Abstract

Increases in population density often are associated with a change in mating system structure in numerous taxa. Typically, male interactions are minimal in extremely low density populations. As density increases, males exhibit territoriality but if density becomes too high, the energetic cost of defending a territory will eventually outweigh the reproductive benefits associated with territoriality. Consequently, males in high density populations may abandon territoriality and adopt dominance polygyny, lekking behavior, or scramble competition. We investigated the relationship between population density and mating system structure in three populations of the chuckwalla, Sauromalus obesus (= ater), near Phoenix, Arizona. Densities in the Phoenix Mountains (2.7 chuckwallas/ha) were lower than any population previously studied. In the Santan Mountains (10.9 chuckwallas/ha), densities were similar to populations studied in the Mojave Desert, and in the South Mountains (65 chuckwallas/ha), densities were the highest yet recorded. Male mating behavior was examined by determining home range overlap and by making direct behavioral observations. Male home range size decreased with increasing population density. There was little overlap in home ranges among males in all three populations, whereas home ranges of males and females consistently overlapped, indicating that males were strictly territorial. This conclusion was supported by behavioral observations of interactions among individuals in a natural setting. The number of females wihin male territories was correlated with food resources (plants) in all three populations. Female home range size appeard to be related to food resources whereas male home ranges appeared to be related to female distribution, population density, and geology. The retention of territoriality in spite of high population densities raises new questions about the relationship between density and resource defense.

Since the seminal work of Emlen and Oring (1977), it has been widely accepted that mating system structure is influenced by a variety of ecological variables, especially population density (Davies, 1991; Maher and Lott, 2000; Travis et al., 1995; Vehrencamp and Bradbury, 1984; Warner and Hoffman, 1980a,b). At both extremely low and high densities, energetic costs of defense should outweigh any reproductive benefits associated with territoriality (Brown, 1964; Emlen and Oring, 1977; Wilson, 1975). Consequently, males in low density populations may exhibit no site defense, males in moderate density populations exhibit territoriality (for review of definitions, see Maher and Lott, 1995), and males in high density populations may abandon territoriality and exhibit dominance polygyny, leks, or scramble competition. Examples of shifts in mating system due to variation in population density can be found in a diversity of taxa. In many ungulates, such as topi (Damaliscus lunatum), Uganda kob (Kobus kob), and fallow deer (Cervus dama), males in low density populations exhibit female (harem) defense polygyny (sensu Emlen and Oring, 1977). As density increases, males exhibit resource defense polygyny, in which males defend non-overlapping territories containing resources important to females (Höglund and Alatolo, 1995; Langbein and Thirgood, 1989). At even higher densities, males exhibit dominance polygyny (sensu Emlen and Oring, 1977), in which males overlap spatially and form dominance hierarchies, and at the highest densities, males form leks (Clutton-Brock et al., 1993; Höglund and Alatolo, 1995; Langbein and Thirgood, 1989). Similar associations between mating system and density exist in odonates, where males exhibit patrolling, territoriality, and territoriality with satellite males, or scramble competition, as density increases (Banks and Thompson, 1985; Kaiser, 1985; Sherman, 1983), and anurans, where some males abandon active defense of calling sites and become satellites under high densities (e.g., Sullivan, 1989).

Variation in mating system structure is common in lizards, especially the herbivorous iguanids (Stamps, 1977, 1983). Many male iguanids defend non-overlapping territories to control access to mates during a well defined mating season (reviewed by Dugan and Wiewandt, 1982; Stamps, 1977, 1983). Some anecdotal evidence suggests that males may also defend resources important to females rather than females per se. For example, male Galapagos land iguanas, Conolophus subcristatus, defend territories based partly on food resources, which they guard before the arrival of females (Werner, 1982). Territoriality in iguanids often breaks down when population densities become high, causing males of numerous species to exhibit dominance polygyny (Berry, 1974; Brattstrom, 1974; Dugan and Wiewandt, 1982; Evans, 1951; Prieto and Ryan, 1978; Ryan, 1982). One often-cited example of the association between density and mating system structure is the chuckwalla, Sauromalus obesus (considered S. ater by some; Hollingsworth, 1998), from the Mojave and Sonoran Deserts (Berry, 1974; Prieto and Ryan, 1978; Ryan, 1982). Mating system structure in chuckwallas is variable; typically, males are territorial (no spatial overlap; Johnson, 1965; Prieto and Ryan, 1978), although at high densities, males appear to adopt dominance polygyny in which they have overlapping home ranges and form dominance hierarchies (Berry, 1974; Prieto and Ryan, 1978; Ryan, 1982).

Chuckwalla population densities vary considerably in the Sonoran Desert in the vicinity of Phoenix, Arizona (Sullivan and Flowers, 1998). Densities at some locations are similar to those of Mojave populations, whereas others represent the lowest and highest yet recorded. This variation makes Sonoran Desert chuckwallas ideal for investigating the relationship between population density and mating system structure. Based on empirical data from field and captive chuckwalla populations, Ryan (1982) developed a model predicting that increases in population density will result in changes in mating system structure. Accordingly, we predicted that males would be territorial in low and moderate density populations; that is, there would be little overlap in male home ranges because males would defend areas with females or areas with resources (i.e., plant food sources) that attract females. We also predicted that males in high density populations should exhibit male dominance polygyny because exclusive defense of resources or females would become too costly in high density populations (Berry, 1974; Prieto and Ryan, 1978; Ryan, 1982); that is, home ranges of two or more males would consistently overlap.

METHODS

Study organism and sites

Chuckwallas are large (snout-vent length 137-211 mm), herbivorous lizards that are strictly associated with rock outcrops throughout the Mojave and Sonoran deserts of North America. Their unique means of evading predators appears to influence habitat association; chuckwallas retreat into rock crevices and inflate their bodies for long time periods when harassed. Hence, areas that provide adequate food (i.e., plants) and refugia (i.e., rock crevices) are centers of social activity. Social interactions (i.e., aggressive interactions or mating behavior) occur almost exclusively during spring and early summer, from mid-March to late June (personal observations; Berry, 1974; Johnson, 1965).

We established study sites in rocky areas for three chuckwalla populations near Phoenix, Arizona, USA: Santan Mountains, South Mountains, and Phoenix Mountains. Much of the surface area in the South Mountains is covered by igneous boulders composed of granodiorite which are uniquely foliated due to movement along a fault zone during cooling (Reynolds, 1985). Consequently, a high concentration of suitable crevices can serve as refugia. The Santan Mountains study site is also primarily composed of granodiorite, but the boulders are not as extensive or foliated as in the South Mountains. The Phoenix Mountains are primarily intrusive basalt (Shank, 1973), which has a fine grained, dense structure and is even less foliated than the Santan Mountains. Study sites varied in area from 10.8 hectares (ha) at Phoenix Mountains, 5.8 ha at the Santan Mountains, and 1.8 ha in the South Mountains. This variation in size was primarily a factor of acquiring enough lizards for radiotracking (see below).

Population density

Population density estimates were calculated using mark-recapture techniques (Sullivan and Flowers, 1998). Study sites were surveyed during the spring and early summer, from early March to late June. Any chuckwalla captured on the study site was given a unique, permanent identification by toe clipping. Toe clipping has minimal effects on lizard species that regularly lose appendages in a natural setting (Hudson, 1996), even in species that climb vertical surfaces (Paulissen and Meyer, 2000). In chuckwallas, individuals often lose appendages naturally (Berry, 1974; personal observations). If more than one toe had to be removed, no more than one toe per limb was clipped. No chuckwalla had more than three toes clipped, and only two males in the South Mountains had more than two toes clipped. Individuals were often observed exhibiting normal basking and displaying behaviors soon after toe clipping. Individuals were also given a paint mark on the tip of the tail to facilitate identification without having to recapture the lizard. All marking and radio tracking (see below) techniques were approved by the Institutional Animal Care and Use Committee at Arizona State University.

The number of recaptures on any given day was typically below eight, at which point density estimates may be biased (Sutherland, 1996). Hence, recapture data from a year were pooled and recapture rates were compared across years (Marvin, 1996). Mark-recapture data from two of the study sites (Phoenix and South Mountains) were collected over a 4 year period (spring of 1995-fall of 1998). For these two populations, estimates of population size were calculated using the Schnabel method, which is appropriate for recapture data collected on several occasions (Sutherland, 1996). At the third study site, the Santan Mountains, animals were marked in 1998 and recaptured in 1999; hence, recapture data for this population were collected in a single period. Accordingly, estimates of population size for the Santan Mountains were calculated using the Petersen method (Sutherland, 1996). Population size estimates were divided by the area of the study site to obtain population density estimates.

Home range

Home ranges of individual chuckwallas were determined using two techniques. First, some recapture data were collected using the permanently marked individuals described above. However, the majority of home range data were collected using radio telemetry.

Chuckwallas from the South Mountains (14 males, 3 females), Santan Mountains (12 males, 5 females) and Phoenix Mountains (8 males, 8 females; Figure 1) were fitted with an external radio transmitter (4.5 g, AVM Instrument Co., Ltd., Livermore, CA, modified model SM1-H, 164 MHz) held in place over their pelvic region with waxed nylon thread. Chafing occurred in three animals at the junction of the rear thigh with the torso during the first 3 days of tracking; however, this was remedied by loosening the ties. The transmitters rarely fell off the animals until the thread holding them wore through. This typically took 2-3 weeks, however, allowing multiple relocations of the chuckwallas. Although chuckwallas constantly enter and emerge from crevices, the transmitters were relatively flat (5 mm), and did not prevent movement in and out of most refugia.

Figure 1

Correlation between plant score and the number of females found on male territories in a low density (Phoenix Mountain), moderate density (Santan Mountains), and high density population (South Mountains).

Figure 1

Correlation between plant score and the number of females found on male territories in a low density (Phoenix Mountain), moderate density (Santan Mountains), and high density population (South Mountains).

Chuckwallas were tracked between early March and late June, in 1998 and 1999. We attempted to relocate each lizard at least 12 times, although transmitters often fell off the animals before 12 relocations could be obtained (range of 6-25 relocations). However, incremental area analysis, where home range size is calculated after each successive relocation (see software description below), revealed that a mean of six relocations was necessary to describe 90% of an individual's home range. Because accurate estimates of home range could be determined quickly, transmitters were often used on more than one individual when either the transmitter was dropped and the lizard could not be recaptured, or when the number of relocations exceeded twelve (the mean number of relocations where home range size ceased to increase was 10) and the lizard could be caught. If the lizard could not be caught, we continued to track the animal until it was recaptured. To ensure that we detected any shifts in home range after a transmitter had been removed or fell off, we continued to monitor whether chuckwallas were within their home range by visual observations and hand recaptures. Locality data was collected at 1-3 day intervals throughout the study period.

All home range analyses were done using Ranges V home range software (Institute of Terrestrial Ecology, Wareham, Dorset, UK). Home range size was determined using the convex polygon method with the outer (100%) edges reported, which is most comparable to previous studies on chuckwalla home range (Abts, 1985; Berry, 1974; Johnson, 1965). Calculating the area of a polygon requires grid coordinates (X and Y) for each lizard locality. We obtained X,Y coordinates by placing transparent grids over aerial photographs (Landiscor, Phoenix, Arizona, USA) of each study site. Although this method did not take into account changes in elevation, it was quite accurate otherwise because even individual small rocks that were chuckwalla refugia could be easily identified. The photographs were scaled by measuring the distance between two objects in the field that were readily identifiable in the photographs. The scales for the three photographs were: Phoenix Mountains, 1:680, Santan Mountains, 1:320, and South Mountains, 1:610. Incremental area analysis was used to determine when home range size ceased to increase. The Ranges V software was used to calculate an overlap matrix showing the percent that home ranges overlapped among all individuals. From the overlap matrix, we calculated the mean percent of male and female home ranges that were overlapped by one or more males.

Statistical analyses of home range, behavioral (see below), and resource data (see below) were conducted using SYSTAT 9.0. When parametric assumptions could not be met, non-parametric analyses were used.

Behavioral observations

Scan sampling observations were made during each visit to a study site for radiotracking to determine whether males exhibited behavior that would imply dominant-subordinate interactions. Each male's territory was searched for other chuckwallas by first scanning the area from a distance with binoculars or a spotting scope before radio tracking. Scanning bouts lasted 10-30 min and were done upon arrival to the study site and periodically during tracking. If behavioral interactions between two or more chuckwallas were observed, the behavior was recorded (see below) until all individuals returned to basking. We defined dominant-subordinate interactions as any behavioral interaction between two or more males with overlapping home ranges that was clearly not an attempt at displacement. For example, if male home ranges overlapped and (1) males were observed on the same boulder pile at the same time or (2) a male clearly tolerated another male in close proximity, without chasing or fighting, yet one exhibited assertion displays (see below; Berry, 1974) and the other subordinate (see Berry, 1974 for description), the interaction was defined as dominant-subordinate or despotism (sensu Emlen and Oring, 1977; Evans, 1951; Maher and Lott, 1995; Wilson, 1975). We also required that dominant-subordinate behaviors be maintained regardless of location (e.g., Evans, 1951); site-specific behavior would be an indication of territoriality. Males were considered territorial if at least one of two conditions were met. First, males were territorial if they exhibited exclusive use of an area; that is, there was little home range overlap among males (an ecological definition sensu Maher and Lott, 1995). We selected a maximum value of 25% home range overlap as a criterion for territoriality, which typically is the most overlap observed in territorial males (e.g., Maher and Lott, 1995; Marvin, 1998). Second, males were territorial if they exhibited site defense (a behavioral definition sensu Maher and Lott, 1995; see description below). Under a behavioral definition, a male may intrude onto another male's territory, giving the impression of home range overlap. However, if an intruder was consistently chased off the territory by the resident upon detection, the resident was considered territorial.

After the resident male had been located by radiotelemetry, each male's territory was searched each day, as thoroughly as possible, by examining crevices available within the territory for other chuckwallas. Data from focal observations (30 min/individual) conducted for a simultaneous study on behaviour were also used to quantify behavior in the three populations. We quantified aggressive interactions following Berry (1974), which included: (1) assertion display: dewlap withdrawn, body partially inflated, push-ups and head-bobbing; (2) threat display: dewlap extended, body inflated, high amplitude head-bobbing; (3) challenge display: dewlap fully extended, body fully inflated, back arched, circling, open-mouthed charging; (4) chasing (may follow threat or challenge display); and (5) attacks: includes biting, follows challenge display (Berry, 1974). The number of hours of behavioral observations for the three sites were 360 h over 4 years for Phoenix Mountains, 112 h over 2 years for the Santan Mountains, and 247 h over 4 years for the South Mountains.

Plant resources

Once a male's home range was determined, plant resources within that area were quantified by counting the number of plants known to be preferred by chuckwallas (from now on referred to as “plant number”). Plants were identified as “preferred” based on feeding observations taken during focal observations of chuckwallas from March to late June in which 30 feeding bouts were recorded, most of which were observed in the South Mountains. A more accurate representation of plant resources on male territories would take into account feeding preferences that females may exhibit. Accordingly, the “plant score” for a male's home range was calculated as Σ f Pi, where P is the number of plants of species i, and f is the relative frequency for which chuckwallas were observed feeding on species i. Hence, plant species scores were weighted according to how often chuckwallas were observed feeding on them. Some of the most abundant and widely distributed plant species were never eaten (see Results below), and, hence, were not counted. Whether plant number or the adjusted plant score were correlated with (1) number of females within male home ranges and (2) male home range area was analyzed using Spearman rank correlation. Because most feeding bouts were observed in the South Mountains, there is potential bias in the plant score given that there was some variation in plant numbers among the three sites (see below). However, this would be a conservative bias because the plants would potentially be ranked inappropriately in the other two sites. Hence, the likelihood of a positive correlation between plant score and females on male territories would decrease for the other two sites.

RESULTS

Population density and home range

Chuckwalla density varied considerably among the three populations. Densities in the Phoenix Mountains were lowest at approximately three animals per ha (Table 1), followed by the Santan Mountains at 11 animals/ha and the South Mountains with 65 animals/ha, which is higher than any other chuckwalla population reported (Table 1). The 95% confidence intervals (CI) did not overlap despite the large range in CI for the South Mountains' estimate (Table 1). There was a negative association between density and home range size with Phoenix Mountains chuckwallas having the largest mean home range, followed by the Santan Mountains, and the South Mountains with a mean home range size approximately 10 times smaller than the Phoenix Mountains (Table 1).

Table 1

Population density, home range size, and mating system structure for various chuckwalla populations

Location   Chuckwalla density (chuckwallas/ha)   Home range (ha) mean ± SE   Mating system   Reference  
Not all population density and home range size estimates were calculated in the same manner. Populations are arranged by increasing density.  
Phoenix Mountains, Arizona   2.7   Male: 0.51 ± 0.12 (n = 8)   Territorial   This study  
  (95% CI = 2.0-4.3)   Female: 0.18 ± 0.10 (n = 8)    
Red Rock Canyon, California   7.1   Male: 0.57   Territorial   Johnson (1965)  
   Female: 0.17    
Black Mountain, California   10   Male: 0.20 ± 0.04   No apparant   Nagy (1973)  
   Female: 0.17 ± 0.06   social structure   
Santan Mountains, Arizona   10.9   Male: 0.08 ± 0.01 (n = 12)   Territorial   This study  
  (95% CI = 7.8-20.3)   Female: 0.02 ± 0.01 (n = 5)    
Lone Butte, California   13.8   Male: 1.90   Dominance   Berry (1974)  
   Female: 0.81   hierarchy   
Chuckwalla Mountains, California   22.8   Male: 0.37 ± 0.08   Dominance   Abts (1985)  
   Female: 0.18 ± 0.02   hierarchy   
South Mountains, Arizona   65   Male: 0.05 ± 0.01 (n = 14)   Territorial   This study  
  (95% CI = 34-551)   Female: 0.02 ± 0.01 (n = 7)    
Location   Chuckwalla density (chuckwallas/ha)   Home range (ha) mean ± SE   Mating system   Reference  
Not all population density and home range size estimates were calculated in the same manner. Populations are arranged by increasing density.  
Phoenix Mountains, Arizona   2.7   Male: 0.51 ± 0.12 (n = 8)   Territorial   This study  
  (95% CI = 2.0-4.3)   Female: 0.18 ± 0.10 (n = 8)    
Red Rock Canyon, California   7.1   Male: 0.57   Territorial   Johnson (1965)  
   Female: 0.17    
Black Mountain, California   10   Male: 0.20 ± 0.04   No apparant   Nagy (1973)  
   Female: 0.17 ± 0.06   social structure   
Santan Mountains, Arizona   10.9   Male: 0.08 ± 0.01 (n = 12)   Territorial   This study  
  (95% CI = 7.8-20.3)   Female: 0.02 ± 0.01 (n = 5)    
Lone Butte, California   13.8   Male: 1.90   Dominance   Berry (1974)  
   Female: 0.81   hierarchy   
Chuckwalla Mountains, California   22.8   Male: 0.37 ± 0.08   Dominance   Abts (1985)  
   Female: 0.18 ± 0.02   hierarchy   
South Mountains, Arizona   65   Male: 0.05 ± 0.01 (n = 14)   Territorial   This study  
  (95% CI = 34-551)   Female: 0.02 ± 0.01 (n = 7)    

Male home ranges were at least twice as large as those of females in all three populations (Table 1). Females in the Phoenix Mountains had a mean home range almost 10 times larger than females in both the Santan and South Mountains (Table 1); however, differences among the three populations were not statistically significant (Kruskal-Wallis test, H2,20 = 5.258, p =.07). Males in the Phoenix Mountains had a mean home range approximately six times larger than males in the Santan Mountains, and 10 times larger than males in the South Mountains (Table 1). Male home ranges were significantly different among populations (Kruskal-Wallis test, H2,30 = 16.63, p =.003). Home ranges of males from the Santan and South Mountains were very small (often a single boulder pile) and similar in size, despite dramatic differences in lizard density. Home range sizes were not significantly different (Mann-Whitney, U = 24.0, p =.1) in these two areas, but home ranges in both the Santan (U = 4.5, p =.006) and South Mountains (U = 1.0, p =.001) were significantly smaller than in the Phoenix Mountains.

There was typically little to no overlap in male home ranges at any site (Table 2) and the mean percent of a male's home range that was overlapped by one or more males did not vary among sites (Kruskal-Wallis test, H2,29 = 4.01, p = 0.135). Even in the South Mountains, despite high population density, males maintained small, non-overlapping territories. Contrary to our predictions, overlap was highest in the Phoenix Mountains, where density was lowest. The few cases in which there was considerable overlap between male home ranges were the result of temporal differences in home range occupancy that occurred when one male disappeared and his neighbor(s) began defending the vacant area. For example, in the South Mountains, one male disappeared in the middle of the mating season, and his territory was divided up by three neighboring males. In both the Phoenix and Santan Mountains, what appeared to be relatively large home range overlap between two males was actually the result of territory shifts where one male “took over” part of another male's territory. In the Phoenix Mountains, where overlap of male home ranges was the largest, an immature male appeared to overlap the territory of another male. However, this was an effect of the immature male moving to different sides of the other male's territory along its borders. Calculating the convex polygon for the immature male resulted in an inaccurate representation of his movement patterns.

Table 2

Mean (± SE) percent of male and female chuckwalla home ranges that were overlapped by another male

Location   Males overlapped by males   Famales overlapped by males   Females overlapped by females  
Values for male—male overlap are liberal because the few cases (n = 1 in both the Phoenix and Santan Mountains) of what appeared to be floater males are included. However, effects of territory take-over (e.g., when a male disappeared and his territory was divided up among his neighbors) have been removed.  
Phoenix Mountains   22.5 ± 10.8 (n = 8)   83.7 ± 10.2 (n = 6)   12.0 ± 10.5 (n = 6)  
Santan Mountains   7.8 ± 6.1 (n = 12)   83.6 ± 7.1 (n = 5)   16.1 ± 14.8 (n = 5)  
South Mountains   11.0 ± 3.8 (n = 14)   89.0 ± 5.6 (n = 5)   15.6 ± 10.2 (n = 5)  
Location   Males overlapped by males   Famales overlapped by males   Females overlapped by females  
Values for male—male overlap are liberal because the few cases (n = 1 in both the Phoenix and Santan Mountains) of what appeared to be floater males are included. However, effects of territory take-over (e.g., when a male disappeared and his territory was divided up among his neighbors) have been removed.  
Phoenix Mountains   22.5 ± 10.8 (n = 8)   83.7 ± 10.2 (n = 6)   12.0 ± 10.5 (n = 6)  
Santan Mountains   7.8 ± 6.1 (n = 12)   83.6 ± 7.1 (n = 5)   16.1 ± 14.8 (n = 5)  
South Mountains   11.0 ± 3.8 (n = 14)   89.0 ± 5.6 (n = 5)   15.6 ± 10.2 (n = 5)  

The entire home range of most females was overlapped by a male home range and often was overlapped by another female (Table 2). Home range overlap involving females was never the effect of temporal or spatial inaccuracies as found with the few males described above. There were no significant differences among sites in the mean percent of a female's home range overlapped by a male (Kruskal-Wallis test, H2,16 = 0.68, p =.710) or female (H2,16 = 0.149, p =.928).

Behavioral observations

Male behavior was similar in all three populations and was consistent with territory defense polygyny (i.e., defending a site from other males while allowing overlap with multiple females). Male displays were frequently observed in the Santan and South Mountains, but did not typically proceed beyond “assertion” head bobbing (sensu Berry, 1974), although one fight between two males occurred in the South Mountains population. Males displayed at a rate (mean ± SE) of 5.8 ± 2.9 displays/h (n = 6) in the Phoenix Mountains, 6.1 ± 1.6 (n = 6) in the Santan Mountains, and 19.1 ± 5.0 (n = 13) in the South Mountains. Although display rates were higher in the high density population, there were no significant differences among populations (Kruskal-Wallis test, H2,17 = 5.166, p =.08).

No male chuckwalla was observed in another male's territory, let alone on the same boulder pile at the same time, and no dominant-subordinate interactions were ever observed. In contrast, females were consistently found in male territories and males and females were typically seen on the same boulder pile (often the same boulder). In the Santan (13.3% of observations) and South Mountains (18.8% of observations), two or three females sometimes were observed basking on the same boulder pile simultaneously, or were found together in the same crevice. Behavioral interactions that involved close contact between males and females were seen often in the form of mating displays (sensu Berry, 1974), or when males and females basked on the same spot, with the female lying on the male. The number of females within a male's territory varied within populations, but not across populations (Kruskal-Wallis test, H = 1.034, p =.595). The mean number of females on a male's territory in the three populations was 1.7 for the Phoenix Mountains (range = 0-5), 1.1 for the Santan Mountains (range = 0-3), and 1.1 for the South Mountains (range = 0-3).

Plant resources

Chuckwallas were observed feeding on eight perennial plant species (Table 3), all of which exhibited a relatively patchy distribution (see Discussion below). Surprisingly, no feedings were observed of the most abundant plant species that were found throughout the study sites (i.e., Ambrosia deltoidea and Encelia farinosa), suggesting that chuckwallas are selective regarding which plants are consumed. However, the extreme abundance of some uneaten species (A. deltoidea) precluded quantification of their abundance within some home ranges; hence, selectivity indices could not be calculated. Yet there was no correlation between male territory size and either plant numbers or adjusted plant scores at any of the three study sites (Table 4), indicating that plant resources were patchy. If plants were distributed evenly, plant resources should have increased linearly with increasing territory size.

Table 3

Plant assessments for species eaten by chuckwallas during behavioral observations

  Plants/ha  
Plant species   South   Santan   Phoenix  
Numbers in parentheses represent the relative frequency that chuckwallas were observed feeding on that plant species.  
Cercidium microphyllum (0.30)   70.0   69.2   43.4  
Sphaeralcea ambigua (0.23)   122.8   195.4   15.6  
Trixis californica/Viguiera deltoidea (0.20)   119.3   4.6   9.1  
Fouquieria splendens (0.13)   8.8   3.1   0  
Hyptis emoryi (0.07)   22.8   52.3   25.4  
Lycium sp. (0.07)   35.1   23.1   24.8  
Total plants   378.8   347.7   118.3  
  Plants/ha  
Plant species   South   Santan   Phoenix  
Numbers in parentheses represent the relative frequency that chuckwallas were observed feeding on that plant species.  
Cercidium microphyllum (0.30)   70.0   69.2   43.4  
Sphaeralcea ambigua (0.23)   122.8   195.4   15.6  
Trixis californica/Viguiera deltoidea (0.20)   119.3   4.6   9.1  
Fouquieria splendens (0.13)   8.8   3.1   0  
Hyptis emoryi (0.07)   22.8   52.3   25.4  
Lycium sp. (0.07)   35.1   23.1   24.8  
Total plants   378.8   347.7   118.3  

Table 4

Spearman rank correlations of male territory characteristics

Correlation   Population  N  Spearman's rho  p 
Probability values were adjusted using the sequential Bonferroni test (Rice, 1989): ** indicates p values that were significant after Bonferroni adjustment; * indicates p values that were not significant after Bonferroni adjustment but were close to significant.  
Male territory area versus plant number   Phoenix   7   0.179   0.702  
  Santan   9   0.250   0.516  
  South   13   0.186   0.562  
Male territory area versus adjusted plant score   Phoenix   7   0.179   0.702  
  Santan   9   0.267   0.488  
  South   13   0.148   0.645  
Male territory area versus female number   Phoenix   7   0.148   0.751  
  Santan   9   0.561   0.19  
  South   13   0.035   0.909  
Plant number versus female number   Phoenix   7   0.945   0.001** 
  Santan   9   0.823   0.006* 
  South   13   0.884   0.0001** 
Adjusted plant score versus female number   Phoenix   7   0.945   0.001** 
  Santan   9   0.832   0.005* 
  South   13   0.886   0.0001** 
Correlation   Population  N  Spearman's rho  p 
Probability values were adjusted using the sequential Bonferroni test (Rice, 1989): ** indicates p values that were significant after Bonferroni adjustment; * indicates p values that were not significant after Bonferroni adjustment but were close to significant.  
Male territory area versus plant number   Phoenix   7   0.179   0.702  
  Santan   9   0.250   0.516  
  South   13   0.186   0.562  
Male territory area versus adjusted plant score   Phoenix   7   0.179   0.702  
  Santan   9   0.267   0.488  
  South   13   0.148   0.645  
Male territory area versus female number   Phoenix   7   0.148   0.751  
  Santan   9   0.561   0.19  
  South   13   0.035   0.909  
Plant number versus female number   Phoenix   7   0.945   0.001** 
  Santan   9   0.823   0.006* 
  South   13   0.884   0.0001** 
Adjusted plant score versus female number   Phoenix   7   0.945   0.001** 
  Santan   9   0.832   0.005* 
  South   13   0.886   0.0001** 

All feeding bouts took place within male territories. There was variation in territory plant scores both within and among populations; male territory plant scores were highest in the Phoenix Mountains (mean ± SD = 11.46 ± 7.90), followed by the Santan Mountains (6.75 ± 2.90) and the South Mountains (3.73 ± 1.95). These population differences were statistically different (ANOVA, F2,26 = 7.37, p =.003); Bonferroni post-hoc tests revealed that the South and Santan Mountains were not different from each other (p =.75), but that both were different from the Phoenix Mountains (p =.001 and.035, respectively). Despite small sample sizes, the number of females on male territories was positively correlated with both plant number and adjusted plant scores in all three populations (Figure 1 and Table 4); hence, plants were likely an important resource to females. This correlation was not simply the result of increases in male territory size as there was no correlation between male territory size and number of females in any population (Table 4).

As noted above, males patrolled larger territories in the Phoenix and Santan Mountains. When controlling for area, plant scores followed an opposite trend to that noted above with the South Mountains highest (mean ± SD = 90.6 ± 66.5/ha), followed by the Santan Mountains (85.6 ± 50.9/ha) and the Phoenix Mountains (35 ± 35.9/ha), although these differences were not statistically different (F2,26 = 2.046, p =.15). However, the pattern suggests that Phoenix Mountains plant resources were approximately one third as abundant as those in the Santan and South Mountains. This may explain why absolute plant numbers (Table 3) for male territories were not as high for the Phoenix Mountains, even though male territories were approximately 10 times larger than those in the South Mountains (see above; mean plant scores should have been 10 times higher, roughly = 37.3, rather than 11.5).

DISCUSSION

Despite wide variation in density among three chuckwalla populations in the Sonoran Desert, males in all populations exhibited strict territoriality. As population density increased, male territory size decreased and varied by as much as a factor of 10; extremely small territories were observed in the high density population. Population density did not appear to influence polygyny levels since the mean number of females per male territory did not differ among the three sites. Consequently, male territory size appears influenced by tradeoffs that maximize the number of females in the territory and minimize territory defense costs associated with population density. In the Santan Mountains, where female home range size and plant scores per ha were equal to the South Mountains, male territories were slightly larger (albeit not significantly) than in the South Mountains. Hence, it may be that the extremely high population density in the South Mountains limits male territory size, as has been found in some birds (e.g., Myers et al., 1979).

The strict territoriality exhibited by chuckwallas under high density conditions in our study contradicts a pattern commonly seen in numerous taxa, including lizards, in which males shift from territoriality to some other form of mating system as population density increases (Brattstrom, 1974; Clutton-Brock et al., 1993; Davies, 1991; Dugan and Wiewandt, 1982; Höglund and Alatolo, 1995; Langbein and Thirgood, 1989; Ryan, 1982; Stamps, 1977, 1983; Sullivan, 1989; Travis et al., 1995). However, maintenance of territories under high densities, similar to the chuckwallas of this study, has been observed in other organisms (Maher and Lott, 2000; Ruby and Dunham, 1987; Wilson, 1975). For example, hermit thrushes at higher densities defend smaller territories, without forming dominance hierarchies (Brown et al., 2000). Similarly, bannertail kangaroo rat (Dipodomys spectabilis) social structure remains stable despite variation in population density; males maintain territoriality in both low and high density populations (Randall, 1984). The maintenance of territorial polygyny under high density conditions may be possible if resources important to females are not uniformly distributed. Clumped resources will facilitate the defense of solitary females if they remain at high resource patches and do not travel frequently to other patches. Additionally, clumped resources favor the formation of female groups (Travis and Slobodchikoff, 1993; Travis et al., 1995). If females form groups within a small area because resources are concentrated there, defense of those females may not be too costly for a male even if population density is high (Davies and Houston, 1984).

In chuckwallas, females likely remain around patches of refugia (i.e., crevices) and food resources (i.e., plants) regardless of whether females are solitary or in groups. The plant species that chuckwallas consumed exhibit a patchy distribution in the Upland Sonoran Desert habitat characteristic of our study sites (Brown, 1982). Further evidence of the patchy distribution of these plants species was demonstrated by the lack of relationship between male territory size and plant resources within populations. Despite the patchiness of resources in all populations, plant availability apparently influenced female chuckwalla home range size. Female home range size was the same in the Santan and South Mountains where plant resources per ha were equivalent. Plant scores per ha in the Phoenix Mountains were approximately one third that of the Santan and South Mountains and female home range size was considerably larger. Hence, female spacing is centered around clumps of resources, but resource availability of each clump varies among populations. In areas of richer resource clumps (i.e., the Santan and South Mountains), females had smaller home ranges. Female home range size, in turn, appears to have influenced male territory size. Male territories were much larger in the Phoenix Mountains, yet did not exceed polygyny levels of the Santan and South Mountains. In the Santan and South Mountains, male territories were associated primarily with granodiorite boulders in close proximity to preferred plant resources. Hence, the patchy nature of the rich resources and, therefore, females, has apparently allowed male chuckwallas to maintain territoriality even under high densities.

Without conducting female removal experiments (e.g., M'Closkey et al., 1987a, b), it is unclear whether male chuckwallas defend females, resources important to females, or both. However, our data on plant distributions suggest that male territoriality may be related to resource defense. In contrast, Berry (1974) suggested plant resources did not vary among male territories in Mojave chuckwallas; however, Berry (1974) did not present data to support such a conclusion. Ryan (1982) concluded that “ resource defense is probably a consequence of (male) territoriality and has had little importance in the evolution of this behavior.” Ryan was influenced by Nagy's (1973) findings that chuckwallas did not increase their home range size as food supplies decreased during a drought, which would be expected if territoriality was based on plant resource defense. However, we found that home range sizes of males and females were similar in the Santan and South Mountains, which had the same plant scores when controlling for area. Plant resources in the Phoenix Mountains were much scarcer and males and females had much larger home ranges, a result consistent with resource defense territoriality (Simon, 1975). Alternatively, resource defense still may be a consequence of female defense in the three populations we studied. If female home ranges are small, or if females exhibit a clumped distribution, then males may not need to defend large territories (Stamps, 1983). In the Santan and South Mountains, females had home ranges that were approximately ten fold smaller than females in the Phoenix Mountains; male home ranges may be small in the Santan and South Mountains because a large territory is not needed to defend females clumped in a small area. Indeed, female distribution is expected to be influenced by resources while male distribution should be a function of female distribution (Davies, 1991; Emlen and Oring, 1977).

This work was supported in part by a Heritage Fund grant from the Arizona Department of Game and Fish. We thank J. Alcock, D. DeNardo, M. Grober, R. Rutowski, and four anonymous reviewers for critically reading the manuscript, and R. Bowker, E. Stitt, and T. Tuchak for help in the field.

References

Abts ML,
1985
.
The life history strategy of the saxicolous desert lizard, Sauromalus obesus (PhD dissertation)
 . Portland, Oregon: Portland State University.
Banks MJ, Thompson DJ,
1985
. Lifetime mating success in the damselfly Coenagrion puella.
Anim Behav
 
33
:
1175
-1183.
Berry KH,
1974
. The ecology and social behavior of the chuckwalla, Sauromalus obesus obesus Baird.
Univ Calif Publ Zool
 
101
:
1
-60.
Brattstrom BH,
1974
. The evolution of reptilian social behavior.
Am Zool
 
14
:
35
-49.
Brown DE,
1982
. Biotic communities of the American southwest—United States and Mexico.
Desert Plants
 
4
:
1
-342.
Brown DR, Stouffer PC, Strong CM,
2000
. Movement and territoriality of wintering hermit thrushes in southeastern Louisiana.
Wilson Bull
 
112
:
347
-353.
Brown JL,
1964
. The evolution of diversity in avian territorial systems.
Wilson Bull
 
76
:
160
-169.
Clutton-Brock TH, Deutsch JC, Nefdt RJC,
1993
. The evolution of ungulate leks.
Anim Behav
 
46
:
1121
-1138.
Davies NB,
1991
. Mating systems. In:
Behavioural ecology: an evolutionary approach
  , 3rd ed (Krebs JR, Davies NB, eds). Oxford: Blackwell.
Davies NB, Houston AI,
1984
. Territory economics. In:
Behavioural ecology: an evolutionary approach
  , 2nd ed (Krebs JR, Davies NB, eds). Oxford: Blackwell.
Dugan B, Wiewandt TV,
1982
. Socio-ecological determinants of mating strategies in iguanine lizards. In:
Iguanas of the world: their behavior, ecology, and conservation
  (Burghardt GM, Rand AS, eds). Park Ridge, New Jersey: Noyes.
Emlen ST, Oring LW,
1977
. Ecology, sexual selection, and the evolution of mating systems.
Science
 
197
:
215
-223.
Evans LT,
1951
. Field study of the social behavior of the black lizard, Ctenosaura pectinata.
Am Mus Novit
 
1493
:
1
-26.
Höglund J, Alatalo RV,
1995
.
Leks
 . Princeton, New Jersey: Princeton University Press.
Hollingsworth BD,
1998
. The systematics of chuckwallas (Sauromalus) with a phylogenetic analysis of other iguanid lizards.
Herpetol Mono
 
12
:
38
-191.
Hudson S,
1996
. Natural toe loss in southeastern Australian skinks: implications for marking lizards by toe-clipping.
J Herpetol
 
30
:
106
-110.
Johnson SR,
1965
. An ecological study of the chuckwalla, Sauromalus obesus Baird, in the western Mojave Desert.
Am Midl Nat
 
73
:
1
-29.
Kaiser H,
1985
. Availability of receptive females at the mating place and mating chances of males in the dragonfly Aeschna cyanea.
Behav Ecol Sociobiol
 
18
:
1
-7
Langbein J, Thirgood SJ,
1989
. Variation in mating systems of fallow deer (Dama dama) in relation to ecology.
Ethology
 
83
:
195
-214.
Maher CR, Lott DF,
1995
. Definitions of territoriality used in the study of variation in vertebrate spacing systems.
Anim Behav
 
49
:
1581
-1597.
Maher CR, Lott DF,
2000
. A review of ecological determinants of territoriality within vertebrate species.
Am Midl Nat
 
143
:
1
-29.
Marvin GA,
1996
. Life history and population characteristics of the salamander Plethodon kentucki with a review of Plethodon life histories.
Am Midl Nat
 
136
:
385
-400.
Marvin GA,
1998
. Territorial behavior of the plethodontid salamander Plethodon kentucki: influence of habitat structure and population density.
Oecologia
 
114
:
133
-144.
M'Closkey RT, Baia KA, Russell RW,
1987
. Defense of mates: a territory departure rule for male tree lizards following sex-ratio manipulation.
Oecologia
 
73
:
28
-31.
M'Closkey RT, Baia KA, Russell RW,
1987
. Tree lizard (Urosaurus ornatus) territories: experimental perturbation of the sex ratio.
Ecology
 
68
:
2059
-2062.
Myers JP, Conners PG, Pitelka FA,
1979
. Territory size in wintering sanderlings: the effects of prey abundance and intruder density.
Auk
 
96
:
551
-561.
Nagy KA,
1973
. Behavior, diet and reproduction in a desert lizard, Sauromalus obesus.
Copeia
 
1977
:
93
-102.
Paulissen MA, Meyer HA,
2000
. The effect of toe-clipping on the gecko Hemidactylus turcicus.
J Herpetol
 
34
:
282
-285.
Prieto AA, Ryan MJ,
1978
. Some observations of the social behavior of the Arizona chuckwalla, Sauromalus obesus tumidus (Reptilia, Lacerta, Iguanidae).
J Herpetol
 
12
:
327
-336.
Randall JA,
1984
. Territorial defense and advertisement by footdrumming in bannertail kangaroo rats (Dipodomys spectabilis) at high and low population densities.
Behav Ecol Sociobiol
 
16
:
11
-20.
Reynolds SJ,
1985
. Geologic history of the South Mountains.
Arizona Bur Geo Min Tech Bull
 
195
.
Rice WR,
1989
. Analyzing tables of statistical tests.
Evolution
 
43
:
223
-225.
Ruby DE, Dunham AE,
1987
. Variation in home range size along an elevational gradient in the iguanid lizard Sceloporus merriami.
Oecologia
 
71
:
473
-480.
Ryan MJ,
1982
. Variation in iguanine social organization: mating systems in chuckawallas (Sauromalus). In:
Iguanas of the world: their behavior, ecology, and conservation
  (Burghardt GM, Rand AS, eds). Park Ridge, New Jersey: Noyes.
Shank DC,
1973
.
Environmental geology in the Phoenix Mountains
 , Maricopa County, Arizona (MS thesis). Tempe, Arizona: Department of Geology, Arizona State University.
Sherman KJ,
1983
. The adaptive significance of post-copulatory mate guarding in a dragonfly Pachydiplax longipennis.
Anim Behav
 
31
:
1107
-1115.
Simon CA,
1975
. The influence of food abundance on territory size in the igaunid lizard, Sceloporus jarrovi.
Ecology
 
56
:
993
-998.
Stamps JA,
1977
. Social behavior and spacing patterns in lizards. In:
Biology of the Reptilia
 , vol.
7
(Gans C, Tinkle DW, eds). London: Academic Press.
Stamps JA,
1983
. Sexual selection, sexual dimorphism, and territoriality. In:
Lizard ecology: studies of a model organism
  (Huey RB, Pianka ER, Schoener TW, eds). Cambridge, Massachusetts: Harvard University Press;
169
-204.
Sullivan BK,
1989
. Mating system variation in Woodhouse's toad (Bufo woodhousii).
Ethology
 
83
:
60
-68.
Sullivan BK, Flowers M,
1998
. Large iguanid lizards of urban mountain preserves in northern Phoenix, Arizona.
Herp Nat Hist
 
6
:
13
-22.
Sutherland WJ,
1996
.
Ecological census techniques: a handbook
 . Cambridge: Cambridge University Press.
Travis SE, Slobodchikoff CN,
1993
. Effects of food resource distribution on the social system of Gunnison's prairie dog (Cynomys gunnisoni).
Can J Zool
 
71
:
1186
-1192.
Travis SE, Slobodchikoff CN, Keim P,
1995
. Ecological and demographic effects on intraspecific variation in the social system of prairie dogs.
Ecology
 
76
:
1794
-1803.
Vehrencamp SL, Bradbury JW,
1984
. Mating systems and ecology. In:
Behavioural ecology: an evolutionary approach
 , 2nd ed (Krebs JR, Davies NB, eds). Oxford: Blackwell.
Warner RR, Hoffman SG,
1980
. Local population size as a determinant of mating system and sexual composition in two tropical marine fishes (Thalassoma spp.).
Evolution
 
34
:
508
-518.
Warner RR, Hoffman SG,
1980
. Population density and the economics of territorial defense in a coral reef fish.
Ecology
 
61
:
772
-780.
Werner DI,
1982
. Social organization and ecology of land iguanas, Conolophus subcristatus, on Isla Fernandina, Galápagos. In:
Iguanas of the world: their behavior, ecology, and conservation
  (Burghardt GM, Rand AS, eds). Park Ridge, New Jersey: Noyes.
Wilson EO,
1975
.
Sociobiology
 . Cambridge, Massachusetts: Belknap Press.