Spondylurus semitaeniatus ( Wiegmann 1837 ) Hedges, S. Blair & Conn, Caitlin E., 2012

Hedges, S. Blair & Conn, Caitlin E., 2012, A new skink fauna from Caribbean islands (Squamata, Mabuyidae, Mabuyinae) 3288, Zootaxa 3288 (1), pp. 1-244 : 194-201

Spondylurus semitaeniatus ( Wiegmann 1837) comb. nov.

Lesser Virgin Islands Skink

( Figs. 73D , 76B , 78 )

Euprepes semitaeniatus — Wiegmann, 1837:135. Holotype from "America," ex. coll. Marcus Elieser Bloch (see Remarks).

Gongylus (Eumeces) agilis —Reinhardt & Luetken, 1863:229 (part).

Euprepes semitaeniatus —Peters, 1864:50 (holotype clarified as ZMB 1238, same specimen as examined by Schneider [1801:181], for syntype of Scincus auratus ; originally from collection of Bloch).

E[uprepes] semitaeniatus — Peters, 1871:400 (part, inferred).

Mabuya sloanii — Bocourt, 1879:401 (part).

Mabuia sloanii — Boulenger, 1887:193 (part, inferred).

Mabuya semitaeniata — Stejneger, 1904:610 (claimed as distinct species, but listed incorrect accession number for holotype).

Mabuya sloanii — Barbour, 1914:320 (part).

Mabuya sloanii — Schmidt, 1928:121 (part).

Mabuya sloanii —Barbour, 1930:105 (part).

Mabuya semitaeniatus — Grant, 1931:218 (Culebra and Mona, in error; listed incorrect accession number for holotype).

Mabuya mabouia — Barbour, 1935:129 (part).

Mabuya mabouya sloanii — Dunn, 1936:544 (part; placed Euprepes semitaeniatus in synonymy but listed incorrect accession number for holotype).

Mabuya mabouia — Barbour, 1937:147 (part).

Mabuya mabouya sloanei — Schwartz & Thomas, 1975:141 (part; placed Euprepes semitaeniatus in synonymy but listed incorrect accession number for holotype).

Mabuya mabouya sloanei — MacLean et al., 1977:30–35 (part).

Mabuya mabouya sloanei — Heatwole et al., 1981:34 (part).

Mabuya mabouya sloanei — Schwartz & Henderson, 1988:151 (part; placed Euprepes semitaeniatus in synonymy but listed incorrect accession number for holotype).

Mabuya mabouya sloanei — Schwartz & Henderson, 1991:457 (part).

Mabuya bistriata — Powell et al., 1996:82 (part).

Mabuya sloanii — Mayer & Lazell, 2000:883 (part).

Mabuya sloanii —Miralles, 2005:49 (part).

Mabuya sloanii — Henderson & Powell, 2009:293 (part).

Material examined (n = 53). U.S. Virgin Islands. ZMB 1238 (holotype), from "America," type-locality restricted here to St. Thomas (no specific locality within island) , U.S. Virgin Islands, ex. coll. Marcus Elieser Bloch, probably acquired 1779–1799 (examination of photographs) (see Remarks ); MCZ R-36592–95 and UMMZ 73821 , Chapman Grant , Little Buck Island , between 2 September 1931 and 18 April 1932; MCZ R-42380, Chapman Grant , Capella Island , 29 April 1936; UMMZ 80585 , Chapman Grant , Little Buck Island, 15 March 1936; ZMH R09300–301 , St. Thomas, " Calwood leg. 1885." British Virgin Islands. KU 242064, Albert Schwartz , Great Camanoe Island (between Lee Bay and Cam Bay), 20 August 1964; KU 242071, Albert Schwartz , Virgin Gorda (southeast part of island from Copper Mine Bay ; no collection date available); KU 242072–73 and 242075–78, Albert Schwartz , Virgin Gorda (just north of Garden Rock ; no collection date available); KU 242074 and 242079– 81, Albert Schwartz , Virgin Gorda (inland margin of Salt Pond behind St. Thomas Bay ; no collection date available); MCZ R-166975, James Lazell , Guana Island , 13 July 1984; MCZ R-170883, P. Shelby, Guana Island , 21 October 1984; MCZ R-176327, G. Proctor, Guana Island , 20 November 1986; MCZ R-176328, L. Phipps, Guana Island, 27 December 1987; MCZ R-176329, C. O. Connell, Guana Island (Shangri La), 19 July 1987; MCZ R-176330, E. Azevedo , Tortola ( Zion Hill ), 24 July 1987; MCZ R-176331, R. Jenkins and J. Randall, Necker Island, 25 July 1987; MCZ R-176332, R. Jenkins , Tortola ( Sage Mountain ), 30 August 1988; MCZ R-176739–43, “Hocking Tech,” Tortola ( Sage Mountain ), 30 June 1994; MCZ R-180273, M. Garcia, Little Thatch Island, 9 October 1994; MPM 26275 , Virgin Gorda (inland margin of Salt Pond behind St. Thomas Bay ); RT 947, D. C. Leber and R. Thomas , Virgin Gorda (inland margin of Salt Pond behind St. Thomas Bay), 17 August 1964; UMMZ 80581–82 , 80584 , and 239599–600, Chapman Grant , Virgin Gorda, 30 March–7 April 1936; MCZ R-182093, E. Henry , Tortola ( Little Dicks Hill ), July 1996; UMMZ 200131 , Fred Kraus , Guana Island , 7 March 1991; UPRRP 5489 , Island Project Staff , Salt Island , 24 May 1966; UPRRP 5503 , Island Project Staff, Necker Island, 6 June 1966; UPRRP 5521–22 , Island Project Staff , Virgin Gorda ( Savanna Bay ), 27 June 1966; USNM 304550 , Ginger Island (near South Bay ), 12 October 1975 (no collector information available); MCZ R-185692, Clive Petrovic, Mosquito Island , 3 October 2007; USNM 576304 , K. Lindsay, Mosquito Island, 12 September 2007. West Indies. ZMUC-R 391 , "?Botanical Garden", accessioned 1893 .

Diagnosis. Spondylurus semitaeniatus is characterized by (1) maximum SVL in males, 74.7 mm; (2) maximum SVL in females, 82.9 mm; (3) snout width, 1.99–3.27% SVL; (4) head length, 15.8–19.4% SVL; (5) head width, 11.9–16.2% SVL; (6) ear length, 0.953–2.27% SVL; (7) toe-IV length, 8.33–12.0% SVL; (8) prefrontals, two (98%), four (2%); (9) supraoculars, three (1%), four (99%); (10) supraciliaries, three (2%), four (98%); (11) frontoparietals, two; (12) supralabial below the eye, five (28%), six (72%); (13) nuchal rows, one (14%), two (80%), three (6%); (14) dorsals, 57–65; (15) ventrals, 59–70; (16) dorsals + ventrals, 119–134; (17) midbody scale rows, 31–34; (18) finger-IV lamellae, 10–15; (19) toe-IV lamellae, 13–19; (20) finger-IV + toe-IV lamellae, 23–33; (21) supranasal contact, Y (96%), N (4%); (22) prefrontal contact, N; (23) supraocular-1/frontal contact, Y (38%), N (62%); (24) parietal contact, Y (98%), N (2%); (25) pale middorsal stripe, Y; (26) dark dorsolateral stripe, Y; (27) dark lateral stripe, Y; (28) pale lateral stripe, Y; and (29) palms and soles, pale ( Tables 3–5).

Within the Genus Spondylurus , S. semitaeniatus differs from all other species except S. anegadae sp. nov., S. culebrae sp. nov., S. lineolatus , S. monae sp. nov., and S. sloanii by having a higher dark dorsolateral stripe width/ middorsal stripe width ratio (1.54–3.36 versus 0.115–1.27 in those other species; Fig. 79 ). From S. anegadae sp. nov., it differs by having a narrower, longer nostril (Fig. 58). It is separated from S. culebrae sp. nov. by having a shorter length of combined head scales ( Fig. 62A ). It is distinguished from S. lineolatus by having a longer head (head length 15.8–19.4% SVL versus 12.9–14.4% SVL in S. lineolatus ) and by having two dark lateral stripes and two dark dorsolateral stripes (versus 10 dark stripes in S. lineolatus ). From S. monae sp. nov., S. semitaeniatus differs by having a shorter rostral scale ( Fig. 61 ).

Additionally, Spondylurus semitaeniatus is distinguished from other species in the genus except S. anegadae sp. nov. and S. lineolatus by having a middorsal stripe that is similar in color to the pale dorsolateral stripes (versus a middorsal stripe that is darker in those other species; Figs. 55 and 73 ). It is separated from S. fulgidus by having a higher number of supraciliaries (3–4 versus five in S. fulgidus ), fewer total digital lamellae (178–215 versus 238 in S. fulgidus ), and nearly non-overlapping dorsals + ventrals (119–134 versus 108–120 in S. fulgidus ). It differs from S. macleani ( Fig. 55G ) by having longer dark dorsolateral stripes. It is distinguished from S. monitae sp. nov. (Fig.

There are frequency differences that also separate Spondylurus semitaeniatus from other species. From S.

caicosae sp. nov., it differs by having a higher number of midbody scale rows (31–34 in 94% of specimens versus 27–30 in 92% of specimens belonging to S. caicosae sp. nov.). It differs from S. martinae sp. nov. by having fewer ventral scales (59–67 in 88% of specimens versus 68–71 in S. martinae sp. nov.).

The molecular phylogeny ( Fig. 5 ) shows that Spondylurus sloanii is closer, genetically, to S. culebrae sp. nov., S. macleani , and S. monitae sp. nov. than it is to S. semitaeniatus . However, the greatest confusion in identification of S. semitaeniatus will likely be with S. sloanii because the two species appear superficially similar and occur in close proximity and sympatry in the Virgin Islands. The most reliable character in separating these two species is the width of the dark dorsolateral stripes compared with the pale middorsal stripe as measured at the forelimbs instead of the normal location for this measurement, at the ears ( Fig. 80A ). In both species, the dark dorsolateral stripes taper posteriorly until they eventually disappear. However, in S. sloanii , the dark dorsolateral stripes start tapering more quickly, before the forelimbs (e.g., compare pattern in Fig. 78E with that in Fig. 81F ). The dark dorsolateral stripe/middorsal stripe ratio (at the forelimbs) is 1.25–2.68 in S. semitaeniatus versus 0.43–1.08 in S. sloanii . A second useful character in separating the two species, although not 100% diagnostic in itself, is prefrontal separation (length of frontonasal-frontal suture). Spondylurus semitaeniatus has> 0.3% separation of prefrontals whereas more than two thirds of S. sloanii have contact between prefrontals, or are within 0.3% SVL of contact ( Figure 80B ). In other aspects of pattern, adult S. semitaeniatus usually differ from S. sloanii in having a pale middorsal stripe that is the same color as the pale dorsolateral stripes (darker than the pale dorsolateral stripes in S. sloanii ), a dorsum that does not appear braided (versus dorsum with dark-edged scales giving a braided appearance in S. sloanii ), and a pale lateral stripe (absent or barely evident in S. sloanii ). Both species have been described as bronze or coppery, but the color of living and preserved S. semitaeniatus appears to be less so (more tan) than that of S. sloanii .

Description of holotype ( Figs. 78A–D ). Based on examination of photographs. An unsexed adult in moderate state of preservation, without injuries and without an abdominal slit. SVL 60.5 mm; tail length not measured (broken); HL, HW, SW, EL, and toe-IV length not measured; ear-opening average in size and oval; toe length in the following order: I <V <II <III <IV.

Head scalation. Rostral wider than high, contacting first supralabials, nasals and supranasals. Paired supranasals in median contact, contacting anteriormost loreal. Frontonasal diamond-shaped, wider than long, laterally in contact with anterior loreal scale. A pair of quadrilateral prefrontals, separated medially, and in contact with frontonasal, both anterior and posterior loreals, first supraciliary, first and second supraoculars, and frontal. Frontal heptagonal, in contact with the second supraoculars and paired frontoparietals. Frontoparietals also in contact with parietals and interparietal. Interparietal tetragonal and lanceolate, separated from nuchals by parietals; parietal eye not distinct. Parietals in contact with upper secondary and tertiary temporal scales. Four supraoculars, the second one being the longest and largest. Four supraciliaries, the second the longest. Nostril in posterior part of the nasal. A small postnasal, bordered by supranasal, anterior loreal and first supralabial. Anterior and posterior loreals squarish with posteromedial projection on latter. Two upper preoculars and two lower preoculars. Seven supralabials, the fifth being the widest and forming the lower border of the eyelid. Five moderately enlarged scales behind eye comprising the postoculars; similar to temporal scales. One primary temporal, two secondary temporals, and three tertiary temporals; all imbricate, smooth, cycloid, not distinctly delimited from the scales on the nape and the sides of the neck. Seven infralabials. Mental scale wider than long, posterior margin curved slightly toward tip of snout. Postmental scale and one pair of adjoining chin shields in contact with anterior infralabials. First pair of chin shields in contact medially; second and third pairs separated by a smaller cycloid scale.

Body and limb scalation. One row of paired nuchal scales and one additional right nuchal scale. Other scales on nape similar to dorsals. On lateral sides of neck, scales slightly smaller. Dorsal scales cycloid, imbricate, smooth, 60 in a longitudinal row; ventrals similar to dorsals; 65 in a longitudinal row; scales around midbody not counted. No distinct boundaries between dorsals, laterals and ventrals. Scales on tail and limbs similar to dorsals, except smaller on limbs. Palmar and plantar regions with small tubercles, subequal in size and delimited by a surrounding region of flatter scales. Subdigital lamellae smooth, single, not countable on fingers or toes. Preanal scales similar to ventrals. No enlarged median subcaudal scales on tail.

Pattern and coloration. Pattern elements weakly defined because of age of specimen (232–212 years old). Dorsal ground color medium greenish-brown with no visible spots. Dark dorsolateral stripes present, wide (not measured), dark brown, extending from tip of snout to first third of body. Dark lateral stripes present, dark brown, extending from behind eye to first third of body (pattern no longer visible posterior to this point). Pale middorsal stripe present, narrow (not measured), medium greenish-brown, extending from tip of snout to first third of body. Pale dorsolateral stripes present, medium greenish-brown, extending from tip of snout to first third of body. Pale lateral stripes absent. Ventral surface of body without pattern. Palmar and plantar surfaces unpigmented. No information is available for coloration of the holotype in life.

Variation. In scalation, most specimens resembled the holotype ( Tables 4). In more recent material, where pattern was more visible, dorsal pattern varied in color and in distribution of small dark brown spots. Limbs appear darker than body, and forelimbs are usually mottled. A Virgin Gorda specimen ( KU 242071) was noted by Albert Schwartz as having a copper to bronzy dorsum posteriorly, and venter bluish, becoming grayer posteriorly. A live animal also from Virgin Gorda ( Fig. 78E ) shows a tan or reddish tan dorsum with orange anteriorly in the zone of the dark lateral stripe (which appears broken and spotty) .

Distribution. This species is widely distributed in the U.S. and British Virgin Islands ( Fig. 10E ). In the U.S. Virgin Islands it is known from St. Thomas, and from two islets of St. Thomas (Capella and Little Buck). In the British Virgin Islands it is known from Fallen Jerusalem (literature record), Ginger Island, Great Camanoe Island, Guana Island, Little Thatch Island, Mosquito Island, Necker Island, Round Rock (literature record), Salt Island, Tortola, and Virgin Gorda ( MacLean et al. 1977; MacLean 1982; Lazell 1983; Schwartz & Henderson 1991; Lazell 1995). Reinhardt and Lütken (1863) mentioned the occurrence of skinks on St. John and Jost van Dyke, citing "Mr. Riises," = Albert Heinrich Riise, a Danish pharmacist and naturalist on St. Thomas. The only museum specimens that we could locate from those islands were two Spondylurus spilonotus from St. John (ZMUC-R 93–94).

Ecology and conservation. Recent reviews of the herpetofauna of the Virgin islands implicate the mongoose in declines and extirpations of skinks, especially from the large islands ( Perry & Gerber 2006; Platenberg & Boulon 2006). On the small islets of St. Thomas (Little Buck Island, Capella Island) where Spondylurus semitaeniatus occurs, the habitat is mostly coastal shrub with introduced Guinea Grass Panicum maximum , Turks Cap Cactus Melocactus intortus , and the shrub Oplonia spinosa , interspersed with Sea Grape Coccoloba uvifera , the same as the habitat of S. sloanii . This habitat can be described as low shrubby vegetation or grass, including exposed rocky areas and occasional beaches (R. Platenberg, personal communication).

No specimens of Spondylurus semitaeniatus exist from the large island of St. John, but it probably occurred there, based on its distribution on surrounding islands. Two specimens of the larger species, S. spilonotus , are from St. John, collected in 1846, and that species has not been recorded there since, almost certainly having been extirpated by the mongoose. The same fate can be assumed for S. semitaeniatus on St. John. Spondylurus semitaeniatus has been observed in recent decades on mongoose-free islands such as Virgin Gorda, Guana, Little Buck, and Mosquito as well as mongoose-inhabited Tortola. There is general acceptance that it has declined in numbers ( Perry & Gerber 2006; Platenberg & Boulon 2006), and continued development of the islands will reduce available habitat of a species already living a fragile existence. Some anecdotal evidence suggests that, while the mongoose is the primary factor responsible for decline in skink numbers in the Virgin Islands, it is not the only factor (see also the account for S. magnacruzae sp. nov.).

Based on IUCN Redlist criteria ( IUCN 2011), we assess the conservation status of Spondylurus semitaeniatus as Endangered (EN A2ace). It faces a primary threat from the introduced mongoose, which probably led to its extirpation on St. Thomas, and probably other islands where there are no museum records as evidence, such as St. John. Secondary threats include habitat destruction from agriculture and urbanization, and predation from other introduced mammals, including black rats. Studies are needed to determine the health of remaining populations, and threats to the survival of the species. Captive breeding programs should be considered.

Reproduction. Three females (71.1–77.8 mm SVL) contained 2–4 (mean = 3) developing young. The date of collection for one specimens was 7 April 1936 (no collection dates available for the other specimens) .

Etymology. Not provided in the original description. However, the species name ( semitaeniatus ) is a feminine singular adjective derived from the Latin semi (half) and taenia (ribbon, stripe), hence half-striped, referring quite accurately to the dorsal pattern of this species, although such a pattern is shared with most other species in the Genus Spondylurus .

Remarks. Spondylurus semitaeniatus is an old species name that has had a confusing history, some of which was clarified recently by Bauer et al. (2003). Until now the name has generally been considered a synonym of either Spondylurus sloanii or Mabuya mabouya . The type was acquired by Marcus Elieser Bloch (1723–1799) and was one of several specimens used by Schneider (1801) to describe Scincus auratus , a species now recognized by the name Trachylepis aurata (Linnaeus) and with a new lectotype (Moravec et al. 2006). The same specimen (ZMB 1238) was then used by Wiegmann (1837) to name Euprepes semitaeniatus . No locality was given for the holotype other than "America," which has led to some of the confusion. However, characters of this specimen (see above) associate it with specimens of Mabuyinae representing the species in the Virgin Islands with a relatively wide distribution.

Bloch had collaborators around the world that sent him material ( Paepke 1999), and therefore it is unlikely, but not impossible, that he was the actual collector. His fish collection, at least, was amassed during 1779–1799 from a limited number of locations, with St. Croix being the only location in the entire Caribbean region mentioned ( Paepke 1999). Thus we might assume from this information that the type of Spondylurus semitaeniatus was collected during 1779–1799, in the general region of St. Croix, and sent to Bloch. Because the only skinks known from St. Croix (and its islet Green Cay) are Spondylurus magnacruzae sp. nov. (records in several museums) and Capitellum parvicruzae sp. nov. (one specimen in ZMUC), we suspect that S. semitaeniatus never occurred there, and thus the type material was probably collected elsewhere. Saint Croix is an island isolated by an expanse of deep water from the remainder of the Virgin Islands, although much of its herpetofauna is shared with the Puerto Rico Bank (including the Virgin islands), indicating that overwater dispersal has been frequent in the recent past.

In the late 18th century, St. Croix, St. Thomas, and St. John were all owned by Denmark, and therefore nearby St. Thomas and St. John would be the next most likely places where the collectors for Bloch would have obtained locality of S. semitaeniatus to St. Thomas. If the specimen actually came from another location in the region (e.g., St. Croix or St. John), it would not affect the taxonomy of this or related species.

There has been some additional confusion surrounding the holotype of Spondylurus semitaeniatus as noted by Bauer et al. (2003). Peters (1864) correctly identified the type as ZMB 1238, but Stejneger (1904) listed it as ZMB 5290, an error repeated by later authors ( Grant 1931; Dunn 1936; Schwartz & Thomas 1975; Schwartz & Henderson 1988). It is unclear whether Stejneger examined a different specimen or just listed the number incorrectly. He noticed that the pale middorsal stripe was very narrow, narrower than in what he was calling Mabuya sloanii (which included other species recognized here), leading him to consider S. semitaeniatus and S. sloanii different species. Relying on Stejneger's description of the holotype, Schmidt (1928) went further and associated S. semitaeniatus with the boldly-striped skinks from Culebra (here called S. culebrae sp. nov.). However, he did not feel strongly enough about the association to consider them distinct species and therefore placed S. semitaeniatus in the synonymy of S. sloanii . Grant (1931) went a step further and treated S. semitaeniatus as a distinct species, occurring only on Culebra and Mona islands (not mainland Puerto Rico), but he later reversed his decision and called them S. sloanii ( Grant 1937) .

Figures

FIGURE 5 (continued on the next page)

FIGURE 10. Map showing the distribution of mabuyine skinks occurring on islands in the greater Puerto Rico region. See Figure 1 for location of the islands. The islands and species are: (A) Puerto Rico (Spondylurus nitidus); (B) Mona (Spondylurus monae sp. nov., squares) and Monito (Spondylurus monitae sp. nov., circle); (C) Desecheo (Spondylurus nitidus); (D) Vieques (Spondylurus nitidus, circle), Culebra (Spondylurus nitidus, circles; Spondylurus culebrae sp. nov., squares), Cayo Luis Peña and Cayo Norte (S. nitidus, circles); (E) most of the U.S. and British Virgin Islands (Spondylurus macleani, diamond; Spondylurus sloanii, squares; Spondylurus semitaeniatus, circles; Spondylurus spilonotus, triangles); (F) Anegada (Spondylurus anegadae sp. nov.); and (G) St. Croix and Green Cay (Capitellum parvicruzae sp. nov., circle; Spondylurus magnacruzae sp. nov.,

FIGURE 55. Representative patterns in species of the Genus Spondylurus (top, dorsal view; bottom, side view). (A) S. anegadae sp. nov.; (B) S. caicosae sp. nov.; (C) S. culebrae sp. nov.; (D) S. fulgidus; (E) S. haitiae sp. nov.; (F) S. lineolatus; (G) S. macleani; (H) S. magnacruzae sp. nov.; (I) S. martinae sp. nov.; and (J) S. monae sp. nov.

FIGURE 61. Graph of rostral height versus rostral length in seven species of the Genus Spondylurus.

FIGURE 62. Graphs of scalation differences between Spondylurus culebrae sp. nov. and other species. (A) graph of head scales length (posterior edge of rostral to posterior edge of parietals) versus snout-vent length in S. culebrae sp. nov. and S. semitaeniatus. The lectotype of S. semitaeniatus is indicated with an arrow. (B) graph of finger-IV lamellae versus head length in S. culebrae sp. nov. and S. sloanii.

FIGURE 73. Representative patterns in species of the Genus Spondylurus (top, dorsal view; bottom, side view). (A) S. monitae sp. nov.; (B) S. nitidus; (C) S. powelli sp. nov.; (D) S. semitaeniatus; (E) S. sloanii; (F) S. spilonotus; and (G) S. turksae sp. nov.

FIGURE 76. Head scalation in species of the Genus Spondylurus (left, side view; right, dorsal view). (A) S. powelli sp. nov. (MCZ R-74343, holotype); (B) S. semitaeniatus (UMMZ 80585); (C) S. sloanii (USNM 576305); and (D) S. spilonotus

FIGURE 78. Spondylurus semitaeniatus, from the Virgin Islands. (A–D) ZMB 1238, holotype, restricted to St. John, U.S.

FIGURE 79. Graph of pale middorsal stripe width versus dark dorsolateral stripe width in 16 species of the Genus Spondylurus.

FIGURE 80. Graphs showing characters distinguishing Spondylurus semitaeniatus and S. sloanii. (A) Graph of dark

FIGURE 81. Spondylurus sloanii, from the Virgin Islands. (A–C) MNHN 554, holotype, St. Thomas, U.S. Virgin Islands. Photographs courtesy of Ivan Ineich. (D–F) USNM 576305, Little Saba Island, St. Thomas (F = live). (G) fetus of USNM 576305.