the evolution of flexible individuals potentially unifies and simplifies the large number of explanatory variables of empirically demonstrated within-sex phenotypic plasticity in mate choice behavior ( Table 1).
Studies of within-sex mate choice plasticity usually focus on "choosy-sex" behavior when the choosers vary in intrinsic characteristics such as age and experience, their condition, parasite load, and ecological and social circumstances, including the adult sex ratio (ASR), the operational sex ratio (OSR), density, and predator or parasite risk. Conclusions are thus necessarily about sexes and implied sex differences. But, many of these factors may induce within-individual changes in behavior, and experiments of induced changes can provide evidence of individual flexibility in the ability to sense and respond in adaptive ways (Gowaty and Hubbell 2013) as individual's circumstances change. The factors tested in within-sex mate choice plasticity studies seem relatively easy to measure, but there are a great many of such factors and some are obvious, complex proxies for the fundamental variables of the SPT ( Fig. 1; Table 1).
Our goals with this review are to: (1) suggest how a few simple parameters unify and simplify a seemingly bewildering number of variables associated with withinsex switches in choosy and indiscriminate behavior; (2) draw attention to environmentally induced behavior of individuals rather than sexes; and (3) propose how small methodological changes can evaluate how very simple variables may work to produce changes in behavior of individuals of either sex. We gathered papers testing within-sex phenotypic plasticity in choosy and indiscriminate behavior (Table 2). We then categorized the variables in terms of their potential effects on the SPT's variables of probability of survival s, probability of encountering potential mates e, postmating time-outs o, the number of potential mates in the population n, and the distribution of fitness that would be conferred w-distribution (also in Table 2) and summarized the studies in various ways. Last, we discuss the implications of the reviewed studies taken together.

Methods
To find studies on within-sex phenotypic plasticity in choosy and indiscriminate mating behavior, we used Web of Science. We searched using phrases that we thought were common in the literature of changes in mating behavior including adult sex ratio or ASR, operational sex ratio or OSR, parasite load, predation risk, condition, age, and experience, each in combination with "mate choice or mate preferences," as in "ASR and mate choice or mate preferences." We also searched Web of Science for papers that cited early papers on mate choice flexibility: Losey et al. (1986), Hubbell and Johnson (1987), Kennedy et al. (1987), Houde (1987Houde ( , 1988, Breden and Stoner (1987), Wade and Pruett-Jones (1990), Shuster and Wade (1991), Clutton-Brock and Parker (1992), Dugatkin (1992a), Pruett-Jones (1992), and Hedrick and Dill (1993). The searches yielded over 3300 citations, of which 198 were empirical papers on changes in choosy versus indiscriminate behavior (Table 2). Box 1 contains a glossary with the meanings that we used for common terms. We categorized studies in Table 2 under probability of survival s, probability of encountering potential mates e, postmating time-outs o, the number of potential mates in the population n, and the distribution of fitness that would be conferred w-distribution (Gowaty and Hubbell 2009) depending on the information in each study. We coded studies of "audience effects" and "sperm competition risk" with question marks. We categorized some studies under multiple SPT parameters. In addition, 16 studies (Table 3) reported negative evidence of phenotypic plasticity.
The justifications follow for placing common explanations (such as predation risk, mating status, OSR, condition, and age) into categories representing encounters with potential mates e, likelihood of survival s, duration of latency before reentering receptivity after mating o, the number of potential mates in the population n, and the likely fitness conferred from any mating or decision to accept a mating w-distribution. Predation risk is an ecological correlate of changes from choosy to random mating (Breden and Stoner 1987). Predation risk logically may represent variation in probability of survival s, probability of encountering potential mates e, postmating time-outs o, and the number of potential mates in the population n (Gowaty and Hubbell 2009) (Table 1). Predation risk very likely reduces individual instantaneous probability of s, but prudent prey may modify their behavior in the presence of predators, modifying their behavior to reduce their own conspicuousness, which is likely also to decrease their e, encounters with potential mates, as well as the local number of potential mates n that they or others may respond to. Experimental laboratory studies of predation risk almost always implicitly controlled for variation in probability of encounter of potential mates e and the number of potential mates in the population n, while remaining silent on variation in subjects' prior breeding experience, their ages, condition, and any reproductive success that might have accrued among individuals with different patterns of acceptances or rejections of potential mates. Thus, we categorized most studies of predation risk under probability of survival s, or probabilities of survival s and encounter e unless investigators provided other evidence that e or n varied (usually in studies of wild-living subjects). Age (Kodric-Brown and Nicoletto 2001) is intuitively important to reproductive decisionmaking. But, age is a fuzzy proxy for an individual's probability of survival, s, and/or the effects of prior experience that can have effects on subjects' knowledge about the fitness that potential mates could confer, w-distribution. In the studies of age effects in Table 2, investigators sometimes controlled for variation in experience. In models of individual flexibility in reproductive decision-making, age is often correlated to variation in the duration of postmating time-outs or latency, o, which in the absence of previous selection on choosy and indiscriminate mating will have no effect on virgins but will on nonvirgins. If virgins are always or often younger than nonvirgins, age may correlate with individual duration of time-outs, o. Because virgins have never mated, the duration of time-out is necessarily zero for virgins. We categorized studies that examined age effects on mate choice behavior under probability of survival, s. Mating status ) effects on switches from choosy to random are still infrequently tested. However, in state-dependent, discreet time models, such as H&J's mating theory (1987) or in the SPT (Gowaty and Hubbell 2009), the difference between virgins and mated individuals is captured with parameter o, the duration of postmating time-outs. For virgins, the duration of postmating time-outs always equals zero, effectively having nothing to do with the individual flexibility until after an individual's first

Box 1. Glossary with definitions of inducing variables and terms indicating reproductive decisions and mating behavior
Accepting refers to the behavior of mating or accepting a mating solicitation; it may be associated with subtle motor patterns: simply staying still may be an acceptance signal (Markow 1987) or stereotypical postures or calls. Accepting a potential mating differs from appetitive behavior that may be associated with assessment of alternative potential mates.
Assessment of alternative potential mates is a cognitive process and thus very difficult to operationalize or standardize. Ecologists and evolutionary biologists infer that individuals are assessing (something) by defined variation in appetitive or approach behavior. Neurobiologists may in the future evaluate assessment via imaging of neurological patterns.
Consensus mate preference occurs when all or most individuals of one sex prefer the same opposite-sex individual (which is in contrast to "individual" mate preference, defined below). For example, investigators of mallards inferred consensus mate preferences when female mallards displayed to dominant males on the wintering grounds (Cunningham and Russell 2000).
Choosiness is defined as the effort an individual invests in mate assessment (Jennions and Petrie 1997), a definition without defined operational criteria.
Choosy refers to the sensory ability of individuals to assess alternative potential mates or to motor patterns indicating rejection of some potential mates, but not others. In organisms in unrestricted field populations, investigators often assign the label "choosy" to subjects who reject some potential mates, but accept others. Like "choosiness," "choosy" is a relatively loose term with many, often nonoverlapping meanings and is often difficult to operationalize, because it embeds and confounds cognitive and motor processes.
Encountering a potential mate is a behavioral state of opposite-sex individuals who are close enough for others to send or receive solicitation signals, rejection signals, or for individuals to otherwise sense characteristics of the potential mate. Empirical studies depend on operationalized definitions of "encountering" that may vary depending on the study species.
Indiscriminate most often refers to individuals who accept copulations with alternative potential mates at random with respect to characteristics that investigators suspect are key traits choosers discriminate (songs, plumage, size, or other phenotypes). Thus, investigators' should perhaps label their subjects as "indiscriminate" relative to the particular tested traits in those being tested between.
Individual flexibility refers to an extreme form of developmental variation, a type of plasticity induced by changing ecological and social circumstances of individuals in real time, not evolutionary time, and perhaps moment to moment. The term captures the idea that an individual may choose to do this or that or something else altogether, changing behavior moment to moment as circumstances change. It stresses the possibility of within-individual changes, not just between-individual changes. Many behavioral studies are about variation in individual behavior, for example, individual flexibility in foods taken, stored, and manner of retrieval.
Individual mate preferences are those that are self-referential so that preferences for potential mates are weighted or conditioned on the traits of the individual expressing "the preference." Individual mate preferences could reflect "consensus mate preference" under some conditions. In practice, investigators of nonhuman animals infer "mate preferences" from subjects' behavioral variation, such as proximity to alternative potential mates, often in controlled situations such as "mate preference arenas".
Mate assessment is a cognitive evaluation based on individuals' abilities to sense differences between alternative potential mates, and in terms of Gowaty and Hubbell's (2009) switch point theorem (SPT), to rank alternative potential mates along chooserunique-ranked axis of fitness that would be conferred by mating with any potential mate.
Mate choice is a fuzzy term implying both cognitive and motor acts in which a focal individual accepts or rejects copulation with a potential mate. In practice, it is sometimes defined more narrowly as "any pattern of behavior, shown by members of one sex, that leads to their being more likely to mate with certain members of the opposite sex rather than others" (p. 4, Halliday, 1983). However, the later definition confounds mate preferences and/or mate assessments with other potential mediators of mating such as intrasexual interactions or sexual coercion.
Preferences, including mate preferences, indicate cognitive states of an individual. Investigators characterize focal individual behaviormoving toward or orienting toward others, as indicating a preference for individuals or for individuals with different traits (e.g., plumage, calls). In other words, investigators infer cognitive states from behavioral correlates. mating, but for remating individuals, the duration of their postmating time-outs o may be important. If all else is equal, that is, holding e, s, n, and the w-distribution constant, the SPT and similar state-dependent models predict that virgins mate on encounter more frequently than already-mated individuals who are predicted more often to wait for a better option. Thus, we categorized studies investigating the effects of mating status, whether virgin or remating under duration of postmating timeouts o. OSR (Berglund and Rosenqvist 1993) has been linked with within-sex phenotypic plasticity in changes from accepting to rejecting potential mates. OSR may be a complex proxy for an individual's encounter probability with potential mates e as many investigators have argued, so we categorized studies of OSR with e, or under e or n, or e and s as OSR also contains information about number of potential mates and the instantaneous survival probability of decision-makers. Disease state, parasite load, condition, body size, and "attractiveness" are usually linked not to choosers but to those individuals that choosers are assessing (Andersson 1994). More recently, focus has changed, so that investigators are asking whether variation in the "choosiness" of individuals of "the choosy sex" depends on the chooser's condition (Kodric-Brown 1995), chooser's disease state or parasite load (Lopez 1999), or chooser's attractiveness (Itzkowitz and Haley 1999). Usually implicitly, investigators assume that condition, disease state, and parasite load affect within-individual energy trade-offs affecting the hypothesized costs of mate choice behavior. Intuitively, individual condition, disease state, body size, and parasite load may indicate variation in likelihood of survival, s, or the likelihood of encountering potential mates, e. We categorized condition, disease state, body size, and parasite load under s, e, or s and e depending on the information available in given papers. Attractiveness of resources (Itzkowitz and  Phenotypic plasticity is a term sometimes used to characterize moment-to-moment changes in phenotypes and thus overlaps in usage with individual flexibility. Here, we make a distinction between developmental conditions that induce usually fixed changes in phenotypes when individuals change sex in response to variation in the adult sex ratio. In contrast, individual flexibility is a term to indicate changes in behavioral phenotypes even moment to moment, as happens when individuals hide from predators. The color camouflage of octopus is an example of individual flexibility in moment-to-moment changes in phenotype.
Rejecting refers to the behavior of individuals refusing to accept a copulation solicitation or a copulation attempt; it may be associated with failure to respond to copulation solicitation postures, or more active behavioral indicators, such as aggressive rejections or moving away from a soliciting opposite-sex conspecific.
Reproductive decisions refer to alternative motor acts of accepting a potential mate on encounter (which might appear "indiscriminate") or waiting for a better mate (which might a appear as "choosy"). However, the SPT assumes that both decisionseither to mate on encounter or to wait for a better mateare conditioned by an individual's prior assessment of the fitness that would be conferred by mating.  Table 2. Studies reporting within-sex phenotypic switches from choosy to indiscriminate mating under variation in population density, OSR, ASR, chooser's condition, chooser's resources, predation risk, disease risk, and other factors sorted by the SPT's hypothesized inducers of individual flexibility: survival probability s, encounter probability e, duration of latency to remating l, the number of potential mates n, and the distribution of fitness that would be conferred under random mating, the w-distribution.           Males were consistent in their preferred (larger) mate without current (reject more), but inconsistent when exposed to current, thus increasing cost of staying close to initially preferred female (accept more).  Kim et al. (2005) Proximity to potential male versus occurrence of mating when competition and interaction was allowed (between two potential mates and the chooser). Both females and males preferred larger partners in nontouch testing arenas, but in actual mating when competition and interaction were allowed larger individuals did not mate more frequently.    Females from low-density population mated sooner (accepted more) than those from the high-density population (rejected more). when density is low, mating occurs on the surface, when high mating is more often underground. Females differed in searching strategies between high-and low-density sites, wandering around in high density and staying at own burrow in low. Mating males were larger in high density. Lehmann (2007) Rejection by walking away from potential mate. In high-density populations, females encountered more males and rejected more; in low-density populations, females encountered fewer males and accepted more. Females in low-density populations were less likely to reject males later in the night when the mating period came close to an end.

Gryllus integer
Hedrick and Dill (1993) Moving toward short-bout sound with cover or long-bout sound without cover. Females prefer long-bout sound. Increased predation made females move toward the short-bout sound with cover instead (accept more).

Guppies
Poecilia reticulata Godin and Briggs (1996) Proximity to one of two males (colorful/dull) presented with or without predator exposure. Females from high-predation river reduced level of sexual activity and preference for a particular male during predation risk (accept more). Females from lowpredation site did not change behavior when exposed to predator.   Rand et al. (1997) Female proximity to speakers with simple or complex call was recorded in dim light (higher predation risk) or darkness (lower predation risk). Females moved toward speakers more in the dark (accepted more) and more often toward a simple call close by than a more complex preferred call further away (accepted more). of females leading to pair formation increased and females resisted less with increased predation risk (accept more). Males tended to mate more often with larger females at no predation risk (reject more). Predation risk implies reduced activity, which leads to reduced encounter rate.
Laboratory ♀ Yes s and e Predation risk ♂ Broad-nosed pipefish Syngnathus typhle Berglund (1993) Time spent dancing in front of female and number of copulations with large or small female was recorded in the presence/absence of predator. Males preferred large females in the absence of predator (reject more), and during predation risk, they were less active and did not differentiate females with relation to body size (accept more).   Simcox et al. (2005) Mating attempts with familiar/unfamiliar females. In laboratory: males attempted to mate more with unfamiliar females, but did not distinguish between females when given only visual access.
In field, males in populations with predation preferred unfamiliar females only when light levels were dim and not bright (at lower predation risk), males from low predation risk populations preferred unfamiliar females only during brighter but not dim conditions.  Haley 1999) sometimes affects chooser's encounters with potential mates; it might be a way that chooser's exploit the preexisting sensory biases of those they choose between, increasing their own encounters with potential mates by bringing potential mates to them in a passive way (West-Eberhard 1984;Rodd et al. 2002). We categorized studies about the attractiveness of a chooser's territorial resources under probability of encounter with potential mates e. Density (McLain 1992) is a complex variable indicating the number of individuals in a given area; thus, it is possible that the salient inducers of change from rejecting to accepting potential mates are encounter probability e or the number of potential mates n. Whenever survival is negatively density dependent, parsing of density effects may require attention to the chooser's probability of survival s as well. Mate choice copying is a phenomenon that students of lekking species frequently speculated about even before Losey et al.'s (1986) model or the carefully controlled experiments of Dugatkin (1992a). We agree with Losey et al. that a key to the puzzle of mate choice copying is the fitness that would be conferred on a chooser versus copier females. Thus, we categorized mate-copying studies under the fitness that would be conferred, w-distribution. Table 2 categorizes the 182 studies that reported evidence of phenotypic plasticity in "choosy" and "indiscriminate" behavior. Fifty-four of the studies are on insects, eight on crustaceans, 13 on arachnids, 73 on fishes, seven on amphibians, four on lizards and snakes, 21 on birds, and two on mammals. Twenty-eight of the studies were field studies, 150 were laboratory studies and four studies combined both laboratory and field components. One hundred and sixty-nine were manipulative experimental studies and thirteen were observational studies describing how variation in ecological or social circumstances correlated with changes in focal subjects' "choosiness." One hundred and twelve studies focused on females' mate choice behavior, 34 on males' mate choice behavior, and 36 on both sexes. Within-sex phenotypic plasticity (Table 2) occurred under changes in the OSR (12 studies), the ASR (three studies), population density (six studies), habitat/diet quality (15 studies), encounter rate (five studies), predation risk (20 studies), and changing cost of mate search (four studies). Also linked to within-sex changes from choosy to indiscriminate mating behavior are: chooser's condition (21 studies, seven of which were on parasite load), chooser's body size (11 studies), chooser's resources (two studies), chooser's attractiveness (one study), chooser's experience (67 studies), mating status, that is, virgin/ mated (four studies), and chooser's age (eight studies). Both sexes investigated in the study but changes in focal subject "choosiness" were found in only one sex as described in the  Guppies, Poecilia reticulata Laboratory Inbreeding no effect. Inbreeding level does not induce female discrimination between sibs and unrelated males in guppies, that is inbreeding level did not influence female preference for unrelated males ♀

Laboratory
Results illustrate both the variability among populations that prevents results obtained from one population from being generalized to the entire species, and the plasticity of sexually selected traits in relation to local conditions. Heavily parasitized females (field caught) take less time to first visit males than lightly parasitized or unparasitized females in one population, but reversed pattern in one other, and no effect in three populations. We provisionally categorized Table 2 studies under the SPT parameters: 84 under the focal chooser's probability of survival s; 55 under chooser's probability of encounter with potential mates e, 11 under the number of potential mates n, 10 under the duration of postmating time-outs or latency o, and 76 by the w-distribution (and we categorized some studies under multiple SPT parameters).
In addition to the Table 2 studies, we found 16 studies ( Table 3) that looked for but did not find evidence of phenotypic plasticity.

Discussion
From Table 2, one may conclude that within-sex phenotypic plasticity in choosy versus indiscriminate behavior is (1) easily manipulated in laboratory studies and relatively easy to observe under naturally occurring conditions in the field. (2) Within-sex phenotypic plasticity is common, widespread taxonomically occurring in insects, spiders, crabs, birds, amphibians, and fish. (3) Individuals of "the choosy sex" were phenotypically plastic in tested species and sometimes at random with respect to traits in the chosen sex usually suspected of evolving under sexual selection. Of the 198 studies reviewed here, only 16 (Table 3) reported no within-sex phenotypic plasticity in reproductive decision-making. Either within-sex phenotypic plasticity is very common, at least wherever investigators look for it, or there is some remaining publication bias against studies that looked for but did not find within-sex phenotypic plasticity of reproductive decisions. Almost all studies provisionally fit one or more of the SPT's simplifying parameters of survival probability s, probability of encountering potential mates e, the number of potential mates in the population n, the duration of postmating time-outs or latencies to receptivity o, and the distribution of fitness that would be conferred w-distribution.
Chooser's traits and their social and ecological circumstances matter Table 2 shows that in a variety of species and situations, variation in subjects' mating decisions depends as much or perhaps more on choosing subjects' intrinsic variation and their ecological and social situations as on the apparent variation in those being chosen between (which receives the most attention in studies of the evolution of traits, mostly in males). The review shows that choosers, both female and male, can and do modify their reproductive decisions under ecological and social contingencies that often have little or nothing to do with traits of chosen individuals. Most notable perhaps was that the ecological conditions or intrinsic variation of the subjects doing the choosing seemed often to matter more than characteristics of individuals among whom choosers' discriminated.
The SPT explicitly makes the point that the social and ecological situation (expressed as the focal individual's probability of survival s, probability of encounter with potential mates e, the duration of any postmating timeouts o, the number of potential mates in the population n, and w-distribution, Fig. 1) of the focal individual determines their reproductive decisions to accept or reject a given potential mate. The SPT not only changes the subject to individuals making reproductive decisions, but also takes attention off of the supposedly sexually selected traits of those being chosen. Instead of the usual attention to fancy male traits that supposedly exploit female choosers' sensory biases, the SPT refocuses on the fitness that would be conferred on the chooser under time costs rendered by variation in chooser's own likelihood of survival, their probability of encounters with potential mates, and the durations of any postmating time-outs they have experienced. Paying attention to choosers' probability of survival brings attention to chooser's health status, chooser's parasite loads, chooser's current risk of being predated, and chooser's fancy traits that increase their probability of encounters with potential mates, which the studies here have also attended to but from different organizing perspectives than that of the SPT.
The parameters: the focal individual's probability of survival s, probability of encounters with potential mates e, the number of potential mates in the population n, the duration of any postmating time-outs o, and the distribution of fitness that would be conferred, the w-distribution, either separately or together (Table 1), may explain the results of a large number of the studies in Table 2. The SPT argues that a specific type of trade-off underlies switches between acceptable and unacceptable mates, namely the trade-off between an individual's real-time fitness accrual and time available for mating. And, the SPT parameters are simpler than complex proxies such as OSR, "predator risk," or "sperm competition risk." For example, OSR may not adequately capture the inducing variable of individual behavior: Even if there are more sexually available females than males or males than females, the rarer sex may flexibly adjust reproductive decisions given their own probability of survival s, their probability of encounters with potential mates e, the number of potential mates in the population n, or the wdistribution. Likewise, predator risk is complex, because the number of potential mates n may decline in the presence of a predator, an individual's immediate survival probability s surely decreases, and their encounters with potential mates e may decline. The SPT has shown that there is a hierarchy among inducing variables that have an order of magnitude or more effect on expected lifetime reproductive success. For example, individuals' survival probability, s, has a relatively huge effect compared to probabilities of encounter with potential mates on expected mean lifetime mating success and individuals' reproductive decisions: Theoretically, when an individual's probability of survival s declines, choosy individuals may accept potential mates that were previously unacceptable. Similarly, if an individual's encounter probability e with potential mates increases enough, they may reject potential mates that were previously acceptable. Sensitivity analyses (Gowaty and Hubbell 2009) of the SPT indicated that variation in survival probability theoretically has a significantly bigger effect than encounter probability on the expected mean lifetime mating success of an individual and thus on their switch points. In the case of the complex proxies reviewed herein, we argue that what may be inducing changes in individuals' reproductive decisions may be their probabilities of survivals, probabilities of encounters with potential matese, the durations of their postmating time-outso, variation in the number of potential matesn, or the w-distribution.

A sea change from Darwin's observations of coy females and indiscriminate males
Widespread evidence of within-sex phenotypic plasticity in choosy and indiscriminate behavior is quite different (Knight 2002) from Darwin's observation of coy females and indiscriminate males. In contrast to expectations from anisogamy and parental investment hypotheses, within-sex phenotypic plasticity in reproductive decisions to accept or reject potential mateschanges according to availability of potential mates, ecological conditions, social interactions, and the attractiveness and/or health status of the choosing individual or the resources available to the choosing individual. Yet, so far we can only safely conclude that subjects of the expected or usual "choosy sex" differ, which allows the conclusion of phenotypic plasticity among individuals in "the choosy sex", but says nothing about individuals in "the indiscriminate sex." How often is the "indiscriminate sex" phenotypically plastic? An empirical answer is unlikely as long as investigators confine their questions to individuals of "the choosy sex," because observations of phenotypic plasticity in one sex are uninformative about the other sex.
Most if not all of the observations of and conclusions about within-sex phenotypic plasticity are inconsistent with parental investment and anisogamy theories predicting fixed sex differences in mate choice behavior because of ancient selection pressures on gamete morphology or parental investment patterns.
We argue for consideration of an alternative possibility that within-sex plasticity is actually individual flexibility, meaning that individuals rather than being "fixed" use the same rules of induced behavior, and alter their reproductive decisions regardless of their sexes, to fit current demographic and social circumstances, so that their behavior is individually flexible and adaptive in their current environments.
Are there fitness consequences of switching between choosy and indiscriminate mating behavior?
The answer is impossible to tell from the studies in Table 2. The SPT (Gowaty and Hubbell 2009) like the earlier H&J mating model (1987) predicts the expected reproductive success for an individual switching from accepting to rejecting potential mates. Thus, it is worth mentioning that very few investigators have tackled empirically questions about fitness costs and benefits to individuals that switch from choosy to random mating, which is particularly interesting given that we were able to categorize 42% of the reviewed studies as provisionally being explained by variation in the fitness that would be conferred, that is the w-distribution. We find most curious the absence of attention to the fitness payouts or relative fitness costs to benefits of switching in the remarkably sophisticated empirical studies of mate choice copying. Losey et al. (1986) said almost 30 years ago ". . .copying may be advantageous. Copying may also result in all the other females copying a poor choice. We can find no empirical or theoretical demonstration of the relative fitness of these two alternatives strategies in the literature (p 654)". Losey et al. then provided a theoretical answer: if chooser females do choose fitter males and if choosers are more common than copiers, the fitness differences between chooser and copier females may be a wash. Since then, a game theoretic model has demonstrated that the adaptive significance of mate choice copying could reside in the ratio of costs and benefits to "active mate searching" (Pruett-Jones 1992), which the SPT expresses as variation in probabilities of survival s and/or probability of encounters with potential mates e or s and e. And, the question remains: Are there fitness consequences of being a chooser versus a copier?

Implications of variability in methods of laboratory and field studies
Methods and criteria for demonstrating variation in "the choosy sex" depend on the study species, whether the study was experimental or observational, field, or laboratory study (Table 2). By design, laboratory studies are relatively more similar in methods of evaluating individual subjects' decisions to accept or reject particular mates. Most laboratory studies of mate preferences depend on controlled testing conditions to exclude the simultaneous operation of within-sex interactions and between-sex coercion on the "preference" behavior of subjects (Kingett et al. 1981;Lambert et al. 1982); most use measures of approach (counts) or duration of time in proximity to assign relative preferences of a focal individual. In contrast, the operational definitions (Box 1) indicating "mate choice," which investigators use under more complex social conditions where it is difficult or impossible to eliminate same-sex interference or sexually coercive behaviors, are far more variable than the operational definitions used in laboratory studies. Thus, in practice, investigators use different definitions of "mate choice" and "mate preferences". In Table 2, some studies used proximity to a potential mate or to alternative potential mates, moving toward potential mates, time near alternative potential mates, "courtship signals," or mating with alternative potential mates, time from introduction to courtship or mating. Investigators usually attempt to adjust any behavioral indicator of a given reproductive decision to the behavior of the study organism, but these different definitions present a number of problems particularly for consistency: Do we always measure what we mean to measure? How comparable are different measures of mate choice to one another? For example, measuring preference as positive if copulation occurs is a motor acta behaviorinvolving interactions with potential mates, which may be quite different from a focal individual's cognitive assessments of fitness that would be conferred by the mating.

Is mate assessment ("choosiness") costly?
A common definition of "choosiness" is the effort an individual is prepared to invest in mate assessment (Jennions and Petrie 1997), which assumes that assessment is often costly. The SPT explicitly assumes that any "effort" associated with cognitive assessment occurs before individuals enter receptivity the first time, or that assessment of encountered potential mates is without time costs, happening immediately on encounter with potential mates. This explicit assumption is embedded in the SPT, because there is no time-eating state prior to encountering a potential mate during which assessment of alternative potential mates occurs. Gowaty and Hubbell (2009) informally assumed that individuals could make assessments of potential mates using information and criteria about themselves and others gained during prereproductive life stages, well before the onset of first-time receptivity to mating. Their simplifying assumption kept the SPT simple. Nonetheless, modifying the SPT to include a time-eating state of assessment that individuals could enter after encounter and before mating would under most conditions lower expected mean lifetime mating success and, depending on variation in the other parameters, such as the w-distribution, increase the number of potential mates an individual would rank as acceptable, that is, it would move the switch point toward potential mates conferring lower fitness. In other words, the current structure of SPT suggests that after entering receptivity for the first time, investing time in assessments would likely be selected against. Future, modified, more complex versions of the SPT could include a assessment state after encountering a potential mate but before accepting or rejecting a potential mate. We could compete the new results with the results from our current versions to evaluate our intuition that selection would act against a time-eating assessment state.

Recommendations for the future
What is the extent of within-individual variation in reproductive decision-making, that is, in individuals of both sexes? Is it trivial in one sex, occasional or common in both sexes? Needed are study designs that include individuals independent of their expected parental investment patterns or gamete size (Naud et al. 2008). How often do individuals in "the indiscriminate sex" switch from indiscriminate to choosy? We recommend that investigators adopt (1) a symmetrical approach to simultaneously evaluate in both sexes within-individual flexible reproductive decision-making behavior, and (2) a more operational approach that differentiates cognitive processes from motor acts when studying variation in "choosy" and "indiscriminate" behavior (Box 1). In addition and most important, (3) we need information on whether flexible mate choice affects individual fitness (McLain 1991) is observed individual flexibility adaptive or not (McLain 2005)? Finally, (4) more investigations are needed in many more species to find out how and when males reject potential mates (Bateman and Fleming 2006).