Agrionemys sp. (Fig. 3) LOCALITY AND AGE. — Sherullah 9, Khordkabul basin, Afghanistan, late Miocene, late Vallesian-basal Turolian transition. MN10/11. MATERIAL EXAMINED. — AFG 130, left epiplastron (Fig. 3F); AFG 131, peripheral plate, fragment without distal border (Fig. 3J); AFG 132, pygal plate fragment (Fig. 3A); AFG 133, peripheral plate, fragment with free border; AFG 134, peripheral plate, fragment (Fig. 3K); AFG 135, neural 5 (Fig. 3C); AFG 136 [+ 138 (10)], fragmentary right xiphiplastron (Fig. 3G, H 2); AFG 137, plate fragment; AFG 138 including 10 specimens: 138(1) and 138(2), peripheral plate fragments with preserved distal border; 138(3), peripheral plate fragment without the border; 138(4) (5) (6) (7) (8), plate fragments; 138(9), fragmentary left epiplastron (Fig. 3D); 138(11) fragment of a xiphiplastral anal part (Fig. 3I, H 1). DESCRIPTION The specimens belonged to several individuals of moderate size (shells c. 17-19 cm long). The fragmentary right xiphiplastron AFG 136 (+ 138[10])(Fig. 3G, H 2) (4.4 cm long) belonged to a plastron of the size of an extant Agrionemys horsfieldii (Gray, 1844) (specimen REP 59) from Molayan, Khordkabul basin of Afghanistan [“ Kabul, Afghanistan ”, being the type locality of the living species (Iverson 1992)], which is 15.7 cm long, and it was barely smaller than the plastron of the epiplastron AFG 130 (Fig. 3F) (3.2 cm wide). The left epiplastral fragment AFG 138(9) (Fig. 3D) is barely larger than AFG 130. The fragment of xiphiplastral anal part AFG 138 (11) seems to have belonged to an individual as large as the fragment AFG 136 (138[10]) (Fig. 3H). The neural 5 (Fig. 3C) corresponds to a dorsal shell of an A. kazachstanica Čhkhikvadze, 1988 specimen (MNHN.RA) of 17. 3 cm long, also corresponding to the plastral length of REP 59 and xiphiplastron AFG 136. It was from a slightly smaller individual than the pygal AFG 132 (this maybe of a carapace long of 18.5-19 cm). The texture of the bones is typical for testudinines: the external surface is apparently relatively smooth with minute points and small granulations. The morphology of the best preserved scute sulci is also typical for testudinines: a fine acute groove limited on each border by a crest, which is especially clear on AFG 130 (Fig. 3F 2) for the gularo-humeral sulcus. Most of the fragments do not provide any particular information within testudinines; their identification is inferred from their conjunction, the best preserved ones being figured, conforming to the genus Agrionemys. The neural is a neural 5, short, wider than long, hexagonal. It has short sides in front and anterior borders wider than posterior borders. It is anteriorly transversally crossed by the sulcus between vertebrals 3 and 4, at the level of its smaller anterolateral sides. The preserved peripherals are all broken. The plates are massive, thick for their length. Two of them are figured (AFG 131, Fig. 3J and AFG 134, Fig. 3K). They are slightly dorsally incurved, being posterior peripherals, probably the 9 th and 10 th. The fragmentary pygal (AFG 132, Fig. 3A) is long, thick, slightly convexe (Fig. 3A3) from right to left and concave interiorly, apparently being from a male (Fig. 3A 2, B). The unique supracaudal inner overlap is very long, nearly reaching the suprapygal-pygal suture, as in Agrionemys living specimens and in a Maragheh fossil specimen (Iran, Turolian) (Maragha in Lapparent de Broin et al. 2006b, c). In the Fig. 3B, the fossil Afghan partial plate is superposed on the fossil Maragheh pygal The left epiplastron AFG 130 (Fig. 3F) is slightly broken, missing a small posterolateral angle at the suture with the hyoplastron. Its posterolateral external surface part is smooth due to a worn surface, then artificially limited by a line nearly parallel to the gularohumeral sulcus. The specimen is relatively wide for its full length. Its anterior border is transversal, with a wide gular border which is barely protruded anteriorly to the humerals. The oblique inclination of the specimen lateral border indicates the anterior lobe was trapezoid. The lateral borders are moderately converging, as in the drawing of the Fig. 3E (drawn on the specimen REP57 of extant A. horsfieldii from Khordkabul basin) and not rounded. The dorsal epiplastral lip is wide, slightly shorter than wide anteriorly, dorsally convex and thick (Fig. 3F 1, F 3, F 4), and it rather abruptly overhangs the posterior epiplastral surface, with a very minute gular pocket, anterior to the entoplastral suture (Fig. 3F 1, F 3): the wider than long posterior lip border is convex and ends in front of the entoplastron. Ventrally, the gular scute is as long as wide and it was barely prolonged on the entoplastron. The gular is not anteriorly protruding and its surface does not protrude in relation to the humeral. The gularohumeral sulcus is straight, not being anterioly slightly incurved, as it is sometimes the case on one side in Chersine and in living Agrionemys. On the whole, the anterior lobe of AFG 130 was not wide on each side of the wide entoplastron and not much anteriorly narrowed. The other left (fragmentary) epiplastron AFG 138(9) (Fig. 3E) is a fragment at the boundary of the gular and the humeral, at the lateral beginning of the dorsal epiplastral lip: it is thick, and the lateral beginning of the dorsal lip shows it was well elevated anterior to the entoplastron (Fig. 3D3) and perhaps the gular was more anteriorly protruding than in AFG 130; the anterior lobe borders were more anteriorly converging than in AFG 130. The xiphiplastron AFG 136 (Fig. 3G) is preserved by the lateral part only. It is anteriorly limited by the hypoxiphiplastral suture which was not transformed in a hinge. It is referred to testudinines by the strong dorsal elevation of the bone below the dorsal overlap of the femoral, much raising from the external border towards the inner scute limit (Fig. 3G 2), and from the posterior extremity up to the hypoxiphiplastral suture (Fig. 3G3); this elevation had to follow on the hypoplastron to end in the inguinal notch without a hinge. The xiphiplastral posterior anal extremity is angular (95°). Associating the two xiphiplastral extremities (that of AFG 136 and that of AFG 138[11]), which are convenient for a same individual size, the anal notch was a wide and short triangle: this is shown when superposing (Fig. 3H) these two specimens on the xiphiplastra of the extant Agrionemys horsfieldii (specimen REP 57) of Khordkabul basin. The xiphiplastral had to be approximately anteriorly as wide (at hypoxiphiplastral suture) as long (medially), with lateral borders much converging to the anal notch, as in living Agrionemys. As in Agrionemys, the anal scute ventral overlap was less long than the femoral overlap part on the xiphiplastron. COMMENTS The specimens clearly show that the turtle from Sherullah 9 belonged to the group of Palaearctic terrestrial testudinids, the Testudininae, integrated in the past in the waste basket genus “ Testudo Linnaeus, 1758 s.l. ” (as in Broin 1977). The group includes the living – Testudo (s.s.), type-species T. graeca Linnaeus, 1758, – Chersine Merrem, 1820, type species Testudo hermanni Gmelin, 1789 [or Eurotestudo Lapparent de Broin, Bour, Parham & Perälä, 2006 (see Bour & Ohler 2008)] and Agrionemys, type species Testudo horsfieldii Gray, 1844 (see Lapparent de Broin et al. 2006a, b, c) (among which genus are notably the species A. horsfieldii and A. kazachstanica) (Iverson 1992). Western fossil species were included in Testudo (s.l.) among which those attributed to the fossil genus Paleotestudo Lapparent de Broin, 2000. This has for type species Testudo canetotiana Lartet, 1851 from the middle Miocene of Sansan (France). Sherullah 9 material is here particularly compared with living specimens (REP and MNHN.RA coll.). It is also compared with fossil material from Maragheh (Iran, MNHN.F. MAR2424, MAR2425, 1905-10 coll.), late Miocene, early Turolian. This is approximately of the same age as Molayan (Sen 1998). Maragheh material has been attributed without comments to Agrionemys cf. bessarabica Riabinin, 1918 in Lapparent de Broin et al. (2006b, c) and it is confirmed here as being at least an Agrionemys sp. Sherullah 9 material was already mentioned as belonging to Agrionemys in Lapparent de Broin et al. (2006b, c) (without description). During the late Miocene, the testudinines are well diversified in the Eurasiatic area in the three genera Testudo, Chersine and Agrionemys (see Lapparent de Broin et al. 2006a, b, c). Testudo is distinguished from both others by the derived presence of a hinge between the xiphiplastron and the hypoplastron, in males as in females, with a longer posterior lobe only composed of the xiphiplastra, being narrower with less converging lateral borders. Here, the xiphiplastron AFG 136 is clearly not hinged as in Agrionemys and Chersine. Agrionemy s has a rather wide posterior lobe at inguinal notches, as Chersine, eventually wider according to the species and subspecies and sex, and this conforms to Sherullah 9 partial xiphiplastron (Fig. 3H); in conjunction with the wide posterior lobe at its base, the shape of the carapace of living Agrionemys is wide. It may be wider for its length according to the species or subspecies and sex but, on the whole, it is wider than in Testudo. This cannot be confirmed for Sherullah 9, the carapace plates being too poorly preserved, but the xiphiplastra are congruent for this greater width. Furthermore, Agrionemy s has a short dorsal gular lip anterior to the entoplastron and the gular pocket is always slight, as in Sherullah 9 specimens. Although also anterior to the entoplastron, the epiplastral gular lip is variably long and deep in Chersine (such a variability being known in Plio-Pleistocene Ch. hermanni spp. - subspp. group, in Lapparent de Broin et al. 2006a, b, c). In Testudo, the pocket is most often larger and overlapped by a long dorsal epiplastral lip, often extending over the entoplastron. The discriminant character of Chersine is a derived secondary division of the supracaudal scute, either only dorsally or only ventrally (in fossil species and subspecies) or both dorsally and ventrally in living representatives. The pygal from Sherullah 9 AFG 132 does not share this character: there is only one supracaudal as in Testudo and Agrionemys (Lapparent de Broin et al. 2006a, b, c). Conforming to Agrionemys: the neural 5 AFG 135 conforms to that observed in extant compared Agrionemys spp. specimens. It is either shorter for its width and/or more hexagono-trapezoid and less rectangular or quadratic than in Testudo and Chersine observed specimens. However, the sulcus V3-V4 position is variable and here it is anterior to the middle length. The pygal is relatively long, according to males of Agrionemys (but also in males of Chersine) but it is much thicker than in living specimens. The single supracaudal ventral overlap of AFG 132 is long as in Agrionemys male specimens (Fig. 3B: AFG 132 pygal superposed on Maragheh pygal). The lateral borders of the plastral lobes are more or less convex in Agrionemys, including the straighter condition of Sherullah 9. In the living species, the anterior lobe is always wide at its base (in relation to the wide carapace) but it may be more or less anteriorly narrowed and protruding at gular lip: there is a bony shape variability which is not yet clearly defined in its relation to sex, individual and speciation. The epiplastron AFG 130 indicates an anterior lobe with moderately converging lateral borders anteriorly, and it is anteriorly straight with wide gulars but without a protruding lip in relation to Agrionemys specimens. However, AFG 138(9) gulars might have been slightly protruding in relation to humerals at least as in some living specimens (Fig. 3E). The lateral posterior lobe borders of AFG 136 are barely rounded with a small femoroanal indentation as in the specimen of A. horsfieldii of Khordkabul basin REP 57. AFG 136 is superposed on extant REP 57 in Fig. 3H, this being presented with its scutes: the scutes are visible lateroposteriorly exceeding the bone, in black. Sherullah 9 AFG 136 and 138(11) share with A. horsfieldii a short and straight anal notch. Other characters which can allow Agrionemys and Chersine to be distinct from Testudo, such as femur trochanter junction (Lapparent de Broin & Antunes 2000) are not represented here. Similarly discriminant, but not fully preserved here, is the carapace shape which is rounded and wide in living Agrionemys, and which is particularly short for its length in known A. horsfieldii specimens (type species of the genus). Also discriminant is the fact the Agrionemys shell is much higher than in Chersine and Testudo, with a higher bridge and a more elevated anterior lobe extremity. However, the shell is lower in A. horsfieldii than in A. kazachstanica (Lapparent de Broin et al. 2006a: fig. 11c). Testudo and Chersine lineages have a more western distribution (around Mediterranean countries; Chersine being only European) compared to living Agrionemys which ranges from Georgia to western Xinjiang (eastern China) and is mostly present in eastern Asia (Iverson 1992; Rhodin et al. 2017). The three lineages have been separated for a long time. Testudo (s.s.) is diversified since the late Miocene, being present in the Vallesian of Greece at the locality “Ravin de la Pluie” [MN10] (Garcia et al. 2011) and the Tortonian of Greece (Gaudry 1862, 1862-1867; Georgalis & Kear 2013; Garcia et al. 2016) at Nikiti [MN 11] and Pikermi, Turolian [MN 12-13]) and seemingly at the same time in Northern Africa (Djebel Semène, Late Vallesian, Tunisia) (Gmira et al. 2013). This wide area indicates a much older diversification. Chersine is present with the double supracaudal (yet not dorsally and ventrally stabilised as in living specimens) in the Pliocene (Ruscinian, MN 15) of Perpignan (Ch. Pyrenaica (Depéret & Donnezan, 1890)), and of Poland. Chersine lineage is considered as dating back to the middle part of the Miocene at Langhian-Serravallian epochs, being considered as firstly represented by the species Paleotestudo canetotiana from Sansan (middle Miocene, early Astaracian, MN6) (Broin 1977; Lapparent de Broin 2000, 2001). The lineage of Paleotestudo is here considered sensu Lapparent de Broin et al. 2006a, b, c, being grounded on this oldest Paleotestudo species (P. canetotiana), which is not considered as a synonym of T. antiqua Bronn, 1831 sensu Corsini et al. 2014 and of T. catalaunica Bataller, 1926 sensu Pérez-García et al. 2016 (MN 7-8, late Astaracian). Notably, the contour of the anterior lobe and the shape of the epiplastral gular lip of these species are different, and whatever the individual considered variability. However, these species are all considered as part of the same larger lineage leading to Chersine (Lapparent de Broin et al. 2006a, b, c; Luján et al. 2016; Pérez-García 2017), being distinguished from the lineage of Agrionemys. The first known representative of Agrionemys lineage seems to be Testudo bessarabica (Lapparent de Broin 2000, 2001; Lapparent de Broin et al. 2006a, b, c; Macarovici 1930; Riabinin 1918) from the late Miocene (Turolian, MN 12) of Taraklia and Ciobirciu (Bessarabia, Moldavia). This species is the type species of Protestudo Č khkikvadze, 1970, for us a potential junior synonym of Agrionemys. It was recombined as Agrionemys bessarabica by Lapparent de Broin et al. (2006b). It has the same high shell due to a high bridge and wide and short xiphiplastron, in a posterior lobe with posteriorly well converging, straight or slightly convexe, lateral borders. The narrow anterior lobe has much converging borders and much protruded gulars. As Agrionemys, it has a long ventral gular overlap on the entoplastron, however longer with regard to Sherullah specimen AFG 130. Agrionemys bessarabica differs from living Agrionemys by the more posterior humeropectoral sulcus which is just posterior or clearly posterior to the entoplastron and not at all overlapping it: this is missing in Sherullah 9, but this is a character also known in some Chersine and Testudo specimens (Gmira 1995; Gmira et al. 2013; Lapparent de Broin et al. 2006b) and that is then potentially specific rather than generic. Agrionemys bessarabica much differs by the gular lip which is anteriorly more narrowly protruding with converging lateral borders, as seen in the four ventrally figured specimens of Riabinin (1918: figs 3-4). This is different from Sherullah 9 AFG 130 specimen (Fig. 3E) which is devoid of gular protrusion as in some Agrionemys horsfieldii and A. kazachtanica living specimens. While, several other preserved living A. horsfieldii specimens from Khordkabul basin present more protruding gulars (REP coll.); they are then closer to A. bessarabica although the protrusion is less long, less anteriorly narrow and externally rounder. The entoplastron seems considerably larger than in Chersine and Testudo, at least it is longer in A. horsfieldii and in A. bessarabica known specimens. The entoplastron is visible in four of the five figured specimens of A. bessarabica (Riabinin 1918: figs 3-4), appearing as more important in the anterior lobe, which is narrower on each side of this entoplastral bone (as in Sherullah 9 AFG 130), but anteriorly pointed: in Sherullah 9 AFG 130, the widely curved posteromedial border of the epiplastron (sutured with the entoplastron) in relation to its lateral width and the symmetrical juxtaposition Fig. 3F 1, F 2, seems also to indicate the presence of a wide entoplastron. Agrionemys bessarabica is different from both A. horsfieldii spp. and Sherullah 9 species by the deeper anal notch. Agrionemys sp. (or previously cf. bessarabica) is considered as present at Maragheh, Iran (late Miocene, early Turolian), by the specimen MNHN.F. MAR2424, MAR2425 (coll. Mecquenem-Morgan 1905-10) (Lapparent de Broin et al. 2006c). In this country, as in Afghanistan, lives A. horsfieldii and not Chersine. The fossil is preserved by a part of posterior carapace of a male with the curved suprapygal-pygal area (Fig. 3L), and a fragment of hypoplastron; this specimen shows the inguinal notch bears a triangular medial inguinal scute which is well separated from the sulcus abdominofemoral, itself well separated from the hypoxiphiplastral suture (Fig. 3M). As Chersine, A. bessarabica and living Agrionemys, the Maragheh specimen differs from Testudo by the primitive absence of hypoxiphiplastral hinge, the suture being well separated from the abdominofemoral sulcus. It also conforms to A. bessarabica and living Agrione, Published as part of Lapparent, France de, Bailon, Salvador, Augé, Marc Louis & Rage, Jean-Claude, 2020, Amphibians and reptiles from the Neogene of Afghanistan, pp. 409-426 in Geodiversitas 42 (22) on pages 412-417, DOI: 10.5252/geodiversitas2020v42a22, http://zenodo.org/record/4447648, {"references":["IVERSON J. B. 1992. - A Revised Checklist with Distribution Maps of the Turtles of the World. Privately Print, Richmond, Indiana, 363 p.","LAPPARENT DE BROIN F. DE, BOUR R. & PERALA J. F. 2006 b. - Morphological definition of Eurotestudo (Testudinidae, Chelonii): First part. Annales de Paleontologie 92 (3): 255 - 304. https: // doi. org / 10.1016 / j. annpal. 2006.07.001","BROIN F. DE 1977. - Contribution a l'etude des Cheloniens. Cheloniens continentaux du Cretace et du Tertiaire de France. Museum national d'Histoire naturelle, Paris, 115 p. (Memoires du Museum national d'Histoire naturelle, Ser. C - Sciences de la Terre; 38).","MERREM B. 1820. - Versuch eines systems der Amphibien. Vol. 8. J. C. Krieger, Marbuch, 191 p.","LAPPARENT DE BROIN F. DE, BOUR R., PARHAM J. F. & PERALA J. F. 2006 a. - Eurotestudo, a new genus for the species Testudo hermanni Gmelin, 1789 (Chelonii, Testudinidae). Comptes Rendus Palevol 5: 803 - 811. https: // doi. org / 10.1016 / j. crpv. 2006.03.002","BOUR R. & OHLER A. - M. 2008. - Chersine Merrem, 1820 and Chersina Gray, 1831: a nomenclatural survey. Zootaxa 1752: 66 - 68. https: // doi. org / 10.11646 / zootaxa. 1752.1.4","SEN S. 1998. - The age of the Molayan mammal locality, Afghanistan. Geobios 31: 385 - 391. https: // doi. org / 10.1016 / S 0016 - 6995 (98) 80021 - 1","RIABININ A. N. 1918. - [On the turtles from Meotian deposits of Bessarabia]. Trudy Geologiceskogo I Mineralogiceskogo Muzejaimeni Petra Velikago Imperatorskoj Akademii Nauk (St Petersburg) (1915) 1: 1 - 16 (in Russian).","LAPPARENT DE BROIN F. DE & ANTUNES M. T. 2000. - Pleistocene Chelonians from Gruta da Figueira Brava (Arrabida, Portugal), in TELLES ANTUNES M. (ed.), Colloquium: Last Neanderthals in Portugal. Odontologic and other evidence, January 18 th 1999. Memorias da Academia das Ciencias de Lisboa, Classe de Ciencias 38: 101 - 112.","RHODIN A. G. J., IVERSON J. B., VAN DIJK P. P., SAUMURE R. A., BUHLMANN K. A., PRITCHARD P. C. H. & MITTERMEIER R. A. 2017. - Turtles of the world. Annotated Checklist and Atlas of Taxonomy, Synonymy, Distribution, and Conservation Status. 8 th ed. TTWG, Conservation Biology of Freshwater Turtles and Tortoises. Chelonian Research Monographs 7, 292 p. https: // doi. org / 10.3854 / crm. 7. checklist. atlas. v 8.2017","GARCIA G., LAPPARENT DE BROIN F. DE, BONIS L. DE, KOUFOS G. D., VALENTIN X., KOSTOPOULOS D. S. & MERCERON G. 2011. - A new terrestrial Testudinidae from the Late Miocene hominoid locality \" Ravin de la pluie \" (Axios valley, Macedonia, Greece). 9 th European Association of Vertebrate Palaeontologists Annual Meeting, 14 - 19 June 2011, Heraklion Crete, Greece, Abstract, 1 p.","GEORGALIS G. L. & KEAR B. P. 2013. - The fossil turtles of Greece: An overview of taxonomy and distribution. Geobios 46: 299 - 311. https: // doi. org / 10.1016 / j. geobios. 2013.05.001","GMIRA S., LAPPARENT DE BROIN F. DE, GERAADS D., LEFEVRE D., MOHIB A. & RAYNAL J. - P. 2013. - Les Tortues du Pliocene d'Ahl al Oughlam (Casablanca, Maroc) et de localites avoisinantes. Geodiversitas 35 (3): 691 - 733. https: // doi. org / 10.5252 / g 2013 n 3 a 9","CORSINI J. A., BOHME M. & JOYCE W. G. 2014. - Reappraisal of Testudo antiqua (Testudines, Testudinidae) from the Miocene of Hohenhowen, Germany. Journal of Paleontology 88: 948 - 966. https: // doi. org / 10.1666 / 13 - 036","LUJAN A., DELFINO M., ROBLES J. M. & ALBA D. M. 2016. - The Miocene tortoise Testudo catalaunica Bataller, 1926, and a revised phylogeny of extinct species of genus Testudo (Testudines: Testudinidae). Zoological Journal of the Linnean Society 178 (2): 312 - 342. https: // doi. org / 10.1111 / zoj. 12414","MACAROVICI N. G. 1930. - Sur une Testudo bessarabica Riab. trouvee dans les couches meotiques de Cioburciu (Dep. Tighina-Roumanie). Bulletin de la Section scientifique de l'Academie roumaine, 13 e ann., 1 - 2: 26 - 30.","GMIRA S. 1995. - Etude des Cheloniens fossiles du Maroc. Anatomie. Systematique. Phylogenie. Museum national d'Histoire naturelle, Paris, 174 p. (Cahiers de Paleontologie; 25).","LAPPARENT DE BROIN F. DE, BOUR R. & PERALA J. F. 2006 c. - Morphological definition of Eurotestudo (Testudinidae, Chelonii): Second part. Annales de Paleontologie 92 (4): 325 - 357. https: // doi. org / 10.1016 / j. annpal. 2006.07.003","VLACHOS E. & TSOUKALA E. 2015. - The diverse fossil chelonians from Milia (Late Pliocene, Grevena, Greece) with a new species of Testudo Linnaeus, 1758 (Testudines: Testudinidae). Papers in Paleontology. https: // doi. org / 10.1002 / spp 2.1031","MILLER L. 1779. - Icones animalium et plantarum. Letterpress, London, 10 p. [published 1776 - 1782]. https: // doi. org / 10.5962 / bhl. title. 112488","LAPPARENT DE BROIN F. DE 2003. - The Miocene chelonians from the southern Namibia, in SENUT B. & PICKFORD M. (eds), Faunas from the southern Namibia. Memoirs of the Geological Survey of Namibia 19: 67 - 102."]}