À la suite de Laplace, Camps [4] récolte à Mansourah des pièces lithiques variées, mais sans en noter la provenance stratigraphique. Elles ne sont certainement pas toutes contemporaines.

La collection Laplace regroupe des objets provenant des deux niveaux inférieurs (notés Mansourah) et de l'ensemble supérieur (notés L.R.) et fait donc foi de l'association stratigraphique de matériel taillé avec la faune. Le matériel des niveaux inférieurs comporte presque exclusivement des objets façonnés essentiellement en quartzite, de technologie simple, choppers unifaciaux et bifaciaux, voisinant avec des galets cassés, quelques éclats, des polyèdres irréguliers et des nucléus sur gros galets (Fig. 1). Les polyèdres sont de dimensions variées (80 à 180 mm de plus grand diamètre) et sont dans des états d'émoussé variés (frais à très roulé) ; les plus frais comportent des plages naturelles. De très rares pièces triédriques témoignent d'une chaîne de fabrication spécifique, mais, à l'exception d'une seule, dépourvue d'indication, toutes proviennent des niveaux supérieurs. La même superposition se retrouve d'ailleurs à l'Aïn Hanech [21]. Aucun objet du niveau le plus ancien ne peut être qualifié de pièce bifaciale à quelque stade d'élaboration que ce soit, ni de hachereau, et ne se rapporte donc à l'Acheuléen brièvement décrit et figuré par Camps [4 (pp. 25–28, figs. 11 & 13)]. La matière première (quartzite) de la presque totalité du matériel est certainement partiellement responsable de son aspect ancien, mais de très rares pièces en calcaire ne sont pas plus élaborées. L'industrie collectée in situ dans les travertins de Mansourah par G. Laplace appartient à une tradition technique sans biface, ni hachereau, échappant à la variabilité de l'Acheuléen régional et doit donc être rapportée à l'Oldowayen sensu lato.

La majeure partie de la faune figurée ou mentionnée par les auteurs antérieurs semble avoir disparu, et la plupart des pièces décrites ci-dessous n'ont jamais été mentionnées dans la littérature. Elles appartiennent aux collections du musée de la faculté des sciences de l'université d'Alger (FSUA), du musée national Cirta de Constantine (MNCC), et de la faculté des sciences de l'université de Rabat (FSUR).

Cf. Panthera pardus: une diaphyse de tibia, de la taille de celui d'une panthère.

Elephas sp. : deux fragments de molaires peuvent être rapportés à une espèce évoluée de ce genre, peut-être E. recki.

Ceratotherium mauritanicum (Pomel, 1888) : une prémolaire inférieure a été figurée par Gervais [12 (pl. 19, fig. 2)].

Hipparion (s.l.) sp. : selon Thomas [25], le genre coexiste à Mansourah avec Equus, et même si aucune pièce ne subsiste en collection, sa détermination peut être retenue. Son extinction en Afrique du Nord est antérieure à l'âge de l'Aïn Hanech.

Equus cf. tabeti Arambourg, 1970 : les dimensions dentaires correspondent à celles de E. numidicus (Aïn Boucherit), d'un grand E. tabeti (Aïn Hanech), ou de E. mauritanicus (Tighenif), mais les restes sont trop incomplets pour une détermination spécifique.

Kolpochoerus cf. maroccanus (Ennouchi, 1953) : trois m³ possèdent trois paires de tubercules, plus un talon équivalant presque à une quatrième paire, chaque paire étant séparée de la suivante dans le plan médian par deux tubercules accolés (Fig. 2). Les vallées intermédiaires sont presque dépourvues de colonnettes accessoires. Les dents sont basses, mais l'une au moins porte des traces de cément. Le Kolpochoerus de Mansourah semble distinct des lignées est et sud-africaines, ainsi que du K. phacochoeroides de l'Aïn Jourdel et de Ahl al Oughlam [9], qui ne possède qu'un tubercule médian entre les 2^e et 3^e paires de piliers, mais davantage de piliers accessoires. Les M3 de K. maroccanus, du Pléistocène inférieur marocain, sont grandes, mais simples ; des m3 qui pourraient être rattachées à cette même espèce sont connues à l'Aïn Hanech et à El Kherba [21,22], et c'est avec elles que les dents de Mansourah s'accordent le mieux.

Tragelaphus cf. gaudryi (Thomas, 1884) : quelques chevilles de cornes incomplètes appartiennent à un grand Tragelaphini. Elles sont plus proches de T. gaudryi de l'Aïn Jourdel que de T. algericus Geraads, 1981, de Tighenif. Toutes ces formes pourraient former une même lignée, bien distincte de celle d'Afrique orientale.

Bovini indet. cf. Bos bubaloides Arambourg, 1979 : un fragment de corne et un fragment de crâne appartiennent sans doute à l'une des espèces créées par Arambourg [2], mais on ne peut en tirer de conclusion biochronologique.

Oryx cf. gazella L. : deux cornes dépourvues de compression transversale doivent donc être rattachées aux oryx de type moderne, qui apparaît au niveau de l'Aïn Hanech.

Kobus aff. kob Erxleben, 1777 : l'antilope la plus commune est un cobe voisin de la forme actuelle et pléistocène K. kob, qui est très polymorphe. Les cornes courtes et trapues semblent néanmoins indiquer une variété distincte.

Alcelaphini indet. Un fragment de molaire supérieure.

Gazella cf. pomeli Arambourg, 1979 : deux cornes peuvent être rattachées à cette espèce, définie à l'Aïn Hanech.

Hippopotamus cf. sirensis Pomel, 1896 : un grand hippopotame tétraprotodonte, voisin de l'espèce actuelle, est commun. Il faut sans doute lui rattacher une symphyse mandibulaire décrite par Pomel comme provenant d'un autre site, sous le nom de H. hipponensis.

Bien qu'elle soit biaisée en faveur des grands mammifères, quelques indications peuvent être tirées de la faune de Mansourah. Les témoins de milieux ouverts et secs y sont plus rares que dans la plupart des autres sites nord-africains de cette période, tandis que Tragelaphini et Reduncini parlent en faveur d'une végétation mixte humide de buissons et prairie.

Plusieurs espèces de Mansourah sont également connues à l'Aïn Hanech. Tels sont Equus cf. tabeti, Gazella cf. pomeli et Kolpochoerus cf. maroccanus. Dans l'ensemble, un âge voisin de celui de l'Aïn Hanech nous semble pouvoir être retenu, mais la présence possible d'Hipparion, ainsi que du Tragelaphus cf. gaudryi, intermédiaire entre celui du Pliocène supérieur et le T. algericus de Tighenif, et d'un Kobus probablement distinct de Kobus kob tend à vieillir le site.

Nous avons déjà eu l'occasion de discuter l'ancienneté de la présence humaine au Maghreb, que ce soit au Maroc [19,20] ou en Algérie [5,11]. Dans cette région, il est apparu clairement que les données biostratigraphiques étaient à ce jour essentielles pour estimer la position chronologique relative des différents gisements. La richesse et l'originalité de la grande faune des travertins de Mansourah situent clairement leur mise en place au cours du Pléistocène inférieur. La présence d'une industrie lithique, associée rapportable à l'Oldowayen sensu lato, fait de ce site un jalon essentiel pour la connaissance des premiers peuplements d'Afrique du Nord, dont il convient à présent de confirmer le détail stratigraphique et archéologique par des fouilles modernes.

Cf. Panthera pardus. A single tibial diaphysis matches that of a panther.

Elephas sp. Gervais illustrated a molar fragment with two plates [12 (pl. 19, fig. 1)]. A distal fragment in FSUR has three plates, a width of 63 mm, and a height of 114 mm. Both are not of Loxodonta (the common form at Tighenif), and they must belong to some derived form of Elephas. Arambourg [1] gave the name E. moghrebiensis to the Aïn Hanech elephant, and reported it also from Mansourah, but it might be identical to the East African E. recki.

Ceratotherium mauritanicum (Pomel, 1888). Gervais [12 (pl. 19, fig. 2)] had figured a left lower p3, now lost. It is definitely less derived than that of C. simum and it must belong to C. mauritanicum, an extinct species defined at Tighenif, but which survived late in the Maghreb [10].

Hipparion sp. Thomas [25] stated that Hipparion teeth occur at Mansourah, together with Equus, but did not provide any description or figure, and no specimen is presently available. However, given their hardly mistakable morphology, this occurrence must be taken as very likely. In East Africa, Hipparion s.l. became extinct in the early Middle Pleistocene; its LAD in North Africa is more imprecise, but probably earlier. It is present at Aïn Boucherit, close to the Plio-Pleistocene boundary, but the two teeth found at Aïn Hanech are probably reworked from the lower level [1]; Hipparion was not found in the recent excavations at the latter site [21]. Thus, if its occurrence in the lower layer of Mansourah is taken for granted, it points to an age earlier than that of Aïn Hanech.

Equus cf. tabeti Arambourg, 1970. Several teeth (mainly upper) and incomplete limb-bones in Algiers belong to at least one species of Equus whose size would fit E. numidicus, a large form of E. tabeti, or E. mauritanicus, three North African forms best known from Aïn Boucherit, Aïn Hanech and Tighenif, respectively [1], [11] and [23]. The outline of the external wall is more like asinine horses, and the long protocone is unlike that of E. numidicus, but isolated teeth of Equus are hardly identifiable to species.

Kolpochoerus cf. maroccanus (Ennouchi, 1953). The material consists of three m3s (basal length = 57.1, 57.0, 57.2; width = 20.8, 22.2, 22.2, respectively) and a talus, in Constantine. The last two m3s are unworn (height = 21.8, 21.9), and almost certainly from the same individual (Fig. 2); the third tooth is slightly worn, and very similar. There are three pairs of rather widely spaced pillars, plus a strong ‘talonid’, almost as wide as the preceding pair, and consisting of two closely appressed sub-equal pillars. There are two small accessory pillars between the talonid and the third pair, on the lingual side, and a small poorly distinct pillar between the first two pairs, on the labial side; all other valleys are free of accessory pillars. Along the midline of the teeth, the successive pairs of pillars are separated by two median pillars. They are sub-equal in size between pairs 1–2, but the anterior one is larger between pairs 2–3. Between the third pair and the talonid, the posterior one is larger and quickly fuses with the anterior one. The teeth are brachyodont, the unworn teeth being very slightly broader than high. Some cement coating remains on the worn tooth only, but the other teeth lack roots and were unerupted, so that cement might not have deposited yet.

Kolpochoerus is one of the most common suids in the Plio-Pleistocene of Africa, but the Mansourah molars differ from all eastern and southern forms. The m3s are definitely shorter than those of K. paiceae from South Africa or the latest forms of K. heseloni from Koobi Fora, longer than those of K. afarensis from the Pliocene of Ethiopia, or than the early forms of K. heseloni from the Turkana basin, and longer and relatively narrower than those of K. majus from the Pleistocene of East Africa. Furthermore, the m3s of K. paiceae always have more than three pairs of pillars, and all posterior tubercles have a simple morphology. On the contrary, K. afarensis and early forms of K. heseloni have more simple m3s. The size of the teeth from Mansourah is within the variation range of K. heseloni from the Omo member G or the Koobi Fora Upper Burgi and KBS members (Fig. 3). However, m3s of comparable size in the Turkana basin are more complex and more hypsodont than those from Mansourah, while those of similar complexity are smaller (see [13 (pl. 6.12.H and pl. 6.12.D)] as examples of the former and the latter, respectively). Specimens from the Lower Pleistocene of Konso, Ethiopia, of similar size, are also more complex than those from Mansourah, with a talonid consisting of at least three pillars ([24], figs. 3A, D). Thus, although the m3 pattern is similar, the Mansourah Kolpochoerus is not on the evolutionary lineage of K. heseloni.

Kolpochoerus has a sparse record in North Africa. Thomas [25] first recorded K. phacochoeroides from the Late Pliocene of Aïn el Bey, Algeria. This species is best known from the Late Pliocene of Ahl al Oughlam in Morocco [8] and [9], which yielded hundreds of specimens. Among these are about 25 m3s; they are more hypsodont than at Mansourah, there is a single median pillar between the 2^nd and 3^rd pairs of pillars, the successive pairs are less clearly separated, with the valleys often blocked by extra pillars, and the talonid consists of several small pillars. Thus, the Mansourah Kolpochoerus is clearly distinct from K. phacochoeroides.

K. maroccanus (Ennouchi, 1953) is the only other named North African form. The type, kept in FSUR, consists of an incomplete M3, plus P3–P4 probably from the same individual, of unknown geological age. The size of M3 matches the largest teeth of K. phacochoeroides, but the tooth is as simple as the smallest teeth of this species, and is slightly less hypsodont. Another M3 from level L of Thomas quarry [11], of Early Pleistocene age, has a slightly broader talon but is otherwise similar. We believe that K. maroccanus is a species different from K. phacochoeroides, but the shared disto-lingual fovea on P4 suggest that they belong to the same clade. Thus, K. maroccanus is an enlarged version of the smaller morph of K. phacochoeroides, somewhat like K. majus is an enlarged version of a simple morph of K. heseloni in East Africa. The m3s from Mansourah look very much like what could be expected of the m3s of K. maroccanus: a rather large size with a relatively simple morphology. The same species is very probably present at Aïn Hanech and El Kherba, where incomplete m3s were reported [21] and [22], but Kolpochoerus is unknown in later sites (Tighenif, and the Middle Pleistocene).

Tragelaphus cf. gaudryi (Thomas, 1884). A frontlet of a kudu-like antelope from Mansourah, first mentioned by Bayle [3], was figured by Gervais [12 (pl. 19, fig. 4)], but is now probably lost. No scale was given by Gervais, and the only point of his description not visible on the figure is that there is only one keel on the horn-core. This keel is anterior, and conspicuous, but the outline of the horn-core suggests that it was not very prominent, and that the cross-section was more or less rounded. This frontlet was referred by Gentry [7] to Tragelaphus gaudryi (Thomas, 1884), the type horn-core of which, from Aïn Jourdel, has a clear but not very prominent anterior keel, and an almost round cross section (about 50 × 50 mm). Two more Tragelaphus horn-cores are preserved in the FSUR, and two in Constantine (Measurements: APD × TD: FSUR: 66 × 55 and 65 × 67; MNCC: 56 × 56.5 and 66 × 58.5). They were much inclined caudally and rather divergent. The only keel is the anterior one, which is faint to moderate. They are similar to the type of T. gaudryi, but their larger size and weak anterior keel make this animal resemble T. algericus Geraads, 1981, from Tighenif, which is probably its descendant. However, the single horn-core from Tighenif is still larger (basal index: 82 × 75), more tightly spiralled, and has an incipient postero-medial keel, of which there is only a hint at Mansourah. Thus, in North Africa, the evolution from primitive T. gaudryi was clearly divergent from the East African one, where it is marked by an increase in transverse compression and strengthening of the anterior keel [7].

Cf. ‘Bos’ bubaloides Arambourg, 1979. Thomas ([25], pl. 4, fig. 6) referred a horn-core tip to ‘Bubalus’ antiquus, but this species is of much later age. Arambourg [2] described three species in the Lower Pleistocene of North Africa: Bos bubaloides and B. praeafricanus from Aïn Hanech are one and the same species, doubtfully distinct from B. palaethiopicus from Aïn Boucherit. As the horn-core is too dorso-ventrally compressed to belong to Bos s.l., the Mansourah bovine is certainly one of these; a similar horn-core tip has been found in the Late Pliocene of Ahl al Oughlam. A cranial fragment with the base of the horn-core in Constantine comes probably from the same species, but it is not clear yet whether it belongs to the African or Eurasian group.

Oryx cf. gazella L. A horn-core from Mansourah, now in the FSUR, was figured [15 (pl. 1, fig. 1)] as Oryx leucoryx, and as more slender than it really is. The diameters of its lowermost part, several centimetres above the missing base, are 42.5 × 36.8. Another horn-core, in Constantine, is preserved down to the base, which is almost round in cross-section (index: 48.5 × 49), but the transverse compression increases upwards, suggesting that the basal cross-section of the FSUR specimen was probably also almost circular. They are slightly curved (FSUR) or almost straight (MNCC). By their lack of transverse compression at the base (a feature probably linked with greater inclination backwards), they differ from oryx horn-cores from Late Pliocene or Early Pleistocene sites in the Turkana basin, but are similar to the modern type, known from Aïn Hanech onwards.

Kobus aff. kob Erxleben, 1777. The most common antelope at Mansourah is close to the modern K. kob. A horn-core was figured [12 (pl. 19, fig. 5)], and there are several specimens in Rabat and Algiers. Fossil and modern forms of K. kob display a large range of variations in size and morphology that encompass those of the Mansourah form, except that horn-cores of the latter are quite short, as in the allied genus Redunca. The transverse compression and backward curvature are unlike the latter genus, and they certainly belong either to an extinct sub-species of K. kob or to a new species.

Alcelaphini indet. A single fragment of upper molar is definitely alcelaphine.

Gazella cf. pomeli Arambourg, 1979. Two incomplete horn-cores, rather short, slightly curved and little compressed, match those of this species, best known from Aïn Hanech.

Hippopotamus cf. sirensis Pomel, 1896. The size and morphology of the hippo, common in FSUA and FSUR, are those of the living H. amphibius or of the Pleistocene H. sirensis (= H. gorgops?). Mention must be made of an anterior fragment of mandible, described and illustrated [18 (pl. 4, Fig. 1, Fig. 2 and Fig. 3)] as coming from the ‘route des Beni Foudda’, but which is labelled ‘Mansourah’; its light yellow colour better fits the latter site. As already shown by Arambourg, it clearly does not belong to H. hipponensis, a hexaprotodont species.

Although the sample is limited and certainly biased towards large animals, some conclusions can be drawn from the faunal assemblage. Hippos are common in all open-air Plio-Pleistocene African assemblages, but tell us nothing about the environment away from the lake or river margin. Alcelaphines, indicators of open dry environments, are rare at Mansourah, in contrast to most Plio-Pleistocene North-African sites. The best only indicators of open dry country are the Oryx and Gazella horn-cores. Tragelaphini and Reduncini form the bulk of the Mansourah bovid fauna, as in many East African sites, but they are not common components of North-African Plio-Pleistocene faunas. Modern species of the former tribe feed on leaves and are therefore indicators of small trees or bushes, while the latter tribe inhabits wet grasslands. Kolpochoerus co-occurs in large numbers with these tribes in the Turkana basin, and was found to be mostly a grazer [14], but more water-dependent than Metridiochoerus. Hence, the most consistent picture for Mansourah is that of mixed wet grassland and bushland.

Among the fauna from the travertines, some taxa have poor biochronological significance. Large tetraprotodont hippos are known from the Late Pliocene onwards; the elephant and Bovini are too poorly known to be useful; Ceratotherium mauritanicum survives late in North Africa. Oryx cf. gazella is probably the only living species, but it is present as early as Aïn Hanech. Kobus kob is known from the Late Pliocene onwards, but the features of the Mansourah horn-cores suggest an extinct variety or species. Hipparion would definitely point to an earliest Pleistocene age, but a doubt remains as to its occurrence in the travertines. Some taxa have more specific affinities with those from Aïn Hanech. Such is Equus cf. tabeti, also known from Ubeidiyeh in Israel, roughly of the same age, Gazella cf. pomeli, and Kolpochoerus cf. maroccanus. Perhaps the most interesting taxon is Tragelaphus cf. gaudryi, intermediate between the Late Pliocene type of this species, and the Tragelaphus algericus from the late Early/early Middle Pleistocene of Tighenif. On the whole, the fauna best matches that of Aïn Hanech, but the latter site has neither Tragelaphini nor Reduncini, and therefore lacks these evidences of early age that are present at Mansourah. The features of the lithic industry are in accordance with an Early Pleistocene age.