III. Age of the Continental Intercalaire.
In a penetrating note, Conrad Kilian  posed the fundamental scheme of the succession of the continental series of the Sahara. He distinguished two great continental complexes there, surrounded by fossiliferous marine series: the “Continental Intercalaire”, generally included between the last Paleozoic marine beds and the first transgressive Upper Cretaceous marine level; the “Terminal Continental” succeeds the marine Cretaceous and corresponds to the Tertiary period. Moreover, he introduced and later specified [Kilian and Langlois, 1938] the notion of a “Continental Post-Tassilian”, representing the upper part of the Paleozoic beneath a non-marine facies.
However, the age of the Continental Intercalaire was not well specified. “Albian sandstones”, as authors such as J. Savornin [1934, 1947] called it? Or a comprehensive series encompassing a continuous group from the Permian to the Cenomanian, as C. Kilian himself was inclined to think? We attempted a synthesis of this question on the occasion of the XIXth International Geological Congress [Lapparent, 1952], taking account of the observations acquired to that date. But these things can be still better specified now, thanks to all the recent paleontological discoveries.
It is certain that continental complexes of Carboniferous age exist, which conclude the Paleozoic beds of the Sahara. Their age can be specified, which is variable according to the place: Namurian at Tindouf; lower or middle Westphalian at Chebbi, Berga, and Edjelé; Stephanian at Colomb-Béchar. Corresponding to Kilian’s Continental Post-Tassilian, these series are however less developed than had been thought: they are found localized in certain determined regions.
The Continental Intercalaire proper, which is now easily distinguished from the Continental Post-Tassilian1, appears in two different ways in the central Sahara. In the first and most frequent case, the clayey-sandy series rests in clear discordance on the Paleozoic or Precambrian, and up to the heart of the old massifs as at Sérouénout [Nordet, 1954]. At the base and diverse levels, it contains a fauna and flora that are a little delicate to interpret, but all the same rather clearly indicate the Lower Cretaceous. In this case, the Continental Intercalaire is an enlarged equivalent of the Wealden of Europe, in other words a fluvio-lacustrine facies covering all or part of the Lower Cretaceous and within which it is illusory to distinguish stages. In the central Sahara, in the opinion of all it also encompasses the lower Cenomanian, because the transgression which surmounts it begins everywhere, in the north as in the south, with the upper Cenomanian with Neolobites vibreyani, a characteristic ammonite.
In the second case, the Carboniferous seems to be followed by a strong clayey-sandy series whose upper part only contains the vertebrates and silicified wood of the Continental Intercalaire, indicating the Cretaceous. It is particularly thus in the Ohanet-Bourarhet region. However, the very detailed researches of the C. R. E. P. S. geologists brought a solution whose elements we did not have in 1952. Thanks to aerial photos and bold explorations that were able to penetrate the Issouane Erg, these geologists discovered a cartographic discordance in the Zarzaïtine series on various ends of the Carboniferous [C. R. E. P. S., 1957, 1960]. Moreover, at the base of this Zarzaïtine series Messrs. Claracq and Chavand discovered remarkable vertebrate remains, determined by Mr. Lehman  as belonging to capitosaurid stegocephalians, exclusively Triassic animals. Furthermore, Mr. Claracq recovered in 1958 two teratosaurid teeth, again with stegocephalians, in the sandstones at the summit of the Zarzaïtine cliff north of Maison Rouge; such an association indicates the Keuper. Therefore the Ohanet-Bourarhet section, noted for the first time in 1948 [Lapparent and Lelubre, 1948], can now be interpreted in the following manner.
The Carboniferous sea must have retreated little by little during the Moscovian; some continental clays and sandstones (our Tiguentourine series) represent the end of this stage.1 After a gap in sedimentation, during which the Hercynian foldings were produced, the sandstones and clays of the Zarzaïtine series were deposited in slight discordance on the Paleozoic: without doubt they would be Middle Triassic for the lower part and Upper Triassic for their higher part (presence of capitosaurids and teratosaurids). The successive clayey-dolomitic group, or upper Zarzaïtine series, has been shown to be sterile up to now: by its position, it perhaps corresponds to the Liassic and Middle Jurassic.2
The Taouratine series, containing the large sauropod Brachiosaurus [Lapparent, Claracq, and Nougarède, 1958] and a flora of clearly Jurassic character [Boureau and Caillon, 1958], belongs without doubt to the Upper Jurassic. It is surmounted by the clayey-sandy beds of the In Akhamil series with fishes, crocodiles, dinosaurs, and plants, which acknowledge some marine lappings coming from the north, with Lower Cretaceous molluscs (Chavan in Lapparent and Lelubre ). Along with the gypsum clays of the Djoua, the group encompass the Lower Cretaceous and lower Cenomanian.
But the Zarzaïtine and Taouratine series are folded and faulted in the region between Alrar and Edjelé. Because of this, before the recent vertebrate discoveries we were led to attach them to the Continental Post-Tassilian, in which we were followed by other authors [Freulon, 1955; Rumeau, Deffrenne and Decremps, 1956]. We recognize that our argument was not solid, and that once more only paleontology, even that of vertebrates, permits fixing the age of the terrain. Therefore from now on it is necessary to admit the notable reworkings and true post-Cretaceous foldings in this region, in addition to the Hercynian foldings, which agrees with more recent data on Saharan geology.
In our 1952 journey in the Nubian sandstones, we also mentioned Messak’s section. Having studied it for his thesis, J. Freulon now thinks that its lower portion could be equivalent to the Zarzaïtine series. It is undoubtedly not excluded that in some other points of the Sahara the base of the Continental Intercalaire can comprise, at least under a reduced thickness, beds that would be Jurassic or Triassic [Boureau and Freulon, 1959].
Moreover, in the eastern Sahara, geologists such as H. Faure, J. Freulon, and J. Ph. Lefranc think that the Continental Intercalaire sandstones encompass a continental equivalent of the Upper Cretaceous toward the top; but these elevated levels have furnished neither vertebrates nor invertebrates.
In conclusion of this analysis, we accept that the dinosaur fauna which I will describe belongs essentially to the Lower Cretaceous; it reaches into the Albian and perhaps up to the lower Cenomanian. This is well confirmed by the existence of this same fauna in extreme southern Tunisia, at diverse levels in the Dahar cliff, surrounded by the marine Upper Jurassic below and the marine upper Cenomanian above [Lapparent, 1953a; Basse, 1953]. The dinosaurs of Baharija in the Egyptian desert and those of Kem Kem in southern Morocco are contemporaneous with more elevated levels, their localities being situated immediately below the marine Cenomanian or being intercalated with the oyster beds.
There will be only two exceptions in the paleontological descriptions to follow, one for the carnosaur teeth at the summit of the Zarzaïtine cliff, attributed to the Upper Triassic, the other for the gigantic bones of Brachiosaurus from ZR.2, which would have an Upper Jurassic age.
The paleontological study follows the classification of dinosaurs generally adopted by specialists (cf. Piveteau [1955, p. 785-962]). To facilitate the reading of this chapter, I indicate the place where the dinosaurs of the Sahara belong in this classification.
Everyone admits that the term dinosaur, consecrated by usage, artificially groups two very different phyletic series whose origins were distinct. On the one hand there are the saurischians, with a typically saurian pelvis bearing three bony branches; on the other the ornithischians, with a characteristic pelvis bearing four branches, the fourth being called the postpubis.
The following table of the systematic arrangement of the dinosaurs of the Sahara is then obtained.
1. Suborder THEROPODA, essentially carnivorous.
— Superfamily CARNOSAURIA, powerful predators.
— Family Teratosauridae: Teratosaurus sp.
— Family Megalosauridae: Baharijasaurus ingens STROMER,
Inosaurus tedreftensis, nov. gen. nov. sp.
— Family Tyrannosauridae: Carcharodontosaurus saharicus (DEPÉRET).
— Superfamily COELUROSAURIA, gracile animals with slender, hollow bones.
— Family Coelurosauridae: Elaphrosaurus iguidiensis, nov. sp.,
Elaphrosaurus gautieri, nov. sp.
2. Suborder SAUROPODA, heavy herbivores with four columnar limbs.
— Family Camarasauridae (?): Rebbachisaurus tamesnensis, nov. sp.
— Family Astrodontidae: Astrodon sp.
— Family Titanosauridae: Aegyptosaurus baharijensis STROMER.
— Family Brachiosauridae: Brachiosaurus nougaredi, nov. sp.
— Superfamily ORNITHOPODA, herbivores and bipeds.
— Family Iguanodontidae: Iguanodon mantelli MEYER.
— Superfamily STEGOSAURIA, bearing bony armor.
— Family Acanthopholidae: genus indeterminate.
The fossil elements studied and described in the following pages number more than five hundred. They include either bones examined at the locality and which for various reasons had to be left in place after measurements, sketches, and photographs, or above all samples brought back to Paris. All of these were initially pooled at the geology laboratory of the Catholic Institute, where most of this work was accomplished. Naturally, comparisons were often made with the collections of the Museum in Paris, and we have also profited by those of the British Museum (Natural History) in London. But it was not possible to directly compare the heavy sauropod bones with the skeletons mounted in the American or German museums, and we had to content ourselves with figures in books or certain molds.
The study ended, we made a gift all the material brought from the central Sahara to the National Museum of Natural History. We thank Mr. J. P. Lehman for his always benevolent reception, and for agreeing to take care of this new collection, which will be well placed among the rich African paleontological collections in this national establishment.
Remarks on fossilization. — The dinosaur bones from the Sahara are easily recognizable on the terrain thanks to their habitual whitish color. Red colors are added to it due to iron oxides often filling the Haversian canals, and rather frequently bluish tints due to the presence of phosphates. Certain bones from In Gall, In Tedreft, and Marandet are very dark, nearly black, with a bluish tint on the surface.
Rarely the fossil bone became spongy. Usually it is heavy, mineralized by chalk phosphates. In these exceptional cases, bones are encountered alongside fossil wood that are calcified together: we noted this fact in the south of the Tanezrouft and in the Tilemsi.
The microstructure of the bone is frequently well preserved. Recall that in 1938, Meyendorff sent some bone fragments recovered by him in Gourara to Dr. Camp at Berkeley (California). By examining the thin sections, C. L. Camp recognized there the reptilian bone structure of the archosaur group, and in fact of dinosaurs (unpublished document).
We made the same observation on some well-preserved bones, in particular on a fragment of a Rebbachisaurus tamesnensis femur from Chebbi (Aïn Cheikh), a bone from In Tedreft, and a sauropod long bone from Tébéhic (Niger). The structure is identical in these three cases. In Pl. IX, fig. 6-7 I figure some microphotographs of the last. The detail of the Haversian canals surrounded by bone cells is seen there in transverse section, in the internal layer of the bone. A typical example of secondary ossification is developed there, in a circular arrangement around the Haversian canals; it is known that it is very preponderant in sauropod bones. The other thin section shows remains of fibrous primary ossification coming from the periosteum, widely invaded by secondary ossification. The appearance of the preparations entirely matches what is known of the microstructure of dinosaur bones.
Note that the sauropod metacarpal bones recovered at In Tedreft show under magnification a fine interlaced structure that makes the elements, on which rest the columnar limbs of an enormous animal, more solid.
1. SUBORDER THEROPODA
It is thanks to the discovery of two large carnivorous dinosaur teeth at Timimoun that the existence of theropods was made known for the first time [Depéret and Savornin, 1925] in the pre-Cenomanian continental sandstones of the central Sahara. Since then, our researches have greatly increased perspectives on the knowledge, diversity, and distribution of carnivorous dinosaurs. Not only are the remains of a large carnosaur revealed to be abundant, but other carnivores have been discovered along with the always rare traces of fragile coelurosaurs, for a total of six theropod species.
In order to more easily recognize the samples described later, I have designated the theropod remains by the letter T, adding the letters A, B, C, D, E, and F for each of the species studied, and these indications are reported on the labels of the elements in the collection. The equivalences are then as follows:
TA = Teratosaurus sp.
TB = Baharijasaurus ingens
TC = Inosaurus tedreftensis
TD = Carcharodontosaurus saharicus
TE = Elaphrosaurus iguidiensis
TF = Elaphrosaurus gautieri
Teratosaurus sp. [theropod A].
Pl. IV, fig. 5 and 6.
Following the paleontological order, I describe next two theropod teeth that must be put aside from the rest of the fauna studied in the subsequent part of the present work. These specimens, in good condition, come from the top of the Zarzaïtine cliff, which is a stratigraphic level clearly prior to those that furnished the other dinosaur remains; they were found by P. Claracq 2 km northeast of Maison Rouge, the C. R. E. P. S. base of Edjelé [Lapparent, Claracq and Nougarède, 1958].
The largest tooth measures 43 mm long and 16 mm wide at the base. It is slightly asymmetrical and must have occupied a rather anterior position in the jaw. The serrations ornamenting the trenchant edges are of the classic megalosaurid type; they are extended along the posterior edge and also the anterior edge. The smaller has the same characters; it measures 22 mm long and 9 wide at the base, where the trace of a rather well-marked neck is seen [Pl. IV, fig. 5-6].
Similarities and differences. — These two teeth are very similar to those of Megalosaurus found rather often in the Jurassic, and at first sight it is tempting to refer them to the family Megalosauridae. But a stratigraphical argument must invite prudence in affirming this. Indeed, the discoverer of these fossils brought back some stegocephalian bones, jaws, and teeth that he had recovered in the same locality. This discovery is necessarily referred to the Upper Triassic, or at most Rhaetian, summit of the Zarzaïtine cliff. I then consulted Mr. Fr. von Huene of the University of Tübingen, a specialist in Triassic dinosaurs. He remarked that our specimens are equally similar to the teeth of teratosaurids, carnosaurs from the Keuper and Rhaetian that are not very rare in southern Germany [von Huene, 1907-1908]. In his opinion, the Zarzaïtine teeth are referred as well to Teratosaurus or Pachysaurus as to Gresslyosaurus. Note that very close, if not identical, forms are known under the name Orinosaurus in the higher Triassic of South Africa. I underline that the discovery of teratosaurids in the central Sahara is a very new fact of great importance.
Baharijasaurus ingens STROMER [theropod B].
Pl. V, fig. 4, 16, and 17; VI, fig. 7.
At Mount Iguallala and In Abangarit, we found a dinosaur with compact vertebrae that seems very distinct from the other theropod described below. It is represented by six caudal vertebrae, of which here is the description.
One anterior caudal vertebra is amphicoelous, more hollowed anteriorly than posteriorly; it measures 6.5 cm long (Pl. VI, fig. 7). A very deep neural canal excavation will be noted.
Two middle caudal vertebrae, entirely of the same type, are 6 cm and 5.5 cm long (Pl. V, fig. 16-17). The articular surfaces for the chevrons are very prominent on the ventral posterior part.
Three other middle caudals measure 6, 5.5, and 5 cm long respectively (Pl. V, fig. 4).
We refer these vertebrae with probability to the Egyptian theropod Baharijasaurus ingens STROMER , moreover rather poorly known. But the resemblance between these rather massive vertebrae and those figured by Gilmore [1920, fig. 66, p. 117 and pl. 42] equally recall the American genus Dryptosaurus, also from the Lower Cretaceous. It is evidently very desirable that new discoveries can permit a better understanding of this carnivore, whose presence in the fauna of the Continental Intercalaire is certain, but still in the state of indications.
Inosaurus tedreftensis, nov. gen., nov. sp. [theropod C].
Pl. VII, fig. 1; XI, fig. 1.
While sorting the elements recovered in the course of our various missions, I noted four vertebrae from In Abangarit that did not appear to be attributable to a known form of dinosaur. They are all equally remarkable by their weak elongation.
First there is a last sacral vertebra, still fused to a portion of the preceding one; the centrum is only 4.4 cm long and the disc measures 3 x 3.5 cm.
A 4 cm long anterior caudal vertebra has the same characters (Pl. VII, fig. 1). The diameters of the anterior disc are 4.5 x 4.5 cm, so that it offers a nearly square aspect. The attachment points of the chevrons are situated very low, which has the effect of elongating the posterior disc from top to bottom; it measures 3 x 4 cm. The vertebra is narrow in is central part, which separates it from vertebra of Astrodon and makes it resemble a theropod.
A third, deteriorated and smaller (length = 3 cm), presents some rather close characters. The fourth is only 1.2 cm long and the keel is very sharp; it must belong to the end of the tail.
Stromer [1934, pl. II, fig. 20, 22, 23] figured three indeterminate caudal vertebrae from Baharija in Egypt that seem close to those I have described. They suggest that there was a new carnivorous dinosaur in the Continental Intercalaire of the Sahara, which later discoveries could make better known.
But precisely, in 1959 we recovered an entire series of elements of the same type at In Tedreft; they were found grouped together in the locality and perhaps belonged to a single individual. It comprises a lot of 18 vertebrae and a tibial fragment. The vertebrae are distributed in the following manner:
Two dorso-lumbar vertebrae have a very compact form; their length is 3.3 cm; the disc measures 4 x 4 cm. Two others are smaller (length = 3 cm). The ventral surface of the vertebral centrum shows a sharp keel between two deep depressions.
Two vertebrae fused together belong to the sacrum.
Five middle or perhaps posterior caudal vertebrae are 5 cm long (Pl. XI, fig. 1); the ventral surface is narrow and bears two keels. The first has preserved a good part of its neurapophysis; the total height of the vertebra must be 10 cm. The marrow mold, preserved on another, is flattened and measures 2 cm wide on the front of the vertebra. To these will be added seven fragments of caudal vertebrae.
The left tibia is represented by its proximal portion. It is entirely of the theropod tibia type, but its size is small: the greatest width of the superior surface is 6.5 cm.
As a result of these discoveries, we now propose for convenience to distinguish this theropod that appears in three Saharan localities by a new designation, Inosaurus tedreftensis, nov. gen., nov. sp. It seems to be related to the family Megalosauridae. Here are its characters:
Small theropod, incompletely known by twenty-five vertebrae and a portion of the tibia. The dorso-lumbar vertebrae are remarkable by their massive and nearly square appearance; the middle caudals bear two keels and are twice as tall as long.
Carcharodontosaurus saharicus (DEPÉRET) [theropod D].
Pl IV, fig. 1 to 4, 7, 9, and 10; V, fig. 1 to 6, 10 to 13; VII, fig. 6; VIII, fig. 5; IX, fig. 1.
Some teeth and bones from a large carnivore are relatively frequent in the dinosaur localities of the Sahara. Here is the description of the pieces gathered one by one and grouped for study.
SKULL. — The head is represented by two joined fragments recovered by Captain L. Archier at In Abangarit in 1952.
An internal mold of this piece, made under the care of Mr. J. P. Lehman, gave part of the endocranium. Comparisons were tried with the previously figured brains of tyrannosaurids, either Tyrannosaurus [Osborn 1912] or Carcharodontosaurus itself [Stromer, 1931]. But, in spite of a very visible symmetry, the interpretation—as much for the bony elements as for the partial endocranial mold—proved very arduous because the considered elements are very incomplete.
TEETH. — The teeth of this carnivore are numerous, although they are found in only four localities: 12 come from Timimoun, 5 from Guermessa, 137 from In Abangarit, and 2 from El Rhaz. If the 2 described originally by Depéret and 6 incomplete fragments are added, there is at a total of 164. This is a minimum figure, because such or such person collected them at In Abangarit and kept them as a precious Saharan souvenir; thus some were found in the collections of the French Institute of Black Africa at Dakar.
Those from Timimoun, recovered by us and Mr. Orengo, are referred very precisely to the specimens described by Depéret  from the same locality, by the size, the form, the oblique folds on the two faces, and the arrangement of denticles along the entire length of both edges. Our largest measures 42 mm wide at the base (Pl. IV, fig. 9). The length of the largest must be between 12 and 14 cm.
Those from Guermessa are smaller, generally rather worn, but with the characteristic thickness and form.
The In Abangarit region has furnished the most important lot (Pl. IV, fig. 1-4 and 7; V, fig. 1-3). It includes the largest, and here are the measurements of several teeth:
length……………. 125 mm 7 mm 105 mm 90 mm 87 mm
width at the base… 47 45 40 37 36
There are also interesting varieties to note. Indeed, if more than 100 are wide and similar to those from Timimoun, with oblique folds on both faces, 20 others are narrower and present variably shown characteristic torsion. These last teeth very probably correspond to a more anterior position in the maxillae, as kinds of canine teeth placed at the turning of the jaw (Pl. IV, fig. 7; V, fig. 3). These more specialized teeth were only rarely noted and have never been well studied; however they are known from Tendaguru, Portugal, southern Morocco, and in the jaws of Tyrannosaurus precisely at this location. Here are the measurements of some of these canine teeth:
length…………….. 80 mm 77 mm 62 mm 64 mm 54 mm
width at the base… 33 31 28 27 26
A good canine tooth 70 mm long was recovered by H. Faure 34 km southwest of EL Rhaz. A tooth fragment, analogous to the specimens from Timimoun and clearly showing the oblique folds, comes from 100 km southeast of El Rhaz.
VERTEBRAL COLUMN. — Thirty vertebrae of this large carnivore are reported.
Cervical vertebrae. A portion of a cervical vertebra was recovered at In Tedreft; it is strongly convex anteriorly. The neural canal is narrow and very deep, and this significant character will be found on the following ones.
Dorsal vertebrae. A very powerful vertebral centrum from Aoulef (Pl. VI, fig. 1) presents the following measurements: length = 7.5 cm; posterior diameter = 12 x 12 cm. The vertebra is concave anteriorly and flat posteriorly. It is slightly larger than the corresponding vertebrae of Allosaurus valens [Gilmore, 1920, p. 43]; but a little less than those of Gorgosaurus [Lambe, 1917, p. 24], and indicates a large theropod 6 to 7 m long.
Two strong dorsal vertebrae come from In Tedreft; the largest is 12 cm long and shows a deep, narrow neural canal.
Sacrum. A 28 cm long portion of a sacrum was found at In Tedreft, formed by two strong and solidly coossified vertebrae. The marrow cavity has a diameter of 3 cm. Without doubt these are only the last two sacral vertebrae, which could be modified caudals. They are larger than those of Allosaurus from America [Gilmore, 1920, p. 45 and pl. 8-10], which is in agreement with the very large size of our Saharan carnivore.
Caudal vertebrae. From the Aoulef locality, with the same color and fossilization characteristics as the dorsal vertebra above, comes an anterior caudal vertebral centrum (Pl. VI, fig. 3) that could be from the same animal. It is concave anteriorly and flat posteriorly. Its length is 10.5 cm; its minimum width 5.5; the disc diameters are 11 cm tall by 9 cm wide.
A vertebra from In Abangarit is an anterior caudal (Pl. VII, fig. 6); it is amphicoelous and measures 12 cm long. The total height of the vertebra bearing its neural spine is 23 cm. This process is notably flattened posteriorly and this differentiates our animal from Gorgosaurus; on the other hand it is much less elongated than the process on the anterior caudal vertebrae of Allosaurus. A middle caudal from In Tedreft measuring 11.5 cm long belongs to a large individual.
Two rather massive caudal vertebrae were brought by H. Faure from Oued Baouet (Pl. VI, fig. 2-3), east of the Aïr (Téfidet locality). They are two middle caudals typical of a theropod, but by their slenderness and development in height (perhaps exaggerated by a flattening during the course of fossilization) they seem to be of a different type than the two preceding vertebrae. They measure 8.5 cm long. Another of the same type from 100 km southeast of El Rhaz is slightly larger. We have hesitated to attribute them to Carcharodontosaurus, because they rather resemble some caudal vertebrae of Ceratosaurus (?) or an indeterminate theropod from Tendaguru [Janensch, 1925]. However, they seem to be of the same type as the vertebrae from the Carcharodontosaurus skeleton from Baharija described by Stromer [1931, pl. I, fig. 10], and we have no reason to distinguish them from it. One seizes on this occasion the fragility of theropod determinations that rest exclusively on caudal vertebra of banal type.
Five other middle caudal vertebrae were found at Timimoun (Pl. VI, fig. 6) and Aoulef; their lengths vary from 10 to 7 cm. A very deep neural canal is noted there, and, on one of them, a furrow for the passage of vasculature on each side anteriorly. An 8 cm long vertebra from Aoulef, which must be a caudal from this theropod, was showed to us by Mr. Augiéras in his personal collection from “Buffalo Bordj” at El Goléa. Three caudal vertebrae from In Tedreft measure 12, 11, and 10 cm long respectively; two chevrons from posterior caudal vertebrae come from this same locality. A 6 cm long vertebra from Alrar shows a well-marked flat part above the point of departure of the zygapophyses. Finally, I note a rather damaged vertebral centrum from In Abangarit, remarkable for its narrowness (2.5 cm) and the single, well-defined keel; it is a posterior caudal.
FORELIMB. — We have a portion of a humerus, recognizable by its characteristic torsion; the diameter of the median part is 4 cm. If this bone does belong to this large theropod, it underlines the forelimb reduction classic in this group. The same character of reduction is noted for a bone from El Rhaz that we interpret as a radius (calculated length: 11 cm).
A second phalanx from digit II of the left manus is perfectly preserved (Pl. VIII, fig. 5). It is 6 cm long; the width of the distal trochlea is 2.4 cm; the minimum diameter of the shaft: 1.8 cm. We also have the distal end of another phalanx, a little stronger than the preceding; the width of the distal trochlea is 2.8 cm.
A good manual ungual phalanx or claw is 8.5 cm long (Pl. VI, fig. 11). It is strongly recurved, with two deeply marked lateral furrows. Note that the lower process for muscular attachment is 2 cm long, which provided great force for depression of the claw into the flesh of the prey. Another, of the same length but more massive, is strongly arched and slightly asymmetrical (Pl. VI, fig. 13); it comes from El Rhaz.
HIND LIMB. — Different from those of the forelimb, the bones of the pes are long and strong. We have the distal portion of a metatarsal indicating a bone more than 30 cm long (Pl. IX, fig. 1). The shaft is hollow. The articulation with the first phalanx is a deep pulley, which must have permitted great flexibility of the pes. A large phalanx from El Rhaz is 8.5 cm long (Pl. IV, fig. 10).
A strong claw, probably from digit III, comes from Alrar (Pl. VI, fig. 12). Its length was at least 10 cm; its width on the side of the articular surface is 4 x 4 cm. The ventral surface is flattened, indicating that it belongs to a pes destined for walking and not a manus for grasping prey. Another, less massive claw, 9 cm long and a little asymmetrical, is perhaps from digit II (Pl. VI, fig. 10); it is rather worn.
Similarities and differences. — The bones and teeth described from the central Sahara bring interesting elements to our understanding of this carnivorous dinosaur. Recall that it was noted for the first time under the name Megalosaurus saharicus by Depéret and Savornin , who figured two teeth from Timimoun. Some analogous teeth, a portion of skull, and various bones were recovered from Baharija in 1911 by the Munich Museum expedition. When he described them, Stromer  separated this form from the genus Megalosaurus by creating a new genus, Carcharodontosaurus. We confirm the utility of this name (if not its elegance) for this Cretaceous carnivore, which could have arisen from Jurassic megalosaurids by an increase in size and more forceful development; it belongs to the family Tyrannosauridae, which is developed in the Cretaceous. R. Lavocat  found numerous teeth of this animal in southern Morocco.
By its large, trenchant, and serrated teeth, and according to its vertebrae, Carcharodontosaurus saharicus (DEPÉRET) must be intermediate in size between Jurassic carnivores and Cretaceous giants such as Gorgosaurus [Lambe, 1917], Tyrannosaurus [cf. Osborn, 1916], or Tarbosaurus [Maleev, 1955].
The teeth of Carcharodontosaurus, very well described by Depéret , seem a little less slender than those of the Tendaguru theropod [Janensch, 1925]; but it must be acknowledged that a distinction of the two forms based solely on teeth is rather subtle.
The forelimb is extremely reduced relative to the hind limb. The recurved, sharp manus claws, animated by solidly implanted muscles, made it a redoubtable predatory organ.
Known in Egypt and in southern Morocco, present in southern Tunisia at Hourara, as in the east and west of Niger, this large predator was an important element in the Saharan vertebrate fauna during Cretaceous times.
Some very slender theropods, with hollow and fragile bones, have left significant remains full of interest among the varied bones that we have recovered in the Sahara. They seems to be distributed in two different types, which are each new species; but we have very few comparative terms to study them.
Elaphrosaurus iguidiensis, nov. sp. [theropod E].
Pl IV, fig. 8; V, fig. 7 to 11; XI, fig. 9.
TEETH. — Forty-nine theropod teeth, smaller, more elongate, clearly more curved, and thinner than the large teeth of Carcharodontosaurus, probably come from a coelurosaur. I say probably because certain specimens could be alveolar teeth from a carnosaur; the difficulty in correctly attributing isolated teeth of carnivorous dinosaurs is well known.
The Ebrechko (Niger) locality furnished thirty-one of them alone, recovered over several square meters (Pl. V, fig. 8-12); it is the manifestation of a torrential delta. The other tooth localities are distributed principally in two regions: extreme southern Tunisia (Pl. IV, fig. 8; V, fig. 13) and Niger. But one very typical tooth was recovered isolated south of Alrar (Pl. V; fig. 14), which indicates a vast geographic distribution in total.
VERTEBRAL COLUMN. — We have eight very elongated caudal vertebrae of very characteristic aspect; the vertebral centrum is hollow. They come from three localities: In Abangarit, Ifayen Ignère, and Timimoun, and this arrangement of rare and fragile elements is significant for the presence of this coelurid across the central Sahara. The most complete is 8 cm long for only 2.5 cm in height of the posterior disc (Pl. XI, fig. 9). Another smaller, incomplete one must measure 6.5 cm long for 2 cm in height of the posterior disc. The same proportions (5.5 cm long for 1.5 cm in height for the disc) exist for the one from Ifayen Ignère. The two smallest are only 4 cm long.
FORELIMB. — A small, 3 cm long claw comes from El Rhaz; two narrow and deep gouges, serving to lodge the retractor muscles, are situated very high and have a very different shape from those of large theropods.
HIND LIMB. — We have the damaged distal end of a small right femur. A more complete element is a 35 cm long tibia; the bone is hollow; the ends were partially destroyed.
Similarities and differences. — The characters of the bones described above indicate the family Coeluridae. The largest caudal vertebra (Pl. XI, fig. 9) is of the same type as that from the Upper Jurassic of Boulogne-sur-Mer figured by Sauvage [1897, pl. VII, fig. 7-8]. Similarities are to be sought alongside Elaphrosaurus bambergi JANENSCH from Tendaguru; but it seems that a constantly lesser size and some accentuated differences make it another species.
Elaphrosaurus gautieri, nov. sp. [theropod F].
Pl. V, fig. 5 and 6; X, fig. 2, 4 to 8, 10, and 11.
The In Tedreft locality has furnished important coelurid remains that seem to be larger and clearly different from the preceding one. Here is a description of the elements.
VERTEBRAL COLUMN. — Initially there is a lot of sixteen vertebrae distributed in the following manner.
One cervical vertebra is well preserved (Pl. XI, fig. 5). It measures 8 cm long; it is convex anteriorly and very hollow posteriorly, and this accentuated arrangement suggests a very mobile neck. The insertion points for the cervical ribs are noted anteriorly. The neurapophysis is largely preserved. The diameter of the marrow at the posterior exit of the neural canal is 18 mm.
Two dorsal vertebrae, 7 and 8 cm long, are characterized by extremely deep neural canals (Pl. X, fig. 5; XI, fig. 2). The ventral surface of the vertebral centrum is very excavated, which has the effect of placing the two articular surfaces into strong relief. Four other vertebral fragments belong to dorsals.
A massive and compact vertebra, 5 cm long, with wide and short zygapophyses, seems to be the dorso-lumbar vertebra. Indeed, its posterior face is widened and inclined obliquely (Pl. V, fig. 6).
Sacral vertebrae. Three bony pieces are revealed as half-vertebrae fused together; they are square in their median parts. They are evidently sacral vertebrae.
Caudal vertebrae. Three well-preserved caudal vertebrae are 8.5 to 8 cm long (Pl. V, fig. 5; XI, fig. 4). They are procoelous; the centrum is hollowed by lateral cavities which lighten it. The keel bears a single strong carina. Two other fragments are from caudal vertebrae of the same type, square in the middle.
FORELIMB. — I consider a short (20 cm long), wide (10 cm for the proximal part and 8.5 cm for the distal part) bone as a left humerus from this dinosaur (Pl. XI, fig. 10). It seems quite different from the very slender humerus of Ornitholestes [Osborn, 1903, 1916] but the same as that of Elaphrosaurus bambergi [Janensch, 1925]; by its robustness it indicates a strongly grasping forelimb.
A 30 cm long bone was recovered isolated; it seems that it must be an ulna from this theropod.
PELVIS. — The distal end of a right pubis shows the considerable widening of this bone so characteristic of theropods. The articular surface with the other pubis is oval, measuring 10.5 x 5.5 cm (Pl. XI, fig. 6).
HIND LIMB. — We have several pieces that testify to a hind limb adapted to running and much more developed than the forelimb. First there is the distal end of a femur. Then a right tibia, a good hollow bone of which we have the two well-preserved ends. The proximal end is strong and triangular, measuring 12.5 x 8.5 cm; the distal end is 8.5 x 5.5 cm. I also cite a fibula represented by its two ends, proximal and distal. The four metatarsal ends that we have recovered indicate extremely long, hollow, and very slender bones. The same character is manifest in a half-phalanx.
Finally, I think that a hollow bone, flat below, presenting an excavated and well-preserved articular surface can well be allotted to the proximal part of a metatarsal from this species; the element was recovered at In Abangarit.
Single individual. — At In Tedreft, we have also recovered bones from this same theropod; but they were found grouped together and belonged to a single individual; from whence came the interest to examine this lot of rare elements separately.
VERTEBRAL COLUMN. — Besides an isolated neural arch from a cervical vertebra, two dorsal vertebrae and one damaged caudal, one very interesting element is a 14 cm long sacrum. It is formed by two vertebrae fused solidly together (Pl. XI, fig. 7-8). The vertebral centra are widened considerably in the zone where they are united. The neural canal is wide (2 cm) and deep. The vertebrae are convex anteriorly and excavated posteriorly.
Does the sacrum really only include two vertebrae fused and thus reinforced by this unaccustomed widening? This can be asked; but in the absence of comparative elements, only the fact can be noted. In any case, the exact same type of widening of fused vertebrae is found in one of the elements described on the preceding page.
FORELIMB. — We have three small claws, lengths 4, 4.5, and 6 cm respectively; this last, with a more elongate form, is very asymmetrical at the base.
HIND LIMB. — A good tibia was complete in the locality and measures 70 cm; the bone is hollow. There is also a right tibia (Pl. XI, fig. 11), nearly exactly the same size as that described above.
We also have the distal end of a fibula, the proximal portion of a metatarsal, and four fragments of phalanges. These bones are extremely elongate and their form shows a great flexibility of articulation.
Similarities and differences. — Thus there are enough indications in recent finds made at In Tedreft to think that a second coelurid species lived in the Sahara in Cretaceous times. The consistently greater size of all the bony elements and the form of the vertebrae seem to us to justify the creation of a species distinct from Elaphrosaurus iguidiensis, which was recognized first. I name it E. gautieri as a sign of the recognition that owed to Mr. Francis Gautier, because he agreed to spend two days on the immense Tamesna reg seeking the In Tedreft locality, which he discovered the preceding year and subsequently led us to himself. It can be asked whether this species had not already been noted twice in other sectors of the Sahara: a tibia from Baharija [Stromer, 1934, pl. III, fig. 1-2] and a tibia recovered by R. Lavocat  in southern Morocco; the length of the latter (64 cm) is in rather good agreement with ours.
According to the dimensions of the bones, and referring to the attempted reconstruction of Elaphrosaurus bambergi [Janensch, 1925, pl. I], E. gautieri can be thought of as a slightly larger animal, being around 6 m long.
2. SUBORDER SAUROPODA
The examination of a first lot of fifty similar vertebrae from the Sahara initially resulted in distinguishing two rather different types of sauropods, above all among the bones from Mount Iguallala in Niger. The first, also the most frequent, was a large animal whose caudal vertebrae had the tendency to be greatly elongated. The other was notably smaller, and its caudal vertebrae, with compact centra, immediately recalled the family Titanosauridae. In the following, the number of vertebrae examined is notably increased, which permits completing my first remarks. Thus I describe this large sauropod and this titanosaurid, but also a third, very rare genus of sauropod, Astrodon, which is presently revealed at In Gall, and a fourth, gigantic and very interesting, Brachiosaurus, from a slightly older stratigraphic level.
As for the description of the theropods, I designate each species by a letter reproduced on the elements in the collection, S indicating Sauropoda:
SA = Rebbachisaurus tamesnensis
SB = Astrodon sp.
SC = Aegyptosaurus baharijensis
SD = Brachiosaurus nougaredi