Revision of the Middle-Eastern dwarf-snakes

16 downloads 0 Views 1MB Size Report
Key words. Eirenis coronelloides, systematics, taxonomy, Middle East, Fertile Crescent. ... Compared to viperids and the larger colubrids, these snakes have at- tracted little ... to BOULENGER'S coronella, but did not use it as a synonym. ...... The few specimens from Iraq are from the vicinity ofthe Tigris River in cen- tral and ...
Revision of the Middle-Eastern dwarf-snakes commonly assigned to Eirenis coronella (Colubridae) By Naomi Sivan and Yehudah L. Werner

Abstract. The systematics of the Middle-Eastern dwarf-snake Eirenis coronella has caused confusion in the past, but since 1978 this name has been used for a small dwarf-snake with 15 dorsal scale rows, dorsally coloured beige with dark transverse bands. Its distribution: Sinai in the west to southwestern Iran in the east; southern Turkey in the north to Saudi Arabia in the south. Two juveniles from Jordan in the Herpetology Collection of the Hebrew University of Jerusalem, having a distinct "crown" on the head and a dark ventral stripe, stimulated the question whether E. coronella may comprise more than one taxon. To answer this question, we examined snakes from the whole distribution area (n = 110) and analysed the results for males and females separately by principal coordinate analysis (PCO). In both sexes the PCO divided the snakes into two main groups, given species status: 1- Eirenis coronella, comprising all specimens from Sinai, Israel and western Saudi Arabia, and some from Jordan, Iraq and Syria. The Sinai population was given subspecies status, based on differences in scale counts and body proportions. The Western Saudi Arabia population kept its sub-species status, based on its colour pattern. 2- E. coronelloides, distinguished from E. coronella by dark crown, ventral stripe or both; and comprising all specimens from Turkey and the remaining specimens from Jordan, Iraq and Syria. A scatter-diagram of ventrals against sub-caudals separates the taxa almost diagnostically. While E. coronella occupy the northern extremes of the Arabian desert, the range of E. coronelloides is more northern and follows the "Fertile Crescent". Kurzfassung. Der Name Eirenis coronella wurde seit 1978 fur eine kleine Zwergnatter mit 15 Dorsalschuppenreihen und beiger Grundfarbe mit dunklen dorsalen Querbandern verwendet. Ihre Verbreitung: Yom Sinai im Westen zum siidwestlichen Iran im Osten und von der siidlichen Tiirkei im Norden bis nach Saudi-Arabien im Siiden. Zwei Jungtiere aus Jordanien mit einer deutlichen "Krone" auf dem Kopf und einem dunklen Ventralstreifen warfen die Frage auf, ob E. coronella vielleicht mehr als ein Taxon umfasst. Urn diese Frage zu beantworten, untersuchten wir Schlangen aus dem ganzen Ausbreitungsgebiet (n = 110) und analysierten die Resultate durch PCO ('Principal Coordinate analysis'), und zwar getrennt fur Mannchen und Weibchen. Die PCO teilte die Schlangen in zwei Gruppen, die Spezies-Status erhalten: 1- Eirenis coronella, darunter aile Exemplare vom Sinai, aus Israel und westlichem Saudi Arabien, wobei einige Tiere aus Jordanien, Irak und Syrien stammen. Die Exemplare vom Sinai erhielten den Status einer Subspezies. 2- E. coronelloides, unterschieden von E. coronella durch das Vorhandensein einer Krone, Ventralstreifen, oder beides; dazu gehoren aile Exemplare aus der Tiirkei zusammen mit den restlichen Exemplaren aus Jordanien, Irak und Syrien. Ein Streudiagram von Ventralia gegen Subcaudalia ergab eine beinahe diagnostische Trennung der Taxa. Key words. Eirenis coronelloides, systematics, taxonomy, Middle East, Fertile Crescent.

Zoology in the Middle East 28, 2003: 39-59. ISSN 0939-7140 © Kasparek Verlag, Heidelberg

40

Zoology in the Middle East 28, 2003

Introduction The genus Eirenis, commonly referred to as "dwarf snakes", is distributed in Asia and southeastern Europe (GASPERETTI 1988). Its centre of distribution is probably Anatolia, from where 11 species have been described (SCHMIDTLER & LANZA 1990, SCHMIDTLER & EISELT 1991, SCHMIDTLER 1993, 1997). The snakes are small to medium-sized and non-aggressive (Eirene in Greek means calmness or peace, and is also the name of the Goddess of Peace in Greek mythology). Compared to viperids and the larger colubrids, these snakes have attracted little research, and remain poorly known. A snake species, Calamaria coronella, was described by SCHLEGEL (1837) without locality data. The only characters which were not based on obscure comparison with another species (probably Coronella austriacai were: dorsal scale-rows 15, temporals three, basic colour grey-yellow to brown with 50 narrow black bands, and a black collar. JAN (1862a, 1862b,1863, 1865, 1866) described and illustrated three species which all have a dark collar and dark dorsal bands or blotches: Eirenis fasciatus Jan from Tiberias, with light head and belly (JAN 1862b; 1866, pl. 5, fig. 2); Homalosoma coronelloides Jan from Morea, Syria, with dark crown and dark ventral stripe (JAN 1862a; 1865, pl. 3, fig. 5.); and Homalosoma coronella Schlegel (JAN 1862a; 1865, pl. 4 , fig. 3) from Jerusalem, belly dotted with brown spots. According to JAN'S descriptions the genus Homalosoma and the species Eirenisfasciatus all have 15 dorsal scale-rows at midbody (JAN 1862a, 1863), although his H. coronella (today Eirenis lineomaculatai actually has 17. JAN'S illustration of E. fasciatus seems much closer to SCHLEGEL'S C. coronella than does his H. coronella, and this has caused much confusion. BOULENGER (1894) called a snake from Lebanon and northern Israel, with 17 scale rows, Contia coronella (Schlegel), continuing JAN'S misapplication of the name coronella. BARBOUR (1914) reported a series of snake specimens from Petra and Mt. Sinai, all with 15 scale rows, as Eirenis coronella, and noticed that one specimen resemble Jan's Homalosoma coronelloides, while the others resemble JAN's Eirenis fasciatus. He suggested all these three to be synonyms. SCHMIDT (1939) described a new species from Palestine with 17 dorsal scale rows and named it Eirenis lineomaculata. He mentioned the similarity of this snake to BOULENGER'S coronella, but did not use it as a synonym. SCHMIDTLER & SCHMIDTLER (1978) reviewed the genus Eirenis in Turkey, included a discussion concerning E. coronella and solved the problem of distinguishing lineomaculatus from SCHLEGEL'S coronella. Since then the name E. lineomaculatus is usually used for the snake with 17 dorsal scale-rows which resembles JAN'S illustration of H. coronella Schlegel. This is a Mediterranean (mesic) species, distributed from Turkey to central Israel (ARNOLD 1982, WERNER 1995: p. 33, fig. 6). The name E. coronella has usually been used for the snakes with 15 dorsal scale-rows described by JAN as E. fasciatus and H. coronelloides (WERNER 1995: p. 58, fig. 8). E. coronella, in this broad sense, is distributed from Sinai in the west to southwest Iran in the east; from Turkey in the north to Saudi Arabia in the south. In Israel it occurs only in desert areas, but in Jordan, Syria and Turkey it occurs also in the Mediterranean (mesic) region. Later the unambiguous alternative name, E. coronelloides, has been used for the same snake by some authors (MARX 1968, WERNER 1982), retaining the name E. coronella for E. lineomaculatus (HOOFIEN 1972, WERNER 1988). CHERNOV (1948) used the names Contia coronella for E. lineomaculatus, and C. coronelloides for E. coronella. DOTSENKO (1989) subdivided the genus Eirenis into two subgenera: Eirenis, containing

Reptilia

41

all species with 17 dorsals at midbody; and Collaria, comprising the species with 15 dorsals. Accordingly E. lineomaculata would belong in Eirenis, and E. coronella in Collaria. SCHMIDTLER & EISELT (1991) reject this subdivision. They believe that 17 dorsals is the original state of the genus, and that a reduction to 15 dorsals may have evolved many times. Systematic reports of E. coronella Schlegel usually compared geographically isolated populations with specimens from the rest of the distribution area. SCHMIDT (1939) described Eirenis coronella fraseri from Iraq (type locality Rutba). HAAS (1961) described Eirenis arabicus from Saudi Arabia (type locality Abqaiq), based on one specimen. ARNOLD (1982) questioned the status of E. c. fraseri, explained why E. arabicus is con specific with E. coronella, and described Eirenis coronella fennelli, from a mountain area in southwest Saudi Arabia. Other studies have reported E. coronella in surveys of selected geographic regions with little attention to systematic problems: Sinai (SCHMIDT & MARX 1956, WERNER 1982), Israel (WERNER 1988), Jordan and Israel (HAAS 1943), Jordan (Drsi et al. 1988, AMR et al. 1994, EL-ORAN et al. 1994), Turkey (BARAN 1978, 1982), Saudi Arabia (GASPERETTI 1988), Southwestern Asia (HAAS & WERNER 1969) and Iran (LATIF! 1991). This listing excludes mere checklists. Recently SHWA YAT (1998) performed a comprehensive study of the genus Eirenis in Jordan, using traditional morphological methods and kariotypes, protein electrophoresis and hemipenis morphology. Within Eirenis coronella she noticed differences in scale counts, body proportions and absolute size between specimens from different parts of Jordan, but drew no taxonomic conclusions. In the Herpetology Collection at the Hebrew University of Jerusalem the specimens from Israel, Sinai and Jordan listed as E. coronella or coronelloides (15 scale rows) vary in colouration. Those from Israel, Sinai and southern Jordan all have a beige background colour with dark dorsal cross-bands, dark collar and beige belly, similar to the snake depicted by JAN (1866, pI. 5, fig. 2) as E. fasciatus. The head has some irregular dark blotches. Young specimens resemble the adults. The specimens from northern Jordan look different. Two juveniles have a distinct dark "crown" on the head, continuous with the collar, and a dark ventral stripe, resembling JAN'S (1865, pI. 3, fig. 5) H. coronelloides. These differences between specimens from different geographical regions stimulated the question, whether this group of snakes, known collectively as E. coronella, may comprise more than one taxon. Therefore we present here a comprehensive study of this group of snakes from the whole distribution area, except Iran, using most available museum material.

Material and methods Material examined

We attempted to sample the entire geographic range of the group, by borrowing specimens from museums. But these small reclusive snakes are rare in collections. Most areas are poorly represented. Iran is represented by only one specimen. We examined all the material in the Hebrew University of Jerusalem (HUJ-R), some from Tel-Aviv University (TAU-R), and most available material from BMNH, FMNH, MCZ, NMP6V (National Museum, Prague), NMW, SMF, ZDEU (Ege University, Izmir) and ZSM (abbreviations follow LEVITON et al. 1985). The material (n = 110) originated from Sinai (7), Israel (48), Saudi Arabia (8), Jordan (13), Syria (8), Turkey (II), Iraq (9), and Iran (I). One of us (YLW) also cursorily examined specimens from Jordan in the Jordan University Museum, Amman. For most specimens the exact locality of collecting was available; but some could be attributed only to a geographical area (like "Horns-Palmyra").

42

Zoology in the Middle East 28, 2003

Eirenis coronella coronella (63): Israel (48): Negev: Har 'Arif HUJ-R 3292, 3849; Har Ramon: HUJ-R 3848; Wadi Nafha: HUJ-R 4501; En Avedat: HUJ-R 4635; Avedat: HUJ-R 8218; Shivta: HlJJ-R 3281; Hatrurim: TAU-R 908, 926; Har Hursha: TAlJ-R 194; Hamakhtesh Hagadol: HUJR 8217; Yeruham: HUJ-R 3392, MCZ 51663; Tel Yeruham HlJJ-R 3847, TAU-R 195; Zomet Hanegev: HUJ-R 8404,8405,8410; 19 km S of Be'er Sheva towards Yeruham: HUJ-R 8071; Dimona: HUJ-R 4094, 4096, 4426; Nahal Ze'elim: TAlJ-R 8992;1 km W of Be'erot 'Aroer: HUJ-R 8117; Near Beer Sheva, 3 km towards Hazerim HUJ-R 3836; Beer Sheva: HUJ-R 21265; Bc'er Sheva - Asluj: MCZ 56492; Sedom road: TAU-R 1953, 1954; Arad: HUJ-R 8066, 8362; 5 km from Arad towards Mezada: HlJJ-R 8394, TAU-R 10438, 11781, HUJ-R Arad road: HUJ-R 20958, 20959, 20960; Zomet Shoqet - Arad road: HlJJ-R 16956, 16957, 16958; 4 km from Zomet Shoqet: HUJ-R 8538; Be'er Tilla near Lahav: HUJ-R 8062; Nahal Kwir: TAlJ-R 12837; Judean Desert: Near Nabi Musa: HlJJ-R 8519, Mishor Adumim: HlJJ-R 16332, 16333; Wadi Qelt: HUJ-R 3147; Maale Adumim HlJJ-R 8714. - Jordan (5): Petra: MCZ 9682, 9684; Sisa-W. Musa: HUJ-R 3148; Sisa-Amman: HUJ-R 3138; Amman-Catarana: HlJJ-R 16968. Syria (5): Palmyra: NMPGV 34898; Homs-Palmyra: FMNH 19585, 19587, 19589, 19590. Iraq (5): Rutba area: FMNH 11362, 11365, 11366, 19576; T1 pipeline station: FMNH 68807. Eirenis coronella fenneli (6, all SW Saudi Arabia): Bani Mashor: BMNH 1982.226, 1982.227, 1982.228; Bani Sur: BMNH 1980.329, 1982.225; Wadi Fatina region: 1982.229. Eirenis coronella ibrahimi (7, Sinai, Egypt): Ar Rabbah valley, near Monastery of St. Catherine: HlJJ-R 8645; W. Klah, Near Monastery ofSt. Catherine: HlJJ-R 8788; Farsh el Fara: HUJR 8797; Monastery of St. Catherine area, ca. 5000': FMNH 72113; Monastery of St. Catherine area, Ar Rabbah, ca. 5000': FMNH 72114; Monastery ofSt. Catherine: FMNH 119652, 119653. Eirenis coronelloides (26): Jordan (8): Amman: BMNH 1927.816.6; HUJ-R 3145; Sisa: HlJJ-R 3146; Sisa-Amman: HUJ-R 16967; Near Amman: HUJ-R 16676; Mafraq: FMNH 19577, Ashaqif: FMNH 19582, N of Zarka: HUJ-R 3143. - Syria (3): Damascus: SMF 19293, Homs: FMNH 19593, Djerablis: NHMW 32025-3 (The locality "Djerablis, Armenia" is most probably Djerabulus near the Euphrate River in northem Syria (PIETSCHMANN, 1915). - Iraq (4): Hsitsche, Khabur: NHMW 32025/2, Karmeliter near Bagdad: 32025/1, Kirkuk: NHMW 20184, 32025/4. Turkey (11): Birecik-lJrfa: HUJ-R 21223 (former ZDElJ 122/1975/2), ZDElJ 79/1977/1, 79/1977/3, 122/1975/1, 122/1975/4, 122/1975/5; lJrfa: HUJ-R 21222 (former ZDElJ 116/1976/3), ZDEU 116/1976/2, 116/1976/4, ZMS 465/1976; SW Urfa: ZMS 103/1983. Eirenis sp., excluded: Qariya, Saudi Arabia: BMNH 1964.150, 1964.151. Mahar Birinji, Iran: MCZ 123158.

Characters Mensural characters: Body length: rostrum to anus (ra, WERNER 1971) (mm); - Tail length: distance from cloaca to tip of tail, if complete (mm); - Head width: greatest width of head (.1 mm); - Eye diameter: longest diameter, average of right and left eye (.1 mm); - Pileus length: length from snout to posterior tip of parietal, parallel to long axis of head (.1 mm); - Frontal length: greatest length of frontal (.1 mm); - Frontal width: greatest width of frontal (.1 mm); Snout length: distance from anterior tip of frontal to tip of snout (.1 mm); -Five-scale length: length of five consecutive scales of the mid-dorsal scale row, measured halfway between tip of snout and cloaca (.1 mm). All except the first three were measured under a stereo-microscope including an ocular micrometer. Head length could not be measured in a reliable way, and was therefore excluded. Tail length, head width and five-scale length were compared in terms of perera (percent of ra, WERNER 1971). For the other characters, other proportions were computed.

Reptilia

43

Meristic characters: Supralabials, sublabials, preoculars, and postoculars (for these four, the scales were counted on the right and left sides, and the average was used). - Collar width index: number of dark scales of three mid-dorsal rows of the neck collar. -- Dorsals at midbody: number of dorsal scale rows at half length between tip of snout and cloaca. - Dorsal bands: number of transverse dark bands between neck collar and cloaca (tail excluded). - Band index: On the middorsal scale-row, at half length between tip of snout and cloaca, we counted the number of light and dark scales over a distance covering three dorsal bands and three "interbands". The band index is the numerical proportion of dark to light scales. - Ventrals: Number of ventral scales according to Dowling's method (PETERS 1964, Fig. 7), excluding the anal plate. - Subcaudals: Both right and left row of subcaudal scales were counted, and the average was used. The tip of the tail was included. Qualitative characters: Presence or absence of a dark crown on head (the crown was defined as a continuous dark patch, connected to the neck collar and covering the frontal, supraoculars and about half the area of the parietals; see .IAN I865, pI. 3, fig. 5). Irregular dark blotches on the head were not defined as crown. Crown index is the percentage of specimens with a crown in a given sample. - Presence or absence of loreal. - Chin shield index: degree of separation of posterior chin shields (0 = full contact, I = contact only in anterior corner, 2 = separated by one scale, 3 = separated by more than one scale). - Ventral stripe index: extent of ventral dark stripe (0 = absent; 0.5 = a few dark spots in anterior part; I = weak, usually present only in anterior part; 2 = distinct; 3 = very prominent). Taxonomic approach and statistical analysis Since early studies of this group it has been clear that it has strong sexual dimorphism (SCHMIDT 1939). All statistical analyses were therefore done separately for males and females. In this group the number of ventrals and subcaudals varies geographically (ARNOLD 1982, SCHWA YAT 1998). We tried to apply a simple scatter diagram of subcaudals against ventrals to reveal any possible groupings, but the separation achieved was not complete. To strengthen the separation of the groups, we applied principal coordinates analysis (PCO) using the multivariate statistical package of Kovach Computing Services (MVSP). Seven characters, mensural, meristic and qualitative, were included: Tail length percra, ventral scales, subcaudals, eye/pileus, head width percra, ventral stripe index and crown (presence/absence). The two-dimensional plottings prepared by this procedure revealed separate groups of specimens. In order to clarify the taxonomic distinction of these groups, the means of characters of geographically neighbouring groups were compared by two-tailed t-test, using the Microsoft Excel program. For the taxa which emerged, dendrograms (prepared with MVSP) were used to show the relations among them, based on taxon averages of the characters.

Results In both sexes the PCO scatter diagrams divided the material into two main groups, left and right (Fig. l). The group to the left comprises all the specimens from Sinai, Israel and Arabia, and part of the specimens from Jordan, Iraq and Syria. Within this group the specimens from Sinai are clearly distinguished as a separate sub-group. The snakes from southwestern Arabia defined by ARNOLD (1982) as Eirenis coronellafenneli, cluster together on the PCO plots inside the main part of the left group. The group to the right comprises all the specimens from Turkey and the rest of the specimens from Jordan, Iraq and Syria. All the specimens with a crown fall into this group. They also tend to have a relatively high stripe index. The group includes also specimens without crown but with some degree of ventral stripe.

44

Zoology in the Middle East 28, 2003

Peecase scores(Standardized Euclidean)

• ...• ...

• flO

~



0.0

-3.~+,.•- -+----+---+---+:---:-:--c+:---+----+--_ Axis 1

peocase scores(StandardizedEuclidean)



Fig. 1. PCO-diagrams of male and female Eirenis (Based on 7 characters of individuals as explained in the text). Circle. Eirenis coronella coronella; rhomb, E. c.fennelli; square, E. c. ibrahimi; triangle, E. coronelloides.

.... . ... .:, .. .:

... e•

.:

.

..

..

In order to gain some insight into these PCO-subdivisions, we mapped the snakes, and an interesting picture emerged (Fig. 2). The left-hand PCO group is geographically continuous, ranging from southern Israel through southern Jordan, Syria and western Iraq, and in addition includes Sinai and south-west Arabia. The right-hand group ("crowned") is relatively northern, limited to a continuous stripe running from northern Jordan through Syria, Turkey, to north-eastern Iraq. The two groups are geographically very close, and they may even be sympatric in Jordan and Syria. Based on the significant biometrical differences (Tab. I) to be discussed below, we propose to give the two PCO-subdivisions species status. The left-hand group seems identical with SCHLEGEL'S Calamaria coronel/a and JAN'S Eirenis fasciatus, and we propose to retain the original name Eirenis coronel/a (Schlegel) for this species. The right-hand group ("crowned") seems identical with JAN'S Homalosoma coronelloides, and we propose for it the name Eirenis coronelloides (Jan). Tab. I shows average values for assorted characters in the two species and the significance of differences between them (separate for males and females). Not included are dorsal scale rows at midbody which, as expected, were 15 in all the snakes. Other characters which were nearly constant, and therefore excluded, were supralabials (7), sublabials (8), preoculars (I), postoculars (2) and loreal (usually present). Deviation from the norm, often on one side only, occurred occasionally throughout the geographical range, and seemed to lack systematic significance. As Tab. I shows, the two species, in both sexes, differ significantly (t-test) in tail perera (larger in E. coronellai, head-width perera (larger in E. coronelloidesi, number of ventrals and subcaudals (both larger in E. coronel/a), ventral stripe index and crown index (both larger in E. coronel/oides).

Reptilia

45

, ',

,,'-0

..... ·