Vanessa S. Quinn, Department of Biology, Indiana State University, Terre Haute, IN 47809, USA. E-mail: firstname.lastname@example.org
The study of multicomponent signals in the context of social systems has generated interesting results demonstrating that complex signals are used in many communication systems. The multicomponent signal in the majority of Sceloporus lizards consists of a color signal (blue abdominal coloration) and the behavioral display of the color signal (fullshow behavior). In a small number of species, males have lost the color signal. We staged outdoor trials between conspecific males in two closely related Sceloporus lizards that differ in the presence of blue abdominal coloration. In the species with the evolutionary loss of blue abdominal coloration (Sceloporus virgatus), fullshow behavior is present but reduced compared with that of the species with male abdominal coloration (Sceloporus undulatus consobrinus). In comparison, the mean rates of other behaviors that do not display abdominal skin (push-up, head-bob) did not differ between these species. We also found that S. virgatus males were more likely to show a neutral response following the first fullshow during the 60-min trial. While, S. u. consobrinus males were more likely to respond to the first fullshow with an aggressive response. Thus, in this case, the color signal and the behavioral signal are evolutionarily decoupled because in S. virgatus the loss of the color signal is not coincident with the loss of the behavioral signal.
Animal communication systems are complex interactions of senders and receivers emitting signals through the environment (Bradbury & Vehrencamp 1998). These systems are critical in both heterospecific and conspecific interactions and in inter- and intrasexual communication. Interactions where communication systems have been studied are the inter- and intraspecific signals involved in courtship and reproduction. In these interaction, animals must correctly determine species, sex, and reproductive condition of other organisms in their environment (Stratton & Uetz 1981, 1983; Endler 1992; Uetz et al. 2009). When a pool of potential mates is identified individuals have the opportunity to send signals that may influence the choice of potential mates. Considerable research has examined these issues using single sensory modalities including pheromonal (Houck et al. 1998, 2007), seismic (De Luca & Morris 1998), acoustic (Lindström & Lugli 2000; Zuk et al. 2008), and visual signals (Zuk et al. 1990). Visual signals are also classified as color (LeBas & Marshall 2000), morphological (Persons & Uetz 2005), or behavioral signals (Arbutnott & Crespi 2009). Research using sensory specialists (individuals using a one signaling modality) has generated interesting hypotheses to explain the evolution of signals important in courtship and mating (Endler 1992).
It has become clear, however, that communication systems are more complex than what is represented by one signal. Multicomponent signals are a category of complex signals comprised of several cues within the same signaling modality (Candolin 2003). One could argue that color and behavior are not within the same signaling modality because color and vision may be processed in different parts of the brain. However, color and movement are detected using the same sensory structures and therefore one could argue that these cues are in the same signaling modality. This issue requires additional study and is not the main focus of this research. Therefore, we will consider this as a multicomponent signal.
Cues are traits that are assessed by the receiver and influence the receiver in some manner (Candolin 2003). Cues can evolve and be maintained in a context other than the communication system. Signals, on the other hand, are cues that have been modified for communication systems or are only important in communication systems (Candolin 2003). We will use signal throughout this paper to refer to the different component of this aggressive display.
In contrast to co-evolving signals, two independent signals can also provide different information to receivers and evolve independently regardless of the signaling modality. The aim of this study is to determine if the abdominal coloration and the behavior are evolving independently or color and the behavior are co-evolving.
Generally, the studies of multicomponent signals have focused on the use of different signals within a single species (for reviews see Candolin 2003; Hebets & Papaj 2005; Partan & Marler 2005). However, closely related species show variation in the presence and absence of different parts of multicomponent signals. A handful of studies have shown that the loss of one cue is not always evolutionarily coupled with the loss of another signal when evolutionary changes to the multicomponent signal exist (Prum 1990; Wiens 2000; Kimball et al. 2001; Yeh et al. 2006). Even less attention (none to our knowledge) has been paid to the response of the receiver when one signal has been lost. Comparing the presence and absence of different signals in closely related species as well as the response to these signals can be useful in understanding the evolution of these diverse and complex signals and how the different components of communication systems are altered when one of the signals is lost.
We tested the hypothesis that two signals (color and behavior) within a multicomponent signaling system are evolutionarily decoupled in two Sceloporus lizards (Wiens 2000). In most species, male Sceloporus lizards have a pair of blue abdominal patches, and these patches vary in presence, size, and hue. A phylogeny, based on molecular and morphological data, has been proposed for Sceloporus lizards (Reeder & Wiens 1996; Wiens & Reeder 1997). The hypothesis that sexual dimorphism in abdominal coloration is the ancestral character state in Sceloporus is strongly supported by parsimony analysis (Wiens 1999). That analysis also showed that males have lost this trait in several species and females have gained it in other species. Thus, this genus has three character states: monomorphic white (derived, male loss), dimorphic (ancestral), and monomorphic blue (derived, female gain).
We compared the rate and duration of fullshow behavior in Sceloporus virgatus (males have lost the blue abdominal coloration) and Sceloporus undulatus consobrinus (male show the ancestral condition of male expression of blue abdominal coloration). If the behavioral signal (fullshow) and the color signal (blue abdominal coloration) are evolutionarily coupled than the loss of blue abdominal coloration in S. virgatus should coincide with the loss of fullshow; a highly aggressive behavior that reveals (when present) the location of abdominal coloration. If no such relationship is detected than the loss of abdominal coloration is evolutionarily decoupled from the loss of the behavioral signal. We compared the rate and duration of fullshows, push-ups, and head-bobs between pairs of male S. u. consobrinus (with abdominal blue) and pairs of male S. virgatus (without abdominal blue). We also compared the behavior of the receiver from both species after the first fullshow to determine if the receiver’s response is present or absent to determine if the receiver’s behavior is also evolutionarily decoupled from the signaler’s phenotype (Quinn & Hews 2000). We are examining the behavior of the receiver after the first fullshow because individual behaviors may not be statistically independent of each other and therefore cannot be analyzed using traditional parametric statistics (Martin & Bateson 1993). Using the first behavior as the measure of interest will ensure that the behavior measured is in response to the signals of interest rather than to future behavioral interactions between signaler and receiver (Quinn & Hews 2000).
In May and June of 1999 and 2000, we collected adult male (n = 28) and female (n = 14) S. virgatus (>50 mm snout-vent length, SVL; Vinegar 1975b) in the Chiricahua Mountains of southeastern Arizona near the Southwestern Research Station of the American Museum of Natural History. Collections were made within generalized ranges of S. u. consobrinus (Lemos-Espinal et al. 1999; Smith et al. 1999; Leache & Reeder 2002). Adult male (n = 36) and female (n = 18) S. u. consobrinus (>52 mm SVL males and females;Vinegar 1975a) was collected along New Mexico State Road 533 in June and July of 1999 and 2000. Sceloporus virgatus was captured the day before use in a trial, maintained in individual cloth bags, and released the next day. Sceloporus u. consobrinus was less abundant and more difficult to capture and was held for 1–10 days prior to use in a trial. Females were held in large outdoor enclosures (120 × 108 × 33 cm) in groups of five and males were held individually in 10-gallon aquaria with dirt and bark substrate. Tanks were kept on a screened porch such that air temperature and light conditions followed daily fluctuations. A 40-W lamp illuminated one end of the tank. Lizards received water and crickets ad libitum. Individuals were returned to their site of capture when in captivity for <48 h. Individuals in captivity for >48 h were returned to Indiana State University for use in other studies.
From 18 May to 11 June 1999, we conducted 14 trials using S. virgatus in large outdoor arenas (6 × 10 m) between 08:00 and 12:00 h (period of high activity for S. virgatus; Rose 1981). From 27 June to 8 July 2000, we conducted 18 trials using S. u. consobrinus between 08:00 and 12:00 h. We paired males by matching individuals with similar SVL (within 2 mm), tail status (broken or intact), and mass (within 1 g). Females were included in these trials to enhance the likelihood of male–male interaction (c.f. Thompson & Moore 1991; Quinn & Hews 2000). Fifteen minutes prior to the start of each trial, lizards were cooled to approx. 15°C to reduce stress of handling and placed on a board (20 × 60 cm) in the arena with the female between the males (Quinn & Hews 2000). For identification, each male was individually marked with a small dot of color paint on the dorsum (Quinn et al. 2001). Individuals did not require a period of time to acclimate to these large enclosures (Quinn & Hews 2000).
The behavior of both of the males was viewed and recorded by V. S. Quinn and a technician for 1-h through a one-way glass window to reduce observer effects. All lizard behavior from both was recorded into a microcassette recorder and transcribed at a later date. Although we recorded the full behavioral repertoire during the 1-h observation period we have reported the results of three behaviors – head-bobs, push-ups, and fullshows. Rates and duration of other behaviors are reported elsewhere (Quinn 2001). Behaviors analyzed included:
1 Head-bobs: The head is raised and lowered rapidly. Head-bobs have been observed in several contexts for Sceloporus lizards including courtship (Martins 1993) and general greeting (Quinn & Hews 2000). In other species, head-bobs have been observed in antipredator (Leal 1999) and appeasement (Decourcy & Jenssen 1994) contexts.
We were also interested in the behavioral responses of the receivers to the behavioral signal of this multicomponent signal. Because an individual’s behavior can be influence by its social environment (Bradbury & Vehrencamp 1998) we chose the first individual to perform a fullshow as the signaler and the other individual as the receiver. We analyzed the first behavior of the receiver immediately following the signaler’s first fullshow and compared this response between the two species using a chi-square test. We categorized the response to the opponent’s first fullshow using the following scheme from Quinn & Hews (2000): (i) aggressive (includes moving towards another individual, performing a fullshow, and biting the other individual); (ii) submissive (includes moving away from the other individual); or (iii) neutral (includes nose tapping or licking the substrate and head-bobs). Moves away and retreats involve movement away from the signaler. Nose tapping and licking the substrate are chemoreception behaviors which occur in both social and non-social contexts (Carpenter 1978).
Male S. u. consobrinus spent more time in fullshow than S. virgatus and this difference was significant (t29 =2.146; p = 0.010; Fig. 1). Rates of fullshow also differed between S. u. consobrinus and S. virgatus (t21 =3.60, p = 0.003), with males of the blue-bellied S. u. consobrinus displaying significantly more fullshows than males of the white-bellied S. virgatus (Fig. 2). By contrast, there were no significant species differences in the rates of the two behaviors that do not display abdominal coloration, push-up (t21 =2.131, p = 0.482), and head-bob (t21 =2.08, p = 0.569; Fig. 2).
In addition to these overall differences in the duration and rates of key display behaviors, the species differed in how the receiver responded to the first fullshow of an opponent. There was a significant association between species and the response of the receiver immediately following the first fullshow by the opponent (χ2=6.897, p = 0.038; Fig. 3). In response to the opponent’s first fullshow, S. virgatus males were more likely to respond with neutral behaviors (i.e., chemoreception behavior) than with aggressive (fullshow or biting) or submissive behaviors (retreating). In contrast, male S. u. consobrinus receivers were more likely to respond with aggressive behavior, than with neutral or submissive behavior, following the first fullshow by their opponent.
We tested the hypothesis that different signals of a multicomponent signal in two species of Sceloporus lizards are evolutionarily decoupled by examining the color signals (blue abdominal coloration) and the behavioral signal (fullshow) that displays the color signal. We compared the rate and duration of fullshow behavior in two closely related species of Sceloporus lizards differing in the presence of the color signal. We also compared the receiver’s response to determine if the receiver’s behavior has remained coupled or decoupled with the presence or absence of the multicomponent signal. Our results show that in the species that lost blue abdominal coloration (S. virgatus) fullshow behavior is present, but is reduced compared with that of the ancestral species (S. u. consobrinus). Thus, in this case, the color signal and the behavioral signal are evolutionarily decoupled because in S. virgatus the loss of the color signal is not coincident with the loss of the behavioral signal.
We were also interested in the receiver’s response to this multicomponent signal. The initial response of male S. u. consobrinus to the first fullshow by their opponent was more likely to be aggressive, such as fullshow or biting, compared with neutral or submissive responses. Conversely, the initial response of male S. virgatus to the first fullshow was more likely to be a neutral behavior such as chemoreceptive behavior or head-bob. Thus, the color signal and the response to the behavioral signal are evolutionarily coupled. The loss of blue abdominal coloration in S. virgatus is coincident with the loss of the response to the behavioral signal. Similar results were shown when male S. virgatus was presented with conspecifics with blue-painted, white-painted, and a spotted abdomen. Males presented with either white-painted or spotted males responded with neutral behavior suggesting that the loss of the signal was coincident with the reduction in the response to the signal but not its loss. When males were presented with blue-painted conspecifics they responded to the multicomponent signal with submissive behavior. These results combined with the results presented herein suggest that the response of the receiver is dependent upon the presence of the color signal (blue abdominal coloration). When S. virgatus males detect both signals in this multicomponent signal the response is submissive in nature. When S. virgatus males detect only the behavioral signal the response is neutral (Quinn & Hews 2000). This adds to the growing body of literature that shows that the signals of multicomponent signals, in some cases, provide different types of information to receivers (Candolin 2003).
The information provided by these different signals may be informative or non-informative (Candolin 2003). Under this system, the behavioral signal – fullshow –could be viewed as an informative signal providing the receiver information about the sender’s species identity because different Sceloporus species perform unique fullshow behaviors that vary in several components (Carpenter 1978; Martins 1991, 1993; Jenssen & Nunex 1998; Wiens 2000). Fullshow behavior could also be an informative signal providing information on the quality of the signaler. Although the duration and rate of fullshow are less in S. virgatus (the species that has lost the morphological color signal) it has not been eliminated from the behavioral repertoire. Fullshow occurs toward the end of an escalated sequence of behaviors terminating in biting (Carpenter 1978; Martins 1991, 1993; Jenssen & Nunex 1998). The threat display in Uta stansburiana lizards, which is similar and probably homologous to the Sceloporus fullshow, is an honest signal of endurance (Brandt 2003); lizards with longer duration threat displays had greater physiological endurance. This leads to the speculation that physiological traits important in male–male competition in S. u. consobrinus and S. virgatus differ. In this case, the behavioral signal of this multicomponent signal is potentially a condition-dependent signal if males in better condition can perform fullshow behavior at a higher rate and for a longer duration.
It is less clear if the blue abdominal color signal is an informative or uninformative signal. The blue color signal in phrynosomatid lizards shows great variation in its expression within the Sceloporus lizards (Wiens 1999) and therefore could be an informative signal of species identity. It is produced when the dermal chromatophore unit has both a layer of iridophores (which reflect blue light into the environment) and a layer of melanin (which absorbs other wavelengths of light; Quinn & Hews 2003). Iridophores are pigment cells with stacked plates of guanine acting to reflect light when properly oriented (Bagnara et al. 1968). Blue coloration, unlike carotenoid-dependent signals, is not condition dependent (Hill & Montgomerie 1994; Linville & Breitwisch 1997; Olson & Owens 1998; Grether 2000; McGraw & Hill 2000) and may not provide an honest signal of condition and thus may be an uninformative signal. Further research into understanding the information content of the blue abdominal coloration may shed light onto the status of this signal.
Our results support the idea that senders and receivers of signals are under different selection pressures and face different costs and benefits associated with changes in signals in communication systems. Perhaps the costs for the receiver associated with ignoring aggressive visual signals, such as fullshow and blue abdominal coloration, are great and selection has favored the maintenance of the response by the receiver to aggressive visual signals. On the other hand, the costs for senders that are associated with performing fullshow when the color signal is absent could be high and therefore selection pressure may maintain the behavioral signal in this system.
It has been suggested that the evolution of different signaling modalities should be evolutionarily coupled with behavioral signals (Geist 1966; Prum 1990; Endler 1992; Marchetti 1993; McLennan 1996). Phylogenetic analyses and behavioral studies have examined multicomponent signals (Prum 1990; Wiens 2000; Kimball et al. 2001; Yeh et al. 2006). Because of the small number of studies it is difficult to detect patterns at this time. Phylogenetic results (Wiens 2000) and the experimental results reported in this paper show that the evolutionary loss of the color signal is decoupled from the behavioral signal in Sceloporus lizards. Prum (1990) showed that the syringeal bone in manikins and behavioral displays are decoupled when one of the signals are lost. Conversely, in peacock-pheasant group (Galliformes: Polyplectron spp.) the loss of ocelli (ornamental eye-spots) and display behaviors are coupled (Kimball et al. 2001). And, in two species of Drosophila, the presence of the morphological trait (wing spot size) and the behavioral trait (courtship wing display) are evolutionarily coupled (Yeh et al. 2006).
The study of animal communication systems using complex signals examined in a phylogenetic context is important to have a more complete understanding of animal communication. We suggest that socially relevant and complex behavioral and morphological signals should be considered in future phylogentic studies to detect patterns in the evolutionary coupling or decoupling of signals used social signals. Such patterns may assist in the understanding of the factors affecting the evolution of communication systems.
We thank C.J. Amlaner, E. Fernandez-Juricic, B.M. Graves, S.L. Lima, W.J. Mattson, E. Tuttle, M. Wikelski, P.A. Zollner, and anonymous reviewers for comments on earlier versions of this manuscript. The volunteers and staff of the Southwestern Research Station of the American Museum of Natural History provided assistance with fieldwork. The National Science Foundation Dissertation Improvement Grant (IBN-0073203), Indiana Academy of Science, Indiana State University, Southwestern Research Station Graduate Student Fund, and the American Museum of Natural History, Theodore Roosevelt Fund (to VSQ) and National Institute of Health (to DKH: IBN-9629783), funded portions of this research.