C. R. Palevol 5 (2006) 803–811 http://france.elsevier.com/direct/PALEVO/ Systematic Paleontology (Vertebrate Paleontology) Eurotestudo, a new genus for the species Testudo hermanni Gmelin, 1789 (Chelonii, Testudinidae) France de Lapparent de Broin a,*, Roger Bour b, James F. Parham c, Jarmo Perälä d a d « Paléobiodiversité et paléoenvironnements », département « Histoire de la Terre » du Muséum national d’histoire naturelle, UMS 203, UMR 5143 du CNRS, CP 38, 8, rue Buffon, 75231 Paris cedex 05, France b Laboratoire des reptiles et amphibiens, département « systématique et évolution » du Muséum national d’histoire naturelle, UMS 602, CP 30, 25 rue Cuvier, 75231 Paris cedex 05, France c Museum of Paleontology, University of California, Berkeley, CA, 94720, USA/ Department of Evolutionary Genomics, DOE Joint Genome Institute and Lawrence Berkeley National Laboratory, 2800 Mitchell Drive, Walnut Creek, CA, 94598, USA Department of Biological and Environmental Sciences, PO Box 65 (Biocenter 3, Viikinkaari 1), FIN-00014 university of Helsinki, Finland Received 18 November 2005; accepted after revision 21 February 2006 Available online 02 May 2006 Presented by Philippe Taquet Abstract The new genus is created to include the species of the hermanni group, which is within Testudo s.l., a Palearctic genus, consequently separated from both Testudo s.s. and Agrionemys. A preliminary cladistic analysis of the osteological characters, including fossil species, demonstrated the splitting of the three lineages, probably since the Oligocene and surely at the Upper Miocene. Diagnosis of the new genus is based on a collection of features. The main stages of evolution leading to the three lineages are provided. We also describe external characters of the extant species that could be considered as diagnostic. However, phylogenetic relationships between genera are not definitively established. To cite this article: F. de Lapparent de Broin et al., C. R. Palevol 5 (2006). © 2006 Académie des sciences. Published by Elsevier SAS. All rights reserved. Résumé Eurotestudo, nouveau genre pour l’espèce Testudo hermanni Gmelin, 1789 (Chelonii, Testudinidae). Ce nouveau genre est créé pour le groupe d’espèces hermanni, le séparant de Testudo s.s. et d’Agrionemys au sein de Testudo s.l., genre paléarctique. Une analyse cladistique des caractères ostéologiques, menée au préalable et incluant des espèces fossiles, a montré la séparation des trois lignées, probablement depuis l’Oligocène et sûrement le Miocène supérieur. La diagnose du nouveau genre est établie sur une conjonction de caractères. Les principales étapes de l’évolution menant aux trois genres sont données. Les caractères externes des espèces actuelles pouvant participer à la diagnose sont examinés. Les relations phylétiques entre les genres ne sont pas établies définitivement. Pour citer cet article : F. de Lapparent de Broin et al., C. R. Palevol 5 (2006). © 2006 Académie des sciences. Published by Elsevier SAS. All rights reserved. Keywords: Turtles; Eurotestudo n.g.; Testudo s.l.; Testudinidae; Europe; Tertiary–Extant Mots clés : Tortues ; Eurotestudo n.g. ; Testudo s.l. ; Testudinidae ; Europe ; Tertiaire–Actuel * Auteur correspondant. E-mail address: fdelap@mnhn.fr (F. de Lapparent de Broin). 1631-0683/$ - see front matter © 2006 Académie des sciences. Published by Elsevier SAS. All rights reserved. doi:10.1016/j.crpv.2006.03.002 804 F. de Lapparent de Broin et al. / C. R. Palevol 5 (2006) 803–811 Version française abrégée Introduction Un nouveau genre est créé pour le groupe d’espèces hermanni, en le séparant de Testudo s.s. [20]. Il existe trois lignées distinctes, séparées au sein de Testudo s.l. (sensu [17,23,24]), Testudinidae terrestre d’origine paléarctique : (a) Testudo s.s., connu en Europe [1,24] depuis le Miocène supérieur au moins [17], et en Afrique sûrement depuis le Pliocène, mais probablement dès le Miocène [15]; (b) Agrionemys [12], connu en Afghanistan et en Moldavie au Miocène supérieur [8,17,30], lié ou non à des formes orientales plus anciennes dont « T. » turgaica du Miocène « moyen » [17,31,34] ; (c) la lignée d’Eurotestudo n.g., représentée en Europe au plus tard depuis le Miocène supérieur et moyen [5,16,32] de France et d’Allemagne et probablement dès l’Oligocène et actuellement représentée en Europe au moins par « T. » hermanni Gmelin, 1789 et « T. » boettgeri Mojsisovics, 1889 [1–3,28]. Des formes du Miocène inférieur [5] peuvent se situer, soit dans cette lignée, soit dans une lignée commune avec celle de Testudo s.s., suivant leur point de séparation [20]. La définition du nouveau genre proposée ici est fondée sur des caractères ostéologiques [5,6,9,10,18,19,23,24], grâce à une nouvelle analyse cladistique incluant des espèces fossiles [5,7,16,30–32] et actuelles de Testudo s.l., ce qui a permis de polariser les caractères et de faire la part des homoplasies. Certains caractères externes des espèces actuelles peuvent participer à la diagnose [1–4,9,23,24,27]. Systématique (voir la version anglaise) Eurotestudo n.g. Espèce type: Testudo hermanni Gmelin, 1789 Espèces valides incluses Le groupe hermanni: Eu. hermanni (Fig. 1), Eu. boettgeri (dont Eu. hercegovinensis tend à être séparé [28]), Eu. pyrenaica, Eu. globosa, Eu. lunellensis, Eu. szalai. Les espèces actuelles et les fossiles Eu. pyrenaica et Eu. lunellensis, sont ou peuvent être bien définies. Eu. globosa (un seul spécimen, mâle, os épais) se présente comme un représentant de Eu. hermanni. Eu. szalai (fragments isolés) n’est pas assez préservé pour être diagnostiqué [1–3,5,11,13,21]. Diagnose du genre Eurotestudo n.g. est caractérisé par la nécessaire combinaison de (1) la série des vertébrales étrécie ; (2) la fusion des suprapygales en un trapèze à bord postérieur rectiligne (ou l’état de la plus forte tendance à la fusion précédant celui de celle-ci) ; (3) la pygale quadrangulaire tendant à devenir hexagonale, à petits côtés antérieurs (alors souvent chevauchés par les marginales 11) ; (4) la tendance à la division de la supracaudale, externe (souvent) et interne (moins souvent) ; (5) la surface ventrale des gulaires formant un triangle dirigé postérieurement et souvent saillant ventralement, souvent avec inflexion médiane antérieure du bord, saillant à l’avant, des gulaires. Les caractères 2 à 5 sont indépendamment homoplasiques chez Agrionemys et/ ou Testudo s.s., rarement, et jamais tous ensemble (Fig. 1A, B, C, D). Caractères externes additifs, non fossilisés (synapomorphies des espèces actuelles potentiellement génériques) : (1) sur la face frontale du bras, distale par rapport aux grandes écailles, une aire antéro-distale limitée, avec, soit des écailles petites et irrégulières (Eu. hermanni), soit uniquement de nombreuses écailles très petites (Eu. boettgeri), alors que toutes les écailles sont grandes et régulières chez les autres espèces de Testudo s.l. ; (2) écaille frontale fragmentée, presque indistincte ; (3) patron de coloration du plastron avec deux bandes foncées parasagittales, entières ou fragmentées [1,3,4]. Matériel référé au genre: tous les spécimens référés à T. hermanni et à Testudo sp. du Quaternaire d’Europe ayant les caractères donnés dans la diagnose et notamment ceux de l’Escale, Lunel-Viel [10,11] et Soave [33]. Les populations de Lunel-Viel et de Soave représentent de bonnes espèces, suffisamment préservées pour être diagnostiquées. La lignée hermanni : elle débute avec Paleotestudo canetotiana [16,19] par la tendance, plus complète que chez les autres espèces de Testudo s.l., à la fusion des trochanters du fémur, puis avec « T. » antiqua [32] par les tendances conjointes à la division externe de la supracaudale et à la fusion des suprapygales, enfin avec Eurotestudo n.g., où tous les caractères sont menés à leur terme [20]. Comparaisons morphologiques Étude cladistique préalable. Une étude cladistique préalable, détaillée par ailleurs [20], inclut un nombre significatif de spécimens des espèces des lignées de Testudo s.s., d’Agrionemys et d’hermanni (voir la version en anglais) et certains de ses possibles alliés [5,16], Paleotestudo canetotiana et Testudo promarginata. Les outgroups sont Manouria impressa, Indotestudo elongata et « Ergilemys » [7] (sensu [5]) bruneti. Tous par- F. de Lapparent de Broin et al. / C. R. Palevol 5 (2006) 803–811 tagent des caractères de Testudinidae terrestres [20]. Testudo s.l. partage des caractères avec Indotestudo et « Er. bruneti » et d’autres avec le seul « Er. bruneti ». Le genre Testudo s.l. est défini, les caractères partagés sont donnés ainsi que des particularités des formes de la lignée d’Eurotestudo n.g. [5,10,11,13,16–18,32]. Le point de séparation de Testudo s.s. et de la lignée Eurotestudo par rapport à Paleotestudo canetotiana [16] est examiné, ainsi que l’intégration de « T. » antiqua [32] dans la lignée. Caractères externes des représentants actuels pouvant appuyer la séparation générique. Les caractères, mentionnés ci-dessus et retenus, sont examinés, ainsi que d’autres, à écarter de la diagnose, tels les tubercules des cuisses (apomorphie de Testudo s.s.), l’éperon caudal présent, mais variable, chez Agrionemys, Eurotestudo et T. kleinmanni et les modes de réduction de la main: réduction à quatre doigts (Agrionemys) ou seulement partielle chez Eurotestudo (ongle du 1er et/ou du 5e doigts, éventuellement réduits à absents) [1,3,4,9,23,24,27]. Discussion et conclusion Les relations phylétiques des trois lignées par les différentes approches (morphologie des actuels ou/et des fossiles [9,23,24], analyse moléculaire [14,22, etc.]), ne peuvent être définies. Il apparaît que, suivant les taxons inclus et en fonction des méthodes utilisées, le groupe actuel hermanni (hermanni seule ou avec boettgeri) peut, soit être rapproché d’Agrionemys [9], mais aussi d’Indotestudo et d’autres taxons [14,22], soit être le groupe frère d’Agrionemys et de Testudo s.s. [24]. D’après l’étude sur laquelle est fondée la présente diagnose [20], les trois lignées sont bien séparées, après Manouria impressa, Indotestudo et «Er. » bruneti, en un groupe « Testudo s.l. ». Soit la lignée d’Eurotestudo n.g. est rapprochée de Testudo s.s., soit les trois lignées de Testudo s.l. sont en irrésolution, si l’on supprime le taxon asiatique fossile « T. » turgaica, moins bien connu. Testudo s.s. et Eurotestudo n.g. acquièrent un même mode de recourbement progressif du bourrelet épiplastral dorsal. Le caractère est constamment mené à son terme chez Testudo s.s. dès son apparition (présence d’une poche gulaire, recourbement jusqu’à l’entoplastron), moins souvent chez Eurotestudo n.g. Agrionemys présente le stade le moins avancé du processus évolutif et dans une conformation différente du lobe antérieur plastral (plus large avec bords latéraux plus convergents et entoplastron moins réduit). Il y a hétérochronie dans l’apparition des stades évolutifs de plusieurs caractères homoplasiques dans les deux groupes. Le patron de 805 coloration commun de la carapace de type « Testudo » milite aussi en faveur de l’union de Testudo s.s. et Eurotestudo. En tout état de cause, le point de séparation de la lignée d’Eurotestudo n.g. par rapport aux formes asiatiques originelles remonte à une époque indéterminée, mais antérieure à l’Oligocène. 1. Introduction The principal aim of this work is to create a new genus Eurotestudo for the so-called hermanni group of testudinids, because it forms a distinct evolutionary lineage without an available name. Some valid names that seemed available for Testudo hermanni Gmelin, 1789 such as Chersine Merrem, 1820 and Medaestia Wussow, 1916, have Testudo graeca Linnaeus, 1758 as type species ([23], A. Rhodin in litt. to J.P.). Eurotestudo n.g. is part of Testudo s.l. (sensu [17,23,24]). This is a diverse group of terrestrial Palaearctic testudinids which, besides (a) Eu. hermanni and affiliated taxa [1–6,10,11,13,16,17,19,24,32] includes (b) the western hinged form Testudo Linnaeus, 1758, s.s., type species T. graeca, a genus extant in the southern-oriental Mediterranean Basin eastward to the Middle East [1,6,8,9, 15,17,24], and (c) Agrionemys Khozatsky & Mlynarski, 1966 [13], type species Testudo horsfieldii Gray, 1844, a western Central Asiatic extant genus, only represented in Europe as fossil (eastern part). Some recent studies [23–26] have elevated many subspecies to the rank of species within Agrionemys, Testudo s.s., and ‘T.’ hermanni Gmelin, 1789, while new species have also been described recently [29]. T. hermanni (osteological Fig. 1) was separated from T. boettgeri Mojsisovics, 1889 (osteological Fig. in [9] as T. hermanni), and the name T. hercegovinensis Werner, 1899 was resurrected for a population previously attributed to boettgeri [28] (not included in the analysis in [20]). The taxon of upper rank (according to the ICZN) to unite the extant and fossil species in the hermanni complex is a genus, necessary in accordance with previous opinions [16,17, 23,24], that agree with the various hypotheses about phylogenetic relationships among the three groups [14,17,22,24]. Examination of fossil lineages, into which we can integrate the extant species, shows that there is a clear separation of the three groups, each one inclusive of a succession of valid species: the separation occurred, at least, since the Upper Miocene, but probably the Oligocene. It is the date of the appearance of the oldest attested Testudo s.s., Testudo marmorum Gaudry, 1862 (Greece). In Africa, Testudo s.s. is definitely known from the Pliocene (Morocco). However, Testudo (‘s.l.’) semenensis Bergounioux, 1955, from 806 F. de Lapparent de Broin et al. / C. R. Palevol 5 (2006) 803–811 Fig. 1. Eurotestudo hermanni Gmelin, 1789, Collobrières, France. Carapace, views: A, dorsal, B, ventral, C, posterior. Plastron, anterior lobe, D, dorsal view. Fig. 1. Eurotestudo hermanni Gmelin, 1789, Collobrières, France. Carapace, vues: A, dorsale, B, ventrale, C, postérieure. Lobe antérieur du plastron, D, vue dorsale. the Upper Miocene (Tunisia), may be attributable to Testudo s.s [15,20]. Agrionemys is firstly known from the Upper Miocene of the Republic of Moldova, by ‘Testudo’ bessarabica Riabinin 1915 [30] (Protestudo Chkhikvadze, 1970) [6], and of Afghanistan ([17] and Rage & Lapparent de Broin, in prep.]. Agrionemys might be related to ‘Testudo’ turgaica Riabinin, 1926 [31], from the ‘Middle’ Miocene of Khazakstan as well as to other Asiatic or eastern European forms [8,34]. The stem lineage of Eurotestudo n.g. is identified in the ‘Middle’ Miocene with the appearance of Paleotestudo canetotiana (Lartet, 1851), France [5,16] and in the Upper Miocene with ‘Testudo’ antiqua Bronn, 1831, Germany) [32]. Other older extinct western European species, such as ‘Testudo’ promarginata Reinach, 1900, from the Lower Miocene (Germany, France) [5], may also be on the stem of Eurotestudo n.g. However, ‘T’. promarginata may predate the split between Eurotestudo n.g. and Testudo s.s. [20]. The diagnosis of the extant group hermanni has already been established on the basis of the morphological study of osteology [5,6,9,10,18,19,23,24] and external characters such as horny appendices, coloration and scales [1–4,24,26]. The present diagnosis of Eurotestudo n.g. is principally based on characters of the carapace, preserved in the fossils: plates and scute outlines as well as proportions, features which no doubt characterize the whole genus. Many specimens of the fossil species and extant populations of the lineages of Testudo, Agrionemys and Eurotestudo n.g. have been exam- ined. The diagnosis of each lineage includes some unambiguous characters and various homoplastic characters: their appearance in the lineages is asynchronous and their variability has been established for each population. 2. Systematics Order Chelonii Brongniart (Latreille), 1800 Superfamily Testudinoidea Batsch, 1788 Family Testudinidae Batsch, 1788 Infrafamily Testudininei Batsch, 1788 Eurotestudo new genus Etymology: from ‘Europe’, the continent of biogeographic origin, and ‘Testudo’ Type species: Testudo hermanni Gmelin, 1789, type locality: Collobrières, Var, France 2.1. Included species Named valid species (sensu ICZN): The ‘Eurotestudo’ hermanni group: extant Eu. hermanni (Fig. 1) and boettgeri from which ‘T.’ hercegovinensis Werner, 1899 may be disassociated [28], and the fossil Eu. pyrenaica (Depéret & Donnezan, 1890), Pliocene of Perpignan (MN 15), Eu. globosa (Portis, 1890), Plio-Pleistocene boundary, Le Ville, Upper Valdarno, Eu. lunellensis (Almera & Bofill, 1903), ‘Middle’ Pleistocene of Caverna de Gràcia, and Eu. szalai Mlynarski, 1955, Pliocene of Weze (MN 15). The extant species plus Eu. pyrenaica and Eu. lunellensis are or can be F. de Lapparent de Broin et al. / C. R. Palevol 5 (2006) 803–811 (respectively) well diagnosed. Eu. globosa (one male specimen in Le Ville, thick bones; referred fragments in other localities from Valdarno) may be a junior synonym of Eu. hermanni. Eu. szalai (some fragments) cannot be sufficiently diagnosed [1,4,5,10–12,21]. Material referred to the genus: all the specimens in the literature referred to T. hermanni and Testudo sp. from the Quaternary of Europe, which present the characters of Eu. hermanni exposed in the present diagnosis: in particular the populations of T. hermanni from the Quaternary of France, especially the populations from l’Escale (ca 0.6 Myr) and Lunel-Viel (ca 0.3 to 0.34 Myr), and ‘Testudo cf. hermanni’ from Soave (Zoppega 2, Italy) (early Middle Pleistocene) [10,11, 33]. The unnamed Lunel-Viel and Soave populations certainly represent distinct diagnosable species. The ‘hermanni lineage’ initially includes Paleotestudo canetotiana [16] by the trend towards the fusion of the trochanters, more complete than in Agrionemys and Testudo s.s. [19], then ‘T.’ antiqua [32] by the common trend towards an external division of the supracaudal and fusion of the suprapygals and finally Eurotestudo n.g. [20] where these characters are the best realized. 2.2. Diagnosis Eurotestudo n.g. is diagnosed by the obligatory combination of the following characters: (1) narrowed vertebral series, narrower than the costal series as a whole (in all populations; an apomorphic character); (2) fusion of the suprapygals into a trapezoid with a straight posterior border: the fusion varies from occasional in fossils (but often incompletely preserved) to most often present in extant populations (a rare homoplasy in Testudo s.s. and Agrionemys); (3) the quadrangular pygal becomes hexagonal with small latero-anterior sides (often present in all populations, a rare homoplasy in Testudo s.s. and Agrionemys), and which are sometimes covered by the 11th marginals; (4) tendency to having a divided supracaudal, externally and eventually internally; frequency of inner division of the supracaudal varies from rare to frequent depending on population (present in ‘Ergilemys’, but presumably not by the same evolutionary process, see [20]); external division of the supracaudal occasional to constant, according to population, constant in hermanni and in the majority of cases in boettgeri (a rare homoplasy in Testudo s.s. and extant Agrionemys and in some other Testudininei such as Pyxis); (5) ventral surface of the gulars, making a posteriorly pointed triangle, frequently ventrally in relief relative to horizontal plan, often present in all 807 populations (a homoplasy in extant Agrionemys); there is often a medial anterior bend between the gulars (a homoplasy in Testudo s.s. and in extant Agrionemys) (Fig. 1). Additional characters: external characters, not fossilized, possibly generic. Synapomorphies for the extant species (unique among the Testudinidae) are: (1) the small scales on the outer area of the front face of the forearm (in addition to the large and regular ones): irregular antero-distal area of small scales in Eu. hermanni, all smaller and very numerous in Eu. boettgeri [3]; (2) the fragmented, almost indistinct frontal scale; (3) the color pattern of the plastron with two parasagittal dark bands, each one whole or broken up [1,3,4]. 2.3. Morphological comparisons A cladistic analysis, previously performed and detailed elsewhere [20], includes, in the ingroup the species: Testudo turgaica, Agrionemys bessarabica, A. horsfieldii, A. kazachstanica Chkhikvadze, 1988; T. marmorum, T. marginata Schoepff, 1793, T. weissingeri Bour, 1995, T. antakyensis Perälä, 1996, T. kenitrensis Gmira 1993, T. graeca (s.l.) from the Maghreb, T. promarginata, Paleotestudo canetotiana, T. antiqua and the hermanni group (above mentioned valid species), and three outgroup taxa of terrestrial testudinids: Manouria impressa (Günther, 1882), Indotestudo elongata (Blyth, 1853) and ‘Er.’ bruneti Broin, 1977, a species attributed to the genus Ergilemys Chkikvadze, 1972 [7] sensu [5], Oligocene, La Milloque, France. Many Miocene and Oligocene fossil species, all insufficiently known, although potentially belonging to the lineage of Eurotestudo n.g., were disregarded. Among them, some Oligo-Miocene fragmentary specimens from France, attributed to ‘Ergilemys’ sp., have a hinge similar to that of Testudo s.s. only [5 (pl. 25, 28)], even in relatively young adults. The relationships of these specimens with ‘Er.’ bruneti and the Eurotestudo n.g. lineage are unclear. All the above taxa are Testudininei by characters given in the analysis [20]. The characters of the clades are present in some other Testudininei. Out of the 18 characters of the analysis, Testudo s.l. shares with Indotestudo and ‘Er.’ bruneti: (1) the coincidence of the costal-marginal scute sulci and the pleural-peripheral sutures and (2) the fusion of the two 12th marginals into a supracaudal. With ‘Er. bruneti’, Testudo s.l. shares the shell form: more elevated than that found in M. impressa, with elevated peripherals and marginals, arched with domed lateral pleural slopes 808 F. de Lapparent de Broin et al. / C. R. Palevol 5 (2006) 803–811 and with two anterior and posterior slopes meeting at a domed more or less flattened part, and more or less strong protuberances below the vertebrals 2 or 3 or 4 and eventually the costals; basically quadrangular and moderately wide, not looking narrow or round; posterior border moderately postero-laterally expanded in dorsal view. In Agrionemys, the shell is rounded and shortened at the level of the bridge, with an elevated bridge and more convergent plastral lobes and a larger entoplastron. In the T. marginata group, the shell becomes elongated (much postero-laterally expanded at the peripheral border), differently from I. elongata (postero-medially expanded). In the antiqua group, the shell widens. In Paleotestudo, the posterior border is not at all expanded. As a member of Testudo s.l., the Eurotestudo n.g. lineage shares with Testudo s.s. and Agrionemys particularly the following characters: (1) the posteriorly ascending dorsal epiplastral lip with a slightly convex surface: stopping its ascension abruptly (a) and, being more or less curved (b), located above the posterior surface of the epiplastron which is not thickened (c) – elements (a), (b), and (c) differentiate these chelonians from Indotestudo (Fig. 1D, figures in [5,9,20]) –; (2) the typically sinuous sulcus between the abdominal and femoral, with the latero-anterior sinuosity clearly extended on the hypoplastron and anterior to the inguinal notch (Fig. 1B); however, in Testudo s.s., the curve tends to be reduced, with the presence of the hypo-xiphiplastral hinge, particularly in the marginata group; in Agrionemys the hypoplastral overlap by the femorals is apparently more extensive, partly because the hypoplastron is shortened; (3) the possible posterior reduction of the series of eight neurals (Fig. 1A) to 7 or 6; this character is very rare in Eurotestudo n.g., but it is the norm in extant Testudo species (in time after the fossil species T. kenitrensis and T. marmorum and some fossil T. graeca from Morocco), and in the extant Agrionemys species (evolving after the fossil species bessarabica); (4) the ‘Testudo s.l.’ type of suprapygal–pygal, as opposed to the ‘geoemydine’ (in Manouria impressa) and ‘Geochelone’ (in Indotestudo and ‘Er.’ bruneti) types: both suprapygals constitute one trapezoid structure, with straight borders, in front of the pygal (Fig. 1C), that is completely elongated throughout its width and not only laterally as in the ‘Geochelone’ type (see [5,9,18] and other references included); consequently, the posterior border of the vertebral 5 is confluent with the limits of the suprapygal–pygal structure (complete coincidence of sutures and sulci); however, in extant Agrionemys, vertebral 5 is slightly shorter pos- tero-medially so that the supracaudal slightly covers the suprapygal (particularly in A. kazachstanica) with a sinuosity [9] and the vertebral 5 may overlap the pygal, often in Eurotestudo, sometimes in Testudo; (5) the narrowing of the lateral scute border on the dorsal epiplastron. In Testudo s.l., the suprapygals (two in general) are divided by a semicircular (primitively) or a semicircular–semitransversal, or a transverse line, according to the following evolutionary stages; the most derived stage is the fusion of the suprapygals into a trapezoid. The fusion of the suprapygals into a trapezoid with a posterior straight border is mostly known in Eurotestudo n.g., although it also occurs rarely in some species of Testudo and Agrionemys. The three genera evolved, in parallel, the following homoplastic characters (that are in general very frequently witnessed in Testudininei): (1) the partial to complete reduction of the dorsal cervical (constantly or occasionally present in a population); (2) a tendency for the pectorals to extend medioanteriorly toward the entoplastron and onto the entoplastron (without meeting each other anteromedially), more or less frequently according to population, and not only in Agrionemys and Eurotestudo n.g. (particularly in the boettgeri, Lunel-Viel and Soave populations), from which taxa this character is well known, but also in T. graeca [9]. With Testudo, Eurotestudo n.g. shares an epiplastral lip that curves onto the entoplastron, overhanging the dorsal surface. Below this, there is a depressed gular pocket. A tendency toward a gular pocket is obvious in Eurotestudo n.g.: a narrow and weak gular pocket is particularly found in the Lunel-Viel population, and one is often present in P. canetotiana. P. canetotiana (figures in [5,16]) is considered as belonging to the Eurotestudo n.g. lineage despite its similarity with Testudo. The differentiating conditions are the acquisition in Testudo of a characteristic hinge, in both sexes, between the hypo-and xiphiplastra, with (a) a correlative elongation of the posterior lobe, (b) the fusion of the lateral extremities of the suture (at the hinge) and of the abdomino-femoral sulcus (except in juveniles and in the small-sized T. kenitrensis), and (c) the tendency to shorten the femorals on the hypoplastron (particularly in the marginata group). In Testudo, the gular pocket is constant, small to strong [6,9,23,24], except in T. antakyensis Perälä, 1996 (the lip is often not even curved; Fig. in [9] as T. terrestris Forsskål, 1775). In Agrionemys (figure in [9]) (unknown in A. bessarabica), the epiplastral lip is never curved up to entoplastron, as in fossils of the Eurotestudo n.g. lineage (‘T.’ antiqua, F. de Lapparent de Broin et al. / C. R. Palevol 5 (2006) 803–811 Eu. pyrenaica, figures in [5,32]) and the epiplastral–entoplastral surface is rarely depressed into a gular pocket. On the other hand, the ventral surface of the gulars in relief can also be found in Agrionemys, but never in Testudo. The process of modification of the suprapygal area, with the tendency to complete fusion of the plates, is most achieved in Eurotestudo n.g., making it unique among Palaearctic forms (present in some African small endemics [18]). Meanwhile, the process of fusion of the last neurals is more achieved in the two other genera, although homoplastically because the fusion progressively develops in each lineage separately. In the lineage of Eurotestudo n.g., there is no single species accumulating all the more derived states of the homoplasies. P. canetotiana has a higher tendency towards a gular pocket. The species have their particularities: ‘T.’ antiqua has a wider shell [32]; ‘Eu.’ pyrenaica has a triangular or trapezoid notch at the nuchal, not affecting the adjacent peripherals and the cervical is completely lacking, as in the Upper Miocene ‘T.’ amberiacensis Depéret, 1894, France [5], which might belong to a pyrenaica group if it is confirmed that it belongs to Eurotestudo. The Eu. aff. hermanni populations from Lunel-Viel and Soave are more derived by the progression of the pectorals on the entoplastron. The Soave population and some elements from the Quaternary of the Iberian Peninsula (references in [11,19,20] have the most developed epiplastral lip, thick and often very protruding, but lacking a gular pocket. The extant Eu. hermanni is most advanced by having a trapezoid suprapygal and the more consistently divided supracaudal (externally and internally), and perhaps also by featuring the very occasional presence of seven neurals (the series is not well enough known in other populations to make comparisons). Eu. boettgeri has the femorals much shortened. Concerning fossil relatives of Eurotestudo n.g., Paleotestudo canetotiana is considered as belonging to the Eurotestudo lineage by the more advanced fusion of the femoral trochanters and its full aspect; some specimens have the gulars in relief ventrally and one has a hexagonal pygal. This is also congruent with its geographical context. As seen above, the species is also rather similar to Testudo graeca (s.l.) except for the absence of hinge and shorter posterior lobe. But it lacks the protuberances and the posteriorly expanded border of the shell. P. canetotiana is the first European form which presents the most derived evolutionary state of the anterior lobe shape: the trapezoid lobe with anteriorly prominent gulars, well laterally exposed, becomes widened at the anterior border and the gulars 809 do not participate in the lateral borders; presently the gulars occupy either only the complete anterior border of the lobe, or a narrow slice (protruding or not) in its medial part, the humeral lateral borders being rounded. This morphology is also present in Testudo s.s. (always) and in A. kazachstanica and some A. horsfieldii, but not in ‘T.’ turgaica and A. bessarabica. However, in Agrionemys, the anterior lobe always has more converging lateral borders.‘T.’ antiqua belongs more confidently to the Eurotestudo n.g. lineage: tendency to feature a divided supracaudal, possibility for having fused suprapygals, general aspect of the shell which is of a quadrangular form, posteriorly expanded according to the norm in Eurotestudo n.g. contrarily to Paleotestudo. The epiplastral lip is wide and long, never curved up to the entoplaston and there is no gular pocket. The shell is particularly wide (width/length). Its femur is undescribed, and the possible fusion of the suprapygals needs to be confirmed. A revision of ‘T.’ antiqua (as for some close fossil species) is necessary to reconsider its phylogenetic position with respect to its possible integration into the genus [20]. 2.4. External characters indicating generic status The extant species in the new genus Eurotestudo are unique among the Testudinidae by the following characters: (1) the scalation of the front face of the forearm includes a distal area of small and irregular (small scales in Eu. hermanni, very small and numerous in Eu. boettgeri), while there are only large and regular scales in other Testudinidae; (2) the frontal scale is fragmented, almost indistinct, while most tortoises have a large and well-delimited frontal, following two elongated prefrontals; (3) the color pattern of the plastron: from the basic pattern of postero-lateral dark spots, originating from the areolar zone (basically radiated), develops a system of parasagittal dark bands that is unknown among other chelonians. Another external character supports the separation of the new genus Eurotestudo: the thigh tubercle (‘thigh-spur’) may constitute a basic autapomorphy in the extant Testudo sp.; it is absent in Agrionemys and Eurotestudo. In return, the color pattern of the dorsal carapace of the ‘Testudo’ type may constitute a basic synapomorphy linking Eurotestudo–Testudo s.s. Some characters (previously considered as synapomorphies) are actually weakly homoplastic: – the caudal spur is moderate in Agrionemys, strong and lengthened in Eurotestudo and very small in T. kleinmanni plus T. werneri. It is present and morpho- 810 F. de Lapparent de Broin et al. / C. R. Palevol 5 (2006) 803–811 logically variable in various lineages of terrestrial Testudinidae, but also in other chelonians (Kinosternon, the extinct Meiolania); – the reduction of the fingers of the hand, considered as shared by Agrionemys and Eurotestudo n.g. is neither a synapomorphy nor a homoplasy: it is not the same character: ● four fingers in Agrionemys; ● five fingers, but nails 1 and/or 5 are often reduced, rarely absent in Eu. hermanni s.s. Testudo has primitively five fingers [1,3,4,9,23,24,27]. Acknowledgements To the Academic Kippis Society; Dr M.F. Bonifay & Prof. E. Bonifay (Marseille, France); Dr S. Calzada Badia, Dra A. Masriera (Barcelona, Spain); Dr M. Delfino (Florence, Italy); Dr A. Rhodin (Lunenburg, MA, USA); Dr L. Sorbini †, Dra A. Vaccari (Verona, Italy); to Dr M. Pickford (English correction); Dr J.-C. Rage and Prof. P. Taquet (review of the note). References [1] 3. Discussion and conclusion None of the approaches, either morphological [9,23, 24] or molecular [14,22,etc.], provides a strong hypothesis of inter-relationships of the lineages or within species in the genera. The results differ according to the authors, the taxa included, the type and amount of genetic material, the number of specimens and the method employed. The hermanni group (including Eu. hermanni alone or with Eu. boettgeri) may be sister to Agrionemys [9,14] or else with Indotestudo or others [22]. Or it may be placed as the sister of a clade with Agrionemys and Testudo [24]. In recently constructed cladograms [20], the three genera are well separated after ‘Er’. bruneti in every hypothesis: ‘T.’ promarginata is either the sister group of both Testudo s.s. and Eurotestudo or the sister taxon of the three genera in politomy if the poorly preserved ‘T.’ turgaica is excluded from the analysis. Even if the exact link-point between the three lineages is not definitely established, their separation and differentiation is well established. We can hypothesize that the shared origin of Testudo and Eurotestudo n.g. is more probable than that of Eurotestudo n.g. and Agrionemys, a previously proposed hypothesis [9]: according to the new analysis [20], similarities remaining between Eu. hermanni and extant Agrionemys spp. are mostly primitive. Beside the derived more curved epiplastral lip, shared by Testudo and Eurotestudo n.g, the derived color pattern of the dorsal carapace of the ‘Testudo’ type, shared by extant species, is also significant. In any case, various homoplastic characters have evolved asynchronously in these three lineages. Although the relationships between Eurotestudo n.g., Testudo, and Agrionemys cannot be firmly established, these lineages are clearly and consistently separated according to all the approaches. Whatever their inter-relationships may be, the common origin of the three genera is in Asia before the Oligocene. [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] R. Bour, L’identité des Tortues terrestres européennes : spécimens-types et localités-types, Rev. Fr. Aquariol. 13 (4) (1987) 111–122. R. Bour, Introduction, Manouria 7 (22) (2004) 2. R. Bour, Testudo boettgeri, Manouria 7 (22) (2004) 9–10. R. Bour, A new character for the identification of populations of the Hermann’s tortoise, Testudo hermanni Gmelin, 1789 (Chelonii, Testudinidae), Salamandra (Frankf.) 40 (1) (2004) 59–66. F. de Broin, Contribution à l’étude des Chéloniens. Chéloniens continentaux du Crétacé et du Tertiaire de France, Mém. Mus. natl Hist. nat. Paris, Ser. C 38 (1977) i–ix, 1–366. V.M. Chkhikvadze, [Sur l’origine des tortues terrestres paléarctiques actuelles], Bull. Acad. Sci. Georgian SSR 57 (1) (1970) 245–247 (in Russian). V.M. Chkhikvadze, Sur la position systématique des Tortues terrestres géantes du Tertiaire du Paléarctique, Soovshch, Akad. Nauk. Gruznuskoj. 65 (3) (1972) 745–748 (in Russian). V.M. Chkhikvadze, in: Les tortues fossiles du Caucase et du Nord de la mer Noire, Metzniereba, Tbilissi, 1983, pp. 1–149 (in Russian). S. Gmira, Étude des Chéloniens fossiles du Maroc. Anatomie. Systématique. Phylogénie, Cahiers de Paléontologie, Éditions du CNRS, Paris, 1995 (140 p.) S. Hervet, Les Tortues du Quaternaire de France : critères de détermination, répartitions chronologique et géographique, Mésogée 58 (2000) 3–47. E. Jiménez Fuentes, J.-L. Cardoso, E.G. Crespo, Presencia de Agrionemys (=Testudo) hermanni (Gmelin, 14789) en el Paleolítico medio de la Gruta Nova da Columbeira (Bombarral, Provincia de Estremadura, Portugal), Stud. Geol. Salmant. 34 (1998) 123–139. L.I. Khozatsky, M. Mlynarski, Agrionemys – Nouveau genre de tortues terrestres (Testudinidae), Bull. Acad. Pol. Sci., II, Ser. Sci. Biol. 14 (2) (1966) 123–125. T. Kotsakis, Révision des tortues (Emydidae, Testudinidae, Trionychidae) du Plio-Pléistocène de Valdarno supérieur (Toscane, Italie), Quaternaria 32 (1980) 11–37. A.C. van der Kuyl, D.L. Ballasina, †J.T. Dekker, J. Maas, R.E. Willemsen, †J. Goudsmit, Phylogenetic relationships among the species of the genus Testudo (Testudines: Testudinidae) inferred from mitochondrial 12S rRNA gene sequences, Mol. Phylogenet. Evol. 22 (2) (2002) 174–183. F. de Lapparent de Broin, African chelonians from the Jurassic to the Present. A preliminary catalog of the African fossil chelonians, Paleontol. Afr. 36 (2000) 43–82. F. de Lapparent de Broin, Les Chéloniens de Sansan, Mém. Mus. natn. Hist. nat., Paris 183 (11) (2000) 219–261. F. de Lapparent de Broin et al. / C. R. Palevol 5 (2006) 803–811 [17] F. de Lapparent de Broin, The European turtle fauna from the Triassic to the Present, Dumerilia 4 (3) (2001) 155–216. [18] F. de Lapparent de Broin, The Miocene chelonians from the southern Namibia, B. Senut, M. Pickford (Eds.), Faunas from the southern Namibia, Mem. Geol. Surv. Namibia 19 (2003) 67–102. [19] F. de Lapparent de Broin, M.T. Antunes, Pleistocene Chelonians from Gruta da Figueira Brava (Arrábida, Portugal), Mem. Acad. Cienc. Lisb. Cl. Cienc. 38 (2000) 101–112. [20] F. de Lapparent de Broin, R. Bour, J. Perälä, Definition of the genus Eurotestudo (Testudinidae, Chelonii): a morphological approach, Ann. Paleontol (in press). [21] M. Mlynarski, Tortoises from the Pliocene of Poland, Acta Geol. Pol. 5 (2) (1955) 161–214 (Consp., 46–62). [22] J. Parham, J.R. Macey, T.J. Papenfuss, C.R. Feldman, O. Türkozan, R. Polymeni, J. Boore, The phylogeny of Mediterranean and their close relatives based on complete mitochondrial genome sequences from museum specimens, Mol. Phylogenet. Evol. 38 (2006) 50–64. [23] J. Perälä, The genus Testudo Linnaeus, 1758 sensu lato (Testudines: Testudinidae): Phylogeny, Taxonomy, Conservation, Ph. D. thesis, University of Bristol, School of Biological Sciences, 2002 (i–xiv + 328 p.) [24] J. Perälä, The genus Testudo (Testudines: Testudinidae): Phylogenetic inferences, Chelonii 3 (2002) 32–39. [25] J. Perälä, Biodiversity in relatively neglected taxa of Testudo L., 1758 s.l, Chelonii 3 (2002) 40–53. 811 [26] J. Perälä, Morphological variation among Middle Eastern Testudo graeca L., 1758 (sensu lato), with a focus on taxonomy, Chelonii 3 (2002) 78–108. [27] J. Perälä, Occurrence and taxonomic significance of thigh-spurs in Testudo marginata Schoepff, 1792 and Testudo weissingeri Bour, 1995, Herpetozoa 14 (3/4) (2002) 123–126. [28] J. Perälä, Testudo hercegovinensis, Manouria 7 (22) (2004) 19– 20. [29] A. Pieh, J. Perälä, Variabilität der Maurischen Landschildkröten (Testudo graeca Linnaeus, 1758 – Komplex) im zentralen Marokko mit Beschreibung zweier neuer Taxa, Herpetozoa 17 (1/2) (2004) 19–47. [30] A.N. Riabinin, [Sur les tortues des dépôts méotiens de Bessarabie], Trudy Geol. min. Muz. Petra Velikago Imper. Akad. Nauk 1 (1915–1918) 1–16 (in Russian). [31] A.N. Riabinin, [Testudo turgaica nov. sp. du Miocène moyen de la région de Turgai], Trav. Mus. Géol., Acad. Sci. URSS 1 (1926) 53–62 (in Russian). [32] H.H. Schleich, Neogene Testudines of Germany. Their stratigraphic and ecological evaluation, Stud. Palaeochel. I (1985) 249–267. [33] L. Sorbini, M.V. Durante Pasa, Le collezioni paleontologiche quaternarie del Museo Civico di Storia Naturale di Verona, Origine, inventario, bibliografia, Mus. Civ. Stor. Nat. Verona, Ser. Cataloghi 1 (1974) 1–53. [34] H.-k. Yeh, Fossil turtles of China, Paleontol. Sin. N.S.C. 150 (18) (1963) 1–112.