HOW TO COMPENSATE FOR COSTLY SEXUALLY SELECTED TAILS: THE ORIGIN OF SEXUALLY DIMORPHIC WINGS IN LONG-TAILED BIRDS

Recent work on birds suggests that certain morphological differences between the sexes may have evolved as an indirect consequence of sexual selection because they offset the cost of bearing extravagant ornaments used for fighting or mate attraction. For example, long-tailed male sunbirds and widowbirds also have longer wings than females, perhaps to compensate for the aerodynamic costs of tail elaboration. We used comparative data from 57 species to investigate whether this link between sexual dimorphism in wing and tail length is widespread among long-tailed birds. We found that within long-tailed families, variation in the extent of tail dimorphism was associated with corresponding variation in wing dimorphism. One nonfunctional explanation of this result is simply that the growth of wings and tails is controlled by a common developmental mechanism, such that long-tailed individuals inevitably grow long wings as well. However, this hypothesis cannot account for a second pattern in our data set: as predicted by aerodynamic theory, we found that, comparing across long-tailed families, sexual dimorphism in wing length varied with tail shape as well as with sex differences in tail length. Thus, wing dimorphism was generally greater in species with aerodynamically costly graduated tails than in birds with cheaper, streamer-shaped tails. This result was not caused by confounding phylogenetic effects, because it persisted when phylogeny was controlled for, using an independent comparisons method. Our findings therefore confirm that certain aspects of sexual dimorphism may sometimes have evolved through selection for traits that reduce the costs of elaborate sexually selected characters. We suggest that future work aimed at understanding sexual selection by investigating patterns of sexual dimorphism should attempt to differentiate between the direct and indirect consequences of sexual selection.     

Long tails affect swimming performance and habitat choice in the male guppy

Sexual selection often favors male secondary sexual traits, although in some cases the elaborate traits incur costs to the males with respect to natural selection. Males of the guppy Poecilia reticulata have longer tails (caudal fins) than females, and the long tails contribute to the mating success of the males through female mate choice. We examined the effect of tail length on the swimming performance of male and female guppies. In a laboratory experiment, males with longer tails exhibited poorer swimming performance than those with shorter tails. However, this effect was not apparent in females. In addition, in a feral population, tail length of males was negatively correlated with water flow velocity in their microhabitats. Although body size of females was negatively correlated with water flow velocity in their microhabitats, tail length of females showed no significant correlation with degree of water flow. These results suggest that the long tail of male guppies incurs costs, such as a decrease in swimming performance, to the males with respect to natural selection and consequently limits their choice of habitats to those with slow water flow.     

Experimental analyses of sexual and natural selection on short tails in a polygynous warbler

We believe that no experimental study has yet tested Darwin's idea that, as well as generating trait elaboration, intersexual selection might sometimes drive sex-biased trait reduction. Here we present the results of two experiments exploring the negative relationship between tail length and reproductive success in male golden-headed cisticolas (Cisticola exilis). In the first experiment, artificially shortening a male's tail produced a dramatic increase in his reproductive success, measured as either the number of females nesting or number of chicks Hedged on his territory. A second experiment, in which manipulated birds were flown through a maze, revealed that short tails also impose costs by reducing aerodynamic performance during slow-speed foraging flight. Because tail shortening yields reproductive benefits and viability costs, we conclude it has evolved via sexual selection. Disentangling exactly how short tails enhance male reproductive success is more difficult. Male-male competition appears partly responsible: aerodynamic theory predicts that tail reduction enhances high-speed flight and, in line with this, shortened-tail males spent more time engaged in high-speed aerial chases of rivals and defended higher-quality territories. However, shortened-tail males had higher reproductive success independent of territory quality and spent more time in aerial displays which may be directed at females. This suggests that tail shortening is also favoured via female choice based on male phenotype.     

Evolution of the structure of tail feathers: implications for the theory of sexual selection

Bird tails are extraordinarily variable in length and functionality. In some species, males have evolved exaggeratedly long tails as a result of sexual selection. Changes in tail length should be associated with changes in feather structure. The study of the evolution of feather structure in bird tails could give insight to understand the causes and means of evolution in relation to processes of sexual selection. In theory, three possible means of tail length evolution in relation to structural components might be expected: (1) a positive relationship between the increase in length and size of structural components maintaining the mechanical properties of the feather; (2) no relationship; that is, enlarging feather length without changes in the structural components; and (3) a negative relationship; that is, enlarging feather length by reducing structural components. These hypotheses were tested using phylogenetic analyses to examine changes in both degree of exaggeration in tail length and structural characteristics of tail feathers (rachis width and density of barbs) in 36 species, including those dimorphic and nondimorphic in tail length. The degree of sexual dimorphism in tail length was negatively correlated with both rachis width and density of barbs in males but not in females. Reinforcing this result, we found that dimorphism in tail length was negatively associated with dimorphism in tail feather structure (rachis width and density of barbs). These results support the third hypothesis, in which the evolution of long feathers occurs at the expense of making them simpler and therefore less costly to produce. However, we do not know the effects of enfeeblement on the costs of bearing. If the total costs increased, the enfeeblement of feathers could be explained as a reinforcement of the honesty of the signal. Alternatively, if total costs were reduced, the strategy could be explained by cheating processes. The study of female preferences for fragile tail feathers is essential to test these two hypotheses. Preferences for fragile tails would support the evolution of reinforcement of honesty, whereas female indifference would indicate the existence of cheating in certain stages of the evolutionary process.     

Female choice, male interference, and sperm precedence in the red-spotted newt

Darwin first identified female choice and male—male competitionas forms of sexual selection resulting in the evolution of conspicuoussexual dimorphism, but it has proven challenging to separatetheir effects. Their effects on sexual selection become evenmore complicated when sperm competition occurs because spermprecedence may be either a form of cryptic female choice ora form of male—male competition. We examined the effectsof tail height on male—male competition and female choiceusing the sexually dimorphic red-spotted newt (Notophthalmusviridescens viridescens). Experiment 1 examined whether maletail height influenced male mating success. Males with deeptails were more successful at mating with females than thosewith shallow tails. Successful, deep-tailed males also were bigger(snout-vent length; SVL) than unsuccessful, shallow-tailed males,but they did not vary in tail length or body condition. Of these,only tail height and tail length are sexually dimorphic traits.Experiment 2 tested the hypothesis that the differential successof males with deeper tails was due to female choice by examiningboth simultaneous female preference for association and sequentialfemale choice. We found no evidence of female choice. When maleswere not competing to mate with females, tail height did notinfluence male mating success. Successful males did not havedifferent SVL and tail lengths than unsuccessful males. Thus,tail height in male red-spotted newts appears to be an intrasexuallyselected secondary sexual characteristic. Experiment 3 usedpaternity exclusion analyses based on molecular genetic markersto examine the effect of sperm precedence on sperm competitionin doubly-mated females. Sperm precedence likely does not havea pervasive and consistent effect on fertilization success becausewe found evidence of first, last, and mixed sperm usage.     

Sexual selection in the barn swallow Hirundo rustica. VI. Aerodynamic adaptations

Secondary sexual characters are assumed to be costly to produce and maintain, and this will select for morphological modifications that reduce the magnitude of such costs. Here we test whether a feather ornament, the sexually exaggerated outermost tail feathers of male barn swallows Hirundo rustica, a trait currently subject to a directional female mate preference, and other aspects of the morphology used for flight have been modified to increase aerodynamic performance. This was done by making comparisons among sexes within populations, among individuals varying in tail length within populations, and among populations from different parts of Europe. Male barn swallows experienced reduced drag from their elongated tail feathers by morphological modifications of the ornamental feathers as compared to females. Morphological features of the outermost tail feathers were unrelated to tail length in both males and females within populations. Wing and tail morphology (length of central tail feathers and wings, wing span, wing area, wing loading, and aspect ratio) was modified in males compared to females. Barn swallows with long tails had morphological tail and wing modifications that reduced the cost of a large ornament, and similar modifications were seen among populations. The costs of the exaggerated secondary sexual character were therefore reduced by the presence of cost-reducing morphological modifications. The assumptions of reliable signalling theory, that signals should be costly, but more so to low than to high quality individuals, were not violated because long-tailed male barn swallows had the largest cost-reducing morphological characters.     

Tail Length and Mutual Mate Choice in Bearded Tits (Panurus biarmicus)

Direct sexual selection via mutual mate choice can result in both sexes showing conspicuous traits. We experimentally tested whether this hypothesis can explain tail length in the bearded tit (Panurus biarmicus). In this species, both sexes have a long, graduated tail. Males have, however, a longer tail than females, suggesting perhaps that females are choosier than males in selecting mates. We used two choice set‐ups for each sex: shortened vs. control tail individuals and elongated vs. control tail individuals. We found that direct sexual selection seems to operate differently in the two sexes. In both set‐ups, females spent more time with the male with the longest tail, and they also showed sexual display behaviour only towards these males. Males spent more time with control than with short‐tailed females, but they did not discriminate between control and long‐tailed females. Moreover, males displayed preference towards both short‐ and long‐tailed females. Thus, females preferred long‐tailed males, whereas males did not always prefer long‐tailed females. Our study suggests that mutual mate choice has played a role in the evolution of long tails in bearded tits. It also suggests that the sexual dimorphism in tail length has evolved because mate choice exerts a stronger sexual selection pressure on males than on females.     

Haematocrit correlates with tail ornament size in three populations of the Barn Swallow (Hirundo rustica)

1. Handicap models of sexual selection propose that male ornaments are indicators of male quality and that honesty is enforced by the costs imposed by the exaggerated ornamental traits. In long-distance migratory birds that feed on the wing, the aerodynamic cost of exaggerated ornamental characters should be particularly high because the size of the ornaments deviates from the natural selection optimum. During migration, birds are expected to raise their oxygen consumption in relation to the energetic demands imposed by their morphology. An increase of haematocrit is an adaptive response to enhance oxygen uptake and efficiency of transfer to the muscular tissues during spells of intense muscular activity.
2. The change of haematocrit of Barn Swallows ( Hirundo rustica ) after their arrival to the breeding sites, and the relationships between haematocrit values recorded after migration and the size of ordinary and sexually selected morphological characters in three Barn Swallow populations were analysed.
3. Males had higher haematocrit values than females. Individual haematocrit values declined after arrival to the breeding sites. Haematocrit values of males were significantly and positively correlated with the size of their ornamental tail but not correlated with other characters, thus suggesting that well-ornamented males, in order to arrive early, have to raise their haematocrit above the level of short-tailed males.
4. Males and females of similar tail length did not differ in their haematocrit, thus suggesting that sexual dimorphism in haematocrit might be functionally related to dimorphism in tail length.
5. Our results are consistent with the handicap principle because long-tailed males experience lower mortality and larger seasonal reproductive success compared with short-tailed males.     

Sexual dimorphism in snake tail length: sexual selection,natural selection,or morphological constraint?

Male snakes typically have longer tails relative to body length than females, but the extent of this dimorphism varies among species. Three hypotheses have been suggested to explain tail dimorphism. The Morphological Constraint Hypothesis proposes that males have relatively longer tails to accommodate hemipenes and retractor muscles. The Female Reproductive Output Hypothesis proposes that females have relatively shorter tails as a secondary result of natural selection for increased reproductive capacity. The Male Mating Ability Hypothesis proposes that sexual selection favours relatively longer tails in males during courtship. These hypotheses make different predictions about the relationships among tail length, body size, male reproductive morphology, female reproductive output, mode of reproduction, and male mating behaviour among and within taxa. Predictions were tested using published data for 56 genera in the family Colubridae and original data for the water snake, Nerodia sipedon. Tail length dimorphism was more male-biased in tam having relatively short tails (r=–0.52, P < 0.001), hemipenes and retractor muscles occupied a greater proportion of the tail in taxa having relatively short tails (r=– 0.71, P < 0.00l and r=– 0.66, P = 0.001, respectively), and tail length dimorphism was more male-biased in taxa in which body size dimorphism was more female-biased (r=– 0.60, P < 0.001). These results support both the Morphological Constraint Hypotheses and the Female Reproductive Output Hypothesis. However, tests of other predictions, including those regarding patterns within N. sipedon , failed to support any of the three hypotheses. Comparisons among taxa suggest several species in which further tests of these hypotheses would be especially appropriate.     

Why do male snakes have longer tails than females?

In most snake species, males have longer tails than females of the same body length. The adaptive significance of this widespread dimorphism has attracted much speculation, but few tests. We took advantage of huge mating aggregations of red-sided gartersnakes (Thamnophis sirtalis parietalis) in southern Manitoba to test two (non-exclusive) hypotheses about the selective forces responsible for this dimorphism. Our data support both hypotheses. First, relative tail length affects the size of the male copulatory organs (hemipenes). Males with longer tails relative to body length have longer hemipenes, presumably because of the additional space available (the hemipenes are housed inside the tail base). Second, relative tail length affects male mating success. Males with partial tail loss (due to predation or misadventure) experienced a threefold reduction in mating success. Among males with intact tails, we detected strong stabilizing selection on relative tail length in one of the two years of our study. Thus, our data support the notion that sex divergence in tail length relative to body length in snakes reflects the action of sexual selection for male mating success.     

Sexual selection in the barn swallow (Hirundo rustica). V. Geographic variation in ornament size

Models of sexual selection in a cline predict the patterns of clinal variation in female mate preference and male secondary sexual characters. These predictions were tested for the nominate subspecies of the barn swallow Hirundo rustica which demonstrates clinal variation in morphology, with several characters in both sexes showing increasing size at higher latitudes. Sexual size dimorphism in the length of the tail ornament and the short, central tail feathers increase with increasing latitude while size dimorphism in other morphological characters is independent of latitude. The main reason for the two divergent patterns of sexual size dimorphism appears to be the higher foraging cost of having a long tail ornamental at low latitudes. The control of development decreases with increasing latitude as demonstrated by an increasing latitudinal cline in fluctuating asymmetry of tail length. Phenotypic variance in tail length increases with latitude in males, but not in females, as shown by the coefficients of variation. Clinal variation in morphology is not due to natural selection associated with a latitudinal increase in the distance between breeding and wintering areas. The geographic patterns of morphological variation suggest that the tail character has diverged geographically as a result of a sexual process of reliable signalling.     

Is natural selection promoting sexual dimorphism in the Green-backed Firecrown Hummingbird ( Sephanoides Sephaniodes )?

In many hummingbird species there is an opposite pattern of sexual dimorphism in bill length and other morphometric measures of body size. These differences seem to be closely related with differences in foraging ecology directly associated with a different resource exploitation strategy. The aim of this study was to assess if natural selection is acting on wing length and bill size in hummingbird males and females with different resource exploitation strategies (i.e., territorial males and non-territorial females). If competition for resources promotes sexual dimorphism as a selective pressure, males should be subjected to negative directional selection pressure for wing length and no selection pressure over bill size, while females should undergo positive directional selection pressure for both bill size and wing length. The morphometric data we collected suggests that there is no selection for wing length and bill size in male hummingbirds. In contrast, our females exhibited positive directional selection for both wing length and bill size. Although we cannot reject sexual selection acting on sexually dimorphic traits, this study suggests that natural selection may promote sexual dimorphism in traits that are closely related with hummingbird foraging ecology and resource exploitation strategies.     

Sexual dimorphism of tail length in lacertid lizards: test of a morphological constraint hypothesis

Males of many lizard species have longer tails than similarly-sized females. We hypothesized that this dimorphism is induced by a longer non-autotomous tail part in males, which is associated with the presence of the copulatory organs at the tail base, and presumably reduces the males' ability to escape predation by tail shedding. A compensatory mechanism would be an increase of total tail length in males, to achieve equal lengths of the autotomous tail part in both sexes. A critical prediction of this 'morphological constraint' hypothesis is that the extent of dimorphism in total tail length increases with the magnitude of sexual differences in length of the non-autotomous tail base. We tested this prediction through a comparative study in a small clade of lacertid lizards. Within each of nine species, sexual differences in length of the non-autotomous tail base and in total tail length do not change with body size. All species, except one, exhibit a clear male-biased dimorphism in length of the non-breakable tail base. In all species studied, males have longer tails than females. We used the method of phylogenetically independent contrasts to explore the interspecific relation between dimorphism in length of the tail base and sexual differences in total tail length. Contrary to our prediction, we found no evidence for a positive correlation between the extent of dimorphism in both traits. Thus, constraints imposed by the male copulatory organs on tail autotomy do not seem to be a significant factor in the evolution of dimorphism in tail length in this clade of lacertid lizards.     

No evidence of sexual dimorphism in the tails of the swallowtail butterflies Papilio machaon gorganus and P. m. britannicus

The European swallowtail butterfly (Papilio machaon) is so named, because of the long and narrow prominences extending from the trailing edge of their hindwings and, although not a true tail, they are referred to as such. Despite being a defining feature, an unequivocal function for the tails is yet to be determined, with predator avoidance (diverting an attack from the rest of the body), and enhancement of aerodynamic performance suggested. The swallowtail, however, is sexually size dimorphic with females larger than males, but whether the tail is also sexually dimorphic is unknown. Here, museum specimens were used to determine whether sexual selection has played a role in the evolution of the swallowtail butterfly tails in a similar way to that seen in the tail streamers of the barn swallow (Hirundo rustica), where the males have longer streamers than those of the females. Previously identified sexual dimorphism in swallowtail butterfly size was replicated, but no evidence for dimorphism in tail length was found. If evolved to mimic antennae and a head to divert a predatory attack, and if an absolute tail size was the most effective for this, then the tail would probably be invariant with butterfly hindwing size. The slope of the relationship between tail length and size, however, although close to zero, was nonetheless statistically significantly above (tail length ∝ hindwing area ^{0.107 ± 0.011}). The slope also did not equate to that expected for geometric similarity (tail length ∝ hindwing area^1/2) suggesting that tail morphology is not solely driven by aerodynamics. It seems likely then, that tail morphology is primarily determined by, and perhaps a compromise of several, factors associated with predator avoidance (e.g. false head mimicry and a startling function). Of course, experimental data are required to confirm this.     

Sexual selection, antennae length and the mating advantage of large males in Asellus aquaticus

In crustacean species with precopulatory mate-guarding, sexual size dimorphism has most often been regarded as the consequence of a large male advantage in contest competition for access to females. However, large body size in males may also be favoured indirectly through scramble competition. This might partly be the case if the actual target of selection is a morphological character, closely correlated with body size, involved in the detection of receptive females. We studied sexual selection on body size and antennae length in natural populations of Asellus aquaticus, an isopod species with precopulatory mate guarding. In this species, males are larger than females and male pairing success is positively related to body size. However, males also have longer antennae, relative to body size, than females, suggesting that this character may also be favoured by sexual selection. We used multivariate analysis of selection to assess the relative influences of body size and antennae length in five different populations in the field. Selection gradients indicated that, overall, body size was a better predictor of male pairing success than antennae length, although some variation was observed between sites. We then manipulated male antennae length in a series of experiments conducted in the lab, and compared the pairing ability of males with short or long antennae. Males with short antennae were less likely to detect, orient to, and to pair with a receptive female compared with males with long antennae. We discuss the implications of our results for studies of male body size and sexual dimorphism in relation to sexual selection in crustaceans.     

Sexual selection, antennae length and the mating advantage of large males in Asellus aquaticus

In crustacean species with precopulatory mate-guarding, sexual size dimorphism has most often been regarded as the consequence of a large male advantage in contest competition for access to females. However, large body size in males may also be favoured indirectly through scramble competition. This might partly be the case if the actual target of selection is a morphological character, closely correlated with body size, involved in the detection of receptive females. We studied sexual selection on body size and antennae length in natural populations of Asellus aquaticus, an isopod species with precopulatory mate guarding. In this species, males are larger than females and male pairing success is positively related to body size. However, males also have longer antennae, relative to body size, than females, suggesting that this character may also be favoured by sexual selection. We used multivariate analysis of selection to assess the relative influences of body size and antennae length in five different populations in the field. Selection gradients indicated that overall body size was a better predictor of male pairing success than antennae length, although some variation was observed between sites. We then manipulated male antennae length in a series of experiments conducted in the laboratory, and compared the pairing ability of males with short or long antennae. Males with short antennae were less likely to detect, orient to and to pair with a receptive female compared to males with long antennae. We discuss the implications of our results for studies of male body size and sexual dimorphism in relation to sexual selection in crustaceans.     

Male reproductive success and sexual selection in northern water snakes determined by microsatellite DNA analysis

Male northern water snakes (Nerodia sipedon) have high variancein reproductive success relative to females. We used DNA-basedpaternity analyses from a 3-year study of two marsh populationsof water snakes to investigate the factors that contributeto variation in male success. Male traits investigated includedbody size, condition, tail length, home range size, activityduring the mating season, and genetic profile (genetic similarityto females, heterozygosity, and genetic variability [d²]).We successfully assigned > 80% of offspring to sires froma sample of 811 offspring from 45 litters. Male reproductivesuccess did not vary significantly with body size, tail length,condition, home range size, or the number of microsatelliteloci at which males were heterozygous, nor with other featuresof their genetic profiles. However, we found evidence of positive assortative mating by size in the marsh in which receptive femaleswere not spatially clumped. Also, males that were most activeduring the mating season were more successful, particularlywhere females were not clumped. We failed to find evidenceof selection acting on male size through variance in reproductivesuccess, indicating that sexual selection does not have an important influence on sexual size dimorphism in this species(males are smaller than females). We propose that males aresmaller than females because the lack of advantage to largesize allows males to adopt a low-energy, low-growth strategythat reduces their risk of predation outside the mating season.     

Sexual selection and tail streamers in the barn swallow

The functional significance of elongated, narrow tips of the tail feathers of certain birds, so-called tail streamers, has recently been discussed from an aerodynamic point of view, and the effects of sexual selection on such traits have been questioned. We review our long-term field studies using observational and experimental approaches to investigate natural and sexual selection in the barn swallow, Hirundo rustica, which has sexually size-dimorphic outermost tail feathers. Experimental manipulation of the length of the outermost tail feathers has demonstrated sexual selection advantages of tail elongation and disadvantages of tail shortening, with opposite effects for natural selection in terms of foraging efficiency, haematocrit and survival. These findings are contrary to the prediction of a general deterioration from both shortening and elongation, if the tail trait was determined solely by its effects on aerodynamic efficiency and flight manoeuvrability. Patterns of sexual selection in manipulated birds conform with patterns in unmanipulated birds, and selection differentials for different components of sexual selection in manipulated birds are strongly positively correlated with differentials in unmanipulated birds. Age and sex differences in tail length, and geographical patterns of sexual size dimorphism, are also consistent with sexual selection theory, but inconsistent with a purely natural selection advantage of long outermost tail feathers in male barn swallows.     

The relative importance of male tail length and nuptial plumage on social dominance and mate choice in the red-backed fairy-wren Malurus melanocephalus: evidence for the multiple receiver hypothesis

Understanding why males of many species exhibit two or more sexual ornaments depends upon identifying both the information conveyed and the intended receiver(s) for each signal. Here we focus on identifying the intended receivers for two sexual signals exhibited by male red-backed fairy-wrens Malurus melanocephalus , extent of nuptial plumage and tail length. In doing so we test the multiple receiver hypothesis, which predicts that each trait is directed toward a different type of receiver (e.g., males vs females). Male red-backed fairy-wrens in nuptial plumage exhibit reversed sexual dimorphism for tail length in the breeding season, when their tails are significantly shorter than those of females or males in eclipse plumage. Using both aviary-based experiments and indices of mate choice and social dominance from a natural population, we found that extent of nuptial plumage and age primarily affected female mate choice and that shorter tails were primarily associated with male:male dominance signaling. The field and aviary studies combined are consistent with the multiple receiver hypothesis, in that each trait appears to be directed primarily to a different set of receivers (plumage for females and tail length for males), though each trait may also signal information to the other set of receivers as well. We propose that sexual selection may favor shorter tail lengths in male red-backed fairy-wrens through social competition mechanisms.     

The influence of natural selection and sexual selection on the tails of sea-snakes (Laticauda colubrina)

Many phenotypic traits perform more than one function, and so can influence organismal fitness in more than one way. Sexually dimorphic traits offer an exceptional opportunity to clarify such complexity, especially if the trait involved is subject to natural as well as sexual selection, and if the sexes differ in ecology as well as reproductive behaviour. Relative tail length in sea-snakes fulfils these conditions. Our field studies on a Fijian population of yellow-lipped sea kraits ( Laticauda colubrina ) show that relative tail lengths in male sea kraits have strong consequences for individual fitness, both via natural and sexual selection. Males have much longer tails (relative to snout-vent length) than do females. Mark-recapture studies revealed a trade-off between growth and survival: males with relatively longer tails grew more slowly, but were more likely to survive, than were shorter-tailed males. A male snake's tail length relative to body length influenced not only his growth rate and probability of survival, but also his locomotor ability and mating success. Relative tail length in male sea kraits was thus under a complex combination of selective forces. These forces included directional natural selection (through effects on survival, growth and swimming speed) as well as stabilizing natural selection (males with average-length tails swam faster) and stabilizing sexual selection (males with average-length tails obtained more matings). In contrast, our study did not detect significant selection on relative tail length in females. This sex difference may reflect the fact that females use their tails primarily for swimming, whereas males also must frequently use the tail in terrestrial locomotion and in courtship as well as for swimming.