Printer Friendly

Mate choice in the sailfin molly, Poecilia latipinna.

Evolutionary biologists have demonstrated repeatedly the importance of sexual selection as a major force molding the distributions of secondary sex characters and male behaviors in species of insects (Claridge et al. 1988; Ritchie 1991), salamanders (Verrell and Arnold 1989; Verrell 1995), frogs (Ryan and Wilczynski 1991; Gerhardt 1991; Ryan et al. 1992b), birds (Moller 1988, 1989; Andersson 1982, 1993), and fishes (Moodie 1982; Rowland 1989a, b; Warner and Schultz 1992; Basolo 1990a,b, 1995; McPeek 1992; Downhower and Lank 1994). In particular, sexual selection has shaped character variation among intraspecific populations through either local variation in the traits that are the targets of sexual selection (Sappington and Taylor 1990; Endler and Houde 1995) or variation in the nature of selection on the same trait (Endler 1983; Kodric-Brown 1985, 1989; Breden and Stoner 1987; Houde 1987, 1988a, b; Stoner and Breden 1988; Long and Houde 1989; Houde and Endler 1990). Although the importance of sexual selection is clear, its precise role often is not, especially its role in molding variation in suites of correlated traits within and among populations.

There are many systems in which several behavioral and morphological traits vary concordantly among individuals in a population, often as a result of shared genetic control (reviewed by Travis 1994b). The genetic variation for these complexes is protected by a balance between natural selection and sexual selection (Ryan et al. 1992a; Shuster and Wade 1991). Although sexual selection favors some multivariate phenotypes over others in these cases, it is often unclear which specific characters are the targets of that selection (but see Endler and Houde 1995; Wiernasz 1995). This point is critical because in many cases some of the characters in the suite vary in their distribution among populations while others do not, yet the high covariances among individuals within populations are preserved (Travis 1994a,b). Here we wish to understand how sexual selection promotes population divergence in a subset of the traits involved in these complexes.

Patterns of character variation within and among populations of the sailfin molly, Poecilia latipinna, suggest just such a role for sexual selection (for fuller reviews, see Travis 1994a,b). There is extensive variation in adult male body size based on variation in the size at maturity. Size at maturity is determined by a Y-linked allelic series (Travis 1989a, 1994b). Male sizes vary within populations, sometimes by up to a factor of three. Populations throughout the species range exhibit repeatable, extensive variation in these body-size distributions; average male size can vary over twofold among populations. Several traits covary strongly with body size within any single population. For example, larger males mature much later and have disproportionately longer and higher dorsal fins and disproportionately smaller gonopodia (the modified anal fin that serves as an intromittent organ) and testes. In most cases larger males exhibit disproportionately higher rates of courtship displays to females and attempt to inseminate females without any courtship (gonopodial thrusting) at disproportionately lower rates. The genetic variation for size-independent behavioral rates within a population is low. However, the tight covariances appear to have a strong genetic basis in the Y-linked control of size (Travis 1994b).

The covariances of trait means among populations, however, do not reflect the covariances among individuals within populations, because some of the traits vary more widely among populations than others. In particular, males of the same body size from different populations can have strikingly different values for some of the traits. For example, males from different local populations in northern Florida exhibit over threefold variation in the size-specific rates of courtship displays and forced insemination attempts (Ptacek and Travis 1996). Similar variation exists in morphological traits used in the courtship display: males of the same size from different local populations can vary nearly twofold in dorsal fin length and height (Travis, Ptacek, and Martin, unpubl. data). Moreover, the mean values of these traits, adjusted for body size, are not correlated among populations, whereas there is a strong partial correlation between these traits (body size held constant) within populations (Travis 1994a).

In this paper we describe the action of sexual selection in sailfin mollies in order to evaluate its potential role in promoting the evident breaking up of tight covariances that has occurred as molly populations have diverged. In particular, we report on the roles of female and male mate choice and discuss those roles in the context of other aspects of the mating system that have already been reported (Farr and Travis 1986; Travis and Woodward 1989; Sumner et al. 1994; Travis 1994b). We synthesize these results to suggest that mate choice plays a prominent role in molding the patterns of character variation within and among molly populations.

MATERIALS AND METHODS

Molly Mating Behaviors

Male mollies direct three principal behaviors to females. In the courtship display, which is an attempt to elicit female cooperation in mating, the enlarged dorsal fin is erected and presented to the female often accompanied by a sigmoid curving of the body and a tilting of the body toward the female (Parzefall 1969, 1979; Farr et al. 1986). Gonopodial thrusting is an attempt at forced insemination without female cooperation; a male orients himself behind a female, brings his gonopodium to a forward position, and attempts to insert it forcefully into the female's gonopore. A third behavior, gonoporal nibbling, is an attempt by a male to make nasal or oral contact with the female's gonopore. Its complete function is unclear, but it appears to aid a male in determining a female's reproductive condition (Farr and Travis 1986; Travis and Woodward 1989). Males of all sizes show all three behaviors, but in general, larger males engage in courtship displays more often and resort to gonopodial thrusting less often than smaller males (Farr et al. 1986; Ptacek and Travis 1996). The behavioral profile of an individual of a particular body size from a specific population is highly repeatable (Travis and Woodward 1989).

Origin and Maintenance of Experimental Fish

We conducted five experiments designed to assay mate choice in sailfin mollies. We collected sail fin mollies from two populations, Lighthouse Pond and Live Oak (Wakulla County, Florida), during the spring seasons from 1986 to 1990 for experiment 1. These fish were used to create and supplement stock populations from each locality that were maintained in the laboratory. We collected fish in February 1993 from three populations, Mounds, Pinhook, and Live Oak (Wakulla County, Florida) for experiment 2. We collected as random a sample as possible of 40 males and 40 females from each population, from which we haphazardly selected a subset for experimentation. We collected mollies from Melanie's Pond (Franklin County, Florida) in June of 1987 for experiment 3. For each of experiments 4 and 5 we collected as random a sample as possible of 40 males and 40 females from each of Mounds and Live Oak populations on two separate occasions. The fish used in experiment 4 were collected in September 1994, and the fish used in experiment 5 were collected in September 1995.

We maintained fish in the laboratory in 75-L aquaria, approximately 10 males and 10 females per aquarium, at a salinity of 6 parts per 1000 and a 14:10 light:dark photoperiod at 25 [degrees] C. We kept individual fish for three to six weeks before assaying male and female responses. Fish used as "choices" were never housed in the same aquarium as the "tested" individuals of the opposite sex to whom they were presented, so no familiarity through acclimatization to laboratory culture could be established between individuals whose interactions we would subsequently monitor. We measured all of the fish to the nearest millimeter standard length (straight-line distance from the tip of the snout to the end of the last vertebra at the base of the caudal fin) to estimate body size distributions for each population.

Observations of Behavior

Behavioral observations in all experiments were made between 0800 and 1400 hours. We tried to minimize disturbance to the fish due to the presence of the observer by covering three sides and the bottom of all observation aquaria with an inner layer of black paper (to minimize reflections from the glass) and an outer layer of aluminum foil. The front glass of each aquarium was covered with one-way film. After each trial of each experiment, water was drained and the test aquarium was washed with fresh water.

We kept the males in isolation for 24 hours prior to testing in all experiments except experiment 4. This protocol appears to motivate males of all body sizes in a similar fashion and yields rates of behaviors typically observed in natural populations (Travis 1994b). The focal males used in experiment 4 were kept isolated from females between tests but were housed with a social group of two other males (one smaller and one larger) to maintain similar levels of behavior over the four consecutive days of testing. Previous work has shown this procedure to induce consistent behavior across days (Travis and Woodward 1989).

We controlled for reproductive history and receptivity of females in the female-choice tests (some trials of experiment 1 and all trials in experiment 2). Female mollies are receptive for a two- to three-day period either upon initial maturation or immediately after producing a brood, after which they are unreceptive to fertilization for 28 to 35 days (Farr and Travis 1986; Snelson et al. 1986; Travis 1989a). Virgin females can only be obtained from laboratory rearings. We controlled for receptivity in nonvirgin females by maintaining individual gravid females in 3.75-L aquaria until they gave birth; these females were tested within 24 hours of releasing their broods.

Only visibly gravid, nonreceptive females were used as object females in the male mate-choice experiments (experiments 3-5). Female mollies signal their receptivity immediately following release of a brood through some type of direct contact cue (Farr and Travis 1986; Sumner et al. 1994). The chemical nature of this cue is unknown and the amount of variation among females within or between populations in this advertisement of receptivity has not been quantified. Once females have been inseminated, males lower their rates of sexual behavior, yet still show all three mating behaviors [TABULAR DATA FOR TABLE 1 OMITTED] at rates that their body size would predict (Sumner et al. 1994). We used only gravid females in the male choice experiments in an attempt to reduce variation in male behavior due to variation in female condition. The use of nonreceptive females also assured that male choice was not confounded with female choice in these direct contact experiments (experiments 3-5).

We conducted female choice tests (experiments 1 and 2) in a 75-L aquarium (122 x 32 x 52 cm) that was divided into five sections of equal size. The two sections at the ends of the tank were partitioned with panels of clear plexiglass with slits at the top of each that allowed water flow. These two end sections constituted the two male compartments. The three central sections were delineated only by markings on the front side of the tank so that the observer could accurately record the movements of the female in this compartment. The female was able to move freely among these three central sections.

The treatment combinations were designed to control for any biases by females toward one or the other side of the choice tank. Some females in experiment 2 saw the same male in more than one combination; other females did not, depending upon the size of the object males. Object males were chosen to match the size of the female [+ or -] 10 mm in experiment 2. This was done to maximize the potential for male-female interactions. Previous observations have shown that large males direct little attention toward very small females and conversely, very small males are less likely to approach very large females (Travis 1994b).

The protocol for the female-choice observations was the same for both sets of experiments (experiments 1 and 2). The two object males were first placed in the choice tank, one in each of the end compartments, and given a 15-min acclimation period. The test female was then placed in the center compartment and given a further 15-min acclimation period during which she was allowed to move freely about the central compartment. The behavioral observations were then conducted for 10 min. The observer recorded the time in seconds the female spent on the left or right side of the tank and the time in the neutral central section of the tank, and for experiment 2, the number of courtship displays exhibited by each object male when the female was on his side of the tank.

The protocol for the male-choice observations differed between the three experiments. A single male was placed in a 37.5-L test aquarium with one large and one small female in experiment 3. The two females could always be distinguished by the observer. We tested each male in experiment 4 on four consecutive days with one of four different female phenotypes that differed in size and population of origin. A single male was placed in a 37.5-L test aquarium with a single female of a particular phenotype for each of the four trials. In experiment 5, a single male was placed in a 37.5-L test aquarium with two females of equal size, one from the male's own population and the other from a different population. We injected a small dot of either red or black acrylic paint subcutaneously (e.g., Thresher and Gronell 1978) near the top of the caudal peduncle on each side of the female to distinguish between the two females during the trial. Half of the females used as object native females were marked with red and the other half with black; the same was true for foreign females. The mark was retained by most females throughout the course of the experiment and marked females appeared to suffer no long-term effects from the treatment.

The protocol for the behavioral observations was similar for all three experiments (experiments 3-5). We placed the test male in the test aquarium at the beginning of each trial and allowed a 15-min acclimation period. We then introduced the two object females (experiments 3 and 5) or the single object female (experiment 4) and allowed another 15-min acclimation period. We observed males for 30 min in experiment 3 and 10 min in experiments 4 and 5. During each observation we recorded the number of courtship displays, gonopodial thrusts, and gonoporal nibbles that the test male performed to each object female (experiments 3 and 5) or the single object female (experiment 4).

Analysis of Experiments

Female Mate Choice

In experiment 1, a total of 57 females from two different populations were tested, 39 from Live Oak and 18 from Lighthouse Pond. We analyzed data from each of the five repetitions (= series) of this experiment separately; females in different series varied in population of origin and several other attributes (Table 1). Within each series, an individual female's preference score was calculated as the time spent near the larger male minus the time spent near the smaller male. Scores could range from -600 to +600 (the observation period was 600 s); a diagnosis of choice for each series was made if the average score was significantly positive by a one-tailed Student's t-test. We used a one-tailed test because all previous studies of size preference in poeciliids, including P. latipinna, have demonstrated preference for larger males (Ryan and Wagner 1987; Zimmerer and Kallman 1989; Basolo 1990a; Ryan et al. 1990; McPeek 1992; Schlupp et al. 1994; Endler and Houde 1995).

We used the proportion of time spent in the neutral, center portion of the aquarium as a measure of a female's level of disinterest or "apathy." This measure can range from 0 to 1.00 and helps distinguish between females that move continually back and forth across the tank showing no clear preference for one male over the other (average preference score and "apathy level" near zero) and females that spend most of the trial in the center of the tank, rarely interacting with either male (average preference score near zero but "apathy level" high).

The t-tests for significance were performed on raw preference scores. When a significant preference was determined, we examined the correlation of preference score with various attributes of the males and the individual females with standard product-moment correlations. No transformations were employed; all assumptions of parametric tests were met.

In experiment 2, 12 females from each of three populations (Mounds, Pinhook, and Live Oak), were each tested three times with three different combinations of males matched for body size (two combinations included a native male, and in one combination both males were foreign). We used a repeated-measures, "crossover" design (Cochran and Cox 1957; Travis and Woodward 1989), which allowed us to test for any effects of order of observation (first, second, or third combination experienced by a female) or sequence in which the combinations were offered.

We measured a female's preference as the difference between the time spent near the native male and the time spent near the foreign male. In those combinations where both males were foreign, the time spent near the male on the right side of the tank was subtracted from the time spent near the male on the left side of the tank. Each female's level of "apathy" was calculated.

We analyzed data from this experiment in three steps. First, we performed an analysis of variance on the preference scores of the 108 trials to test for the effects of order and sequence. We examined residuals from the raw preference scores to verify that assumptions of parametric tests were met. Second, after verifying that these effects were not significant, we used analysis of variance to test preference scores for effects of a female's population of origin, female identity within population, the combination of males offered, and the interaction of the combination of males with a female's population of origin; the female x treatment interaction serves as the residual error term in this analysis. Third, after finding a significant interaction of population of origin with male combination, we examined the nine individual average preference scores for significant deviations from zero using two-tailed t-tests to diagnose which combinations induced females to exercise a preference. This procedure follows the spirit of Fisher's Protected LSD test.

Male Mate Choice

Thirty-seven males were each tested once with a pair of females from the same population in experiment 3. Although we recorded all behaviors exhibited by each male to each female, we used the proportion of all gonoporal nibbles directed toward the larger female as a measure of male choice. We chose to analyze this male behavior as a measure of preference for two reasons. First, prior data indicated that gonoporal nibbling rate is associated with a male's level of interest (Sumner et al. 1994). Second, this behavior varies the least as a function of male size in this population (r = -0.16, P [greater than] 0.10) and in other populations (Farr et al. 1986; Ptacek and Travis 1996). It is arguable that male thrusting rates are better measures of male preference than rates of gonoporal nibbling. The dilemma with using thrusting rates is that they vary inversely with male body size; this inverse relationship makes this metric a powerful one for use with smaller males but potentially a weak measure for larger males, who offer relatively few thrusts. While we believe that rates of gonoporal nibbling provide a metric with comparable statistical power for males of disparate sizes, we will report the results from tests of male preference using both behaviors.

We tested the proportion of nibbles or thrusts directed toward the larger female for a significant deviation from no preference, or a proportion of 0.50, with a one-tailed Student's t-test. The directional alternative for choice at the cohort level was based on our earlier results that demonstrated male preference for larger females (Sumner et al. 1994; Travis 1994b). Proportions were subjected to the arcsine transformation.

We used regression analysis to examine whether a male's preference score could be depicted as a function of his own body size (on a natural logarithmic scale), the ratio of the sizes of the available females (expressed as the difference between the natural logs of the sizes of the two females), or the ratio of the size of the male to that of the larger female (expressed as the difference in natural log sizes). Forward and backward multiple regression and all-possible-subsets regression all yielded the same best fit equation as diagnosed by Mallow's Cp-statistic (Draper and Smith 1981), so we present only that equation in the results. We chose the particular transformations used in this and all of our other analyses of male mate choice by trying several variations and choosing the transformations that yielded the best pattern of residuals to meet assumptions of parametric tests.

In experiment 4, we tested 26 males from each of the Mounds and Live Oak populations to quantify their responses to four different types of females: large native female, large foreign female, small native female, and small foreign female. Small females were less than 33 mm and large females were greater than 40 mm for both populations. We used a repeated-measures crossover design, similar to experiment 2, so that we could test for the order of treatment presentation and the sequence of treatment presentation.

Test males were divided into three distinct size classes. "Small" males were defined in each population as those more than 1.5 standard deviations below the average standard length of males in that population. "Large" males were more than 1.5 standard deviations above the average standard length. "Medium" males were within 0.5 standard deviations above or below the average standard length. We used these categories in an analysis of variance, instead of simply measuring male size and using it as a covariate, to maximize power to assess whether males of different sizes respond to differences in female phenotype in a predictable pattern. We performed analyses of variance on the natural logarithm of the rate of each male mating behavior, as well as the natural logarithm of the sum of all behaviors, following the protocol described above for experiment 2.

In experiment 5, we tested 20 males from each of the Mounds and Live Oak populations to see if they would distinguish between native and foreign females matched exactly for body size. To increase our power to detect size-dependent choice patterns in males, we chose them nonrandomly, testing the eight largest males in each population, the eight smallest, and four intermediate in size (for the rationale for this procedure, see Ptacek and Travis 1996). Our design incorporated the ability to test for any effect on the behaviors of males of the acrylic paint mark placed on the females. The males from the Mounds population ranged in size from 25 to 45 mm SL, whereas those from Live Oak ranged from 25 to 51 mm SL. Object females were chosen to match the size of the test male [+ or -] 10 mm. Half of the males from each population were tested with the native female marked with a black paint dot and the foreign female marked with a red paint dot; the markings were reversed for the other half of the males.

We analyzed the natural logarithms of the rates of each behavior, as well as the natural logarithms of the total behavioral effort, as an incomplete block design with two treatments, female identity and color (Box et al. 1978). Each male was considered a block; in this case, the "block" effect accounts for the covariance between each male's responses to his two females, analogous to removing the "subject" effect in a repeated-measures design. The treatment effects must be considered additive because all interactions are confounded with block effect. This confounding is a direct consequence of how the color markings had to be used.

Rationale of Experiments

We conducted two types of female-choice tests. Experiment 1 was designed to assay female preference within a population by giving females from two different populations a choice between males of different size, each from the same population as the female being tested. This test does not isolate a preference for a specific character because several characters covary with male body size within a population. Experiment 2 was designed to assay female preference among populations by offering females from three different populations a choice between two males of the same body size presented in three different population combinations, two that included a male from the same population as the female being tested and one where neither male was from the female's population of origin. This experiment was designed to investigate whether females would discriminate against foreign males and whether any such discrimination was a simple function of divergent preferences for male trait variation.

We conducted both of these experiments using a choice tank design where males were partitioned from direct contact with females and each other. The design allowed for visual and diffuse water-borne chemical cues (water flow was allowed between all three sections) to maximize cues available to females from object males. The partitions enclosing the object males kept them from interacting with each other; a male only performed courtship displays when the female was adjacent to his compartment. We used this design to keep female choice from being confounded with male-male competition.

We performed three sets of experiments designed to establish whether male mollies exhibit mate choice. The first male-choice experiment, experiment 3, was designed to assay male preferences within a single population for females of different body size but common reproductive condition. Experiment 4 was designed to assess whether males alter their rates of various mating behaviors in response to native and foreign females of different sizes when presented individually. Given that substantial evidence suggests males will exercise a choice among females from their own population (Sumner et al. 1994; Travis 1994b), they may also discriminate between native and foreign females, and such discrimination would not be evident when a male was presented with only a single female as in experiment 4. Experiment 5 tested this hypothesis by testing males from two different populations with two females of equal size, one native and one foreign, presented simultaneously.

Male-choice tests were conducted with males and females in direct contact. Males attempt gonoporal nibbles and gonopodial thrusts only when in direct contact with females and perform only courtship displays when partitioned from females in choice-tanks (Sumner et al. 1994). We chose a direct contact design over a choice-tank design for our experiments because male choice may incorporate changes in any of the three behaviors and males of different sizes rely on different behaviors when interacting with females (Farr et al. 1986; Ptacek and Travis 1996). We wanted to maximize our ability to detect preference for particular female phenotypes due to adjustments by males in any of the three mating behaviors. Female-female interactions were minimized by our use of gravid, nonreceptive object females. Females used in these experiments interacted minimally with either the focal male or each other.

RESULTS

Female Mate Choice: Experiment 1

Nonreceptive females did not exercise any preference (Table 1, series 3 and 4); average preference scores for these females were low, variance was low, and the apathy levels were very high. These patterns contrast with those seen in receptive females from the same population (series 1 and 2), where average preference scores were four- to eightfold higher, and apathy levels were, at most, only half as high.

In the three series with receptive females where choice was evident, females preferred larger males; average preference scores were high and apathy levels were low. The strongest preference was evident in laboratory-raised virgins from Live Oak (series 1), and the weakest in wild-caught experienced females from Lighthouse Pond (series 5). Larger females exhibited stronger preferences in series 1 (r = 0.66, P [less than] 0.05) [TABULAR DATA FOR TABLE 2 OMITTED] but not in any of the others (correlations of 0.11 and 0.32 for series 2 and 5, respectively). The amount of size difference between the males appeared not to exert any substantial influence on the level of preference; the correlations for series 1, 2, and 5 were -0.42, -0.12, and -0.08. In series 1, stronger preferences were associated with a larger ratio between a female's size and that of the larger male (r = 0.74, P [less than] 0.05): the larger the female relative to the larger male, the stronger her preference.

Female Mate Choice: Experiment 2

Female preference scores showed no effect of the order of observation (F = 1.96; df = 8, 60; P [greater than] 0.05) or of the sequence in which treatment combinations were experienced (F = 0.81; df = 2, 60; P [greater than] 0.05). We therefore partitioned the variance in preference scores among four sources. Neither female identity (F = 1.41; df = 33, 66, P [greater than] 0.05) nor population of origin (F = 0.32; df = 2, 33; P [greater than] 0.05) consistently affected preference scores, nor was there any strong main effect of male combination (F = 2.81; df = 2, 66; 0.10 [greater than] P [greater than] 0.05), but the interaction between population of origin and male combination was highly significant (F = 5.05; df = 4, 66; P [less than] 0.001). This suggests that each specific male combination means something distinct to females from each population (Table 2).

Females, on average, spent a higher percentage of time with native males than with foreign males in some but not all combinations [ILLUSTRATION FOR FIGURE 1 OMITTED]. Mounds females significantly preferred native males when the foreign males were from Live Oak (t = 2.54, df = 11, P = 0.027) and although only marginally significant, preferred native males to males from Pinhook (t = 2.16, df = 11, P = 0.054) as well. Live Oak females significantly preferred their own males over Pinhook males (t = 5.81, df = 11, P = 0.0001), but showed no significant preference when the foreign males were from Mounds (t = 1.12, df = 11, P = 0.286). Pinhook females significantly preferred their own males to Live Oak males (t = 2.94, df = 11, P = 0.013), but showed no significant preference when the foreign males were from Mounds (t = 0.52, df = 11, P = 0.614). No preference was ever exhibited by females confronted with a choice between foreign males from two locations.

Females did not exhibit a general preference for higher rates of size-specific courtship display. There was no relationship between male size and courtship display rate for males from Mounds (slope = 0.14, ns), but males from Pinhook and Live Oak exhibited a significantly positive relationship (slope = 1.58 and 1.05, respectively, P [less than] 0.05). Although a full analysis of covariance is not warranted because of heterogeneity in slopes among populations, qualitative comparisons can be presented to support our claim.

Live Oak and Pinhook males of equivalent size exhibited similar numbers of courtship displays per trial (e.g., a 45-mm Pinhook male directed 3.4 displays per minute, whereas a Live Oak male of the same size directed 3.8 displays per minute), yet females from each population showed a strong preference for native males. Mounds females provided even more convincing evidence; even though a Mounds male of 45 mm directed only 2.6 displays per minute, fewer than either of the other two populations, Mounds females strongly preferred native males over Live Oak males, which would suggest they preferred lower display rates.

However, even divergent directional preferences for size-specific display rates cannot account for the results. For example, Live Oak females did not prefer native males over Mounds males, and if they preferred higher display rates, as implied by their choice of native males against Pinhook alternatives, they should have exercised a clear preference. If Pinhook females were choosing solely on the basis of lower size-specific display rates, as implied by their behavior when the alternative to their native males were Live Oak males, they should have preferred Mounds males to their own. The surest conclusion is that females are using some multivariate visual cue to distinguish native from foreign males, and for some population pairs those cues are not reliable.

Male Mate Choice: Experiment 3

Male mollies preferred larger females over smaller ones. They directed a significantly higher proportion of gonoporal nibbles to the larger of the two females (t = 4.87, df = 36, P [less than] 0.0001) as well as a significantly higher proportion of gonopodial thrusts to the larger female (t = 2.63, df = 36, P [less than] 0.025). Thus either metric supports the interpretation that males prefer larger females. Because of the correlative relationships with body size described previously, we restrict further analysis to the use of the gonoporal nibbling metric. The 37 males used in this experiment ranged from 26 to 67 mm SL; the larger females ranged from 43 to 65 mm SL and the smaller ones from 31 to 53 mm SL. The difference in size between the females available to a male ranged from 6 to 20 mm SL.

Unlike that of females, the variation among males in their level of preference is clearly associated with two attributes. Larger males had higher preference scores (t = 3.07, P [less than] 0.004; [ILLUSTRATION FOR FIGURE 2 OMITTED]), an effect that accounts for 21% of the variance. Moreover, preference scores increased as the ratio of female sizes increased (t = 2.13, P [less than] 0.05), which accounted for an additional 6% of the variance. The best fit regression of the transformed variables yielded a slope of 0.389 for the natural log of male size and a slope of 1.144 for the difference in natural logarithms of female sizes. This result is not likely due to female-female competition, since larger females were not observed to chase smaller females or keep them from interacting with males.

Figures 2 and 3 illustrate this result. We can construct an upper limit to a score that would indicate "no preference" by adding a value equal to twice the standard error in preference scores to 0.785, which is the arcsine-transformed value for 50% direction of behavior to each female [ILLUSTRATION FOR FIGURE 2 OMITTED]. For illustrative purposes, any score above this value would indicate that the male had a decided preference, and any score below it would indicate a lack of preference. With this preference score, the regression equation can be rearranged to describe a line on the axes of male size and the ratio of female sizes that separates the combinations of male size and the disparity in female sizes where a preference will be found from those combinations where it will not. Translated back to the original scale of measurement [ILLUSTRATION FOR FIGURE 3 OMITTED], the line indicates that larger males ([greater than] 50 mm SL) are much more discerning than smaller males, which is consistent with our prior results (Sumner et al. 1994; Travis 1994b). Larger males will exhibit a preference between females that differ by as little as about 8%, whereas smaller males ([less than] 20 mm SL) require over a 45% difference in female sizes before they exhibit a clear preference.

Male Mate Choice: Experiments 4 and 5

There was no evidence that males discriminate native from foreign females when presented with them either singly or in pairs. In experiment 4, where males were presented with single females representing four different female phenotypes, there was no evidence, for any of the three male behaviors, that either the order of presentation or different sequences affected the rates of behavior (data and six analyses of variance available upon request).

Males from the two populations tested in experiment 4 exhibited the same patterns of size-specific behavioral variation, as well as significant variation among males within size classes, that have been reported previously for each population (Travis and Woodward 1989; Ptacek and Travis 1996; [ILLUSTRATION FOR FIGURE 4 OMITTED]; Table 3). There was no significant effect of the different female phenotypes on the rates of any of the three behaviors for Live Oak males, nor were any of the possible [TABULAR DATA FOR TABLE 3 OMITTED] interactions between male and female phenotype significant for any behavior (Table 3). For Mounds males, there was no significant effect of the different female phenotypes on the rates of any of the three behaviors, but the interaction between male size and female size was significant for the rate of gonoporal nibbling. Smaller males directed more nibbles to small females, and larger males directed more nibbles to large females [ILLUSTRATION FOR FIGURE 4 OMITTED]. No other interactions were significant for any other behavior.

Total behavioral rates were not influenced by either male size or different female phenotypes in this experiment (Table 3). There is no evidence that males from either population alter the rates of specific behaviors or their total activity level in response to different female phenotypes. These results imply that preferences of females for native males observed in experiment 2 are solely a property of female choice and not of a specific male-female pair.

There was no evidence in experiment 5, where males were presented simultaneously with a native and a foreign female, that males from Mounds or Live Oak discriminate against foreign females. The two male populations did differ in the levels of behaviors directed to the different females, and the female's color mark did appear to affect male behavior.

Mounds males directed, on average, 26% more displays to foreign females than to native females and 37% more displays to females marked red than to females marked black ([ILLUSTRATION FOR FIGURE 5 OMITTED], Table 4). Of the six males who directed at least 50% more displays to foreign than to native females, five of these encountered the foreign female with a red mark. In contrast, Mounds males directed 55% more thrusts to females marked black than to females marked red [ILLUSTRATION FOR FIGURE 5 OMITTED]. There were no effects of female identity for nibbles, thrusts, or total behavioral effort, nor were there any effects of color mark on rates of nibbles or total behavioral effort (Table 4). Mounds males were heterogeneous for rates of each behavior; they were also heterogeneous for total behavioral effort.

Live Oak males exhibited remarkably little discrimination between females on any basis. The only statistically significant effect was one that paralleled a pattern seen for thrusting in Mounds males: Live Oak males directed 45% more nibbles to females marked black than to females marked red ([ILLUSTRATION FOR FIGURE 5 OMITTED], Table 4). Males from the Live Oak population were not significantly heterogeneous for any behavior; although this result is at odds with previous ones, it is in part attributable to the fact that we were forced to use smaller males from Live Oak than we used in previous studies to keep object females within 10 mm of the test males' size. It is very difficult to find females from Mounds that are larger than 40 mm.

DISCUSSION

Mate Choice In Sailfin Mollies Female Mate Choice

Nonreceptive females tested in choice tanks in experiment 1 showed no preference for either large or small males and had apathy levels two to three times higher than receptive females. This finding is in contrast to other studies of poeciliid fishes where females have shown consistent patterns of choice behavior regardless of receptive state (e.g., guppies: Houde 1987, 1988a; Houde and Endler 1990; Reynolds and Gross 1992; Dugatkin 1996; swordtails: Clark et al. 1954; Basolo 1990b; Morris et al. 1996; mosquitofish: McPeek 1992). Female sailfin mollies signal their receptivity to males upon maturation or immediately following release of a brood and this signal is known to produce alterations in the rates of male mating behaviors (Farr and Travis 1986; Sumner et al. 1994). The existence of fertility advertisement by receptive females of this species may explain why only receptive females showed preferences in our experiment. We know of only one other study that has investigated female choice in P. latipinna (Schlupp et al. 1994). This study also demonstrated female preference for larger males, but females were standardized for receptivity only by keeping them isolated from males for more than one interbrood interval (M. Ryan, pers. comm.). Further investigation of the influence of fertility advertisement by females during their receptive period on male mating behaviors and male-male interactions is warranted.

[TABULAR DATA FOR TABLE 4 OMITTED]

Our results demonstrate that sexually receptive females prefer larger males over smaller ones when both are from their own population. The preference of receptive females (virgins or lab-reared or wild-caught nonvirgins) for larger males was consistent among the cohorts that we tested; further work is necessary to clarify whether the strength of this preference is variable among populations. The preference for larger males that we observed is consistent with that described by other workers for P. latipinna (Schlupp et al. 1994).

Female preference for larger males appears to be a general characteristic of poeciliid fishes. It has been demonstrated for P. reticulata (Endler and Houde 1995), the gynogenetic species P. formosa (Schlupp et al. 1994; Travis, unpubl. data), several Xiphophorus species (Ryan and Wagner 1987; Zimmerer and Kallman 1989; Basolo 1990a; Ryan et al. 1990), and Gambusia holbrooki (McPeek 1992). In most of these examples, increases in male size are highly correlated with increases in courtship display rates and/or the size of dorsal and caudal fins. Ryan and Keddy-Hector (1992) have argued that females prefer the higher values of these associated traits rather than larger size per se. Further experiments are needed to determine if such a pattern is true for females of P. latipinna as well.

Females from the three different populations tested preferred native over foreign males in four of the six population combinations. Preferences for native over foreign males across all combinations were not always symmetric. For example, while the Live Oak-Pinhook results were symmetric, Mounds females preferred their own males over Live Oak males, but Live Oak females did not choose native males consistently when presented with the same alternative. These results suggest that females are using some multivariate visual cue to distinguish native from foreign males, and for some population pairs those cues are not reliable.

It is unclear which combination of size-specific traits were used by females to distinguish native from foreign males. Males from these populations differ in both size-specific behavioral rates (Ptacek and Travis 1996) and size-specific morphometry, particularly in the dimensions of the dorsal fin and gonopodium (Travis, Ptacek, and Martin, unpubl. data), and no differences in any single character can be invoked in post hoc fashion to account for the behaviors of females in our experiments. In general, molly populations differ by up to fivefold in some size-specific behavioral rates and up to two-fold in size-specific morphometric traits, remarkable magnitudes in the face of the extensive gene flow that connects these populations (Trexler 1988; Trexler et al. 1990). The mere existence of these differences implies some level of divergent sexual or natural selection on these trait values.

Our results do not rule out the possibility that females were merely choosing familiar male patterns over unfamiliar ones. Preferences for native over foreign males have been demonstrated repeatedly in guppies (Haskins and Haskins 1950; Crow and Liley 1979; Houde 1988a; Luyten and Liley 1991; Endler and Houde 1995) and have been observed in a variety of other taxa as well (reviews by Mayr 1963; Endler 1977). In these cases females appear to choose particular phenotypic traits that are characteristic of males from their own population, such as population-specific orange coloration patterns in guppies (Endler 1983; Houde 1987, 1988a,b; Long and Houde 1989; Kodric-Brown 1985, 1989; Endler and Houde 1995). An important role for sexual selection through female choice is still implied by this hypothesis: female choice imposes optimizing selection on these trait values (Verrell 1988; Travis 1989b).

Our results from the experiments on female choice thus have two possible interpretations that we cannot yet distinguish. First, females from different populations exercise divergent directional preferences for size-specific trait values that are associated with differences among males in those values. This result implies an active role for sexual selection in either causing or at least contributing to the maintenance of the behavioral or morphological distinctions among males. Second, females from different populations exert optimizing selection for different average size-specific trait values. This mechanism still implies an important role for sexual selection in maintaining trait differences among populations. These interpretations can be distinguished only by further experimental work that manipulates size-specific character variation in males, but our results thus far support Ryan and Rand's (1993) diagnosis of sexual selection and mate recognition as two facets of the single, general problem of mate choice (see also Verrell 1988).

Male Mate Choice

Males ought to mate selectively if females vary in fertility, if paternity can be assured with some confidence, or if there is considerable investment in mating, such as sperm are limiting or there is a high cost to mating in terms of search time or risk of predation (Ridley 1983). The life history of sailfin mollies suggests that at least two of these criteria apply: females differ dramatically in fertility (Travis 1989a; Travis et al. 1990), and male mollies exhibit phenomenally high rates of sexual behavior (Travis 1994b).

Male mollies exhibited mate choice within populations, preferring larger females over smaller ones, and larger males exercised a stronger preference than smaller ones. These patterns are consistent with two observations in the field: larger females are more attractive to males than are smaller females, and larger males spend more time associated with larger females than with smaller females (Travis 1994b). The fertility advantage gained by mating with a larger female provides a plausible basis for this preference.

In contrast to the pattern of female preference, the disparity in size between the object females distinctly influenced male preference. In populations where the range of female sizes is narrow, males will not distinguish among females on the basis of size. In turn, such a pattern would suggest that the importance of male preference will differ among populations that differ characteristically in female size-structure (Travis 1989a).

These results are also consistent with female-female competition based on body size and intrasexual selection cannot be entirely discounted. Foran and Ryan (1994) assessed female-female competition for males of P. latipinna using females of P. latipinna and P. formosa. They found that females of P. latipinna significantly increased levels of blocking (a female swimming between another female and the male), in the presence of a male, but the other three aggressive behaviors, butting, biting and chasing, only increased significantly in the absence of males. The females used in these experiments, however, were matched for size, so competition based upon larger size cannot be addressed by this study.

We found no evidence that males distinguished females from different populations when encountered individually in experiment 4. Mounds males displayed more to females with a red color mark and males from both populations increased rates of gonoporal nibbling or gonopodial thrusting towards females marked black, regardless of population of origin in experiment 5. The markings were placed on the caudal peduncle of the female and might have resembled the black brood spot that occurs dorsal to the anus in females once they reach sexual maturity (Constantz 1989). In gambusiines, the anal spot seems to provide a cue for gonopodial orientation during the male's thrust (Peden 1973). Its importance in Poecilia is not entirely clear. However, other species have shown preferences for experimentally induced colors. For example, female estrildid finches show preferences for males wearing certain colors of leg-bands. Females of each species prefer band colors that are associated with conspecific markings (e.g., Burley 1986). Yet despite some influence of the color markings on certain male behaviors, male mollies clearly did not prefer native females to foreign females regardless of color mark.

Male sailfin mollies can clearly distinguish conspecific from heterospecific females, preferring conspecifics over the sexually parasitic gynogen P. formosa (Woodhead and Armstrong 1985; Schlupp et al. 1991; Ryan et al. 1996; Travis, unpubl. data) and over female P. mexicana (Woodhead and Armstrong 1985; Ryan et al. 1996), the other parental species of the gynogen. There is an obvious fitness disadvantage for P. latipinna males that mate with P. formosa females and probably a fitness disadvantage for any interspecific mating. The lack of intraspecific discrimination implies either that males cannot distinguish among females or that evolution has not favored such discrimination. Females exhibit morphometric differentiation among populations, but it is much smaller than the magnitude of male differentiation, so cues may be insufficient. On the other hand, chemical communication is so vital to males that it seems unreasonable that some olfactory cue does not exist.

We suggest there is no strong reason for males to discriminate among females from different populations. The stable local differentiation in male body-size among molly populations implies that "population outcrossing" could be disadvantageous to males, but the situation is complex because most of the male body size variation is due to a Y-linked gene (Travis 1994a). Local selection on body size provides no rationale for a male to discriminate against an immigrant female: he, and not she, carries the favored allele. If a male moves to a population in which his size is at a disadvantage, his "choices" are limited: he could attempt to return to his natal site or he could attempt to mate with resident females and "hope for the best" for his sons. All things considered, such as the probability of moving to a "wrong" site, the cost of return migration, the lack of any pressure to discriminate against an immigrant female, etc., selection for male discrimination is likely to be extremely weak.

It is easier to envision why females should distinguish native from foreign males. If a female mates with an immigrant male from a population whose body size is not optimal in her population, her sons will carry a gene for body size that is actively selected against at her natal site. A female could recognize such an immigrant male if his size-specific behavior or morphology differed from that of native males. There is clear evidence that males migrate between populations and contribute to gene flow (Trexler et al. 1990), so this is a genuine dilemma for females. If a female immigrates into a population in which males are very different in size from those of her natal site, the body-size problem gives her no reason to discriminate because resident males are more likely than not to carry the locally favored allele. Thus, although not all circumstances suggest a benefit to discrimination, at least one known situation does.

Mating System and Maintenance of Body-Size Variation within Populations

These results, along with others, illuminate the mating system of the sailfin molly and carry several important implications about how the balance between sexual and natural selection acts to maintain male size variation within many populations and why it may not do so in some (Travis 1994a). Females clearly prefer the phenotype of the larger males, regardless of the trait or combination of traits on which they base the choice. One of us (JT) has suggested in earlier work that female choice works indirectly in this system: females incite male competition by advertising their fertility and accept the results of that competition. Larger males respond more strongly to this advertisement than smaller males (Sumner et al. 1994), and larger males can impede the access of smaller males to receptive females (Baird 1968; Parzefall 1973, 1979). The results of this paper are certainly in accord with this suggestion, but they also imply that females could be exerting further influence by actively associating with larger males or by offering greater cooperation to them.

In natural populations, a few females receive most of the attention from males (Travis 1994b, unpubl. data), in accord with a focus by males on receptive females. Several males usually pursue a single female (as in P. mexicana, Balsano et al. 1985), and her options for choice seem limited to differential cooperation with males of different sizes. Active female choice cannot be completely responsible for the behavioral advantage of larger males because it is not universally successful: gene flow from immigrant males often occurs in the face of active female discrimination against them (see above). Nonetheless, regardless of the mechanism and the relative importance of direct and indirect choice, the behavioral elements of sexual selection through both female choice (sensu lato) and male-male interaction clearly favor the larger male phenotypes.

The net mating success of larger males could be increased by their choices among females. Males prefer larger females, and the strength of that preference increases as male size increases. If larger males focused their efforts on larger females, regardless of their receptive state, leaving the smaller females to mate with smaller males, size-assortative mating would result. The much higher fertility of the larger females would give larger males a much higher level of per capita reproductive success than they would accrue through randomly choosing females.

All of these lines of evidence indicate that larger males accrue a fitness advantage over smaller males through sexual selection, but this advantage could be eroded if other aspects of the mating system work in favor of smaller males. Although larger males clearly dominate access to larger, receptive females, smaller males move continuously among females and attempt forcibly to inseminate nonreceptive and receptive females alike (Sumner et al. 1994; Travis 1994b, unpubl. data). If sperm competition was significant, then smaller males as a genetic group might minimize their apparent disadvantages in the sexual selection process. Studies of sperm precedence in poeciliids (reviewed by Travis et al. 1990) suggest that sperm received during the receptive period are more likely to fertilize ova than sperm retained from prior inseminations. Clearly the net advantage to larger males then depends on the outcome of sperm competition, which must integrate the relative frequencies of each male genotype, the volume of retained sperm and newly received sperm, the distribution of newly received sperm among male genotypes (i.e., how effectively a larger male monopolizes receptive females), and any genotype-specific variation in sperm motility or effectiveness. Although we cannot resolve this issue, our behavioral results have limited the remaining questions about the role of sexual selection in maintaining male body-size variation within populations to the domain of sperm competition, because the behavioral evidence indicates a clear advantage to larger males through mate-choice mechanisms.

ACKNOWLEDGMENTS

We thank C. Johnson and J. Leips for field assistance in collecting fish. We also thank N. Martin for assistance in behavioral observations and M. Childress for providing event-recorder software written by J. Baylis, University of Wisconsin. The manuscript benefited from comments by E Breden, A. Houde, M. Kirkpatrick, M. Ryan, A. Thistle, and two anonymous reviewers. This research was sponsored in part by National Science Foundation grant DEB 92-20849 to JT.

LITERATURE CITED

ANDERSSON, M. B. 1982. Female choice selects for extreme tail length in a widowbird. Nature (Lond.) 299:818-820.

-----. 1993. Sexual dimorphism and modes of sexual selection in lekking Jackson's widowbirds Euplectes jacksoni (Ploceinae). Biol. J. Linn. Soc. Lond. 49:1-17.

BAIRD, R. C. 1968. Aggressive behavior and social organization in Mollynesia latipinna (LeSueur). Texas J. Sci. 20:157-176.

BALSANO, J. S., E. J. RANDLE, E. M. RASCH, AND P. J. MONACO. 1985. Reproductive behavior and the maintenance of all-female Poecilia. Environ. Biol. Fish. 12:251-263.

BASOLO, A. L. 1990a. Male swords and female preferences. Science 253:1426-1427.

-----. 1990b. Female preference for male sword length in the green swordtail, Xiphophorus helleri (Pisces: Poeciliidae). Anim. Behav. 40:332-338.

-----. 1995. A further examination of a pre-existing bias favoring a sword in the genus Xiphophorus. Anim. Behav. 50:365-375.

BOX, G. E. P., W. G. HUNTER, AND J. S. HUNTER. 1978. Statistics for experimenters. Wiley, New York.

BREDEN, F., AND G. STONER. 1987. Male predation risk determines female preference in the Trinidad guppy. Nature 329:831-835.

BURLEY, N. 1986. Comparison of the band-colour preferences of two species of estrildid finches. Anita. Behav. 34:1732-1741.

CLARK, E., L. R. ARONSON, AND M. GORDON. 1954. Mating behavior patterns in two sympatric species of xiphophorin fishes: their inheritance and significance in sexual isolation. Bull. Am. Mus. Nat. Hist. 103:135-226.

CLARIDGE, M. F., J. DEN HOLLANDER, AND J. C. MORGAN. 1988. Variation in hostplant relations and courtship signals of weed-associated populations of the brown planthopper, Nilaparvata lugens (Stal), from Australia and Asia: a test of the recognition species concept. Biol. J. Linn. Soc. Lond. 35:790-793.

COCHRAN, W. G., AND G. M. Cox. 1957. Experimental designs. Wiley, New York.

CONSTANTZ, G. D. 1989. Reproductive biology of poeciliid fishes. Pp. 33-50 in G. K. Meffe, and F. F. Snelson Jr., eds. Ecology and evolution of livebearing fishes (Poeciliidae). Prentice-Hall, New York.

CROW, J. F., AND N. R. LILEY. 1979. A sexual pheromone in the guppy, Poecilia reticulata (Peters). Can. J. Zool. 57:184-188.

DOWNHOWER, J. F., AND D. B. LANK. 1994. Effect of previous experience on mate choice by female mottled sculpins. Anim. Behav. 47:369-372.

DRAPER, N. R. AND H. SMITH. 1981. Applied regression analysis. 2d ed. Wiley, New York.

DUGATKIN, L. A. 1996. Interface between culturally based preferences and genetic preferences: female mate choice in Poecilia reticulata. Proc. Nat. Acad. Sci. USA 93:2770-2773.

ENDLER, J. A. 1977. Geographic variation, speciation, and clines. Princeton Univ. Press, Princeton, NJ.

-----. 1983. Natural and sexual selection on color patterns in poeciliid fishes. Environ. Biol. Fish. 9:173-190.

ENDLER, J. A., AND A. E. HOUDE. 1995. Geographic variation in female preferences for male traits in Poecilia reticulata. Evolution 49:456-468.

FARR, J. A., AND J. TRAVIS. 1986. Fertility advertisement by female sailfin mollies, Poecilia latipinna (Pisces: Poeciliidae). Copeia 1986:467-472.

FARR, J. A., J. TRAVIS, AND J. C. TREXLER. 1986. Behavioural allometry and interdemic variation in sexual behaviour of the sailfin molly, Poecilia latipinna (Pisces: Poeciliidae). Anim. Behav. 34:497-509.

FORAN, C. M., AND M. J. RYAN. 1994. Female-female competition in a unisexual/bisexual complex of mollies. Copeia 1994:504-508.

GERHARDT, H. C. 1991. Female mate choice in treefrogs: static and dynamic acoustic criteria. Anim. Behav. 42:615-635.

HASKINS, C. P., AND E. F. HASKINS. 1950. Factors governing sexual selection as an isolating mechanism in the poeciliid fish Lebistes reticulatus. Proc. Nat. Acad. Sci. USA 36:464-476.

HOUDE, A. 1987. Mate choice based upon naturally occurring color-pattern variation in a guppy population. Evolution 41:1-10.

-----. 1988a. Genetic difference in female choice between two guppy populations. Anim. Behav. 36:510-516.

-----. 1988b. The effects of female choice and male-male competition on the mating success of male guppies. Anim. Behav. 36:888-896.

HOUDE, A.E., AND J. A. ENDLER. 1990. Correlated evolution of female mating preferences and male color patterns in the guppy, Poecilia reticulata. Science 248:1405-1408.

HUBBS, C. 1964. Interactions between a bisexual fish species and its gynogenetic sexual parasite. Bull. Texas Mem. Mus. 8:1-72.

KODRIC-BROWN, A. 1985. Female preference and sexual selection for male coloration in the guppy (Poecilia reticulata). Behav. Ecol. Sociobiol. 17:199-206.

-----. 1989. Dietary carotenoids and male mating success in the guppy: an environmental component to female choice. Behav. Ecol. Sociobiol. 25:393-401.

LONG, K. D., AND A. HOUDE. 1989. Orange spots as a visual cue for female mate choice in the guppy (Poecilia reticulata). Ethology 82:316-324.

LUYTEN, P. H., AND N. R. LILEY. 1991. Sexual selection and competitive mating success of male guppies (Poecilia reticulata) from four Trinidad populations. Behav. Ecol. Sociobiol. 28:329-336.

MAYR, E. 1963. Animal species and evolution. Belknap, Press, Cambridge, MA.

MCPEEK, M. A. 1992. Mechanisms of sexual selection operating on body size in the mosquitofish (Gambusia holbrooki). Behav. Ecol. 3:1-12.

MOLLER, A. P. 1988. Female choice selects for male sexual tail ornament in the monogamous swallow. Nature (Lond.) 332:640-642.

-----. 1989. Natural and sexual selection on a plumage signal of status and on morphology in house sparrows, Passer domesticus. J. Evol. Biol. 2:125-140.

MOODIE, G. E. E. 1982. Why asymmetric mating preferences may not show the direction of evolution. Evolution 36:1096-1097.

MORRIS, M. R., W. E. WAGNER JR., AND M. J. RYAN. 1996. A negative correlation between trait and mate preference in Xiphophorus pigmaeus. Anim. Behav. 52:1193-1203.

PARZEFALL, J. 1969. Zur vergeichenden Ethologie verschiedener Mollynesia-Arten einschliesslich einer Hohlenform von M. sphenops. Behaviour 33:1-37.

-----. 1973. Attraction and sexual cycle of poeciliids. Pp. 177-183 in J. H.

Schroder, ed. Genetics and mutagenesis of fish. Springer-Verlag, Berlin, Germany.

-----. 1979. Zur genetik und biologischen Bedeutung des Aggressionsueshaltens von Poecilia sphenops (Pisces, Poeciliidae). Z. Tierpsychol. 50:399-422.

PEDEN, A. E. 1973. Variation in anal spot expression in gambusiin females and its effect on male courtship. Copeia 1973:250-263.

PTACEK, M. B., AND J. TRAVIS. 1996. Interpopulation variation in male mating behaviours in the sailfin molly, Poecilia latipinna. Anim. Behav. 52:59-71.

REYNOLDS, J. D., AND M. R. GROSS. 1992. Female mate preference enhances offspring growth and reproduction in a fish, Poecilia reticulata. Proc. R. Soc. Lond. Ser. B Biol. Sci. 250:57-62.

RIDLEY, M. 1983. The explanation of organic diversity. Oxford Univ. Press, Oxford.

RITCHIE, M. G. 1991. Female preference for "song races" of Ephippiger ephippiger (Orthoptera: Tettigoniidae). Anim. Behav. 42: 518-520.

ROWLAND, W. J. 1989a. Mate choice and the supernormality effect in female sticklebacks (Gasterosteus aculeatus). Behav. Ecol. Sociobiol. 24:433-438.

-----. 1989b. The ethological basis of mate choice in male three-spine sticklebacks, Gasterosteus aculeatus. Anim. Behav. 38: 112-120.

RYAN, M. J., AND A. KEDDY-HECTOR. 1992. Directional patterns of female mate choice and the role of sensory biases. Am. Nat. 139:S4-S35.

RYAN, M. J., AND A. S. RAND. 1993. Species recognition and sexual selection as a unitary problem in animal communication. Evolution 47:647-657.

RYAN, M. J., AND W. E. WAGNER JR. 1987. Asymmetries in mating preferences between species: female swordtails prefer heterospecific mates. Science 236:595-597.

RYAN, M. J., AND W. WILCZYNSKI. 1991. Evolution of intraspecific variation in the advertisement call of a cricket frog (Acris crepitans, Hylidae). Biol. J. Linn. Soc. Lond. 44:249-271.

RYAN, M. J., D. K. HEWS, AND W. E. WAGNER JR. 1990. Sexual selection on alleles that determine body size in the swordtail Xiphophorus nigrensis. Behav. Ecol. Sociobiol. 26:231-237.

RYAN, M. J., C. M. PEASE, AND M. R. MORRIS. 1992a. A genetic polymorphism in the swordtail Xiphophorus nigrensis: testing the prediction of equal fitnesses. Am. Nat. 139:21-31.

RYAN, M. J., S. A. PERRILL, AND W. WILCZYNSKL 1992b. Auditory tuning and call frequency predict population-based mating preferences in the cricket frog, Acris crepitans. Am. Nat. 139:1370-1383.

RYAN, M. J., L. A. DRIES, P. BATRA, AND D. M. HILLIS. 1996. Male mate preferences in a gynogenetic species complex of Amazon mollies. Anim. Behav. 52:1225-1236.

SAPPINGTON, T. W., AND O. R. TAYLOR. 1990. Disruptive sexual selection in Colias eurytheme butterflies. Proc. Nat. Acad. Sci. USA 87:6132-6135.

SCHLUPP, I., J. PARZEFALL, AND M. SCHARTL. 1991. Male mate choice in mixed bisexual/unisexual breeding complexes of Poecilia (Teleostei: Poeciliidae). Ethology 88:215-222.

SCHLUPP, I., C. MARLER, AND M. J. RYAN. 1994. Benefit to male sailfin mollies of mating with heterospecific females. Science 263:373-374.

SHUSTER, S. M., AND M. J. WADE. 1991. Female copying and sexual selection in a marine isopod crustacean, Paracerceis sculpta. Anim. Behav. 41:1071-1078.

SNELSON, F. F., JR., J. D. WETHERINGTON, AND H. L. LARGE. 1986. The relationship between interbrood interval and yolk loading in a generalized poeciliid fish, Poecilia latipinna. Copeia 1986: 295-304.

STONER, G., AND F. BREDEN. 1988. Phenotypic differentiation in female preference related to geographic variation in male predation risk in the Trinidad guppy (Poecilia reticulata). Behav. Ecol. Sociobiol. 22:285-291.

SUMNER, I. T., J. TRAVIS, AND C. D. WILSON. 1994. Methods of female fertility advertisement and variation among males in responsiveness in the sailfin molly (Poecilia latipinna). Copeia 1994:27-34.

THRESHER, R. E., AND A. M. GRONELL. 1978. Subcutaneous tagging of small reef fishes. Copeia 1978:352-353.

TRAVIS, J. 1989a. Ecological genetics of life-history traits in poeciliid fishes. Pp. 185-200 in G. K. Meffe and F. F. Snelson Jr., eds. Ecology and evolution of livebearing fishes (Poeciliidae). Prentice-Hall, New York.

-----. 1989b. The role of optimizing selection in natural populations. Annu. Rev. Ecol. Syst. 20:279-296.

-----. 1994a. The interplay of life-history variation and sexual selection in sailfin mollies. Pp. 205-232 in L. A. Real, ed. Ecological genetics. Univ. of North Carolina Press, Chapel Hill.

-----. 1994b. Size-dependent behavioral variation and its genetic control within and among populations. Pp. 165-187 in C.R.B. Boake, ed. Quantitative genetic approaches to animal behavior. Univ. of Chicago Press, Chicago.

TRAVIS, J., AND B. D. WOODWARD. 1989. Social context and courtship flexibility in male sailfin mollies, Poecilia latipinna (Pisces: Poeciliidae). Anim. Behav. 38:1001-1011.

TRAVIS J., J. C. TREXLER AND M. M. MULVEV. 1990. Multiple paternity and its correlates in female Poecilia latipinna (Pisces, Poeciliidae). Copeia 1990:722-729.

TREXLER, J. C. 1988. Hierarchical organization of genetic variation in the sailfin molly, Poecilia latipinna (Pisces: Poeciliidae). Evolution 42:1006-1017.

TREXLER, J. C., J. TRAVIS, AND M. TREXLER. 1990. Phenotypic plasticity in the sailfin molly, Poecilia latipinna (Pisces: Poeciliidae). II. Laboratory experiment. Evolution 44:157-167.

VERRELL, P. A. 1988. Stabilizing selection, sexual selection, and speciation: a view of specific-mate recognition systems. Syst. Zool. 37:209-215.

-----. 1995. Males chooses larger females as mates in the salamander Desmognathus santeelah. Ethology 99:162-171.

VERRELL, P. A., AND S. J. ARNOLD. 1989. Behavioral observations of sexual isolation among allopatric populations of the mountain dusky salamander, Desmognathus ochrophaeus. Evolution 43: 745-755.

WARNER, R. R., AND E. T. SCHULTZ. 1992. Sexual selection and male characteristics in the bluehead wrasse, Thalassoma bifasciatum: mating site acquisition, mating site defense and female choice. Evolution 46:1421-1442.

WIERNASZ, D. C. 1995. Male choice on the basis of female melanin pattern in Pieris butterflies. Anim. Behav. 49:45-51.

WOODHEAD, A. D., AND N. ARMSTRONG. 1985. Aspects of mating behavior of male mollies (Poecilia spp.). J. Fish Biol. 27:593-601.

ZIMMERER, E. J., AND K. D. KALLMAN. 1989. Genetic basis for alternative reproductive tactics in the pygmy swordtail, Xiphophorus nigrensis. Evolution 43:1298-1307.
COPYRIGHT 1997 Society for the Study of Evolution
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1997 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Ptacek, Margaret B.; Travis, Joseph
Publication:Evolution
Date:Aug 1, 1997
Words:11029
Previous Article:Butterfly eyespots: the genetics and development of the color rings.
Next Article:Genetic and color interactions at a contact zone of Acanthochromis polyacanthus: a marine fish lacking pelagic larvae.
Topics:

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters