The snake fauna of Togo: systematics, distribution and biogeography, with remarks on selected taxonomic problems

Segniagbeto G. H., Trape J. F., David P., Ohler A., Dubois A. & Glitho I. A. 2011. — The snake fauna of Togo: systematics, distribution and biogeography, with remarks on selected taxonomic problems. Zoosystema 33 (3): 325–360. ABSTRACT We present here an annotated list of the 91 snake species currently recorded from Togo, West Africa. Seven species are here recorded for the first time from this country: Calabaria reinhardtii, Hapsidophrys lineatus, Lycophidion nigromaculatum, Philothamnus carinatus, Leptotyphlops cf. narirostris, Letheobia crossi and Typhlops lineolatus. Main morphological data of examined specimens are provided. Some taxonomical problems are pointed out and discussed. The distribution of these species is detailed. We also provide a short discussion on the snake trade in Togo.


INTRODUCTION
First investigations on snakes of Togo date back to the time of the German colonisation and Sternfeld (Werner 1898(Werner , 1899(Werner , 1902(Werner , 1929Sternfeld 1908aSternfeld , b, 1909. During the same period, the fi rst list of Togolose snakes was established by Sternfeld (1908b), who included 75 species. However, many taxa listed by Sternfeld are no longer valid or refer to taxa that inhabit other African regions. Since Sternfeld's work, no extensive review of the Togolese snake fauna had been undertaken. During the years 1930-1950, Loveridge (1939, 1940, 1944) mentioned a few species occurring in Togo in his revisions of some African snake genera.
In the 1970s, three Belgian fi eld missions investigated the reptile fauna of Togo (Hulselmans & Verheyen 1970;Hulselmans et al. , 1971 and allowed the publication of a second list of Togolese snakes on the basis of specimens collected. Roman (1984) published a list of the 97 snake species recorded in the "Pays de l'Entente", namely Benin, Burkina Faso, Ivory Coast, Niger, and Togo. Th is work recorded 42 species from Togo.
In 2002, the monograph of the national survey of the biological diversity mentioned 101 snake species from the country (Anonymous 2002). However, a critical analysis of this work shows that it includes taxonomic mistakes and, just like Sternfeld (1909), listed nominal species not known from Togo, species unknown from this part of Africa and species that have now been synonymised or renamed in recent taxonomic studies, such as that of Lenk et al. (1999). As a consequence, no recent overview of the Togolese snake fauna is available. In the present paper, we provide an up to date review of the snakes of Togo mainly based on voucher specimens.
On this basis, we undertook a detailed survey of the snake species occurring in Togo. Th is paper presents the fi rst synthesis of our results. It is essentially based on the examination of preserved specimens as well as on a critical analysis of the literature. Main morphological characters of exami ned specimens are summarized. Some taxonomical problems are pointed out. Th e distributions of these species are detailed. Th e present paper is a continuation of works undertaken by the fi rst author on the whole reptile fauna of Togo (Segniagbeto et al. 2007). However, despite our extensive fi eld work and literature analysis, this paper still represents a preliminary analysis; much remains to be discovered about the snake fauna of Togo.

GEOGRAPHY, CLIMATE AND ECOLOGY
Togo is a West African country bordering the Gulf of Guinea (Fig. 1). It is made up of a long strip of land located between a latitude of 6°-11°N and a longitude of 0°-2°E. Th e country stretches over 660 km from north to south. It is only 50 km wide along the coast, east-west. Its maximal width is 120 km between 7 and 8°N. Th e landscape is largely a gently undulating plain, with the exception of the Atakora range ("chaîne de l'Atakora"), which crosses the country in a northeast-southwest direction. Some summits peak over 900 m in the southern part of the range. Peneplains between 100 and 400 m above sea level are found in the northern, central and southern parts of the country. From these peneplains rise tablelands such as Dapaong plateau and Bombouaka plateau, made of cuesta landscapes which reach 500 m on their northern edge. Bassar plateau ("plateau de Bassar") is largely dominated by Voltaian sedimentary geologic beds of shale. Th e Kante hills is composed of shales of the structural beds of the Atakora range and, in its southernmost part, by the Akposso-Akebou plateau ("plateau d'Akposso-Akebou"). Th e main feature of Togolese hydrography is the basin of the Oti river and its tributaries (Keran, Kara, Assoukoko rivers and so on), which altogether cover nearly 45% of the northern part of the country. Other main hydrographic features are the Mono basin (21 300 km) in the center and south of the country, and the Zio and Haho basins in the South.
Two kinds of wind predominate in Togo: the dry, hot northeastern trade winds ("alizés") known as the Harmattan, and the wet and warm trade winds locally known as the Monsoon ("la Mousson"). Th e confl ict between these two kinds of wind induces the Intertropical Convergence Zone (ITCZ) or the Intertropical Front. Th e variation in its location along the year induces two main climatic regimes: the tropical climate in the North with one rainy and one dry seasons, and the Guinean climate in the South. Th is latter climate is characterized by two rainy seasons and two dry seasons, all unequal. Between these two main climates, a transitional area presents a single rainy season with a slight decrease in rainfall in August or September.
As a consequence of its location, the Togolese landscape consists, from south to north, of a succession of various ecosystems ranging from coastal grasslands to equatorial and wet tropical forests and ending in Sudan savannahs in the North. According to Ern (1979), the vegetation of Togo can be divided into fi ve ecological regions (Fig. 1). From north to south, the country is successively made up of ecological region I or zone of Sudan savannahs where leguminous plants of the family Mimosoideae (Acacia spp.) or Combretaceae (Termi nalia spp.,  Region IV corresponds to the southern part of the country. It is characterized by a wet tropical climate similar to the equatorial climate and was originally largely covered with true tropical wet forests or semi-deciduous forests. Region V is restricted to the littoral area. It is a strongly disturbed landscape of littoral bushes, halophilous or marshy grasslands and mangroves. Th is diversity in the vegetal ecosystems is highly favourable to a great diversity of animal species, especially the snakes, most species of which are confi ned to precise biotopes.

INVESTIGATIONS IN THE FIELD
During two years (from March 2006 to January 2008), we collected specimens of snakes in as many localities in Togo as possible, spread over all ecological areas of the country ( Fig. 1; Table 1). We put a special emphasis on collections in the ecological region IV, in which collection localities are the most numerous. We deposited cans or buckets half fi lled with a 4% formaline solution or 95% ethanol in villages. Local villagers were made aware of our research and regularly deposited in these containers snakes encountered during their farming activities. Regular visits to these villages were organized to retrieve the specimens. Th is proved a highly effi cient way of gathering large numbers of specimens. During these visits, information on the habitat and ecology of the species were recorded. Geographic coordinates of each locality were obtained by a GPS receiver.
Besides the assistance of villagers, GHS and JFT also collected during fi eld trips. Collections were made both by day and at night. Specimens collected alive were photographed in the laboratory. Specimens killed in the fi eld, especially venomous species, were immediately photographed. All specimens were deposited in the collections of the "Département de Zoologie et de Biologie animale" of the Lomé University (Togo). We also made mere visual observations when we met species easy to identify or if we considered that the species was already adequately present in our collections. Lastly, we visited some snake farms housing snakes for the pet trade.

INVESTIGATIONS IN THE LABORATORY
All specimens collected during our fi eld investigations were examined. We also had access to specimens deposited in the collections of some other museums. Th is study is based on a total of 917 snake specimens. In the results given below, we indicate collection numbers and localities of all specimens. Main morphological characters, depending on the snake family, were recorded in all specimens, especially head scales, number of dorsal scale rows, ventral and subcaudal scales and measurements.

REMARK
DISTRIBUTION. -Th is species is mostly known from the forested area of Togo but additional specimens were recently collected slightly farther north from Aledjo and other localities. Th e distribution of this taxon in Togo is probably wider. Previously it had been recorded from the country by Sternfeld (1908bSternfeld ( , 1909, Hulselmans & Verheyen (1970), Hulselmans et al. ( , 1971 and Roman (1984).
DISTRIBUTION. -Th is species is widespread throughout the country. It mainly occurs in savannahs but also in wooded areas. Other specimens not yet deposited in collection were obtained from Aledjo, Bassar and Huilehui. Th e species was mentioned by Werner (1898,1902) and Sternfeld (1908bSternfeld ( , 1909; from Missahohe, Kete-Kratchi, Bismarkburg, Atakpame, Sokode, and Mango). Hulselmans & Verheyen (1970),  and Leaché et al. (2006) established the range of this species in ER II, III and IV of the country.
DISTRIBUTION. -First record from Togo. Th is recently described species is quite common. It was observed in the regions of Fazao and Aledjo as well as in the South of the country. Th is species was long confused with Dasypeltis scabra. DISTRIBUTION. -Th is species occurs mostly in forested areas. In Togo it is known only from Yo. Its taxonomic status and its distribution are far from being clear.  collected a specimen at Ahoue-Houe, in the forest region, which may probably be referred to this species. DISTRIBUTION. -Th is species is widespread in Togo, in the South as well as in the North. Our northernmost specimen is from Aledjo but this species might be present at Bassar and Kara, and in ER I. Its occurrence in Togo had previously been mentioned by Sternfeld (1908bSternfeld ( , 1909, Hulselmans & Verheyen (1970), Roman (1984), Spawls & Branch (1995) and David & Ineich (1999).
DISTRIBUTION. -Th is species was recorded from Togo by Werner (1898), Sternfeld (1908bSternfeld ( , 1909 from the locality of Missahohe, and by Anonymous (2002). According to Chippaux (2006), the distribution of this species is at the subregional levels (Western and Central Africa). Rasmussen, 1993 MATERIAL EXAMINED. -No specimen was collected.
DISTRIBUTION. -Th is arboreal species is very common in ER IV and also quite frequently encountered in ER II, III and V. Th e confi rmed northern limit of its range is Aledjo but it might also occur in the region of the Kara, especially in forests of Djamdè and Sarakawa. Sternfeld (1909: 24)  DISTRIBUTION. -On the basis of available voucher specimens, this species is present throughout the country. Previously, it had been mentioned from Togo by Broadley (1971aBroadley ( , 1998, Harding & Welch (1980), Welch (1982), Roman (1984), Golay (1985), Golay et al. (1993), Spawls & Branch (1995) and David & Ineich (1999). Sternfeld (1908bSternfeld ( : 220, 1909

REMARK
Th is species presents two morphs in Togo, namely a "savannah morph" observed in Aledjo (T 192) area and a "forest morph" in ER IV. Th ey diff er by the dorsal yellow rings, which are more contrasting in specimens of the savannah morph. It has also been mentioned from Togo by Sternfeld (1908bSternfeld ( , 1909, Broadley (1968), Hulselmans & Verheyen (1970), Hulselmans et al. ( , 1971, Roman DISTRIBUTION. -Th is species is mostly a forestdweller. Our voucher specimens were obtained in ER IV. Sternfeld (1908bSternfeld ( , 1909 recorded this species from Missahohe. It has also been mentioned from Togo by Hughes (1976), Harding & Welch (1980), Roman (1984, Golay (1985), Golay et al. (1993), Spawls & Branch (1995) and David & Ineich (1999 Parker, 1933, with 170 to 174 ventral scales in Togo, might be the valid name for a western subspecies which would diff er by a higher number of ventrals from Aparallactus lunulatus nigricollaris Chabanaud, 1916. Th is latter taxon has 133 to 140 ventrals and is known from Congo to Uganda. Furthermore, West African specimens are larger than those from Eastern Africa.

Naja nigricollis
We compared our specimens of Aparallactus lunulatus from Togo with others from Central and East Africa (RCA: Seko, MNHN 1994.3234, MNHN 1996Soudan: Boma, MNHN 2001.0141, MNHN 2001Congo: MNHN 1916.263-264;Kenya: MNHN 1901.451, MNHN 1904. Specimens from West (Togo) and Central Africa (RCA and Congo) agree very well with the diagnosis of Aparallactus lunulatus (Peters, 1854), the type locality of which is Tete, Mozambique. Th ey diff er from specimens from Kenya and Sudan by 1) lack of black nuchal collar, 2) specimens from Kenya have a suture between the prefrontal and the 2nd supralabial, which is lacking in specimens from Sudan, and 3) specimens from Kenya have a stouter body than those from Sudan and West Africa. In contrast, variation in the number of ventral scales as compiled by de Witte & Laurent (1947) does not seem to be Segniagbeto G. H. et al. sound enough to regard these taxa as full species. It should be underlined that the descriptions provided by Chabanaud (1916) and Parker (1933) agree well with that of Aparallactus lunulatus (Peters, 1854).

Atractaspis irregularis
DISTRIBUTION. -All collect localities are in the forested area of the country. Th is species is still unknown out of the forests but it might be present in the southern part of ER II and in ER V. Previously it had been mentioned from Togo by Sternfeld (1908bSternfeld ( , 1909, Hulselmans & Verheyen (1970), , Harding & Welch (1980), Golay et al. (1993) and Chippaux (2006 DISTRIBUTION. -Th is species inhabits mostly forested areas and was recorded from ER IV. It might be present in the South of ER II. Previously it had been mentioned from Togo by Sternfeld (1908bSternfeld ( : 212, 1909 and Loveridge (1939). MORPHOLOGY. -TL from 375 to 448 mm; 19 MSR, keeled; 168-174 Ven, smooth; 85-97 Sc, paired; anal single; 8 supralabials; 9 infralabials; 1 preocular; 2 postoculars; temporal scale formula 1 or 2 + 2 + 2.

Gonionotophis klingi
DISTRIBUTION. -Th e type specimen of this species described by Matschie (1893) was collected at Bismarkburg, currently Yegue, in Adele. Th is species is largely a forest-dweller occurring in ER IV. It might be present in the southern part of ER II. It had also been recorded from forested areas by Matschie (1893), Werner (1902), Sternfeld (1908bSternfeld ( , 1909 from Atakpame, Loveridge (1939), Hulselmans et al. (1971) and Leaché et al. (2006).

REMARK
Specimen MRAC 29670 has the loreal fused with the preocular. Th is character was already mentioned by Roux-Estève (1969).
DISTRIBUTION.  mentioned the collection of a specimen of Hemirhagerrhis aff . nototaenia (locality not given) during the third Belgian zoological mission to Togo. Th ese authors stated that this specimen was to be described soon. To our best knowledge this specimen was never mentioned again. It was probably referrable to H. nototaenia. Th is species was also cited from Togo by Roman (1984

REMARKS
Th e genus is currently under revision by the senior author of this paper; the results will be presented elsewhere. Nevertheless, in Togo and West Africa, the characters of both Lamprophis fuliginosus and Lamprophis lineatus are quite constant. Both species can be easily separated by the condition of the contact between the upper part of the preocular and the frontal. In all (57) examined specimens of L. fuliginosus from throughout West Africa, the preocular is in broad contact with the frontal, whereas these scales are separated in all (12)  DISTRIBUTION. -Th is species, mainly nocturnal, has been observed in the forested area near water where it feeds on small amphibians and lizards. Th is snake inhabits forested areas and Guinean savannahs. It is present in ER II, III, IV, and V of the country. Previously, this species had been recorded from Togo by Matschie (1893), Werner (1898( , 1902( ), Sternfeld (1908b DISTRIBUTION. -Lamprophis olivaceus has been mentioned from Togo by Sternfeld (1908bSternfeld ( , 1909, from Missahohe. Chippaux (2006) also cited this species from Togo, probably on the basis of Sternfeld (1908bSternfeld ( : 213, 1909. It is mostly a forest dwelling species which is widespread in Western and Central Africa. MORPHOLOGY. -TL from 240 to 860 mm; 23-25 MSR, smooth; 198-216 Ven, smooth, 43-62 Sc, paired; anal single; loreal longer than high; 8 supralabials; 8-10 infralabials; 2, sometimes 1 preoculars; 2 postoculars; temporal scale formula 1 + 2 + 3. DISTRIBUTION. -Th is species is mainly a forest-dweller. Th e northern limit of its range is in the region of Aledjo where new specimens were recently collected. It might also be present in the region of Kara. Previously this species had been recorded from Togo by Matschie (1893), Sternfeld (1908bSternfeld ( , 1909, Hulselmans & Verheyen (1970), and Hulselmans et al. (1971). MORPHOLOGY. -TL from 168 to 325 mm; 17 MSR, smooth; 159-184 Ven, smooth; 37-47 Sc; 8-9 supralabials; 9-10 infralabials; 1 preocular; 2 postoculars; temporal scale formula 1 + 2 + 3.
DISTRIBUTION. -Th is species is mostly a savannahdweller but it also occurs in the forested area. It is present throughout the country. Previously it had been recorded from Togo by Hulselmans & Verheyen (1970), Roman (1984 DISTRIBUTION. -Th is species is mainly a forest-dweller. It had been recorded from Togo by Matschie (1893), Sternfeld (1908b), Loveridge (1944from Missahohe), and Hulselmans & Verheyen (1970). Leaché et al. (2006) also mentioned the occurrence of the species in the forest area between Ghana and Togo.

REMARK
In spite of the limited sample of the taxon acutus available to us, we can separate Psammophylax acutus (Günther, 1888) from Psammophylax togoensis (Matschie, 1893) both on the basis of constant morphological characters, such as the presence of ventrolateral stripes, and of their geographical distribution. Th e non-parametric Mann-Whitney U test, implemented in SPSS 16.0 to compare the diff erence between the two species in the "BL/TaL" variable, shows a signifi cant diff erence between the two samples (Mann-Whitney U test: 2 = 31.00; P = 0.03; P < 0.05). Based on these results, we follow Chirio & Lebreton (2007) in considering the taxon togoensis to be a valid species. Recently, Ramphiophis acutus was referred to the genus Psammophylax Fitzinger, 1843 by Kelly et al. (2008).
DISTRIBUTION. -Ramphiophis oxyrhynchus had previously been mentioned from Togo by Sternfeld (1908b), Hulselmans & Verheyen (1970), , 1971), and Roman (1984. Th is species is present in all ecological regions of the country but, on the basis of recently collected specimens, it seems to be more common in ER II along the Aledjo Mountain range.
DISTRIBUTION. -Th e sole specimen examined, from Agou at the limit between ER III and IV, does not allow us to defi ne the distribution of this species in Togo. Pending the collection of additional specimens, we consider it to be present in ER III and IV. Th e species was cited from Togo by Broadley (1994) from Agou, Atakpame, Ezime and Tohoun. DISTRIBUTION. -Additional specimens not yet deposited were obtained from Fazao and Aledjo (Aledjo Kadara). Th is species is thus present in ER II and IV but may also occur in other ecological regions of the country. Previously it had been mentioned from Togo by Matschie (1893), Sternfeld (1908bSternfeld ( , 1909, Hulselmans & Verheyen (1970) on the basis of specimens collected at Kolokope in ER III. Trape & Mané (2006b) provided a regional distribution of this species.
DISTRIBUTION. -First record from Togo. Th ese fi ve specimens were collected at Alédjo Kadara (ER II). According to Chirio & Lebreton (2007), this species of dry forests and wet savannahs is distributed in West and Central Africa. Additional collection would allow us to establish the distribution of the species in Togo. However, results obtained during the fi eld work suggest that the occurrence of this species in ER I, III and IV is likely. DISTRIBUTION. -Leptotyphlops sundevalli was mentioned from Togo by Sternfeld (1909: 8;as Glauconia sundevalli Jan, 1862) from specimens collected at Missahohe, and Hughes & Barry (1969). Sternfeld (1908bSternfeld ( , 1909 as Glauconia conjuncta Jan, 1861) also mentioned the presence in Togo of Leptotyphlops conjunctus (Jan, 1861), a valid species known from Southern Africa. Pending the re-examination of specimens obtained by Sternfeld, we cannot ascertain the identifi cation of this specimen, which may be Leptotyphlops bicolor, or L. sundevalli or even another taxon.
DISTRIBUTION. -Th is species is largely aquatic. It can often be seen hunting amphibians on the banks of streams in forested areas. Th is species, very common in ER II and IV, had previously been mentioned from forested areas of Togo by Werner (1902), Sternfeld (1908bfrom Missahohe and Atakpame), Hulselmans & Verheyen (1970), Hulselmans et al. ( , 1971 and Leaché et al. (2006).

Natriciteres olivacea (Peters
DISTRIBUTION. -Th e distribution of this species in Togo remains poorly known. Th e sole precise locality is Sodo Zion, in the Akposso. Sternfeld (1908bSternfeld ( , 1909 recorded it from Missahohe. Th is species might be present in ER III and V. Its occurrence in Togo was also mentioned in Anonymous (2002)  DISTRIBUTION. -Th is species is very common throughout the country, especially in the South. Th e Ewe people consider it to be the embodiment of a divinity. It is abundant around Lake Togo and in the Mono. Th is species is as common in the forested areas of ER IV as in the Guinean savannahs where it is the most frequently encountered. Visual observations were obtained in the South as well as in the North. Recorded morphological characters on collected specimens are identical with those of Sternfeld (1908bSternfeld ( , 1909, Hulselmans & Verheyen (1970), , 1971), Villiers (1975), Roman (1984, Chippaux (2006), Leaché et al. (2006)  DISTRIBUTION. -Although the species occurs throughout the country, it is more common in the forested area and in the South around the Mono and Lake Togo. It had previously been recorded from Togo by Sternfeld (1908bSternfeld ( , 1909, Werner (1898), ), Villiers (1975), Roman (1984, Chippaux (2006) (1893), Werner (1898, 1902), Sternfeld (1908b, Hulselmans & Verheyen (1970), Hulselmans et al. ( , 1971 and Leaché et al. (2006).

Family VIPERIDAE Oppel, 1811
Atheris chlorechis (Pel, 1851) ( Fig. 10 DISTRIBUTION. -Th is species is essentially a forestdweller. It has previously been mentioned from Togo by Sternfeld (1908bSternfeld ( , 1909, Harding & Welch (1980), Golay et al. (1993), David & Ineich (1999), Anonymous All specimens of these species new to the country were obtained in the forested area, with the exception of those of Letheobia crossi which were obtained in localities belonging to ER II. Additional collections in the forested parts of the country (ER III and IV) are necessary to improve our knowledge of the Togolese snake fauna. Th e survey of the Togolese snake fauna being still in its preliminary stages, the list will undoubtedly increase as more specimens are collected.

SPECIES PROBABLY PRESENT IN TOGO
On the basis of data in Villiers (1975), Roman (1980, Chippaux (2006)

BIOGEOGRAPHY OF TOGOLESE SNAKE SPECIES
Togo hosts a very diverse snake fauna. Th e diversity of habitats allowed Ern (1979) to recognize fi ve ecological regions from the South to the North. Th is ecological diversity is at the origin of the faunal richness. Some snake species are restricted to one particular ecological region, although most species are present throughout the country. As a rule, ecological region IV, which covers the forested part of the country, is the richest in snake diversity (Fig. 12). As a rule, forested areas represent a highly favourable habitat for snakes, a factor which explains their high diversity. However, many species inhabiting savannahs are also quite common in the forested regions. Th is abundance of savannah dwelling species could result from the transformation of Togolese forest ecosystems into savannahs due to the intense exploitation of tree species. Th is phenomenon could explain both the diversity of species and the abundance of specimens in ER IV.

Species of the Sudanese savannahs
Th is category includes those species present only in ecological region I. It includes Gongylophis muelleri, Bamanophis dorri, Hemirhagerrhis nototaenia, Aparallactus lunulatus, Naja katiensis, and so on. Besides these typical species, other more widespread taxa are also very abundant in this region, such as Psammophis lineatus, Elapsoidea semiannulata, Echis ocellatus, and so on.

Species of the Guinean savannahs
Th ese species, typical inhabitants of African Guinean savannahs, present a wide distribution over the whole of the ecological regions of the country. Th ey are also present in Sudanese savannahs and even frequently found in the forested region. Th is latter occurrence is one of the reasons of the high specifi c diversity of ER IV.

DISTRIBUTION OF THE SPECIES
Togo is located in the "Dahomey gap" as defi ned by Jenik (1994), Salzmann & Hoelzmann (2005 and Ullenbruch et al. (2010). Th is gap is regarded as a major biogeographical barrier which separates the fauna of the West African forest blocks from the fauna of Central African forests (Booth 1958;Schiøtz 1967;Hamilton 1976). However, the forest belt between Togo and Ghana which covers the Togo mounts lies in the middle of the Dahomey gap. Th is range, which peaks over 900 m asl, receives a substantial amount of precipitation. Th e vegetation of this area is composed of dense tropical semideciduous and montane forests. According to Leaché et al. (2006), this mountain range constitutes a virtual "island" surrounded by Guinean savannahs. It hosts many forest species which are thus isolated from the large tracts of tropical forests located on the western and eastern margins of the gap. Th is is especially noteworthy for amphibian species such as Conraua derooi, Hyperolius baumanni and H. torrentis. Th e survey of the Togolese herpetological fauna being still is in infancy, it is possible that some snake or reptile species are restricted to this montane forested area. As far as we know, most West African forest snake species have a wide range covering Sierra Leone, Guinea, southern Ivory Coast, and Ghana as well as Nigeria and Cameroon. So, if no snake species is really endemic to this mountain range, it nevertheless acts as a fi lter in the distribution of many other taxa. According to Villiers (1975), Lenk et al. (1999) and Chippaux (2006), the forested region between Togo and Ghana marks the eastern limit of the range of several species typical of the West African fauna: Dendroaspis viridis, Lycophidion nigromaculatum, Aparallactus lineatus, Aparallactus niger, Atheris chlorechis, and Bitis rhinoceros. In a similar way, the range marks the western limit of the range of largely Central African species: Dendroaspis jamesoni, Bitis gabonica and Atheris squamigera.
It is interesting to note that several typical savannah dwelling species such as Psammophis praeornatus, Psammophis lineatus, Meizodon coronatus, Scaphiophis albopunctatus, Prosymna meleagris, and so on, reach the coast in the regions of southern Togo, Benin and of south-eastern coastal Ghana, whereas this is not observed in Liberia, Ivory Coast, in the south-western coastal Ghana, in Nigeria and in Cameroon. Th e Dahomey gap makes possible the southwards projection towards the coast of northern savannahs at the level of Togo and Benin. As a result, these savannahs separate the West African forest into two blocks.

INFLUENCE OF THE FRAGMENTATION OF FORESTS ON THE DISTRIBUTION OF SPECIES
It is clear that the Dahomey gap allows species of the savannahs to reach the southern coastal areas of the country. However, the strong presence of these species in the forested ER IV is also due to the heavy fragmentation of forest ecosystems. It is well known that West African forests host a very large biodiversity (Myers et al. 2000). However, these ecosystems are nowadays heavily damaged by human activities. In Togo, 43.6% of forest ecosystems have been destroyed since 1990 (FAO 2006). Furthermore, according to this same source, between 2000 and 2005, Togo had the third highest rate of deforestation in the world, 4.5% per year. Th e forest region between Togo and Ghana is already severely disturbed if not destroyed. Much of its surface has been transformed into a cultivated area according to Rödel & Agyei (2003). Only some highly anthropically-impacted patches of forests remain in several localities of ER IV, such as Missahohe, Akposso-Akebou plateaus, Danyi mounts and Adele plateau. Th ese patches are the last evidence of recently fl ourishing forests. Th e immediate result of the destruction of forest ecosystems in Togo is the transformation of ER IV into an area of savannahs. Th is phenomenon is accompanied by the extension of the range of typical savannah-dwellers into this (formerly) forested region. Th e corollary of this extension is that strict forest dwelling species are in regression and are replaced by ubiquitous, wide-ranging species. Our collections from the tropical forests only produced a few specimens of Calabaria reinhardtii, Dasypeltis  Our observations on the impact of the deforestation on the amphibian fauna of Togo are even more alarming and are addressed elsewhere (Hillers et al. 2009). According to the current rate of the destruction of natural biotopes of Togo, it is mandatory to survey the herpetological fauna and the whole of the biodiversity of the country before it is too late.

IN THE INTERNATIONAL PET TRADE
Th e trade of the wildlife has fl ourished in the whole of West Africa since the colonial period. From 1965 onwards, reptiles have been increasingly exploited for the pet trade, with a consequent increase in specimens exported. According to Roe et al. (2002), the development of commercial aviation throughout the world was one of the factors that allowed the increase of the pet trade during the last two decades. European countries, the United States and Japan constitute the most important outlet for this business. In these countries, the utilisation of wild animals as exotic pets is increasing. In spite of the drafting of the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES) in March 1973, aimed at the regulation of the international trade of the wildlife, the number of exported specimens is ever increasing. Th e diversity in exported species is also increasing. In 1990, Togo exported a record number of 80 000 specimens of the Ball python (Python regius), more than any other country worldwide (Aff o 2001). Th is record led the CITES to impose to Togo an export quota for reptile species. Table 3 summarizes the trends in the trade of three selected species during the fi ve recent years. In considering the record fi gures of 1990 and the results of our own investigations for the last fi ve years, we note that the export of living reptile specimens is still signifi cant for these three species.
Th e collection of a large number of living specimens of many species may prove to be prejudicial to the long-term survival of the animals in the wild. What is the impact of these collects on Togolese populations? In spite of the collection of a large number of specimens of P. regius in this trade, this species is largely distributed and very abundant in all ecosystems of the country. On the other hand, C. reinhardtii will be very threatened by this exploitation because of the destruction of it habitats.

CONCLUSION
Th is present survey should be considered as preliminary. In this study, we added seven species to the snake fauna of Togo. No species is endemic to this country. However, we have no doubt that new taxa will be added to the list presented here, inasmuch as both collections in the fi eld and the examination of preserved specimens are still in progress. Nevertheless, these results already demonstrate that the snake fauna of Togo is rich and diverse. Among the 90 species considered in our study, a total of 80 species are based on voucher specimens, plus eight species mentioned in the literature. Th e present paper completes and updates the previous studies on the Togolese snake fauna published by Matschie (1893), Werner (1898, 1902), Sternfeld (1908b, Hulselmans & Verheyen (1970),   and Anonymous (2002). Th is latter publication listed a higher number of species than our own counts but this report included either taxa currently considered to be synonyms or others, which occur in Central, Eastern and Souther Africa, but are defi nitely not present in Western Africa. On the other hand, forest dwelling specimens of wide ranging species such as Dasypeltis fasciata, Psammophis sibilans, P. phillipsi, Mehelya guirali, Philothamnus heterodermus and Naja melanoleuca show interesting taxonomic variation. However, due to the high rate of destruction of forest ecosystems caused by the exploitation of wood out of any control by the relevant authorities, we may have concerns about the rapid disappearance of these forest habitats before the Togolese herpetofauna can be adequately surveyed. Besides the destruction of forest ecosystems, another potentially serious threat to the snake fauna is the international trade of wildlife. In Anonymous (2002) it was stated that "although the international trade of wildlife is regulated by the CITES convention, the collection in the wild of a high number of species or of specimens of the same species is harmful to the survival and the sustainable exploitation of the biological diversity in the wild".
An improved knowledge on the diversity, distribution and abundance of the venomous taxa might lead to improved medical treatments, for example by the production of specifi c antivenomous serums, and a decrease in human fatal cases in these areas.