First report on Cretaceous vertebrates from the Algerian Kem Kem beds. A new procoelous salamander from the Cenomanian, with remarks on African Caudata Tannina Alloul, Jean-Claude Rage, Rachid Hamdidouche, Nour-Eddine Jalil To cite this version: Tannina Alloul, Jean-Claude Rage, Rachid Hamdidouche, Nour-Eddine Jalil. First report on Cretaceous vertebrates from the Algerian Kem Kem beds. A new procoelous salamander from the Cenomanian, with remarks on African Caudata. Cretaceous Research, 2018, 84, pp.384-388. ￿10.1016/j.cretres.2017.11.019￿. ￿hal-01675293￿ HAL Id: hal-01675293 https://hal.sorbonne-universite.fr/hal-01675293 Submitted on 4 Jan 2018 HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. 1 ACCEPTED MANUSCRIPT 1 First report on Cretaceous vertebrates from the Algerian Kem Kem beds. A new procoelous 2 salamander from the Cenomanian, with remarks on African Caudata. 3 4 Tannina Alloula,*, Jean-Claude Rageb, Rachid Hamdidouchea, Nour-Eddine Jalilb,c T 5 P 6 aUniversity of Sciences and Technology Houari-Boumèdiène, Laboratory of Geoynamic I R 7 Bassins and Orognesis, Algiers, Algeria C 8 bSorbonne Universities, CR2P CNRS-MNHN-UPMC Paris 6, Department Origins and S 9 evolution, National Museum of Natural History, CP 38, 57 rue Cuvier, 75005 Paris, France U 10 cFaculty of Sciences Semlalia, University Cadi Ayyad, Marrakesh, Morocco N 11 A 12 *corresponding author: [email protected] M 13 E-mail addresses: [email protected] (T. Alloul), [email protected] (J.C. Rage), 14 [email protected] (R. Hamdid ouche), [email protected]. (N.E. Jalil). D 15 E 16 Abstract T 17 In northwestern Africa, the Kem Kem plateau is a major source of continental Cenomanian P 18 fossils. The plateau extends across the Algerian-Moroccan border but, unlike the intensely E 19 worked Moroccan part, the Algerian side of the Kem Kem beds has received less attention. C 20 However, recent field work in Algeria resulted in the recovery of a locality that yielded a C 21 promising vertebrate assemblage. Among the fossils is a trunk vertebra belonging to a A 22 salamander, a group whose remains are extremely rare in Africa. The vertebra is procoelous 23 and it presents combination of characters that suggest it belongs to a new taxon of unknown 24 affinities. Although the putative new taxon is represented by a single specimen that is too 25 poorly preserved to be formally named, the discovery is important for showing that 26 salamanders were more diversified than expected in the Cretaceous of Africa. 2 ACCEPTED MANUSCRIPT 27 Keywords 28 Kem Kem 29 Algeria 30 Cenomanian T 31 Amphibia P 32 Caudata I R 33 C 34 1. Introduction S 35 Caudata (i.e. salamanders) are primarily Laurasian amphibians (Bailon et al., 2011; U 36 Gardner and Rage, 2016). Most extant and extinct species occur in North America, Europe N 37 and Asia, which are the territories that made up the former Laurasia. These past and present A 38 ranges suggest that salamanders originated in Laurasia (Gardner and Rage, 2016). Outside of M 39 Laurasia, living salamanders occur only in the northern parts of three former Gondwanan D 40 continents (Frost, 2017): South America, the African Plate (Africa plus the Arabian Peninsula E 41 and Middle East) and India. The colonization of India does not appear problematic, because T 42 the Indian Plate has been in contact with Eurasia for a longer time (latest Cretaceous to early P 43 Eocene? Kapur and Khosla, 2016; Verma et al., 2016) than South America and Africa, both of E 44 which have been linked to Laurasia only recently, during the Neogene. The colonization of C 45 these two southern continents by modern salamanders appears to be a Neogene phenomenon. C 46 In Africa, fossiAls of both living and extinct taxa are known; they are rare but range from the 47 Middle Jurassic to the Pleistocene (Gardner and Rage, 2016, and references therein). 48 Here we describe the first salamander from the upper Cretaceous Kem Kem beds of Algeria. 49 This specimen was collected from a locality known as Oued Bou Seroual. 50 51 3 ACCEPTED MANUSCRIPT 52 2. African Caudata 53 A few fossils document the recent history of African Caudata. The earliest known of 54 these fossils comes from the early Pleistocene of Morocco and was referred to as Pleurodeles 55 cf. waltl (Bailon et al., 2011). P. waltl is a living species of European affinities, which is T 56 consistent with the Laurasian (Eurasian) origin for living salamanders inhabiting P 57 northernmost Africa. The date of dispersal of living salamanders into Africa is unknown, the I R 58 only certainty is to assume that it is older than the early Pleistocene. C 59 Aside from Pleistocene fossils, patchy older remains document a history clearly distinct S 60 from the recent colonization. They all come from the northern part of Africa and range from U 61 the middle Jurassic (Bathonian; Haddoumi et al., 2016) to the early-middle Eocene (Gardner N 62 and Rage, 2016). The relationships of these fossils are either unknown or disputed. Of A 63 particular importance are the remains from the Cenomanian-Santonian interval (Late M 64 Cretaceous) assigned to the endemic genus Kababisha (Evans et al., 1996) or to a closely D 65 related form, cf. Kababisha (Rage and Dutheil, 2008; Gardner and Rage, 2016). Their E 66 presence in the Late Cretaceous of Africa was regarded either as the result of vicariance (Rage T 67 et al., 1993) or of a dispersal from Laurasia (Evans et al., 1996). Here, we report on a new P 68 specimen from the Cenomanian of Africa, which likely represents a salamander distinct from E 69 Kababisha. C 70 C 71 3. The AlgerianA Kem Kem and the fossiliferous locality 72 The Kem Kem plateau of Algeria is located in the western part of the Saharan platform 73 at the junction between the mountain chain of Ougarta and the Moroccan Anti-Atlas (Zellouf, 74 1987) (Fig. 1A). The name Kem Kem has a Berber origin meaning torn or shredded (Lavocat, 75 1954); the name Hammada is also used, which means a vast and rocky plateau. This plateau is 76 almost tabular with a slight inclination to the north, it is semi-desertic and excavated by a very 4 ACCEPTED MANUSCRIPT 77 dense river network (Joly, 1962). It is 200 km long extending NE to SW from the village of 78 Taouz in southeastern Morocco, to the village of Zegdou in southwestern Algeria (Joly, 1962) 79 (Fig.1B). It is located approximately 1400 km southwest of Algiers and 350 km southwest of 80 Bechar (Fig. 1A). T 81 The wadis (i.e. rivers) that incise the surface of the Kem Kem as a dense network P 82 typically are not deep enough to expose the underlying marlstones and sandstones. In the I R 83 Oued Bou Seroual area, however, the wadi Daoura does cut into the sandstone layer. A C 84 deposit rich in disarticulated micro-vertebrates was recovered recently in this region. It is S 85 situated in the central part of the Kem Kem plateau, 90 km northeast of Zegdou and 50 km U 86 east of the famous Gara Sbaa locality (Cenomanian, Morocco; Lavocat, 1948; Cavin et al., N 87 2010) (Fig. 1B). The preliminary and unpublished list of vertebrates includes: Chondrichthyes A 88 (Onchopristis dunklei, O. numidus), Actinopterygii (Polypteriformes, Semionotiformes), M 89 Actinistia, Dipnoi, Amphibia (Anura), Squamata, Crocodylomorpha, Sauropoda, Theropoda, D 90 Pterosauria and, as reported here, a salamander. E 91 T 92 4. Geological setting P 93 The Cretaceous series of the Hammada, along the Algerian-Moroccan border was first E 94 and briefly described, on the Moroccan side, by Clariond (1933) during field work throughout C 95 the Hammada of Taouz. He described the following succession, from bottom to top: 120 m C 96 thick whitish and pinkish soft sandstone; 3 m thick calcareous sandstone with crystals of A 97 calcites and manganese spots, attributed to the Albian on the basis of the presence of the 98 echinoderm Dorocidaris taouzensis; and a thick layer of limestones, which he divided into 99 two parts, a lower part assigned to the Cenomanian, due to the presence of the ammonite 100 Neolobites vibrayanus, and an upper part dated as Turonian on the basis of the presence of the 101 gastropod Nerinea requieni. 5 ACCEPTED MANUSCRIPT 102 Later Choubert (1948), Lavocat (1948, 1954), and Dubar (1949) divided the Kem 103 Kem beds into three formations: a lower continental formation commonly called ‘Grès 104 infracénomanien’ or ‘Formation d’Ifezouane’ assigned to the Albian (Choubert, 1948; Dubar, 105 1949; Ettachfini and Andreu, 2004); a second, lagoonal formation composed of colorful T 106 marlstones with gypsum, assigned to the lower Cenomanian and called ‘Marne versicolore à P 107 gypse’ (Choubert, 1948) or ‘Formation d’ Aoufous’ (Dubar, 1949); and a third, marine I R 108 formation of Cenomanian-Turonian age, comprised of white marly-limestones including C 109 flints, called ‘Formation d’Akabou’ (Dubar 1949). S 110 Sereno et al. (1996) united the two lower formations of Dubar (1949), namely the U 111 Ifezouane and Aoufous formations into a single unit informally named the ‘Kem Kem beds’. N 112 The Kem Kem beds were assigned to the lower Cenomanian (Sereno et al., 1996; Cavin et al., A 113 2010) on the basis of close similarity between the vertebrate assemblage of these beds and M 114 that of Bahariya, in Egypt (Catuneanu et al., 2006). D 115 The Kem Kem beds in Oued Bou Seroual, Algeria, are reported here for the first time E 116 and consist mainly of sandstone. The lower part includes thin reddish sandstones and T 117 yellowish coarse sandstones, overlaid by reddish coarse sandstones; all these sandstones show P 118 oblique and horizontal stratifications. The upper level comprises yellowish coarse sandstones E 119 interspersed with greenish coarse friable sandstone; this is the richest level in terms of the C 120 number of vertebrate fossils. C 121 A 122 5. Material and methods 123 The poorly consolidated sandstones were screen washed using 1 mm, 800, 500 and 124 400 µm mesh-size sieves. Three kilograms of matrix from the Oued Bou Seroual area were 125 processed. In spite of this small sample, the collected and treated sedimentary rocks delivered 126 diverse vertebrate assemblage, which includes about a hundred remains identifiable at high 6 ACCEPTED MANUSCRIPT 127 taxonomic level. The vertebrate micro-remains were subsequently sorted under 128 stereomicroscope (model Leica A60). The described specimen is housed in the 129 palaeontological collections of the Museum of the University of Sciences and Technology 130 Houari Boumediene (MUHB), Algeria. T 131 P 132 6. Systematic Palaeontology I R 133 Lissamphibia Haeckel, 1866 C 134 Caudata Scopoli, 1777 S 135 Family indeterminate U 136 Material: one trunk or anteriormost caudal vertebra (MUHB 1010001). N 137 A 138 6.1. Description M 139 MUHB 1010001 (Fig. 2A-J) is a small, slightly distorted vertebra (maximum length D 140 from anterior rim of prezygapophysis to posterior rim of postzygapophysis = 2.1 mm). Its E 141 main characteristic is the procoelous nature of its centrum. In dorsal aspect, the vertebra is T 142 elongate and narrow. The prezygapophyses are well developed, but their shape cannot be P 143 determined precisely. The neural spine is very low. It appears as a ridge that runs along the E 144 entire length of the neural arch; posteriorly, the ridge forms a low, triangular tubercle, but C 145 anteriorly the ridge is so shallow that it is scarcely perceivable. The distal portions of the C 146 transverse procAesses are broken off. Only their bases are preserved; those are broad and 147 positioned relatively posteriorly. In anterior view, the neural canal is large and the 148 prezygapophyses are approximately level with the top of the canal. The anterior cotyle is 149 filled by matrix. Short but strong anterior basapophyses are present on either side, 150 lateroventral to the cotyle. The bases of the transverse processes are directed lateroventrally. 151 They are not thick and they do not include a dorsal and a ventral elements; in other words, 7 ACCEPTED MANUSCRIPT 152 they are not true rib-bearers. In lateral view, the base of the transverse process is attached 153 obliquely (anterodorsally to posteroventrally) to the lateral wall of the neural arch. A low 154 ridge extends between the transverse process and the ventral part of the posterior condyle, but 155 there are no accessory ridges or flanges buttressing the process anteriorly. No vertebrarterial T 156 foramen pierces the basis of the transverse process and the vertebra lacks spinal foramina. The P 157 condyle clearly projects posteriorly; it appears as a bony continuation of the centrum and not I R 158 as a calcified infilling of a posterior cotyle. On the ventral face, a shallow but sharp keel C 159 occupies the posterior two-thirds of the centrum length. There are no foramina on the ventral S 160 surface. In posterior view, the condyle shows a large notochordal pit. U 161 N 162 6.2. Remarks A 163 MUHB 1010001 shows a combination of characters that is encountered only in M 164 Caudata: presence of basapophyses; absence of buttresses on either side of the cotyle as a D 165 result of the high position of the prezygapophyses; marked anterior orientation of the E 166 prezygapophyses, which renders the interzygapophyseal constriction very shallow; presence T 167 of a ridge extending between the transverse process and the condyle; and condyle non- P 168 hemispheric, flat posteriorly, with a large notochordal pit. The vertebra lacks haemapophyses, E 169 therefore it comes either from the trunk or the anteriormost caudal regions. In addition, the C 170 absence of a double-processed rib-bearer (instead, seemingly replaced by a simple transverse C 171 process) suggesAts that the vertebra belongs to an elongate, snake-like salamander. The 172 procoelous nature of the vertebra enables to narrow comparisons to procoelous salamanders, 173 which are inferred to be snake-like forms. 174 The vertebrae of Caudata are either amphicoelous or opisthocoelous, with a very few 175 exceptions that may be labelled procoelous. The nature of the posterior vertebral condyle, 176 which renders the vertebrae procoelous, has been disputed (Evans et al., 1996). Rage et al. 8 ACCEPTED MANUSCRIPT 177 (1993) regarded vertebrae with posterior condyles as really procoelous. However, according 178 to Evans et al. (1996), the posterior condyle is not a true condyle; instead, it is made up by the 179 infilling of the cotyle by calcified material and the vertebrae would be ‘pseudoprocoelous’. It 180 is true that the posterior condyle is made by additional material in large vertebrae, but in small T 181 vertebrae it appears to be a true, osseous condyle that is continuous with the centrum. The P 182 nature of the posterior condyle of these salamanders remains to be really investigated, but this I R 183 issue is beyond the scope of our study. C 184 Among Caudata, except some extant species of Ambystoma that are pseudoprocoelous S 185 (Evans et al., 1996), a posterior condyle occurs only in some extinct Gondwanan taxa, which U 186 are restricted to the Late Cretaceous. These fossils are Kababisha humarensis and K. N 187 sudanensis fom the Cenomanian (or perhaps Campanian-Maastrichtian; Eisawi, 2015) of A 188 Wadi Abu Hashim, Sudan (Evans et al., 1996), cf. Kababisha from the Cenomanian of M 189 Morocco (Rage and Dutheil, 2008) and from the Coniacian-Santonian of Niger (Rage et al., D 190 1993; Gardner and Rage, 2016; JCR, work in progress), and Noterpeton bolivianum from the E 191 Maastrichtian of Bolivia, South America (Rage et al., 1993). T 192 The vertebra from the Algerian Kem Kem beds is readily distinguished from those of P 193 other procoelous and/or pseudoprocoelous Caudata in being relatively more depressed and E 194 less narrow (Fig. 2A-J vs 2K-M), in lacking vertebrarterial foramina and the anterior C 195 accessory crests that buttress the transverse process, and in having strong anterior C 196 basapophyses. It should be noted, incidentally, that the presence or absence of vertebrarterial A 197 foramina was not addressed in the description of Noterpeton (Rage et al., 1993). Based on 198 well-preserved specimens, it may be stated here that such foramina are present in Noterpeton 199 as they are in other procoelous caudatans, except in the taxon from Oued Bou Seroual. The 200 Algerian vertebra further differs from those of K. humarensis and K. sudanensis in being 201 relatively more elongate (Fig. 2L, M). However, its elongation is somewhat reminiscent of a 9 ACCEPTED MANUSCRIPT 202 vertebra referred to a juvenile individual of Kababisha by Evans et al. (1996: text-fig. 9G-J). 203 Nevertheless, elongation of the specimen from Algeria is not related to a juvenile, age as 204 demonstrated by its well-developed prezygapophyses and the moderate size of its neural 205 canal. T 206 In addition, the Algerian vertebra does not represent an intracolumnar variant in P 207 Kababisha or Noterpeton. In these two genera, all post-atlantal vertebrae are clearly taller and I R 208 have anterior accessory crests or flanges. Consequently, we regard the salamander from Oued C 209 Bou Seroual as representing a new taxon, but defer naming it because currently only a single S 210 and incomplete vertebra is available. U 211 N 212 7. Discussion A 213 Assuming that Wadi Abu Hashim in Sudan is really Cenomanian in age as originally M 214 reported (Werner, 1994), and not Campanian-Maastrichtian as recently suggested (Eisawi, D 215 2015), then the specimen from Oued Bou Seroual, cf. Kababisha from the Moroccan Kem E 216 Kem, and Kababisha humarensis and K. sudanensis from Sudan are the only salamanders T 217 known from the Cenomanian of Africa, and the only salamanders known from the Callovian- P 218 Turonian of Gondwana, an interval of approximately 76 million years. These Cenomanian E 219 taxa represent the earliest known procoelous or pseudoprocoelous salamanders, an C 220 assemblage that extends up to the Maastrichtian. All known Cenomanian salamanders from C 221 Gondwana havAe procoelous (or pseudoprocoelous) vertebrae, whereas those from the 222 Cenomanian of Laurasia have amphicoelous vertebrae (Gardner and DeMar, 2013; Skutschas, 223 2013). 224 Unfortunately, the new salamander does not help to resolve origin of the Gondwanan 225 procoelous salamanders, i.e. either the result of vicariance or of a dispersal from Laurasia. 226 This will remain unresolved until new palaeontologic discoveries.