|Home | About | Journals | Submit | Contact Us | Français|
Previous experiments suggest that males spend more time with the more receptive of 2 novel females or the one with the higher fitness potential. However, males often court individual females repeatedly over a season; for example, male lizards sequentially visit familiar females as they patrol territorial boundaries. It may benefit males to vary display intensity as they move between multiple females. In this study, we explored the factors influencing amount of male courtship to familiar females in the sagebrush lizard, Sceloporus graciosus. We tested whether males vary the amount of courtship exhibited due to individual differences among males, female reproductive state, or female fitness potential. Each male was allowed to interact separately, but repeatedly, with 2 females until both females laid eggs. Male courtship behavior with each of the 2 females was assayed at an intermediate point, after 3 weeks of interaction. We found that individual differences among males were considerable. The number of male courtship displays was also positively correlated with female latency to lay eggs, with males displaying more often toward females with eggs that had not yet been fertilized. Courtship behavior was not well predicted by the number of eggs laid or by female width, both measures of female quality. Thus, male S. graciosus appear to alter courtship intensity more in response to signals of female reproductive state than in response to variation in potential female fitness.
Courtship is costly; it requires considerable time and energy (e.g., Shine and Mason 2005; Sozou and Seymour 2005) and exposes males to higher predation risk (e.g., Daly 1978; Pruden and Uetz 2004). Nevertheless, male animals of many species spend quite a bit of time in courtship, often visiting the same females repeatedly as they move about their territories (e.g., in lizards: Davis and Ford 1983; in fish: Norman and Jones 1984). This high cost of repeated courtship may be a requisite component of reproductive behavior. For example, female primates prefer the familiar odors of males that have previously marked in their territory (Fisher et al. 2003). Males may also be able to minimize costs by varying the amount of courtship depending on the female's reproductive state. Sometimes, only females in the appropriate physiological condition will respond to male courtship advances (e.g., in fish: Clement et al. 2004; in birds: Maney et al. 2006). Repeated courtship may also accelerate mating. For example, female serins, Serinus serinus, exposed to more male song increased the time spent building the nest (Mota and Depraz 2004). Female sagebrush lizards perform less rejection behavior toward displaying males when previously exposed to more courtship bouts (Kelso and Martins 2008). In this study, we extend recent findings on the factors influencing amount of male courtship to the context of repeated interactions with familiar females and test whether male Sceloporus graciosus lizards vary courtship due to 1) individual differences, 2) to match female reproductive state, or 3) to match female quality.
Most morphological and behavioral traits show individual differences, perhaps as a result of underlying genetic (Diatchenko et al. 2005; Tang et al. 2007) or hormonal variation (Enstrom et al. 1997; Cawthorn et al. 1998). Some of these differences can be used for individual recognition, showing small within-individual variation but larger between-individual variation (e.g., in lizards: Brandt and Allen 2004; in bats: Russ and Racey 2007). Different individuals may also have distinct ranges of behavior, such as some may tend to display much more often than others in a range of situations (Sih et al. 2004). Individual differences are necessary for selection to have an effect; this variation is essential for evolution to occur. Individual differences may also limit the degree to which males can respond differentially to the condition or fitness quality of multiple female partners.
Males assess whether unfamiliar females are receptive through visual (LeBas and Marshall 2000; Barelli et al. 2007), chemical (Cooper and Perez-Mellado 2002; Carazo et al. 2004; Ferkin et al. 2004; Head et al. 2005), and other cues (Komers et al. 1994; Semple and McComb 2000; Murai et al. 2002). In many species, males prefer novel females to those they may have already had the chance to inseminate (Tokarz 2006). For example, male anoles, Anolis carolinensis, increase their display rate when exposed to novel females both in the laboratory and in the field (Orrell and Jenssen 2002). Males attend to signals of reproductive condition in novel individuals and adjust their courtship tactics accordingly (Kelso and Verrell 2002). For example, male displays increase toward females with reproductive coloration in collard lizards (Baird 2004) and in budgies (Eda-Fujiwara et al. 2003). However, these studies tested changes in male behavior when presented with novel females, and in nature, male lizards more naturally encounter multiple familiar females sequentially (e.g., Haenel et al. 2003). A male would benefit by courting females that are reproductively responsive and that have not yet been mated in order to provide him with higher possible reproductive fitness.
Alternatively, recent studies have found that males prefer females of higher quality, likely to have higher evolutionary fitness (e.g., Torres and Velando 2004). Males assess the quality of unfamiliar females and prefer to mate with females who possess certain characteristics, such as those that are larger (e.g., Whiting and Bateman 1999) or more colorful (e.g., Amundsen and Forsgren 2001). These morphological characteristics may convey information about the potential fecundity or fitness quality of a potential mate. Males may thus not vary their courtship to match reproductive state of their female partners if those partners vary also in fitness. Instead, courtship amount may depend only on more static qualities that correspond to female fitness.
Sceloporus graciosus is a small, territorial lizard in which male and female territories overlap (Martins 1991); thus, polygamous males may be continuously exposed to familiar females. Females consistently exhibit rejection displays toward courting males, and they may require frequent exhibition of courtship displays before they will allow male insemination (Kelso and Martins 2008). Courtship consists of headbob and shudder displays, which require considerable energy expenditure to perform (Brandt 2003). Male S. graciosus exhibit preferences for certain females, for example, those that are from their own population (Bissell and Martins 2006). However, the way in which males alter courtship to respond to differences among females is not known.
In this study, we consider 3 possible predictors of male courtship behavior in S. graciosus: 1) individual variation among males, 2) female reproductive state, and 3) female fitness. We hypothesize that 1) individual males differ in the amount of courtship they exhibit; 2) males respond to the reproductive state of familiar individuals and exhibit more courtship displays toward females that are not yet gravid; and 3) males respond to differences in female morphology and exhibit more courtship toward females with higher fitness, as measured by the number of eggs laid.
We collected S. graciosus from Southern California (42 females, 21 males) in early May 2006. Experiments began later that month in our laboratory at Indiana University, Bloomington. All lizards were housed individually in 21-L aquaria (41 × 21 × 26 cm), with sand substrate, basking rock, heat lamp, and water dish. Lizards were fed crickets or mealworms every other day.
Female S. graciosus start depositing yolk in their eggs as early as late April to early May (Goldberg 1975), as they emerge from hibernation. Between June and July, oviductal eggs are present and laying begins during this time (Jameson 1974; Goldberg 1975). Although females are fairly synchronous in their cycles (Goldberg 1975), larger females tend to ovulate earlier and sometimes are able to produce a second clutch (Jameson 1974). Females usually lay at least 3 eggs and hatchlings emerge after mid-August (Jameson 1974).
We placed each male into the home aquarium of one female, allowing him to interact freely with her for 2 days. After 2 days, we moved the male to the home tank of a second female (of similar size to the first) and continued alternating the male between this pair of females every 2 days for up to 3 weeks after both females had laid eggs. We paired each male with 2 females to get a second measure of male behavior that could be used in testing for individual differences among males. We checked the aquaria daily for eggs, recording the number of eggs and date laid. If a single female laid eggs on more than one day (never more than 12 consecutive days), we counted the total number of eggs across days and recorded laying date as the first date in which we observed eggs.
At the beginning of the experiment, we measured snout-to-vent length (SVL) and total length (TL) for each lizard. We then held each lizard against a Plexiglas sheet, flattening the ventral surface against the clear surface and photographing with a Nikon coolpix 7600 digital camera. We repeated the photographs on day 17, when most females were substantially gravid. We used ImageJ (Meyers et al. 2006) to measure the width of each lizard from the photographs at 3 transverse sections across the abdomen (anterior, middle, and posterior). Sceloporus graciosus are sexually dimorphic in coloration, with males having much brighter blue coloration than females, although light patches are also present in females. We measured the total area of the 2 ventral blue patches and the total area of the gular (throat) patch of both male and female lizards as well as the width of the white band between the abdominal blue patches. For females, we also measured the total area of the 2 abdominal orange patches, which increases as eggs grow within the female's abdomen (Zucker and Boecklen 1990).
Three weeks after the first pairing of a male with one of the females (about halfway through the experiment), we introduced each male and one of his female partners into an empty and novel 21-L aquarium to measure courtship behavior. We recorded the subsequent interaction for 40 min using a Panasonic AG-188 VHS camcorder. Male S. graciosus direct headbob and shudder displays during courtship (Martins 1993; Kelso and Martins 2008). During the assay, we thus counted the total number of headbob and shudder displays performed by the male. Headbob displays are stereotyped series of stereotyped single and double headbobs, whereas shudder displays are series of short, rapid headbobs (Martins 1993, 1994). In both cases, we counted full displays rather than individual headbobs. Although female lizards sometimes exhibit rejection behavior that may offer a cue to their reproductive condition (e.g., sidlehopping, Kelso and Martins 2008), only a few of the females in the current study exhibited any measurable behavior. We thus scored, but did not include, female behavior in any of the statistical tests. We did not observe any copulation attempts during this study. We tested each male again 2 days later with the second female for a similar courtship assay.
We used separate hypothesis tests to determine whether male courtship rate (number of displays per minute) was well explained by 1) individual differences among males, 2) female reproductive state, and 3) morphological or other static features of the female. We used multiple regression to test these hypotheses, focusing on the behavioral assay as the unit of analysis and including a factor for male identity because each male was measured twice (once with each of 2 females) much as would be done in a paired t-test or repeated-measures analysis of variance (ANOVA). We also calculated Pearson product–moment correlations to measure effect size, as needed. All analyses were conducted using the glm module of SAS (2002) and Type III sums of squares. For each model, we examined residual plots to confirm that data conformed to the usual ANOVA and regression assumptions of normality and homoscedasticity without statistical transformation.
We began by testing the importance of male identity as a predictor of male courtship rate. We then tested for correlates between male morphological features and courtship rate that might lend further insight into individual differences among males. To do this, we used principal components analysis (PCA) with Varimax rotation to combine 14 measures of male morphology into a smaller number of composite variables. Specifically, we combined measures of SVL and TL with 6 morphological measures (3 of body width and 3 of color patch area) from each of 2 photographs (before and 2.5 weeks into paired encounters). We limited our attention to PCA axes with eigenvalues greater than 1.0 and then used multiple regression to test the predictive ability of these morphological PCA axes to explain variation in male courtship rate.
To test the importance of female reproductive state in explaining the amount of male courtship, we focused on aspects of female morphology that change with reproduction. As eggs grow within a female, she becomes wider and develops a bright orange color that slowly spreads in area across her entire abdomen. We first used a PCA (with Varimax rotation) to combine the 3 measures of female width (anterior, middle, and posterior) into a single composite measure of width. We then used separate multiple regressions to test the relative importance of female width (PCA factor), area of abdominal orange, and first laying date on male display rate.
Finally, to test the importance of female quality on male courtship, we concentrated on more general morphological and fitness measures. First, we combined 15 female morphological measures (the same 14 measures as for males above plus the area of abdominal orange) into a smaller number of composite predictor variables and used multiple regression to test the ability of these composite factors to explain male courtship rate. Second, we tested whether we could predict male courtship by the number of eggs laid by an individual female. Note that our results for this hypothesis may not be directly comparable to those of mate choice studies asking similar questions because we did not conduct simultaneous choice tests in which males could compare and show a preference between 2 novel females. In our study, we mimicked the field condition for this species closely, never allowing the male to interact with more than one female at a time and testing males only after they had become familiar with the 2 test females over 3 weeks of repeated interactions.
Males produced an average of 21 (standard error [SE] = 4.0) displays per hour toward females in courtship tests. Male display rate varied considerably, with one male displaying as much as 65 times per 40-min assay and others not displaying at all (Figure 1). Although males also responded differently toward each of their 2 females, directing as many as 3 times as many displays toward one female as to the other (Figure 1), male identity was a statistically significant predictor of courtship rate (F20,20 = 3.28, P < 0.01).
The 14 male morphological characters were well described by 5 composite factors that together explained 85% of the variation (Table 1). The first PCA axis combined the 3 measures of abdominal width and the single measure of ventral blue from photographs taken in the middle of the male–female pairing procedure (second sampling period), all of which are positively associated with each other. The second axis combined the width of the white strip separating the left and right ventral blue patches (from photographs at both sampling periods) with SVL and the area of the ventral blue patch (from the first sampling period). SVL and the area of the blue patch were negatively associated with the width of the white strip. Factor 3 combined the 3 measures of abdominal width taken from photographs in the first sampling period. Factor 4 summarized variation in the total area of the gular patch in the 2 sampling periods, and Factor 5 consisted almost entirely of TL. None of these male morphological factors significantly explained courtship rate, whether or not male identity was also included in the statistical model (Table 1).
Female laying date explained a significant proportion of the variation in male display rate, even after male identity had been included in the model (female laying date: F1,7 = 8.24, P < 0.03; male identity: F18,7 = 9.09, P < 0.01). Considering only females that laid eggs, males displayed more toward females that laid eggs later in the season (Figure 2; r = 0.41, degrees of freedom [df] = 26, P < 0.05). The reduced df in these analyses reflect the fact that only 27 of the 42 females laid eggs during the course of this experiment. Females who laid eggs produced an average of 3.3 (SE = 0.2) eggs, beginning to lay approximately 47 days (SE = 1.7, range = 33–70) after the beginning of the repeated encounters between males and females. The first eggs were thus laid 2 weeks after behavioral tests were conducted.
Males displayed less often toward females exhibiting more ventral orange coloration (Figure 3; r = −0.39, df = 24, P < 0.05). Female ventral orange coloration was a significant predictor of male display rate (F1,26 = 4.78, P < 0.04). Note that although the df may be somewhat inflated in this analysis because 7 of the males were measured twice, the relationship is still significant (P < 0.05) if we use only 1, 19 df (comparable to averaging the 2 data points for each male). Although an optimal analysis would include male identity in a repeated-measures model, male identity was related to female ventral coloration, confounding the interpretation of these predictor variables. Although either factor was a significant predictor alone, neither was significant when included together.
Males did not display more toward wider females (F1,38 = 0.05, P = 0.82). Female width explained less than 1% of the variation in male display rate, a result that was also true when male identity was included in the model (female width: F1,18 = 0.34, P = 0.56; male identity: F20,18 = 3.04, P = 0.01).
Females performed little measurable behavior during the behavioral assays. For example, only 11 of the 42 females engaged in any sidlehopping, and no copulations were observed. Thus, we did not test female behavior as a statistical predictor of male courtship behavior.
Female morphology did not predict male courtship rate. The 15 female morphological characters were well described by 5 composite factors, explaining 84% of the variation (Table 2), but the details of these axes were quite different from that found in the PCA for males. The first PCA axis for females combined the area of the ventral blue patch and the width of the ventral white (from photographs at both sampling periods). The width of the ventral blue patch was negatively associated with the width of the white strip. The second axis combined the 3 measures of abdominal width from the second sampling period with abdominal orange. Factor 3 combined the 3 measures of abdominal width taken from photographs in the first sampling period with SVL. Factor 4 summarized variation in the total area of the gular patch in the 2 sampling periods, and Factor 5 consisted almost entirely of TL. None of the combined female morphological factors significantly explained male courtship rate (Table 2).
Number of eggs was also not a significant predictor of male display rate whether including all females (F1,19 = 1.27, P = 0.2729; male identity: F20,19 = 3.07, P < 0.01) or considering only those females that laid eggs (F1,7 = 0.12, P = 0.7409; male identity: F18,7 = 3.49, P < 0.05).
Our results show that although male sagebrush lizards show considerable individual differences in courtship behavior, they also tailor their courtship behavior to reflect the current reproductive condition of familiar female partners. Using our unique experimental design (repeated interactions with single females rather than a simultaneous choice test), we found no evidence that males vary courtship to match differences in female fitness or other measures of quality.
Individual differences have been increasingly recognized as an important feature of behavior. Although morphological characteristics did not predict the amount of courtship exhibited by males in this study, individual differences may reflect underlying genetic or hormonal differences that influence male fitness. For example, wild tits show consistent individual variation in exploratory behavior, and this behavior is heritable (Dingemanse et al. 2002). Individual differences may also reflect relationships with other traits in forming a behavioral syndrome or “personality” (as reviewed by Bell 2007; e.g., Kralj-Fiser et al. 2007; Stamps 2007). Male Japanese quail repeatedly vary in the quantity and quality of courtship exhibited toward a lure or live female (Lumineau et al. 2005). Similarly, a previous study found that sagebrush lizards in the field show consistent individual differences in display frequency (Martins 1991). Here, we confirm that result and find also that individual differences in display frequency are comparable in magnitude to the variation exhibited by individual males interacting with familiar females at different stages of reproductive condition.
Male S. graciosus in our study displayed more intensely toward females that were less far along in the reproductive cycle—laying eggs later in the season and showing less orange coloration at the time of behavioral tests. This result concurs, for example, with those in Ojanguren and Magurran (2004) who found that guppy males court nonpregnant females more than they do pregnant females and that size did not predict the amount of courtship. Male wolf spiders also direct more courtship toward unmated females, determining the reproductive status of female conspecifics through chemical cues in spider silk (Roberts and Uetz 2005). Females of many species signal reproductive condition with color changes (Weiss 2002; Setchell et al. 2006), and others have found that males respond to coloration cues when making decisions on who and how much to court (e.g., Amundsen and Forsgren 2001). In lizards, males are known to decrease courtship toward females with more intense red coloration (Watkins 1997) and to be able to detect reproductive state through coloration in unfamiliar females (LeBas and Marshall 2000; Weiss 2002). Here, we show that males fine-tune their behavior to match female reproductive state, even when interacting with familiar females over several weeks. As in Bissell (2001), we found no evidence that males attend to female width as a sign of female gravidity.
Unlike other recent studies, we found no evidence of male preferences for females with particular morphological characteristics. For example, in striped plateau lizards, males spend more time with orange females and female orange coloration is correlated with egg mass (Weiss 2006). In the current study, however, males did not display more intensely toward females that laid more eggs or to females with particular morphological features. As with any preference study, we may not have found significant effects because of inadvertently not measuring the most salient aspects of female morphology (for discussion, e.g., see Hamilton and Sullivan 2005). The difference in results is also likely to be due, in part, to our use of sequential rather than simultaneous choice tests. As shown in Bissell and Martins (2006), preference patterns can differ markedly when lizards are offered a sequential as opposed to a simultaneous choice. Finally, our experimental design differs because we tested males interacting with familiar females, after 3 weeks of repeated interactions between the pair. Additional studies are needed to confirm that repeated, but sequential, interactions more accurately represent the situation encountered by male sagebrush lizards in the wild.
Our study suggests that instead of choosing among females, Sceloporus males minimize the cost of courtship by interacting frequently with each female and tailoring courtship effort to match her condition. Future studies are needed to determine whether males increase and decrease courtship in a complex interaction or whether courtship drops precipitously once a female is inseminated.
Summer Research Experience for Undergraduates in Animal Behavior, Indiana University's Center for the Integrative Study of Animal Behavior, National Science Foundation (0453403 to E.D.); Common Themes in Reproductive Diversity Training Grant, awarded to Indiana University Bloomington, National Institutes of Health (to M.R.).
We thank Erin Kelso, Saúl Nava, and Stephanie Dowdy-Nava for assistance in capturing the Sceloporus graciosus used in this study. We thank Arián Avalos, Winnie Ho, and Deanna Soper for assistance in laboratory setup and maintenance. We thank Greg Demas, Ellen Ketterson, and Mike Wade for comments on early drafts. The research reported in this article conformed to all laws of the United States and was approved through Bloomington IACUC # 06-015. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.