SEARCH

SEARCH BY CITATION

Keywords:

  • colour variation;
  • ecological speciation;
  • intrasexual selection;
  • sensory drive

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The evolution of intersexual interactions, like mate choice, during ecological speciation has received widespread attention. However, changes in intrasexual interactions, like male territoriality, during ecological divergence are largely unexamined. We conducted field experiments with adaptively diverged populations of the eastern fence lizard (Sceloporus undulatus) to determine whether territorial males behaved differently towards ecologically similar vs. dissimilar intruders. We performed trials with light-coloured males from White Sands, New Mexico and dark-coloured males from the surrounding desert. We found that intruders from White Sands elicited more aggression than intruders from dark-soil habitat. We also documented a case of ‘sex confusion’ where white-sand males courted dark-soil intruders. We found population differences in signalling patch size that can explain both aggression bias and sex misidentification. We argue that direct selection (for population recognition or optimal signal transmission) and indirect selection (by-products of ecological adaptation) should influence both intersexual and intrasexual interactions during ecological speciation.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Models of ecological speciation focus on the interaction between adaptive changes, which allow exploitation of novel resources, and intersexual interactions, specifically the evolution of mate choice (e.g. Seehausen et al., 2008; Servedio, 2004; Schluter, 2009). However, in many cases, the same conditions that promote shifts in intersexual interactions during ecological speciation should also influence intrasexual interactions (Seehausen & Schluter, 2004; Dijkstra et al., 2008). For example, ecological speciation may often involve direct selection for mechanisms of population recognition given the importance of identifying not only appropriate mates (Seehausen et al., 2008) but also appropriate competitors. Additionally, if the signal transmission environment (e.g. display microhabitat, light environment) differs for incipient ecological species, colonization of novel habitats can lead to evolution of signals for optimal transmission (i.e. sensory drive; Boughman, 2002; Endler, 1992; Leal & Fleishman, 2004), altering both inter- and intrasexual interactions that rely on visual signalling. Finally, in a novel environment, selection on ecological traits can have by-product effects on traits used for both inter- and intrasexual interactions (Dijkstra et al., 2008; Maan et al., 2010). We therefore assert that the same processes underlying ecological speciation (e.g. population recognition, sensory drive, and by-product effects) are likely to influence intrasexual interactions, particularly for territorial species that use visual signals for intraspecific communication (Plath & Strecker, 2008; Seehausen & Schluter, 2004).

In many vertebrate species, territorial behaviour is a key intrasexual interaction because it is an important determinant of the fitness of males. For lizards, successful territory acquisition and defense directly influence the reproductive success of males because larger territories give male lizards access to more females (Haenel et al., 2003a,b). In territorial lizards with polygynous mating systems, male territory size is often determined by morphological, physiological, and behavioural characteristics (Meyers et al., 2006; Husak et al., 2007). For example, morphological characteristics like body size and signal patch colour are correlated with the outcome of male competitive interactions in Iguanian lizards (Ord et al., 2001; Stapley & Keogh 2006). Similarly, physiological measures like plasma testosterone levels are associated with dominance in territorial lizards (Tokarz, 1995; Watt et al., 2003). Finally, behaviours, such as dominance displays and weapon performance (Crotophytus collaris: Lappin & Husak 2005; Rankinia diemensis: Stuart-Smith et al., 2007), can also be strong predictors of the fitness of males in territorial lizards.

Because territoriality is costly to males, the accurate identification of threats is important. Territory defense and intrasexual aggression are energetically demanding (Ydenberg et al., 1988; Vehrencamp et al., 1989; Whiting, 1999). Conspicuous visual signals displayed during male–male combat can also attract predators, further placing males at risk. To avoid these costs, territorial males should discriminate between males that threaten the stability of their territory and males that pose little threat. In other vertebrates, territorial males are often more aggressive to ecologically similar males that are competitors for access to local females than they are to ecologically dissimilar males (Seehausen & Schluter, 2004; Dijkstra et al., 2008). Some studies of intrasexual behaviour with territorial lizards indicate that males can discriminate between threatening and nonthreatening intruders (Van Dyk & Evans, 2007; Carazo et al., 2008), but whether territorial lizards can discriminate between ecologically similar and ecologically dissimilar males remains untested.

Our study of male territoriality focuses on rapidly diverging populations of Sceloporus undulatus (eastern fence lizard; see Leache, 2009 for discussion of taxonomy) across a substrate colour ecotone in southern New Mexico. The dunes of White Sands are a recent geological formation (2000–6000 years) composed of 275 square miles of gypsum sand (Kocurek et al., 2007). The stark white-sand dunes are in striking contrast to the dark substrate of the surrounding Chihuahuan dark-soil habitat. White Sands populations of S. undulatus exhibit cryptic, blanched colouration, presumably evolved through natural selection to avoid predation. The blanched forms are easily distinguished from dark-coloured conspecifics in the surrounding dark-soil habitat. Dorsal colouration in this system is heritable (Rosenblum, 2004), and the genetic basis of differences in melanin has been identified (Rosenblum et al., 2010). A single dominant amino acid substitution in the Mc1r gene is responsible for shifts in melanin in white-sand S. undulatus (Rosenblum et al., 2010). Melanin production is a key determinant of both dorsal and signalling colours (e.g. Quinn & Hews, 2003) and has a simple genetic basis. Therefore, natural selection on dorsal colouration may have a by-product effect on colour patches used in male territorial displays. Indeed, divergence in dorsal colouration is accompanied by concordant divergence in S. undulatus social signalling colour patches. The bright blue ventral colours in S. undulatus have a demonstrated social function in male–male territorial displays (Cooper & Burns, 1987) and differ between white-sand and dark-soil populations in brightness, chroma, and hue (Robertson & Rosenblum, 2009). Intrasexual territorial interactions are important in this species, and males defend territories with highly stereotyped behaviour (see Methods).

Given the rapid divergence between white-sand and dark-soil populations of S. undulatus, the importance of territoriality in this species, and the evolution of colour patches with potential relevance for both natural and sexual selection, we were interested in whether adaptation to the novel habitat is associated with shifts in intrasexual interactions. We conducted behavioural trials with both white-sand and dark-soil S. undulatus to determine whether males discriminate between ‘home’ (ecologically similar) and ‘away’ (ecologically dissimilar) intruders. We found that (i) white-sand males behaved differently towards home and away intruders, (ii) white-sand males exhibited inappropriate courtship behaviour towards away males, and (iii) population and sex differences in social signal patches may explain patterns of aggression and courtship. These results argue for increased attention to changes in intrasexual interactions during ecological divergence following invasion of novel habitats.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Experimental encounters

We studied male behaviour in two populations of S. undulatus in southern New Mexico: the ‘dark-soil’ form at Jornada Long-Term Ecological Research Station and the ‘white-sand’ form at White Sands National Monument. We chose focal populations with obvious phenotypic and habitat differences to determine whether males could distinguish between ecologically similar and ecologically dissimilar competitors. Although we used allopatric populations, our experiment design is evolutionarily relevant. There is a narrow transition zone between white-sand and dark-soil populations, where S. undulatus individuals are intermediate in dorsal colouration. Molecular studies have also revealed some migration across the ecotone (Rosenblum, 2006; Rosenblum et al., 2007), indicating that dark-soil and white-sand populations are not genetically isolated and could interact at the ecotone.

We quantified and compared male territoriality and aggression in response to home (ecologically similar) and away (ecologically dissimilar) intruder males using a 2 × 2 reciprocal design (Fig. 1). We conducted a total of 80 trials with 20 trials in each treatment group. We conducted behavioural trials in the natural territory of each focal male. Staged encounters in the natural territory of the focal male provide an advantage over trials conducted in the laboratory because of the motivation of the focal male to defend a natural territory. We conducted trials from May to July 2009 during the breeding season when territorial behaviour is highest (Smith & John-Alder, 1999) and during the hours of 06:30–14:00 when males are most active. All males in the study were determined to be reproductively mature based on body size (SVL ≥ 5.6 cm, mass ≥5.9 g).

image

Figure 1.  Staged arena encounters occurred in the natural territory of dark-soil and white-sand focal males. We used a 2 × 2 reciprocal design to examine the response of focal males to both ‘home’ and ‘away’ intruder males. We conducted 20 trials for each of four treatment groups, for a total of 80 trials.

Download figure to PowerPoint

Territorial males were spotted by eye, captured by noose or hand, and immediately placed in a circular metal-flashing arena (diameter = 0.85 m; height = 0.35 m) at the site of capture. The arena had no floor, so all interaction occurred on natural substrate. The inside wall of the arena was painted light brown to eliminate reflectance. After a five-minute acclimation of the focal male to the experimental arena, we lowered the intruder male into the arena approximately 40 cm from the focal male. We tethered the intruder by string to the rod used for lowering, and the tether was long enough for intruder males to behave freely in the arena. Intruder males were tethered for consistency of placement within the arena across trials and to reduce focal male exposure to researchers. We did not tether focal males in any trial. Similar experiments with S. undulatus and other Iguanian lizards have shown that tethering does not affect the behaviour of the tethered lizard or the free-roaming lizards to which they are presented (e.g. Cooper & Burns, 1987, Rosenblum, 2008). We recorded each five-minute trial using a digital video camera (Canon FS11, Canon, Lake Success, NY, USA). We recorded focal male behaviour in the field and subsequently rescored and verified behavioural observations from the videos.

Because other studies on Iguanian lizards have shown that territorial males exhibit less aggression towards familiar males (Whiting, 1999), we ensured that intruder males were not neighbours to the focal male. In addition, we size-matched focal and intruder males to reduce a possible confounding effect of size difference in agonistic behaviour, as male dominance is associated with larger body size in many lizards (Jenssen et al., 2005; Stuart-Smith et al., 2007; Sacchi et al., 2009). In 80 trials, we used 40 different intruder males (19 and 21 males from dark-soil and white-sand, respectively). Most males participated in 1–3 trials (mean ± SD = 1.68 ± 0.73), but two males were included in six and eight trials. Most intruder males were used 1–2 times per day (mean ± SD = 1.45 ± 0.66).

After each trial, the focal male was brought to the field-station laboratory for morphological measurements and digital photography. The male was then kept in captivity and used as an intruder male in subsequent trials, or returned directly to the field (after marking to avoid resampling). Each male was housed individually with access to food, water, heat, and UV light. We blocked all cage walls to eliminate male interactions while in captivity.

Because white-sand males exhibited behaviour typical of courtship to dark-soil males (see Results), we conducted 60 male–female trials to characterize the response of male S. undulatus to reproductively mature females. The evolution of intersexual behaviour was not the focus of this study but male–female trials were important for documenting the species-typical courtship response for these populations. Finally, we also presented white-sand focal males with a syntopic heterospecific lizard, Holbrookia maculata in preliminary trials. Focal males did not respond to the heterospecific intruder male, indicating appropriate species recognition.

Behavioural analyses

Sceloporus undulatus behaviour is highly stereotyped and can be broadly categorized as territorial, courtship, and exploratory (Cooper & Burns, 1987). Some behaviours are employed in more than one context, whereas other behaviours are context-specific (Martins et al., 2005). Territorial displays often are comprised of a short-burst of shudder bobs followed by a gular (throat) extension and/or lateral compression (Carpenter & Ferguson, 1977; Martins, 1993), but can also include tail whips, head butts, and biting (Carpenter & Ferguson, 1977; Martins, 1993). An extreme and unambiguous territorial display, termed ‘full show’ is characterized by elevation of the body with full gular extension and lateral compression (Carpenter & Ferguson, 1977). In contrast, courtship displays are characterized by sustained shudder bobs (rapid up–down movement of head), licking of the tail and neck, and mounting. Push-ups are used in many different contexts and are considered exploratory behaviour when performed alone (Carpenter & Ferguson, 1977).

We used two statistical approaches to test for population differences in territorial and courtship behaviour (i.e. categorical and continuous analyses detailed below). Both analyses are based on an unambiguous measure of territoriality (full show) and an unambiguous measure of courtship (shudder bob and mounting). We also measured a number of additional variables such as latency to and total number of territorial and courtship bouts. Results of analyses with these variables are not reported but were all consistent with those presented here. Individuals that did not engage in any behaviour were excluded from analyses.

Categorical analysis of behavioural repertoire

For the categorical repertoire analyses, we categorized the behaviour of each territorial male in response to home and away intruders into one of three categories: exploratory behaviour only, territorial behaviour, and courtship behaviour only. For these categorical analyses, the data were coded as follows: individuals that engaged exclusively in territorial aggression (full show) or initiated a response with shudder bobs, but subsequently engaged in territorial aggression (full show) were coded as ‘territorial’. Individuals that exclusively engaged in courtship (shudder bobs and mounting only) were coded as ‘courtship’. Individuals that responded with push-ups but did not engage in full show or courtship were coded as ‘exploratory’. We used contingency analyses to compare the frequency of each behavioural category among the four treatment groups and tested for significance using χ2 tests. All analyses were conducted in jmp ver 8 (SAS Institute Inc., Cary, NC, USA).

Continuous analysis of time in territorial and courtship display

Next, we quantified how much time territorial males spent in territorial and courtship displays in response to home and away intruders. For these continuous analyses, the data were coded as the total time the focal male spent in territoriality and total time the focal male spent in courtship (range = 0–300 s). Territorial and courtship responses were analysed separately. We used standard least squares regression using restricted maximum likelihood to test for the effect of intruder male (random effect), intruder male population, focal male population, and their interactions on behavioural response by the focal male. From each model, the F ratio and associated degrees of freedom were used to generate a P-value to test for significance of each effect. We log-transformed continuous variables prior to analyses for normality. We initially tested for an effect of the differences in body size (SVL) between the focal and intruder male, but – as expected because males were size-matched – SVL differences were not significant in any model and therefore not included in the final analyses.

Social signalling colour patch size variation

White-sand and dark-soil populations diverge in a suite of phenotypic traits including dorsal and social signal colouration (Rosenblum, 2006; Robertson & Rosenblum, 2009), yet no study has yet quantified whether there are differences between populations in patch size. Here, we compared the patch size of male and female lizards from white-sand and dark-soil populations to determine population and sex differences in patch size. Patch size was measured in a larger sample than the individuals in the behavioural trials and included 19 females (9 dark-soil, 10 white-sand) and 77 males (40 dark-soil, 37 white-sand). To quantify patch size, we photographed the ventral side of all individuals with a NikonD70 SLR digital camera (Nikon Inc., Melville, NY, USA). Photographs were imported into Adobe Photoshop CS (version 9.0.2, Adobe Systems Inc., San Jose, CA, USA). We corrected for ambient light conditions by reference to a grey standard (QP101) in the background of every photograph. We outlined and selected the entire ventral surface of each individual excluding the limbs and tail using the magnetic lasso tool. We then used the magic wand function to select the social signal patch for each individual. The optipix filter in Photoshop (Reindeer Graphics®; Wide Histogram, Reindeer Graphics, Ashenville, NC, USA) provides a count for the total number of pixels in the selected portion of a photograph. We calculated the signalling patch size as the percentage of blue and black pixels relative to the number of pixels in the total ventral area. Patch size was arcsine-square-root-transformed prior to analysis for normality. We used an one-way anova to determine whether social signalling patch size varied between populations and between sexes and Tukey–Kramer HSD to determine pairwise differences in patch size. We also tested for an effect of patch size on focal male behaviour with the subset of samples used for behavioural trials. We used an one-way anova to determine whether there was a difference between the patch size of intruder males when focal males displayed aggressive or courtship behaviour.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

We conducted 40 trials with dark-soil focal males (20 trials each with home and away intruders) and 40 trials with white-sand focal males (20 trials each with home and away intruders) for a total of 80 trials (Fig. 1). Focal males responded behaviourally to intruders in 59 trials and remained motionless in 21 trials. The trials without interaction were evenly distributed across treatment groups. As expected for motivated, territorial males, both white-sand and dark-soil focal males exhibited predominately territorial behaviour in response to home intruders. We analysed the data from all trials together and also by omitting the 21 trials with no behaviour. Because we found no change in results, we report the analyses from the 59 trials with behavioural interaction. These 59 trials included 36 different intruder males (16 and 20 males from dark-soil and white-sand, respectively).

Categorical analysis of behavioural repertoire

The frequency of each behavioural response (exploratory, courtship, or territorial) differed among the four treatments (Fig. 2; χ26 = 34.18, < 0.00001). In response to home males, dark-soil and white-sand focal males both responded with either exploratory or territorial behaviour. In response to away males, dark-soil focal males exclusively engaged in territorial behaviour whereas white-sand focal males responded with courtship in more than 50% of trials (Fig. 2). The courtship response by white-sand focal males when presented with dark-soil intruders – characterized by sustained shudder bobs and mounting – is commonly observed when S. undulatus males are presented with conspecific females during the breeding season (Fig. 2). Although some males from each treatment group displayed an initial short-burst of shudder bobs to intruders, in all other treatment groups these shutter bobs were always followed by territorial display behaviour (Fig. 2).

image

Figure 2.  Categorical behavioural response of focal males towards home and away intruders varied among treatment groups (χ26 = 34.18, < 0.001). Focal male behaviour was categorized as: push-ups only (exploratory), full show (territorial), and shudder bobs and mounting only (courtship). White-sand males exhibited courtship behaviour towards away males in over 50% of trials (left panel). Typical male courtship response in 60 male–female trials (right panel).

Download figure to PowerPoint

Continuous analysis of time in territorial and courtship display

We detected a strong effect of intruder male on the time focal males spent in both territorial and courtship displays. Focal males spent more time in territoriality when the intruder male was from white-sand habitat. In contrast, focal males spent more time in courtship when the intruder male was from dark-soil (Table 1, Fig. 3). The interaction effect between focal and intruder male was not significant for analyses of territoriality, indicating that both white-sand and dark-soil males responded to white-sand intruders with increased aggression. However, there was a significant interaction effect between focal and intruder male in courtship analyses (Table 1). The significant interaction effect in courtship analyses was because of the fact that time in courtship was longest when white-sand focal males were presented with dark-soil intruders, consistent with categorical behavioural repertoire differences (Fig. 2).

Table 1.   Standard least squares regression using restricted maximum likelihood method for predicting the effect of intruder male, focal male, and their interaction on territoriality and courtship in dark-soil and white-sand Sceloporus undulatus. Both models include a random effect for intruder male individual. F ratio and associated degrees of freedom (d.f.) were used to generate a P-value to test for effect significance. DS, dark-soil.
SourceTerritorialityCourtship
Estimated.f.F ratioP-valueEstimated.f.F ratioP-value
Focal male DS0.641,401.840.181−0.751,474.2320.045
Intruder male DS−1.821,1912.860.0011.411,1613.3840.002
Focal male DS*intruder male DS0.8751,403.390.072−0.751,474.2320.045
image

Figure 3.  Time spent in territorial and courtship displays varied among treatment groups. Territoriality (left panel) represents the average time in full show (±1 SEM), and courtship (right panel) represents the average time in shudder bobs and mounting (±1 SEM). For territoriality, we detected an effect of intruder male, indicating that focal males were more aggressive to white-sand intruders. For courtship, we detected an effect of intruder male and an interaction effect between focal and intruder, indicating that time in courtship was longest when white-sand focal males were presented with dark-soil intruders.

Download figure to PowerPoint

Social signalling colour patch size variation

For the population-level analyses with expanded sampling, ventral signalling patch size varied by both population and sex (anova: F3,92 = 63.72, P < 0.0001; Fig. 4). Average patch size was largest for white-sand males (x ± SD = 39.08 ± 9.90%), intermediate for dark-soil males (20.46 ± 5.35%) and white-sand females (14.58 ± 5.72%), and smallest for dark-soil females (4.81 ± 5.72%). Tukey–Kramer HSD post hoc tests showed that the means for all pairwise comparisons were significantly different (all P < 0.05). For the experimental trials, ventral signalling patch size of the intruder male was associated with focal male behaviour. Focal males exhibited aggressive behaviour towards intruder males with larger patches (anova: F1,55 = 6.6, P = 0.0128), and focal males exhibited courtship behaviour towards intruder males with smaller patches (anova: F1,55 = 5.2, P = 0.026).

image

Figure 4.  Social signal patch size varied by population and sex. The average patch size (±1 SD) was largest for white-sand males and smallest for dark-soil females. A grey-scale image (drawn from digital photographs) represents the average patch size for each group (percentage of vential). All groups differed from each other in Tukey–Kramer post hoc tests (P < 0.05). ws, white-sand; ds, dark-soil.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Here, we show changes in male territorial behaviour for lizard populations adapted to divergent habitats, suggesting that ecological divergence can have strong consequences for male–male interactions. Fitness in male Iguanian lizards depends strongly on territoriality (e.g. territory size, quality, and access to breeding females). Because territorial behaviour is costly, males should be able to appropriately identify potential competitors. Here, we show that S. undulatus populations across the white-sand ecotone in southern New Mexico behave differently towards ecologically similar and dissimilar males. We also describe an unusual case of sex confusion whereby males in the novel habitat appear to misidentify ecological dissimilar male intruders as females. We propose that both aggression and misidentification could be related to the evolution of social signalling colour patches in the white-sand lizards.

Patterns of male territorial behaviour

We found that both white-sand and dark-soil males responded differently towards ecologically similar and ecologically dissimilar male intruders (Figs 2 and 3). Thus, despite the recent time since divergence (<6000 years), the differences between white-sand and dark-soil populations are meaningful for intrasexual interactions. Although males in our study behaved differently towards different classes of intruders, there was not a simple home vs. away effect. Both white-sand and dark-soil males were more territorial towards white-sand males [i.e. white-sand males were more aggressive towards home males, whereas dark-soil males were more aggressive towards away males (Figs 2 and 3)]. The fact that focal males adjusted their behavioural response to intruders with different phenotypes is relevant for understanding the early stages of ecological speciation in territorial species and suggests the potential for rapid evolution of intrasexual interactions (e.g. Martins et al., 1998).

An unusual pattern observed in our experiment was that white-sand focal males were not only less aggressive towards dark-soil intruder males, but actually exhibited stereotyped courtship behaviour towards these intruders (Fig. 2). White-sand males were clearly motivated to defend their territories from intruders as evidenced by their aggressive behaviour towards home intruders. Unlike dark-soil focal males, which engaged in territorial behaviour in most trials, white-sand males did not recognize male dark-soil intruders as threats, and in fact exhibited courtship behaviour to them in more than 50% of trials. There are examples in other lizard species of alternative male reproductive tactics that lead to sex misidentification [e.g. sneaker-male behaviour and female mimicry to gain access to reproductive females (Zamudio & Sinervo, 2000, Whiting et al., 2009, Calsbeek & Sinervo, 2008)]. The ‘sex confusion’ observed for S. undulatus is instead almost certainly a nonadaptive by-product effect of ecological divergence and demonstrates how adaptation to novel environments can have rapid and dramatic effects on appropriate identification of potential competitors. Understanding the specific ecological or evolutionary effects of this sex confusion requires additional study to determine whether the misidentification persists following repeated bouts of interaction.

Cues influencing male behaviour

Why did white-sand lizards elicit more aggression than dark-soil lizards and why did white-sand males court dark-soil males? Territorial response depends on the focal male perceiving an intruder, determining whether the intruder is a threat, and mounting an appropriate behavioural response (Endler, 1992; Dijkstra et al., 2008). The most obvious difference between white-sand and dark-soil lizards is their dorsal colour, but this phenotype is unlikely to explain patterns of territorial behaviour. Given the lateral compression and male body posture during S. undulatus territorial displays, dorsal colour is not the primary colour signal transmitted between interacting males (Cooper & Burns, 1987; Sheldahl & Martins, 2000).

The most likely mechanism explaining differences in S. undulatus behaviour towards different classes of intruder males involves social signalling colour patches. Ventral colour patches on the belly and gular region are important for intrasexual interactions and prominently displayed during conspecific interactions in S. undulatus (Rand & Williams, 1970; Losos, 1985; Cooper & Burns, 1987; Sheldahl & Martins, 2000). White-sand and dark-soil males exhibit significant differences in signal patch colour and size. White-sand males have larger, greener signal patches whereas dark-soil males have smaller, bluer signal patches (Robertson & Rosenblum, 2009; Fig. 4). In our behavioural trials, focal males were more likely to exhibit aggressive behaviour towards intruder males with larger ventral patches. Our results are thus consistent with behavioural data from other Iguanian lizards that suggest males with larger belly patches should be perceived as a greater threat (Hover, 1985; Zucker, 1989; Macedonia et al., 2004). Aggression bias towards white-sand intruder males could be explained by larger patch size; however, behavioural trials that specifically manipulate patch size are needed to confirm this hypothesis.

Social signalling colour patches are important not only for population recognition but also for sex discrimination (Cooper & Burns, 1987; Kwiatkowski & Sullivan, 2002). Sceloporus undulatus males generally have prominent blue ventral signalling patches, whereas females have reduced or absent ventral colouration. A previous experiment that manipulated S. undulatus ventral colouration provided compelling evidence that sex recognition in this species is based on the extent of ventral colouration (Cooper & Burns, 1987). Males whose vents were artificially painted to resemble female social signal colouration were courted, whereas females whose vents were painted to resemble males incited aggression (Cooper & Burns, 1987). Our study provides a natural example of how evolution of social signal colouration could affect behaviour in this species. As presented earlier, white-sand males have significantly larger signalling patches than dark-soil males. In fact, the ventral patches of dark-soil males are more similar in size to those of white-sand females than those of white-sand males (Fig. 4). Further, in our behavioural trials, focal males were more likely to court intruder males with smaller ventral patches. Thus, the most likely explanation for why white-sand males exhibited courtship instead of territorial behaviour to dark-soil males is that they perceived dark-soil males as females because of signal colour patch evolution.

One caveat is that we did not explicitly manipulate any of the visual, behavioural, or chemical cues that could influence male territorial behaviour. Our study sought to determine whether patterns of intrasexual interaction have changed in diverging lizard populations, and follow-up studies will be necessary to determine the specific cues that mediate the observed differential response. Behavioural trials that explicitly manipulated S. undulatus social signal colouration demonstrated a strong effect of signal colouration on male behaviour (Cooper & Burns, 1987). However, combined manipulation of social signals and other cues in treatment groups is required to unambiguously determine the effect of multiple signals on S. undulatus territorial response.

Sensory drive and ecological speciation

The sensory drive hypothesis of speciation involves the divergent evolution of male social signalling traits in distinct habitats and the evolution of female preference for those traits (e.g. Endler, 1992; Boughman, 2002). Sensory drive relies on the coordinated evolution of the signal, the sensory system and perceptual ability of the receiver, and the effective transmission of the signal in each habitat. Sensory drive is particularly relevant for models of ecological speciation (Leal & Fleishman, 2004; Maan et al., 2006; Cummings, 2007). Many examples of ecological speciation involve changes in habitat characteristics (e.g. habitat complexity, light environment) and organismal phenotype (e.g. colour, morphology) that are relevant to sensory drive (Boughman, 2002). For example, stickleback benthic and limnetic ecomorphs have different advertisement colours (Boughman, 2001), and Lake Victoria cichlids have evolved visual systems to detect variable male phenotypes under a light gradient (Maan et al., 2006; Seehausen et al., 2008).

Although sensory drive has been primarily demonstrated for female choice during ecological speciation, the transmission and receipt of male social signals to other territorial males should operate under the same principle. Thus, understanding shifts in intrasexual interactions during ecological speciation will require additional attention to the signalling environment, the cues of intruder males and the sensory system of focal males. There are several reasons why the white-sand system will be fruitful for future studies of sensory bias as an explanation of variable male territorial behaviour. First, the white-sand and dark-soil habitats have extremely different transmission environments, both in terms of display background (substrate colour) and light environment (Robertson & Rosenblum, 2009). Second, the social signalling colours of white-sand and dark-soil S. undulatus are divergent in colour and size and are sexually dimorphic (Robertson & Rosenblum, 2009; Fig. 4). Third, there is evidence for differences in visual acuity between white-sand and dark-soil S. undulatus populations with white-sand lizards exhibiting reduced visual sensitivity (S. Nava, unpublished data). Fourth, the simple genetic basis for changes in melanin that affect both dorsal and signalling colours may provide a direct link between natural and sexual selection in this system. Evolution of social signal colouration at White Sands could therefore be a by-product of natural selection for changes in melanin production and/or could be a direct target of selection because of shifts in display environment and visual acuity in the novel habitat. More generally, we expect that changes in intersexual interactions during ecological speciation might often be accompanied by changes in intrasexual interactions and that these changes merit increased attention.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

We thank Emília Martins, Saul Nava, Diana Hews, and Luke Harmon for helpful discussions about experimental design and Michael Gründler for assistance in the field. The Rosenblum and Harmon laboratories and two reviewers provided critical feedback on the manuscript. We thank White Sands National Monument and Jornada Long-term Ecological Research Station for logistical support in the field and permission to conduct research, and New Mexico Game and Fish for research and collecting permits. All experiments were conducted under approved University of Idaho ACUP 2009-37.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References