Population frequencies of alternative male phenotypes in tree lizards: geographic variation and common-garden rearing studies

Transcription

1 Behav Ecol Sociobiol (1997) 41: 371±380 Ó Springer-Verlag 1997 Diana K. Hews á Christopher W. Thompson Ignacio T. Moore á Michael C. Moore Population frequencies of alternative male phenotypes in tree lizards: geographic variation and common-garden rearing studies Received: 17 January 1997 / Accepted after revision: 30 August 1997 Abstract Tree lizards (Urosaurus ornatus) vary in throat fan (dewlap) color. Earlier, we described ve dewlap types (Orange, Orange-Blue, Yellow, Yellow-Blue, and Blue), and reported that only males had blue in the dewlap and that presence or absence of a discrete blue patch was correlated with male alternative reproductive phenotypes in a central Arizona population. Here, with a modi ed scheme characterizing two dewlap elements, background color (orange, yellow, blue) and blue patch occurrence, we assessed: (1) sexual, annual, and geographic variation in the frequencies of dewlap elements; (2) simple habitat correlates; and (3) the e ects of laboratory rearing regime on dewlap type. Within a population, frequencies of males and females expressing orange or yellow backgrounds did not di er, suggesting that control of background is similar in the sexes. Within several populations, frequencies of the dewlap elements did not di er across years (and probably generations), indicating that phenotype frequencies are relatively stable. Among ve populations frequencies of background colors varied, as did frequencies of male types (blue patch present or absent). Dewlap frequencies did not correlate with habitat (boulders or mesquite trees), although few populations were sampled. In male and female o spring reared from eggs to sexual maturity in a common-garden laboratory study, background color D.K. Hews (&) Department of Life Sciences, Indiana State University, Terre Haute, IN 47809, USA Fax: (812) ; LSHEWS@scifac.indstate.edu C.W. Thompson Burke Museum, DB-10, University of Washington, Seattle, WA 98195, USA I.T. Moore Department of Zoology, Oregon State University, Corvallis, OR 97331, USA M.C. Moore Department of Zoology, Arizona State University, Tempe, AZ , USA frequencies in both sexes and blue patch frequencies in males di ered among o spring from di erent populations. O spring frequencies matched respective parental population frequencies. Results suggest that amongpopulation variation in frequencies of the two dewlap elements are mediated by di erences in genetics, in maternal e ects, or both. Thus, di erences in male behavior functionally linked to the blue patch also may be controlled by genetic or maternal e ects. Key words Alternative male reproductive tactics á Color polymorphism á Urosaurus ornatus á Geographic variation á Common-garden analysis Introduction Species may exhibit relatively discrete, within-sex variation in reproductive behavior and traits associated with mating. Typically males are the variable sex and examples of such alternative reproductive tactics include: territorial versus nomadic male lizards (Thompson and Moore 1991b, 1992), courting versus noncourting male swordtail sh (Ryan and Causey 1989; Zimmerer and Kallman 1989) and calling versus silent toads (Howard 1984; Arak 1988). Male alternatives are receiving increasing attention from evolutionary biologists (reviews in Rubenstein 1980; Austad 1984; Gross 1984, 1996; Lott 1991). Relatively little is known about variation in frequencies of male alternatives and about the proximate developmental basis of male alternatives, the two topics of this study. The extent of geographic and annual variation in frequencies of male alternatives is not well studied (Travis 1994; Carroll and Corneli 1995; Sinervo and Lively 1996). To address evolutionary hypotheses about male alternatives researchers need to assess the spatial and temporal variation within and among populations in the expression of male alternatives, and to seek environmental correlates of any such variation.

2 372 This was the rst general goal of this study. Our second goal was to examine the development of male alternatives. Using a common-garden design, we reared o spring from two populations that di ered in frequencies of two male types to determine if variation in rearing environment a ected the frequencies at which male alternatives were expressed. Male alternatives could represent conditional strategies in response to environmentally induced variation in traits that a ect mating success (Gross 1996). A special case of environmental e ects possibly mediating male alternatives are non-genetic maternal e ects. For example, in some birds male embryos experience di erences in levels of maternally derived testosterone in the egg yolk, and these levels correlate with levels of aggression in o spring (Schwabl 1993). In contrast, male di erences could arise at least in part from genetic di erences. If so, several evolutionary mechanisms could maintain the genetic variation (e.g., Shuster and Wade 1991; Ryan et al. 1992; Sinervo and Lively 1996). We analyzed variation in throat fan (dewlap) color in the tree lizard, Urosaurus ornatus, a species in which male dewlap colors correlate with alternative reproductive tactics. In central Arizona (Thompson and Moore 1991a) and in some New Mexico populations (Hover 1982, 1985; Zucker and Boecklen 1990), the dewlap can have two color elements. The rst, background color, is expressed in both sexes and can be yellow or orange. The second, a central blue patch (Fig. 1A), can be present, but only in males. We considered these two aspects of dewlap color separately because of the sex di erences in expression and the di erences in hormonal control: blue patch occurrence in adults is altered by manipulating androgens in hatchlings (Hews et al. 1994) but the background hue is not (Hews et al. 1994; Hews and Moore 1995; D. Hews and M. Moore, unpublished work). Male tree lizards display the brightly colored dewlap during interactions with individuals of both sexes, as in many lizards (reviewed in Cooper and Greenberg 1992). Limited behavioral data suggest that the orange or yellow background coloration does not correlate with behavioral di erences (L. Oliver, D. Hews and M. Moore, unpublished work). Much more is known about the functional signi cance of the central blue patch expressed in some males. The patch can function as a status signal (Hover 1985; Thompson and Moore 1991b) and is a component of the male alternatives (Moore 1991; Thompson and Moore 1991a,b, 1992). Field data indicate presence/absence of the blue patch correlates with di erences in patterns of spatial use and aggressive behavior (Hover 1985; Thompson and Moore 1991a,b, 1992; Knapp and Moore 1995; D. Hews and M. Moore, unpublished work), and with di erences in hormonal responses to aggressive encounters (Knapp and Moore 1996). Males with blue dewlap patches are territorial and more aggressive than males lacking the patch, who are either nomads or satellites, foregoing territory defense. In male lizards, territoriality has major consequences for male reproduction (e.g., Stamps 1983; Hews 1990, 1993), and di erences in male tree lizards are thought to represent alternative male reproductive tactics (Moore 1991; Thompson et al. 1993). Further, males without the blue dewlap patch have faster growth rates in snout-vent length (SVL) under laboratory conditions (Thompson et al. 1993; Hews et al. 1994). Laboratory-reared hatchlings (Hews et al. 1994) and free-living adults (Hews and Moore 1994) of the two male types also di er in body mass, for a given SVL. The behavioral and color alternatives in males are xed developmentally: adult males have not been observed to switch between color types (expressing or not expressing the blue patch) or to switch between Fig. 1 A Ventral surface of dewlap of a male expressing the central blue patch (dark area); thin line indicates extent of area typically covered by the background color. B Locations of study sites in relation to major drainages and mountain ranges. Abbreviations as in Table 1, except FM (Fort McDowell Indian Reservation), which is a site surveyed in Thompson and Moore (1991a)

3 373 the behavioral tactics of defending or not defending a territory (Thompson et al. 1993; D. Hews and M. Moore unpublished work; R. Knapp and M. Moore unpublished work). We asked the following questions about variation in this color signal in tree lizards: 1. Within a site, do the frequencies of males and females expressing the background colors di er? 2. Within a site, do the frequencies of background color and of blue patch di er among years? 3. Among sites, do the frequencies of background colors and of the blue patch di er? 4. Do two simple habitat descriptors, boulder elds versus mesquite trees, correlate with the among-site variation of either of the two dewlap elements? 5. Do o spring from populations di ering in frequencies of each of the two dewlap elements still di er in their expression when they are incubated, hatched and reared in a common laboratory environment? Methods Thompson and Moore (1991a) surveyed dewlap frequencies in U. ornatus at two central Arizona sites, Sycamore Creek and Fort McDowell. In this study we re-surveyed frequencies at the Sycamore Creek site and surveyed four new sites in central Arizona. These sites, from west to east (Fig. 1, Table 1), were Sierra Estrella (SE), South Mountain (SM), Fort McDowell (FM), Sycamore Creek (SC), Coon Blu (CB), and Top of the World (TW). All sites were within 120 km or less of each other and exhibited varying degrees of geographic isolation as they were in di erent drainages or in di erent mountain ranges. Elevation of all sites was between 300±400 m except TW, which was near 1400 m. To assess annual variation we surveyed three sites in multiple years: TW in 1990, 1992, 1993, CB in 1990, 1991, 1993, and SC in 1990±1993. We included a site (Fort McDowell, FM) on the map in Fig. 1 to present previously published frequency data (1986±1988 combined, from Thompson and Moore 1991a) for comparisons. Data for FM, however, were not included in any of the among-population statistical analyses. Tree lizards were especially common at sites with mature stands of velvet mesquite trees along riparian corridors and at sites with extensive areas of large boulders. Although the vegetation di ered among sites, we partitioned these sites into two groups, tree versus boulder, based on the primary substrate used by tree lizards (Table 1). We sampled one mesquite site (CB) and four boulder sites (SE, SM, SC, TW). Boulder sites di ered in vegetation. Three of the boulder sites had vegetation characteristic of Arizona Uplands subdivision of Sonoran desert scrub (Brown 1982; palo verde trees, saguaro cacti, bursage and jojoba shrubs). The fourth higher elevation boulder site, TW, had Interior Chaparral vegetation (Brown 1982; scrub oak and manzanita shrubs). We systematically surveyed areas where lizards were often seen, capturing adult males and gravid females by noosing. We determined sex using scale characters that all adult males express. We immediately scored dewlap color at capture, while the lizards were still at their active (operating) body temperature because the appearance of blue dewlap color alters with temperature. Tests of repeatability within and among observers revealed no betweenobserver variation in assignment of males to dewlap types. Except for animals brought into the laboratory, we scored and released all lizards after giving them a diagnostic mark (toe-clip) to avoid resampling. We censused all sites between April and July. At sites we sampled in multiple years, it is likely that we sampled di erent generations because mortality is high (about 20% of animals resident on a study grid survived to the following spring, at both the SC and CB sites; R. Knapp, D. Hews and M. Moore, unpublished work; D. Hews, L. Oliver and M. Moore, unpublished work). Criteria used to assign male dewlap type followed Thompson and Moore (1991a). Brie y, males varied continuously in the size of the dewlap patch from males with none or few blue scales to males with blue patches covering more than 50% of dewlap area. But males whose dewlap surface was more than 10% blue had the blue scales organized into a discrete, centrally located patch, whereas males with 10% or less of the dewlap blue had blue scales scattered about the dewlap (Thompson and Moore 1991a). Thus, we operationally de ned the former as ``blue patch'' type and the latter as ``not blue'' type (see also Thompson and Moore 1991b; Thompson et al. 1993). Some males had solid blue dewlaps without any orange or yellow background color, although in all populations but one (TW) they occurred at very low frequencies (see Results). These males presented problems for the analysis of dewlap frequencies. It was unclear whether they should have been scored as having or not having a central blue patch, and as having or not having a colored background (they could be considered as having no background color and a central blue ``patch'' that was larger than typical). We did two analyses for each comparison, omitting or including solid blue males. When included, solid blue males were scored as having a third background type (blue) and the central blue patch. In all cases, conclusions were not altered by the inclusion or exclusion of males with solid blue dewlaps, presumably because they occurred at such low frequencies. In 1990 we examined whether the development of dewlap type was resistant to variation in egg and hatchling rearing environment. If o spring from populations di ering in dewlap types reared in a common laboratory environment develop population-typical dewlaps, genetic di erences or maternal environmental e ects are implicated, in contrast to other possible environmental factors acting during egg (e.g., temperature, moisture) or juvenile (e.g., social environment) stages. In the laboratory we rst determined the frequency of background types in adult males and females reared from eggs from di erent populations. Second, we compared the presence or absence of the blue patch in adult male o spring. The hypothesis that all between-population variation in dewlap type is due to environmental e ects predicts that the frequencies of individuals expressing each type should not di er among the o spring Table 1 Description of Urosaurus ornatus collection sites in central Arizona, including habitat type and the primary substrate on which tree lizards were found. For each site, the abbreviation used in the text and gures is in parentheses Site name Habitat type Lizard substrate Sycamore Creek (SC) Coon Blu (CB) South Mountain (SM) Sierra Estrella (SE) Top of the World (TW) Upper Sonoran desert vegetation Mature stand of velvet mesquite trees (riparian) Upper Sonoran desert vegetation Upper Sonoran desert vegetation Scrub oak, juniper, and manzanita Large boulders Mesquite trees Large boulders Large boulders Large boulders

4 374 of the populations when reared in a common environment. Alternatively, if some of the population di erences are due to genetic di erences, the frequencies of types should di er among the o spring of these populations. For these rearing studies we selected three populations, two in which most animals had yellow backgrounds (SM, TW), and one in which the majority had orange backgrounds (SC). We captured gravid females as randomly as possible (with respect to male territories). Eggs from these gravid females were incubated in individual containers with moistened vermiculite (for methods see Hews et al. 1994), resulting in 96% hatching success. On the day of hatching we assessed sex using an external scale character and individually marked animals with a unique toe-clip. We housed hatchlings in same-sex groups of six to ten individuals in terraria and provided them with ad libitum water and food (crickets and wingless fruit ies, dusted with bone meal), access to incandescent heat lamps for thermoregulation, and full spectrum and UV lighting over each tank (see Hews et al for details). O spring were reared until adult dewlap background color could be scored. For males, this was when blue ventrolateral ``belly'' patches were fully expressed, usually by 90 days post hatching. This is an androgen-dependent (Hews et al. 1994; Hews and Moore 1995), maletypical trait. For females nal dewlap color was scored at 150 days post-hatching, when laboratory-reared females were reproductively mature (previously con rmed via laparotomy). All dewlap frequency data were compared as count data, using log likelihood ratio v 2 tests, computed using SYSTAT statistical software (Wilkinson 1988). When data were used in two comparisons (e.g., when comparing males among sites, and comparing the sexes within a site), Bonferroni corrections (e.g., Rice 1989) were not used because in all such cases, di erences were highly signi cant and remained so if corrected. di er signi cantly from males, in which 79% (71/90) had orange backgrounds (v 2 = 1.309, df =1,P< 0.253). Thus frequencies of orange and yellow background types in males and females closely paralleled each other at all four sites where we examined both sexes, although in some comparisons there were small but statistically signi cant di erences. Annual variation Frequencies of dewlap types varied little across years. At both sites with multi-year samples, there was no signi- cant variation across years either in frequencies of males with orange, yellow, or blue backgrounds (Fig. 2A; TW, v 2 = 5.384, df =4, P= 0.250; CB, v 2 = 7.086, df =4,P= 0.131), and in frequencies of the blue patch when solid blue males were scored as expressing the patch (Fig. 2B; TW, v 2 = 4.230, df =2, P= 0.121; CB, v 2 = 0.175, df =2,P= 0.916). When solid blue males were excluded (data not shown) the same results were obtained for background comparisons (TW, v 2 = 2.536, df =2,P= 0.281; CB, v 2 = 2.186, df = 2, P = 0.335) and for the comparisons of blue patch frequencies (TW, v 2 = 5.458, df =2,P= 0.065; CB, v 2 = 0.031, df =2,P= 0.985). Results Male-female comparisons of background color Frequencies of orange and yellow background types in males and females were similar at the SC site, although in some years there were small but statistically signi cant di erences. In 1990 and 1993, the sexes did not di er signi cantly in the frequencies of orange or yellow dewlap backgrounds (1990: yellow background, males, 17% (13/77), females 13% (11/87), likelihood ratio v 2 test, P = 0.486; 1993: males, 7% (15/198), females 6% (12/202), P = 0.542). In the other two years SC males and females di ered slightly (1991: yellow background, males, 3% (6/199), females 9% (12/132), P = 0.018; 1992: males, 2% (1/52), females 10% (18/164), P = 0.034). As with the SC population, there were slight but statistically signi cant sex di erences in the frequencies of background color at the other three sites, although the data are more limited because we sampled both males and females in only one year (1990) at these sites. In the SM population 90% (37/41) of the adult females had a solid yellow dewlap and this percentage di ered from the 100% (68/68) in males (v 2 = 8.077, df =1, P< 0.004). In the TW population 93% (51/55) of adult females had a solid yellow dewlap, statistically lower than the 100% (65/65) seen in males (v 2 = 6.405, df =1,P< 0.011). In the CB population, 87% (33/38) of adult females had orange dewlaps, which did not Among-site variation and habitat correlates Dewlap color patterns in tree lizards varied substantially among central Arizona populations. Combining data for the three years for CB and for TW, we compared fre- Fig. 2A, B Geographic variation in dewlap color patterns in adult tree lizards, including males with a solid blue dewlap, who were scored as expressing a blue background and the blue dewlap patch. A Percentages of males with yellow (open bar), orange (shaded bar) or blue (solid bar) background color, B Percentages of males with (solid bar) and without (open bar) the central blue dewlap patch. Sample sizes for both A and B are in each bar in A. Numbers above a site abbreviation indicate the year of sampling, for sites sampled in multiple years

5 375 quencies among all ve populations. Including males with solid blue dewlaps, there was signi cant heterogeneity among the ve populations in background frequencies (Fig. 2A, v 2 = 396.5, df =8, P< ) and in blue patch frequencies (Fig. 2b, v 2 = 73.81, df = 4, P < ). Results excluding solid blue males were identical (data not shown): populations varied both in the frequencies of background types (v 2 = , df =4,P< ) and in the frequencies of the blue patch (v 2 = 58.31, df =4, P< ). Thus, two populations, TW and SM, expressed yellow backgrounds almost exclusively, two others, SC and CB, expressed orange backgrounds at relatively high frequencies, and the SE population had intermediate frequencies of individuals with orange backgrounds. These SC results are as Thompson and Moore (1991a) previously reported for both the SC and FM populations. Finally, males with the blue patch were more common than those without at TW, SE, and CB, but males without the blue dewlap patch were more common at SM. We censused lizards in two habitat types where tree lizards were especially common: stands of mature mesquite trees and more open desert ``boulder elds''. At the two mesquite sites, CB (this study) and FM (Thompson and Moore 1991a), there were relatively high frequencies of the orange background and of the blue dewlap patch (Fig. 2). These two sites are within a few kilometers of each other (Fig. 1). Both background color and blue patch expression varied among the boulder populations. In two boulder populations (SM and TW) 98±100% of the lizards had yellow backgrounds, but in two other boulder populations (SE, SC), orange background was the more common type, and was expressed by 60±90% of lizards. Similarly, the frequency of males with the blue patch varied among boulder populations: around 30% at SM, around 50% at SC, and around 80% at both SE and TW had the patch. Thus there was no clear correlation between habitat (boulders, mesquite trees) and variation in the two dewlap elements. Finally, one result suggests that the two dewlap elements do not covary. The two populations with predominantly yellow backgrounds, SM and TW, di ered the most in the frequency of the blue patch. colors as adults (Fig. 3A, v 2 = , df =2, P< ; no analyses with solid blue males because none were produced). We obtained similar results for females (Fig. 3B): adult female o spring of the three populations di ered in background frequencies (v 2 = 33.38, df =2;P< ), as did parental females of the three populations (v 2 = , df =2, P< ). Further, background frequencies in male and female o spring were not signi cantly di erent Fig. 3A, B Percentages of tree lizards expressing each dewlap background color in the common garden experiment, which compared o spring reared from three source populations di ering in the proportions of tree lizards with each background type. A Males, B Females. Sample sizes are in each bar. Frequencies for parental populations are presented for comparison (includes males with solid blue dewlaps) (abbreviations: O sp o spring, Pop parental population) Rearing studies To examine e ects of rearing environment on the expression of dewlap color, we reared male and female o spring from a primarily ``orange'' population (SC) and from two ``yellow'' populations (SM, TW). Adult males captured in these three parental populations differed in the frequencies of background colors (including solid blue males, Fig. 3A, v 2 = , df =2, P< ; excluding solid blue males, v 2 = 230.5, df =2,P< ). Though reared in a single environment, male o spring from these three populations di ered signi cantly in the frequencies of background Fig. 4 Percentages of adult males with (solid bar) or without (open bar) the blue dewlap patch that are o spring from di erent populations incubated and reared in the same laboratory environment. Frequencies for parental populations are presented for comparison (includes males with solid blue dewlaps). Sample sizes are in each bar. (abbreviations O sp o spring, Pop parental population)

6 376 from frequencies in the parental populations: progeny from SM and TW had exclusively yellow backgrounds, whereas a higher percentage of the SC progeny expressed orange, similar to the SC parental population. We compared the frequencies of males with and without the blue patch (Fig. 4), using only two populations (SC and TW) from the rearing experiments. Statistical power for the analysis of male dewlap types in o spring from the third SM population would have been low for comparing proportions of types within one sex, and only nine males were reared from SM (seven had the blue patch). In 1990 the parental SC and TW populations di ered signi cantly in frequencies of captured adult males with or without the blue patch (including solid blue males, Fig. 4, v 2 = , df =1, P< ; excluding solid blue males, v 2 = 5.426, df = 1, P < 0.020). These frequencies also di ered in the lab-reared adult male o spring (v 2 = 6.525, df =1, P< 0.011; no solid blue males produced). Finally, frequencies of adult male o spring with the blue patch did not di er signi cantly from those in adult males of the two respective parental populations. Discussion Sexual variation The sexes di ered little in frequencies of individuals with orange versus yellow background colors. This suggests that control of background color may be similar in males and females. Hormonal studies support this hypothesis, as androgen manipulations in males and in females, either in adults or in juveniles, did not alter the population-typical frequencies of background colors (Hews et al. 1994; Hews and Moore 1995; D. Hews and M. Moore, unpublished work). Control of orange and yellow background color may di er between Arizona and New Mexico populations. Zucker and Boecklen (1990) suggested that in a New Mexico population of U. ornatus males and females di ered in the control of orange and yellow background colors because dewlap color of adult females varied with reproductive state and yellow dewlaps changed to orange in gravid females. We collected only gravid females with mature eggs, many of which were yellow females (from SC and CB). These females were held in captivity for up to 2 weeks until oviposition, and we never observed dewlap color to change between yellow and orange. Annual variation Frequencies of dewlap elements were relatively constant among years. In 4 years of sampling at SC and in 3 years of sampling at TW and CB there was little yearly variation in frequencies of the two dewlap elements. Frequencies of the male blue dewlap patch for SC during 1990±1993 also were very similar to earlier reports for SC during 1986±1988 by Thompson and Moore (1991a), when the central blue patch frequency varied between 55% and 60%. Thus, over 8 years at the SC site, there was no signi cant variation in frequencies of male dewlap types. Geographic variation and habitat correlates Dewlap elements in tree lizards varied over relatively small distances (less than or equal to 120 km). Adding to analyses by Thompson and Moore (1991a), who reported dewlap frequencies for SC and FM, we sampled four central Arizona populations. Some populations had yellow backgrounds almost exclusively (TW, SM), while other had either high (SE, SC in 1991±1993) or intermediate (CB) frequencies of the orange background. Although the number of sites we sampled was too small for rigorously assessing habitat correlations, there was no simple pattern of variation associated with habitat. For example, boulder sites varied in frequencies of both the dewlap elements. Dewlap frequencies at the two mesquite populations were similar (FM population, Thompson and Moore 1991a; CB, this study; Fig. 2). These two sites, however, were within 10 km of each other and were connected by a riparian corridor, and may experience gene ow. Di erences between habitats in blue patch frequencies might be due to biased sampling of the two male types, if the types di er in behavior. For example, at the CB mesquite population nonterritorial males (which lacked the blue patch) were more commonly resighted higher in the tree canopy than were territorial males (which had the blue patch), which were more often resighted lower on the tree trunks (L. Oliver, D. Hews, and M. Moore, unpublished work). Frequencies varied among boulder populations, but this probably was not due to sampling biases. At SC, a boulder population, males with and without the blue patch did not di er in microhabitat use (D. Hews and M. Moore, unpublished work). In addition, our success rate at capturing lizards once they were sighted was very high at all boulder sites. Thus, in the less complex boulder habitat, we might have been more likely to get an unbiased sample. Variation in U. ornatus dewlap colors is also evident on a larger geographic scale. In the eastern part of the species' range (west Texas; some parts of New Mexico), populations are monomorphic for male dewlap type, with all males having solid blue dewlaps (Mittleman 1942; Smith 1946). In contrast, solid blue was a rare dewlap type in AZ populations except at TW where 15± 19% males were this type. Further, males in some New Mexico (NM) populations (Carpenter 1995) and in our central AZ populations were polymorphic. Dewlap type nomenclature in Carpenter (1995) di ered from ours, but photographs in Carpenter (1995) depicted two types of male dewlaps termed Blue and Blue-Green at Aguirre Springs, NM. These dewlap types appeared similar to

7 377 Blue and Yellow-Blue males at our eastern-most site, TW, on the west side of an interconnected series of mountains extending into NM. Dewlap frequencies of captive-reared males from Aguirre Springs (Carpenter 1995) were similar to frequencies found at our TW site, if dewlaps scored by Carpenter as having green were considered analogous to ``blue patch present''. We scored green hues in our populations as ``blue'' because overlapping layers of yellow and blue, or layers of orange and blue, appear to produce the green (see Carpenter 1995 for color photos). Rearing studies and determination of male types The developmentally permanent male types in the SC population could arise from genetic di erences, from di erences in e ects of environmental factors in development, or from both. We manipulated hormones in hatchling males and altered the adult dewlap type (Hews et al. 1994). Genetic di erences could produce such hormonal di erences. Alternatively, two possible environmental hypotheses also involve hormone di erences due to di erences in juvenile social environment or in maternally-derived yolk steroid hormones. Social experience can alter plasma pro les of gonadal and adrenal hormones, which mediate agonistic interactions and dominance relationships in many vertebrates (Nelson 1995), and social environment can a ect adoption of behavioral alternatives (Lawrence 1987; Crespi 1988; Kodric-Brown 1988; Travis and Woodward 1989; Verrell 1983; Arak 1988; Wagner 1992). The rearing results of this study and other data for our SC population suggest that social environment does not alter dewlap type. Males reared in the laboratory under di erent social conditions expressed dewlap types at a frequency identical to the parental population (Thompson et al. 1993). Maternal e ects are a particularly likely environmental factor a ecting the developmentally xed male alternatives in tree lizards. Maternal e ects on o spring phenotype are well known (Mousseau and Dingle 1991; White 1991; Moore 1995; Bernardo 1996; Pennisi 1996). For example, egg size variation could a ect body size, which might in uence determination of male types. However, egg size and size at hatching do not di er between the male types in tree lizards (Hews et al. 1994). Hormonal maternal e ects can be substantial in vertebrates (Clark and Galef 1995; Moore 1995), even in oviparous species (Palmer and Guillette 1991). In some birds, the social rank of juveniles correlates with the level of maternally-derived testosterone in the yolk of the eggs from which they hatched (Schwabl 1993). In tree lizards, di erences in maternal yolk testosterone could determine the male types and this deserves study. Five sets of results for tree lizards from this and earlier studies are equally consistent with the genetic di erences and the maternal e ects hypotheses: 1. Adult males never changed dewlap type (Thompson and Moore 1992). 2. Hatchling and juvenile social environment did not a ect adult dewlap types frequencies (Thompson et al. 1993). 3. Populations di ered in frequencies of dewlap types and did so consistently across years and presumably generations (this study). 4. When reared in common incubation and posthatching laboratory environments, adult o spring of di erent populations di ered in frequencies of the blue dewlap patch, and did so in a population-typical manner (this study). These rearing results do not support the environmental hypothesis involving incubation or non-social post-hatching stimuli. 5. Finally, when female hatchlings from the SC population were given implants with the androgen dihydrotestosterone (DHT), they expressed the maletypical blue dewlap patch as adults, and did so at a frequency not di erent from that in SC males (Hews and Moore 1995). Thus DHT may have revealed the underlying genotype of females (suggesting autosomal inheritance), or it may have activated an underlying phenotype that was hormonally organized earlier in development by maternally-derived eggyolk hormones but not normally expressed in females because female vertebrates typically have low plasma DHT levels. For discussion of hormonal organization (permanent e ects) and activation (temporary e ects) and development of male alternatives in vertebrates see Moore (1991). If maternal hormonal e ects play a role in tree lizard male types, the e ects may be complex. Single matings of virgin females in captivity can result in clutches (5 of 9 females) containing both types of males when reared in the laboratory (SC population; D. Hews and M. Moore, unpublished work). An example of complexity in maternal hormone e ects is found in some birds, where levels of maternally-derived testosterone in eggs increase with order of laying within a clutch (Schwabl 1993). An additional complexity is that female tree lizards could have environmentally and/or genetically based di erences in the amount of steroid hormones in egg yolk, contributing to the population-speci c frequencies of male types. Species with male alternatives vary extensively in the degree to which expression of the male types is in uenced by environmental factors (Caro and Bateson 1986; Gross 1996). Environmental factors appear to contribute little to the di erences characterizing male types in some species, including ru s, Philomachus pugnax (van Rhijn 1983, 1991; Lank et al. 1995), swordtails, Xiphophorus spp. (Kallman 1984, 1989; Zimmerer and Kallman 1989), bluegill sun sh, Lepomis (Dominey 1980; Kindler et al. 1989), crickets, Gryllus integer (Cade 1981), and a marine isopod crustacean, Paracerceis sculpta (Baitoo et al. 1988; Shuster 1991; Shuster and Wade 1991). In other species with male alternatives,

8 378 environmental e ects probably are major contributors to the male types (Gross 1996). Social environment (see citations above) and energetic state (Gross 1996) are common examples of environmental e ects on male alternatives. Many environmental factors can also contribute to di erences in body size, which often correlate with adoption of male alternatives (Gross 1996). Size di erences occur commonly in species with developmentally plastic (sensu Austad 1984, Moore 1991) alternatives, often representing options that change ontogenetically (e.g., Pratt et al. 1994). Comparative and evolutionary considerations Rearing studies by Carpenter (1995) reveal similarities and di erences between AZ and NM tree lizard populations. Carpenter (1995) put gravid females in outdoor enclosures where resulting hatchlings were reared. As adults, some male Aguirre Springs o spring had dewlap types not found in the eld, suggesting that rearing environment a ected dewlap color development or that di erent male dewlap types di ered in survivorship in the wild. Adult male o spring similarly reared from a second population, DonÄ a Ana, had dewlap frequencies that did not di er from adults captured at this site, consistent with both the genetic or maternal e ects hypotheses. Some males also altered dewlap colors at a sexually mature size (Carpenter 1995). Because of differences in scoring criteria we could not determine whether, using our criteria, these mature individuals would have ``switched'' dewlap types. In sum, however, these and other data (see above section on sexual differences) suggest that AZ and NM populations di er in both developmental and functional aspects of the male dewlap types. Evolutionary understanding of male alternatives would be furthered with increased information on variation in the alternatives. Studies of such variation are lacking in two respects. First, compared to work on species with developmentally plastic types, little work has focused on annual or geographic variation in species with developmentally xed alternatives (Sinervo and Lively 1996). For example, many studies have found that adoption of alternative male behaviors is correlated with variation in male-bias to the operational sex ratio (Verrell 1983; Lawrence 1987; Arak 1988; Crespi 1988; Kodric-Brown 1988; Moore 1989; Travis and Woodward 1989; Sullivan 1989; Wagner 1992). A rare study of geographic variation found that two populations of soapberry bugs di ered in the frequencies of two plastic male alternative behaviors, guarding and non-guarding, and that the frequencies correlated with di erences in the population sex ratio (Carroll 1993). Experiments suggested that populations di ered genetically in male ability to respond to sex ratio variation (Carroll and Corneli 1995). Data on relative tnesses of male types are also key in testing evolutionary models and such data are rare (e.g., Ryan et al. 1992; Sinervo and Lively 1996). The tree lizard rearing data may indicate di erences in mating success. Assuming male alternatives actually are due to genetic di erences, then the laboratory rearing data should re ect mating success of the male types if gravid females were captured as randomly as possible (with respect to male territories), which we attempted to do. Mark-recapture data also indicate that gravid females often made extensive forays o their usual homeranges in search of suitable nest-sites (L. Oliver and M. Moore, unpublished work). Thus if ``territorials'' (blue patch present) and ``rovers'' (blue patch absent) have equal tnesses, then the laboratory frequencies should be the same as in nature. But if territorials have greater mating success, then laboratory frequencies should be higher than in nature. In both populations we had a slight bias in favor of higher frequencies of the territorial type in the laboratory. Territorials may thus have higher mating success. However frequency-dependence in the system might keep the tness di erences small and di cult to measure without large sample sizes (Gross 1996). Lizards may be especially good systems for determining both the developmental mechanisms and the tness consequences of male alternatives. Male behavioral alternatives occur in other lizards and may vary geographically. In the lizard Sceloporus undulatus erythrocheilus, male dewlap colors correlated with territorial and nonterritorial alternatives in one population, and male color frequencies varied geographically (Rand 1990, 1991). Recapture data and adult hormone manipulations suggested that male types were developmentally permanent (Rand 1992). Throat color varied among males in a population of side-blotched lizards, Uta stansburiana, and Sinervo and Lively (1996) suggested that color correlated with behavioral and genetic di erences. The extent of geographic variation in male types is unclear, although other U. stansburiana populations can lack some or all of these color variants (e.g., Tinkle 1967). Relative mating success of male types in lizards could be assessed using genetic paternity analyses, and survivorship can be assessed using markresighting techniques. Thus, future work on these and other systems can focus on determining the key genetic and/or environmental cue(s) controlling such developmental trajectories and the tness consequences of the resulting male alternative phenotypes. Acknowledgements This work was supported by NIMH Individual NRSA MH10074 PYB to D.K.H., and by NIMH MH PYB to M.C.M. We thank A. Bradley, L. Oliver, R. Knapp, M. Stewart, S. Woodley for assistance, and L. Neinaber for valuable animal care assistance. M. Gross and an anonymous reviewer made helpful comments on this manuscript. This work was conducted under ASU Animal Care Protocol 90±180R. We collected with Arizona Game and Fish Department and the U.S. Forest Service permits. We followed the ASIH/SSAR/HL 1987 Guidelines for Use of Live Amphibians and Reptiles in Field Research.

9 379 References Arak A (1988) Callers and satellites in the natterjack toad: evolutionary stable decision rules. Anim Behav 36:416±432 Austad SN (1984) A classi cation of alternative reproductive behaviors and methods of eld-testing ESS models. Am Zool 24:309±319 Baitoo HS, Shuster SM, Sassaman CA (1988) Polymorphic male molting and growth schedules in a marine isopod crustacean. Am Zool 28:134A Bernardo J (1996) Maternal e ects in animal ecology. Am Zool 36:83±105 Brown DE (ed) (1982) Biotic communities of the American southwest± United States and Mexico. Desert Plants 4 (special issue):1±342 Cade WH (1981) Alternative male strategies: genetic di erences in crickets. Science 212:563±564 Caro TM, Bateson P (1986) Organization and ontogeny of alternative tactics. Anim Behav 34:1483±1499 Carpenter GC (1995) The ontogeny of a variable social badge: throat color development in tree lizards (Urosaurus ornatus). J Herpetol 29:7±13 Carroll SP (1993) Divergence in male mating tactics between two populations of the soapberry bug. I. Guarding versus nonguarding. Behav Ecol 4:156±164 Carroll SP, Corneli PS (1995) Divergence in male mating tactics between two populations of the soapberry bug. II. Genetic change and the evolution of a plastic reaction norm in a variable social environment. Behav Ecol 6:46±56 Clark MM, Galef BG Jr (1995) Prenatal in uences on reproductive life history strategies. Trends Ecol Evol 10:151±153 Cooper WE, Greenberg NB (1992) Reptilian coloration and behavior. In: Gans C, Crews D (eds) Hormones, brain and behavior. Biology of the reptilia, vol 18, Physiology E. University of Chicago Press, Chicago, pp 298±422 Crespi B (1988) Alternative male mating tactics in a thrip: e ects of sex ratio variation and body size. Am Midl Nat 119:83±92 Dominey WJ (1980) Female mimicry in male bluegill sun sh: a genetic polymorphism? Nature 284:546±548 Gross M (1984) Sun sh, salmon, and the evolution of alternative reproductive strategies in shes. In: Potts GW, Wooton RJ (eds) Fish reproduction: strategies and tactics. Academic Press, New York, pp 55±75 Gross M (1996) Alternative reproductive strategies and tactics: diversity within sexes. Trends Ecol Evol 11:92±98 Hews DK (1990) Examining hypotheses generated by eld measures of sexual selection on male lizards, Uta palmeri. Evolution 44:1956±1966 Hews DK (1993) Food resources a ect female distribution and male mating opportunities in the iguanian lizard, Uta palmeri. Anim Behav 46:279±291 Hews DK, Moore MC (1994) Relative body mass but not fat body or liver mass di ers between male reproductive alternatives of the tree lizard, Urosaurus ornatus. Am Zool 34:120A Hews DK, Moore MC (1995) In uence of androgens on the development of secondary sexual characters in tree lizards (Urosaurus ornatus). Gen Comp Endocrinol 97:86±102 Hews DK, Knapp R, Moore MC (1994) Early exposure to androgens a ects adult expression of alternative male types in tree lizards. Horm Behav 28:96±115 Hover EL (1982) Behavioral correlates of a throat color polymorphism in a lizards, Urosaurus ornatus. PhD dissertation, University of Michigan, Ann Arbor Hover EL (1985) Di erences in aggressive behavior between two throat color morphs in a lizard, Urosaurus ornatus. Copeia 1985:933±940 Howard DR (1984) Alternative mating behaviors in young male bullfrogs. Am Zool 24:397±406 Kallman KD (1984) A new look at sex determination in poeciliid shes. In: Turner BJ (ed) Evolutionary genetics of shes. Plenum, New York, pp 95±171 Kallman KD (1989) Genetic control of size at maturity in Xiphophorus. In: Me e GK, Snelson FF Jr (eds) Ecology and evolution of livebearing shes (Poeciliidae). Prentice Hall, Englewood Cli s, pp 163±184 Kindler PM, Phillip DP, Gross MR, Bahr JM (1989) Serum 11- ketotestosterone and testosterone concentrations associated with reproduction in male bluegill (Lepomis macrochirus: Centrarchidae). Gen Comp Endocrinol 75:446±453 Knapp R, Moore MC (1995) Hormonal responses to aggression vary in di erent types of agonistic encounters in male tree lizards, Urosaurus ornatus. Horm Behav 29:85±105 Knapp R, Moore MC (1996) Male morphs in tree lizards, Urosaurus ornatus, have di erent delayed hormonal responses to aggressive encounters. Anim Behav 52:1045±1055 Kodric-Brown A (1988) E ects of sex ratio manipulation on territoriality and spawning success of male pup sh, Cyprinodon pecosensis. Anim Behav 36:1136±1144 Lank DB, Smith CM, Hanotte O, Burke T, Cooke F (1995) Genetic polymorphism for alternative mating behaviour in lekking male ru Philomachus pugnax. Nature 378:59±62 Lawrence WS (1987) Dispersal: an alternative mating tactic conditional on sex ratio and body size. Behav Ecol Sociobiol 21:367±373 Lott D (1991) Intraspeci c variation in the social systems of wild vertebrates. Cambridge University Press, Cambridge Mittleman MB (1942) A summary of the iguanid genus Urosaurus. Bull Mus Comp Zool 91:105±181 Moore AJ (1989) The behavioral ecology of Libellula luctuosa (Burtmeister) (Odonata: Libellulidae). III. Male density, OSR, and male and female mating behavior. Ethology 80:120±136 Moore CL (1995) Maternal contributions to mammalian reproductive development and the divergence of males and females. In: Slater PJB, Rosenblatt JS, Snowdon CT, Milinski M (eds) Advances in the study of behavior, vol 24. Academic Press, New York, pp 47±118 Moore MC (1991) Application of organization-activation theory to alternative male reproductive strategies: a review. Horm Behav 25:154±179 Mousseau TA, Dingle H (1991) Maternal e ects in insect life histories. Ann Rev Entomol 36:511±534 Nelson RJ (1995) An introduction to behavioral endocrinology. Sinauer Assoc, Sunderland Palmer BD, Guillette LJ Jr (1991) Oviductal proteins and their in uence on embryonic development in birds and reptiles. In: Deeming DC, Ferguson MWJ (eds) Egg incubation: its e ects on embryonic development in birds and reptiles. Cambridge University Press, Cambridge, pp 29±46 Pennisi E (1996) A new look at maternal guidance. Science 273:1334±1336 Pratt NC, Phillips JA, Alberts AC, Bolda KS (1994) Functional versus physiological puberty: an analysis of sexual bimaturism in the green iguana, Iguana iguana. Anim Behav 47:1101±1114 Rand MS (1990) Polymorphic sexual coloration in the lizard Sceloporus undulatus erythrocheilus. Am Midl Nat 125:352±359 Rand MS (1991) Behavioral function and hormonal control of polymorphic sexual coloration in the lizard Sceloporus undulatus erythrocheilus. PhD dissertation, University of Colorado, Boulder Rand MS (1992) Hormonal control of polymorphic and sexually dimorphic coloration in the lizard Sceloporus undulatus erythrocheilus. Gen Comp Endocrinol 88:461±468 Rhijn JG van (1983) On the maintenance and origin of alternative strategies in the ru Philomachus pugnax. Ibis 125:482±498 Rhijn JG van (1991) The ru, individuality in a gregarious wading bird. Academic Press, New York Rice WR (1989) Analyzing tables of statistical tests. Evolution 43:223±225

10 380 Rubenstein DI (1980) On the evolution of alternative mating strategies. In: Staddon JER (ed) Limits to action: the allocation of individual behavior. Academic Press, New York, pp 65±100 Ryan MJ, Causey BA (1989) ``Alternative'' mating behavior in the swordtails Xiphophorus nigrensis and Xiphophorus pygmaeus (Pisces: Poeciliidae). Behav Ecol Sociobiol 24:341±348 Ryan MJ, Pease CM, Morris MR (1992) A genetic polymorphism in the swordtail Xiphophorus nigrensis: testing the prediction of equal tness. Am Nat 139:21±31 Schwabl H (1993) Yolk is a source of maternal testosterone for developing birds. Proc Natl Acad Sci USA 90:11446±11450 Shuster SM (1991) The ecology of breeding females and the evolution of polygyny in Paracerceis sculpta, a marine isopod crustacean. In: Bauer RT, Martin JW (eds) Crustacean sexual biology. Columbia University Press, New York, pp 91±110 Shuster SM, Wade MJ (1991) Equal mating success among male reproductive strategies in a marine isopod. Nature 350:608±610 Sinervo B, Lively CM (1996) The rock-paper-scissors game and the evolution of alternative male strategies. Nature 380:240±243 Smith HM (1946) Handbook of lizards. Comstock, Cornell Stamps JA (1983) Sexual selection, sexual dimorphism and territoriality. In: Huey RB, Pianka ER, Schoener TW (eds) Lizard ecology: studies of a model organism. Harvard University Press, Cambridge, pp 169±204 Sullivan BK (1989) Mating system variation in Woodhouse's toad (Bufo woodhousii). Ethology 83:60±68 Thompson CW, Moore MC (1991a) Syntopic presence of multiple dewlap color morphs in male tree lizards. Copeia 1991:493±503 Thompson CW, Moore MC (1991b) Throat color reliably signals status in male tree lizards, Urosaurus ornatus. Anim Behav 42:745±753 Thompson CW, Moore MC (1992) Behavioral and hormonal correlates of alternative reproductive strategies in a polygynous lizards: tests of the relative plasticity and challenge hypotheses. Horm Behav 26:568±585 Thompson CW, Moore IT, Moore MC (1993) Social, environmental and genetic factors in the ontogeny of phenotypic differentiation in a lizard with alternative male reproductive strategies. Behav Ecol Sociobiol 33:137±146 Tinkle DW (1967) The life and demography of the side-blotched lizard, Uta stansburiana. Misc Pub Mus Zool Univ Mich 132:1± 182 Travis J (1994) Size-dependent behavioral variation and its genetic control within and among populations. In: Boake CR (ed) Quantitative genetic studies of behavioral evolution. University of Chicago Press, Chicago, pp 165±187 Travis J, Woodward BD (1989) Social context and courtship exibility in male sail n mollies, Poecilia latipinna (Pisces: Poeciliidae). Anim Behav 38:1001±1011 Verrell PA (1983) The in uence of the ambient sex ratio and intermale competition on the sexual behavior if the red-spotted newt, Notophthalmus viridescens (Amphibia: Urodela: Salamandridae). Behav Ecol Sociobiol 13:307±313 Wagner RH (1992) Extra-pair copulations in a lek: a secondary mating system. Behav Ecol Sociobiol 31:63±71 White HB III (1991) Maternal diet, maternal protein and egg quality. In: Deeming DC, Ferguson MWJ (eds) Egg incubation: its e ects on embryonic development in birds and reptiles. Cambridge University Press, Cambridge, pp 1±16 Wilkinson L (1988) SYSTAT: the system for statistics. Systat, Evanston Zimmerer EJ, Kallman KD (1989) The genetic basis for alternative reproductive tactics in the pygmy swordtail, Xiphophorus nigrensis. Evolution 43:1298±1307 Zucker N, Boecklen W (1990) Variation in female throat coloration in the tree lizard (Urosaurus ornatus): relation to reproductive cycle and fecundity. Herpetologica 46:387±394 Communicated by P.J. Weatherhead Note added in proof Due to an oversight the following data were omitted from the results section: The percentage of males (including solid blue males) with the blue patch in 1990±1993 were, respectively, 48% (38/79), 48% (101/210), 63% (32/52), and 54% (114/ 213) at the SC site. This annual variation was not statistically signi cant (v 2 ˆ 3:808; df ˆ 3; P < 0:283 :