Study site

Zamami-jima Island is located 35 km west of Okinawa-jima Island and is 6.7 km² in area. Four native diurnal lizards (Takydromus smaragdinus [Lacertidae], Diploderma polygonatum polygonatum [Agamidae], Plestiodon marginatus [Scincidae], and Ateuchosaurus okinavensis [Scincidae]) occur on the island (Maenosono and Toda, 2007). In 2020, the occurrence of A. carolinensis was limited to the southern part of the island (Furuzamami Area) (NCDDEAO, 2024). This study was conducted at two types of sites: areas where the anoles are present (GAP) at the Furuzamami site, and areas where the anoles are absent (GAA) at the Ama, Asa, and Toumahama sites (Fig. 1). The GAP site comprised secondary forests, coastal forests, meadows, and wastelands. Common tree species in secondary forests included Pinus luchuensis, Ficus erecta, Macaranga tanarius, and Toxicodendron succedaneum, whereas Casuarina sp., Pandanus odorifer, Leucaena leucocephala and Hibiscus tiliaceus were common in coastal forests. Meadows are dominated by the grass Pennisetum purpureum. Similar secondary and coastal forest habitats were found in the GAA sites. In addition, the Ama site featured wetlands with Typha latifolia, along with human houses surrounded by hedges and gardens, whereas the Asa site contains meadows dominated by P. purpureum.

Fig. 1.

A map of Zamami-jima Island showing location of the study sites. Circle with unbroken margins indicates Anolis carolinensis present site (Furuzamami site) and circles with broken margins indicate A. carolinensis absence sites (Ama, Asa, and Toumahama sites).

Table 1.

Census route length and time in each study route.

Route census surveys

Census routes were established along paved and unpaved roads at the study sites. The lengths of these routes were 1,300, 480, 230, and 470 m in the Furuzamami, Ama, Asa, and Toumahama sites, respectively (Table 1). Although a few sections of these routes were covered with canopy of tall trees, they mostly passed through areas with shorter trees, grass land, and patches of Bidens sp.

Census route surveys were conducted on September 20–22, 2021; September 9–10, 2022; and September 29–30, 2022. Two or three persons (authors) conducted surveys during the daytime (0930–1630 h) and nighttime (1930–0100 h), during which we walked along the routes and searched for the anoles and the four native diurnal lizard species both on tracks and in bordering vegetation. The total census times were 29 h 30 min, 5 h 15 min, 1 h 40 min, and 5 h 10 min in the Furuzamami, Ama, Asa, and Toumahama sites, respectively (Table 1). Whenever lizards of any five species were found we recorded species, perch height above the ground, plant species on which the lizard was perched, and perch substrate types (plant parts, ground or artificial substrates, such as glass house stainless frames, plastic buckets, and concrete blocks). Lizards were not captured. The perch height was measured using a digital laser distance meter (Leica DISTOTM X310; Leica, Germany).

Data analysis

Due to small sample sizes in each green anole absent site, data from these sites were combined. Differences in substrate types between GAP and GAA areas for each species were examined using Fisher's exact test. In addition, niche breadth (B) and niche overlap (O_ab) of substrate types were calculated following Pianka (1986) and Pianka (1973), respectively. Each formula is shown below:

where i indicates i-th substrate type number I, Pi indicates the ratio of substrate type number I to the total, and n indicates the total number of substrate types. These scores are relative indexes, so that the absolute baseline was not determined. Differences between species in perch heights above the ground were tested by Tukey's HSD test. All statistical analyses were performed using R version 4.2.1 (R Core Team, 2022).

Microhabitat uses during daytime

All four native diurnal lizards known from Zamami-jima Island were recorded in this study: A. carolinensis (GAP area; day: N=50, night: N=93), T. smaragdinus (GAP area; day: N=54, night: N=30, GAA area; day: N=36, night: N=75), D. p. polygonatum (GAP area; day: N=20, night: N=4, GAA area; day: N=24, night: N=11), P. marginatus (GAP area; day: N=11, GAA area; day: N=4), A. okinavensis (GAP area; day: N=3) (Table 2). Ateuchosaurus okinavensis was not observed in the GAA areas, and neither P. marginatus nor A. okinavensis were observed at night.

During the daytime A. carolinensis utilized plants and artificial steel frames, which were remnants of old greenhouses. Eight plant species were identified as being used by the anoles, with P. odorifer having the highest frequency (Table 2). Anolis carolinensis predominantly used leaves (84.0%), followed by branches and veins (12.0%) (Fig. 2a). The recorded height above the ground for the anoles was 1.12±0.70 (mean±SD) m. Takydromus smaragdinus was the most frequently recorded native species, most often on Bidens sp. in daytime (Table 2). It was frequently observed on leaves (57.4%) and the ground (37.0%) in the GAP area (Fig. 2a), and the niche breadth of the substrate types was 1.386 (Table 3). The plant species (Table 2) and substrate types (Fig. 2a) used by T. smaragdinus at the GAP area were not significantly different from those used by conspecifics at the GAA area (Fisher's exact test; P=0.54), and the niche overlap score for substrate types between GAP T. smaragdinus and GAA T. smaragdinus was high (O_ab=0.980). The height above ground for T. smaragdinus in the GAP area (0.34±0.63 m) was not significantly different from that in the GAA area (0.26±0.34 m) (Tukey's HSD test; P>0.01, Fig. 3a). Diploderma polygonatum polygonatum was frequently observed on tree trunks at both the GAP and GAA areas (GAP area: 60.0%, GAA area: 64.0%, Fisher's exact test; P=0.17, Fig. 2a), and the niche overlap score obtained between the GAP and GAA areas was high (O_ab=0.931), with niche breadths of 2.138 and 2.258 in the GAP and GAA areas, respectively (Table 3). Diploderma p. polygonatum also used dead branches, herb stems, leaves at the GAA area, but not at the GAP area. No significant difference was observed in height above ground between the GAP area (1.22±1.10 m) and the GAA area in this species (0.79±0.62 m) (Tukey's HSD test; P>0.01, Fig. 3a). Plestiodon marginatus and A. okinavensis were only observed on the ground in both areas (Fig. 2a).

Table 2.

Number of observed plant species and parts used by each lizard and skink species during daytime and nighttime. GAP: the green anole present site, GAA: combined data in three green anole absent sites, D: daytime, N: nighttime.

Fig. 2.

Recorded substrate types of the lizards. A: daytime, B: nighttime. Ano: Anolis carolinensis, Tak: Takydromus smaragdinus, Dip: Diploderma polygonatum polygonatum, Ple: Plestiodon marginatus, At: Ateuchosaurus pellopleurus. Numbers in parentheses indicate sample size.

Table 3.

Niche width of microhabitat in each species at day and night.

The niche overlap score between A. carolinensis and T. smaragdinus was the highest for GAP area (O_ab=0.834, Table 4), but the frequency of use for substrate types between these two species was significantly different (Fisher's exact test; P<0.01). The observed height of A. carolinensis was significantly higher than that of T. smaragdinus (Tukey's HSD test; P<0.01), whereas it was not significantly different from that of D. p. polygonatum (Tukey's HSD test; P=0.99) (Fig. 3a). Niche overlap scores between A. carolinensis and T. smaragdinus or D. p. polygonatum in GAP (Table 4) were smaller than those in GAA (Table 5).

Fig. 3.

Perch height above the ground for the five lizard species studied. A: daytime, B: nighttime. Ano: Anolis carolinensis, Tak: Takydromus smaragdinus, Dip: Diploderma polygonatum polygonatum, Ple: Plestiodon marginatus, At: Ateuchosaurus pellopleurus. Numbers in parentheses indicate sample size.

Resting places at nighttime

Plestiodon marginatus and A. okinavensis were not found at night. All lizards found at night did not move when approached. Anolis carolinensis was most frequently found on P. purpureum (Table 2), and 97.9% of resting substrate types was on leaves (Fig. 2b). The niche breadth of the anole was 1.043. The mean resting height above ground for A. carolinensis was 1.38±0.58 m, which was not significantly different from that of T. smaragdinus (1.53±0.58 m) and D. p. polygonatum (2.01±0.71 m) in the GAP area (Tukey's HSD test; P>0.01, Fig. 3b). Takydromus smaragdinus was frequently observed on P. purpureum at the GAA area (49.3%), whereas it was frequently observed on Casuarina spp. (40.0%) instead of P. purpureum (0%) at the GAP area (Table 2). The most common resting substrate types of T. smaragdinus were leaves at both sites (GAP area: 87.5%, GAA area: 81.3%, Fig. 2b), and the niche breadths were 1.280 and 1.484 in the GAP and GAA areas, respectively (Table 3). The resting substrate types of T. smaragdinus did not differ significantly between the GAP and GAA areas (Fisher's exact test; P=0.06), with the niche overlap score of substrate types between those areas was 0.997. The mean resting height above ground for T. smaragdinus did not differ between the GAP and GAA areas (1.26±0.53 m) (Tukey's HSD test; P>0.01, Fig. 3b). The most common resting substrate types of D. p. polygonatum were branches and veins at both the GAP and GAA areas (Fig. 2b) and the niche breadths were 2.000 and 2.373 at the GAP and GAA areas, respectively (Table 3). The resting substrate types of D. p. polygonatum did not differ significantly between the GAP and GAA areas (Fisher's exact test; P=1.00), with the niche overlap score of substrate types between those areas was 0.990. The mean resting height above ground for D. p. polygonatum did not differ between the GAP and GAA areas (1.50±0.55 m) (Tukey's HSD test; P>0.01, Fig. 3b).

Table 4.

Niche overlap score of microhabitat in daytime (upper diagonal) and nighttime (lower diagonal) in the green anole present (GAP) area. NA indicates not able to compare due to one or both species were not observed.

Table 5.

Niche overlap score of microhabitat in daytime (upper diagonal) and nighttime (lower diagonal) in the green anole absent (GAA) area. NA indicates not able to compare due to one or both species were not observed.

The niche overlap scores in the GAP area obtained for A. carolinensis and T. smaragdinus (O_ab=0.993) were high compared with that for A. carolinensis and D. p. polygonatum (O_ab=0.722, Table 4), but the frequency of substrate type differ significantly among the three species (Fisher's exact test; P<0.01). Unlike during the daytime, the nighttime niche overlap of the invasive anole and these two native lizards at the GAP area (Table 4) did not differ from that at the GAA area (Table 5). The frequency of use of particular plant species was different among these three species within the GAP area, and also differed across GAP and GAA areas for the native species (Table 2). The resting heights above the ground did not differ significantly among all the observed lizards (Tukey's HSD test; P>0.01, Fig. 3b).

Table 6.

SVLs based on The Herpetological Society of Japan (2021) and habitat types of lizard species.

Microhabitats use by the lizards during daytime

Habitat selection by a species is determined by its own ecology and by interactions with other species, such as inter-specific competition and resource partitioning (Krebs, 2001). A previous study confirmed that the snout-vent lengths of native T. smaragdinus, D. p. polygonatum, and P. marginatus were similar (Table 6). Although detailed food habits have not been identified, these lizards all primarily consume invertebrates such as insects (The Herpetological Society of Japan, 2021). In the present study, the utilized plant species, parts, and perch heights above ground of native lizards differed in the GAA area (see Table 5). This suggests that these native lizards primarily utilize different microhabitats, allowing them to coexist on the small island, although detailed comparisons of food habits are required.

Anolis carolinensis prefers relatively light and open areas, with its habitat being the trunk-crown (Jenssen et al., 1995). In the present study, this species was frequently observed in places where sunlight penetrated, such as the leaves and branches. This suggests that the introduced green anole utilizes habitats similar to its original habitats in the native region. A possible reason for this is that the anoles are likely to use habitats that native lizards do not fully occupy or as a result of competition with native lizards.

Important factors for coexistence in lizards are segregations in temporal, spatial, trophic, and thermal resources (Pianka, 1986; Watson and Gough, 2012). When segregations in these resources are not enough, alien species may drive away native species (Downes and Bauwens, 2002; Damas-Moreira et al., 2020). In the present study, the plant species used by A. carolinensis differed from those frequently used by native lizards, but the niche overlap score of substrate types between A. carolinensis and T. smaragdinus was relatively high and the scores were near to 1. In addition, the perch height above the ground of A. carolinensis was similar to that of D. p. polygonatum. These results suggest that parts of the green anole microhabitat overlap with those of the native species. Furthermore, the niche overlap scores between the anole and native lizards (T. smaragdinus and D. p. polygonatum) at the GAP area were lower than those at the GAA rea. These results suggest that A. carolinensis has provided negative effects on these native lizards on Zamami-jima Island through competition. Although the results of Fisher's exact test indicated that there are some differences in substrate types between the invasive anole and the native lizards, it is difficult to evaluate at the moment whether the differences are large enough to allow these species to coexist. Anzai et al. (2017) and Mitani et al. (2020) demonstrated that patterns of microhabitat usage in A. carolinensis varied between males and females, and between juveniles and adults in the Ogasawara Islands. Therefore, further studies are required to more precisely evaluate the extent of inter-specific competitions between A. carolinensis and native lizards in particular age classes and or different sexes.

Resting place selection

The mean resting height above the ground of A. carolinensis, T. smaragdinus, and D. p. polygonatum did not differ between the GAP and GAA areas. With respect to the substrate types of the resting places, the niche overlap pattern between A. carolinensis and T. smaragdinus differ between the GAP and GAA areas. Takydromus smaragdinus seldom utilized the leaves of P. purpureum as a resting place in the GAP area, whereas it commonly used that species at the GAA areas (Table 2). This suggests that A. carolinensis and T. smaragdinus may compete for resting places.

The selection of resting places is crucial for avoiding predation and accessing the habitat utilized during the daytime, as well as for providing benefits for social behaviors (Christian et al., 1984; Clark and Gillingham, 1990; Lima et al., 2005; Singhal et al., 2007). Although studies on the resting places of lizards are limited, several authors have focused on Anolis lizards. For instance, three Anolis species (A. lineatopus, A. grahami, and A. valencienni) that inhabit different daytime habitats, also exhibit differences in their resting perch characteristics and height above the ground at night (Singhal et al., 2007). Combined together, the putative shift in plant species used as resting sites in GAP T. smaragdinus may reduce the accessibility of preferable daytime habitat, increase the predation pressure, and change their social behaviors. Nonetheless, the leaves of P. purpureum, the most frequent resting place of A. carolinensis, are abundant at the GAP area and appear to be still available for T. smaragdinus. Thus, exact cause of the putative effects of A. carolinensis on the substrate plant for resting in T. smaragdinus should be carefully examined in the future studies.

Although previous studies (e.g., Lima et al., 2005; Singhal et al., 2007; Asato et al., 2021) have primarily focused on resource utilization by native lizards with no comparisons with alien species, the present study suggests resting place competition and displacement between A. carolinensis and one species of native lizard. Given the limited research on resting place selection and competition, further studies examining potential changes of predation pressure and social behaviors are needed to understand potential negative effects of A. carolinensis on T. smaragdinus and D. p. polygonatum.

Concerns for the future

Anolis carolinensis was introduced to Zamami-jima Island approximately 10 years ago, in 2013. A control program for this species has been ongoing since its initial establishment (Ministry of the Environment, Government of Japan, 2023; Okinawa Prefectural Government, 2023). These efforts seem helpful in preventing population expansion of the anole, although they may not be sufficient to eradicate it. The intensity of interspecific competition is inversely proportional to the amount of resources allocated between species (Pacala and Roughgarden, 1982). Microhabitat overlap between the green anole and native lizards was observed in this study, and the direct and indirect effects on the native species might increase if the anole expands its range or if the population density rises. Controlling green anoles requires significant effort (Toda et al., 2009; Toda, 2022), but further efforts to monitor and eradicate A. carolinensis, as well as to monitor impacts on native species, are necessary.