A new species of burrowing snake (Serpentes: Dipsadidae: Apostolepis) from the state of Mato Grosso, Central-West region of Brazil

During a faunal rescue conducted at a hydroelectric power station constructed in a Cerrado savanna area in the state of Mato Grosso, a sample of five small stripe-patterned individuals of snakes of the genus Apostolepis Cope, 1862 document the existence of an undescribed species, which is named herein. The new species can be distinguished from its congeners by a combination of scale counts, number of maxillary teeth and color pattern. The new species is most similar to Apostolepis borellii Peracca, 1904, A. lineata Cope, 1887, A. nelsonjorgei Lema & Renner, 2004, A. nigroterminata Boulenger, 1896, A. serrana Lema & Renner, 2006 and A. underwoodi Lema & Campbell, 2017 in its coloration pattern. However, it is distinguished from these species by having a pair of triangular blotches covering portions of the third to sixth supralabials, a white nuchal collar, the shape of the fourth supralabial and the shape of the tip of tail, the number of supralabials in contact with parietals, the size of the anterior chinshields, the color pattern of the paraventral side, parietal and terminal scales, the width of dorsal stripes, and a distinct number of subcaudals. The new species occurs in areas within the Cerrado biome.

Apostolepis is particularly speciose in Brazil, where 32 species have already been recorded (Costa and Bérnils 2018).Contributions to the taxonomy of these burrowing snakes in Central-West region of Brazil were initially made by Cope (1887), who described two species -Apostolepis lineata and A. vittatafrom the municipality of Chapada dos Guimarães, state of Mato Grosso.Subsequently, Koslowsky (1898) described A. intermedia from Miranda, state of Mato Grosso do Sul.A few years later, Peracca (1904) described A. borellii based on a specimen from Urucum massif, near Corumbá, Mato Grosso do Sul; Amaral (1925) described A. rondoni based on a specimen from "Matto Grosso" (=the former name of a huge region which presently comprises three Brazilian states: Mato Grosso do Sul, Mato Grosso, and Rondônia), and Prado (1942) described A. goiasensis based on a specimen from Rio Verde, Goiás.More recently, Lema (2002aLema ( ,b, 2003a)), and Ferrarezzi et al. (2005), respectively, described A. albicollaris from Brasília, Distrito Federal, A. christineae from Barra do Bugres, presently Porto Estrela, Mato Grosso, A. cerradoensis from Minaçú, Goiás, and A. ammodites from Palmas, Tocantins, whereas Lema and Renner (2006) and Lema (2016) described two new species from Serra do Roncador, Mato Grosso, A. serrana and A. roncadori, respectively.Additionally, Martins and Lema (2015) and Lema and Renner (2016) removed A. borellii from the synonymy of A. nigroterminata.
During a faunal rescue conducted at a hydroelectric power station constructed in a Cerrado savanna area in the state of Mato Grosso, a sample of five small stripe-patterned individuals of snakes of the genus Apostolepis document the existence of an undescribed species, which is named herein.
All measurements were made to the nearest 0.1 mm using digital calipers, except for snout-vent (SVL) and tail (TL) lengths, which were taken with a flexible ruler to the nearest 1.0 mm.Ventral scales were counted according to Dowling (1951).Bilateral variation is reported as right/left.When no everted hemipenis was available, the sex of each specimen was determined by making a post-cloacal incision between the eighth and tenth subcaudals to verify the presence of the hemipenes.The hemipenial description is based on right organs from two preserved specimens (UFMT-R 1933 and MCP 14524), which were prepared according to the method described in Pesantes (1994) and Zaher and Prudente (2003).UFMT-R 1933 was not expanded or inflated since it was irreversibly damaged during filling of the hemipenial body.On the other hand, MCP 14524 was fully inflated with red petroleum jelly.Terminology for hemipenial morphology followed Dowling and Savage (1960) and Zaher (1999).The distribution map was made using the free Quantum GIS software.We use the names A. quinquelineata and A. nigrolineata in the sense of Lema (1997) and Lema and Renner (1998), respectively (but see Curcio et al. 2011).
Preserved specimens of Apostolepis kikoi sp.nov.are most similar to A. borellii, A. lineata, A. nelsonjorgei, A. nigroterminata, A. serrana and A. underwoodi in its general pholidosis and coloration pattern.However, the new species is distinguished from A. borellii by having its paraventral sides cream and unblemished (vs.blackish), a pair of triangular blotches covering portions of the third, fourth, fifth and sixth supralabials (vs. a small, trapezoidal blotch covering only the posterior half of the third and the entire fourth supralabial); tip of the tail conical (vs.rounded) and fewer subcaudals (vs.26-30 vs. 32).Apostolepis kikoi sp.nov.can be distinguished from A. nigroterminata by having a triangular blotch covering portions of the third, fourth, fifth and sixth supralabials (vs.an irregular blotch on the posterior margin of the third to the anterior margin of the fourth supralabial -see also Harvey 1999: 401, fig. 7 in Lema and Renner 2016, and Figs 4-5 below), a blackish blotch on rostral scale adjacent to anterior border of prefrontals (vs.blotch absent), darker parietals (vs.light blotches irregularly distributed on parietal scales) and a distinct number of maxillary teeth (4 + 2 vs. 3 + 2).Further, it differs from A. nigroterminata in having its background color beige (in living specimens) (vs.background color red-orange); paravertebrals distinct (vs.paravertebrals indistinct); first and fifth stripes wider, covering upper half of third and lower half of fourth row on each side (vs.first and fifth stripes also wider, but covering first and about 50% of fourth row of scales on each side).Apostolepis kikoi sp.nov.can be distinguished from A. underwoodi by having a vertebral stripe one scale wide (vs.vertebral stripe narrow, running on the medial line of each vertebral scale); paravertebral stripes covering sixth row on each side (vs.paravertebral stripes covering half of fifth and half of sixth rows).Finally, Apostolepis kikoi sp.nov. is distinguished from A. lineata, A. nelsonjorgei and A. serrana by having its fourth supralabial scale rectangular (vs.triangular in A. nelsonjorgei), 4-6 supralabials contacting parietals (vs.5-6 contacting parietals in A. nelsonjorgei), terminal scale black dorsolaterally (vs.terminal scale entirely white in A. lineata and A. nelsonjorgei), anterior chinshields longer than posterior (vs.anterior and posterior chinshields of about the same size in A. nelsonjorgei), the presence of a white nuchal collar (vs.white nuchal collar absent in A. lineata and A. serrana) and a distinct number of subcaudals (27-30 vs. 40-46 in A. nelsonjorgei and 33 in A. serrana).
Description of holotype (Figs 1-3).A small female, possibly juvenile, SVL 262 mm, TL 26 mm (9.92% of SVL).Body subcylindrical.Tail very short with tip conical and laterally compressed.Terminal scale pointed.Head slightly distinct from neck, narrower than diameter of midbody.Head length from quadrate-articular jaw joint to tip of snout (in lateral view) 6.56 mm (2.5% of SVL), 4 mm at widest point (60.9% of head length).Snout rounded in dorsal and lateral views, slender and slightly projected beyond jaws; snout length from tip of snout to anterior margin of right orbit 2.37 mm (36.12% of head Five occipitals wider than long; median occipital positioned between posterior tips of parietals, smaller than adjacent vertebral; two pairs of lateral occipitals twice as large as dorsals, contacting posterior edges of sixth supralabials.Supralabials 6/6, 1 contacting rostral, 1-2 contacting nasal, 2 contacting nasal and preocular, 2-3 entering orbit, 3-4 contacting single postocular, 4-6 contacting parietal.Supralabials in ascending order of size, with sixth supralabial higher and longer (1.46 mm long, 1.29 mm high) than remaining supralabials.Infralabials 7/7, 1-4 contacting anterior chinshields, 4-7 contacting posterior chinshields.Mental subtriangular, wider (1.14 mm) than long (0.82 mm), separated from anterior chinshields by contact between first infralabials.Anterior chinshields elongated, longer than posterior chinshields.Suture between chinshields 1.82 mm.Chinshields separated from ventrals by four gulars and two preventrals.Gulars in four rows between last supralabial and Coloration of holotype in life (Fig. 2).Head dorsally blackish with light blotches irregularly distributed on rostral, nasal, prefrontal, frontal, supraocular and parietal scales.Suture between prefrontals brownish.Dorsally, rostral scale has a blackish blotch adjacent to anterior border of prefrontals; ventrally, rostral scale has a blackish spot in its convex portion.Posterior upper margin of first and fourth supralabials blackish.Second supralabial blackish.Anterior portion and upper half of third supralabial blackish.Blackish head cap covers upper half of fifth and sixth (except for its lower anterior portion) supralabials.Posterior portion of nasal, preocular and postocular scales blackish.Infralabials and chinshields cream, except for small blackish spots on third and fourth scales.Brown pigment of throat region restricted to more laterally positioned gular scales (i.e., evidencing an incomplete gular band).White nuchal collar two and half to three dorsals long.Black nuchal collar absent.Background color beige, with five black stripes; vertebral stripe one scale wide; paravertebral stripes covering sixth and tenth rows, less distinct than remaining dorsal stripes; first and fifth stripes wider, covering upper half of third and lower half of fourth row on each side.Paraventral sides and venter unblemished.Black band on tail extends for nine scales dorsally; seven subcaudals are black.Terminal scale with band of melanophores dorsolaterally and entirely clear ventrally.
Color of holotype in preservative (Fig. 3).Head dorsally brownish with light blotches irregularly distributed on rostral, nasal, prefrontal, frontal, supraocular and parietal scales.Dorsally, rostral scale brown pigmented; this blotch covers anterior border of prefrontals; ventrally, rostral scale with brownish spot on its convex portion.Anterior and posterior parts of prefrontals brown.Anterior and posterior parts of suture between prefrontals brown.Posterior upper margin of first and fourth supralabials brown.Second supralabial brown.Anterior portion and upper half of third supralabial brown.Brown head cap covers upper half of fifth and sixth (except for its lower anterior portion) supralabials.Posterior portion of nasal, preocular and postocular scales brown.Infralabials and chinshields cream, except for small brown spots on third and fourth scales.Brown pigment of throat region restricted to more laterally positioned gular scales (i.e., evidencing an incomplete gular band).White nuchal collar two and half to three dorsals long.Black nuchal collar absent.Background color light brown, with five brown stripes; paravertebral stripes light brown and less distinct than remaining dorsal stripes.Venter immaculate.Dark brown band on tail extends for nine scales dorsally; seven subcaudals are dark brown.Terminal scale with band of melanophores dorsolaterally and entirely clear ventrally.
Variation.Measurements and morphological variation are summarized in Table 1 7).Retracted organs extend for length of nine subcaudals.Everted hemipenes subcylindrical, unilobed, unicapitate and noncalyculate.Basal region on sulcate side bears numerous spines of similar size.Several moderate-sized spines present on lateral region of hemipenial body in its absulcate side.Sulcus spermaticus bifurcates about two-thirds before end of organ; branches -which extend centrolineallyreach distal tip of lobe.Basal region on absulcate side also bears numerous spines of similar size, but these are abruptly replaced by two larger spines in middle region of hemipenial body, larger than those disposed on lateral region of hemipenial body.Capitulum confined to sulcate side.Distal region of absulcate side bears transverse papillate flounces, without calyces, whereas a small number of papillae are concentrated above flounces.Distribution (Fig. 8).All individuals of the type series of Apostolepis kikoi sp.nov.were obtained in the area presently occupied by the Manso multi-use reservoir and hydroelectrical power plant (APM Manso), in most part situated in the municipality of Chapada dos Guimarães, Mato Grosso, Brazil.With nearly 428 km 2 , the reservoir of APM Manso inundated -from December 1999 -many different Cerrado physiognomies (see Conceição 2000), established over terrains up to 287 m.a.s.l.from the confluence of the Casca and Manso rivers (approximately at 14°52' S, 55°48' W) upwards.Manso is the main tributary to the Cuiabá River, a major tributary of the left bank of the Paraguay River.The local climate is generally hot and semi-humid (classified as "Aw climate" in the Köppen's climate classification map for Brazil, see Alvares et al. 2014), with a well-marked seasonality and rains concentrated in the summer, from October/November to April/May.Mean annual precipitation is 1350 mm; mean annual temperature is around 26 °C.There are four to five months of drought (May to September), and relative air humidity may drop to less than 30% from July to September (Strüssmann and Mott 2009).The Manso River basin is included in the morphostructural domain of the Paraná River sedimentary basin.It is cut across two distinct lithostratigraphic units in nonconformity: the Permeable Mesozoic sandstones belonging to the São Bento (or Botucatu) Group and the inclined Precambrian phyllites and gneisses of the Cuiabá Group, frequently with a lateritic horizon at or near the surface (Barros et al. 1982;Vieira Jr. et al. 2012).
Etymology.The specific epithet honors Francisco Luís Franco ("Kiko"), a specialist in Brazilian snakes, as a tribute to his relentless friendship, dedication and enthusiasm as curator of Herpetological Collection Alphonse Richard Hoge of Instituto Butantan, São Paulo, Brazil (partially and tragically destroyed by fire on 15 May 2010).
Remarks.MCP 12096 was selected as the holotype because the general color in life of A. kikoi was described from the live holotype before it was euthanized.
Apostolepis nigroterminata was described by Boulenger (1896) after a single specimen from "Cayaria" (= Callaria, Departamento de Ucayali), eastern Peru.Harvey's (1999) placement of A. borellii in the synonymy of A. nigroterminata resulted in the inclusion of the latter species in a former Brazilian list of reptiles (see Costa and Bérnils 2015).However, it should be noted that Harvey's (1999) nomenclatural act was based on the analysis of the holotype of A. borellii (a specimen from Urucum massif, Mato Grosso do Sul, Brazil), one specimen collected at the confluence of Rio Araguaia and Tapirapé, Tapirapé Village, Mato Grosso do Sul (AMNH 87942) -that was subsequently re-identified as A. phillipsi by Martins and Lema (2015) -and several other specimens from Bolivia and Peru.Recently, Lema and Renner (2016) removed A. borellii from the synonymy of A. nigroterminata and restricted the distribution of the latter to some localities in Peru.However, these authors also listed a specimen of A. nigroterminata from "Brazil: Acre: Rio Branco" in Appendix as "UFAC w/n".The species was also included in an updated list of Brazilian reptiles and referred to occur in the states of Acre, Mato Grosso, and Pará (Costa and Bérnils 2018).Besides the unvouchered mention to Acre in Lema and Renner (2016), Lema et al. (2017) presented a picture of a specimen from Acre without a clear locality description or voucher number, which also provides little evidence for the occurrence of this species in Brazil.The record for Pará is also unvouchered (Maschio et al. 2012), while the specimen (UFMT 10672) from Nobres, Mato Grosso, referred to the species by Santos et al. (2011) was examined and is reidentified here as Apostolepis sp.Therefore, we argue that the specimens examined herein (UFAC 383, UFAC 397, UFAC 504) represent not only three locality records of A. nigroterminata for the state of Acre, Brazil, but also the first documented record of the species for the country.The specimens collected in Rio Branco (UFAC 397, UFAC 504) extend the geographic distribution of A. nigroterminata about 760 km northeastward from Callaria.In particular, the specimen depicted in Figs 4-5 matches the original description of Boulenger (1896) and that given by Lema and Renner (2016) in most details of scalation and color pattern of this species (see Table 2).Two other species of Apostolepis -A.lineata and A. vittatawere also described from Chapada dos Guimarães (Cope 1887).Although the only existing syntype of A. lineata is in very bad condition, Harvey's (1999) redescription is sufficiently complete to allow it to be unambiguously distinguished from Apostolepis kikoi sp.nov.Together with A. assimilis (Fig. 9) and A. vittata (Fig. 10), the description of Apostolepis kikoi sp.nov.increases the number of species of Apostolepis reported to occur sympatrically in the Manso reservoir area to three.An attempt was made to identify all the specimens of Apostolepis kikoi sp.nov.using Nogueira et al.'s (2012) key.However, the specimens could not be characterized beyond couplet 1, because of the many overlapping characters presented in the couplets.The assignment of Apostolepis kikoi sp.nov.into a formal group should await a more comprehensive phylogenetic arrangement than is available for the genus.occurrence of Apostolepis nigroterminata in Brazil.Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) provided a fellowship to FMS during his graduate studies at the Universidade Federal de Mato Grosso do Sul and to OME-N during his undergraduate studies at the Universidade Federal do Rio Grande, and a productivity research grant to CS (CNPq process 309541/2012-3).