Elevation of Divergent Color Polymorphic and Monomorphic Lizard Lineages (Squamata: Agamidae) to Species Level

The Australian Tawny Dragon lizard (Ctenophorus decresii), as currently recognized, comprises two genetically divergent lineages, northern and southern, that differ notably in male coloration. A narrow contact zone exists between the lineages with asymmetric and limited hybridization, indicating incompatibility and highlighting the need for further taxonomic assessment. Here, we evaluate morphological variation in C. decresii and elevate the lineages to separate species. The southern lineage retains the name C. decresii (Duméril and Bibron, 1837), and we formally reinstate C. modestus for the northern lineage, which was previously a synonym of C. decresii (Amphibolurus modestus, Ahl, 1926). We redescribe C. modestus and C. decresii and highlight important considerations for reevaluation of their conservation statuses. Recognition of C. modestus represents another species that may have differentiated in relatively mesic mountainous refugia during Pleistocene glacial-interglacial cycles, and underscores the importance of divergence in male coloration as a driver of speciation in the rock dragon species group.

C TENOPHORUS is the largest genus of Australian agamid lizards, comprising a radiation of 33 currently recognized species (Melville and Wilson, 2019;Sadlier et al., 2019). Within the genus, there is a monophyletic group of six saxicolous species: C. decresii, C. fionni, C. mirrityana, C. rufescens, C. tjantjalka, and C. vadnappa. All are found in South Australia, with the exception of C. mirrityana which is endemic to an adjacent area of western New South Wales. Additionally, all of these species are sexually dimorphic with slightly larger, more brightly colored and patterned males and cryptically colored females and juveniles (Houston, 1974;Johnston, 1992;McLean et al., 2013). Male throat and dorsolateral coloration are distinguishing characteristics between species and function as important sexual signals (Stuart-Fox and Johnston, 2005;Osborne et al., 2012). These traits are emphasized during territorial and courtship displays involving dorsolateral compression of the body and extension of the throat, alongside head-bobbing and hindlimb push-ups (Gibbons, 1979;Ramos and Peters, 2016).
The Tawny Dragon (C. decresii) comprises two largely allopatric lineages which differ notably in male throat and dorsolateral color and pattern (Houston, 1974;McLean et al., 2014a). The northern lineage occurs in the Flinders and Olary Ranges, while the southern lineage occurs in the Mount Lofty Ranges, on the Fleurieu Peninsula, and on Kangaroo Island. Divergence between the two lineages is consistent with range contraction to allopatric refugia during glacial-interglacial Pleistocene cycles (Byrne, 2008;McLean et al., 2014a), and a secondary contact zone exists near the Barossa Valley, between the Olary and Mount Lofty Ranges (McLean et al., 2014a;Dong et al., 2019a: bioRxiv 840678). Differing refugium-specific selective pressures may have led to the evolution of divergent phenotypes between lineages (e.g., McLean et al., 2014b). Males of the southern lineage are monomorphic with blue throat coloration with an ultraviolet (UV) reflectance peak, and some yellow reticulation in the Kangaroo Island population (McLean et al., 2014a;Fig. 1A, B). The northern lineage is color polymorphic with four discrete male throat color morphs: orange, yellow, orangeyellow (yellow with an orange central patch), and gray (Teasdale et al., 2013;Fig. 1D-G). McLean et al. (2013McLean et al. ( , 2014a characterized phenotypic and genetic divergence among C. decresii (northern and southern), C. fionni, C. mirrityana, and C. vadnappa, and highlighted the need to further investigate the taxonomic statuses of the C. decresii lineages. McLean et al. (2014a) found phylogeographic discordance: mitochondrial and nuclear sequence data suggested that populations in the northern Flinders Ranges (northern-most range) were more closely related to the southern lineage than the geographically proximate northern lineage, while this was not supported by microsatellite data. Since then, comprehensive genomic data (1,333 single nucleotide polymorphism loci; SNPs) for populations across the range of C. decresii have shown that the northern Flinders Ranges populations clearly group with the northern lineage (Fig. 2). Furthermore, an in-depth study of the contact zone indicates that intrinsic genetic incompatibilities and sexual signal divergence restrict gene flow between the lineages (Dong et al., 2019a: bioRxiv 840678). Hybridization is asymmetric and limited; hybrids backcross almost exclusively to the northern lineage and predominantly have northern lineage mtDNA haplotypes and northern throat phenotypes (in terms of coloration and UV reflectance; Dong et al., 2019a: bioRxiv 840678). Together, these patterns indicate reproductive incompatibility between Male dorsolateral patterning and coloration differ where (C) C. decresii sensu stricto has a more ''pinched'' or broken lateral stripe with a greater extent of bordering orange or yellow coloration, and (H) C. modestus has a relatively straight edged lateral stripe with cream and orange coloration terminating at the shoulder. the two lineages, which restricts the direction and extent of hybridization.
Here, we formally recognize and redescribe the northern and southern lineages of C. decresii as separate species based on notable morphological, phenotypic, and genetic differentiation and evidence of reproductive incompatibilities between the lineages. The type specimens of C. decresii originate from Kangaroo Island and refer to the southern lineage; decresii gives reference to the French name for Kangaroo Island, L'Île de Decrès. The northern lineage is referred to as Amphibolurus modestus (Ahl, 1926), previously regarded as a synonym of C. decresii by Cogger et al. (1983).

MATERIALS AND METHODS
We compared morphometric measurements and meristic characters (Table 1) of northern (n ¼ 70; male ¼ 45, female ¼ 25) and southern (n ¼ 40; male ¼ 21, female ¼ 19) C. decresii, as well as C. fionni (n ¼ 15; male ¼ 10, female ¼ 5) and C. vadnappa (n ¼ 15; male ¼ 10, female ¼ 5; Fig. 3A; Table S1; see Data Accessibility), as they are the most closely related and morphologically similar species within the rock dragon group (McLean et al., 2013). Data are from McLean et al. (2013) and were collected from adult (SVL 65 mm for males and 60 mm for females) preserved specimens held in collections at the Australian Museum, Sydney (AMS), South Australian Museum (SAM), and Museums Victoria (NMV). Full details of data collection and additional univariate data analyses are given in McLean et al. (2013). To assess whether A. modestus applied to the northern or southern lineage of C. decresii, we obtained photographs, measurements, and meristics of the two syntypes of A. modestus (ZMB 30092) held in the Museum für Naturkunde in Berlin, Germany. The syntypes included a subadult male which was re-registered as ZMB54516 in 1995 and a female which retained the original registration number (ZMB30092; Frank Tillack, in lit., 12 December 2019). To confirm that the type specimens of C. decresii represent the southern lineage, we obtained photographs, measurements, and meristics of the two syntypes, an adult male (MNHN 6545) and adult female (MNHN 6545A) held in the Muséum national d'Histoire naturelle in Paris, France. Institutional abbreviations follow Sabaj (2020).
Linear discriminant analyses (LDA) were used to test whether individuals could be reliably assigned to the four a priori taxa. Males and females were analyzed separately to account for sexual dimorphism, using 23 variables for both (Table 1). Morphometric measurements were expressed as a percentage of snout-vent length (SVL). Tail length (TL) was excluded from both analyses because many individuals had damaged tails (due to predation or intraspecific aggression). Femoral pore count was excluded for females because they could only be counted reliably for males, and SVL was excluded for males because ZMB54516 is a subadult (SVL ¼ 52.9 mm). An additional LDA was conducted on males without the specimen of A. modestus in order to include SVL as a variable (24 variables used). For all analyses, species was used as a dependent variable (i.e., northern C. decresii, southern C. decresii, C. fionni, C. vadnappa) with morphological data as independent variables. Data were centered and scaled, and analyses were conducted using the R package MASS (v7.3.50; Venables and Ripley, 2002) in R v 3.5.1 (R Core Team, 2018).

RESULTS
Discrimination among males of the four taxa (northern C. decresii, southern C. decresii, C. fionni, and C. vadnappa) was highly significant (males: Wilks' k ¼ 0.095, F 69,180 ¼ 3.13, P , 0.0001). Linear discriminant analysis (LDA) on male morphological variables indicated that the male type specimens of A. modestus and C. decresii fell within the 95% confidence ellipses of the northern and southern lineages of C. decresii, respectively (Fig. 3B). Differentiation along the first axis (LD1; variance explained ¼ 52.4%) is driven primarily by number of femoral pores, number of infralabial scales, tibia length, and head length ( Table 2). The second axis (LD2) showed differentiation (variance explained ¼ 28.5%) among northern C. decresii, C. fionni, and C. vadnappa, driven primarily by tibia length and foot length (Table 2; Fig. 3B). Similarly, discrimination of the four taxa including SVL as a variable (excluding the type specimen of A. modestus) for males was highly significant (Wilks' k ¼ 0.071, F 72,174 ¼ 3.44, P , 0.0001). Differentiation along LD1 (variance explained ¼ 60.7%) is driven primarily by SVL, number of femoral pores, and hand length (Table S2; see Data Accessibility). LD2 showed differentiation (variance explained ¼ 23.7%) among northern C. decresii, C. fionni, and C. vadnappa, driven primarily by foot length and tibia length (Table S2; Fig. S1; see Data Accessibility). In both analyses of males, northern C. decresii and southern C. decresii were strongly differentiated along LD1 with 95% confidence ellipses that do not overlap ( Fig. 3B; Fig. S1; see Data Accessibility). The 95% confidence ellipse of northern C. decresii partially overlaps with that of C. fionni but not that of C. vadnappa; the ellipses of C. fionni and C. vadnappa partially overlap with each other; the ellipse of southern C. decresii does not overlap with any other taxon ( Fig. 3B; Fig. S1; see Data Accessibility). For females, discrimination among the four taxa was also significant (Wilks' k ¼ 0.104, F 69,90 ¼ 1.48, P ¼ 0.039). LDA on morphological variables showed differentiation along LD1 (variance explained ¼ 45.7%) driven by tibia length, number of scales between rostral and nasal, and humerus length (Table 2; Fig. 3C). Differentiation along LD2 (variance explained ¼ 34.4%) was driven by SVL, the number of infralabial scales, and hand length (Table 2; Fig. 3C). LDA indicated that the female type specimen of A. modestus fell within the 95% confidence ellipse of northern C. decresii (Fig.  3C). The female type specimen of C. decresii falls in the overlap between ellipses of southern C. decresii and C. vadnappa (Fig. 3C). However, the collection locality of the specimen (i.e., Kangaroo Island) precludes the possibility of the specimen being C. vadnappa. The 95% confidence ellipse of southern C. decresii partially overlaps with that of both northern C. decresii and C. vadnappa; however, the ellipses of northern C. decresii and C. vadnappa do not overlap with each other. The 95% confidence ellipse of C. fionni very marginally overlaps with that of C. vadnappa only. Grammatophora decresii (Duméril and Bibron, 1837: 472-474). Amphibolurus decresii 'southern race' (Houston, 1974: 51). Ctenophorus decresii 'southern lineage' (McLean et al., 2013: 52).
Types.-The original description was based on two syntypes, a male and a female, from L'Île de Decrès (Kangaroo Island) Diagnosis.-Ctenophorus decresii is a member of the C. decresii species group (Houston, 1974;Houston and Hutchinson, 1998). Within the C. decresii species group, C. decresii is distinguishable by the following combination of characters: snout scales smooth to longitudinally keeled; vertebral scales flat; presence of a broad dark undulating lateral stripe, often constricted into blotches and interrupted to form two separate blotches posterior to the tympanum; and in males, the throat is blue or blue and yellow with yellow to orange margins.
Description.-A moderately sized lizard reaching a maximum SVL of approximately 88 mm and total length of 230 mm. Head and body depressed. Tail long and tapered to a fine tip. Forelimbs moderately long, almost reaching groin when adpressed. Hindlimbs long and reaching or almost reaching snout when adpressed. Finger lengths: 4 . 3 . 5 . 2 . 1; toe lengths: 4 . 3 . 5 . 2 . 1. Characteristic of the genus, a distinct row of enlarged keeled scales extending from the nostril, below the eye to above the tympanum (Houston and Hutchinson, 1998). Ear membranes exposed and scale-free. Scales on snout are smooth to longitudinally keeled. 12-18 supralabial scales; 11-16 infralabial scales; 4-6 scales between rostral and nasal; 4-6 scales between supralabial and nasal; 10-13 internasal scales; 21-29 subdigital lamellae on the fourth toe (McLean et al., 2013). Dorsal crest is small, no more than a series of keeled scales, becoming smaller posteriorly and not reaching base of tail. Nuchal crest low or absent. Scales on the dorsal surfaces of the limbs and tail are keeled. The gular fold is strongly developed and extends across the shoulders. 40-46 femoral and precloacal pores are arranged in a straight line along the thighs. Pores are present but smaller in females. Adult male base color varies from gray-blue to brown, with pale blue flanks (Figs. 1C, 4A). The dorsolateral surface of the head is orange to yellow, extending past the shoulders. The extent of this coloration varies among individuals and depends on location but does not extend further than the mid-point of the axilla-groin length. A broad black lateral stripe begins posterior to the tympanum and terminates at the groin (Figs. 1C, 4A). The lateral stripe forms two separate blotches on the shoulder, one circular and enclosed entirely by orange to yellow coloration, while the other is elongated and wraps partly around the other. The remaining length of the lateral stripe is wavy edged or broken. On the ventral surface, males are cream with orange flushes on the belly, hindlimbs, and the length of the tail during the breeding season. A dark chest patch tapers to a point midbody and sometimes extends along the undersides of the forelimbs. This dark patch may vary in intensity, from black to gray. Male throat coloration in mainland populations is ultraviolet-blue with yellow to orange coloration along the gular fold ( Adult females are cryptically colored, predominantly graybrown to orange-brown, densely patterned with darker and lighter speckling (Fig. 4B) and sometimes a thin, pale vertebral line. Similar to males, a black lateral stripe runs along the flanks, similar in shape to that of males but sometimes less prominent. White to cream on the ventral surface. Varying degrees of gray reticulations on the throat, sometimes with a blue tinge. Orange to yellow flushes on the belly and gular region during the breeding season. Juveniles resemble adult females in coloration and pattern. Color patterning remains visible in fresh spirit-preserved specimens but fades with time.  (Figs. 2C, 3A, 6). These lizards occupy temperate habitats with varying levels of vegetation. Adults occur almost exclusively on rocks, perching conspicuously on elevated sites to bask and/or perform territorial and courtship displays. They are active in hot weather and, if disturbed, will run into a crevice or beneath a rock slab. Amphibolurus modestus (Ahl, 1926: 187-188). Amphibolurus decresii 'northern race' (Houston, 1974: 52). Ctenophorus decresii 'northern lineage' (McLean et al., 2013: 52).
Types.-The original description was based on two syntypes, both from 'Australien' and registered as ZMB30092 (Ahl, 1926). We designate the subadult male as lectotype (now registered as ZMB 54516; Fig. 5B) and the adult female as paralectotype (ZMB 30092; Fig. S3; see Data Accessibility) because only the male exhibits the diagnostic coloration. No exact locality is available for these specimens.
Diagnosis.-Ctenophorus modestus is a member of the C. decresii species group (Houston, 1974;Houston and Hutchinson, 1998). Within the C. decresii species group, C. modestus is distinguishable by the following combination of characters: snout scales smooth to longitudinally keeled; vertebral scales flat; a smooth-edged dark lateral stripe, sometimes interrupted to form a separate blotch posterior to the tympanum; and in males, throat coloration is polymorphic (orange, yellow, orange-yellow, or gray) with or without an overlay of gray reticulations and a small central black patch.
Description.-A moderately sized lizard reaching a maximum SVL of approximately 85 mm and total length of 255 mm. Head and body depressed. Tail long and tapered to a fine tip. Forelimbs moderately long, almost reaching groin when adpressed. Hindlimbs long and reaching or almost reaching snout when adpressed. Finger lengths: 4 . 3 . 5 . 2 . 1; toe lengths: 4 . 3 . 5 . 2 . 1. Characteristic of the genus, a distinct row of enlarged keeled scales extending from the nostril, below the eye to above the tympanum (Houston and Hutchinson, 1998). Ear membranes exposed and scale-free. Scales on snout are smooth to longitudinally keeled. 13-18 supralabial scales; 13-18 infralabial scales; 4-6 scales between rostral and nasal; 4-6 scales between supralabial and nasal; 9-13 internasal scales; 19-27 subdigital lamellae on the fourth toe (McLean et al., 2013). Small clustered pale yellow to orange spines anterior to tympanum. Dorsal crest is small, no more than a series of keeled scales, becoming smaller posteriorly and not reaching base of tail. Nuchal crest low or absent. Scales on the dorsal surfaces of the limbs and tail are keeled. The gular fold is strongly developed and extends across the shoulders. 34-45 femoral and precloacal pores are arranged in a straight line along the thighs. Pores are present but smaller in females. Adult male base color varies from gray-blue to brown, with pale blue flanks (Figs. 1H, 4C). The dorsolateral surface of the head is brown, with orange around the eyes, nostrils, and along the upper jaw, and beneath the tympanum terminating at the neck. The extent of this coloration varies among individuals and depends on location. A broad black lateral stripe begins posterior to the eye, becoming thicker posterior to the tympanum and terminating at the groin. A thin cream, yellow, or orange stripe begins at the tympanum and terminates just posterior to the shoulder (Figs. 1H, 4C). On the ventral surface, males are cream with orange flushes on the belly, hindlimbs, and the length of the tail during the breeding season. A dark chest patch tapers to a point midbody and sometimes extends along the undersides of the forelimbs. This dark patch may vary in intensity, from black to gray. Male throat coloration varies among individuals with distinct gray, orange, yellow, and orange-yellow morphs (Teasdale et al., 2013; Fig. 1D-G). Base throat color sometimes overlain with gray reticulations or with a distinct black mark in the center of the throat.
Adult females are cryptically colored, predominantly graybrown to orange-brown, densely patterned with darker and lighter speckling (Fig. 4D) and sometimes a thin, pale vertebral line. Similar to males, a black lateral stripe runs along the flanks, but sometimes less prominently than in males. White to cream on the ventral surface, with varying degrees of gray reticulations on the throat. Orange to yellow flushes on the belly and gular region during the breeding season. Juveniles resemble adult females in coloration and pattern. Color patterning remains visible in fresh spiritpreserved specimens but fades with time.   6. Geographic distribution of Ctenophorus modestus (squares) and C. decresii (circles) based on specimens in Australian Museums from the Atlas of Living Australia (data: https://doi.org/10.26197/ 5d91626857226); coordinates which were likely inaccurate (e.g., unsuitable habitat) were removed. Protected areas described in the Collaborative Australian Protected Areas Database (CAPAD, 2018) are shown as gray areas. The extent of occurrence is shown for C. modestus (49,102 km 2 ; orange lines) and C. decresii (6,604 km 2 total [mainland: 5,772 km 2 ; Kangaroo Island: 832 km 2 ], blue lines). SVL,52.9;TL,95.6;AGL,24.3;HL,18.1;HW,14.0;HD,7.8;EYE,3.8;SL,6.6;JL,8.3;NW,5.3;HUML,7.7;RADL,8.5;HANL,11.7;FINL,5.6;FEML,14.5;TIBL,17.4;FOOT,21.3;TOE,11.5 Distribution and natural history.-Ctenophorus modestus inhabits rocky areas and outcrops, sheltering in rock crevices, similar to other members of the C. decresii group (Gibbons, 1979). The species occurs throughout the Flinders and Olary Ranges in South Australia, Australia (Figs. 2C, 3A, 6). These lizards occupy habitats ranging from well-watered to semiarid, with sparse vegetation and are active in hot weather (Houston, 1974;Gibbons, 1979). Adults occur almost exclusively on rocks, but juveniles may be found among surrounding vegetation up to 100 meters away (Houston, 1974;Gibbons and Lillywhite, 1981). Adults perch conspicuously on elevated sites to bask and/or perform territorial and courtship displays (Gibbons, 1979). If disturbed, they will run into a crevice or beneath a rock slab. Predators are predominantly birds but likely also include elapid snakes (Gibbons and Lillywhite, 1981;Stuart-Fox et al., 2004).

Measurements (mm) and meristic characters of lectotype.-(Bilateral meristics as left [L] and right [R])
Comparisons among species in the C. decresii complex Similar species.-Members of the C. decresii species group differ most markedly in male coloration and geographic distribution. The distribution of C. modestus does not overlap with four other species in the C. decresii species group (C. decresii, C. mirrityana, C. rufescens, C. tjantjalka). The dorsal color pattern of C. modestus is very similar to that of C. mirrityana which occurs in the Barrier Ranges of western New South Wales, but can be distinguished by the absence of a thinner non-continuous orange stripe within the black lateral stripe, absence of a dark median throat stripe in males, and a larger head relative to SVL (McLean et al., 2013). Ctenophorus tjantjalka occurs in the low stony hills and eroded creek lines of the Lake Eyre Basin in northern South Australia and has a shorter, deeper head relative to C. modestus. Male C. tjantjalka are uniform light blue above with salmon flanks and lack the dark lateral stripe of C. modestus. Ctenophorus rufescens occurs primarily in the granitic ranges of north-western South Australia. Both sexes of C. rufescens are rusty to medium brown, lacking both the bright yellow, orange, or blue coloring on the head of male C. modestus and the dark lateral stripe in juvenile and female C. modestus. It is most likely that C. modestus may be confused with C. fionni, which occurs west of Spencer Gulf and Lake Torrens in South Australia. Male C. fionni exhibit extensive geographic variation in coloration (Houston, 1974) and do not differ consistently in any morphological trait from C. modestus. Available keys (e.g., Cogger, 2018) differentiate between C. modestus and C. fionni based on the presence of scattered pale tubercules on the flanks on the former; however, this is primarily diagnostic of C. decresii sensu stricto. Additional diagnostic traits are needed to allow C. fionni to be distinguished from most other members of the C. decresii group independent of a collection locality.
In the northern-most part of its geographic range, C. modestus occurs in sympatry with C. vadnappa. These species can be distinguished by adult male coloration, juvenile and female coloration, and morphology of the snout and vertebral scales. Adult male C. vadnappa have red variegations on the sides of the body that tend to align in vertical bars and no dark lateral stripe, whereas adult male C. modestus lack red reticulations and have a dark lateral stripe. Juvenile and adult female C. vadnappa also lack the dark lateral stripe, which is present in C. modestus. Ctenophorus vadnappa have wrinkled snout scales, whereas the snout scales are keeled in C. modestus. In C. vadnappa, the vertebral scales are keeled, whereas they are smooth in C. modestus.
There are notable differences in male throat coloration between C. modestus and C. decresii sensu stricto. Male C. modestus have orange, yellow, orange-yellow, or gray throats (Fig. 1D-G), whereas male C. decresii have blue throats (Fig.  1A-B). The black lateral stripe is relatively straight-edged in C. modestus compared with the irregular edge in C. decresii (Fig. 1C, H). Ctenophorus modestus and C. decresii also differ significantly in six morphological characters, with both sexes of C. modestus having higher SVLs (C. modestus male: 66.3-84.6 mm, female: 65.1-82.9 mm; C. decresii male: 65.2-77.5, mm female: 61.9-74.13 mm), greater numbers of supralabial (C. modestus male: 13-18, female: 13-17; C. decresii male: 12-16, female: 12-18) and infralabial scales (C. modestus male: 13-18, female: 14-18; C. decresii male: 12-16, female: 11-16), and males of C. modestus having longer jaws (C.  Melville et al., 2018). The taxonomic separation of the northern and southern lineages into C. modestus and C. decresii, respectively, warrants a reevaluation of each species' conservation status. Both have distributions that are more restricted than the original distribution of C. decresii sensu lato. To measure the extent of occurrence (km 2 ) for C. modestus and C. decresii sensu stricto, we drew minimum convex polygons encompassing occurrences based on all specimens in Australian Museums, downloaded from the Atlas of Living Australia. The extent of occurrence for C. modestus is 49,102 km 2 (Fig. 6) and falls almost entirely within the Flinders Lofty Block (FLB) bioregion as delineated by the Interim Biogeographic Regionalisation for Australia (IBRA version 7; Department of the Environment, 2012). The extent of occurrence for C. decresii is 6,604 km 2 (mainland: 5,772 km 2 ; Kangaroo Island: 832 km 2 ; Fig. 6), distributed within both the FLB and Kanmantoo (KAN) bioregions. The FLB and KAN bioregions consist of 8.9% and 20.3% protected areas (CAPAD, 2018; Table 3). However, the distributions of C. modestus and C. decresii overlap minimally with protected areas, except in the northern Flinders Ranges and on Kangaroo Island (Fig. 6).
The species is locally abundant in suitable habitat in the southern Flinders Ranges and Olary Ranges (Gibbons and Lillywhite, 1981); however, suitable rocky habitat is fragmented and most vegetation has been cleared in the southern part of their range. Further, the population of C. modestus north of Blinman (31805 0 37.1 00 S, 138840 0 41 00 E) in the northern Flinders Ranges is predicted to decline under future climate scenarios (Walker et al., 2014). The IUCN assessment acknowledges a decreasing population trend due to this range contraction (Melville et al., 2018). A better understanding of temporal variation in population density is needed to assess population trends in the northern part of the geographic distribution of both species.
Ctenophorus decresii may merit conservation management due to its more restricted and fragmented distribution. The total extent of occurrence for the species is less than 20,000 km 2 (i.e., 6,604 km 2 ). Further, the habitat is fragmented, given that the species occurs only in rocky areas with low to moderate vegetation cover (criteria B1-a). Ctenophorus decresii is particularly vulnerable to habitat loss due to human development and bush rock removal (Michael et al., 2010). Although the FLB and KAN bioregions have both experienced grazing pressure from sheep, cattle, and feral goats leading to habitat degradation and clearing, it has occurred more severely in the KAN bioregion (Environment Australia, 2000;Bastin, 2008). Therefore, C. decresii likely has a higher risk conservation status than C. modestus. Land clearing has had detrimental impacts on genetic variation in ecologically similar species (Driscoll, 2004;Driscoll and Hardy, 2005;Levy et al., 2010), and vegetation clearance around rock outcrops reduces dispersal between patches of suitable habitat in saxicolous species (Stow et al., 2001;Levy et al., 2010).
Bush rock removal has been implicated in the decline of several threatened reptile species (Michael et al., 2010;Fitzsimons and Michael, 2017;Michael and Lindenmayer, 2018). Additionally, habitat loss and fragmentation resulting from agriculture creates remnant edge habitats where reptiles experience significantly higher predation risk (Hansen et al., 2019). Furthermore, the degree to which populations of C. decresii have been impacted by recent extensive bushfires on Kangaroo Island and in Cudlee Creek on the mainland (ca. December 2019-January 2020) are still unknown. Although continued monitoring is necessary, the bushfires are likely to negatively impact many subcriteria of B1-b (i-ii: geographic range, iii: area, extent, and/or quality of habitat, iv: number of locations or subpopulations, v: number of mature individuals).
Throat polymorphism.-The throat color polymorphism in C. modestus is well characterized by a large body of literature. Throat coloration is a sexual signal, develops before sexual maturity, is fixed for life, and does not vary with body size (Gibbons and Lillywhite, 1981;Stuart-Fox, 2002), and may function in rival recognition (Osborne et al., 2012). Four discrete male throat morphs (i.e., orange, yellow, orangeyellow, gray) are present in all surveyed populations (Teasdale et al., 2013) with morph frequencies influenced by local environmental conditions (mean annual aridity index and vegetation cover; McLean et al., 2015). The four morphs employ different behavioral strategies, with the gray and orange morphs being the least and most aggressive and bold, respectively, and the yellow morph's behavior dependent on the competitor's morph (Yewers et al., 2016). These alternative behavioral strategies also correspond to different hormone profiles (Yewers et al., 2017). Throat morphs are associated with MHC supertypes with gray morphs twice as likely to possess supertype four (Hacking et al., 2018).
The genetic and biochemical basis of throat coloration in C. modestus has been characterized. Throat morphs are autosomally inherited and controlled by at least two loci. The extent of orange or yellow within morphs behaves as a quantitative (polygenic) trait (Rankin et al., 2016). Orange and yellow skin coloration are produced from a combination of pteridine pigments (self-synthesized) and carotenoid pigments (acquired through the diet), whereas gray skin is produced from melanin pigments (McLean et al., 2017).
Ctenophorus modestus and C. decresii share tetrachromatic visual systems with well-developed color vision and no significant differences in spectral sensitivity or opsin expression (Yewers et al., 2015;Dong et al., 2019b). Throat morphs of C. modestus vary in conspicuousness, as perceived by agamid lizards (Teasdale et al., 2013). Gray morphs are the least chromatically conspicuous and orange and orangeyellow morphs are the most conspicuous against rock and lichen backgrounds. The throat colors of C. decresii and C. modestus are locally adapted to be more conspicuous against native lichen than non-native lichen as perceived by agamid lizards (McLean et al., 2014b).
Biology.-The mating season of C. modestus is from September to December. Females produce clutches of between 3-7 eggs, commonly laying two clutches in one breeding season (maximum of three clutches; Dong, Rankin, McLean, Stuart-Fox, unpubl.). At oviposition, eggs have a mass of 1.1660.19 g (mean6SE) and offspring hatch in 52-65 days when incubated at 28-28.58C (Rankin et al., 2016;Dong et al., unpubl.). There are low rates of multiple paternity within clutches: a 4.3% rate (1 in 23 clutches) in a wild population (Hacking et al., 2017), and rates of 3.7% (1 in 27 clutches; Dong et al., unpubl.) and 0% (0 in 25 clutches; Rankin et al., 2016) in captive populations. An incubation temperature of 28.58C yielded a sex ratio of 1:0.72 (males:females; n ¼ 58, 3 unsexed; Rankin et al., 2016) and 288C yielded a sex ratio of 1:0.57 (males:females; n ¼ 58, 11 unsexed; Dong et al., unpubl.). It is not definitively known whether the species exhibits genetic sex determination (GSD) or temperaturedependent sex determination (TSD). Previous studies on the reproductive biology of C. decresii sensu lato demonstrating TSD in the species were conducted on individuals from New South Wales (Harlow, 2000;McFadden and Harlow, 2007), Table 3. Amounts and types of protected areas relevant to the distribution of Ctenophorus decresii sensu lato. For the Flinders Lofty Block (FLB) and Kanmantoo (KAN) IBRA bioregions: total area (ha) of the region, total protected area (ha) within the region, percentage of area protected of total area, and percentage of total protected area within each IUCN Protected Area Category: Ia ('Strict Nature Reserve'), Ib ('Wilderness Area'), II ('National Park'), III ('National Monument or Feature'), IV ('Habitat/Species Management Area'), V ('Protected Landscape/Seascape'), VI ('Protected area with sustainable use of natural resources'). Data from CAPAD (2018).

Bioregion
Total area ( et al., 2013). Dragon lizards in the family Agamidae include both GSD and TSD species, with both occurring within the same genera (see references in Ezaz et al., 2009). Further empirical studies are needed to establish the sex-determining mechanisms of both C. modestus and C. decresii. Few published studies have focused solely on C. decresii sensu stricto. The species has a similar breeding season of September-December. Females produce one clutch per breeding season (rarely a second clutch) of between 3-6 eggs (Dong et al., unpubl.). At oviposition, eggs have a mass of 1.5160.21 g and will hatch in 57-61 days when incubated at 2860.128C (mean6SE; Dong et al., unpubl.). There was no evidence for multiple paternity from 14 clutches in a captive population (Dong et al., unpubl.), and an incubation temperature of 288C yielded a sex ratio of 1:0.71 (males:females; n ¼ 58, 11 unsexed; Dong et al., unpubl.). (1985) previously resurrected C. modestus (Ahl, 1926) from synonymy with C. decresii and nominated the larger of the two syntypes as a lectotype. However, they provided no rationale or evidence to support this decision. We regard Wells and Wellington's (1985) recognition of C. modestus as unavailable for application for the following reasons: (1) they did not provide a ''. . . definition that states in words characters that are purported to differentiate the taxon'' as required under Article 13(a)(i-ii) of the third edition of the Code of Zoological Nomenclature applicable at that time;

Wells and Wellington
(2) Ahl's (1926) description of C. modestus also did not meet this criterion, so Wells and Wellington (1985) also did not provide a ''. . . bibliographic reference to such a published statement. . . '' as required by the same article of the Code. Suggesting that the larger female syntype be regarded as the lectotype suggests Wells and Wellington may not have viewed the syntypes of C. modestus because it was well understood at the time that adult male coloration is a key character used to differentiate species in the C. decresii group (Houston, 1974;Gibbons and Lillywhite, 1981). The importance of male coloration in species diagnosis in this group has become even clearer since then (Johnston, 1992;McLean et al., 2013) and is the reason we have nominated the smaller male syntype as the lectotype. While Ahl's (1926) description of C. modestus does not provide a clear means of ''. . . differentiating the taxon'' required of publications in 1985, it does conform to the requirements of the Code for species descriptions prior to 1931. Therefore, the name Amphibolurus modestus Ahl (1926; now Ctenophorus modestus) is available for application. Furthermore, Ahl's male syntype of C. modestus is clearly conspecific with the taxon referred to previously as the ''northern race'' (Houston, 1974) or ''northern lineage'' (McLean et al., 2013(McLean et al., , 2014a of C. decresii sensu lato based on morphology (Fig. 3B) and dorsal patterning (Fig. 5B). The female syntype is consistent with the ''northern race/lineage'' of C. decresii sensu lato based on morphology, although it is not objectively determinable based on phenotype.
We have resurrected C. modestus from synonymy with C. decresii on the basis of genetic distinctiveness and diagnostic differences in male coloration, mensural, and meristic characters. Historical changes in the distribution of suitable habitat appear to have had an influence on the evolution of the C. decresii species group. Extensive rock in the past presumably allowed ancestors of the C. decresii group to disperse widely in southern Australia with subsequent isolation leading to the evolution of different species (Houston, 1974). Further, the combination of sexual selection and local adaptation to environmental differences in climate, vegetation, and native rock and lichen coloration have likely contributed to divergence in male coloration between species (McLean et al., 2014b).

DATA ACCESSIBILITY
Supplemental material is available at https://www. ichthyologyandherpetology.org/h2020064. Unless otherwise indicated in the figure caption, the published images and illustrations in this article are licensed by the American Society of Ichthyologists and Herpetologists for use if the use includes a citation to the original source in accordance with the Creative Commons Attribution CC BY License.