Behavioral Ecology Advance Access originally published online on October 10, 2007
Behavioral Ecology 2008 19(1):1-8; doi:10.1093/beheco/arm094
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Female mating biases for bright ultraviolet iridescence in the butterfly Eurema hecabe (Pieridae)
School of Marine and Tropical Biology, James Cook University, PO Box 6811, Cairns, Queensland 4870, Australia
Address correspondence to D.J. Kemp. E-mail: darrell.kemp{at}jcu.edu.au.
Received 6 July 2007; revised 12 September 2007; accepted 12 September 2007.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Exaggerated male-limited coloration is widespread among butterflies, yet convincing demonstrations of intraspecific mating preferences for signal brightness and/or chromaticity are relatively rare in this group. Here, I couple behavioral experiments involving manipulations of ambient light environments and male reflectance patterns with observation of wild mating patterns to investigate visual mating biases in the large grass yellow (Eurema hecabe). Males in this species possess exaggerated, limited-view ultraviolet (UV) iridescence across most of their dorsal wing surface that has putative sexual signaling function. In the first experiment, conducted in small (0.7–m3) cages, individuals were significantly less likely to copulate when the UV portion of natural ambient illumination (i.e., 300–400 nm) was strongly reduced. In 2 subsequent experiments, conducted under full-spectrum sunlight in small and large (5 x 6 x 4 m) cages, males with their UV signal artificially dulled by 25% consistently copulated with fewer, and smaller, females than sham-control individuals. Importantly, the manipulated levels of UV brightness in these experiments fall well within the naturally occurring bounds of variation in male UV reflectance. These findings therefore unanimously support the presence of a UV signal–based female bias. In apparent contrast, comparison of 161 in-copula and 188 free-flying males from a high-density field assemblage revealed that copulating males were significantly older and henceforth actually possessed (subtly) less UV bright wings. Copulating male UV brightness was, however, positively related to the size of their mate, which echoes the experimental findings and may represent a signature of mutual mate choice. I discuss these results in light of the full complexities of the butterfly mating system and the potential signaling value of iridescent coloration in butterflies and animals generally.
Key words: coloration, female mate choice, Lepidoptera, ornamentation, sexual selection, signaling, structural color.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
The evolution of extravagant sexual ornamentation has intrigued biologists since at least Darwin (1874)
and continues to pose considerable empirical and theoretical challenges (Kokko et al. 2006). Our understanding of why and how colorful sexual ornaments evolve has been greatly advanced by studies in model vertebrate systems, such as guppies (e.g., Endler 1980, 1983), blue tits (e.g., Andersson et al. 1998; Sheldon et al. 1999), and house finches (e.g., Hill 1990, 1991). Working off a firmly established empirical basis for the existence of mating biases, researchers have used these (and other) systems to explore theories for the indirect and direct evolution of mating preferences. Guppies, by virtue of their relatively short generation times and suitability for laboratory breeding, have proved particularly useful for assessing the quantitative genetics of mate choice and colorful sexual ornamentation and for evaluating genetic mate choice models (e.g., Houde 1994; Brooks and Endler 2001). Invertebrate groups such as butterflies offer similar empirical opportunities, but these have tended to figure less prominently in studies of exaggerated color ornaments and the evolution of color-based mating preferences.
Color-based mate choice in butterflies has been most frequently investigated from the perspectives of reproductive isolation and the mechanisms underlying speciation (e.g., Jiggins et al. 2001; Mavarez et al. 2006). Fewer studies have sought to evaluate whether preferences for intraspecific levels of variation exist and/or whether such preferences could drive the exaggeration of male-limited bright color patterns (although see Robertson and Monteiro 2005; Kemp 2007). As a result, presently topical issues in the study of sexual selection, such as phenotypic and genetic condition dependence, the costs of signal expression and mate choice, and direct and indirect mating benefits, have rarely been addressed using model butterfly systems. This contrasts starkly with the ever-expanding and strongly theoretically grounded body of research into avian plumage coloration (e.g., Johnsen et al. 2003; Parker and Garant 2004; Qvarnstrom et al. 2006). Despite the popularity and suitability of birds for behavioral study, however, this group is less amenable (than many invertebrates) to certain types of empirical investigation. Large, laboratory-based manipulative and quantitative genetic experiments, for example, are more readily achieved with butterflies than birds (although genetic studies are certainly not impossible in the latter system; see Parker and Garant 2004; Qvarnstrom et al. 2006). This places colorful butterflies as worthy candidate systems for—among other things—evaluating predictions of indirect benefits of mate choice models. However, their utmost utility is contingent on whether or not color ornament–based mating preferences actually exist and whether such preferences are likely to have contributed to color pattern exaggeration in this group.
The most convincing demonstrations of the evolutionary significance of mating preferences are those in which the results of manipulative experiments are matched by observations in the wild or on wild-caught populations. In house finches, for example, observations of female preferences for bright red plumage among wild-caught finches (Hill 1990) were matched by the outcome of field-based manipulative experiments (Hill 1991). Similarly, in guppies, Endler (1983) famously replicated the outcome of field introduction experiments (in which guppies placed in a predator-free environment evolved more conspicuous ornamentation) in separate experiments carried out in a semicontrolled artificial stream environment. Guppies have since been shown to express similar mating preferences under controlled laboratory conditions (Endler and Houde 1995; Brooks and Endler 2001). The analytical coupling of laboratory manipulation with field observation is powerful because although the former approach unequivocally indicates causality, the latter approach suggests likely relevance of the causal element under natural conditions.
In this study, I employ a diversity of manipulative and observational approaches to investigating female mating preferences in a polyandrous pierid butterfly, the large grass yellow (Eurema hecabe). Similar to other coliadines, the males of this sexually dimorphic species possess a bright, chromatic, and iridescent ultraviolet (UV) signal across most of their dorsal wing surface. This iridescent UV overlays a diffuse pigment–based yellow and is framed at the wing margin by a melanic black band (Figure 1). Females exhibit a similar dorsal wing pattern except that their UV markings are less bright and restricted to a small proximal region of the forewing (Kemp DJ, Rutowski RL, Macedonia JM, unpublished data). The ventral surfaces of both sexes are strongly UV absorbent and appear bright yellow (males) or whitish yellow (females). In Colias eurytheme, a visually similar relative, females are thought to prefer mates that possess bright UV iridescence (Silberglied and Taylor 1973, 1978; Papke et al. 2007), which is visually amplified by the underlying pigmentary yellowish orange (Rutowski et al. 2005). Comparative and phylogenetic analyses are also consistent with a sexually selected origin for iridescent wing coloration within the Coliadinae (Brunton and Majerus 1995; Brunton 1998; Kemp et al. 2005). However, no study has yet used direct manipulations to assess the importance of iridescent signal brightness per se in this group.
|
I evaluated whether female E. hecabe prefer bright UV iridescence using experimental manipulations of the ambient light environment (experiment 1, below), manipulations of male wing coloration in small (experiment 2) and large (experiment 3) enclosures, and field collection of copulating individuals. This analytically multifarious approach facilitated both experimental and field-based assessments of the role of male UV coloration in mate choice. Because body size may offer a phenotypic marker of mate quality in both male and female butterflies and because both sexes may be choosy (Rutowski 1985
), I incorporated this parameter along with UV reflectance in the assessment of mating patterns and potential mechanisms of male and female mate choice (investigation of other potential mate choice cues, such as pheromones [Costanzo and Monteiro 2007; Papke et al. 2007] is beyond the scope of this study). |
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Experiment 1: ambient light manipulation
This experiment, designed to assess the effect of ambient UV illumination on mating activity, was conducted from 26 February to 1 March 2006 using 2 aluminum-framed 0.7-m square cages located in an outdoor compound at James Cook University, Cairns (Australia; 16°53'S, 145°45'E). One cage, the designated UV cage, was covered on the top and 3 sides with a single layer of Rosco (No. 3114) plastic film, which strongly absorbs light below 400 nm (Figure 2). The second cage was similarly covered with a layer of clear plastic ("Seran wrap"), which proved equivalently transparent to the Rosco film at human-visible wavelengths (i.e., 400–700 nm) but without absorbing appreciably in the UV range (Figure 2). These plastic films were chosen so that the 2 cages varied dramatically in only their ambient UV illumination. Each cage contained potted nectar plants (Stachytarpheta spp. [Verbanaceae]) and larval foodplants (Aeschynomene indica [Fabaceae]).
|
Initially, 30 virgin females and 40 virgin males were released in each cage. Observations were subsequently made at 10-min intervals with mating pairs removed and replaced with a single virgin male and female. In this experiment, as in the experiments below, subjects were the 2- to 4-day-old laboratory-reared offspring of N = 20 mothers captured around Cairns in late January 2006. Individuals were reared in groups of approximately 100 within a series of 0.7-m square mesh cages located within a semithermally controlled room (temperature range 23–30 °C, ambient photoperiod regime). Potted A. indica was used as the larval host. On eclosion, individuals were all initially assessed for adult body size by measuring the length of their right forewing, from apex to thoracic insertion, to the nearest 0.1 mm using digital calipers. There is a strong positive relationship between forewing length and body mass in this species (i.e., r = 0.87; see Table 2 of Jones [1992]
). No individual was used more than once, either within or across experiments.
|
Experiment 2: wing UV manipulation (small enclosures)
This experiment, designed to directly assess the relevance of male UV to mating success, was conducted in 3 blocks: 2–4 March, 9–10 March, and 17–18 March 2006, using a single 0.7-m square cage (as used above, but with "fly wire" covering). On each occasion, 30 virgin females were liberated into the cage along with 20 virgin males in each of 2 treatment groups. Males in the "UV" treatment group had their dorsal UV dulled by approximately 25% (Figure 3a) using a solution of rutin dissolved in ethanol. An important point to note regarding the effect of this manipulation is that it falls within the range of naturally occurring UV brightness variation in this species (compare panels a and b of Figure 3). The rutin solution was obtained by agitating 1 g of rutin (Life Extension Corp, Fort Lauderdale, FL) in 10 ml of ethanol at 60 °C for 20 min (see Kemp 2007
). Males were treated on the day before each experimental block, inside a 4 °C walk-in cold room in which their wings could be dorsally spread and the solution lightly applied with a fine artists' paintbrush. Once dry (within 1 min), individuals were removed to room temperature, where they recovered with no apparent ill effects. Males in the second experimental group, the "control" males, were handled identically except that the rutin solution was applied to an equivalent area of their ventral wing surfaces. For recognition purposes, treatment males were marked with a small black dot (Sharpie ultrafine marker) near the apex of their right forewing, whereas control males were marked on their left forewing. The cage was observed at 10-min intervals from 10:00 h until 13:00 h, after which all males were removed and stored in temporary holding cages or killed (at the end of each experimental block). All mating pairs were immediately removed and replaced with a virgin female and a male from the corresponding experimental group.
|
Experiment 3: wing UV manipulation (large enclosure)
This experiment, conducted from 13 to 16 March 2006, was identical to experiment 2 except for being performed in a large (6 x 5 x 4 m), semicylindrical outdoor experimental enclosure. The enclosure was overlain with 33% UV-absorbing shade cloth and outfitted with tropical palms, potted larval foodplants, and adult nectar plants. Males were manipulated as in experiment 2. At the beginning of this experiment (ca. 10:00 h, March 13), a total of 50 virgin females were released into the cage, along with 50 males in each of the 2 experimental groups (i.e., UV and control males). The cage was monitored at 10-min intervals, and mating pairs were replaced with virgin individuals. Observations were suspended each day at 17:00 h (when the cage was shaded and butterfly activity ceased) and resumed at 08:00 h the following day.
Field mating patterns
This component of the study was carried out between 2 March and 28 April 2006 (between the hours of 10:00 and 15:00) at several field breeding sites in the vicinity of Cairns. The sites consisted of poorly drained vacant lots or roadside ditches containing abundant growth of the locally favored larval hostplant, A. indica. Eurema hecabe breed continuously and are extremely abundant in all developmental stages at these localities during the mid-late wet season (i.e., February–May). I walked around these sites and encouraged in-copula pairs to reveal themselves by gently tapping the vegetation with a butterfly net. Pairs were then netted, along with the nearest free-flying male that was seen to be actively engaging in mate-locating behavior. Equal numbers of mating pairs and free-flying males were collected on each sampling day, with all butterflies stored initially within 30-ml plastic screw-top specimen jars within an ice-filled cooler. Individuals were immediately frozen on returning to the laboratory; intact in-copula pairs were frozen only once separated. As in the experiments, I measured forewing length to the nearest 0.1 mm as a measure of body size, and I assessed male age using well-established protocols based on 5 categories of damage to wing margins (i.e., wing wear; refer to Kemp 2006a, 2006b and references therein). Contrasts involving wing wear are based on the assumption that the processes involved with copulation do not contribute to additional wing wear, which seems reasonable given that copulating pairs seek cryptic locations among vegetation.
The assessment of field mating patterns may be complicated by the occurrence of pupal matings; that is, situations where emerging or freshly emerged adult females are located and mated before their adult body structures (e.g., wings) are fully hardened. Precopulatory female mate choice is circumvented in these instances, which increase along with population density (Silberglied and Taylor 1973, 1978). Pupal matings do occur in high-density populations of E. hecabe (Kemp DJ, personal observation) but are difficult to reliably identify during field collection. Pairs that were unequivocally the result of a pupal mating, that is, those involving soft-winged females, were excluded from further analysis. For all other pairs, I dissected females and counted the number of spermatophores contained within their bursae copulatrix. Various workers have established that male butterflies always donate a single spermatophore during an undisturbed copulation (refer to Drummond 1984); hence, a count of spermatophores can indicate a female's mating history. I used this information to analyze restricted subsets of data that were known not to result from pupal matings.
Assessment of wing coloration
As a limited-view iridescent signal, the UV component of male wing coloration varies both in overall brightness (i.e., reflectance intensity) and in the above-wing angular range over which it is visible. Depending on how these attributes are integrated by the female visual system, they could both contribute to UV signal perception. I therefore used 2 separate but complimentary approaches to quantifying male wing coloration (using each specimen's left and right forewings). First, I measured wing reflectance using an Ocean Optics USB-4000 spectrometer coupled with a PX-2 pulsed xenon light source. The spectrometric measurement methods are both well established and detailed/illustrated elsewhere (Kemp 2006a). Briefly, the set-up consisted of a 5-mm collimated light beam provided at 90° to the horizontal, with the probe situated at 45° and focused to capture input from a 2-mm circular area. Forewings were measured on a dual axis stage with their proximal–distal axis parallel with the plane of the probe and their base closest to the probe (see Figure 1 in Kemp 2006a). Reflectance was measured relative to a magnesium oxide standard, which reflects brightly and evenly across the 300- to 700-nm range.
Second, I used established whole-wing video viewing techniques to quantify the angular visibility of each specimen's iridescent UV. For these assessments, samples were illuminated with a tungsten–halogen fiber-optic source and viewed with a video camera fitted with both a Tiffen 18A and low-pass filter, which collectively transmit light in only a narrow UV band (ca. 350–400 nm). The physical arrangement of forewing, light source, and collector probe were as used during spectrometry. I rotated the wing around its proximal–distal axis and—while viewing the camera output in real time—measured the angle spanning the point at which UV reflectance first became visible to the point at which reflectance was no longer visible. This is a measure of the limited-view nature of male UV reflectance, hereafter referred to as "UV angular breadth" (see also Kemp and Rutowski 2007).
Data summary and statistical analysis
Because information on the visual sensitivities of E. hecabe are not yet available, I summarized the spectral data obtained from field-caught specimens using a series of variables that together describe the most variable physical properties of the male reflectance curve (refer to Figure 1). Of most relevance here, according to the experimental results, was UV "brightness," a variable calculated as the average reflectance from 300- to 400-nm (i.e., R300–400). I also calculated mid-wave (R400–500) and long-wave (R550–700) brightness to summarize the main visual features of the diffuse yellow. UV "hue" was calculated as the position (to the nearest 1 nm) of the UV peak, and yellow hue was calculated as the wavelength corresponding to the point midway between the long-wave "plateau" and the mid-wave "nadir" (refer to Figure 3 in Kemp and Rutowski 2007). Finally, because "chroma" or "saturation" is adjudged as the relative stimulation of different photoreceptor classes (Endler 1990), I calculated 2 candidate measures of chroma, each consisting of contrasts of discrete reflectance regions. UV chroma (Cuv) and yellow chroma (Cy) were calculated as
 | (1) |
 | (2) |
Standard statistical procedures are used throughout except for the observational analyses investigating the phenotypes of paired versus free-flying field males and those relating quantitative aspects of male phenotype with the size of their in-copula mates. Here, I adopted an information–theoretic approach in which the best fitting (i.e., most informative) model was identified as the linear combination of predictor variables that minimized the value of Akaike's Information Criterion (AIC; Burnham and Anderson 2002). This criterion improves on simply using log-likelihood values because it accounts for the size of the candidate models. An information–theoretic approach is recommended for modeling observational data because traditional analyses are theoretically unjustified and perform poorly in simulations (Burnham and Anderson 2002).
The most informative model in each case was selected using the maximum likelihood–based generalized linear/nonlinear modeling function of STATISTICA (v7.0). Model significance and the significance of incumbent parameters were assessed using log-likelihood and Wald tests, respectively. I also calculated semipartial r values to estimate the size and direction of individual effects. In the analysis comparing paired versus free-flying males, in which the dependent variable is binary, I specified a binomial distribution and logit-link function, whereas in analyses involving female body size as the dependent variable, I specified a normal distribution. Means are accompanied by 95% confidence intervals.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Experiment 1: ambient light manipulation
A total of 22 matings were observed in this experiment, of which 19 (86%) occurred in the UV+ cage (
= 11.6, P < 0.001). This is consistent with a role for UV wing coloration (or at the very least, full-spectrum solar illumination) during mating interactions. There was no evidence for size-assortative mating (Pearson's correlation between male and female wing length—all matings: r = 0.078, N = 22, P = 0.729; matings in the UV+ cage only: r = 0.008, N = 19, P = 0.973).
Experiment 2: wing UV manipulation (small enclosures)
Of a total 54 matings observed in this experiment, 34 (63%) involved control males, which represents a marginally nonsignificant departure from unity ( = 3.63, P = 0.057) and suggests a tendency for males with brighter wing UV to copulate more readily. As with experiment 1, evidence for size-assortative mating was lacking (r = 0.209, N = 54, P = 0.130). However, control males secured copulations with females that were clearly larger, on average, than the mates of treatment males (t52 = 3.44, P < 0.005; Figure 4a).
|
Experiment 3: wing UV manipulation (large enclosures)
The results of this experiment mirrored those of the small cage experiment; except in this case the difference in mating probability between control (33 matings; 69%) versus treatment (15 matings) males achieved significance (
= 6.75, P < 0.01). Although there was no evidence for size-assortative mating (Pearson's r = 0.024, N = 48, P = 0.870), control males again mated with larger females than treatment males (t45 = 2.14, P < 0.05; Figure 4b). Combining data across the 2 rutin experiments (2 and 3) indicated no difference between them in the probability of reduced UV males mating (i.e., the effect of the wing UV manipulation was equivalent; logistic regression: G1 = 0.67, P = 0.413). Factorial analysis of female body size revealed (as observed separately for each experiment) an effect of male wing UV treatment (F1,98 = 14.9, P < 0.0001), no effect of experiment (i.e., the average size of the females that mated in each experiment was statistically indistinguishable; F1,98 = 0.74, P = 0.39), and no treatment x experiment interaction (F1,98 = 0.15, P = 0.70). Thus, the observed covariance between male treatment and female body size was similar across the 2 experiments (Figure 4), regardless of potentially biologically important differences involving sex ratios and the size of the experimental arenas.
Field mating patterns
I collected 161 copulating pairs and 188 free-flying courting males. Interestingly, across all sampled males, UV brightness appeared to be the most variable aspect of their dorsal forewing coloration (Figure 3b). Single variable contrasts (Table 1) indicated that in-copula males were significantly older and possessed less UV bright wings than their free-flying counterparts. There was also a marginally nonsignificant tendency for in-copula males to possess less bright and more whitish long-wave reflectance. In all cases, however, the observed wing color differences were relatively subtle (refer to the effect sizes in Table 1). The most informative multivariable logistic model of male provenance (i.e., paired or free flying), calculated on all wing color variables along with forewing length (but excluding wing wear), was one containing UV brightness only (AIC = 481.1, N = 349, G1 = 4.66, P < 0.05), which simply restates the univariate results (Table 1).
|
Because female body size may provide a visual cue of mate quality, and due to the experimental relevance of this variable (Figure 4), I assessed whether male wing traits covaried with the size of their in-copula mate. Single variable contrasts, conducted on the entire data set (Table 2), revealed significant positive covariance between female winglength and male winglength, UV brightness, and long-wave brightness. The most informative multivariable model of female size (AIC = 496.5, N = 161, G2 = 19.3, P < 0.0001) included male forewing length (Wald = 9.98, P < 0.005; semipartial r = 0.231) and UV brightness (Wald = 8.56, P < 0.005; semipartial r = 0.215). Thus, larger males and males with brighter UV wing markings tended to copulate with larger females (male size and UV brightness were themselves unrelated: r = –0.025, N = 349, P = 0.635).
As noted earlier, an unidentified number of copulating pairs probably resulted from pupal matings, a situation in which females have a restricted opportunity to exert precopulatory mate choice. The analytical signature of wing color–based female mate choice could be diluted by these cases, but some insight into their incidence could be given by an examination of female mating history. Dissections revealed that the 161 copulating females contained either zero (n = 31), 1 (n = 107), 2 (n = 16), 3 (n = 6), or 4 (n = 1) spermatophores. The 31 zero spermatophore cases could only have resulted from copulations involving virgin females that were aborted prematurely (approximately 50% of pairs separated on capture) before spermatophore transfer. Only in cases where females carried 1+ spermatophores could the possibility of pupal mating be unequivocally ruled out, so I repeated the above analyses using these restricted cases (N = 23). In-copula males were compared with an equivalent number of free-flying individuals sampled as soon as possible at the same site. This analysis produced a similar pattern of differences to that observed among the larger data set, but owing to reduced statistical power, no single difference achieved statistical significance (e.g., wing wear: U = 202, P = 0.170; UV brightness: t44 = 1.69, P = 0.098; all other variables: t44 < 1.61, P > 0.11). The best fitting of all potential multivariable models in this case was a nonsignificant one including UV brightness only (AIC = 64.9, G1 = 2.88, P = 0.090).
The restricted analysis of female body size revealed increased covariances between female body size and male wing traits (Table 2, Figure 5). Male UV brightness was the strongest single predictor in this case and explained more than 20% of the variance in female size (data were not modeled further due to the reduced sample). Given the greater potential role of precopulatory mate choice in this data set, this analysis further supports the experimental findings in which unmanipulated males consistently mated with larger, putatively higher quality females (Figure 4).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Experimental manipulation has featured extensively in the study of mating preferences, but in order to have ultimate relevance, putative mating biases must be expressed under natural ecological conditions. In this study, I employed a diversity of manipulative experimental approaches to investigate color-based butterfly mating preferences and compared the results with mating dynamics in a wild population. The experiments furnished several complimentary lines of evidence that female E. hecabe prefer mates bearing bright UV iridescence. Although the results of experiment 1 merely implicate normal levels of UV illumination as being important to sexual behavior, experiments 2 and 3 demonstrate that females are sensitive to differences in male UV brightness that fall within naturally occurring levels of variation (Figure 3). This is significant as the first manipulative evidence for qualitative UV-based female mating preferences in the Coliadinae, and as one of only several such demonstrations in the broader Lepidoptera (see also Robertson and Monteiro 2005
; Kemp 2007). Prior research in visually similar coliadines demonstrated a female preference for the presence of UV markings (Silberglied and Taylor 1973, 1978), and both male age and UV brightness have been correlated with attractiveness (Rutowski 1985; Papke et al. 2007), but qualitative signal attributes have never been manipulatively evaluated. In the second component of this study, an assay of field mating patterns was seen to only partly support the experimental findings. Copulating male UV brightness covaried with female body size, particularly among pairs most conducive to the expression of precopulatory choice (Figure 5), which corroborates the persistent experimental finding in which UV-reduced males paired with smaller-than-average females (Figure 4). In comparison with their free-flying counterparts, however, copulating males were older and actually possessed (subtly) less bright UV wing markings, thus running counter to expectations of the opposite scenario. These findings may stand testimony to the full complexity of sexual interactions in the wild and offer several potentially unique insights into the nature of color-based mate choice in polyandrous pierid butterflies.
A holistic appraisal of the experimental and observational data suggests ways in which they might each be reconciled within a common framework of color-based female choice. First, the slightly duller wing markings of in-copula males appears as a consequence of them being older than their free-flying counterparts (refer to the effect sizes in Table 1). Because older males are more likely to have mated previously, and a history of matings extends the functional duration of future copulations (Rutowski et al. 1987; Kemp DJ, unpublished data), it seems possible that a haphazard sample of copulating pairs would be biased toward the collection of older, duller UV males (Kemp 2006a; although see Rutowski 1985). Unfortunately, it was not possible to determine how long individuals had been copulating prior to their encounter, which negated the analysis of copulation duration (approximately 50% of pairs also separated on capture). Second, although I aspired to sample reproductively active free-flying individuals, in practice it is impossible to control for variation in sexual motivation. Some sampled individuals could have been sexually unmotivated due to young age or recent copulation or due to a preoccupation with acquiring adult resources. On this basis, whereas all subjects used in the experiments were virgin and presumably possessed equivalently high motivation to copulate, this assumption would seem unlikely to hold for all males in a heterogeneously aged field population. In retrospect, therefore, a comparison of in-copula versus free-flying individuals may not furnish a viable field-based assay of mating preferences. Similar issues are discussed by Kemp (2006b), who reports the inability of copulating butterflies from a high-density field assemblage to reflect known regimes of UV-based mate choice in C. eurytheme.
It is also possible—if not likely, given that males have limited paternal resources to donate (Rutowski 1985; see below)—that the evolutionary benefit of attractiveness may be conferred via an increased quality rather than quality of matings. In this case, putatively attractive, bright UV males would not be more likely to be sampled in-copula but would instead be more likely to pair with females of high phenotypic quality (i.e., high reproductive value). Relevant to this point is the robust experimental and field-observational link between male UV brightness and female body size (Figures 4 and 5). This effect seems most likely to reflect either or both of 2 potentially interrelated possibilities: 1) that smaller females are less choosy with respect to male wing coloration and/or 2) that males themselves express mating preferences favoring larger females. Supporting the first possibility is the fact that condition-dependent mating preferences are known for insects (Hunt et al. 2005) and could arise due to the costs associated with expressing choice. Supporting the second possibility is the general expectation that males should also be choosy, and the observation that many male insects, including polyandrous pierid butterflies (Rutowski 1985), exhibit mate choice for phenotypic markers of fecundity such as body size (see review in Bonduriansky 2001). If male E. hecabe generally prefer large females and females prefer bright UV, then we would expect the emergent property of assortative pairing on the basis of these 2 attractiveness markers (similarly to observed; Figure 5). This effect would be exacerbated by condition dependence in the strength of mating preferences in either or both sex (as in scenario 1, above). However, whether or not males are indeed choosy remains to be demonstrated. Of additional interest for future experimentation is the field-observed covariance between male and female body size, which could arise if females prefer large as well as UV bright mates (male size and UV brightness varied independently in this data set).
Assuming that male UV brightness is indeed a component of male attractiveness, what mechanisms could underlie the evolution of such a female mating bias? Given that UV iridescence is known to function (at least) as a species-isolating mechanism in the Coliadinae (Silberglied and Taylor 1973, 1978), one possibility is that female E. hecabe simply favor bright UV as an unambiguous signal of species identity. However, this scenario alone would not necessarily predict the observed relationships between male UV brightness and female body size. In terms of potential direct and indirect benefits, recent empirical work has demonstrated that the expression of bright UV structural color in C. eurytheme is heritable and phenotypically condition dependent (Kemp and Rutowski 2007). Like E. hecabe, males of this species generate a bright, limited-view UV signal via a nanoscale architecture of lamellar thin films, which overlays pigment-generated yellowish orange coloration (Ghiradella 1974). Of particular interest from a direct benefits perspective is the fact that adult males who experience a poor quality larval environment are not only smaller overall but also display duller and more angularly restricted UV reflectance (Kemp and Rutowski 2007). The reduction in UV signal characteristics is more pronounced than changes in the coincident pigmentary yellowish orange. On this basis, given the presence of natural variation in the larval environment, female E. hecabe might profitably use UV brightness as a measure of the likely condition of their mate and their opportunity to receive a more nutritious ejaculate. Research on other members of the Coliadinae (again, notably, C. eurytheme) has revealed that male-donated nutrients, in the form of a spermatophore and accessory secretions deposited within the female reproductive tract, contribute to subsequent female fecundity (Boggs and Gilbert 1979). Smaller males, and males that have mated previously or recently produce smaller ejaculates (Rutowski and Gilchrist 1986; see also Hiroki and Obara [1997] for data on this issue in Japanese populations of E. hecabe), and females mated to these males experience reduced reproductive output (Rutowski et al. 1987). The potential for UV-mediated signaling of mate quality C. eurytheme is also supported by the finding that females mated to males with brighter UV markings experience greater lifetime fecundity (Kemp DJ, unpublished data). Research is presently underway to determine whether this type of UV-mediated direct benefits signaling has the potential to operate in E. hecabe.
Finally, the present experimental findings compliment those in another butterfly (Kemp 2007) and across several birds (e.g., Andersson et al. 1998; Johnsen et al. 1998; Sheldon et al. 1999; although see Liu et al. 2007) and fish (e.g., Endler 1980, 1983) in which females evaluate structurally colored visual signals. These signals differ fundamentally from pigment colors in that they result from selective reflection from an optically functional fine-scale surface architecture. Several workers have suggested, on this basis, that structurally colored ornaments may be uniquely informative of genetic mate quality (Fitzpatrick 1998; Kemp and Rutowski 2007). Because the visual "performance" of a structurally colored signal (i.e., its brightness, chromaticity, angular visibility, etc.) will be determined by the precise fine-scale regularity of the surface architecture, it may provide a macroscale visual indicator of developmental stability. The high genetic variance observed for male-limited structural UV in C. eurytheme (Kemp and Rutowski 2007) is at least consistent with this type of indirect (genetic) benefits signaling, but specific empirical studies of signal content are required. It would also be interesting to know whether males with UV markings enhanced beyond normal levels of UV brightness are increasingly favored by females, which would reveal whether male UV is under truly subject to directional—as opposed to stabilising—selection (see Kemp 2007; Liu et al. 2007). Due to their suitability for further behavioral studies, quantitative genetic studies and developmental manipulation, coliadine butterflies—such as E. hecabe—offer an excellent candidate system for further investigation of this hypothesis and for elucidating the broader evolutionary significance of structurally colored ornaments in nature.
|
FUNDING |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Australian Research Council Discovery-Projects program (DP0557190).
|
ACKNOWLEDGEMENTS |
|---|
I thank W. Cresswell and 2 anonymous referees for constructive comments on the manuscript, R.L. Rutowski for providing UV photographs (Figure 1), and J.M. Macedonia for deriving the transmission curves (Figure 2). K.A. Robertson and A. Monteiro provided helpful advice on using Rutin to manipulate wing reflectance.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Andersson S, Ornborg J, Andersson M. Ultraviolet sexual dimorphism and assortative mating in blue tits. Proc R Soc Lond B Biol Sci (1998) 265:445–450.
Boggs CL, Gilbert LE. Male contribution to egg production in butterflies—evidence for transfer of nutrients at mating. Science (1979) 206:83–84.
Bonduriansky R. The evolution of male mate choice in insects: a synthesis of ideas and evidence. Biol Rev (2001) 76:305–339.[Medline]
Brooks R, Endler JA. Direct and indirect sexual selection and quantitative genetics of male traits in guppies (Poecilia reticulata). Evolution (2001) 55:1002–1015.[CrossRef][Web of Science][Medline]
Brunton CFA. The evolution of ultraviolet patterns in European Colias butterflies (Lepidoptera: Pieridae): a phylogeny using mitochondrial DNA. Heredity (1998) 80:611–616.[CrossRef][Web of Science]
Brunton CFA, Majerus MEN. Ultraviolet colors in butterflies: intra- or inter-specific communication? Proc R Soc Lond B Biol Sci (1995) 260:199–204.
Burnham KP, Anderson DR. Model selection and multimodel inference: a practical information-theoretic approach. (2002) New York: Springer.
Costanzo K, Monteiro A. The use of chemical and visual cues in female choice in the butterfly Bicyclus anynana. Proc R Soc Lond B Biol Sci (2007) 274:845–851.[Medline]
Darwin C. The descent of man and selection in relation to sex. (1874) London: John Murray.
Drummond BA. Multiple mating and sperm competition in the Lepidoptera. In: Sperm competition and the evolution of animal mating systems—Smith RL, ed. (1984) London: Academic Press. 291–370.
Endler JA. Natural selection on color patterns in Poecilia reticulata. Evolution (1980) 34:76–91.[CrossRef][Web of Science]
Endler JA. Natural and sexual selection on color patterns in poeciliid fishes. Environ Biol Fishes (1983) 9:173–190.[CrossRef]
Endler JA. On the measurement and classification of color in studies of animal color patterns. Biol J Linn Soc (1990) 41:315–352.[Web of Science]
Endler JA, Houde AE. Geographic variation in female preferences for male traits in Poecilia reticulata. Evolution (1995) 49:456–468.[CrossRef][Web of Science]
Fitzpatrick S. Color schemes for birds: structural coloration and signals of quality in feathers. Ann Zool Fenn (1998) 35:67–77.
Ghiradella H. Development of ultraviolet-reflecting butterfly scales: how to make an interference filter. J Morphol (1974) 142:395–410.[CrossRef][Web of Science]
Hill GE. Female house finches prefer colorful males: sexual selection for a condition-dependent trait. Anim Behav (1990) 40:563–572.[CrossRef][Web of Science]
Hill GE. Plumage coloration is a sexually selected indicator of male quality. Nature (1991) 350:337–339.[CrossRef]
Hiroki M, Obara Y. Delayed mating and its cost to female reproduction in the butterfly, Eurema hecabe. J Ethol (1997) 15:79–85.[CrossRef]
Houde AE. Effect of artificial selection on male color patterns on mating preference of female guppies. Proc R Soc Lond B Biol Sci (1994) 256:125–130.
Hunt J, Brooks R, Jennions MD. Female mate choice as a condition-dependent life-history trait. Am Nat (2005) 166:79–92.[CrossRef][Web of Science][Medline]
Jiggins CD, Naisbit RE, Coe RL, Mallet J. Reproductive isolation caused by color pattern mimicry. Nature (2001) 411:302–305.[CrossRef][Medline]
Johnsen A, Andersson S, Örnborg J, Lifjeld JT. Ultraviolet plumage ornamentation affects social mate choice and sperm competition in bluethroats (Aves: Luscinia s. svecica): a field experiment. Proc R Soc Lond B Biol Sci (1998) 265:1313–1318.
Johnsen A, Delhey K, Andersson S, Kempenaers B. Plumage color in nesting blue tits: sexual dichromatism, condition dependence and genetic effects. Proc R Soc Lond B Biol Sci (2003) 270:1263–1270.[Medline]
Jones RE. Phenotypic variation in Australian Eurema species. Aust J Zool (1992) 40:371–383.[CrossRef][Web of Science]
Kemp DJ. Heightened phenotypic variation and age-based fading of ultraviolet butterfly wing coloration. Evol Ecol Res (2006a) 8:515–527.
Kemp DJ. Ultraviolet ornamentation and male mating success in a high density assemblage of the butterfly Colias eurytheme. J Insect Behav (2006b) 19:669–684.[CrossRef]
Kemp DJ. Female butterflies prefer males bearing bright iridescent ornamentation. Proc R Soc Lond B Biol Sci (2007) 274:1043–1047.[Medline]
Kemp DJ, Rutowski RL. Condition dependence, quantitative genetics, and the potential signal content of iridescent ultraviolet butterfly coloration. Evolution (2007) 61:168–183.[CrossRef][Web of Science][Medline]
Kemp DJ, Rutowski RL, Mendoza M. Color pattern evolution in butterflies: a phylogenetic analysis of structural ultraviolet and melanic markings in North American sulphurs. Evol Ecol Res (2005) 7:133–141.
Kobayashi A, Hiroki M, Kato Y. Sexual isolation between two sympatric types of the butterfly Eurema hecabe (L.). J Insect Behav (2001) 14:353–362.[CrossRef]
Kokko H, Jennions MD, Brooks R. Unifying and testing models of sexual selection. Annu Rev Ecol Evol Syst (2006) 37:43–66.[CrossRef]
Liu M, Siefferman L, Hill GE. An experimental test of female mate choice relative to male structural coloration in eastern bluebirds. Behav Ecol Sociobiol (2007) 61:623–630.[CrossRef][Web of Science]
Mavarez J, Salazar CA, Bermingham E, Salcedo C, Jiggins CD, Linares M. Speciation by hybridization in Heliconius butterflies. Nature (2006) 441:868–871.[CrossRef][Medline]
Papke RS, Kemp DJ, Rutowski RL. Multimodal signalling: structural ultraviolet reflectance predicts male mating success better than pheromones in the butterfly Colias eurytheme L. (Pieridae). Anim Behav (2007) 73:47–54.[CrossRef][Web of Science]
Parker TH, Garant D. Quantitative genetics of sexually dimorphic traits and capture of genetic variance by a sexually-selected condition-dependent ornament in red junglefowl (Gallus gallus). J Evol Biol (2004) 17:1277–1285.[CrossRef][Web of Science][Medline]
Qvarnstrom A, Brommer JE, Gustafsson L. Testing the genetics underlying the co-evolution of mate choice and ornament in the wild. Nature (2006) 441:84–86.[CrossRef][Medline]
Robertson KA, Monteiro A. Female Bicyclus anynana butterflies choose males on the basis of their dorsal UV-reflective eyespot pupils. Proc R Soc Lond B Biol Sci (2005) 272:1541–1546.[Medline]
Rutowski RL. Evidence for mate choice in a sulphur butterfly (Colias eurytheme). Z Tierpsychol (1985) 70:103–114.[Web of Science]
Rutowski RL, Gilchrist GW. Copulation in Colias eurytheme (Lepidoptera: Pieridae): patterns and frequency. J Zool (1986) 207:115–124.
Rutowski RL, Gilchrist GW, Terkanian B. Female butterflies mated with recently mated males show reduced reproductive output. Behav Ecol Sociobiol (1987) 20:319–322.[CrossRef][Web of Science]
Rutowski RL, Macedonia JM, Morehouse N, Taylor-Taft L. Pterin pigments amplify iridescent ultraviolet signal in males of the orange sulphur butterfly, Colias eurytheme. Proc R Soc Lond B Biol Sci (2005) 272:2329–2335.[Medline]
Sheldon BC, Andersson S, Griffith SC, Ornborg J, Sendecka J. Ultraviolet color variation influences blue tit sex ratios. Nature (1999) 402:874–877.
Silberglied RE, Taylor OR. Ultraviolet differences between sulfur butterflies, Colias eurytheme and C. philodice, and a possible isolating mechanism. Nature (1973) 241:406–408.[CrossRef]
Silberglied RE, Taylor OR. Ultraviolet reflection and its behavioral role in the courtship of the sulphur butterflies Colias eurytheme and C. philodice (Lepidoptera, Pieridae). Behav Ecol Sociobiol (1978) 3:203–243.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. G Meadows, M. W Butler, N. I Morehouse, L. A Taylor, M. B Toomey, K. J McGraw, and R. L Rutowski Iridescence: views from many angles J R Soc Interface, April 6, 2009; 6(Suppl_2): S107 - S113. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. M Doucet and M. G Meadows Iridescence: a functional perspective J R Soc Interface, April 6, 2009; 6(Suppl_2): S115 - S132. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||