Celestus striatus Gray, 1839

Schools, Molly & Hedges, Blair, 2024, A new forest lizard fauna from Caribbean islands (Squamata, Diploglossidae, Celestinae), Zootaxa 5554 (1), pp. 1-306 : 100-107

Celestus striatus Gray, 1839

Golden Forest Lizard

(Fig. 33–34)

Celestus striatus — Gray, 1839:288. Holotype: BMNH 1946.8.8.3 (locality unknown).

Diploglossus Cliftii — Duméril & Bibron, 1839:595.

Celestus striatus — Gray, 1845:117.

Celestus striatus — Bocourt, 1879:377

Diploglossus striatus — Boulenger, 1885:289.

Diploglossus striatus —Fischer, 1888:29.

Celestus occiduus — Barbour, 1910:297 (part).

Diploglossus fowleri — Schwartz, 1971a:3. Holotype: MCZ R-125601, collected by Danny C. Fowler at Windsor (near Windsor Cave , “ 153 m ”), Trelawny Parish, Jamaica, on 15 August 1970 (18.35, -77.65).

Celestus fowleri — Schwartz & Henderson, 1991:373.

Celestus fowleri — Hedges et al., 2019:17.

Celestus fowleri — Schools & Hedges, 2021:220.

Celestus striatus — Schools & Hedges, 2021:220.

Celestus fowleri — Landestoy et al., 2022: 204.

Celestus striatus — Landestoy et al., 2022: 205.

Material examined (n=3). JAMAICA. BMNH 1946.8 .8.3. Trelawny. KU 226528, Windsor, 12 July 1961; MCZ R-125601, Danny C. Fowler, Windsor, 15 August 1970 .

Diagnosis. Celestus striatus has (1) a dorsal pattern of absent/chevrons, (2) head markings absent/present, (3) markings in the longitudinal paramedian area absent/present, (4) dots arranged in bars in the lateral band absent, (5) an adult SVL of 145 mm, (6) ventral scale rows, 101–109, (7) midbody scale rows, 41–43, (8) total lamellae on one hand, 59–66, (9) total strigae on ten scales, 279, (10) relative length of all digits on one hindlimb, 37.8 %, (11) relative distance between the angled subocular and mouth, 0.710 %, (12) relative eye length, 3.85 %, (13) relative forelimb length, 26.1 %, (14) relative ear width, 1.30 %, (15) relative rostral height, 1.94 %, (16) relative head length, 18.9 %, (17) relative mental width, unavailable, (18) relative postmental width, unavailable, (19) relative cloacal width, 7.93 %, (20) relative prefrontal width, 5.68 %, (21) relative largest supraocular width, 2.63 %, (22) relative longest finger length, 7.48 %, (23) relative distance between the ear and eye, 9.00 %, (24) relative head width, 82.1 %, (25) relative frontal width, 76.5 %, (26) relative nasal height, 1.08 %, (27) relative angled subocular height, 1.12 %, (28) relative distance between the eye and naris, 6.16 %, (29) relative canthal iii length, 2.12 %, (30) relative angled subocular width, 2.29 %, and (31) relative nasal length, 1.59 %. The species stem time is 3.49 Ma and the species crown time is 0.30 Ma (Fig. 4).

Celestus striatus has a larger relative length of all digits on one hindlimb (37.8), relative forelimb length (26.1), relative prefrontal width (5.68), relative longest finger length (7.48), larger relative head width (82.1), and relative distance between the eye and naris (6.16) than most other species of the genus.

From C. barbouri , we distinguish C. striatus by the adult SVL (145 versus 78.4–93.6), the ventral scale rows (101–109 versus 118–151), the midbody scale rows (41–43 versus 47–56), the total lamellae on one hand (59–66 versus 36–49), the total strigae on ten scales (279 versus 105–136), the relative length of digits on one hindlimb (37.8 versus 18.2–23.5), the relative distance between angled subocular and mouth (0.710 versus 0.437 –0.556), the relative eye length (3.85 versus 2.87–3.63), the relative forelimb length (26.1 versus 15.4–19.0), the relative rostral height (1.94 versus 1.41–1.66), the relative head length (18.9 versus 14.6–16.6), the relative prefrontal width (5.68 versus 3.97–4.33), the relative longest finger length (7.48 versus 2.92–3.81), the relative distance between the ear and eye (9.00 versus 6.23–7.15), the relative head width (82.1 versus 73.8–81.7), the relative distance between the eye and naris (6.16 versus 4.68–4.83), and the relative width of canthal iii (2.12 versus 1.54–1.93). From C. capitulatus sp. nov., we distinguish C. striatus by the dorsal pattern (absent/chevrons versus irregular dots/dots in chevrons), the adult SVL (145 versus 62.1–81.8), the total lamellae on one hand (59–66 versus 25–38), the total strigae on ten scales (279 versus 105–192), the relative length of digits on one hindlimb (37.8 versus 17.6–22.3), the relative eye length (3.85 versus 2.75–3.80), the relative forelimb length (26.1 versus 14.3–18.1), the relative head length (18.9 versus 15.1–17.7), the relative prefrontal width (5.68 versus 4.30–4.72), the relative largest supraocular width (2.63 versus 2.03–2.61), the relative longest finger length (7.48 versus 3.45–3.75), the relative distance between the ear and eye (9.00 versus 6.45–7.84), the relative head width (82.1 versus 71.6–78.6), the relative frontal width (76.5 versus 78.1–81.6), the relative angled subocular height (1.12 versus 0.586–1.01), and the relative distance between the eye and naris (6.16 versus 4.57–5.03). From C. crusculus , we distinguish C. striatus by the adult SVL (145 versus 59.6–77.6), the total lamellae on one hand (59–66 versus 30–39), the total strigae on ten scales (279 versus 106–194), the relative length of digits on one hindlimb (37.8 versus 18.7–24.7), the relative eye length (3.85 versus 2.93–3.61), the relative forelimb length (26.1 versus 12.8–20.7), the relative prefrontal width (5.68 versus 3.93–4.67), the relative longest finger length (7.48 versus 2.94–4.10), the relative distance between the ear and eye (9.00 versus 6.07–8.61), the relative frontal width (76.5 versus 82.6–91.1), the relative distance between the eye and naris (6.16 versus 4.31–4.86), and the relative width of canthal iii (2.12 versus 1.59–2.07). From C. duquesneyi , we distinguish C. striatus by the dorsal pattern (absent/chevrons versus bands), the adult SVL (145 versus 62.1), the total strigae on ten scales (279 versus 130), and the relative ear width (1.30 versus 2.45). From C. hesperius sp. nov., we distinguish C. striatus by the dorsal pattern (absent/chevrons versus dots in chevrons), the adult SVL (145 versus 54.0–62.3), the ventral scale rows (101–109 versus 111–114), the total lamellae on one hand (59–66 versus 29–34), and the total strigae on ten scales (279 versus 95–122). From C. hewardi , we distinguish C. striatus by the dorsal pattern (absent/chevrons versus mottled/bands), the ventral scale rows (101–109 versus 113–137), the relative length of digits on one hindlimb (37.8 versus 24.1–30.6), the relative distance between angled subocular and mouth (0.710 versus 0.744–1.40), the relative forelimb length (26.1 versus 22.2–24.6), the relative ear width (1.30 versus 1.40–1.82), the relative rostral height (1.94 versus 1.50–1.76), the relative cloacal width (7.93 versus 8.81–9.89), the relative prefrontal width (5.68 versus 4.18–4.80), the relative longest finger length (7.48 versus 5.03–5.66), the relative distance between the ear and eye (9.00 versus 6.72–8.73), the relative head width (82.1 versus 68.4–77.1), the relative frontal width (76.5 versus 57.3–75.3), the relative nasal height (1.08 versus 1.21–1.24), the relative distance between the eye and naris (6.16 versus 5.00–5.60), and the relative angled subocular width (2.29 versus 1.63–2.23). From C. jamesbondi sp. nov., we distinguish C. striatus by the adult SVL (145 versus 54.7–72.0), the total lamellae on one hand (59–66 versus 30–36), the total strigae on ten scales (279 versus 101–173), the relative length of digits on one hindlimb (37.8 versus 19.8–26.3), the relative forelimb length (26.1 versus 14.4–19.9), the relative prefrontal width (5.68 versus 4.29–5.09), the relative longest finger length (7.48 versus 3.66–4.33), the relative distance between the ear and eye (9.00 versus 6.92–7.80), the relative head width (82.1 versus 76.0–80.8), the relative nasal height (1.08 versus 1.12–1.21), and the relative distance between the eye and naris (6.16 versus 4.25–5.54). From C. macrolepis , we distinguish C. striatus by the dorsal pattern (absent/chevrons versus bicolored), adult SVL (145 versus 254–316), the ventral scale rows (101–109 versus 112–116), the midbody scale rows (41–43 versus 46–48), the total lamellae on one hand (59–66 versus 52–54), and the relative distance between angled subocular and mouth (0.710 versus 1.39–1.66). From C. macrotus , we distinguish C. striatus by the dots arranged in bars in the lateral areas (absent versus present), the adult SVL (145 versus 60.0–86.1), the ventral scale rows (101–109 versus 87–93), the total lamellae on one hand (59–66 versus 39–40), the total strigae on ten scales (279 versus 64–115). From C. microblepharis , we distinguish C. striatus by the total lamellae on one hand (59–66 versus 30), the total strigae on ten scales (279 versus 165), the relative length of digits on one hindlimb (37.8 versus 16.6), the relative eye length (3.85 versus 1.83), the relative forelimb length (26.1 versus 14.2), the relative ear width (1.30 versus 0.446), and the relative longest finger length (7.48 versus 3.11). From C. molesworthi , we distinguish C. striatus by the dorsal pattern (absent/chevrons versus dots in chevrons), the adult SVL (145 versus 78.1–103), the total lamellae on one hand (59–66 versus 32–44), the total strigae on ten scales (279 versus 138–159), the relative length of digits on one hindlimb (37.8 versus 22.4–29.4), the relative eye length (3.85 versus 3.28–3.70), the relative forelimb length (26.1 versus 17.5–24.2), the relative ear width (1.30 versus 1.37–1.50), and the relative rostral height (1.94 versus 1.72–1.81). From C. occiduus , we distinguish C. striatus by the adult SVL (145 versus 269–367), the midbody scale rows (41–43 versus 46–56), the relative distance between angled subocular and mouth (0.710 versus 1.26–1.27). From C. oligolepis sp. nov., we distinguish C. striatus by the dorsal pattern (absent/chevrons versus dots in chevrons), the ventral scale rows (101–109 versus 98), the midbody scale rows (41–43 versus 35), and the total lamellae on one hand (59–66 versus 30).

Description of holotype. BMNH 1946.8.8.3. An adult; SVL 145 mm; tail slightly compressed, 138 mm (95.2% SVL); axilla-to-groin distance 68.5 mm (47.2% SVL); forelimb length 37.8 mm (26.1% SVL); hindlimb length 53.9 mm (37.2% SVL); head length 27.3 mm (18.8% SVL); head width 22.4 mm (15.4% SVL); head width 82.1% head length; diameter of orbit 5.58 mm (3.85% SVL); horizontal diameter of ear opening 1.88 mm (1.30% SVL); vertical diameter of ear opening 1.59 mm (1.10% SVL); length of all toes on one foot 54.8 mm (37.8% SVL); shortest distance between angled subocular and lip 1.03 mm (0.710% SVL); shortest distance between the ocular and auricular openings 13.1 mm (9.03% SVL); longest finger length 10.9 mm (7.52% SVL); largest supraocular width 3.81 mm (2.63% SVL); cloacal width 11.5 mm (7.93% SVL); prefrontal width 8.23 mm (5.68% SVL); frontal width 76.5% SVL; nasal height 1.56 mm (1.08% SVL); angled subocular height 1.62 mm (1.12% SVL); shortest distance between the eye and naris 8.93 mm (6.16% SVL); canthal iii width 3.08 mm (2.12% SVL); angled subocular width 3.32 mm (2.29% SVL); nasal width 2.30 mm (1.59% SVL); rostral 1.94X 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 concave posterior margin, wider than long, bordered by posterior internasals, 1 st loreals, canthal iii, 1 st median oculars, and the frontal; frontal much 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, which is wider than long; parietal separated from supraoculars by 1 st and 2 nd temporals and frontoparietal (left)/(right); nasal single; nostril just posterior to 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 –5 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); final 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, prefrontal/frontonasal complex, and 1 st and 2 nd loreals (left)/(right); 11 (left)/10 (right) median oculars, 1 st contacting the prefrontal (left)/(right); 1 upper preocular (left)/(right); an irregular anterior supraciliary (left)/(right); 6 lateral oculars (left)/(right); 5 temporals (left)/(right); 2 suboculars (left)/(right); posterior subocular large and elongate (left)/(right); anterior subocular small (left)/(right); 11 supralabials (left)/(right), 7 to level below center of eye (left)/(right); 8 (left)/9 (right) infralabials, 6 (left)/7 (right) to level below center of eye; mental small, followed by a single, much larger postmental; 5 pairs of enlarged chin shields, followed by 1 pair of reduced chin shields; 1 st pair in contact with one another; 2 nd –5 th pairs separated by 1–7 scales; 97 transverse rows of dorsal scales from interoccipital to base of tail; 109 transverse rows of ventral scales from mental to vent; 43 scales around midbody; 5 digits; finger lengths 3>4>2>5>1; 16 lamellae under longest finger (left)/(right); 66 total lamellae on one hand; toe lengths 4>3>5>2>1; 23 (left)/22 (right) lamellae under longest toe; faintly striated and keeled dorsal body and caudal scales; smooth ventral scales; 279 total strigae counted on ten scales.

Color (in alcohol): dorsal surface of head dark gold, patternless; lateral surfaces of head grading from dark gold to slightly paler gold, patternless; dorsal surfaces of the body are dark gold, patternless; dorsal surface of tail the same dark gold as the head, patternless; lateral areas are the same dark gold as the head, patternless; dorsal surfaces of the limbs are dark gold, patternless; lateral and ventral areas of the limbs fade to pale gold, patternless; ventral surfaces of the head, body, and tail are pale gold, patternless.

Color (in life): From the photograph by Rudi Diesel that he provided to SBH in the 1990s, and has since been on the Caribherp website ( Hedges 2024), it can be seen that the ground color is golden tan with dark brown markings and pale tan accenting of the dorsal crossbands/chevrons. The iris is orange-red.

Variation. Both KU 226528 and MCZ R-125601 have a much more pronounced dorsal pattern of chevrons than the holotype, which is patternless (probably faded in preservative) . MCZ R-125601 and KU 226528 have darker outlines on the borders of their head scales . KU 226528 also has large dots in a longitudinal paramedian series and no dots in the lateral band. Measurements and other morphological data for the holotype and other examined material are presented in Table 1 .

Distribution. Celestus striatus is only known from three specimens, two of which were found at Windsor in North-central Jamaica in the Cockpit country, 160 m (Fig. 11). The type locality is simply reported as “West Indies” ( Gray 1845; Boulenger 1885).

Ecology and conservation. Little data exist on the ecology of this species. Both the holotype and paratype of Celestus fowleri (synonymized herein with C. striatus ) were collected from bromeliads in the Cockpit region of Jamaica (near Windsor Cave) up to 2.5 m off the ground ( Schwartz 1971a), suggesting that it is an arboreal species. Celestus fowleri was the first Antillean diploglossid to be found in bromeliads, a niche “virtually unoccupied by Antillean reptiles” ( Schwartz 1971a). At the time of collection, Schwartz (1971a) had little doubt that the bromeliad was a diurnal retreat for the species and mentioned that it was possible that this species was in fact terrestrial, only using the bromeliads as a refuge during the day. This was followed by the caveat that there was in fact no evidence to support this hypothesis. Like Crombie (1999), one of us (SBH) has never seen this species despite examining hundreds of bromeliads during repeated visits to the type locality and many other places in the Cockpits. In the 1990s, researchers studying the bromeliad crab at Windsor encountered and photographed C. striatus in a bromeliad at Windsor (R. Diesel, pers. comm. to SBH) and sent photographs to one of us (SBH), reproduced here (Fig. 34). The species is likely extant but very rare.

We consider the conservation status of Celestus striatus to be Critically Endangered B1ab(iii) because of its extreme rarity and greatly restricted distribution, probably related to some aspect of human impact (habitat alteration, introduced predators, etc.), based on IUCN Redlist criteria ( IUCN 2023). Currently, IUCN 2023 considers this species to be Vulnerable D2 because “this species is endemic to Jamaica, known only from around the type locality, Windsor Cave, in the Cockpit Country, Trelawny Parish; at an elevation near 160 m asl ( Schwartz 1970; Henderson & Powell 2009). The species may occur at other localities in the Cockpits; however, exhaustive searches in bromeliads by herpetologists through the region have not yet yielded any records (S.B. Hedges pers. comm. 2015).” Studies are needed to determine the health and extent of remaining populations and threats to the survival of the species.

Reproduction. No data on reproduction are available for this species.

Etymology. The species name ( striatus ) is a masculine Latin adjective meaning “striate,” in reference to the scales of this species, which were noted by Gray (1839) as striate but not keeled.

Remarks. As we summarized elsewhere ( Schools & Hedges 2021), the taxonomic history of this species is confusing because it has been synonymized with other species including Celestus occiduus (as was C. hewardi ) and Comptus stenurus . It differs morphologically from other species in multiple characters (see Diagnosis section) and we retain its classification as an independent species, clearly part of the Jamaican radiation of Celestus . It is also the type species of the genus Celestus .

Remarkable is the fact that this important taxon, which carries the generic and subfamilial names and was known from a single specimen for most of the last two centuries, has gone from essential obscurity to being extant. Several factors have contributed to this delay or solved the mystery. First, the species was probably limited in distribution or rarely encountered even before introduction of the mongoose. An arboreal species would have been more difficult to collect than terrestrial species like C. occiduus or C. crusculus . Secondly, Straham & Schwartz (1977) considered C. striatus to be Central American based on its unusual scalation, which was an error stemming from a misinterpretation of a photograph of the type specimen (see Schools & Hedges 2021). The third factor, and the one that definitively solved this mystery, was the collection of DNA sequences from these old specimens (holotypes of C. striatus and C. fowleri ), which was made possible by forward-thinking museums and curators, who permitted this research to take place.

FIGURE 33. (A–F) Celestus striatus (BMNH 1946.8.8.3, holotype), SVL 145 mm.

FIGURE 34. (A–B) Celestus striatus (vouchers not available), in life. From near Windsor Cave, Trelawny Parish, Jamaica. Photos by Rudolf Diesel.

The type specimen now has no discernable pattern, possibly because it has been in preservative for nearly two centuries. Gray (1839, 1845) described it as “silvery” and “bleached,” and named it “The Golden Galliwasp.” However, he also noted that it was “brown-varied” and “sides brown-spotted.” Boulenger’s (1885) illustration shows distinct side barring and thin cross-bands across the back, consistent with Gray’s description. Because Boulenger listed synonyms and additional specimens of C. striatus (unsupported by us), the pattern shown in his illustration might be a composite of multiple specimens of different species. However, the habitus of the lizard in the illustration resembles the type specimen and the simplest explanation is that in life the animal originally had a pale ground color (tan or golden) with the markings exactly as shown by Boulenger (1885), and that all remnants of pattern have since faded with time. The photograph in life of Celestus striatus (Fig. 34) is clear only at the anterior end of the lizard, but it shows a golden tan ground color with dark brown marks on the side of the head and thin dark brown crossbands or chevrons not unlike those illustrated by Boulenger.

Herein, we synonymize Celestus fowleri with Celestus striatus based on genetic and morphological data. The original specimen of Celestus fowleri collected was a juvenile in 1961, prior to the collection of the holotype in 1970 ( Schwartz 1971a). This extended time between collection dates led Schwartz (1971a) to speculate that C. fowleri was, “extremely uncommon, remarkably elusive, or ecologically or altitudinally restricted.” As members of the “long-legged” group of Jamaican lizards, Schwartz (1971a) suspected that a close relationship existed between Celestus duquesneyi , C. fowleri , and C. hewardi , suggesting that C. duquesneyi and C. fowleri were either ecological or geographic isolates of C. hewardi .

The holotypes of Celestus striatus (BMNH 1946.8.8.3) and Celestus fowleri (MCZ R-125601) are included in our genetic dataset and their sequences are similar, indicating that they belong to a single species, C. striatus . This is not surprising because of their similar morphology, including scale counts, showing adaptations (long limbs and long digits) for climbing trees. The species 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 of the much larger C. occiduus . Based on our timetree (Fig. 4), C. striatus diverged from its closest relative 3.49 Ma, consistent with typical species of vertebrates (> 0.7 Ma; Hedges et al. 2015). Celestus striatus was recognized as a distinct species by our ASAP analysis.

Genus Comptus Schools & Hedges, 2021

Caribbean Rough-scaled Forest Lizards

(Fig. 35)

Type species. Diploglossus stenurus Cope, 1862:188 .

FIGURE 35. Map showing the distribution of Comptus . The black arrow indicates an instance of sympatry between C. alloeides and C. weinlandi . Hollow symbols indicate unexamined records assignable to species.

Diagnosis. Species of Comptus have (1) a dorsal pattern of absent/irregular dots/dots in series/dots in chevrons/ mottled/chevrons, (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) a maximum SVL of 60.1–161 mm, (6) ventral scale rows, 81–110, (7) midbody scale rows, 36–45, (8) total lamellae on one hand, 32–58, (9) total strigae on ten scales, 143–323, (10) relative length of all digits on one hindlimb, 23.4–39.7 %, (11) relative distance between the angled subocular and mouth, 0.403–1.12 %, (12) relative eye length, 2.83–4.43 %, (13) relative forelimb length, 19.3–27.9 %, (14) relative ear width, 0.451–2.18 %, (15) relative rostral height, 1.46–2.42 %, (16) relative head length, 14.7–20.0 %, (17) relative mental width, 0.840–1.95 %, (18) relative postmental width, 2.32–3.05 %, (19) relative cloacal width, 8.29–10.3 %, (20) relative prefrontal width, 3.95–5.47 %, (21) relative largest supraocular width, 2.23–3.49 %, (22) relative longest finger length, 4.14–7.19 %, (23) relative distance between the ear and eye, 6.03–9.51 %, (24) relative head width, 62.8–82.2 %, (25) relative frontal width, 58.8–86.5 %, (26) relative nasal height, 0.863–1.32 %, (27) relative angled subocular height, 0.568–1.33 %, (28) relative distance between the eye and naris, 4.38–6.77 %, (29) relative canthal iii length, 1.37–2.42 %, (30) relative angled subocular width, 1.64–3.36 %, and (31) relative nasal length, 1.39–2.03 %.

Content. Six species (Table 3): Comptus alloeides , C. arboreus sp. nov., C. badius , C. maculatus , C. stenurus , and C. weinlandi .

Distribution. Comptus occurs in the Cayman Islands, Navassa Island, and throughout most of Hispaniola, including the associated islets of Ile-a-Vache, Île à Cabrit, and Ile Grande Cayemite.