SALVADOR CARRANZA 1* & EDWARD WADE 2. Abstract. Introduction

Transcription

1 Zootaxa : 1 24 (2004) Copyright 2004 Magnolia Press ISSN (print edition) ZOOTAXA ISSN (online edition) Taxonomic revision of Algero-Tunisian Pleurodeles (Caudata: Salamandridae) using molecular and morphological data. Revalidation of the taxon Pleurodeles nebulosus (Guichenot, 1850) SALVADOR CARRANZA 1* & EDWARD WADE 2 1. Department of Zoology, The Natural History Museum, London, SW7 5BD (salc@nhm.ac.uk) * Present address: Departament de Biologia Animal, Universitat de Barcelona, Av. Diagonal 645, E Barcelona, Spain (scarranza@ub.edu) 2. Middlesex University, Cat Hill, Barnet, Hertfordshire, EN4 8HT (e.wade@mdx.ac.uk) Abstract The taxonomic status of Algero-Tunisian Pleurodeles was reanalysed in the light of new molecular and morphological evidence. Mitochondrial DNA sequences (396 bp of the cytochrome b and 369 of the 12S rrna) and the results of the morphometric analysis, indicate that Algero-Tunisian P. poireti consists of two genetically and morphologically distinct forms. One restricted to the Edough Peninsula, and another one covering all the rest of its distribution in Algeria and Tunisia. The name P. poireti (Gervais, 1835) is restricted to the population of the Edough Peninsula, while P. nebulous (Guichenot, 1850) correctly applies to all other populations in the distribution. P. poireti originated approximately 4.2 Myr ago, probably as a result of the Edough Peninsula being a Pliocene fossil island, allowing both forms of Algero-Tunisian Pleurodeles to diverge both genetically and morphologically. Key words: Pleurodeles, Algeria, taxonomy, mitochondrial DNA, 12S rrna, cytochrome b, morphology, Pliocene fossil island Introduction The genus Pleurodeles currently consists of two species. P. waltl Michaelles and P. poireti (Gervais, 1835). Morphology and mitochondrial DNA sequences indicate that, among living forms, the sister taxon of Pleurodeles is Tylototriton from southeast Asia (Titus and Larson, 1995). Despite some fossils originally thought to be related to Pleurodeles dating back to the Upper Oligocene of Germany (Palaeopleurodeles Herre 1941), it is believed that the origin of Pleurodeles is much more recent, having split from its sister taxon (Tylototriton) during the Middle Miocene, some 10 Myr ago (see Carranza and Arnold, 2003 for a critical review of the Palaeontological data). This result is supported by the age of the Accepted by M. Wilkinson: 22 Mar. 2004; published: 13 Apr

2 earliest specimens assigned to Pleurodeles, which date from the Upper Miocene or Lower Pliocene of Spain (Sanchiz, 1977). Pleurodeles waltl is large (up to 300 mm in total length Pasteur, 1958), with usually 15 presacral vertebrae, a tubercular process on rib 3 (and often traces of this structure on ribs 1 and 2), and sharp rib tips that project through a row of glandular swellings on the flanks. P. waltl is common and widely distributed in the southern two thirds of the Iberian Peninsula and occurs with less abundance in northern Morocco, where it is found in the area delineated by Tangiers, Casablanca and Alhoceima (Bons and Geniez, 1996). The second Pleurodeles species, P. poireti, is smaller (up to about 230 mm in total length), with 14 presacral vertebrae, no obvious tubercular processes on the ribs, which are not sharp, and no glandular swellings on the flanks. It occurs in northern Tunisia and Algeria extending westwards as far as Oran. Nomenclatural history The first person to mention the presence of P. poireti in northern Algeria was Poiret (1789) in his Voyage en Barbarie. In there, he gives a brief description of this species under the name of Lacerta palustris. Gervais (1835), analysed material of Lacerta palustris from Oran, Algiers and Bône from Marly and Gerard and recognised it represented a new species distinct from L. palustris of Linné 1758, which he named Triton Poireti (=Pleurodeles poireti). However, he already noted the similarity that existed between Triton Poireti and the Ibero-Moroccan Pleurodeles Waltl, especially in the shape of the head and tail. He concluded that, although T. Poireti lacked the exposed long ribs characteristic of Pleurodeles Waltl, it was linked to it. Despite Gervais (1835) material being collected from widely separated localities, the extant type series of T. Poireti actually consists of two specimens, both from Bône (= Annaba) (MNHNP 4744 and MNHNP 4744A Thireau, 1986), within the general area to which we subsequently refer herein to as the Edough Peninsula (see Fig. 1). Bône, therefore should be considered the type locality (Thireau, 1986). Bonaparte (1839) assigned T. Poireti to the genus Glossoliga. Latter Guichenot (1850) synonymized T. Poireti with the Sardinian Euproctus Rusconii Gené, 1838 (=Euproctus platycephalus; Gravenhorst, 1829). In the same account, Guichenot (1850) described a new species, Triton nebulosus (=Pleurodeles nebulosus). According to Guichenot, T. nebulosus differed from T. Poireti in having longer tail, longer, wider and more depressed head, shorter teeth slightly curved and less prominent skin tubercles, which appear closer together. The type material of T. nebulosus is deposited at the MNHNP and includes a complete female specimen from Algiers collected by Guichenot himself (MNHNP 1442), and two other specimens, which were dissected about 1867 to produce a complete skeleton (MNHNP CD 8) and a skull (MNHNP CB 21) (Thireau, 1986). Gray (1850) placed the Algero-Tunisian newt in the genus Triton (T. Poireti) and synonymized T. nebulosus with T. Poireti. Euproctus Rusconii is listed in Gray (1850) as a synonym of the Sardinian Magnolia Press CARRANZA & WADE

3 endemic Euproctus platycephalus. Gervais (1853) showed that T. Poireti from Algeria and E. Rusconii from Sardinia were different species and argued in favour of Bonaparte s (1839) proposal of altering the generic name of Triton Poireti to Glossoliga. Guichenot s Triton nebulosus and E. Rusconii were both listed as synonyms of T. Poireti by Dumeril and Bibron (1954), who changed its generic name from Triton to Euproctus (as E. Poireti). Lataste (1881), found morphological differences between populations of Glossoliga Poireti. In recognition of such, he described a new morphological variant from Bône (Annaba) as a new species, Glossoliga Hagenmulleri, in honour of Dr. Hagenmüller, who collected all of the 32 specimens (26 males and 6 females) used in the original description (Lataste, 1881). But in fact, Gervais (1835) type material of T. Poireti originated from three localities in Algeria, one of which was Bône. Lataste (1881) was in error as to which of the three localities should apply Gervais (1835) T. Poireti. He decided that the specimens from d Alger in his account (= Algiers) should carry the name Poireti. This appears to be based on an illustration (Gervais, 1853; pl. 14 Fig. 9) of a dorsal view of a skull of a specimen clearly stated to be d Algérie i.e. Algeria, therefore too imprecise to be applicable. Moreover, the types that appear to have survived are from Bône (Thireau, 1986), which should therefore be considered most appropriate as the type locality of Triton Poireti (Thorn,1968; Frost, 1985; Thireau, 1986). Lataste s (1881) type series of Glossoliga Hagenmulleri comprises six specimens from Mount Edough, within the Edough Peninsula (northeast Algeria; see Fig. 1): BMNH (BMNH auct.). According to Lataste (1881), G. Hagenmulleri differed from G. Poireti in having the palatine teeth row forming a V; smaller size; head longer than broad; tongue small and less free behind; snout narrower; contour of jaws semi elliptical; limbs, especially fingers and toes, more slender; gular fold less pronounced; colour olive-brown above, greyish beneath with more or less distinct darker spots. These morphological differences between the two species of Algero-Tunisian newts were later corroborated by Boulenger (1882), who placed them in the genus Molge (M. poireti and M. hagenmulleri). After Boulenger (1882), several authors recognised both forms at the specific or subspecific levels (Camerano, 1885; Woltersttorff, 1905; Noble, 1924) but reassigned their generic name to Triton (Camerano, 1885), to Pleurodeles (Woltersttorff, 1905), and again to Triton (Noble, 1924). In a revision of Lataste s (1881) work, Pasteur (1958) synonymized Pleurodeles hagenmulleri (Lataste, 1881) with Pleurodeles poireti (Gervais, 1835). He analysed some specimens of P. poireti and P. hagenmulleri from the Muséum National d Histoire Naturelle, Paris (MNHNP) for most morphological characters used in the original description (Lataste, 1881), and those of Boulenger (1882) and Doumerge (1901). He concluded that, with the exception of MNHNP , which perfectly fitted Lataste s (1881) description of P. hagenmulleri, he could not clearly differentiate between P. poireti and P. hagenmulleri. As a result of his work, the recognised form of Algero-Tunisian Pleurodeles today is P. poireti. ZOOTAXA REVISION OF PLEURODELES 2004 Magnolia Press 3

4 In this paper, we revise the taxonomy of P. poireti in North Africa using morphology and cytochrome b (cytb) and 12S rrna (12S) mitochondrial sequences. Materials and Methods Relevant data for all the specimens used in the molecular and morphological analyses are presented in Table 1 and Fig. 1. Species used in the molecular study are a selection of some key sequences from a previous phylogeographic study of Pleurodeles published by Carranza and Arnold (2004). TABLE 1. Details of material and sequences used in the present study. Numbers under CODE refer to localities shown in Fig 1. All measurements are in mm. BMNH Natural History Museum, London; MNHNP Muséum National d'histoire Naturelle, Paris; AL Algeria; Cytoch. b Cytochrome b. SPECIES ACCESSION LOCALITY SEX SVL HL HW IOW TL SP3T 3T AL TL/ HL/ HW/ HW/ AL/ Cytoch. b / CODE NUMBER SVL SVL SVL HL SVL 12S rrna P. p o ir e t i 1 MNHNP 4744 (Syntype) 2 MNHNP 4744A (Syntype) Bône (AL) F Bône (AL) M BMNH BMNH BMNH BMNH BMNH BMNH Mount Edough (AL) Mount Edough (AL) Mount Edough (AL) Mount Edough (AL) Mount Edough (AL) Mount Edough (AL) M M M M M M BMNH Bône (AL) F BMNH Bône (AL) F AY222508/ AY BMNH Bône (AL) F AY222507/ AY BMNH Bône (AL) M BMNH Hippone (AL) 14 BMNH Hippone (AL) 15 BMNH Hippone (AL) 16 BMNH Hippone (AL) M M M M continued on the next page Magnolia Press CARRANZA & WADE

5 TABLE 1 (continued) SPECIES ACCESSION LOCALITY SEX SVL HL HW IOW TL SP3T 3T AL TL/ CODE NUMBER SVL HL/ SVL HW/ SVL HW/ HL AL/ SVL Cytoch. b / 12S rrna ZOOTAXA P. p o ir e t i 17 BMNH Hippone (AL) 18 BMNH Hippone (AL) M M BMNH Bône (AL) F BMNH Bône (AL) M BMNH BMNH Bône (AL) M Bône (AL) F P. nebulosus 23 MNHNP 1442 (syntype) 24 BMNH BMNH Algiers (AL) F Algiers (AL) M AY222511/ AY Algiers (AL) F BMNH 1859.? Algiers (AL) F BMNH Algiers (AL) M BMNH Algiers (AL) M BMNH Algiers (AL) F BMNH BMNH Larba (AL) F AY222504/ AY Larba (AL) M AY222510/ AY BMNH Teniet El Had (AL) 33 BMNH Teniet El Had (AL) F AY222544/ AY F AY222545/ AY BMNH Bejaia (AL) M AY222505/ AY BMNH BMNH Constantine (AL) Constantine (AL) M F AY222462/ AY Tabarca 38 Tabarca 39 Tabarca 40 Dam Bou- Heurtma 41 Dam Bou- Heurtma F F M F M DNA analysis 43 DNA analysis Tabarca Tabarca AY222518/ AY AY222519/ AY continued on the next page REVISION OF PLEURODELES 2004 Magnolia Press 5

6 TABLE 1 (continued) SPECIES ACCESSION LOCALITY SEX SVL HL HW IOW TL SP3T 3T AL TL/ CODE NUMBER SVL HL/ SVL HW/ SVL HW/ HL AL/ SVL Cytoch. b / 12S rrna P. nebulosus 44 DNA analysis 45 DNA analysis 46 DNA analysis 47 DNA analysis 48 DNA analysis 49 DNA analysis 50 DNA analysis 51 DNA analysis 52 DNA analysis 53 DNA analysis 54 DNA analysis 55 DNA analysis Tabarca 5Km of Nefza South of Tabarca South of Tabarca Jendouba Near Fernana Ain Draham Dam Bou- Heurtma Dam Bou- Heurtma Dam Bou- Heurtma Dam Bou- Heurtma Dam Bou- Heurtma AY222520/ AY AY222522/ AY AY222543/ AY AY222542/ AY AY222540/ AY AY222541/ AY AY222521/ AY AY222535/ AY AY222538/ AY AY222526/ AY AY22251/ AY AY222539/ AY Phylogenetic analysis In the present study we compared 396 bp of the cytochrome b (cytb) and 369 bp of the 12S rrna mitochondrial genes of representatives of P. poireti from almost all its distribution range, and some P. waltl from the Iberian Peninsula and Morocco. Salamandra s. morenica, S. s. terrestris and Mertensiella luschani were used as outgroups. Sequences were aligned using ClustalX (Thompson et al., 1997) with default parameters. No gaps were necessary to align the cytb sequences and two were included in order to align the three outgroup 12S sequences with the ingroup. Phylogenetic trees were inferred for the two genes together using maximum-likelihood (ML) with the General Time Reversible (GTR) model of sequence evolution, with gamma distributed rates (G) (selected by MOD- ELTEST, Posada and Crandall 1998). ML analyses were performed in PAUP* 4.0b10 (Swofford, 1998) and included heuristic searches involving tree bisection and reconnec Magnolia Press CARRANZA & WADE

7 tion (TBR) branch swapping with 100 random stepwise additions of taxa. Nodal support was assessed by bootstrap (Felsenstein, 1985) involving 1000 pseudo-replications. FIGURE 1. Map of North Africa showing localities of Pleurodeles used in the present study. See Table 1 and Fig. 5 for further details. The dashed line delimits the approximate distribution range of P. poireti. We refer to it in the text as the Edough Peninsula. REVISION OF PLEURODELES 2004 Magnolia Press 7 ZOOTAXA

8 FIGURE 2. Schematic drawing showing all eight morphometric measurements used in the present study. SVL snout-vent length; TL tail length; Sp3T right hind limb shank length; 3T right hand third toe length; AL length of the anterior right arm; HW head width; IOW interorbital width; HL head length. Morphometric analysis A total of 41 P. poireti were included in the analyses. Eight measurements were taken using a calliper to the nearest 0.01 mm (see Fig. 2): snout-vent length (SVL), head length Magnolia Press CARRANZA & WADE

9 (HL), head width (HW), interorbital width mesured between the bonny margins of the orbit (IOW), tail length, shank length (SP3T), third toe length (3T), right forearm length (AL). Some specimens of P. poireti were missing the gular fold (see Pasteur, 1881; Lataste, 1958), so for HL measurements we measured from the tip of the snout to the retroarticular process in the articulation of the mandible with the quadrate (see Fig. 2). Statistica 5.5 was used for statistical analysis. Normality of the data was determined using the Shapiro-Wilks W test implemented in Statistica 5.5. When not normal, data was log-transformed in order to meet better the assumption of normality. In order to assess differences in the vomerine teeth of both species (Lataste, 1881), warmed plasticine was placed in the mouth to the distance of the vomerine tooth bands. Pressure was applied (the jaws closing) leaving the impression of the teeth and the choane. Scaled drawings were made of the impressions using a camera lucida. ZOOTAXA Results The morphological and the molecular analysis clearly indicate that P. poireti consists of two full species, one from the Edough Peninsula and the other one throughout the rest of its range in Algeria and Tunisia. As explained in the introduction, the Edough Peninsula form is the one to which the name poireti Gervais correctly applies; the earliest available name for the other form is nebulosus Guichenot. Examples of some specimens of P. poireti and P. nebulosus are shown in Fig. 3. Pleurodeles poireti (Gervais, 1835) Lacerta palustris: Poiret (part), 1789: 290 Triton Poireti Gervais (part), 1835: 112. Gray, 1850: 18 Glossoliga Poireti: Bonaparte (part), 1839: unnumbered. Gervais, 1853: 312 Euproctus Poireti: Duméril & Bibron (part), 1854: 160 Glossoliga Hagenmulleri Lataste, 1881: 26 Molge hagenmulleri: Boulenger, 1882: : 93 Triturus hagenmulleri: Camerano, 1885: 419. Noble, 1924: 304 Triton (Pleurodeles) Hagenmulleri: Wolterstorff, 1901: 9 Pleurodeles poireti subsp. Hagenmulleri: Wolterstorff, 1905: 263 Pleurodeles poireti: Pasteur (part) 1958: 160. Schleich et al. 1996: 94 Diagnosis: A dwarf form of Pleurodeles usually not exceeding 129 mm in total length. Restricted essentially to the Edough Peninsula and lowland surrounding areas (Carranza and Arnold, 2003; Veith et al, 2004; see Fig. 1). Unique DNA sequences of the cytochrome b and 12S rrna mitochondrial genes (sequences submitted to Genbank for comparison; see Table 1 for accession numbers). Differentiated from the closely related P. nebulosus on account of the smaller size from the snout to the tip of the tail (t-value = REVISION OF PLEURODELES 2004 Magnolia Press 9

10 ; P = ; total size range of P. poireti mm) and an average uncorrected genetic distance of approximately 7.7% for the cytochrome b and 3.7% for the 12S rrna mitochondrial regions sequenced for the present study. Genetic variability within P. poireti is 0% for the cytochrome b and 0% for the 12S rrna although two specimens from the same locality have been analysed (see Table 1 and Fig. 1). Males of P. poireti differ from males of P. nebulosus in having smaller snout-vent length ( mm versus mm), head length ( mm versus mm), head with ( mm versus mm), interorbital width ( mm versus mm), tail length (50 73 mm versus mm), anterior arms length ( mm versus mm), 3 rd toe length ( mm versus mm) and length from the elbow to the third toe ( mm versus mm) (see Table 4 for statistics on all measurements). Females of P. poireti are differentiated from females of P. nebulosus by the first having shorter interorbital width ( mm versus mm) and anterior arms ( mm versus mm) (see Table 5 for statistics on all measurements). Some specimens of P. poireti have clearly a V shaped row of palatine teeth (see Fig. 4). Differentiated from P. waltl by its smaller size, in having usually 14 instead of 15 presacral vertebrae, in lacking sharp rib tips that project through a row of glandular swellings on the flanks and a tubercular process on rib 3 and often traces of this structure on ribs 1 and 2. It also differs from P. waltl in 9.2% of the cytochrome b and 3.7% of the 12S rrna mitochondrial regions sequenced for the present study. The two syntypes were part of a larger series with localities in addition to Bône (Annaba). Therefore, it seems advisable to designate a lectotype. Specimen MNHNP 4744A is selected on the grounds of its being complete (see Fig. 3, specimen 9). Description of lectotype: MNHNP 4744A. Male collected in Bône (Annaba), Algeria by Marley and Gérard; 93 mm in total length, see Table 1 for other relevant measurements. Preserved-hardened accompanied by some shrinkage. Dorsum and most of the tail strongly rugose and tuberculate, diminishing in coarseness towards the belly and the head, which anteriorly is almost smooth. Belly slightly rugose and gular fold pronounced. Testicular swelling moderately conspicuous. Anterior humoral baldges present. Black suffusion dorsally and dorsolaterally, coalescing into marbeling and ill-defined blotches. Colour in alcohol: rust with paler tubercles, lateral glandular protuberance and dorsal protuberances. Throat, belly, underside of the tail and digits pale. Transitions from dark to pale on the sides to the belly rather abrupt. Throat lightly spotted, belly very sparsely so. Description of the paralectotype: MNHNP 4744 (see Fig. 3, specimen 8). Female collected in Bône (Annaba), Algeria by Marley and Gérard. Right scapula and forearm removed. Total length 107 mm. For other measurement see Table 1. Colour and texture similar to the lectotype (MNHNP 4744A). The gular fold is virtually absent Magnolia Press CARRANZA & WADE

11 FIGURE 3. Photograph showing nine specimens of P. pioireti (above) and four P. nebulosus (below). A 23 centimetres scale bar is shown on the left-hand side of the picture; black rectangles and intermediate white spaces all represent 1 cm. Numbers above the specimens refer to: 1. BMNH , largest specimen of P. poireti included in the present study. Female from Bône (Annaba); 2. BMNH , male of P. poireti from Mount Edough; 3. BMNH , male of P. poireti from Mount Edough; 4. BMNH , male of P. poireti from Mount Edough; 5. BMNH , male of P. poireti from Mount Edough; 6. BMNH , male of P. poireti from Mount Edough; 7. BMNH , male of P. poireti from Mount Edough; 8. MNHNP 4744, female, paralectotype of P. poireti from Bône (Annaba); 9. MNHNP 4744A, male, lectotype of P. poireti from Bône (Annaba); 10. BMNH a, largest specimen of P nebulosus recorded to date. Male from N. Africa; 11. BMNH , female of P. nebulosus from Algiers; 12. BMNH , male of P. nebulosus from Algiers; 13. MNHNP 1442, female, lectotype of P. nebulosus from Algiers. Pleurodeles nebulosus (Guichenot, 1850) Lacerta palustris: Poiret (part), 1789: 290 Triton Poireti Gervais (part), 1835: 112. Gray, 1850: 18 Glossoliga Poireti: Bonaparte (part), 1839: unnumbered. Gervais, 1853: 312 REVISION OF PLEURODELES 2004 Magnolia Press 11

12 Euproctus Rusconii: Guichenot (part), 1850: 29 Euproctus Poireti: Duméril & Bibron (part), 1854: 160 Glossoliga Poireti Lataste, 1881: 26 Molge poireti: Boulenger, 1882: : 93 Triturus poireti: Camerano, 1885: 419. Noble, 1924: 304 Triton (Pleurodeles) Poireti: Wolterstorff, 1901: 9 Pleurodeles Poireti subsp. Poireti: Wolterstorff, 1905: 263 Pleurodeles poireti: Pasteur (part) 1958: 160. Schleich et al. 1996: 94 Diagnosis: A medium size Pleurodeles usually up to 180 mm in total length but exceptionally reaching 230 mm (BMNH a. This very large P. nebulosus was discovered among P. waltl specimens at the NHM, london. The results of the x-rays clearly showed that it has 14 presacral vertebrae and that it lacks the typical P. waltl tubercular process on rib 3 and traces of this structure on ribs 1 and 2. Moreover, as in P. nebulosus the specimen also lacks the typical P. waltl sharp rib tips and a row of glandular swellings on the flanks. It was identified as P. nebulosus and not P. poireti by means of its bigger size). Distributed across northern Algeria and Tunisia, except in the Edough Peninsula and surrounding lowland areas (Carranza and Arnold, 2003; Veith et al. 2004; see Fig. 1). Unique DNA sequences of the cytochrome b and 12S rrna mitochondrial genes with low intraspecific variability (sequences submitted to Genbank for comparison; see Table 1 for accession numbers). Differentiated from the closely related P. poireti on account of its bigger total size (t-value = ; P = ; total size range of P. nebulosus mm) and an average uncorrected genetic distance of approximately 7.7% for the cytochrome b and 3.7% for the 12S rrna mitochondrial regions sequenced for the present study. Genetic variability within P. nebulosus is 0.5% for the cytochrome b and 0.8% for the 12S rrna (21 specimens analysed covering almost the whole distributional range of the species; see Table 1 and Fig. 1). Males of P. nebulosus differ from males of P. poireti in having larger snout-vent length (53 69 mm versus mm) head with ( mm versus mm), interorbital width ( mm versus mm), tail length (62 90 mm versus mm), anterior arms length ( mm versus mm), 3 rd toe length ( mm versus mm) and length from the elbow to the third toe ( mm versus mm) (see Table 4 for statistics on all measurements). Females of P. nebulosus are differentiated from females of P. poireti by their larger interorbital width ( mm versus mm) and anterior arms (13 22 mm versus mm) (see Table 5 for statistics on all measurements). Some specimens of P. nebulosus have a clearly U shaped row of palatine teeth; none of the specimens studied presents clearly V shaped row of palatine teeth (see Table 1 and Fig. 4). Differentiated from P. waltl by its smaller size, in having usually 14 instead of 15 presacral vertebrae, in lacking sharp rib tips that project through a row of glandular swellings on the flanks and a tubercular process on rib 3 and often traces of this structure on ribs 1 and 2. It also differs from P. waltl in 9.7% of the cytochrome b and 4.8% of the 12S rrna mitochondrial regions sequenced for the present study Magnolia Press CARRANZA & WADE

13 FIGURE 4. Schematic diagrams of the palatine teeth of P. poireti and P. nebulosus obtained with plasticine casts (see material and methods). Numbers inside the drawings refer to specimens codes from Table 1. The position of the choanae was also obtained in the casts and it has been highlighted above the palatine teeth. Every drawing has a scale bar to the left, which corresponds to 2 mm. REVISION OF PLEURODELES 2004 Magnolia Press 13

14 The type series of P. nebulosus consists of skeletal material and one partially dissected spirit specimen (MNHNP 1442), which we designate as the lectotype. Description of lectotype: MNHNP 1442 (see Fig. 3; specimen 13). Female collected in Algiers by Guichenot; lower jaw and throat dissected, tip of the tail missing; mm in total length; head moderately depressed; dorsally and dorsolaterally moderately to strongly rugose/ turberculate; tubercles tend towards disposition in transverse irregular series separated by creases becoming progressively smaller and more numerous vertebrally and on the belly; the head is smoother; gular fold present. Colour in alcohol: dorsum very dark, tubercles little paler (if at all) becoming increasingly paler towards the belly; throat and palmar aspects of hands paler of yellowish ochre coloration; lateral glandular protuberances barely noticeable; dorsal protuberances inconspicuous anteriorly, slightly evident posteriorly, coalescing at the level of the cloaca to a continuous line on the tail. Phylogenetic analysis The results of the phylogenetic analysis are presented in Fig. 5 and show that the two individuals of P. poireti from Bône (Annaba) region represent an independent phylogenetic lineage sister to P. nebulosus. All three species of Pleurodeles are monophyletic with high bootstrap support. To calibrate the molecular clock we assumed that P. waltl separated from the ancestor of P. poireti and P. nebulosus at the end of the Mesinian Salinity Crisis, some 5.3 Myr ago (Hsu, 1972, 1973; Blondel and Aronson, 1999; Krijgsman et al., 1999; Duggen et al., 2003). This date marks the disappearance of the land bridge that for more than 600,000 years had connected southern Spain with North Africa, and had facilitated greatly the faunal exchange between both continents. Therefore, for many animals that, as the urodels, do not disperse easily across bodies of salt water, the end of the Messinian Salinity Crisis 5.3 Myr ago was a very important vicariant event, and therefore an excellent calibration point (see Carranza and Arnold, 2003 for a discussion on the alternative clock calibration using the less reliable fossils data available for this group of urodels). According to our calibration, P. poireti and P. nebulosus have been evolving as independent lineages for approximately 4.2 Myr (see Carranza & Arnold, 2003). Genetic distances among all three species of Pleurodeles are very similar for the 12S rrna gene (3.7% for P. poireti vs P. waltl; 3.7% for P. poireti vs P. nebulosus and 4.8% for P. waltl vs P. nebulosus) and not so different for the cytochrome b gene (9.2% for P. poireti vs P. waltl; 7.7% for P. poireti vs P. nebulosus and 9.7% for P. waltl vs P. nebulosus) Magnolia Press CARRANZA & WADE

15 FIGURE 5. ML tree (Log likelihood ; GTR+G model of sequence evolution) inferred from cytochrome b and 12S rrna sequences. Bootstrap support is shown by the branches. Numbers separated by a dash from the species names refer to specimen codes from Fig 1 and Table 1. The estimated date for the separation between P. poireti and P. nebulosus and for the calibration point used in the analysis is marked with a filled circle. Morphometric analysis An analysis of sexual dimorphism within P. poireti and P. nebulosus was carried out using a student-t and indicated that males and females of both species differ statistically in some REVISION OF PLEURODELES 2004 Magnolia Press 15

16 of the variables used (see Tables 2 and 3). Our limited sampling suggests sexual dimorphism is more accentuated in P. poireti, with males being significantly smaller (SVL) than females, having shorter and narrower heads, shorter interorbital width but having relatively longer tails and anterior arms. Sexual dimorphism in P. nebulosus is less pronounced than in P. poireti, with males having absolute and proportionally longer tails and longer forearms than females. As a result of the positive results of the sexual dimorphism males and females of both species were analysed separately. The results of the morphological comparisons are presented in Tables 4 (males) and 5 (females). Males of P. poireti and P. nebulosus are significantly different in all 8 variables measured. However, they do not differ in any of the 5 body proportions calculated suggesting that there is has been an isometric reduction in size in males of P. poireti. Values for all 8 variables that are significantly different between males of both species are always higher for P. nebulosus than for P. poireti. Females of P. poireti and P. nebulosus are not so different, with P values being significant for the difference in the interorbital width and the length of the anterior arms. As in males, P. nebulosus females always present higher values than P. poireti females in all measurements. TABLE 2. Analysis of sexual dimorphism between 8 males and 11 females of Pleurodeles nebulosus using a student-t test for 13 discrete measurements and body proportions. Significant P values at the P < 0.05 level have been highlighted with an asterisk and at the P < 0.01 with two asterisks. Variable Mean (Males) Mean (Females) t-value P SVL HL HW IOW TL * SP3T T AL ** TL/SVL ** HL/SVL HW/SVL HW/HL AL/SVL ** Magnolia Press CARRANZA & WADE

17 TABLE 3. Analysis of sexual dimorphism between 16 males and 6 females of Pleurodeles poireti using a student-t test for 13 discrete measurements and body proportions. Significant P values at the P < 0.05 level have been highlighted with an asterisk and at the P < 0.01 with two asterisks. ZOOTAXA Variable Mean (Males) Mean (Females) t-value P SVL ** HL ** HW ** IOW * TL SP3T T AL TL/SVL ** HL/SVL HW/SVL HW/HL AL/SVL ** Discussion The molecular and morphological evidence presented here supports Lataste s (1881) hypothesis that a different species of Pleurodeles lives in the Edough Peninsula in northeast Algeria (see Fig. 1). Our results, are at variance with Pasteur (1958), who recognised one species of Algero-Tunisian Pleurodeles. One of the possible explanations as to why Pasteur (1958) decided to synonymise both forms is because two MNHNP specimens from Tunisia (27.27 and 27.28) labelled as P. hagenmulleri (=P. poireti) included in his study were, in fact, P. nebulosus. According to our analyses, all Algero-Tunisian Pleurodeles populations outside the Edough Peninsula (Fig. 1) are genetically and morphologically P. nebulosus. This includes all 14 Tunisian individuals from 7 different localities (some of these very close to the Algerian border). This has important implications, especially since Pasteur (1958) did not include many P. hagenmulleri (=P. poireti) in his work, and these two Tunisian specimens had the morphological characteristics of P. nebulosus. The much larger size of the two MNHNP Tunisian specimen (135 mm) and (179 mm) was used by Pasteur (1958) to seriously question the validity of Lataste s (1881) new species. According to our measurements, the maximum total size of P. poireti is 129 mm, while the much bigger P. nebulosus can easily reach 180 mm and exceptionally 230 mm (see Fig. 3, specimen 10). This size difference is statistically significant (t-value = - REVISION OF PLEURODELES 2004 Magnolia Press 17

18 ; P = ) indicating that, as suggested by Lataste (1881), size is one of the main diagnostic characters that differentiates between both forms of Algero-Tunisian Pleurodeles (see Tables 4 and 5 and Fig. 3). Other diagnostic characters listed in Lataste (1881) were the shape of the row of palatine teeth (V shaped in the Edough Peninsula species and U shaped in all the rest), and some subtle differences in the general shape of the head, the snout, and the contour of the jaws (semi-elliptical in P. poireti and semi-circular in P. nebulous). The drawings of the palatine teeth shown in Fig. 4 indicate that, some Edough Peninsula P. poireti have a clearly V shaped row of palatine teeth (specimens 11 and 9 in Fig. 4). However, and despite the high level of intraspecific variability, there is not a single specimen of P. nebulosus in our sample with a clearly V shaped row of palatine teeth (see Fig. 4). Therefore, when present, the character palatine teeth forming a clearly V shaped row can be used to differentiate between these two forms (Lataste, 1881; Boulenger, 1882; Doumerge, 1901). A drawing of the skull of an individual of each species is shown in Fig. 6 for comparison. Once more, despite the high level of intraspecific variability (Pasteur, 1958, SC and EW pers. obs.), there is a clear difference between both species in the general shape of the skull, snout and contour of the jaws (Lataste, 1881). FIGURE 6. Ventral aspect of the skull of A.-adult P. poireti, BMNH (Bône) and B. adult P. nebulosus, BMNH 130a (Algiers). The high level of intraspecific morphological variability presented by the Edough Peninsula P. poireti for most of Lataste s (1881) diagnostic characters is due to the pronounced level of sexual dimorphism existent in this species. One of the consequences of this sexual dimorphism (see Table 3), is that the morphological differences between males of P. poireti and P. nebulosus are greater than the differences between females of both spe Magnolia Press CARRANZA & WADE

19 cies (Tables 4 and 5; Lataste, 1881). Pasteur (1958) analysed his material without any apparent regard to sex, appearing to overlook sexual dimorphism in P. poireti and, in consequence, he was not able to correctly evaluate the level of morphological difference between both species. Had he analysed males and females separately, he would have been less likely to discard Lataste s (1881) diagnostic characters (see Table 4). Despite being genetically very different, the females of P. poireti and P. nebulosus included in our analysis are morphologically not so different (see Table 5), and therefore more difficult to differentiate using Lataste s (1881) diagnostic characters. As it is often the case, especially among species of Triturus, the males possess the sexually dimorphic characters, whereas females of closely related species are more uniform and therefore more difficult to differentiate. ZOOTAXA TABLE 4. Results of the t-test for 13 morphological characters between males of P. poireti and P. nebulosus. Significant P values at the P < 0.01 have been highlighted with two asterisks. Variable Mean (P. poireti) Mean (P. nebulosus) t-value P SVL ** HL ** HW ** IOW ** TL ** SP3T ** T ** AL ** TL/SVL HL/SVL HW/SVL HW/HL AL/SVL Differences in sexual dimorphism between populations or, like in this case, closely related species could be caused by many factors, including differences in climatic conditions, resource partitioning with other species, sexual selection, etc. In the Pyrenean Brook newt (Euproctus asper), for instance, differences in sexual dimorphism observed between alpine Central Pyrenean populations and Prepyrenean ones are thought to be caused by increased selection pressures favouring high reproductive efficiency and reduced intersexual competition under alpine, high-mountain climate conditions (Serra-Cobo et al. 2000). Unfortunately, very little is known about the ecology and life history of the populations of REVISION OF PLEURODELES 2004 Magnolia Press 19

20 P. poireti from the Edough Peninsula and the widespread Algero-Tunisian P. nebulosus. Until more data is available, it will not be possible to understand which factors are causing the differences in sexual dimorphism between these two different species. TABLE 5. Results of the t-test for 13 morphological characters between females of P. poireti and P. nebulosus. Significant P values at the P < 0.05 level have been highlighted with an asterisk and at the P < 0.01 with two asterisks. Variable Mean (P. poireti) Mean (P. nebulosus) t-value P SVL HL HW IOW ** TL SP3T T AL * TL/SVL HL/SVL HW/SVL HW/HL AL/SVL Possible origin of P. poireti The data presented here indicates that P. poireti is restricted to a relatively small area in the North-east of Algeria (see Fig. 1). This region, referred throughout the paper as the Edough Peninsula, is a mountainous area with some peaks reaching up to 850 m above the sea level. Unlike other mountainous parts of the Algero-Tunisian coast, the Edough Peninsula is connected to the mainland through an area of lowland marshes with an important fluvial system and a relatively large brackish lake (Lac Fetzara). This area of lowland marshes surrounding the Edough Peninsula, probably represents the southern limit of P. poireti, which extends as far as 25 Km to the southeast of Annaba (Carranza and Arnold, 2003; Veith et al. 2004; see Fig. 1). The molecular analysis indicates that speciation between P. poireti and P. nebulosus occurred approximately 4.2 Myr ago, during the Upper-middle Pliocene. This may be explained by assuming that, shortly after the land-bridge between Africa and Europe disappeared approximately 5.3 Myr ago, a population of the ancestor of P. poireti and P. neb Magnolia Press CARRANZA & WADE

21 ulosus became isolated in the Edough Peninsula. During The Pliocene, global changes in the sea-level, crustal movements and local landform changes affected the planet causing marine transgressions to inundate many coastal areas all over the planet (Lanza, 1984; Morafka, 1976; and references therein). These Pliocene marine transgression may have inundated the lowland area surrounding the Edough Peninsula, which would have become a temporary island. In biogeographical terms, such temporary islands are known as continental fossil islands (Lanza, 1984). The existence of Pliocene fossil islands and their biogeographic importance has been studied in depth for the Toscana region in Italy (Lanza, 1984) and for other regions as far as central coastal California (Santa Lucia and Gabilan Ranges) (Morkafa and Banta, 1973; Morafka, 1976). In both cases, it was shown that the level of endemicity in these regions was much higher than in the surrounding areas. For instance, in coastal California, Peabody and Savage (1958) and Morafka and Banta (1973) showed a clear correlation between Pliocene embayment and subspecific boundaries of herpetofauna in this region. Stebbins and Major (1965) and Jepson (1925) reported the presence of plant endemic species in ranges that were formerly islands (i.e. Santa Lucia and Franciscan). An endemic rodent species (Diplodomys elephantinus) is known for the Gabilan fossil island, in coastal California (Morafka, 1976). In Europe, the level of endemicity in fossil islands is also high and, among animals, it includes several species of gastropods, terrestrial isopodes, coleopterans, a murine rodent and a giant hedgehog (Lanza, 1984, Butler, 1980). The presence of the endemic P. poireti and the geology of the area suggest that the Edough Peninsula might be a Pliocene fossil island. If that were so, we would expect to find other endemic plants and animals. Among amphibians, good candidates to test the fossil island hypothesis are Hyla meridionalis, Salamandra algira, Bufo bufo, Bufo mauritanicus, Bufo calamita and Rana saharica. Of all these species, S. algira has been studied from the genetic point of view (Steinfart et al. 2000). This study showed that the Annaba specimens were genetically very different from the Moroccan ones. But, apart from those from Annaba, no other samples of S. algira from Algeria were included in the study, so it is not possible to know if there is an endemic Salamandra from the Edough Peninsula yet. Future studies should focus in this interesting area of northeastern Algeria in order to see if there are any other animal or plant endemics and to find any geological clues that may support the fossil island hypothesis. ZOOTAXA Acknowledgements We are indebted to E. N. Arnold, J. Roca and S. Arroyo for their help. We appreciate the contribution of D. Donaire, R. Bour (MNHNP), Soumia Fahd, R. Fonoll, B. Clark (BMNH), O. Arribas and D. Siebert in several aspects of this work. We are also very grateful to the authors of the Atlas de Andalucia (J. P. González de la Vega, J. M. Barnstein, D. Donaire and L. García) for all the samples of Pleurodeles from southern Spain used in REVISION OF PLEURODELES 2004 Magnolia Press 21

22 this study. Salvador Carranza is supported by a Ramón y Cajal contract from the Ministerio de Educación y Cultura, Spain. Literature cited Blondel, J. & Aronson, J. (1999) Biology and wildlife of the Mediterranean region. Oxford University Press, New York, 327 pp. Bonaparte, C.L.P. ( ) Iconografia della Fauna Italica, 2. Anfibi. Roma, 26 fascicles. Bons, J. & Geniez, P. (1996) Amphibians & reptiles of Morocco (including Western Sahara), Biogeographical atlas. Asociación Herpetológica Española, Barcelona, 320 pp. Boulenger, G.A. (1882) Catalogue of the Batrachia Gradientia S. Caudata and Batrachia Apoda in the collection of the British Museum. Taylor and Francis, London, UK. VIII, 127 pp. Butler, M. (1980) The giant erinaceid insectivore, Deinogalerix Freudental, from the Upper Miocene of Gargano, Italy. Scripta Geologica, 57, Carranza, S. & Arnold, E.N. (2004) History of West Mediterranean newts, Pleurodeles (Amphibia: Salamandridae), inferred from old and recent DNA sequences. Systematics and Biodiversity, 1 (3), Camerano (1885) Memorie Accademia delle Scienze di Torino. Memorie Classe di Scienze Fisiche, Matematiche, e Naturali, (2), 36, 419. Doumerge, F. (1901) Essai sur la faune herpétologique de l Oranie. Fouque éd, Oran, 404 pp. Extract from Bulletin de Société Géographie et Archéologie d Oran ( ). Duggen, S., Hoernle, K., Bogaard, P. van den, Rüpke, L. & Morgan, J.P. (2003) Deep roots of the Messinian salinity crisis. Nature, 422, Duméril, A.M.C. & Bibron, G. ( ) Erpétologie générale ou histoire naturelle complète des Reptiles. Roret, Paris, XXXII pp. Felsenstein, J. (1985) Confidence-limits on phylogenies an approach using the bootstrap. Evolution, 39, Frost, D.R. (1985) Amphibian species of the world. A taxonomic and geographic reference. Allen Press, Inc. and the Association of Systematics Collections Lawrence, Kansas, USA, 732 pp. Gené (1838) Synopsis reptilium Sardiniae indigenorum. Memorie Accademia delle Scienze di Torino. Memorie Classe di Scienze Fisiche, Matematiche, e Naturali, (2) 1, Gervais, P. (1835) Communication sur les Reptiles de Barbarie. Bulletin de la Société de Sciences Naturelles de France, Séance du , Gervais, P. (1853) Le Glossoliga Poireti et l Euproctus Rusconii. Annales des Sciences Naturelles, Paris 3 rd series, t XX, Gravenhorst, J.L.K. (1829) Deliciae musei zoologici vratislaviensis. 1. Chelonios et Batrachia. Sumptibus L. Vossii, Leipzig, XIV+106 pp. Gray, J.E. (1850) Catalogue of the specimens of Amphibians in the Collection of the British Museum, 2. Batrachia gradientia. British Museum (Natural History), London, 72 pp. Guichenot, A. (1850) Histoire naturelle des Reptiles et des Poissons. Exploration scientifique de l Algérie pendant les annes 1840, 1841, Zoologie. Bibliothèque Français, Paris, 144 pp. Herre, W. (1941) Palaeopleurodeles hauffi nov. gen., nov. spec., ein fossiler Schwanzlurch aus dem Miozän Süddeutschlands. Zoologischer Anzeiger, 134, Hsu, K. J., Ryan, W. B. F. & Cita, M. B. (1973) Late Miocene desiccation of the Mediterranean. Nature, 242, Hsu, K. Montadert, J., Beernouilli, L.D., Cita, M.B., Erickson, A., Garrison, R.E., Kidd, R.B., Melieres, F., Muller, C. & Wright, R. (1977) History of the Mediterranean salinity crisis. Nature, 267, Magnolia Press CARRANZA & WADE