a preliminary analysis of phylogenetic relationships and...

Amphibia-Reptilia 30 (2009): 273-282

A preliminary analysis of phylogenetic relationships andbiogeography of the dangerously venomous Carpet Vipers,

Echis (Squamata, Serpentes, Viperidae) based onmitochondrial DNA sequences

E. Nicholas Arnold1, Michael D. Robinson2, Salvador Carranza3,*

Abstract. Phylogenetic analysis of 1117 bp of mitochondrial DNA sequences (731 bp of cytochrome b and 386 bp of16S rRNA) indicate that Echis consists of four main clades: E. ocellatus, and the E. coloratus, E. pyramidum, and E.carinatus groups. In the E. coloratus group, E. coloratus itself shows substantial genetic divergence from E. omanensis,corroborating their separate species status. In the E. pyramidum clade, E. pyramidum from Egypt and E. leucogaster fromWest Africa are genetically very similar, even though samples are separated by 4000 km. South Arabian populations ofthe E. pyramidum group are much better differentiated from these and two species may be present, animals from Dhofar,southern Oman probably being referable to E. khosatzkii. In the E. carinatus group, specimens of E. carinatus sochurekiand E. multisquamatus are very similar in their DNA. The phylogeny indicates that the split between the main groups ofEchis was followed by separation of African and Arabian members of the E. pyramidum group, and of E. coloratus and E.omanensis. The last disjunction probably took place at the lowlands that run southwest of the North Oman mountains, whichare likely to have been intermittently covered by marine incursions; they also separate the E. pyramidum and E. carinatusgroups and several sister taxa of other reptiles. The E. carinatus group may have spread quite recently from North Oman intoits very extensive southwest Asian range, and there appears to have been similar expansion of E. pyramidum (including E.leucogaster) in North Africa. Both these events are likely to be associated with the marked climatic changes of the Pleistoceneor late Pliocene. Similar dramatic expansions have also recently occurred in three snake species in Iberia.

Keywords: biogeography, Echis, evolution, mitochondrial DNA, snakes, taxonomy, venom.

Introduction

Carpet vipers (Echis) are found across the semi-arid regions of west, northern and east Africa,west, south and east Arabia, parts of Iran andAfghanistan north to Uzbekistan, and in Pak-istan, India and Sri Lanka (fig. 1). They are of-ten abundant and in many areas are a commoncause of fatal snake bite in people (Gasperetti,1988; Spawls and Branch, 1995). It has beenshown that antivenoms raised against venomfrom a population of Echis in one area may be

1 - Department of Zoology, The Natural History Museum,London, Cromwell Road, SW7 5BD, London, UK

2 - Sultan Qaboos University, Department of Biology, Col-lege of Science, P.O. Box 36, Al-Khod, Muscat, Sul-tanate of Oman

3 - Institute of Evolutionary Biology (CSIC-UPF), PasseigMaritim de la Barceloneta 37-49, Barcelona, Spain*Corresponding author; e-mail:[email protected]

ineffective in treating bites elsewhere (Warrelland Arnett, 1976; Gillissen et al., 1994). Some-times this is because different species of Echisare likely to have been involved (Gillissen et al.,1994), for venom chemistry sometimes variesgreatly between them. Consequently, a good ap-preciation of the taxonomy of the genus is es-sential for efficient treatment.

For a long time, just two species of Echiswere recognised: E. coloratus (Günther, 1878)found in Arabia, Jordan, Israel and easternEgypt, and E. carinatus (Schneider, 1801),which was believed to occur over most of therange of the genus. In E. coloratus, populationsfrom Israel and west Jordan have recently beendescribed as a subspecies, E. c. terraesanctae(Babocsay, 2003) and populations from north-ern Oman given full species status as E. oma-nensis (Babocsay, 2004), both taxa being recog-nised on the basis of morphology. The system-atics of Echis carinatus in its broad sense has

© Koninklijke Brill NV, Leiden, 2009. Also available online - www.brill.nl/amre

274 E.N. Arnold, M.D. Robinson, S. Carranza

Figure 1. Distribution of main taxa of Carpet vipers (Echis). Star indicates distinctive population in north Algeria (see below).Maps based on Gasperetti (1988); Spawls and Branch (1995); Babocsay (2004); Geniez et al. (2004); Mazuch (2005); andTrape and Mané (2006).

been confused and unstable in recent decades.Within this assemblage, a number of taxa havebeen described on the basis of external fea-tures, including E. carinatus leakeyi (Stemm-ler and Sochurek, 1969) and E. c. aliaborri(Drewes and Sacherer, 1974) from Kenya. Echisc. sochureki (Stemmler, 1969) was describedfrom Pakistan, Iran and North Oman and E.c. astolae (Mertens, 1971) from Astola Islandnear Pakistan. Echis ocellatus (Stemmler, 1970)of the savannahs of West Africa was recog-nised as occurring from Senegal eastwards to atleast western Chad (Wüster et al., 1997), whilepopulations assigned to E. leucogaster (Roman,1972) are found in drier areas of West Africaand also as isolates in the western Sahara northat least to southern Algeria (Wüster et al., 1997).

A broad study of Echis anatomy, incorporat-ing features of the hemipenis and dorsal scaleshape (Arnold, 1980a), made it clear that E. car-inatus in its broad sense is morphologically dis-junct. The eastern forms (at that time assignedto E. c. carinatus in India and Sri Lanka, E.c. sochureki and E. c. astolae) are very dif-ferent in these aspects from the more westernones, the two assemblages approaching each

other in Oman. There are also differences inreproduction, eastern animals apparently beingviviparous while western ones are egg-laying(Smith, 1943; Minton, 1966; Stemmler, 1971;Spawls and Branch, 1995). Among the west-ern populations, the oldest available name isE. pyramidum (Geoffroy Saint-Hilaire and Ge-offroy Saint-Hilaire, 1827), which is based onEgyptian material. Morphology indicated thatnearly all Echis populations from Egypt toKenya belonging to the western group could beregarded as a single rather variable form, andthat E. leucogaster was very similar to these(Arnold, 1980a). On the other hand, E. ocel-latus appeared to be a separate species distin-guished by scale counts, pattern and hemipenialstructure. It was also noted that an animal fromBiskra, northern Algeria (Natural History Mu-seum, London: BMNH 1907.4.6.55) has a dis-tinctive hemipenis and may represent yet an-other taxon (fig. 1B).

Subsequent to this survey, Echis multisqua-matus Cherlin, 1981, was described from Turk-menistan and Echis was later the subject of anextensive revision (Cherlin, 1990; Cherlin andBorkin, 1990). In this, a total of twelve species

Phylogeny of Echis 275

and a further eight subspecies were recognised,seven of these taxa being new. For example, an-imals from southwest Arabia previously placedin E. pyramidum were assigned to E. khosatzkiiCherlin, 1990 and E. varia borkini Cherlin,1990. Of Cherlin’s new forms, E. multisqua-matus was shown to be close to E. carinatussochureki in its external morphology (Auffen-berg and Rehman, 1991). But many of Cherlin’sother radical taxonomic changes have neverbeen formally reassessed, and the systematics ofthe Echis carinatus complex has remained un-stable. Species-level distinction between E. car-inatus and the more western forms has gener-ally been accepted (see for instance Gasperetti,1988; Schätti and Gasperetti, 1994), but manyauthors use conflicting taxonomic arrangementselsewhere in the genus, especially in Africa.Thus, Largen and Rasmussen (1993), Spawlsand Branch (1995), and Mazuch (2005) all varyin the names they use for some populations,and the last two publications do not incorporatesome of Cherlin’s changes in the regions theycover. Regrettably, confusion in the systemat-ics of Echis has probably been greatest in themedical and toxinological literature (Wüster etal., 1997). For example, less than 50% of exper-imental venoms and animals involved in bitesthat were mentioned could be confidently as-signed to the currently more widely acceptedspecies (Wüster and McCarthy, 1996).

Because of these problems, the relationshipsof some of the more critical populations ofEchis are surveyed here using mitochondrialDNA sequences. So far, their use in the genushas been limited, although individuals assignedto six taxa have been investigated using frag-ments of cytochrome b (cytb) and 16S rRNA(Lenk et al., 2001). In the course of a contin-uing broader molecular study on the Arabianreptile fauna, tissue samples were collected bythe present authors from three taxa of Echis inOman during 2005. The DNA sequences pro-duced from these were combined with othernew ones from Mauritania, Niger and with thoseof Lenk et al. (2001) and Malhotra and Thorpe

(2000) in Genbank, and with another sequenceof E. ocellatus deposited there by W. Wüster.The results of these preliminary analyses wereused to assess aspects of the systematics andbiogeography of Echis.

Material and methods

A total of 17 specimens of the genus Echis were includedin this study, among them representatives of at least sevennominal species covering much of the geographical distrib-ution of the genus. One Cerastes cerastes (Linnaeus, 1758)and one C. vipera (Linnaeus, 1758) were used as outgroups.Specimen data are given in table 1.

Extraction of DNA and PCR amplification

The DNeasy tissue kit, Qiagen, was used to extract genomicDNA from the samples, following the manufacturers’ in-structions. Primers used in both amplification and sequenc-ing were L 14846F1 (5′-CAA CAT CTC AGC ATG ATGAAA CTT CG 3′; Lenk et al., 2001) and S2 (5′-TGG GATTGA TCG TAG GAT GGC GTA-3′) for the cytochrome b

(cytb) gene, and L2510 (5′-CGC CTG TTT ATC AAA AACAT-3′) and H3062 (5′-CCG GTT TGA ACT CAG ATC A-3′) for 16S rRNA (Lenk et al., 2001). The two gene frag-ments were amplified using PCR procedures described byCarranza et al. (1999, 2001) and processed with an ABI377 automated sequencer following the manufacturer’s pro-tocols.

Phylogenetic analyses

DNA sequences were aligned using ClustalX (Thompson etal., 1997) with default parameters (gap opening = 10; gapextension = 0.2). The cytb alignment included no gaps andno stop codons were observed when the sequences weretranslated into amino acids using the vertebrate mitochon-drial code, suggesting that all the cytb sequences analyzedwere functional. Only five gaps had to be postulated to un-ambiguously align all the 16S rRNA sequences; thereforeall the positions were included in the phylogenetic analy-ses.

Topological incongruence among partitions was testedusing the incongruence length difference (ILD) test (Michke-vich and Farris, 1981; Farris et al., 1994). In this, 10 000heuristic searches were carried out after removing all invari-able characters from the dataset (Cunningham, 1997). Totest for incongruence among data sets we also used a recip-rocal 70% bootstrap proportion (Mason-Gamer and Kel-logg, 1996) or a 95% Bayesian posterior probability thresh-old. Topological conflicts were considered significant if twodifferent relationships for the same set of taxa were bothsupported with bootstrap values > 70%, or posterior proba-bility values > 95%.


Table 1. Details of material and sequences used in the present study. The question mark indicates that the cytb and 16SrRNA sequences of the specimen of E. ocellatus by Lenk et al. (2001) are probably the result of a misidentification of an E.leucogaster specimen as an E. ocellatus (see fig. 2).

Taxa Locality Accession numbers Referencescytb/16S rRNA

Cerastes cerastes Erfoud (Morocco) AJ275703/AJ275755 Lenk et al. (2001)Cerastes vipera Djebil (Tunisia) AJ275705/AJ275757 Lenk et al. (2001)

Echis coloratus Wadi Rishrash (Egypt) AJ275708/AJ275760 Lenk et al. (2001)Echis omanensis-1 Jebel Akhdar (north Oman) EU642590/EU642581 E3026.8Echis omanensis-2 North of Tanuf (north Oman) EU642587/EU642578 E3026.9Echis omanensis-3 Wadi Bani Habib (north Oman) EU642588/EU642579 E3026.10Echis omanensis-4 Wadi Bani Khalid (north Oman) EU642589/EU642580 E23116.1Echis carinatus subsp. (Pakistan) AJ275706/AJ275758 Lenk et al. (2001)Echis multisquamatus (Turkmenistan) AJ275702/AJ275763 Lenk et al. (2001)Echis carinatus sochureki Manah (north Oman) EU642591/EU642582 E3026.2Echis ocellatus-1 West Africa AF292568/ Wüster, unpublishedEchis ocellatus-2 West Africa AF191579/ Malhotra and Thorpe (2000)Echis ocellatus-3 10 km N. of Tapoua (Niger) EU642592/FJ808740 SPM0050Echis sp. (Yemen) AJ275707/AJ275759 Lenk et al. (2001)Echis khozatskii NW. of Ayun pools (Dhofar, Oman) EU642584/EU642575 E3026.15Echis pyramidum (Egypt) AJ275709/AJ275761 Lenk et al. (2001)Echis ocellatus? (Mali) AJ275710/AJ275762 Lenk et al. (2001)Echis leucogaster-2 Ayoûn el’Atrous (Mauritania) EU642585/EU642576 E3026.3Echis leucogaster-3 Ayoûn el’Atrous (Mauritania) EU642586/EU642577 E3026.4

Three methods were employed for phylogenetic analysesof the two separate gene fragments and for those in whichthey were combined. These were: maximum-likelihood(ML), maximum-parsimony (MP) and Bayesian analysis.Modeltest (Posada and Crandall, 1998) was used to selectthe most appropriate model of sequence evolution for theML and Bayesian analyses under the Akaike InformationCriterion. This was the GTR model taking into accountgamma distribution and the number of invariable sites foreach independent partition (cytb and 16S) and for the com-bined data set (cytb + 16S).

Both ML and MP analyses were performed inPAUP*4.0b10 (Swofford, 1998) and included searches in-volving tree bisection and reconnection (TBR) branch swap-ping with 10 and 100 step-wise additions of taxa, respec-tively. In the MP analyses, gaps were included as a fifthstate and transitions and transversions were given the sameweight. Bayesian analyses were performed on MRBAYESv.3.1.2 (Huelsenbeck and Ronquist, 2001) and each parti-tion had its own model of sequence evolution and modelparameters (see above). Four incrementally heated Markovchains with the default heating values were used. All analy-ses started with randomly generated trees and ran for 2×106

generations in two independent runs with samplings at in-tervals of 100 generations that produced 20 000 trees. Afterverifying that stationarity had been reached, both in termsof likelihood scores and parameter estimation, the first 5000trees were discarded in both independent runs of the cytband 16S rRNA and the combined analyses and a major-ity rule consensus tree was generated from the remaining15 000 post-burnin trees. The frequency of any particularclade among the individual trees contributing to the con-sensus tree represents the posterior probability of that clade

(Huelsenbeck and Ronquist, 2001); only values equal orabove 95% were considered to indicate sufficient support(Wilcox et al., 2002).

Results

The two gene partitions used (cytb and 16S)were shown to be congruent using the ILD-test (P = 0.90), and independent analyses ofthe two gene partitions confirmed there wereno topological conflicts (Mason-Gamer andKellogg, 1996). Therefore, both mitochondrialfragments were combined for further analyses.Of the 1117 bp of the combined data set (731of cytb and 386 of 16S), 327 were variable and276 parsimony-informative.

The results of the phylogenetic analyses areshown in figure 2. ML, MP and Bayesian phy-logenetic analyses all produced trees with iden-tical topologies in which four well-supportedmain units are discernable. They are the E.ocellatus, E. pyramidum, E. coloratus, and E.carinatus groups which have uncorrected ge-netic divergences of about 12-18% from each


Figure 2. Relationships of Carpet vipers (Echis) based on 731 base pairs of cytochrome b and 386 of 16S rRNA mitochondrialgenes. Two other viperines, Cerastes cerastes and C. vipera are used as outgroups. Maximum likelihood (ML), maximumparsimony (MP) and Bayesian analyses all gave identical topologies. Figures close to nodes are: ML bootstrap value/MPbootstrap value/Bayesian posterior probability value (only values equal or higher than 0.95 are indicated with an asterisk“*”). The question mark indicates that the cytb and 16S rRNA sequences of the specimen of E. ocellatus by Lenk et al. 2001(see table 1) are probably the result of misidentification of an E. leucogaster.

other in the cytb fragment investigated (see ta-

ble 2). In the E. pyramidum group, there is a

quite deep divergence between African and Ara-

bian representatives (11-13%). In North Africa,

a specimen of E. pyramidum from Egypt is very

similar to two assigned to E. leucogaster from

Mauritania (5%). A further snake from Mali,

identified in GenBank as E. ocellatus (Lenk

et al., 2001), is also close to E. pyramidum

(1%) and E. leucogaster (5%) and very diver-

gent from the other E. ocellatus included in this

study (14-16%); it may consequently actually

also be an E. leucogaster. Overall divergence

in these four widely separated North African

snakes is only about 2-5%.

The two south Arabian members of the E.

pyramidum group, from Yemen and Dhofar in

south Oman, are sisters but show a divergence


Table 2. Uncorrected genetic distances for the cytochrome b gene fragment used in this study.

Species 1

1 E. sp. – 22 E. khosatzkii 0.10 – 33 E. ocellatus? 0.12 0.12 – 44 E. pyramidum 0.12 0.11 0.01 – 55 E. leucogaster-2 0.13 0.10 0.05 0.05 – 66 E. leucogaster-3 0.13 0.10 0.05 0.05 0.00 – 77 E. omanensis-2 0.17 0.15 0.15 0.14 0.16 0.16 – 88 E. omanensis-3 0.17 0.15 0.15 0.14 0.16 0.16 0.00 – 99 E. omanensis-4 0.16 0.14 0.15 0.14 0.16 0.16 0.01 0.01 – 1010 E. omanensis-1 0.16 0.14 0.15 0.14 0.16 0.16 0.01 0.01 0.01 – 1111 E. coloratus 0.17 0.15 0.16 0.16 0.16 0.16 0.10 0.10 0.09 0.09 – 1212 E. multisquamatus 0.18 0.16 0.16 0.16 0.16 0.16 0.16 0.16 0.16 0.16 0.16 – 1313 E. carinatus 0.18 0.16 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.16 0.02 – 14

sochureki14 E. carinatus subsp. 0.18 0.16 0.15 0.15 0.16 0.16 0.16 0.16 0.16 0.16 0.15 0.03 0.03 – 1515 E. ocellatus-1 0.18 0.16 0.16 0.15 0.16 0.16 0.17 0.17 0.17 0.17 0.17 0.16 0.16 0.16 – 1616 E. ocellatus-2 0.18 0.16 0.16 0.15 0.16 0.16 0.17 0.17 0.17 0.17 0.17 0.17 0.16 0.16 0.00 – 1717 E. ocellatus-3 0.15 0.15 0.14 0.13 0.12 0.12 0.17 0.17 0.17 0.18 0.17 0.18 0.16 0.18 0.06 0.06 – 18

18 Cerastes cerastes 0.19 0.18 0.18 0.18 0.19 0.19 0.18 0.18 0.18 0.18 0.18 0.21 0.20 0.21 0.20 0.20 0.18 – 1919 Cerastes vipera 0.20 0.18 0.21 0.21 0.20 0.20 0.18 0.18 0.18 0.18 0.17 0.20 0.19 0.20 0.21 0.21 0.21 0.13 –

of about 10%. In the E. coloratus clade, there isa deep dichotomy between the western E. col-oratus itself and E. omanensis of North Oman,with a divergence of about 9-10%. In the E. car-inatus group, animals assigned to E. multisqua-matus from Turkmenistan and E. c. sochurekifrom North Oman are very similar (about 2%)and a further specimen, apparently from Pak-istan shows a divergence of 3% from these.

Discussion

Systematics

The great divergence between samples of E.ocellatus and E. pyramidum considered to bevalid here indicates that they are not closelyrelated as has been previously suggested (Lenket al., 2001). As noted, the specimen assignedto E. ocellatus on which this relationship wasbased is more likely to be assignable to the E.pyramidum group (see fig. 2). It remains to beseen if DNA indicates this is also true for thegenerally similar Kenyan E. pyramidum leakeyi,with which, it is suggested on morphological

grounds, E. p. aliaborri should be synonymised(Mazuch, 2005).

The strong divergence of the two Arabianpopulations of the E. pyramidum group studiedhere, from African ones and from each other,indicates they represent at least one separatespecies. The name E. khosatzkii (Cherlin, 1990)may be available for the population in Dhofar,southern Oman (fig. 1B). The locality of theholotype of E. khosatskii, is ‘Hadhramaut’ (infact, somewhere on an itinerary from Mukallahon the coast to the Hadhramaut valley and backto the sea at Ash Shihr, fide Anderson, 1896).The paratype comes from Ghayl Ba Wazir in thesame general area of eastern Yemen, and a sim-ilar specimen has been reported 170 km to theeast at Sayhut (Schätti and Desvoignes, 1999).These specimens come from 300-500 km westof the Dhofar ones. They share with these a dis-tinctive range of dorsal patterns and, comparedwith most other Arabian animals of the E. pyra-midum group from further west, relatively highventral and subcaudal scale counts (East Yemenand Dhofar: 165-179 ventral scales and 39-46subcaudal scales in males (n = 6), and respec-tively 183-189, and 36-39 in females (n = 4).


Western Yemen and adjoining southwest SaudiArabia: usually 145-169 ventral scales and 29-37 subcaudal scales in males (n = 5) and re-spectively 158-170 and 29-38 in females (n =6) – E.N. Arnold, unpublished data; Gasperetti,1988; Schätti and Gasperetti, 1994). However,adults from eastern Yemen have narrow heads,whereas they are broad in the ones from Dho-far (Arnold, 1980a). The other genetically dis-tinct southern Arabian member of the E. pyra-midum group investigated here does not havea precise locality within Yemen (Lenk et al.,2001), but its genetic divergence makes it un-likely to be conspecific with the Dhofar popu-lation. It could conceivably be a representativeof the populations in western Yemen for whichthe name E. borkini (Cherlin, 1990) (type local-ity Lahej near Aden) may be available (E. sp.in fig. 1B). More DNA sampling of the E. pyra-midum group in Yemen and Oman is requiredbefore these problems of nomenclature can beresolved. However, it is quite possible that atleast two species of Echis could occur in south-west Arabia, for some other reptile taxa havetwo or more closely related allopatric or parap-atric species there distributed from west to east.These include spiny-tailed agamids (Uromastyxyemenensis and U. benti – Wilms and Schmitz,2007) and semaphore geckos (Pristurus carterigroup – Arnold, 1986a). In other cases more dis-tantly related congenerics occur allopatrically inthis region, for example in Hemidactylus andPtyodactylus geckos (Carranza and Arnold, un-published data).

The strong genetic divergence between Echiscoloratus and the recently distinguished E. oma-nensis corroborates the species status of theseforms. In contrast, similarity in DNA sequenceconfirms an assessment from morphology (Auf-fenberg and Rehman, 1991) that Echis mul-tisquamatus is conspecific with E. carinatussochureki. The genetic similarity of E. multi-squamatus and E. c. sochureki has already beennoted by Lenk et al. (2001).

Historical biogeography

Echis belongs to the Viperinae, a group forwhich an African origin has been postulated(Lenk et al., 2001). Causus, the sister groupof the Viperinae is found in Africa and, of thefive units that comprise the subfamily, Atheris,Bitis and Cerastes also occur there with thelast two extending into Arabia, an area that isgeologically part of that continent. The only unitof the Viperinae to occur predominantly outsideAfrica is made up of Vipera, Pseudocerastesand Eristicophis and mainly occurs in Europeand Asia. Echis itself has three of its main unitsentirely in the African-Arabian region and thefourth, the mainly Asian E. carinatus group, isalso represented there in North Oman (fig. 1D).Given the overall distribution pattern of taxawithin the Viperinae, and the topology of thephylogenetic tree from figure 2 in which the E.coloratus group is basal to all the other threegroups, it is most parsimonious to assume anAfrican-Arabian origin for Echis. This contrastswith an Asian genesis in the Irano-Turanianregion suggested by Cherlin (1990). But thathypothesis did not take the distribution of therelations of Echis into account and the putativephylogeny of the genus on which it was based(Cherlin, 1990, fig. 6) is different from the onepresented here.

The DNA phylogeny (fig. 2) suggests thatEchis spread and diversified relatively rapidlyto produce its main groups, while disjunctionof the E. pyramidum group around the RedSea, and the separation of E. coloratus andE. omanensis within Arabia, occurred ratherlater. The Red Sea is a site of disjunction formany other reptile clades, including groundgeckos (Stenodactylus), sand skinks (Scincus),lacertids of the Acanthodactylus scutellatus andMesalina brevirostris-M. rubropunctata groups,sand vipers (Cerastes) and awl-headed snakes(Lytorhynchus) (see Arnold, 1980b, 1986b,1987). Similarly, the disjunction between E.coloratus and E. omanensis in eastern Arabia,is matched by many other groups includingBunopus, Hemidactylus and Pristurus geckos


(Arnold, 1980a; Arnold and Carranza, unpub-lished DNA data), sand skinks (Scincus m.mitranus and S. mitranus muscatensis – Car-ranza et al., 2008), and lacertids (Acanthodacty-lus schmidti and A. blandfordii – Harris et al.,1998). Different degrees of genetic divergencebetween western and eastern representatives ofthese units suggest that there was not a sin-gle vicariance event but a series of temporaryinterruptions that affected different taxonomicgroups at different times. This possibility gainssome support from topography. The highlandregions of Dhofar and North Oman are sep-arated by a lowland area that stretches fromthe Arabian Gulf southeast to the Arabian Sea.A fault-line running through this area and exten-sive inland sabkhas (salt flats) suggest there mayhave been an intermittent shallow sea separat-ing the more elevated regions on either side of it(Gasperetti, 1988). As already noted, a numberof taxa, perhaps including Echis, have differentspecies distributed along the southern coastal ar-eas of southwest Arabia, indicating another re-gion where some vicariance took place. If thisreally applies to Echis, genetic distances (ta-ble 2) suggest separation of the two forms con-cerned might have been around the time whenthe ancestor of E. coloratus and E. omanensisspeciated.

Parsimony suggests E. carinatus may haveoriginated by vicariance in North Oman andsubsequently expanded eastwards, although thetopology of the phylogeny (fig. 2) indicatesthis may have involved more than one invasion,or alternatively an invasion and a later back-colonisation. Such movement between NorthOman and Iran and neighbouring areas mayhave taken place in other reptile groups. Hemi-dactylus persicus and Acanthodactylus blan-fordii appear to have moved from Oman intoIran (Arnold and Carranza, unpublished data)and this may also have happened in Asacccusgeckos (Arnold and Gardener, 1994). Move-ments in the opposite direction are likely to haveoccurred in the ancestor of Bunopus hajarensisand B. spatalurus, and in Ablepharus cf. pan-

nonicus and Pseudocerastes (Arnold and Gal-lagher, 1977).

Echis carinatus may have extended its rangecomparatively recently and rapidly in southwestAsia, as the North Oman and Turkemenistananimals investigated here are genetically simi-lar, even though they are separated by at least1200 km. It remains to see if spread into In-dia was also comparatively recent. Such rapidexpansion by snake species, once the edgeof a new inhabitable region is colonised, hasalso occurred elsewhere. For example, DNAphylogenies show that the Montpellier snake(Malpolon monspessulanus), Horseshoe whipsnake (Hemorrhois hippocrepis) and Westernfalse smooth snake (Macroprotodon brevis)have only reached the Iberian peninsula fromnorth Africa recently, but have spread exten-sively there (Carranza et al., 2004; Carranza andArnold, 2006).

Limited genetic divergence (table 2) suggeststhat quite rapid range extension has also oc-curred in Echis pyramidum (including E. leuco-gaster) in North Africa. This may have hap-pened after restriction by climatic change al-though, if so, the core area where E. pyramidumsurvived is not identifiable. There also appearsto have been some restriction after expansion, asthe range of E. pyramidum is very fragmented.These events are likely to have occurred duringthe well-substantiated climatic fluctuations thatcharacterised the Pleistocene and late Pliocene(Sarnthein, 1978; Schuster et al., 2006). Move-ment of E. carinatus in southwest Asia mayalso have been around the same general pe-riod.

Acknowledgements. We are grateful to Ali Alkiyumiand Salim Al-Saadi and the Department of Diwan affairs,Sultanate of Oman, for permission to collect tissue sam-ples in Oman in 2005 (Permit number: 08/2005). E.O.Z.Wade, J.M. Padial and J. Brito provided tissue samples,J. Roca essential technical assistance, and D. Donaire ac-tive help in the field. E.N. Arnold also thanks Sultan Qa-boos Bin Said for permitting him to join the Oman Floraand Fauna Survey 1977, during which the first specimens ofEchis from Dhofar were collected; and also M.D. Gallagherfor organising this and many other collecting excursions.


B.C. Groombridge and C.J. MacCarthy provided adviceon morphological aspects of this study. Work was fundedby grants from the Natural Environment Research Council(NERC) NER/B/S/2000/00714 and NER/A/S/2001/00511to E.N. Arnold and by project CGL2005-06876/BOS fromthe Ministerio de Educación y Ciencia, Spain and Grant2005SGR00045 (Grup de Recerca Emergent) from the Gen-eralitat de Catalunya. Salvador Carranza is supported by aRamón y Cajal contract from the Ministerio de Educación yCiencia, Spain.

References

Anderson, J. (1896): A Contribution to the Herpetology ofArabia with a Preliminary List of Reptiles and Batrachi-ans of Egypt. London, R. H. Porter.

Arnold, E.N. (1980a): The reptiles and amphibians of Dho-far, southern Arabia. Journal of Oman Studies, SpecialReport 2: 273-332.

Arnold, E.N. (1980b): A review of the lizard genus Sten-odactylus (Reptilia: Gekkonidae). Fauna of Saudi Ara-bia 2: 386-404.

Arnold, E.N. (1986a): New species of Semaphore Gecko(Pristurus: Gekkonidae) from Arabia and Socotra.Fauna of Saudi Arabia 8: 352-377.

Arnold, E.N. (1986b): A key and annotated checklist to thelizards and amphisbaenians of Arabia. Fauna of SaudiArabia 8: 385-435.

Arnold, E.N. (1987): Zoogeography of the reptiles and am-phibians of Arabia. In: Proceedings of the Symposiumon the Fauna and Zoogeography of the Middle East, p.245-256. Krupp, F., Schneider, W., Kinzelbach, R., Eds,Wiesbaden, Verlag.

Arnold, E.N., Gallagher, M.D. (1977): Reptiles and amphib-ians from the mountains of Oman: the scientific resultsof the Oman flora and fauna survey 1975. Journal ofOman Studies, Special Report 1: 59-80.

Arnold, E.N., Gardener, A.S. (1994): A review of the Mid-dle Eastern leaf-toed geckoes (Gekkonidae: Asaccus)with descriptions of two new species from Oman. Faunaof Saudi Arabia 14: 424-441.

Auffenberg, W., Rehman, H. (1991): Studies on Pakistanreptiles. Part 1. The genus Echis (Viperidae). Bull. Flor.Mus. Nat. Hist. Biol. Sci. 35: 263-314.

Babocsay, G. (2003). Geographic variation in Echis col-oratus (Viperidae, Ophidia) in the Levant with the de-scription of a new subspecies. Zool. Middle East 29: 13-32.

Babocsay, G. (2004): A new species of saw-scaled viper ofthe Echis coloratus complex (Ophidia: Viperidae) fromOman, Eastern Arabia. System. Biodivers. 1: 503-514.

Carranza, S., Arnold, E.N. (2006): Systematics, biogeog-raphy, and evolution of Hemidactylus geckos (Reptilia:Gekkonidae) elucidated using mitochondrial DNA se-quences. Mol. Phylogenet. Evol. 38: 531-545.

Carranza, S., Arnold, E.N., Thomas, R.H., Mateo, J.A.,López-Jurado, L.F. (1999): Status of the extinct giantlacertid lizard Gallotia simonyi simonyi (Reptilia: Lac-ertidae) assessed using mtDNA sequences from museumspecimens. Herpetol. J. 9: 83-86.

Carranza, S., Arnold, E.N., Mateo, J.A., López-Jurado, L.F.(2001): Parallel gigantism and complex colonizationpatterns in the Cape Verde scincid lizards Mabuya andMacroscincus (Reptilia: Scincidae) revealed by mito-chondrial DNA sequences. Proc. R. Soc. Lond. Ser. B.268: 1595-1603.

Carranza, S., Arnold, E.N., Wade, E., Fahd, S. (2004):Phylogeography of the false smooth snakes, Macro-protodon (Serpentes, Colubridae): mitochondrial DNAsequences show European populations arrived recentlyfrom Northwest Africa. Mol. Phylogenet. Evol. 33: 523-532.

Carranza, S., Arnold, E.N., Geniez, P., Roca, J., Mateo, J.A.(2008): Radiation, multiple dispersal and parallelism inMoroccan skinks, Chalcides and Sphenops (Squamata:Scincidae), with comments on Scincus, Scincopus andthe age of the Sahara Desert. Mol. Phylogenet. Evol. 46:1071-1094.

Cherlin, V.A. (1990): Taxonomic review of the snake genusEchis (Viperidae). II. An analysis of taxonomy and de-scriptions of new forms. In: Reptiles of Mountain andArid Territories: Systematics and Distribution, Proceed-ings of the Zoological Institute, Vol. 207, p. 193-223.Borkin, L.J., Ed., Leningrad, USSR Academy of Sci-ences.

Cherlin, V.A., Borkin, L.J. (1990): Taxonomic revision ofthe snake genus Echis. In: Reptiles of Mountain and AridTerritories: Systematics and Distribution, Proceedings ofthe Zoological Institute, Vol. 207, p. 175-192. Borkin,L.J., Ed., Leningrad, USSR Academy of Sciences.

Cunningham, C.W. (1997): Is congruence between datapartitions a reliable predictor of phylogenetic accuracy?Empirically testing an iterative procedure for choosingamong phylogenetic methods. Syst. Biol. 46: 464-478.

Drewes, R.C., Sacherer, J.M. (1974): A new population ofcarpet vipers Echis carinatus from northern Kenya. J.East Afr. Nat. Hist. Soc. 145: 1-7.

Farris, J.S., Kallersjo, M., Kluge, A.G., Bult, C. (1994):Testing significance of incongruence. Cladistics 10: 315-319.

Gasperetti, J. (1988): Snakes of Arabia. Fauna of SaudiArabia 9: 169-450.

Geniez, P., Mateo, J.A., Geniez, M., Pether, J. (2004): TheAmphibians and Reptiles of Western Sahara. An Atlasand Field Guide. Frankfurt am Main, Chimaira.

Geoffroy Saint-Hilaire, E., Geoffroy Saint-Hilaire, I.(1827): Descriptions des reptiles qui se trouvent enEgypt. In: Description de l’Egypte: Histoire Naturelle,p. 121-160. Audouin, V., Ed., Paris.

Gillissen, A., Theakston, R.D.G., Barth, J., May, B., Krieg,M., Warrell, D.A. (1994): Neurotoxicity, haemostaticdisturbances and haemolytic anaemia after a bite by aTunisian saw-scaled or carpet viper (Echis ‘pyramidum’complex): Failure of antivenom treatment. Toxicon 32:937-944.

Günther, A.C. (1878): On reptiles from Midian collected byMajor Burton. Proc. Sci. Meetings Zool. Soc. London.(1878/4): 977-978.


Harris, D.J., Arnold, E.N., Thomas, R.H. (1998): Relation-ships of lacertid lizards (Reptilia: Lacertidae) estimatedfrom mitochondrial DNA sequences and morphology.Proc. R. Soc. Ser. B. 265: 1939-1948.

Huelsenbeck, J.P., Ronquist, F. (2001): MRBAYES:Bayesian inference of phylogeny. Bioinformatics. 17:754-755.

Largen, M.J., Rasmussen, J.B. (1993): A catalogue of thesnakes of Ethiopia (Reptilia: Serpentes), including iden-tification keys. Trop. Zool. 6: 313-434.

Lenk, P., Kalyabina, S., Wink, M., Joger, U. (2001): Evo-lutionary relationships among the true vipers (Reptilia:Viperidae) inferred from mitochondrial DNA sequences.Mol. Phylogenet. Evol. 19: 94-104.

Malhotra, A., Thorpe, R.S. (2000): A phylogeny of theTrimeresurus group of pit vipers: new evidence froma mitochondrial gene tree. Mol. Phylogenet. Evol. 16:199-211.

Mason-Gamer, R.J., Kellogg, E.A. (1996): Testing for phy-logenetic conflict among molecular data sets in the tribeTriticeae (Gramineae). Syst. Biol. 45: 524-545.

Mazuch, T. (2005): Taxonomy and biology of the viperEchis pyramidum leakeyi in Kenya. Ajva tera fórum2005: 64-71.

Mertens, R. (1971): Die Amphibien und ReptilienWest-Pakistans, 1. Nachtrag. Stuttgarter Beiträge zurNaturkunde 216: 1-4.

Michkevich, M.F., Farris, J.S. (1981): The implications ofcongruence in Menidia. Syst. Zool. 30: 351-370.

Minton, S.A. (1966): A contribution to the herpetology ofWest Pakistan. Bull. Amer. Mus. Nat. Hist. 134: 27-184.

Posada, D., Crandall, K. (1998): MODELTEST: testing themodel of DNA substitution. Bioinformatics 14: 817-818.

Roman, B. (1972): Deux sous-espèces de la vipère Echiscarinatus (Schneider) dans les térritoires de Haute-Voltaet du Niger: Echis ocellatus Stemmler – Echis carinatusleucogaster n. ssp. Notes et Documents Voltaiques 5: 1-13.

Sarnthein, M. (1978): Sand deserts during glacial maximumand climatic optimum. Nature 272: 43-46.

Schätti, B., Gasperetti, J. (1994): A contribution to the her-petofauna of Southwest Arabia. Fauna of Saudi Arabia14: 348-423.

Schätti, B., Desvoignes, A. (1999): Herpetofauna of South-ern Yemen and the Sokotra Archipelago. Geneva, Mu-seum d’Histoire Naturelle.

Schneider, J.G. (1801): Historiae amphibiorum naturalis etliterariae, Vol. 2. Jena.

Schuster, M., Duringer, P., Ghienne, J.F., Vignaud, P., Mack-aye, H.T., Likius, A., Brunet, M. (2006): The age of theSahara desert. Science 316: 821.

Smith, M.A. (1943): The Fauna of British India, Ceylonand Burma, Including the Whole of the Indo-ChineseSub-Region. Reptilia and Amphibia. Vol. 3: Serpentes.London, Taylor and Francis.

Spawls, S., Branch, B. (1995): The Dangerous Snakes ofAfrica. London, Blanford.

Stemmler, O. (1969): Die Sandrasselotter aus Pakistan:Echis carinatus sochureki subsp. nov. Aquaterra 6: 118-125.

Stemmler, O. (1970): Die sandrasselotter aus Westafrika:Echis carinatus ocellatus subsp. nov. (Serpentes, Viperi-dae). Revue Suisse de Zoologie 77: 273-282.

Stemmler, O. (1971): Zur Haltung von Echis carinatusleakeyi, der Kenya-Sandrasselotter. Zoologischer Garten(N. F.) 40: 200-210.

Stemmler, O., Sochurek, E. (1969): Die Sandrasselotter vonKenya: Echis carinatus leakeyi subsp. nov. Aquaterra. 6:89-94.

Swofford, D.L. (1998): PAUP*: Phylogenetic Analysis Us-ing Parsimony (and Other Methods), v.4.0. Sunderland,MA, Sinauer Associates.

Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmourgin,F., Higgins, D.G. (1997): The clustalX windows inter-face: flexible strategies for multiple sequence alignmentaided by quality analysis tools. Nuc. Acid. Res. 24:4876-4882.

Trape, J.F., Mané, Y. (2006): Guide des Serpents d’AfriqueOccidentale – Savane et Désert. Paris, IRD Éditions.

Warrell, D.A., Arnett, C. (1976): The importance of bitesby the saw-scaled or carpet viper (Echis carinatus).Epidemiological studies in Nigeria and a review of theworld literature. Acta Tropica 33: 307-341.

Wilcox, T.P., Derrick, J.Z., Heath, T.A., Hillis, D.M. (2002):Phylogenetic relationships of the dwarf boas and com-parison of Bayesian and bootstrap measures of phyloge-netic support. Mol. Phylogenet. Evol. 25: 361-371.

Wilms, T.M., Schmitz, A. (2007): A new polytypic speciesof the genus Uromastyx Merrem 1820 (Reptilia: Squa-mata: Agamidae: Leiolepidinae) of southwestern Ara-bia. Zootaxa 1394: 1-23.

Wüster, W., Golay, P., Warrell, D.A. (1997): Synopsis ofrecent developments in venomous snake systematics.Toxicon 35: 319-340.

Wüster, W., McCarthy, C.J. (1996): Venomous snake sys-tematics: Implications for snake bite treatment and toxi-nology. In: Emenomings and Their Treatments, p. 13-23.Bon, E.C., Bon, M., Eds, Lyon, Fondation Mérieux.

Received: April 25, 2008. Accepted: January 26, 2009.