Phylogenetic relationships of a new species of Apostolepis from Brazilian Cerrado with notes on the assimilis group (Serpentes: Colubridae: Xenodontinae: Elapomorphini)

Phylogenetic relationships of a new species of Apostolepis from Brazilian Cerrado with notes on the assimilis group (Serpentes: Colubridae: Xenodontinae: Elapomorphini)

Hebert Ferrarezzi^I; Fausto Erritto Barbo^I; Cristina España Albuquerque^{I, II}

^ILaboratório de Herpetologia, Instituto Butantan. Av. Dr. Vital Brazil, 1500, CEP 05503-900, São Paulo, SP, Brasil. E-mail: [e-mail], [e-mail]

^IIDepartamento de Cirurgia, Faculdade de Medicina Veterinária e Zootecnia Universidade de São Paulo. Av. Prof. Dr. Orlando Marques de Paiva, 87, CEP 05508-900, São Paulo, SP, Brasil. E-mail: [e-mail]

ABSTRACT

The assimilis species group is one of the several assemblages that can be distinguished within the genus Apostolepis. The group is characterized by a uniform red dorsal pattern, broad nucho-cervical collars, enlarged light snout blotch, dark ventral head, and entirely black terminal shield. A new species is here recognized as a member of the assimilis group and its phylogenetic relationships to A. assimilis and A. cearensis is inferred by means of cladistic analysis. Apostolepis ammodites sp. nov. can be easily diagnosed from all congeners by many autapomorphies and an exclusive combination of features. Comparative data are presented on scalation, coloration, dentition, osteology, and geographic distribution, with some notes on feeding and reproductive biology. The new species occurs through central Brazilian Cerrado, filling a gap between the areas occupied by A. assimilis and A. cearensis. Although superficially resembling A. assimilis, parsimony analysis based on morphological data indicates its phylogenetic relationships closest to A. cearensis. Systematic analysis shows that two recently described species, A. freitasi and A. tertulianobeui, must be relegated as junior synonymies of A. cearensis and A. assimilis, respectively.

Keywords: Serpentes, Colubridae, Elapomorphini, Apostolepis, assimilis group, A. ammodites sp. nov., phylogenetic systematics, biogeography.

RESUMO

As espécies do grupo assimilis se caracterizam por um padrão dorsal vermelho uniforme, colares nuco-cervicais largos, mancha clara do focinho larga, região ventral da cabeça densamente pigmentada e escudo terminal completamente preto, é um dos vários grupos que podem ser distintos dentro do gênero Apostolepis. Uma nova espécie é aqui reconhecida como membro do grupo assimilis e suas relações filogenéticas para A. assimilis e A. cearensis são inferidas por meio de uma análise cladística. A. ammodites sp. nov. pode ser diagnosticada facilmente de todos os congeneres por muitas autapomorfias e uma combinação exclusiva de caracteres. Dados comparativos são apresentados sobre escutelação, coloração, dentição, osteologia e distribuição geográfica, junto com notas sobre alimentação e biologia reprodutiva. A espécie nova ocorre através do Cerrado brasileiro central, preenchendo o intervalo entre as áreas ocupadas por A. assimilis e A. cearensis. Embora se assemelhe superficialmente a A. assimilis, análises de parcimônia baseadas em dados morfológicos indicam maior parentesco filogenético com A. cearensis. Análises sistemáticas mostram que duas espécies recentemente descritas, A. freitasi e A. tertulianobeui, devem ser consideradas como sinônimos juniores de A. cearensis e A. assimilis, respectivamente.

Palavras-chave: Serpentes, Colubridae, Elapomorphini, Apostolepis, grupo assimilis, A. ammodites sp. nov., sistemática filogenética, biogeografia.

INTRODUCTION

The South American fossorial snakes of the genus Apostolepis Cope 1862 consist of a highly speciose clade of about 30 species, and a number of yet undescribed ones. The taxonomy of such an already difficult group (Vanzolini, 1986; Ferrarezzi, 1993; Harvey, 1999) has become increasingly complex, not only due to the lack of a comprehensive systematic review but, also because of a recent proliferation of poorly described new taxa (e.g., Lema, 2002, 2003, 2004a,b). This is the case of two new species, lately described by Lema (2004a,b) and based on a single specimen each: Apostolepis freitasi from central Bahia, and Apostolepis tertulianobeui from Minas Gerais, both being undistinguishable from Apostolepis cearensis and Apostolepis assimilis, respectively, but clearly referred to the assimilis group as defined by Ferrarezzi (1993).

Elapomorphus assimilis was described by Reinhardt (1861), from Minas Gerais, Brazil. Latter, Gomes (1915) described Apostolepis cearensis based on four specimens from State of Ceará, Northeastern Brazil, comparing it to A. assimilis and A. coronata (= Elapomorphus lepidus). Apparently unaware of Gomes's species, Werner (1924, 1925) described in two subsequent papers A. sanctae-ritae, based on a single specimen from "Santa Rita", Brazil (Brasilien Expedition V.F. Steindachner, 1903) and A. amarali based on a specimen from unknown South American locality. The last name has undisputedly been considered a junior synonymy of A. cearensis (Amaral, 1930, 1936; Peters & Orejas-Miranda, 1970). Recently, Ferrarezzi (1993), recognized the assimilis species group of Apostolepis to include A. assimilis and A. cearensis, and considered A. sanctaeritae as conspecific to A. cearensis, rather than a synonym of flavotorquata (Amaral, 1930, 1936; Peters & Orejas-Miranda, 1970) or a valid species (Lema & Fernandes, 1997; Lema, 2002).

The Instituto Butantan received, in the last few years, several specimens of a clearly distinctive but unnamed elapomorphine species of the genus Apostolepis, that seems to be quite common within the Cerrado domain of central Brazil. It appears to be intermediate between Apostolepis assimilis and A. cearensis, in superficial morphology and through it geographical distribution. The present paper refers to the formal definition and diagnoses of the assimilis species group along with an attempt to elucidate the taxonomic status and systematic relationships of its component species.

MATERIAL AND METHODS

Specimens: A total of 25 museum specimens known to represent the new species were examined. In addition to published figures and descriptions, representative individuals or samples of most species of Apostolepis were analyzed (Ferrarezzi, 1993), in order to fundament a brief character analyses of the group. Relevant material observed for external features (scalation, color pattern and proportions), cranial osteology, and hemipenial anatomy is listed in ^{Appendix 1}. Specimens from the following institutions with there respective acronyms were analyzed (used in the text, figures, and appendix): IB, Laboratório de Herpetologia, Instituto Butantan, São Paulo; MNRJ, Museu Nacional, Universidade Federal do Rio de Janeiro; CHUNB, Coleção Herpetológica da Universidade de Brasília; MZUSP, Museu de Zoologia da Universidade de São Paulo; FUNED, Fundação Ezequiel Dias, Belo Horizonte, MG.

Along with the description of the new species we report the state observed for nearly all external characters currently thought to be diagnostic for the new species relative to its congeners (Ferrarezzi, 1993; Harvey, 1999; Harvey et al., 2001). Morphological information about members of the dorbigny species group were obtained from Harvey, 1999 and Embert (pers. comm.) on Ventral shields were counted using Dowling's (1951) method.

In addition to the observation of external morphology, we prepared osteological and hemipenial material of the new species and its closest relatives Apostolepis assimilis and A. cearensis, as well as of A. pymi and A. flavotorquata, used as outgroups. Skulls of preserved specimens were prepared either dried (with the aid of dermestid beetles and by chemical maceration). Hemipenes were mounted with agar jelly, following the procedures described by Pesantes (1994) for inverted organs from preserved specimens, or by Manzani & Abe (1988) for freshly killed ones. The nomenclature used for hemipenial structures is based on Dowling & Savage (1960) and Zaher (1999).

Species habitat correlations were made by plotting distribution localities in the South American vegetation map of Hueck & Siebert (1972).

Comparative analyses: Monophyly of an ingroup comprising the assimilis species group (Apostolepis assimilis, A. cearensis and A. ammodites sp. nov. - sensu Ferrarezzi, 1993) has been assumed as a working hypothesis. Representatives of the dorbignyi group (A. dorbignyi, A. multicincta) were included as the closest outgroup, whereas those belonging to the flavotorquata group (A. flavotorquata, A. pymi) were taken as a more distant outgroup to root the tree resultant from cladistic analysis. The empirical justification for this assumption about ingroup-outgroup demarcation can be found in Ferrarezzi (1993), but it is also implicit in our character matrix. Character states observed were numerically coded in discrete binary and multistate transformation series, in order to construct the data matrix ready to be subjected to a parsimony analysis using Hennig86 applicative (Farris, 1982). Statements regarding evolutionary direction (polarity) of morphological transformation series (character states) were only assessed a posteriori to parsimony analyses, following the principles recommended by Farris (1982) and Nixon & Carpenter (1993).

The criterion used to evaluate phylogenetic relationships between the new species and other ingroup taxa was apomorphy based (Hennig, 1966). Systematic relationships and biogeographic history of the new species are briefly discussed in a cladistic and vicariance context.

The assimilis species group

Definition and diagnosis: Apostolepis of medium size; nasal separated from preocular; fifth supralabial scale only in contact with the parietal; temporal 0+1; terminal shield pointed in a sharply edge; large internasal white blotch covering the snout and reaching the frontal shield; moderate to extremely broad and sharply evident white and black nucho-cervical collars; dorsal pattern uniformly red (without any trace of dark stripes); white ventral surface, except the gullar region which is dark pigmented (black spoted) to mostly black; terminal shield black, undifferentiated from the tail tip coloration. Duvernoy's gland moderately developed; harderian gland well developed, reaching the temporal region, adductor jaw muscles moderately developed, reaching the dorsal surface of the brain case.

Distribution: Northeastern, central, and southeastern Brazil to neighboring northeastern Argentina; through the Caatinga and Cerrado domains to mesophytic forest of upper Paraná.

Content: At least Apostolepis assimilis (Reinhardt) (Fig. 1), A. cearensis Gomes (Fig. 2), and A. ammodites sp. nov. (Fig. 3). Perhaps also A. quirogai, A. arenaria and A. gaboi (see discutions and conclusions).

Remarks: The monophyly of the assimilis group, as here defined, is still uncertain due to the complete lack of osteological information for several species not analyzed in the present study.

Key to the Apostolepis of the assimilis species group

1. Dorsal pattern uniform red, without any vestige of dark stripes or lines; white and black nucho-cervical collars relatively wide, the posterior edge of black cervical collar reaching at least the 6^th scale of the vertebral row; nasal well separated from preocular by prefrontal-supralabial contact; light snout blotch large, usually extensive to frontal ................................................................................2

1'. Without the above combination of features ........................other Apostolepis

2. Terminal shield mostly whitish, well differentiated from black tail tip coloration; three infralabials contacting anterior chinshields; temporals 0+0 (5^th and 6^th labial contacting parietal); symphisial aproaching to nearly in contact with anterior chinshields ..........................................................................dorbignyi group

2'. Terminal shield uniformly black, undifferentiated from tail tip coloration; four infralabials contacting anterior chinshields; temporals usually 0+1 (last labial separated from parietal); symphisial usually well separated from chinshields by first infralabials ...........................................................................assimilis group

3. Light snout blotch cream; white supralabial blotch moderate sized; ventral surface of head lightly to moderately pigmented; rostral shortly rounded; nucho-cervical (white+black) collars about 23+34 scales long; ventral counts 236267 in males and 246270 in females (southeastern to central Brazil) ...........................A. assimilis

3'. Light snout blotch reddish; white supralabial blotch small to minute; ventral surface of head heavily pigmented to mostly black; rostral at least slightly prominent (if not pointed); nucho-cervical (white+black) collars about 46+47 scales long; ventral counts 215235 in males, 225252 in females ........................................4

4. With an additional white post-cervical collar; rostral nearly roundish, its portion visible from above at most one half of interprefrontal suture length; (central to adjacent northeastern Brazil) ........................................A. ammodites sp. nov.

4'. Without an additional white post-cervical collar; rostral prominently pointed, its portion visible from above at least two thirds of interprefrontal suture length; (northeastern Brazil) ...................................................................A. cearensis

Apostolepis assimilis (Reinhardt, 1861)

(Figura 1, ^7a, ^9a, ^10a)

Elapomorphus assimilis Reinhardt, 1861:235. Type-locality: Minas Gerais, Brazil.

Apostolepis tertulianobeui Lema, 2004b:155160 (New synonym).

Diagnosis: (Tables 1 and 2; Fig. 1) Rostral short and rounded, not prominent; little visible from above; ventrals 236267 in males, 246270 in females; subcaudals 3139 in males, 2531 in females (Fig. 8), light snout blotch cream (^{Fig. 7a}); white supralabial blotch usually covering at least two supralabials; a single white nuchal color 23 scales long, followed by a black cervical color 34, restricted dorsally.

Distribution: Caatingas, central Brazilian cerrados and interspersing mesophytic forest of southeastern Brazil, but uncertain to northern Argentina (Giraudo & Scrocchi, 1998); upper Paraná, upper São Francisco, and eastern Paraguay basins (Fig. 6).

Natural history:Apostolepis assimilis is an inhabitant of open vegetation areas from Atlantic forest and Cerrado, also considered an urban species (Pereira et al., 2004). It preys on small amphisbaenians and scolecophidian snakes (Ferrarezzi, 1993; Marques et al., pers. comm.). A female had six vitellogenic follicles and the largest measuring was 30,91 mm.

Remarks:A. tertulianobeui Lema, 2004 is clearly a member of the assimilis species group (sensu Ferrarezzi, 1993), judging from color pattern and scalation features. In most relevant diagnostic features presented by Lema (2004b), the holotype of A. tertulianobeui does not differ from the range of variation we have observed in a large sample of A. assimilis (Ferrarezzi, 1993). Eventually, a narrower cervical collar can be found in a few individuals as an intrapopulational polymorphism, but a reduced dark pigmentation, as referred by Lema (2004b), may be better explained by fading in the long time preserved holotype specimen. Although there is no specific type-locality, besides the state of Minas Gerais, the distrbution of A. tertulianobeui is cincident with the distributional area of Apostolepis assimilis (Fig. 6). Therefore, even though we did not examine the holotype of A. tertulianobeui, we have no doubt that this name must be relegated as a junior synonym of A. assimilis.

Apostolepis cearensis Gomes, 1915

(Figuras 2, ^7c, ^9c, ^10c)

Apostolepis cearensis Gomes, 1915:122. Type-locality: Estado do Ceará, Brazil.

Apostolepis sanctae-ritae Werner, 1924:43. Type-locality: Santa Rita, now Ibipetuba, Estado da Bahia, Brazil (New synonym).

Apostolepis amarali Werner, 1925:62. Type-locality: South America (here restricted to Northeastern Brazil).

Apostolepis sanctaeritae Lema & Fernandes, 1997:58.

Apostolepis freitasi Lema, 2004:37. Type-locality: Tanque do Aragão, Bahia, Brazil (New synonym).

Diagnosis: (Fig. 2) Rostral somewhat pointed and prominent; nasal well separated from preocular; temporals usually 0+1, the 5^th supralabial in contact with parietal (occasionally 0+0: 5^th+6^th in contact); ventrals 215-237 in males, 227-248 in females; subcaudals 27-32 in males and 23-31 in females (Table 1; Fig. 8); terminal caudal shield pointed and uniformly black; head black postero-laterally, with a large light area on snout (red in life, reaching frontal) (^{Fig. 7}) and small to indistinct light supralabial spot; a broad white nuchal color (covering 45 scale rows) followed by a similar extensive black cervical color; dorsum uniform colored, without stripes; ventral surface uniform white, except by the heavily dark pigmented infralabial-gular region.

Distribution: Widespread within the Caatinga morphoclimatic domain of northeastern Brazil (Fig. 6).

Remarks: The recently described species A. freitasi Lema, 2004 is clearly a member of the assimilis species group (sensu Ferrarezzi, 1993), judging from the features of color pattern and scalation. In all diagnostic features presented by Lema (2004a), the holotype of A. freitasi is identical to A. cearensis. It is not surprising that Lema (2004a) traces no comparison of his new species with the last one, since if this have been done, the author would have failed to find any diagnostic feature at all. The same can be said about the type-locality which is within of the geographic distribution area of A. cearensis (Fig. 6). Therefore, even without having examined the holotype of A. freitasi, we have no doubt that this name must be included as a junior synonym of A. cearensis.

Apostolepis ammodites sp. nov.

(Figuras 3, 4, ^7b, ^9b, ^10b, ^{appendix 2})

Holotype: BRAZIL: TOCANTINS: Palmas (10º02'S, 47º20'W, 230 m altitude), adult , IB 65267, deposited at "Coleção Herpetológica Alphonse Richard Hoge", Laboratório de Herpetologia, Instituto Butantan.

Paratypes: BRAZIL: BAHIA: Cocos (14º11'S, 44º32'W, 559 m altitude), IB 61525; Correntina (13º20'S, 44º38'W, 561 m) CHUNB s/nº; DISTRITO FEDERAL (15º46'S, 47º55'W), IB 49363; GOIÁS: (IB 15723); Cana Brava, Rio Cana Brava (13º30'S, 48º21'W, 400 m), IB 40478; MZUSP 8007; São Domingos (13º23'S, 46º19'W, 667 m), IB 62593, 67392; MATO GROSSO: São Felix, Lago Itacy (11º37'S, 50º40'W, 195 m), IB 15723; MINAS GERAIS: Betim (19º58'S, 44º11'W, 860 m), FUNED 003; Curvelo (18º45'S, 44º25'W, 632 m), IB 22410; Serra do Cipó (19º24'S, 43º19'W, 689 m), MZUSP 7595; Três Marias (18º12'S, 45º14'W, 538 m), FUNED s/nº; Vazante (17º59'S, 49º54'W, 680 m), IB 48041; TOCANTINS: Gurupi (11º43'S, 49º04'W, 287 m); Lageado (09º45'S, 48º21'W, 202 m), IB 64533 (Adult ), IB 64534 (juvenile ), IB 64571 (Adult ), IB 64613 (Adult ), IB 65680 (juvenile ), IB 65681 (Adult ); Palmas (10º02'S, 48º20'W, 230 m), IB 65261 (Adult ), IB 65262 (Adult ), IB 65263 (Adult ), IB 65264 (Adult ), IB 65265 (Adult ), IB 65266 (Adult ), IB 65268 (juvenile ), IB 65269 (Adult ), IB 65664 (Adult ), IB 65688 (Adult ); Porto Nacional (10º42'S, 48º25'W, 212 m), IB 65682 (Adult ), IB 65683 (Adult ), IB 66166, (Adult ); Santa Isabel, Ilha do Bananal (15º17'S, 49º25'W, 611 m), IB 12324; unspecified locality: IB 71470 (juvenile ), IB 71471 (Adult ).

Etymology: Ambivalent connotation, from the Greek, ammos (ammos) meaning sand plus dytes (dytes, the internal 'y' may be rendered by 'i' in Latin) meaning anyone that likes immersions or diving, but furthermore from the Latin ammos meaning around or having two plus dites meaning richly adorned or sumptuous. In allusion both to the habitat and habits of the new species and to its unique extravagant nucho-cervical color pattern having two white collars around a black collar.

Diagnosis: Rostral nearly roundish but somewhat prominent, its portion visible from above about 1/3 to 1/2 of the interprefrontal suture length; ventrals 221235 in males to 234252 in females; subcaudals 3236 in males and 2631 in females (Table 1; Fig. 8); light snout blotch reddish in life (^{Fig. 7}) (fading to yellowish in preservative) and a small to indistinct light supralabial spot; white nuchal collars about 36 scales long, black cervical collar 47 scales long, followed by an additional postcervical white collar 13 scales long; dorsum uniformly red (without stripes); ventral surface uniformly light (unpigmented), except under the head and neck; premaxillae bounded to maxillae; dorsal laminae of nasal contacting frontal and prefrontal; maxillary teeth 4+2, dentary teeth 910; pterygoid teeth 34 (Table 2).

Description and variation: Relatively large species of Apostolepis (maximum total length for a female 634 mm). Head subcylindrical, not distinct from neck, with somewhat prominent but rounded snout; body elongate cylindrical; tail short (males (n = 10): 911% of total length, mean 10%; females (n = 14): 79% of total length, mean 8%), thick and bluntly pointed (^{Appendix 2}). Eye minute, with a subelliptic (nearly round when dilated) pupil.

Scalation: (Fig. 4) Rostral somewhat prominent but roundish, the portion visible from above about 1/2 to 1/3 of its distance to frontal. Nasal always separated from preocular by a broad prefrontal contact with second supralabial. Six supralabials (second and third entering eye). Temporals 0+1, usually only the fifth supralabial in contact with parietal (occasionally also the forth making a slightly contact). Occipitals 35, the lateral very large. Synfisial separated from chinshields by contact between first infralabials; usually seven infralabials, the first four in contact with anterior chinshields, and the forth and fifth in contact with posterior chinshields; anterior chinshields wider and nearly equal to the posteriors in length. Ventrals 221235 (mean = 229.2) in males (n = 10) to 234250 (mean = 242.6) in females (n = 14); subcaudals 3236 (mean = 33.4) (n = 10) in males and 2631 (mean = 28.5) in females (n = 13) (Table 1; Fig. 8); terminal caudal shield short, high and blunt pointed, in contact with 56 scales around.

Dentition and cranial osteology: (Table 2; ^{Figs. 9} and ¹⁰) Tooth series short and with stout teeth. Maxillary very short, with four solid teeth increasing in size posteriorly followed, after a distinct interspace, by a pair of large deeply grooved fangs, situated bellow the eye. Five palatine followed by three or four pterygoid teeth; the anterior end of pterygoid barely extending beyond its articulation with ectopterygoid, and just touching the palatine. Dentary short, bearing 910 subequal teeth, increasing in lenght to the forth and then gradually decreasing to the last one. Reduced posterior process of dentary extending just a behind splenial-angular hinge; meckelian channel closed along most of the dentary length. Ectopterygoid flattened and deeply forked anteriorly, forming a fairly large maxilo-ectopterygoid fenestra. Premaxillae very thick and robust, the lateral processes bounded to the maxillae, and with a stout dorsal process making very broad contact with the dorsal laminae of nasals, which is very wide and nearly rectangular, broader than and making an extensive transversal suture with the frontal and with the antero-dorsal edge of prefrontal. Supratemporals small, distantly separated from parietal and not reaching the posterior level of occipitals. Postorbital bone absent.

Hemipenis: The organ is subcilindrical and stout with a narrower base. Apex is just slightly bilobate, calyculate, and semicapitate. Spines are nearly uniformly distributed around the body. The sulcus spermaticus is forked slightly above the mid-length of the organ, with the branches running centrolineally to the extreme apex.

Coloration in life: (Fig. 3) Dorsum uniform vivid orange-red (without any vestige of dark stripes); head black postero-laterally (to firsts occipital rows), with a large reddish blotch on snout (covering entirely the prefrontals to reach the anterior end of frontal and supraoculars), a small to indistinct cream supralabial spot posterior to the eye is mostly restricted, if not confined, to the forth supralabial; a broad white nuchal collar (36 scales long) followed by a similar extensive black cervical collar (47 scales long); and an additional white postcervical collar (13 scales long); ventral surface uniformly light (unpigmented), except under the head and neck, where the infralabial, chin and gullar regions are mostly black. The dorso-ventral tail coloration is similar to that of body, and with a distinct black band around the tail tip (711.5 scale long); terminal caudal shield uniformly black (not differentiated from tail tip coloration).

Distribution: All the known distribution records for the new species fall within the central and northeastern portion of the Cerrado morphoclimatic domain (Fig. 6).

Natural History:Apostolepis ammodites sp. nov. is a inhabitant of open vegetation areas of Cerrado (Fig. 5), where it appears to be restricted to sandy soils (F. Curcio, C. Nogueira and M.T. Rodrigues, pers. comm.).

A specimen (IB 64613) contained digested remains of skin of a small unidentified Amphisbaena and the head of an ant of the subfamily Attinae. The ant's head is most probably the secondary remain from the stomach content of the amphisbaenid prey.

The new species is oviparous (A. Silveira, pers. comm.). A female (IB 65261, 502 mm SVL) had three vitellogenic follicles and the largest measuring was 15,99 mm. Five males (IB 62593, IB 65262 IB 65263, IB 65268, and IB 65664) had large testes and convoluted deferent ducts. The largest male (IB 62593, 450 mm SVL) had the largest test, with 13,86 mm and was collected in August/2000. The smallest male (IB 65268) with large testes and opaque deferent ducts had 282 mm SVL. It was collected between December/2001 and January/2002.

Characters and character states coded for the cladistic analysis

Characters and respective character states numerically coded and listed below correspond to those in the data matrix of Table 3, which are optimized in the cladogram presented in Figure 11. Linearly ordered multistate transformation series (denoted by NA) were managed under the assumption of additivity during parsimony analysis.

Characters

Coloration (Figs. 1-4, ⁷):

1. Light snout blotch color (^{Fig. 7}):

(0) cream or whitish

(1) reddish

2. Light snout blotch shape:

(0) moderate sized irregularly edged blotch(es), confined to prefrontals dorsally

1) a broad regular blotch covering all snout, reaching frontal and supraoculars

3. White supralabial blotch:

(0) moderate sized, fairly exceeding fourth labial

(1) small (mostly confined on fourth labial) to minute or eventually absent

4. Pigmentation of the ventral head surface:

(0) virtually immaculate

(1) irregularly dark blotched.

(2) mostly black

5. White nuchal collar:

(0) narrow (12 scales long) if present

(1) moderately broad (24 scales long)

(2) large (46 scales long)

6. Black cervical collar:

(0) narrow (1 scale long)

(1) moderate (23 scales long)

(2) wide (46 scales long)

(3) extremely long (57 scales long)

7. Anterior indentation of black cervical collar along vertebral row:

(0) absent

(1) present

8. Ventral extension of black cervical collar:

(0) absent without vestiges

(1) variable, but usually at list a vestige

9. White post-cervical collar:

(0) absent

(1) present

10. Terminal shield color:

(0) mostly whitish, differentiated from tip of tail

(1) uniformly black, undifferentiated from tip tail

11. Rostral shield:

(0) short and rounded

(1) slightly prominent but roundish

(2) prominently pointed

12. Temporals: Comment: The usual condition in A. cearensis is state (0), although polymorphism has been observed in some populations. A single species of the flavotorquata group exhibits intraspecific polymorphism.

(0) 0+1 (posterior temporal present between last supralabial and parietal)

(1) 0+0 (no temporals, last two supralabial contacting parietal)

13. Infralabials contacting anterior chinshields:

(0) four

(1) three

14. Ventral counts (modal): Comment: The dorbignyi group exhibits interspecific variation and were coded as missing data.

(0) 217232 in males, 227248 in females

(1) 241261 in males, 254265 in females

15. Subcaudal counts (modal):

(0) 3236 in females, 3741 in males

(1) 2731 in females, 3237 in males

(2) 2327 in females, 2831 in males

Cranial osteology (^{Figs. 9}, ¹⁰, Tab. 2):

16. Number of pterygoid teeth:

(0) 34 (mode 4)

(1) 2

17. Number of dentary teeth (mode):

(0) 78

(1) 910

18. Dentary shape:

(0) nearly straight

(1) curve

19. Prearticular crest of compound bone:

(0) higher than surangular crest posteriorly

(1) as low as surangular crest

20. Shape of mentonian foramen:

(0) elongate groove

(1) rounded

21. Shape of premaxillae:

(0) roughly T-shaped, with a distinct ascending process

(1) stoutly enlarged without a distinct ascending process, but not pointed

(2) strongly enlarged and prominently pointed

22. Septomaxillary process of premaxillae:

(0) short and widely separated from each other.

(1) moderately long and close to each other

(2) long and closer to each other

23. Contact between premaxillae and maxillae:

(0) distantly separated

(1) separated

(2) bound to each other

24. Width of facial segment of skull:

(0) relatively wide, conferring a larger palatine-ectopterygoid distance

(1) relatively narrower

25. Ectopterygoid shape:

(0) wider and flat

(1) narrow and sub cylindrical

26. Contact between dorsal laminae of nasals and frontals:

(0) absent

(1) present, forming a transverse suture

27. Shape fronto-parietal suture in dorsal view:

(0) U-shaped, no parietal indentation between frontals

(1) W-shaped, with an antero-median parietal indentation between frontals

28. Width of fronto-parietal suture:

(0) short and wide

(1) narrower and longer

29. Contact between nasal and prefrontal:

(0) absent

(1) present dorsally

(2) present laterally

Ecology (Fig. 5):

30. Habitat (soil):

(0) non sandy soils

(1) sandy soils

31. Habitat (physiognomy):

(0) mesophytic forests

(1) mesic savannas

(2) semiarid caatingas

Results of Cladistic analysis

Parsimony analysis of the data matrix in Table 3 resulted in a single most parsimonious tree, which was rooted in the flavotorquata species group (Fig. 11). This cladogram presents a highly consistent pattern of character congruence, as denoted by its consistency and retention indexes: length = 43; CI = 0.97, RI = 0.97 (the values are the same under both Wagner and Fitch parsimony options applied to ordered multistate characters). Even if the total amount of data is partitioned into different character classes (vis: coloration, scalation, osteology, or ecology), the same single tree topology is obtained from each independent analyses of the four subsets.

The results of the cladistic analysis demonstrate that A. ammodites sp. nov. is closer to A. cearensis than A. assimilis (Fig. 11). This last species appears as the most conservative lineage within the clade. At least with regard to the character sample analyzed here, it remains nearly identical to the common ancestor of the tree species. On the other hand, the lineage leading to A. ammodites sp. nov. and A. cearensis is the most divergent lineage morphologically, judging from an amount of synapomorphic and autapomorphic traits of coloration and, especially, cranial osteology (^{Figs. 9}, ¹⁰ and 11).

Regarding the evolutionary rate of the different classes of morphological data, the scalation features represent the most conservative ones, while the cranial osteology is the most divergent.

DISCUSSION AND CONCLUSIONS

At least three other species exhibiting entirely black terminal shield may be further proven to belong to the assimilis group Ferrarezzi, unpublished data). These are Apostolepis arenaria Rodrigues, 1993, and A. gaboi Rodrigues, 1993, a pair of closely related species that shares apomorphies with both A. ammodites sp. nov. and A. cearensis, plus a few additional features only with the latter, (and thus, may be sister to A. cearensis). On the other hand A. quirogai Giraudo & Scrocchi, 1998 lacks most derived features of the other species, and may represent the sister taxa of the whole group. The phylogenetic relationships of all of them are left to be discussed in a further study. By now, it is sufficient to explain that keeping them out of here will not disqualify our analysis about the pattern of cladistic relationships of the other species, since they have also been excluded from outgroup comparison.

At least one of the apomorphies determined within the assimilis group, a higher degree of rostral prominence, is also shared by several species referred to the dimidiata and arenaria groups of Apostolepis (Ferrarezzi, 1993). Absence of rostral prominence is inferred as a plesiomorphic within Apostolepis since it is the common condition in the closely related genera Elapomorphus and most species of Phalotris (excluding the nasutus group, Ferrarezzi, 1993, 1994). An ad hoc assumption of our present phylogenetic hypothesis is that rostral prominence has evolved independently in several elapomorphine lineages. However, a better understanding of phylogenetic relationships, in addition to detailed information on morphometrics and, hopefully, also on ecology (including detailed descriptions of microhabitat use) are still necessary for reliable interpretations of the possible phylogenetic meaning and adaptive value of an enlarged rostral scale in these groups of Phalotris and Apostolepis.

The present discovery of new diagnostic characters arranged in a simple congruent pattern of relationships, although just for a few species, stands as a first step towards a comprehensive resolution of the whole group. The original synapomorphy squeme proposed here represents a testable phylogenetic hypothesis to be corroborated, modified or (mostly important) even refuted, after an advanced survey on these character states for other putative members of a broad - developed (Fig. 12). The biological area cladogram resulting from the recovered cladistic relationships for the assimilis group of Apostolepis indicates that the northeastern portion of the Cerrado is more closely connected to the Caatinga than to the southwestern Cerrado region. Further area cladograms derived from other groups of organisms exhibiting the same areas of endemism are needed in order to test the generality of this conjecture.

ACKNOWLEDGEMENTS

We are indebted to R.J. Sawaya, for kindly taking countless photographies of preserved specimens and skulls (don't forgetting the "Eye Series"). F. Curcio contributed to the preparation of osteological material. O.A.V. Marques, C. Nogueira, F. Curcio, A.C.O.R. Costa, and M.T. Rodrigues provided images of live specimens and their natural habitats, bringing us important information about species in the field, as well as fruitful discussions about the subject. D. Embert provided new information on the variation of A. multicincta. M.T. Rodrigues furnished valuable suggestions and criticism during the elaboration of the senior author's dissertation on elapomorphine systematics, and P.E. Vanzolini kindly shared his annotated literature on Apostolepis taxa. We are indebted to F.L. Franco for advisory help with literature, V.J. Germano for cooperative work with preserved collection material, A. Dall'Onder for assitance with images, and S.M. Almeida-Santos for comments on reproductive biology. This paper was partially funded by grants from the Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) and from FUNDAP.

Recebido em: 15.06.2005

Aceito em: 26.08.2005

Material Examined

Museum Collections:

IB (Laboratório de Herpetologia, Instituto Butantan, São Paulo);

MZUSP (Museu de Zoologia, Universidade de São Paulo);

MNRJ (Museu Nacional do Rio de Janeiro);

UNB (Dept.Zoologia, Universidade de Brasília);

LAPH (Lab. An. Peç. e Herpet., Universidade Estadual de Feira de Santana, Bahia).

Apostolepis assimilis: BRAZIL: SÃO PAULO: Carapicuíba (IB 72970); Icem/Fronteira (IB 40008); Itu (MZUSP 4180, 6606); São Paulo (IB 24180, 24548, 24873, 27598, 31716, 33316; IB-MSP348), Pirituba (IB 70351 SKULL), Jaraguá (IB 70356 SKULL); Santana do Parnaíba (IB 61761); DISTRITO FEDERAL: Brasília (IB 20566, 28734, CHUNB24456, 24474); MINAS GERAIS: Conceição dos Ouros (IB 33206); Uberlândia (MZUSP 3841, 3845). GOIÁS: Jataí (MZUSP 3783); Mineiros (IB 55495); MATO GROSSO: Buriti, Chapada dos Guimarães (MZUSP 5346); MATO GROSSO DO SUL: Amambaí (IB 41163); Campo Grande (IB 42978, 57222; MZUSP 10155); Nova Andradina (IB 27489); Paranaíba (IB 45615); Ponta-Porã (IB 44065); PARANÁ: Londrina (IB 37462, 40008).

Apostolepis cearensis: BAHIA: Brumado (IB 33685); Camaçarí (LAPH371); Capim Grosso (LAPH294); Feira de Santana (LAPH012, 019, 070, 071, 074 086, 130, 162, 166,203, 277, 310, 315, 429, 434, 463, 464, 505, 515, 615, 624, IB 53700); Mina Caraíba, Jaguarari river (IB 26203); Santo Inácio (MZUSP -MTR.field# 916639, 916640, 916641 (incl.HEMIPENIS); São Gonçalo (LAPH073); CEARÁ: (IB 14948), (IB 0882, holotype, 0910, 0911, 20220 including skull); Crato, Serra do Araripe (IB 20385); Fortaleza (IB 2106, 20020, 55318, 40262); Juazeiro do Norte (IB 20164); Lima Campos, Icó (IB 20190); Limoeiro do Norte (IB 12775); Northeastern Ceará (IB 20097); RIO GRANDE DO NORTE: Natal (IB 43960, 55064-67); PERNAMBUCO: Recife (IB 56983); PIAUÍ: Terezina (IB 49743); PARAÍBA: Cabaceiras (MZUSP 9013); Campina Grande (IB 9050); João Pessoa (MZUSP 7975); (MNRJ r3371a, r3371b).

Apostolepis ammodites sp. nov.: BAHIA: Cocos (IB 61525), Correntina (CHUNB s/nº); DISTRITO FEDERAL (IB 49363); GOIÁS: (IB 15723); Cana Brava, Rio Cana Brava (IB 40478); São Domingos (IB 62593, 67392); MATO GROSSO: São Felix, Lago Itacy (IB 15723); MINAS GERAIS: Betim (FUNED003); Curvelo (IB 22410); Serra do Cipó (MZUSP 7595); Três Marias (FUNED s/nº); Vazante (IB 48041); TOCANTINS: Gurupi (MZUSP 8007); Lajeado (IB 64533, 64534, 64571, 64613, 65680, 65681); Palmas (IB 65261-69, 65664, 65688); Porto Nacional (IB 65682, 65683, 66166); Santa Isabel, Ilha do Bananal (IB 12324); unspecified locality (IB 71470, 71471).

Apostolepis flavotorquata: BAHIA: Coribe (UNB1213); GOIÁS: Anápolis (IB 18851); Nerópolis (UNB0948); MATO GROSSO: Chapada Matogrossense (IB 32530); MATO GROSSO DO SUL: Ilha Solteira (IB 36449, 36450 SKULL).

Apostolepis pymi: AMAZONAS: Balbina (IB 51396); Manaus (MZUSP 8423, 8657); PARÁ: Belém (IB 3033, 3034); Canindé, Rio Guruoi (MZUSP 2001); (uncataloged specimen- SKULL),

• Amaral, A. 1930 [1929]. Estudos sobre ophídios neotrópicos XVII. Valor systemático de várias formas de ophídios neotrópicos. Memórias do Instituto Butantan, 4:32-68.
• Amaral, A. 1936. Contribuição ao conhecimento dos ophídios do Brasil. VIII. Lista remissiva dos ophídios do Brasil. 2Ş ed. Memórias do Instituto Butantan, 10:87-162.
• Dowling, H.G. 1951. A proposed standard system of counting ventrals in snakes. British Journal of Herpetology, 1:97-99.
• Dowling, H.G. & Savage, J.M. 1960. A guide to snake hemipenis: a survey of basic structure and systematic characteristics. Zoologica New York, 45:17-28.
• Farris, J.S. 1982. Outgroups and Parsimony. Systematic Zoology, 31:328-334.
• Ferrarezzi, H. 1993. Sistemática Filogenética de Elapomorphus, Phalotris e Apostolepis (Serpentes: Colubridae: Xenodontinae). Dissertação (Mestrado). Universidade de São Paulo.
• Ferrarezzi, H. 1994 [1993]. Nota sobre o gênero Phalotris com revisão do grupo nasutus e a descrição de três novas espécies (Serpentes: Colubridae: Xenodontinae). Memórias do Instituto Butantan, 55(1):21-38.
• Giraudo, A.R. & Scrocchi, G.J. 1998. A new species of Apostolepis and comments on the genus in Argentina. Herpetologica, 54(4):470-476.
• Gomes, J.F. 1915. Contribution to the knowledge of Brazilian snakes. Description of four new species and a new genus of opisthoglyphs. Annais Paulista de Medicina e Cirurgia, 4:121-129.
• Harvey, M.B. 1999. Revision of Bolivian Apostolepis (Squamata: Colubridae). Copeia, 2:388-409.
• Harvey, M.B.; Gonzales, L.A. & Scrocchi, G.J. 2001. New species of Apostolepis (Squamata: Colubridae) from the Gran Chaco in Southern Bolivia. Copeia, 2:501-507.
• Hennig, W. 1966. Phylogenetic Systematics. University of Illinois Press, Illinois.
• Hueck, K. & Siebert, P. 1972. Vegetationskarte von Südamerika. Gustav Fisher Verlag, Stuttgart.
• Lema, T. 2002. Redescription of Apostolepis sanctaeritae Werner, 1924 (Serpentes, Colubridae), and a comparsion with related species. Copeia, 4:1147-1151.
• Lema, T. 2003. Descrição de nova espécie de Apostolepis Cope, 1861 do cerrado do Brasil, pertencente ao grupo dimidiata (Serpentes, Elapomorphinae). Acta Biológica Leopoldensia, 25(1):123-131.
• Lema, T. 2004a. New species of Apostolepis Cope, 1861, from caatinga belonging to assimilis group (Serpentes, Elapomorphinae). Comunicações do Museu de Ciências da PUCRS. Serie Zoologia, 17(1):3-7.
• Lema, T. 2004b. Description of a new species of Apostolepis Cope, 1861 (Serpentes, Elapomorphinae) from brazilian Cerrado. Acta Biológica Leopoldensia, 26(1):155-160.
• Lema, T. & Fernandes, R. 1997. O status de Apostolepis sanctaeritae Werner, 1924 e sua revalidação (Serpentes: Colubridae: Elapomorphinae). Acta Biológica Leopoldensia, 19:51-59.
• Manzani, P.R. & Abe, A.S. 1988. Sobre dois novos métodos de preparo do hemipênis de serpentes. Memórias do Instituto Butantan, 50:15-20.
• Nixon, K.C. & Carpenter, J.M. 1993. On outgroups. Cladistics, 9:413-426.
• Pereira, D.N.; Rodrigues, M.G.; Barbo, F.E.; Stender-Oliveira, F.; Calleffo, M.E.V. & Marques, O.A.V. 2004. Composição atual da fauna de serpentes da cidade de São Paulo. In: Congresso Brasileiro de Herpetologia, 1. Resumos. Pontifícia Universidade Católica do Paraná, Curitiba. CD-ROM.
• Pesantes, O.S. 1994. A method for preparing the hemipenis of preserved snakes. Journal of Herpetology, 28:93-95.
• Peters, J. A. & Orejas-Miranda, B. 1970. Catalogue of the Neotropical Squamata. Part I. Snakes. Bulletin United States National Museum, 297:1-347.
• Reinhardt, J. 1861. Herpetologiske Middelelser. II. Beskrivelser af nogle nye til Calamariernes Familie henhörende Slänger. Videnskabelige Meddelelser fra Dank Naturhistorisk Forening i Kjöbenhavn, 2:229-250, [1860], pl.4, figs. 1-5.
• Rodrigues, M.T. 1993 [dated 1992]. Herpetofauna das dunas interiores do Rio São Francisco: Bahia: Brasil. V. Duas novas espécies de Apostolepis (Ophidia: Colubridae). Memórias do Instituto Butantan, 54 (2):53-59.
• Vanzolini, P.E. 1986. Addenda and corrigenda to the Catalogue of Neotropical Squamata. Smithsonian Herpetological Service, 70:1-26.
• Werner, F. 1924. Neue oder wenig bekannte schlangen aus dem Naturhistorischen Staatmuseum in Wien. Sitzungsberichte der Akademie der Wissenschaften in Wien, 133:29-56.
• Werner, F. 1925. Neue order wenig bekannte Schlangen aus dem Naturhistorichen Staatsmuseum in Wien (2). Sitzungsberichte der Akademie der Wissenschaften in Wien, 134:45-66.
• Zaher, H. 1999. Hemipenial morphology of the South American xenodontine snakes, with a proposal for a monophyletic Xenodontinae and a reappraisal of colubroid hemipenes. Bullletin of the American Museum of Natural History, 240:1-168.

APPENDIX 1 

appendix 2

Publication Dates

• Publication in this collection
  30 Nov 2005
• Date of issue
  2005

History

• Received
  15 June 2005
• Accepted
  26 Aug 2005