Marisora, Hedges & Conn, 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 : 119-122

Genus Marisora gen. nov.

Middle American Skinks

Type species. Mabuya unimarginata Cope, 1862:187 .

Diagnosis. Species of the Genus Marisora are characterized by (1) frontoparietals, two (rarely three), (2) supraciliaries, four (occasionally three, five, or six), (3) supraoculars, four (rarely three), (4) prefrontal contact, absent (or contact very rarely), (5) parietal contact, present (or occasionally no contact), (6) rows of nuchals, one (rarely two rows), (7) dorsals + ventrals, 109–131, (8) total lamellae, 184–229, (9) a dark middorsal stripe, absent, (10) dark dorsolateral stripes, usually absent (present in M. alliacea comb. nov.), (11) a dark lateral stripe, present, and (12) dark ventral striping, absent. Species of Marisora are medium to large, with a range of maximum body sizes among the species of 82–95 mm SVL (except for one species, M. magnacornae sp. nov., known from a single 77.4 mm specimen; Table 2).

All Marisora have a basic pattern, usually bold and well-defined, of a dark lateral band several scales wide bordered below by a narrow pale stripe usually less than one scale wide. In two species ( M. aurulae sp. nov. and M. falconensis ), this basic pattern is weakly defined, and in another ( M. alliacea ), there are additional (dorsolateral) dark stripes. The absence of dark dorsolateral stripes (except in M. alliacea ) distinguishes this genus from Aspronema , Brasiliscincus (most individuals), Manciola , Orosaura , Panopa , Psychosaura , Spondylurus , and Varzea (most individuals). The presence of one row of nuchals (rarely two) distinguishes the Genus Marisora from Exila and Panopa (2–5 rows) and most Spondylurus (usually 2 rows). The presence of two (rarely three) frontoparietals (instead of one fused scale) distinguishes this genus from Exila , Notomabuya , and Panopa . The presence of a pale lateral stripe and absence of dark ventral striping distinguish this genus from the Genus Alinea . The absence of a middorsal dark stripe further distinguishes this genus from Aspronema . The presence of four (usually) supraciliaries (versus 5–6) distinguishes Marisora from Capitellum and Exila . Contact (usually) of the parietal scales distinguishes this genus from the Genus Copeoglossum (parietals usually not in contact). In having four supraoculars (rarely three), Marisora is separated from two genera with three supraoculars: Aspronema (rarely four) and Mabuya (rarely two or four). In having 184–229 total lamellae, it is distinguished from Manciola (147– 154 lamellae) and Alinea (231–259 total lamellae). From Maracaiba , it differs (weakly) by having a low number of dorsals (50–63 versus 63; only three of 80 Marisora with 63 dorsals).

Content. Seven species are placed in this genus: Marisora alliacea , Marisora aurulae sp. nov., Marisora

Distribution. The Genus Marisora is distributed throughout Middle America from central Mexico (Colima in the west and Veracruz in the east) to southern Panama, in northern South America ( Colombia, and Venezuela; primarily in the Caribbean lowlands), and on Caribbean islands relatively close to mainland areas (Cozumel, Mexico; Bay Islands, Honduras; Great Corn Island, Nicaragua; Trinidad and Tobago; Grenada, the Grenadines, and St. Vincent ( Figs. 1 , 8A , 9E , and 11D, I–J ).

Etymology. The generic name ( Marisora ) is a feminine noun derived from the Latin words maris (sea) and ora (coast, or border), referring to the distribution of this genus occurring predominately in low elevations near the coast (Caribbean, Atlantic, and Pacific), with relatively few inland and upland localities. Three of the seven species occur exclusively on islands.

Remarks. No attempt was made here to conduct a comprehensive revision of mainland mabuyine skinks from Middle America or South America. Nonetheless, it was necessary for us to examine a sufficient number of specimens from those regions to compare with specimens treated here from Caribbean islands. In doing so we were able to evaluate two previously named Middle American species, Mabuya alliacea Cope and Mabuya brachypoda Taylor , that have been recognized by some (Burger 1952; Taylor 1956; Webb 1958; Flores-Villela 1993; Campbell 1998) but not by others ( Dunn 1936; Peters & Donoso-Barros 1970; Savage 2002; Miralles et al. 2009b).

Burger (1952) objected to Dunn's (1936) arrangement of placing nearly all American taxa in Mabuya mabouya . He noted variation in some specimens collected in Middle America and resurrected Mabuya mabouya alliacea as a subspecies. However, he examined only four characters, which he admitted were insufficient. The study by Taylor (1956) was more comprehensive. He examined a larger number of specimens, all from Costa Rica, tabulated measurements and scale counts, and concluded that three species were present in that country, naming one of them Mabuya brachypoda . Taylor (1956) used standard measurements and scale counts and some nonstandard characters (e.g., paired chin shields contacting labials) to diagnose the three taxa.

Taylor (1956) showed that Cope's Marisora alliacea could be diagnosed and that all of the specimens of that species occurred in eastern Costa Rica, on the Atlantic slope of the cordillera. All had long limbs, a dorsal pattern of dark dorsolateral stripes (absent in other Middle American mabuyine skinks), a low number of midbody scale rows (26–29), supranasal separation (or only point contact), and greenish color in life. Except for the dorsal stripe pattern, each of the characters individually can be found in the other two species, albeit rarely or uncommonly, but in combination they are diagnostic. He examined 17 specimens of the species from Costa Rica. We examined five other specimens from eastern Costa Rica (UF 30454, 30459, 30460, 30467, and 30471) and a specimen from adjacent southern Nicaragua (USNM 19542), all conforming to the characterization of this species by Cope (1876) and Taylor (1956). In addition, we note that all of the specimens we examined have dark venters in preservative. We see no evidence that the range of this species occurs beyond eastern Costa Rica and southern Nicaragua, although a more comprehensive examination of material is needed to better define its distribution. It appears to be the only species of Marisora occurring on the Atlantic slope of Costa Rica.

Taylor (1956) divided the remaining mabuyine skinks of Middle America into a long-limbed species, here Marisora unimarginata , with dark dorsal spots and a short-limbed species usually lacking spots, which he described as Mabuya brachypodus (later, corrected to M. brachypoda as the species name is adjectival, "shortfooted"). He considered Marisora unimarginata to occur in Panama and western Costa Rica (Pacific slope) and Marisora brachypoda to also occur in western Costa Rica (Pacific slope) north to Mexico, although the only material he examined of both species was from Costa Rica. He found that in most M. unimarginata the sixth supralabial was below the eye (fifth in M. brachypoda ) and only one pair of chin scales contacted the infralabials (two pairs contacted the infralabials in M. brachypoda ). Both species have primarily 30 or 32 midbody scale rows, but the latter is a rare count in M. brachypoda and the specimens with 32 scale rows may not be M. brachypoda (see below).

Because the holotype of Marisora unimarginata (now unlocated and presumably lost) is from Panama, and Cope (1876) made no mention of dark dorsal flecks or spots, Taylor (1956) was concerned that his well-spotted Costa Rican M. unimarginata might be yet another new species. However, some specimens of Marisora that we have examined from Panama (e.g., CM 43594), otherwise agreeing with M. unimarginata , have dorsal spotting, so apparently it is a variable character. Taylor noted that localities for the two species ( M. brachypoda and M. unimarginata ) in western Costa Rica were as close as 10 km and at the same elevation, suggesting that they do not

Limb length has been a standard character in mabuyine systematics, often used in a non-quantitative manner, by scoring whether the adpressed limbs (arms back, legs forward) overlap or not. This comparison can be accomplished with soft specimens, but the limbs of many preserved specimens are too stiff to bend without damaging them. Nonetheless, it is possible in even those specimens to measure arm and leg lengths. Taylor (1956) did this and presented his results in tables. To those data, we have added limb length measurements of specimens we examined, combined arm and leg lengths, and plotted all against SVL (see below). First, it should be acknowledged that there is significant measurement error, given that few limbs can be perfectly straightened and measured (and, considering error in alternate measurements of bent limbs, using string). Nonetheless, most specimens separate into a short-limbed species ( Marisora brachypoda ) and long-limbed species ( M. alliacea and M. unimarginata ), consistent with other characters.

Savage (2002) disagreed with Taylor (1956) and instead recognized only one Middle American species, Marisora unimarginata , noting that populations of the various species recognized by Taylor (1956) in Costa Rica showed intergradation. However, Savage did not present evidence for this claim, and it disagrees with the evidence provided by Taylor (1956) and by material that we have examined, which includes geographically intermediate populations. The concordant nature of the character variation, and agreement with geography, suggest to us that Cope (1876) was correct in describing M. alliacea , and Taylor (1956) was correct in describing M. brachypoda , and in recognizing all three species.

There is evidence that additional species are present on the mainland of Middle America. Taylor (1956) noted that two individuals from Barracana, Costa Rica, had an unusual pattern of dark lines through each scale. We have also noticed that specimens of Marisora brachypoda from Honduras are similarly lineate and additionally possess distinctive pale ventrolateral stripes. Moreover, three specimens (TCWC 80536, UF 143817, and RT 1729) that we examined from Guanacaste, Cost Rica, differ considerably from other M. brachypoda , including specimens from Puntarenas, Costa Rica. They have shorter toes than any M. brachypoda (7.43–8.68% SVL versus 8.90–12.7%) and shorter heads than all but one M. brachypoda (14.9–15.7% SVL versus 15.6–21.4%). Their limb length is at the low extreme (arm + leg length 45.3–48.3% SVL versus 46.4–61.7%), and dorsals + ventrals are at the high extreme (124–129 versus 109–124). All three specimens have separated supranasals, whereas those scales are in contact in 25 of 29 other M. brachypoda examined, and in the holotype ( Taylor 1956). In pattern, they have a more discontinuous dark lateral stripe than others. Otherwise, they have the key characters of M. brachypoda , including short limbs, two pairs of chin shields in contact with infralabials, and fifth supralabial below the eye. Because the holotype of M. brachypoda is from Guanacaste Province, this potential new species may be sympatric with M. brachypoda . Alternatively, it may represent the true M. brachypoda , in which case populations currently assigned to M. brachypoda from elsewhere will require a new name.

Relatively deep divergences (4–5%) among populations of Marisora brachypoda in the molecular phylogeny ( Fig. 6 ), and paraphyletic branching, also suggests that multiple species are present. A comprehensive review of all Middle American specimens of Marisora is warranted to determine the number of species present and their distributions. However, we believe that it is more useful for systematists and non-systematists to recognize five diagnosable species ( M. alliacea , M. brachypoda , M. magnacornae sp. nov., M. roatanae sp. nov., and M. unimarginata ) in Middle America now, even though one ( M. brachypoda ) is paraphyletic and in need of further study, than to maintain the current taxonomy whereby a single species ( M. unimarginata sensu lato), is recognized and known to be a complex of species.

Mijares-Urrutia and Arends (1997) described Marisora falconensis from the state of Falcón, Venezuela. Miralles et al. (2005a) located additional material of M. falconensis in museum collections that extended the distribution of the species, especially along the northern coast east of Falcón to the state of Sucre, Venezuela. They also assigned a specimen (UMMZ 54793) from Guajira, Colombia to that species. We have examined that Colombian specimen and agree that it is similar to M. falconensis in scalation and is a member of the Genus Marisora . However it has a wider (nearly two scale rows) pale lateral stripe and a narrower dark lateral stripe as compared with M. falconensis . Further comparisons with additional specimens are needed to determine whether M. falconensis or a related species occurs in Colombia.

Marisora falconensis has not yet been compared with M. unimarginata (sensu stricto) by any authors, morphologically or with molecular data. Available DNA sequences of Marisora from Middle America are from M. from all three new species of the Genus Marisora , described below. However, M. falconensis and M. unimarginata share some characters including long limbs, broadly overlapping scale counts, and a pattern that includes (variably) dorsal spotting, although M. falconensis tends to be a darker species, at least in preservative. However, in having a weakly-defined dark lateral stripe, M. falconensis resembles M. aurulae sp. nov. One character that we found to distinguish most specimens of M. falconensis and M. unimarginata is the chin shield character. In M. unimarginata , there is usually (88%) one pair of chin shields posterior to the postmental that touch the infralabials (i.e., are not separated by a sublabial), whereas in M. falconensis , there are usually two (79%) or three (5%) such pairs of chin shields. However, a large number of Venezuelan and other South American specimens of Mabuya exist in museum collections that have not been examined by us or by previous authors. A comprehensive examination of this material is needed to better understand the systematics of mabuyine skinks from Venezuela and elsewhere in South America.

Below we describe three new species of the Genus Marisora from Caribbean islands. One species occurs in the Windward Islands (southern Lesser Antilles, Trinidad and Tobago) and is most closely related to M. falconensis . The other two are from islands off of Middle America (Great Corn and Roatán) and appear to be most closely related to species occurring in that geographic region ( M. alliacea , M. brachypoda , and M. unimarginata ). Morphological data, combined with molecular data for several of the species, show that all seven species form a clade (Genus Marisora ) that is most closely related to two genera also occurring in that general geographic region: Mabuya of the Lesser Antilles and Maracaiba of Venezuela. This phylogenetic relationship has been observed previously using many of the same Genbank sequences ( Miralles et al. 2005a; Miralles & Carranza 2010).

Figures

FIGURE 1. Map of Caribbean region. Neotropical skinks occur, or once occurred, on the islands indicated.

FIGURE 6. Phylogenetic tree of Neotropical skinks (Mabuyidae, Mabuyinae) showing linearized branch lengths (GTR +

FIGURE 8. Distribution of the 16 genera of Neotropical skinks, Family Mabuyidae, Subfamily Mabuyinae. (A) Genera occurring in the greater Caribbean region. (B) Genera occurring exclusively in South America. (C) Additional genera occurring in South America, including one with representatives on Caribbean islands.

FIGURE 9. Map showing the distribution of mabuyine skinks occurring on islands in the central, southern, and western Caribbean region. See Fig. 1 for location of the islands. The islands and species are: (A) Jamaica (Spondylurus fulgidus; hollow circle = specimen not examined); (B) Hispaniola (Spondylurus haitiae sp. nov., circle; Mabuya hispaniolae sp. nov., triangle; Spondylurus lineolatus, squares); (C) Turks and Caicos Islands (Spondylurus caicosae sp. nov., circles; Spondylurus turksae sp. nov., triangles; hollow triangle is Cotton Cay record (see text); (D) Margarita (Copeoglossum margaritae sp. nov.); and (E) Roatán (Marisora roatanae sp. nov.).

FIGURE 11. Map showing the distribution of mabuyine skinks occurring in the Lesser Antilles and Trinidad and Tobago. See Fig. 1 for location of the islands. The islands and species are: (A) Anguilla, Dog Island, and St. Barts (Spondylurus powelli sp. nov., squares), and St. Martin (S. martinae sp. nov., circle); (B) Guadeloupe: Basse-Terre (Mabuya guadeloupae sp. nov., square), Grande-Terre (Mabuya grandisterrae sp. nov., triangle), Îlet à Cochons (Mabuya cochonae sp. nov., star), La Désirade and Terre de Bas, Petite-Terre (Mabuya desiradae sp. nov., diamonds), and Marie-Galante (Capitellum mariagalantae sp. nov., circle); (C) Martinique (Mabuya mabouya, circles; Capitellum metallicum, triangle); (D) St. Vincent, the Grenadines, and Grenada (Copeoglossum aurae sp. nov., circles; Marisora aurulae sp. nov., triangles); (E) Redonda (Copeoglossum redondae sp. nov., circle) and Montserrat (Mabuya montserratae sp. nov., square); (F) Dominica (Mabuya dominicana); (G) St. Lucia (Alinea luciae), (H) Barbados (Alinea lanceolata); (I) Trinidad (Copeoglossum aurae sp. nov., circles; Marisora aurulae sp. nov., triangle), and (J) Tobago (Copeoglossum aurae sp. nov., circles; Marisora aurulae sp. nov., triangles). Hollow symbols in A and J are literature records (specimens not examined). Where no specific localities within an island are known for a species, a symbol is placed centrally on the island: Alinea luciae, Capitellum mariagalantae, Capitellum metallicum, and Copeoglossum redondae sp. nov. The scale in (A) applies to all islands.