Composition and natural history notes of the coastal snake assemblage from Northern Bahia, Brazil

Abstract Information about the snake diversity and their natural history from the Atlantic forest domain in Brazil refer mostly to inland forests than to coastal region. Within the state of Bahia, this knowledge is concentrated to the southeastern coastal stretch. Herein we report on the diversity of snakes from the restinga, ombrophilous forest and anthropogenic environment from the northern Atlantic coast of Bahia. We sampled nine sites for three years and visited four museum collections. Furthermore, we provide anecdotal natural history information, voucher analyses, literature complements, and a key to fascilitate species identification. We report a total of 774 snakes belonging to 50 species and 23 new distribution records for northeastern coast of Bahia, supplemented by new data on feeding and reproduction. The number of detected species is similar to numbers obtained in comparable studies from other Brazilian ecoregions. This study reports and focuses for the first time on all known species of snakes from the northeastern coast of Bahia.


Introduction
Studies on diversity inventories and natural history increase the knowledge of a regional fauna, its interaction with other organisms, and the environment in general (Greene 1994). As such, they provide relevant basis data to better understand the complexity of ecosystems. This is not different with snakes, which represent an integral part of such ecosystems, be them as prey, predator, host and other functions within the network of ecosystem relations. Therefore, snakes, as well as other constituents of a local ecosytems, are important and need to be understood and protected. Hence, novel data about our target group is elementary for further conservation studies on these taxa Losos 1988, Shine andBonnet 2009).
As a first step, pertinent and relevant taxa-related information can be acquired through species inventories and studies describing aspects of natural history that increase our knowledge on the different habitats and habits of regional snakes (Greene 1997). These studies include biodiversity components in terms of richness, abundance, distribution, diet composition, activity periods, reproduction, morphological variation, parasitism, predators and other intrinsic data of the group (Martins and Oliveira 1998). All these aspects enhance the knowledge of snakes (Cadle and Greene 1993), but these studies are not evenly distributed among the different Brazilian ecoregions. For example, there are several contributions on snake communities from the Amazon (e.g. Cunha and Nascimento 1978, Martins and Oliveira 1998, Santos-Costa et al. 2015, as well as the Cerrado (e.g. Recoder andNogueira 2007, Sawaya et al. 2008), Caatinga (e.g. Vanzolini et al. 1980, Vitt and Vanglinder 1983, Guedes et al. 2014 and Pantanal (e.g. Strüssmann and Sazima 1993). Regarding the Atlantic forest, the snake fauna from the Southeast and South of Brazil is well represented (e.g. Marques et al. 2001, Marques and Sazima 2004, Hartmann et al. 2009a, but despite the increase of these studies in recent years, the natural history of these animals from most northeastern states of Brazil has remained poorly documented. Studies involving snakes from northeastern Brazil are often species inventories from the Caatinga and transition areas with other ecoregions (Vanzolini et al. 1980, Vitt and Vangilder 1983, Pereira Filho and Montingelli 2011, Loebmann and Haddad 2010, Ribeiro et al. 2012, Garda et al. 2013, Guedes et al. 2015. Unfortunately, the Atlantic forest in northeastern Brazil is represented by few studies only (Argôlo 2004, Santana et al. 2008, Morato et al. 2011, França et al. 2012, rendering a regional study, such as the one presented herein, the more important. A total of 131 snake species are currently known from the state of Bahia (Hamdan and Lira-da-Silva 2012, Curcio et al. 2012, Fernandes and Hamdan 2014, but there is a lack of detailed studies. Argôlo (2004) provides the most complete compilation of the snake assemblage of the Atlantic forest in southern Bahia, reporting species from different forested, disturbed and open habitats. Dias and Rocha (2014) surveyed the herpetofauna from restinga in the northern and southern coast of the state, while Marques et al. (2011) recorded snakes from a single restinga location in the northern coast of Bahia. Herein we include: (1) a checklist of all snake species from the northern coast of Bahia; (2) natural history information (frequency, size, distribution, habitat, microhabitat, activity and environmental variables, diet, reproduction and defensive behavior); (3) conservation aspects, and (4) an identification key to the snakes from this region.

Study area
The present study was conducted on the northern coast of Bahia, which extends for 220 km from Salvador to the boundaries of the state of Sergipe ( Figure 1). The climate in this region varies from humid to subhumid with a weak dry season in the southernmost portion and a two months dry season in the remaining area. Precipitation fluctuates from 1300 to 1900 mm per year ( Figure 2) (IBGE 2002, SEI 2013. The vegetation in the coastal municipalities (next lower subdivision of the state) consists of pioneer vegetation with marine influence, called restinga. Four typical vegetation types compose the restinga of the region: beach vegetation, flooded plain, shrub vegetation and restinga dry forest (Menezes et al. 2009(Menezes et al. , 2012. Farther inland, the landscape changes to dense ombrophilous forests and semideciduous stationary forests, both with Cerrado enclaves (SEI 2013).
Four sampling sources were used to acquire data on snakes: (1) Visual Encounter Survey (VES) limited by time, i.e. recording visually detected and manually captured snakes; (2) dead specimens found by locals; (3) carcasses and shed skins detected by us; and (4) incidentally encountered and recorded snakes outside the VES. For the VES, we walked for two hours on four 500 m transects placed on four aforementioned vegetation types. The VES covered the period from 06:00 h and 18:00 h in six turns: 06:00-08:00; 08:00-10:00; 10:00-12:00; 12:00-14:00; 14:00-16:00; 16:00-18:00. In the last year VES was performed during nighttime from 19:00 h to 21:00 h at Busca Vida, Baixio, Barra do Itariri and Instituto da Mata, totaling a sample effort of 3024 h (2592 h diurnal and 432 h nocturnal). During the VES and incidental encounters we recorded for each snake the time of activity, predominant vegetation type and microhabitat. Each captured snake was marked and then released at its site of capture. Marking involved fluorescent elastomer applied subcutaneously in the ventral scales. The i th number of the ventral scale, with the first number being the ventral scale anterior the anal plate, corresponding to the i th number of the captured snake (adapted from Hutchens et al. 2008). We used four different colors, one for each vegetation type to see if animals were moving among the vegetation types in case of recapture. Voucher specimens are deposited in the Coleção Herpetológica do Centro de Ecologia e Conservação Animal at the Universidade Católica do Salvador.

b) Animal characterization
All species were identified according to the literature (Peters and Orejas-Miranda 1970;Vanzolini et al. 1980, Cunha and Nascimento 1978, Dixon 1986, 1989, Dixon et al. 1993, Campbell and Lamar 2004, Fernandes 2006, Albuquerque 2008, Lima 2012, Ascenso 2013, Pires et al. 2014. Snout-Vent Length (SVL) and sex was recorded from all captured animals and museum specimens when possible. We observed predation events, and during the last year of surveys, we forced animals to regurgitate their stomach content through manual palpation.
Morphological characteristics of the animals from the region comprising scale counts, body shape and coloration allowed us to elaborate an identification key to all species recorded in this study. We complemented the information of animals with records obtained from the literature, as well as field observations and data from voucher specimens. We did not include natural history notes of introduced (nonnative) species.  (CHECOA, Salvador). Considering the extensive region and lack of surveys from northeastern Bahia, and in particular its inland portions, we included additional records using voucher specimens from 10 neighbouring municipalities: Lauro de Freitas (-12.869 S, -38.315 W), Simões Filho (-12.768 S, -38.407 W), , Pojuca (-12.410 S, -38.321 W), Catu (-12.323 S, -38.420 W), Itanagra (-12.302 S, -38.059 W), Araçás (-12.16 S, -38.123 W), Cardeal da Silva (-12.072 S, -38.98 W), Acajutiba (-11.670 S, -38.017 W) and Rio Real (-11.502 S, -37.955 W) ( Figure 1). No snakes were obtained from the municipalities of Araçás, Acajutiba, Cardeal da Silva and Rio Real. When possible, stomach contents and reproductive status were verified. Vouchers are listed in Appendix 1.

Statistics
Species richness and frequency were computed based on the species sampled during fieldwork only, excluding specimens housed in collections. The frequency of the species from the northern coast of Bahia followed the model adapted by Mesquita et al. (2013) (extracted from Abreu and Nogueira 1989and Luiselli 2006. The frequency of a certain snake species is computed by dividing the number of samples (18 bimonthly field trips) containing the target species by the total number of samples × 100. The resulting percentage classified the species as accidental (0.1%-25%), accessory (25%-50%) or constant (50%-100%). We calculated the dominance with the number of specimens of the target species divided by the total number of snakes recorded during fieldwork × 100, applying the categories: accidental (0%-2.5%), accessory (2.5%-5%) and dominant (5%-100%). The combination of frequency and dominance enables the classification of the species as: 1) very rare (accidental/accidental, dominance < 1%), 2) rare (accidental/accidental), 3) intermediate (constant/accessory, constant/accidental, accessory/accidental, accessory/dominant, accessory/accessory), and 4) common (constant/dominant).

Results
A total of 194 snakes were recorded during fieldwork and 580 obtained from museum collections, totaling 774 snakes of a total of 49 native species. Of those, 32 species were recorded during fieldwork and 17 species were complemented through museum records and literature references. Twenty-three new distributional records of species are added for the northern coastal stretch of Bahia (see Table 1 for details). In the municipality of Camaçari one specimen of an introduced species, Pantherophis guttatus, which is native to North America (see Fonseca et al. 2014) was detected. As it is non-native, any natural history descriptions were excluded, but included in the frequency statistics.
Diurnal activity was recorded for 158 snakes during all surveys with members of the families Typhlopidae, Colubridae (except T. melanocephala) and Elapidae restricted to diurnal activity. Our observations show snakes were mostly active during morning between 08:00 h and 11:59 h and less frequently observed during the afternoon ( Figure 3). During the hottest period of the day (12:00 h -14:00 h) we recorded 17 snakes; eight of those were P. nattereri foraging on the ground. This species might tolerate high temperatures, since we observed an adult (SVL ≥ 800 mm) moving on the sandy soil at 55 °C. We present occasional encounters and nighttime surveys from the last year (18:00 h-21:00 h; Figure 3), when we observed mostly H. angulatus, L. annulata, Oxyrhopus spp. and viperids. We also recorded 13 snakes on roads of which four were killed by cars.

Food items
Stomach contents of 11 species were analysed. Lizards represented 40% of food items, followed by amphibians (20%), mammals and snakes (15% each), fishes and invertebrates (5% each). The lizard Ameivula ocellifera was the most recurrent prey item in restinga with observed predatory events between 09:50 h and 13:06 h. Two isolated events of Oxyrhopus trigeminus feeding on A. ocellifera were recorded, when the snakes (SVL = 320 mm and 315 mm) preyed on three lizards (SVL = 53.8 mm and 51.9 mm; a partially digested specimen, respectively). Another lizard species ingested by O. trigeminus was Tropidurus hygomi. In addition, A. ocellifera was also a food item of Philodryas nattereri in two different events. In the first, the snake (SVL = 390 mm) regurgitated a partially digested A. ocellifera. In the second event, an adult P. nattereri subdued the lizard and disappeared immediately thereafter with it. We report the first predatory event of the Table 1. Snake species from the northern coast of Bahia. Number of specimens from fieldwork (N), frequency (f%), dominance (D%) and number of specimens from museum specimens (NC). New records (*); Introduced species (**); Extracted from Dias and Rocha (2014) (***).  (SVL = 190 mm). We recorded the ingestion of Leptodactylus natalensis by Helicops angulatus in two events, both partially digested. One of these events also revealed frog eggs, possibly of the same prey species. Herein we report the first predation record of Amerotyphlops  (2010) reported a bluish coloration in offsprings. This observation is confirmed from a young museum specimen (SVL ≤ 185 mm) with remnants of juvenile bluish coloration transitioning into brownish coloration. Strüssmann and Sazima (1993) reported insects as part of its diet.

Fig. 4C
A very rare species (see Marques et al. 2012c). It was recorded in an ombrofilous forest remnant on the leaf litter basking in the sun at 14:00 h in September (SVL = 1140 mm). Corallus hortulanus is arboreal, often found on vegetation of different habitats during the night (Cunha andNascimento 1978, Guedes et al. 2014). Feeds on frogs, lizards, birds and small mammals including bats (Henderson et al. 1995, Kok et al. 2006, Pizzatto et al. 2009). We observed this species in an S-coil position, when encountered, followed by strikes and bites when molested.

Epicrates assisi Machado, 1945 Fig. 4D
A single specimen was recorded in the municipality of Mata de São João in ombrophilous forest habitat with a cerrado enclave. After erected to a species level by Passos and , the distribution of E. assisi is restricted to open habitat of northeastern Brazil but it also apparently colonized disturbed habitats that once were forested (Morato et al. 2011, Garda et al. 2013. It is a nocturnal and terrestrial species (Rodrigues and Prudente 2011) which feeds on small mammals (Pizzatto et al. 2009).

Fig. 4E
A rare species, large sized (min SVL = 1000 mm, max SVL = 3340 mm). Recorded at Praia do Forte, Imbassaí, Massarandupió and Baixio. Due to the aquatic life habit, this species was recorded only on flooded plains in the restinga (n = 3) on May and June, during periods of heavy precipitation. The earliest record was at 07:56 h when the animal was basking among reed vegetation (Juncus sp.) in a riverbed, whereas the other observations occurred around 13:00 h. Of those two, one specimen was crossing a road between two lakes and the other was basking in a flooded area among herbaceous vegetation. The species is strongly associated with wetlands, inhabiting rivers and ponds from the Atlantic forest and Amazon, but also moving on land to additionally feed and give birth (Cunha andNascimento 1978, Argôlo 2004). Feeds on fishes, frogs, lizards, crocodilians, birds and mammals (Beebe 1946, Cunha and Nascimento 1978, Henderson et al. 1995, Martins and Oliveira 1998, Pizzatto et al. 2009).

Fig. 4F
Dias and Rocha (2014) recorded the species in restinga habitat at Costa Azul. Despite being most frequently encountered in forested habitats and riparian forests, it also occurs in open habitats (Marques and Sazima 2004, Vaz-Silva et al. 2007, Garda et al. 2013). The species is diurnal and forages on soil and in vegetation. It feeds on frogs (Dixon 1993, Marques et al. 2001.

Chironius carinatus (Linnaeus, 1758) Fig. 4G
A rare species. Recorded at the Instituto da Mata in March and July. Detected in an ombrophilous remnant forest adjacent to a pasture. We observed one large specimen (total length > 1000 mm) swimming in a stream at 16:22 h, while the other record is based on a shed skin. The species inhabits forests, riparian forests, igapós, cocoa plantations, pastures and open habitats. Chironius carinatus is diurnal, semiarboreal and terrestrial (Cunha and Nascimento 1978, Vanzolini et al. 1980, Argôlo 2004. It feeds on frogs, birds and lizards (Beebe 1946, Dixon 1993.

Fig. 4H
A very rare species of medium size (min SVL = 190 mm, max SVL = 620 mm). Recorded at Praia do Forte, Imbassaí, Massarandupió and Costa Azul. We observed one specimen foraging on bromeliads in restinga dry forest at 08:33 h in September.  recorded the species on a flooded plain. Voucher specimens originated from restinga (n = 9) and ombrophilous forest (n = 5). Chironius exoletus also inhabits forested habitats and pastures. The species is diurnal, terrestrial and semiarboreal, feeding on frogs (Marques and Sazima 2004, Bernarde and Abe 2006, 2010, Hartmann et al. 2009b. Some tried to bite when handled.

Fig. 4I
A common species of medium size (min SVL = 274 mm, max SVL = 1000 mm). Recorded at Busca Vida, Guarajuba, Praia do Forte, Imbassaí, Massarandupió, Baixio and Barra do Itariri. Occurs in the whole region in restinga, ombrophilous forest and urban environments. The species inhabits predominantly open habitats, but also forests, urban and disturbed environments (Carvalho and Nogueira 1998, Vaz-Silva et al. 2007, França et al. 2012, Miranda et al. 2012. Observed from February to December, active from 06:43 h to 17:20 h. We observed specimens on soil (n = 9), in bushes (n = 5), on the leaf litter (n = 3), coiled on a branch during nighttime, in bromeliad, herbaceous vegetation, pond and on a fallen trunk (n = 1 each microhabitat). Feeds on frogs, mainly hylids (Pinto et al. 2008). In Cerrado, females contained oviductual eggs from October to March (Pinto et al. 2010). In September we detected a gravid female (SVL = 610 mm; 3 eggs; 28.5 -31.2 mm). The species occasionally bites when handled. One specimen was observed to move from leaf litter to suspended branches to evade capture.

Fig. 4K
Medium sized snake (min SVL = 220 mm, max SVL = 570 mm). Recorded in the municipality of Catu in ombrophilous forest. It inhabits mainly forested habitats, transition areas and disturbed habitats . Diurnal and terrestrial, occasionally resting on vegetation (Martins and Oliveira 1998). It feeds on frogs and lizards Soini 1986, Martins andOliveira 1998).

Fig. 4L
A very rare species of medium size (min SVL = 595 mm, max SVL = 850 mm). Recorded at Reserva Sapiranga and Instituto da Mata. We observed a specimen on a suspended branch at 13:22 h in December. Specimens were recorded in restinga dry forests. Inhabits pastures, restinga and urban environment in other regions (Carvalho and Nogueira 1998, Bernarde and Abe 2006, Miranda et al. 2012. Feed mostly on hylid frogs and occasionally on lizards. The species is diurnal and arboreal, foraging in vegetation (Albuquerque et al. 2007). When handled the species intimidates with mouth gaping and bites.

Fig. 4M
Medium sized snake (min SVL = 300 mm, max SVL = 802 mm). Recorded in the municipalities of Lauro de Freitas, Mata de São João, Catu and Jandaíra, without information on habitat. Terrestrial and diurnal, inhabits forested and deforested environments (Strüssmann and Sazima 1993). Argôlo (2004) and Leite et al. (2007) observed most specimens in plantation and around human habitation. Feeds preferably on frogs, also lizards, snakes, birds and mammals (Strüssmann and Sazima 1993, Leite et al. 2007, Marques and Muriel 2007. Fig. 4N An intermediate frequent species of medium size (min SVL= 236 mm, max SVL = 754 mm). Recorded at Arembepe, Praia do Forte and Imbassaí from June to November. Found in activity between 09:05 h and 12:45 h. We observed individuals on suspended branches (n = 2) and bush (n = 1). In the region, it inhabits restinga and ombrophilous forest. In other ecoregions, it occupies also primary forests, disturbed and open habitats (Keiser 1982, Miranda et al. 2012, Garda et al. 2013). The species is arboreal and diurnal, feeding mostly on lizards and occasionally frogs Oliveira 1998, Marques et al. 2001). We observed mouth gaping or immediate fleeing.

Fig. 4P
A rare species of large size (in SVL = 139.1 mm, max SVL = 1600 mm). Recorded at Busca Vida, Praia do Forte and Imbassaí in July and August. We observed an individual foraging on the leaf litter at 15:46 h. It inhabits forests, open and disturbed habitats, is diurnal and semiarboreal (Vanzolini et al. 1980, Argôlo 2004, Bernarde and Abe 2006. It feeds on frogs, lizards, birds, eggs, marsupials and rodents (Cunha andNascimento 1978, Marques andSazima 2004). We observed immediate fleeing or inflating of the gular region.

Spilotes sulphureus sulphureus (Wagler, 1824) Fig. 4Q
A very rare species. Recorded at Praia do Forte in August. Marques et al. (2011) reported the species from Imbassaí, captured in a restinga dry forest (SVL = 1970 mm). Inhabits primary forest, ombrophilous forest, relict wet forest and cabruca, foraging more on vegetation than on the ground (Martins and Oliveira 1998, Argôlo 2004, Lisboa et al. 2009, Loebmann and Haddad 2010. The species is arboreal and diurnal, feeding on birds and rodents (Beebe 1946, Marques et al. 2001. We observed defensive behaviors of elevating the anterior of the body, striking and biting.

Fig. 4R
A rare species of small size (min SVL = 130 mm, max SVL = 277 mm). Recorded at Busca Vida, Praia do Forte and Imbassaí. Occurred in restinga (n = 9) and ombrophil-ous forest (n = 2). The recorded specimens were active at 06:00 h and 19:00 h in May and June, both foraging on soil. The species is terrestrial, fossorial, diurnal and nocturnal (Vanzolini et al. 1980, Marques and Puorto 1998, Sawaya et al. 2008. It feeds on insects and mostly chilopods. Santos- Costa et al. (2006) report ovoposition of three eggs in February, June and September. We observed the ovoposition of two eggs in May (SVL = 277 mm). The tail of this species apparently breaks easily, since we observed two specimens from different sites with broken tails near the cloaca, not exceeding eight subcaudal scales.

Fig. 5A
A rare species of small size (min SVL = 175 mm, max SVL = 455 mm). Recorded at Instituto da Mata in July and October and the municipalities of Pojuca and Catu. Dias and Rocha (2014) reported this species from the restinga of Costa Azul. We found this species only in ombrophilous forest. It is nocturnal, terrestrial and semiarboreal and feeds exclusively on slugs (Maia-Carneiro et al. 2012). Females from southeastern Brazil exhibited vitellogenesis from July to December and oviductual eggs from August to February (Pizzatto et al. 2008). We observed seven ovarian follicles (3.6 -4.7 mm) in October and oviductual eggs (SVL = 480 mm; 6 eggs; 14.5 -16.6 mm) in April. If threatened, it flattens its head.

Imantodes cenchoa cenchoa (Linnaeus, 1758) Fig. 5B
A very rare species of medium size (min SVL = 455 mm, max SVL = 670 mm). Recorded at Instituto da Mata in November (see Marques et al. 2013) and the municipality of Simões Filho. Occurs in the ombrophilous forest in the study region. The species is nocturnal and we recorded it at 20:23 h. The animal was foraging on a bromeliad. Despite its arboreal habit, I. cenchoa often forages on ground level in forests and disturbed habitats (Cunha and Nascimento 1978). It feeds on frogs and lizards (Marques et al. 2001 Oliveira 1998, Garda et al. 2013). The species is nocturnal and we found it active from 18:05 h to 21:00 h. The snakes were observed on soil (n = 3), in anthropogenic environment (n = 1) and two meters high on vegetation (n = 1), always near ponds or lakes. It feeds on frogs, lizards and snakes (Cantor and Pizzatto 2008). Head flattening and cloacal discharge was observed.

Fig. 5E
An intermediately frequent species of medium size (min SVL = 173 mm, max SVL = 810 mm). Recorded at Praia do Forte, Reserva Sapiranga, Instituto da Mata and Costa Azul in March and from September to November. Occurs in ombrophilous forest (n = 10) and restinga (n = 3). The species can be abundant in urbanized and open environments and less frequent in forest habitats Oliveira 1998, França et al. 2012). We observed snakes in rain puddles, lake shores and rivers and in open areas.
Helicops angulatus is mainly nocturnal, active from 18:38 h to 21:08 h (n = 10) with a single record at 10:42. It feeds on fishes and frogs (Dixon and Soini 1986, Martins and Oliveira 1998, Ford and Ford 2002. We observed defensive behaviors, such as biting, struggling and occasional cloacal discharge.

Helicops leopardinus (Schlegel, 1837) Fig. 5F
A very rare species of medium size (min SVL = 109 mm, max SVL = 638 mm). Recorded at Arembepe, Imbassaí and Baixio in February and October. We found only specimens in the restinga, but vouchers originate also from forested habitats (n = 13). We observed snakes at edges of lakes and only during diurnal activity from 10:00 h to 16:12 h. However, Helicops leopardinus is known to be mainly nocturnal. It feeds on fishes and frogs (Strüssmann andSazima 1993, Ávila et al. 2006). We observed biting when handled.

Philodryas nattereri Steindachner, 1870
Fig. 5G A common species of medium size (min SVL = 295 mm, max SVL = 840 mm). Recorded at Arembepe, Guarajuba, Praia do Forte, Imbassaí, Massarandupió, Baixio, Barra do Itariri and Costa Azul during the year. We found it only in restinga (n = 38) during fieldwork, supported also by voucher specimens. It inhabits mainly open and semiarid habitats, also recorded at brejos de altitude (isolated fragments of humid forests surrounded by arid Caatinga) and disturbed habitats (Vanzolini et al. 1980, Pereira-Filho and Montingelli 2011. The species is diurnal and active from 08:30 h to 17:50 h. Philodryas nattereri is mainly terrestrial, as we observed it on the ground (n = 19), on the leaf litter (n = 14), among bush branches (n = 2), over herbaceous vegetation (n = 1) and on a branch that was four meters suspended (n = 1). It feeds on frogs, lizards, snakes, birds and mammals (Mesquita et al. 2011). We observed the elevating of its anterior body and immediate fleeing behavior, where it moved from the ground to a nearby suspended branch.

Fig. 5H
An intermediately frequent species of medium size (min SVL = 200 mm, max SVL = 770 mm). Recorded at Busca Vida, Arembepe, Instituto da Mata, Imbassaí, Massarandupió and Costa Azul in January and from July to December. We found it in restinga (n = 9) and disturbed areas of ombrophilous forest (n = 2). The species is diurnal and active from 07:56 h to 14:56 h. We observed snakes foraging on ground (n = 3), on leaf litter (n = 2) and in herbaceous vegetation (n = 1). The species is terrestrial and semiarboreal (Marques et al. 2001). It feeds on frogs, birds and rodents (Leite et al. 2009).
In the Caatinga domain, the species is mating from November to January ). We analyzed a fertile male (SVL = 551 mm) in November. We observed cloacal discharge and biting as a defensive behaviour.
Clelia plumbea (Wied, 1820) Fig. 5J A medium sized snake (min SVL = 335 mm, max SVL = 820 mm). Recorded in the municipality of Catu. Voucher specimens were from urban habitation. Inhabits mainly forests, disturbed, and open habitats (Zaher 1996, Argôlo 2004). The species is terrestrial and diurnal, feeding on lizards, snakes and mammals (Bernarde and Abe 2006). The reproductive cycle is continuous and the clutch size varies from four to 29 eggs (Gaiarsa et al. 2013).

Oxyrhopus petolarius digitalis Reuss, 1834
Fig. 5K A rare species of small size (min SVL = 176 mm, max SVL = 445 mm). Recorded at Praia do Forte and Instituto da Mata in July and November. We observed animals on pasture (n = 2) near ombrophilous forest remnants. A voucher specimen was obtained from the restinga. The species is terrestrial, forages on the ground (n = 2) and is nocturnal, detected at 21:10 h. Inhabits forest and disturbed habitats (Argôlo 2004, França et al. 2012. It feeds on lizards, birds, eggs and mammals and has a clutch size of two to 12 eggs (Gaiarsa et al. 2013). We observed winding when handled. A medium sized snake (min SVL = 323 mm, max SVL = 754 mm). Recorded at Guarajuba, Praia do Forte and Imbassaí (see Marques et al. 2012b). Occurs in restinga (n = 3) and anthropogenic habitat (n = 1). Dias and Rocha (2014) report the species from the restinga of Costa Azul. This species inhabits several types of open habitat including urban environments (Carvalho and Nogueira 1998, Sawaya et al. 2008, Valdujo et al. 2009). It feeds on lizards (Sawaya et al. 2008). Gaiarsa et al. (2013) reports clutch size of three to seven eggs. Fig. 5N A medium sized snake (min SVL = 214 mm, max SVL = 870 mm). Recorded at Busca Vida and Jauá, and in the municipalities of Mata de São João and Catu. Occurs in ombrophilous forests and restinga. Pseudoboa nigra also inhabits open and disturbed habitats and forests (Argôlo 2004, Gaiarsa et al. 2013). The species is terrestrial and nocturnal. It feeds on frogs, lizards and mammals (Vanzolini et al. 1980, Vitt and Vanglinder 1983, Orofino et al. 2010, França et al. 2012. Orofino et al. (2010) reported offspring with a SVL of 340 mm. We analyzed offspring with SVL of 214 mm and 274 mm.

Fig. 5O
A very rare species of medium size (SVL = 738 mm). Recorded at Instituto da Mata (see Marques et al. 2013) in May. We found it in ombrophilous forest remnants foraging at ground level on the leaf litter at 21:00 h. The species inhabits forests in Amazonia and Atlantic coast (Guedes et al. 2011). Nocturnal and arboreal, occasionally found on the ground. It feeds mainly on lizards (Martins and Oliveira 1998, Marques et al. 2001, Argôlo 2004. Clutch size varies from 3-12 eggs (Gaiarsa et al. 2013).
Thamnodynastes pallidus (Linnaeus, 1758) Fig. 5P A very rare species of small size (SVL = 363 mm). Recorded at the Instituto da Mata (see Marques et al. 2013). We observed it in ombrophilous forest remnants. We con-firm that this snake is nocturnal, as we found one foraging on the ground at 20:30 h. The female did not contain ovarian follicles in May. Marques et al. (2014) report ovarian follicles in July and eggs in September. Cunha and Nascimento (1978) report T. pallidus inhabiting the humid ground on forests and feeding on frogs and insect larvae. Fig. 5Q A medium sized snake (SVL = 655 mm). Recorded in the municipalities of Lauro de Freitas collected in 1994, without habitat information. Morato et al. (2011) record the species from the state of Sergipe at 175 km from our record. Inhabits forests, disturbed habitats and plantations Abe 1991, Argôlo 2004). This species is terrestrial and diurnal, but occasionally nocturnal, feeding mainly on snakes Abe 1991, Marques andSazima 2004).

Fig. 5R
Rare species of medium size (Min SVL = 136 mm, Max SVL = 450 mm). Recorded at Arembepe, Instituto da Mata, Imbassaí and Baixio in March and October. Occurs in ombrophilous forest (n = 8), restinga (n = 4), pasture and disturbed habitats (n = 1 each). Also recorded from urban environment (França et al. 2012 occurs in ombrophilous forest, but inhabits also cocoa plantations, pastures, swamps, and urbanized habitats (Pontes and Rocha 2008). One specimen was obtained from an industrial complex. In the state of Bahia, it occurs in the area north of Belmonte, being replaced by the nominate taxon Erythrolamprus p. poecilogyrus farther south (Fernandes 2006). The species is diurnal and nocturnal, and feeds on frogs and lizards (Marques et al. 2001, Argôlo 2004, Sawaya et al. 2008. Populations from the Atlantic forest from northeastern Brazil reproduce throughout the year with a clutch size of 3-17 eggs (Pizzatto et al. 2008b). Fig. 6D A very rare species of small size (min SVL = 191 mm, max SVL = 494 mm). Recorded at the Instituto da Mata (see Marques et al. 2013) and the municipalities of Catu and Pojuca. We observed it in ombrophilous forest in February. Commonly found near ponds, forest edges, fields and disturbed habitats (Dixon and Soini 1986). The snake was foraging on the leaf litter at 10:26 h. The species is terrestrial, occasionally using the vegetation to rest (Martins and Oliveira 1998). It feeds on frogs, tadpoles and lizards (Cunha andNascimento 1978, Bernarde andAbe 2010). We detected a gravid female (SVL = 403 mm; 1 egg, 26 mm) in June. We observed dorsoventral flattening.

Erythrolamprus taeniogaster Jan, 1863 Fig. 6E
A very rare species of medium size (min SVL = 147 mm, max SVL = 583 mm). Recorded at Praia do Forte, Imbassaí, and the municipalities of Lauro de Freitas, Simões Filho, Camaçari and Catu. We observed it in restinga in June, with voucher specimens also from restinga (n = 5) and ombrophilous forest (n = 4). In northeastern and southeastern Brazil, the species has a coastal distribution (Fernandes et al. 2002). In southern Bahia, Argôlo (2004) reported the species from sites near flooded areas, dams and swamps. We observed the species on a dirt road near a flooded plain at 15:00 h in urban habitation. The species struggles when handled.

Erythrolamprus viridis viridis (Günther, 1862) Fig. 6F
A small sized snake (SVL = 382 mm). Recorded in the municipality of Mata de São João. Occurred in ombrophilous forest, but also in transition areas from Atlantic forests and agreste (hilly north south chain with hot and sub-humid climate) region (Dixon 1987). The species is terrestrial but forages also in vegetation (Vanzolini et al. 1980). It feeds mainly on frogs and lizards (Vitt 1983.

VIPERIDAE Oppel, 1811
Bothrops erythromelas Amaral, 1923 Fig. 6K A medium sized snake (adult SVL ca. 570 mm). Recorded from the municipalities of Camaçari and Lauro de Freitas (MZUFBA 499 and 1366, respectively). The vegetation of both municipalities varies from restinga to ombrophilous forest with enclaves of cerrado, but no precise information was available on either of the two specimens' habitats. This is a terrestrial and nocturnal species that inhabits arid and semiarid regions, from sandy and rocky areas to deciduous forests, mainly associated with the Caatinga and Cerrado (Campbell and Lamar 2004).

Bothrops leucurus Wagler, 1824 Fig. 6L
An intermediately frequent species of large size (min SVL = 115 mm, max SVL = 1600 mm). Recorded at Arembepe, Guarajuba, Itacimirim, Praia do Forte, Instituto da Mata, Imbassaí, Sauípe, Massarandupió, Baixio and Mangue Seco. We found it in restinga (n = 9), nearby human habitation and pastures (n = 1) from February to October. Bothrops leucurus is more active during nighttime (Argôlo 2004), but we observed it mainly active from 06:10 h to 08:15 h. One adult moved at 15:30 h and two juveniles at 20:00 h and 20:08 h. We observed it mostly on the ground (n = 5), moving over bushes, coiled on short palm trees and on anthropogenic material (n = 1 each). Juveniles feed on frogs and adults mostly on rodents (Argôlo 2004, Fagundes et al. 2009). When threatened, they coil their body, vibrate the tail and occasionally strike. Fig. 6M A small sized snake (SVL = 451 mm). This species is confirmed from north of Salvador by four specimens, three by Lira-da- Silva et al. (2003) near sea level in the municipalities of Camaçari and Dias D'Ávila and one voucher at the Institute Butantan, also from Camaçari (IBSP 959). Terrestrial species, inhabits savannas and higher plateaus in the Cerrado (Loebmann 2009, Campbell andLamar 2004) and Caatinga < 250 m asl. (Loebmann and Haddad 2010). Fig. 6N A rare species of medium size (min SVL = 255 mm, max SVL = 1268 mm). Recorded at Busca Vida, Imbassaí and Massarandupió on March, August and September. Voucher specimens were from ombrophilous forest (n = 9), probably from deforested areas (Bastos et al. 2005) and restinga (n = 4). We observed a specimen crossing a road at 20:00 h while others were found on soil and leaf litter. This species is terrestrial and nocturnal, inhabiting open habitats and feeding on rodents (Vanzolini et al. 1980, Graça Salomão et al. 1995, Tozetti and Martins 2013. When handled the squirting of cloacal discharge can reach 1.5 m and if threatened, they coil their body and vibrate the tail and rattle.

Fig. 6O
A single record exists from the region based on a skin of an animal captured on a farm in the municipality of Entre Rios in 1996. The species inhabits primary forests, riparian forest, lowlands and cocoa plantations in southern Bahia. Occasionally it is found in secondary forest and surroundings. The species is nocturnal and feeds on small rodents (Cunha andNascimento 1978, Argôlo 2004).

Discussion
The number of snake species from the northern coast of Bahia represents 38% of all species known from the state of Bahia (Hamdan and Lira-da-Silva 2012, Curcio et al. 2012, Fernandes and Hamdan 2014 and contains six families. Disregarding the mistaken record of Bothrops jararaca by Marques et al. (2011), we recorded 49 native species in the study region. The absence of 17 species during our fieldwork, confirmed only from older museum specimens without new voucher deposits over at least ten years, suggests they have either become extinct in the region or our search method was insufficient (time spent) or non-appropriate, since our larger effort were on restinga. Three of these missing species (D. corais corais, P. guerini and P. nigra) are known from restinga habitat, whereas other species relate mostly to wooded areas, such as D. dichrous, C. plumbea, E. aesculapii venustissimus, E. miliaris merremii, E. viridis viridis, X. rhabdocephalus rhabdocephalus, M. corallinus and L. muta rhombeata. These assumptions can only be confirmed with additional sampling efforts in the few forest remnants of the region. Three species require taxonomic comments. First, the pitviper Bothrops lutzi from our study region, which appears isolated and far distant from the next known populations of ca. 425 km distance in Petrolina, Pernambuco, which was erroneously report-ed as 325 km in Lira-da- Silva (2000). We measured three specimens from Camaçari and Dias d'Ávila with ventral scale counts from 153-156, confirming Campbell and Lamar's (2004) findings that these are the lowest scale counts known for this species, which usually ranges from 161-179 ventral scales. These low values are more typical for B. erythromelas which exhibits 160 or lower ventral scales, which raises the question of potential mis-identification or hybridization between both viperid species. However, the B. erythromelas from the same region north of Salvador (Camaçari and Lauro de Freitas) yielded species typical 19-19-17 dorsal scales, whereas the sympatric B. lutzi exhibited species specific 21/23-21-19 dorsal scales. Moreover, the color pattern of latter specimens from Camaçari and Dias D'Ávila resembles rather typical B. lutzi than B. erythromelas, including diffuse and narrow trapezoid dorsal blotches (approaching a quadrangle), a diffuse and posteriorly not broadening postorbital stripe beginning behind the eye, and labials mostly darkened or mottled in the dorsum ground color with a few white spots. In contrast, B. erythromelas has comparatively wider angled dorsal blotches (resembling triangles, either pointed or rounded at the top), a larger postorbital stripe beginning below the eye and broadening posteriorly, and labials mostly white with occasional few larger spots. Machado et al. (2014) reported shared haplotypes between the two pitvipers, suggesting an introgressive hybridization, where the two species come into contact in western Bahia. The low ventral counts for the B. lutzi from the municipality of Salvador region may be another example of hybridization/ introgression, but so far this is only suggested due to the low ventral scale count, which alternatively may represent a case of geographically isolated variation.
Furthermore, there are several enclaves of Cerrado type vegetation (Campo Cerrado, Campo-limpo-de-Cerrado) from Camaçari (Salvador region) 200 km north to Ribeiro do Pombal, Bahia (Ab'Saber 1977, IBGE 2004. These enclaves allow the unusual presence of "Cerrado-species" in this area, such as Dendropsophus nanus, D. rubicundulus, Hypsiboas raniceps and Dermatonotus muelleri (Xavier et al. 2015). These remnant patches of Cerrado vegetation propose a scenario of a previous grassland corridor for an expansion of B. lutzi from its western populations in the current Cerrado domain (approx. 450 km distant). Subsequent aridification and expansion of Caatinga domain interrupted the corridor (Ab'Saber 1977) with the populations north of Salvador becoming relictual during the late Pleistocene 0.63 -0.11 mya (Ab'Saber 1977, Prado and Gibbs 1993, Pennington et al. 2000, Werneck 2011, Machado et al. 2014, while the intervening dried area was colonized by B. erythromelas. Drastic urban sprawl and intense agricultural practices may have also extirpated local grassland populations of B. lutzi. This may relate to notes that B. lutzi was once common in the low and humid region of southern Rio Paraguҫu, Recôncavo da Bahia, close to Salvador (B. lutzi labelled as B. n. neuwiedi in Amaral 1925, see cit. in Silva and Rodrigues 2008). Possibly more relictual B. lutzi populations exist on higher elevated grassland within the Caatinga today. For example, the holotype of Bothrops neuwiedi bahiensis Amaral, 1925 (IBSP 3012) originates from Itiuba, Bahia, a town in the midst of mountains that might have yielded grassland suitable for B. lutzi, while it is surrounded by Caatinga, that likely provides habitat for B. erythromelas. Similarly, combined information suggests sympatry of these vipers in central Piauí (Cambell andLamar 2004, Benicio et al. 2015) or Guanambi Rodrigues 2008, Machado et al. 2014). The specimen from Itiuba is perceived as B. neuwiedi by Silva and Rodrigues (2008, and their figs 14A-B), but possibly represents a B. lutzi based on the authors' qualitative characters assigned to B. lutzi, such as diffuse, at least partialy, interspace blotches, white supralabial spots vertically directed on 4 th and posterior supralabials, white spotting more marked posterior the 3 rd supralabial, and lack of dorsal postcephalic stripe. In contrast, this specimen has also well defined dorsal blotches, a neuwiedi character acc. Silva and Rodrigues (2008). However, the first author of the latter publication co-authored new records of B. lutzi from Minas Gerais , that, together with records of an adult B. lutzi from São Desidério in Bahia or a juvenile of approximately 30 cm total length from Grande Sertão Veredas National Park, Minas Gerais (Figure 7), indicate that these neuwiedi-characters are often within the variation of B. lutzi, rendering it more polymorphic than previously described. The suggestion of B. lutzi at Itiuba is also biogeographically plausible, as this specimen is nearest to other B. lutzi, north at Petrolina, Paraíba (Silva and Rodrigues 2008) southeast at Camaçari and Dias d'Avila (Lira-da- Silva et al. 2003) and west at Ibiraba, Bahia (Machado et al. 2014).
Comments on the second species relate to the confusingly polymorphic species Erythrolamprus poecilogyrus divided into multiple subspecies based mainly on color pattern variation that shows also drastic changes throughout ontogeny. Offspring and juvenile pattern resembling two taxa, E. p. poecilogyrus and E. p. schotti, have been recorded by us from the study area. According to Dixon and Markezich (1992) the only key characters to distinguish E. p. poecilogyrus from E. p. schotti that do not change with ontogeny are: (1) dorsal transverse bands from head to posterior including the mid-body section in E. p. poecilogyrus versus changing to a dorsal line or interruption of transverse bands in E. p. schotti; and (2) dark to black edged cephalic shields in E. p. poecilogyrus versus relatively featureless cephalic shields in E. p. schotti. In addition, young (SVL< 350 mm) E. p. schotti from northeastern Brazil are characterized by widely spaced mid-dorsal marking.
In our study area, offspring and juvenile specimens were found displaying characters of color pattern of both taxa noted above, whereas other juveniles exhibited widely spaced dorsal markings and some specimens showed intermediate expression of dark-edging of cephalic shields. This pattern concurs with the polymorphic scenario evaluated by Fernandes (2006), where the pattern of young E. p. schotti predominates in northeastern Brazil, but with some juveniles exhibiting characters of E. p. poecilogyrus (from 16% with conspicuous transverse dorsal bands to 74% with well-defined ventral markings). In contrast stands the color pattern variation of adults, whereby E. p. schotti and E. p. poecilogyrus separate well at or near the Jequitinhonha River, ca. 16° latitude at Belmonte, Bahia (Fernandes 2006). From there north, only E. p. schotti adult color pattern occurs including our study region north of Salvador. This demarcation between these two taxa is consistent with a decrease in the number of ventral scales and maxillary teeth from E. p. poecilogyrus (south) to E. p. schotti (north) found by Dixon and and Markezich (1992). Hence, we conclude that the E. p. poecilogyrus-like color pattern in some offspring and juveniles from north of Salvador has been and still is part of their regional morphological variation, and is not related to true E. p. poecilogyrus, which does not occur in our study area. A first hypothesis states, that the E. p. poecilogyrus-like juvenile color pattern represents simply an ancestral trait, still occurring in small proportions (i.e. has not been lost through selection or genetic drift) in populations in northeastern Brazilian populations of E. p. schotti.
A second hypothesis includes a historical component, whereby the occasional occurrence of E. p. poecilogyrus color pattern characters in the E. p. schotti range reflect a late Pleistocene/Holocene scenario of expansion and contraction of both taxa's range, or a shifting of their contact and overlap zones, triggered by greater paleoclimatic and vegetation shifts (e.g. Ab 'Saber 1977, Xavier et al. 2015. Northeastern Brazil exhibits a mosaic of ecoregions and range limits with taxa and their co-adapted gene complexes associated to Atlantic Forest, Cerrado and Caatinga relicts (e.g., Freitas 2014). Similar to a scenario suggested for the pitvipers Bothrops lutzi and B. erythromelas above, a period with dry climate within the late Pleistocene 0.63 -0.11 mya, in particular during the Last Glacial Maximum 12000-18000 years ago (e.g. Ab'Saber 1977; Pennington et al. 2000), may have caused the expansion of E. p. schotti to current coastal stretches of northeastern Brazil as far south as Belmonte, Bahia. The occurrence of mainly E. p. schotti morphology along the coast of northeastern Brazil indicates, that this species of Cerrado and Caatinga (at least for the northern half of Brazil) has adapted to environmental conditions of the Atlantic Forest. This is possibly facilitated by genetic exchange with E. p. poecilogyrus (an Atlantic Forest species) during postglacial warming from their Pleistocene refuges of Atlantic Forests along the coast (Carnaval and Moritz 2008) or a northward introgression of E. p. poecilogyrus from its core area south of Belmonte. The few color pattern characters by E. p. poecilogyrus in the E. p. schotti range would be such imprints of historical introgression.
As a consequence, the area around Salvador exhibits a large polymorphism in E. poecilogyrus, with specimens of each parental form, and those with prominent or only a few intermediate characters, or a combination of distinctive characters from each species, similar as was described in hybrid zones and the occurrence of relictual in- trogressed morphological characters in North American watersnakes (Mebert 2008(Mebert , 2010. As Dixon and Markezich (1992) noted, character gradients between E. p. schotti and E. p. poecilogyrus in northeastern Brazil are less steep than in southeastern Brazil, which supports the aforementioned relatively recent, late Pleistocene and postglacial expansion/contraction events between these two taxa. However, the occurrence of E. p. poecilogyrus color pattern characters in more western dry areas of Minas Gerais, Bahia, and Piauí is rather in support of the first hypothesis in regards to color pattern (E. p. poecilogyrus is an ancestral trait persisting in E. p. schotti populations), which may not restrict gene flow between both taxa along the southern Bahia coast.
The third species in need of discussion relates to Erythrolamprus aesculapii ssp. Peters and Orejas-Miranda (1970) attribute five subspecies to E. aesculapii, with two possibilities for the state of Bahia (Curcio 2008): Erythrolamprus a. monozoa and E. a. venustissimus. Erythrolamprus aesculapii monozoa does not apply to our specimen according to inter-ring distances. Freitas (2014) recorded E. a. aesculapii also from Lauro de Freitas, but it is unlikely since it is allocated to the Amazon population of the species. Erythrolamprus a. aesculapii differs from E. a. venustissimus by a light head color with dark body bands and red rings of same size as black rings, whereas E. a. venustissimus presents a blackish head with light body bands and red rings twice the size of black rings (Peters and Orejas-Miranda 1970). Therefore, following coloration and geographic distribution, we considered E. a. venustissimus more likely to represent the regional E. aesculapii. The taxonomy of this species is confusing and its population's status should be reviewed soon (F. Curcio, pers. comm.).
When we compare the snake species richness of the northern coast of Bahia with other snake assemblages from studies with similar sampling effort and field days, the region exhibits a higher species richness (see Table 3). Guedes et al. (2014) gathered voucher specimens from the entire Caatinga domain, resulting in 112 species. The Serra do Mar region extends along the coast from the state of Rio de Janeiro to Santa Catarina, containing 74 snake species (Marques et al. 2001). Santos-Costa et al. (2015) reported 53 species from a large area within Amazon forest, while Argôlo (2004) obtained 61 species from the southern coastal region of Bahia, covering several types of habitats.
Stomach content in eleven species was found with lizards representing 40% of food items, amphibians 20%, mammals and snakes 15% each, fishes and invertebrates 5% each. Some of these natural history notes are complemented by references. For example, Mesquita et al. (2011) observed Philodryas nattereri to be more active during the warmest period of the day, as well as the ingestion of Ameivula ocellifera, which probably is a common prey since both species are abundant along the coastal region. In contrast, most other species recorded at mid-day were in vegetation or in water, likely due to cooler ambient temperatures. In regards to our night observations, the principal activity of Helicops angulatus at the margin of lakes and ponds related to the foraging and predation of Leptodactylus natalensis, confirming that this snake occasionally feeds on frogs (Martins and Oliveira 1998). Species belonging to the genus Chironius are diurnal predators specialized on frogs, confirmed by us through an unidentified tadpole (total lenght = 39.4 mm) that was ingested by a juvenile C. exoletus Table 3. Richness of snake species: comparison between the northern coast of Bahia and other regions of Brazil.

References
Richness Region Guedes et al. (2014) 112 Caatinga domain Marques et al. (2001) 74 Serra do Mar (SdM)-RJ-SP-PR-SC Argôlo (2004) 61 Atlantic forest of South of state of Bahia -BA Santos- Costa et al. (2015) 52 Floresta Nacional de Caxiuanã-PA This study 49 North coast of Bahia França and Braz (2013) 47 Parque Nacional Chapada dos Veadeiros- GO Loebmann and Haddad (2010) 44 Planalto do Parnaíba-CE Sawaya et al. (2008) 36 Estação Ecológica de Itirapina-SP Hartmann et al. (2009a) 27 Núcleo Santa Virgínia, SdM-SP Marques and Sazima (2004) 25 Estação Ecológica Juréia Martins-SP Miranda et al. (2012) 24 Parque Nacional dos Lençóis Maranhenses Hartmann et al. (2009b) 24 Núcleo Picinguaba, SdM-SP Strussmann and Sazima (1993) 22 Pantanal Mesquita et al. (2013) 22 Caatinga-CE (SVL = 190 mm) (Dixon et al. 1993, Pinto et al. 2008. However, the ingestion of a Hemidactylus mabouia by C. flavolineatus could also represent an opportunistic, likely nocturnal predation event. Alencar et al. (2012Alencar et al. ( , 2013 state that lizards are the most frequent food item of Oxyrhopus trigeminus, but Bernarde and Abe (2010) classified it as a nocturnal generalist. However, our survey and predation records show the species may also forage during day, as we found two diurnal lizard species in its diet, Ameivula ocellifera and Tropidurus hygomi. Alternatively, these lizards were found sleeping and preyed on during nocturnal foraging. Ophiophagy is also common in the genus Micrurus (Campbell and Lamar 2004), yet there was no previous record of M. ibiboboca ingesting Amerotyphlops brongersmianus. We also report the first cannibalism event of O. trigeminus in which a specimen (SVL = 390 mm) ingested a juvenile (SVL approximately 140 mm). The ingestion of rodents by vipers is widely reported Oliveira 1998, Argôlo 2004) and is confirmed herein with a rat (Rattus sp.) ingested by a juvenile Crotalus durissus cascavela (SVL = 275 mm) and two Rattus norvegicus ingested by adults of Bothrops leucurus. But we also detected the remnants of the lizard Tropidurus hispidus in the stomach of B. leucurus near the Instituto da Mata. Sibynomorphus neuwiedi feeds exclusively on slugs (Maia-Carneiro et al. 2012), which was also confirmed in this study, similar to the ingestion of fish by Helicops leopardinus (Ávila et al. 2006), for which we recorded the fish Geophagus brasiliensis as prey.

Aspects of snake conservation on northern coastal Bahia
The herpetofauna of the north coast of Bahia is mainly threatened by habitat loss and degradation, including the ongoing expansion of residential areas and hotel-resorts (Tinôco 2011 Another deleterious aspect to snake conservation is the negative human perception towards snakes. Religious beliefs and superstition transcend an ancient fear about these animals, combined with the lack of contemporary knowledge (Greene 1997, Shine and Bonnet 2000, Fernandes-Ferreira et al. 2011, which results in countless killing of snakes including non-venomous and mimetic species. The presence of snakes around human habitation often leads to the death of these animals due to the prejudiced perception and low education of residents (Moura et al. 2010). In this study, we observed animals with damaged heads and bodies, mainly viperids (Bothrops erythromelas, B. leucurus, B. lutzi), but also colubrids and dipsadids (Chironius flavolineatus, Erythrolamprus almadensis, Leptodeira annulata, Oxyrhopus trigeminus, O. petolarius, Philodryas olfersii, Spilotes pullatus and Micrurus ibiboboca). Furthermore, accidental and intentional road killing are additional factors that affect snake populations negatively (Weatherhead and Madsen 2009).
Twelve species were classified as common or intermediately frequent. Since snakes are difficult to sample (Sawaya et al. 2008, Gaston andFuller 2007), this emphasizes that common and easily sampled species are relevant to better understand the biology of snakes and to better contribute to the conservation of a region as a whole. This study attempts to report for the first time all known snake species from the northern coast of Bahia. This region exhibits a snake species diversity as high as the southern part of the state of Bahia, which likely is the consequence of its distinct mosaic nature of habitat related species assemblages, including species from three distinct morphoclimatic domains, Cerrado, Caatinga, and Tropical Atlantic domains (Xavier et al. 2015). The 17 undetected species during the three years of fieldwork might have become extinct in the region, or simply require more field effort to be detected. Nonetheless, it reflects the lack of knowledge about the status of the northern coast of Bahia. We recommend continuation of monitoring programs, especially those of the forest fragments and inner portions of the region. This will provide an updated overview on the suggested extinct or threatened species, as well as all other species recorded here for the first time and new records.