Caribicus warreni ( Schwartz 1970 )

Schools, Molly & Hedges, Blair, 2024, A new forest lizard fauna from Caribbean islands (Squamata, Diploglossidae, Celestinae), Zootaxa 5554 (1), pp. 1-306 : 34-39

Caribicus warreni ( Schwartz 1970)

Hispaniolan Giant Forest Lizard

(Fig. 9–10)

Diploglossus warreni Schwartz, 1970:780 . Holotype: AMNH 103215 , collected by C. R. Warren at Palmiste, Tortue Island,

Département du Nord Ouest, Haiti on 27 January 1968 (20.0179, -72.7246; 320 m). Diploglossus carraui — Incháustegui et al., 1985:196. Holotype: USNM 197369 , collected by Niño Gómez and presented by

José Antonio Carrau from Comedero in May 1978 (19.8183, -71.0617; 76 m). Diploglossus carraui — Schwartz & Henderson, 1991:403. Diploglossus warreni — Schwartz & Henderson, 1991:406. Celestus warreni —Powell et al., 1996:66. Celestus carraui —Powell et al., 1996:65. Celestus warreni carraui — Hallermann & Böhme, 2002:169. Celestus warreni warreni — Hallermann & Böhme, 2002:169. Celestus warreni — Hedges et al., 2019:17. Celestus warreni —Langer, 2019:16. Caribicus warreni — Schools & Hedges, 2021:218. Caribicus warreni — Landestoy et al., 2022:204.

Material examined (n=14). DOMINICAN REPUBLIC. ANSP 38502 , Dominican Republic , locality not available (pet trade), 1 August 1990 . MALT (one unnumbered specimen), data received from Miguel Landestoy. Puerto Plata . USNM 197369 , Niño Gómez, Comedero , La Isabela, May 1978 . HAITI. ANSP 38501 , locality not available (pet trade), 1 June 1989 ; Nord-Ouest. AMNH 103215 , C . R. Warren, Tortue Island , Palmiste, 27 January 1968 ; MNHNSD 723–5 , 727–9 , 731 ; KU 227530 , Tortue Island , Palmiste, November 1968 ; USNM 59435 , Riviere des Barres , 21 February 1917 .

Diagnosis. Caribicus warreni has (1) a dorsal pattern of absent/bands, (2) head markings absent/present, (3) markings in the longitudinal paramedian area present, (4) dots arranged in bars in the lateral band absent/present, (5) a maximum SVL of 227–300 mm, (6) ventral scale rows, 78–98, (7) midbody scale rows, 33–43, (8) total lamellae on one hand, 41–47, (9) total strigae on ten scales, 458–500, (10) relative length of all digits on one hindlimb, 27.0–27.3 %, (11) relative distance between the angled subocular and mouth, 1.51–1.57 %, (12) relative eye length, 3.43–3.54 %, (13) relative forelimb length, 19.8–22.0 %, (14) relative ear width, 1.20–1.88 %, (15) relative rostral height, 1.55–1.99 %, (16) relative head length, 19.1–22.1 %, (17) relative mental width, 1.46–1.87 %, (18) relative postmental width, 2.73–3.32 %, (19) relative cloacal width, 9.33–10.3 %, (20) relative prefrontal width, 4.65–7.41 %, (21) relative largest supraocular width, 1.95–2.79 %, (22) relative longest finger length, 3.41–4.71 %, (23) relative distance between the ear and eye, 5.96–9.08 %, (24) relative head width, 71.4–83.7 %, (25) relative frontal width, 82.3 %, (26) relative nasal height, 1.06 %, (27) relative angled subocular height, 0.652–1.34 %, (28) relative distance between the eye and naris, 5.42–6.27 %, (29) relative canthal iii length, 1.90–2.17 %, (30) relative angled subocular width, 2.13–2.91 %, and (31) relative nasal length, 1.49 %. The species stem time is 6.83 Ma and the species crown time is 0.63 Ma (Fig. 4).

Caribicus warreni has the smallest relative longest finger length (3.41–4.71) and relative nasal height (1.06) of the genus. This species also has the largest relative distance between the angled subocular and mouth (1.51–1.57) of the genus.

FIGURE 9. (A–F) Caribicus warreni (AMNH 103215, holotype), SVL 218 mm.

FIGURE 10. Caribicus warreni (SBH 194521), in life. From Puerto Plata Province, Dominican Republic. Photo by SBH.

From Caribicus anelpistus , we distinguish C. warreni by the relative length of digits on one hindlimb (27.0– 27.3 versus 24.6), the relative ear width (1.20–1.88 versus 1.15), and the relative rostral height (1.55–1.99 versus 2.03). From C. darlingtoni , we distinguish C. warreni by the dorsal pattern (absent/bands versus lineate), the SVL (227–300 versus 61.1–74.9), the total lamellae on one hand (41–47 versus 33–39), the total strigae on ten scales (458–500 versus 90–120), the relative distance between angled subocular and mouth (1.51–1.57 versus 0.768–1.13), the relative mental width (1.46–1.87 versus 2.05–2.52), the relative cloacal width (9.33–10.3 versus 7.08–8.48), the relative longest finger length (3.41–4.71 versus 4.86–6.14), the relative frontal width (82.3 versus 74.3–80.7), the relative nasal height (1.06 versus 1.14–1.45), and the relative nasal width (1.49 versus 1.63–1.92).

Description of holotype. AMNH 103215. An adult female; SVL 218 mm; tail nearly cylindrical, broken in life midway, regenerated, 111 mm (50.9% SVL); axilla-to-groin distance 134 mm (61.5% SVL); forelimb length 39.3 mm (18.0% SVL); hindlimb length 52.1 mm (23.9% SVL); head length 34.8 mm (16.0% SVL); head width 24.0 mm (11.0% SVL); head width 69.0% head length; diameter of orbit 6.47 mm (2.97% SVL); horizontal diameter of ear opening 1.69 mm (0.775% SVL); vertical diameter of ear opening 1.97 mm (0.904% SVL); length of all toes on one foot 46.6 mm (21.4% SVL); shortest distance between angled subocular and lip 2.18 mm (1.00% SVL); shortest distance between the ocular and auricular openings 15.7 mm (7.20% SVL); longest finger length 9.50 mm (4.36% SVL); largest supraocular width 4.47 mm (2.05% SVL); prefrontal width 10.4 mm (4.77% SVL); frontal width 79.4% frontal length; nasal height 1.95 mm (0.894% SVL); angled subocular height 2.03 mm (0.931% SVL); shortest distance between the eye and naris 9.92 mm (4.55% SVL); canthal iii width 2.17 mm (0.995% SVL); angled subocular width 4.55 mm (2.09% SVL); nasal width 2.49 mm (1.14% SVL); rostral 1.64X as wide as high, barely visible from above, not in contact with nasals, in contact with 1 st supralabial and anterior internasal (left)/(right); anterior internasals are narrower than posterior ones; frontonasals and prefrontal fused into a single large plate with a slightly concave posterior margin, much wider than long, bordered by posterior internasals, 1 st loreals, canthal iii, 1 st median oculars, and the frontal; frontal longer than wide; a pair of frontoparietals, separated by the posterior prolongation of the frontal and the interparietal plate; interparietal plate smaller than parietals and separating them, posteriorly touching the interoccipital (fused to both parietals), which is longer than wide; parietal separated from supraoculars by 1 st and 2 nd temporals and frontoparietal (left)/(right); nasal single; nostril above suture between 1 st and 2 nd supralabials (left)/(right); 1 postnasal (left)/(right); 2 loreals (left)/(right); 1 st loreal higher than wide (left)/(right), in contact with postnasal, posterior internasal, prefrontal/frontonasal complex, canthal iii, 2 nd loreal, and 3 rd –4 th supralabials (left)/(right); 2 nd loreal shorter than 1 st, approximately as high as wide (left)/ (right), excluded from contact with supraocular by canthal iii (left)/(right); 2 nd loreal posteriorly bordering the upper and lower preoculars (left)/(right); canthal iii wider than high (left)/(right), contacting 1 st median ocular, anterior supraciliary, upper preocular, and 1 st and 2 nd loreals (left)/(right); 10 median oculars (left)/(right), 1 st contacting the prefrontal (left)/(right); 1 (left)/2 (right) upper preoculars; an irregular anterior supraciliary (left)/(right); 7 lateral oculars (left)/(right); 5 temporals (left)/(right); 2 suboculars (left)/(right); posterior subocular large and elongate (left)/(right); anterior subocular small (left)/(right); 9 supralabials (left)/(right), 6 to level below center of eye (left)/ (right); 9 infralabials (left)/(right), 5 (left)/6 (right) to level below center of eye; mental small, followed by a single, larger post mental; 5 pairs of enlarged chin shields, followed by 1 pair of reduced chin shields; 1 st pair in contact with one another, 2 nd pair in contact with one another anteriorly, posteriorly separated by one scale; 90 transverse rows of dorsal scales from interoccipital to base of tail; 92 transverse rows of ventral scales from mental to vent; 37 scales around midbody; 5 digits; finger lengths 3>4>2>5>1; 11 lamellae under longest finger (left)/(right); 44 total lamellae on one hand; toe lengths 4>3>5>2>1; 15 lamellae under longest toe (left)/(right); striate and slightly keeled dorsal body and caudal scales; smooth ventral scales; 436 total strigae counted on ten scales.

Color (in alcohol): dorsal surface of head golden gray-brown, patternless; lateral surfaces of head grading from golden gray-brown to dark yellow-cream; dorsal surfaces of the body are gray-brown with faded brown longitudinal paramedian markings and chevrons; dorsal surface of tail the same as the body; lateral areas grade from gray-brown to dark yellow-cream with continuations of the chevrons from the body; dorsal surfaces of the limbs are gray-brown with darker brown mottling; lateral and ventral areas of the limbs grading from gray-brown to dark yellow-cream; ventral surfaces of the head, body, and tail are dark yellow-cream and patternless.

Variation. Although faded in most cases, the examined material resembles the pattern of the holotype closely with darker bands extending laterally along the dorsum. In the type, these bands most closely resemble chevrons.All examined specimens have faded markings in the longitudinal paramedian area, in most specimens these markings are thick, short longitudinal paramedian lines, whereas in KU 227530 the longitudinal paramedian area is mottled. KU 227530 is unique in having black outlines on its head scales. Dots in the lateral area are either absent or very reduced. Measurements and other morphological data for the holotype and other examined material are presented in Table 1.

Distribution. Caribicus warreni is distributed in southeastern Haiti and the northern regions of the Dominican Republic and Haiti, including Tortue Island, at elevations of 30–690 m (Fig. 5).

Ecology and conservation. Caribicus warreni has been collected in a variety of habitats. Populations were reportedly found in lowland, pine, and cloud forests ( Incháustegui et al. 1985; Franz & Cordier 1986). In these habitats, individuals were found associated with root systems of trees, under rocks, outside of burrows (during the day), and on asphalt roads at night ( Incháustegui et al. 1985). On Tortue Island, individuals were collected in banana groves ( Schwartz & Henderson 1991). Studies of the stomach contents of C. warreni show that the vast majority of prey species are associated with leaf litter ( Incháustegui et al. 1985). Additional studies found that juvenile C. warreni showed preference to cricket stimuli over lettuce, cologne, and water ( Cooper & Bradley 2009).

The IUCN Redlist ( IUCN 2023) considers the conservation status of Caribicus warreni to be Vulnerable B1ab(iii) “due to its limited distribution (with an extent of occurrence of 14,646 km 2), fragmented subpopulations and ongoing threats include expanding agricultural activities, charcoal production, predation by cats, dogs and mongooses. It is killed by local people who mistakenly consider these lizards to be venomous, and it is on the illegal pet trade that continues to decline its extent of occurrence, and quality of habitat, and it is only found in a small protected area.” This species has been successfully bred in captivity ( McGinnity 2002). Unfortunately, eradication of introduced mammalian predators, including black rats, which are also a threat, is currently not possible on large islands such as Hispaniola. Analyses of satellite imagery of forest cover for the countries of Haiti ( Hedges et al. 2018) and the neighboring Dominican Republic ( Sangermano et al. 2015) have shown that protected areas and reserves are often ineffective conservation actions unless accompanied by effective management.

Reproduction. Ovoviviparous. Ten gravid females in August had 8–27 young (SVL 32–47 mm), litter size positively correlated with female SVL ( Incháustegui et al. 1985). The maximum recorded litter size of this species is 34 ( Lawler & Norris 1979).

Etymology. The species name refers to Mr. C. Rhea Warren, the collector of the original material used to describe the species.

Remarks. The original description of Caribicus warreni suggested a close relationship between Caribicus warreni and Comptus stenurus , in part because the paratype of Caribicus warreni was at one point identified as Comptus stenurus ( Schwartz 1970) . This assignment placed Caribicus warreni in a group with Panolopus costatus , in addition to Comptus stenurus ( Schwartz 1970) . Schwartz (1970) also suggested that the fossils of a large diploglossid in a cave at Cerro de San Francisco ( Etheridge 1965) could be C. warreni , which would indicate that C. warreni was previously more widespread.

Caribicus carraui was described as its own species based on morphological characters ( Incháustegui et al. 1985), before it was later designated a subspecies of C. warreni ( Hallermann & Böhme 2002) , along with C. anelpistus . Caribicus anelpistus was later elevated to a full species again ( Powell & Henderson 2003), whereas C. carraui was placed in the synonymy of C. warreni ( Powell & Henderson 2003) . In the type description of C. carraui , the authors speculated that it was an intermediate form between C. anelpistus and C. warreni , but still designated C. carraui as a full species, primarily based on ventral differences in both coloration (cream) and pattern (“randomly placed dark brown blotches or smudges”) ( Incháustegui et al. 1985). Although we identified several diagnostic characters separating C. warreni from C. anelpistus , more data are needed from the latter, especially DNA sequences, to confirm its taxonomic status.

Strahm & Schwartz (1977) speculated that this species appeared early in the geologic history of Hispaniola and that its currently restricted range was a result of the arrival of subsequent species that out-competed them. The other theory proposed to explain the small range of Caribicus warreni was that they had always had a small range, and species that arrived later were more successful because they never had to compete with C. warreni ( Strahm & Schwartz 1977) . Schwartz et al. (1979) speculated that this species was not rare, even though the first specimen was collected in 1917 and not reported again until the collection of the holotype in 1968.

Sexual dimorphism has been reported in this species, with males having larger heads ( Lawler & Norris 1979; Incháustegui et al. 1985). Antagonistic behavior between captive males has also been reported and was speculated to be related to the sexual dimorphism in head size ( Lawler & Norris 1979).

Caribicus warreni is included in our genetic dataset and has significant support in both Bayesian and ML likelihood analyses at the crown node of the species and the stem node that places it as the closest relative to C. darlingtoni . Based on our timetree (Fig. 4), C. darlingtoni diverged from its closest relative 6.83 Ma, consistent with typical species of vertebrates (> 0.7 Ma; Hedges et al. 2015). Caribicus warreni was recognized as a distinct species by our ASAP analysis.

Genus Celestus Gray, 1839

Jamaican Forest Lizards

Celestus Gray, 1839:288 . Type species: Celestus striatus Gray, 1839:288 , by original designation.

Macrogongylus Werner, 1901:299 . Type species Macrogongylus brauni Werner, 1901:299 , by original designation.

Diagnosis. Species of Celestus have (1) a dorsal pattern of absent/flecks in series/irregular dots/dots in chevrons/ mottled/chevrons/bands/bicolored, (2) head markings absent/present, (3) markings in the longitudinal paramedian area absent/present, (4) dots arranged in bars in the lateral band absent/present, (5) an adult SVL of 54.0– 367 mm, (6) ventral scale rows, 87–151, (7) midbody scale rows, 35–59, (8) total lamellae on one hand, 25–66, (9) total strigae on ten scales, 64–398, (10) relative length of all digits on one hindlimb, 16.6–37.8 %, (11) relative distance between the angled subocular and mouth, 0.339–1.66 %, (12) relative eye length, 1.83–5.17 %, (13) relative forelimb length, 12.8–26.7 %, (14) relative ear width, 0.446–2.45 %, (15) relative rostral height, 1.41–2.35 %, (16) relative head length, 14.6–22.9 %, (17) relative mental width, 1.28–2.35 %, (18) relative postmental width, 2.47–3.81 %, (19) relative cloacal width, 6.59–11.2 %, (20) relative prefrontal width, 3.93–5.68 %, (21) relative largest supraocular width, 1.69–4.03 %, (22) relative longest finger length, 2.92–7.48 %, (23) relative distance between the ear and eye, 6.07–10.9 %, (24) relative head width, 64.6–82.1 %, (25) relative frontal width, 57.3–95.5 %, (26) relative nasal height, 0.726–1.62 %, (27) relative angled subocular height, 0.553–1.61 %, (28) relative distance between the eye and naris, 4.25–6.51 %, (29) relative canthal iii length, 1.54–2.16 %, (30) relative angled subocular width, 1.63–2.90 %, and (31) relative nasal length, 1.11–2.33 %.

Content. Fourteen species (Table 3); Celestus barbouri , C. capitulatus sp. nov., C. crusculus , C. duquesneyi , C. hesperius sp. nov., C. hewardi , C. jamesbondi sp. nov., C. macrolepis , C. macrotus , C. microblepharis , C. molesworthi , C. occiduus , C. oligolepis sp. nov., and C. striatus .

Distribution: Celestus occurs almost entirely on Jamaica, with a single species ( C. macrotus ) on Hispaniola (Figs. 11–12).

FIGURE 11. Map showing the distribution of seven species of Celestus in Jamaica (main map) and C. macrotus in Hispaniola (inset, showing the border region between Haiti and the Dominican Republic). Hollow symbols indicate unexamined records. See figure 12 for additional species of Celestus .

FIGURE 12. Map showing the distribution of six species of Celestus in Jamaica. Hollow symbols indicate unexamined records assignable to species. Small black dots indicate unexamined museum records not assignable to species. See figure 11 for additional species of Celestus .