The only detailed anatomical description of
Since 1975, a nearly complete mounted skeleton of
The aim of this research was to uncover the following: (1) to disentangle the history of the discovery of
The focus of this paper is the Venice specimen of
The mounted specimen as exhibited today at the MSNVE. For scale, the right femur is 920 mm long.
This specimen is mounted on a metal frame and in order to photograph and describe it, all of the bones were dismounted from the frame, with the exception of the sacrum, which is fixed to the frame. Pictures of every original bone in its cranial (anterior), caudal (posterior), dorsal, ventral, lateral, and medial views were taken, using a Canon EOS 600D camera with 100 ISO sensitivity and a Tamron 17–50 mm (F/2.8) lens at focal distance 50 mm. The photographs are stored in the archive of the MSNVE, which is accessible to researchers by contacting the responsible people for Research and Scientific Divulgation of the Museum. A 200 mm-long caliper with measurement error of 0.01 mm and a 100 cm-long metric string (measurement error of 0.1 cm) were used to measure the bones. A table with all the measurements is included in the
Caudal vertebrae with pleurapophyses (often reported as caudal ribs or transverse processes in the literature) are considered as proximal caudals; those lacking pleurapophyses but with haemapophyses are middle caudals; distal caudals lack pleurapophyses and hemapophyses. The cervical-dorsal transition in the vertebral column was identified following
Bone surface texture, degree of fusion of the elements and obliteration of the sutures in skulls and vertebrae are the most common approaches to assess the ontogenetic stage of fossil tetrapods (e.g.,
According to
Between 1965 and 1972, five French palaeontological expeditions searched for dinosaurs in the Gadoufaoua area of the Sahara desert in Niger (
During the second expedition (February 25th–April 7th, 1966), GDF 300 (a nearly complete but disarticulated and scattered skeleton) and GDF 381 (“a skeleton two thirds complete”, p. 54) were collected. The following year, those specimens were carried to Paris for preparation and study. GDF 300 became later the holotype of
The third expedition (1969) found some dinosaur material at the In Gall locality (actually outside the Gadoufaoua area), but no
During the fourth expedition (January 5th–March 23rd, 1970), a nearly complete
During the fifth expedition (January 5th–February 25th, 1972), the
In 1971, Giancarlo Ligabue and Cino Boccazzi knew about the Gadoufaoua locality while travelling across the Sahara desert (
In the formal description of the new species,
Philippe Taquet (P Taquet, pers. comm., 2012) confirmed that the Venice specimen is the paratype with the missing bones being casts of the holotype. He also told us that he mapped the paratype bones in the field and that the map is still kept at the MNHN. Ronan Allain sent us a copy of that field map, which confirms that the Venice specimen contains the paratype material.
The field map sent to us by R Allain is divided into two sheets. The first sheet contains the writing “Ouranosaurus nig[eriensis]—Airfield—1970 —(specimen Venice Museum pro parte)” (in French) and the field map of the partially articulated paratype skeleton that is pictured in
Part of sheet 1 with the writing “
The second sheet is the map of a set of bones, which is not clearly identifiable in
The presence of a total of three pubes and possibly three ilia and scapulae, as well as a duplication of segments of the caudal vertebral column, indicates that the two sheets refer to two distinct skeletons.
Each bone in the two sheets is identified by a number in order to identify the elements and reassemble the skeleton once in the laboratory (
Dinosauria |
Ornithischia |
Ornithopoda |
Iguanodontia |
Ankylopollexia |
Styracosterna |
Note: the name
Other potentially diagnostic characters include a circular orbit with the same height as the lower temporal fenestra; dorsal segment of the vertebral column made of only 15 vertebrae.
The original diagnosis by
“Medium-sized iguanodontid (7 metres long)”—This cannot be accepted as a diagnostic feature.
“Bipedal”—According to
“Very long skull, narrow and relatively low, which maximum height occurs at the level of the nasal bulges”—This is a vague statement about elongation that should be supported by measurement ratios. The hadrosaurid
“Long and thin snout ending in a duck bill”—Proportionally, the snout of
“Extremely long, straight and anteriorly expanded premaxillae, which separate posteriorly the nasals from the maxillae”—This is the condition observed also in
“External nares widely visible in dorsal view” and “orifice of convergence between the nasal ducts very back placed”—
The external narial openings of
“Short predentary bone, wider than long”—Many other styracosternans have predentaries that are wider than long (e.g.,
“Low maxilla”—The maxilla of
“Short nasal bearing a rounded dorsal bulge”—The “rounded dorsal bulge” on the nasal is indeed a unique feature of
“Small antorbital fenestra”—The presence of a small antorbital fenestra is a primitive feature within the iguanodontians occurring for example in
“Circular orbit with the same height as the lower temporal fenestra”—This is not due to a larger size of the orbit but to a comparatively small size of the lower temporal fenestra. In nearly all other styracosternans, the lower temporal fenestra is higher than the orbit (e.g.,
“Straight and horizontal posterolateral process of squamosal that very little overlaps the paroccipital process”—This condition is found also in
“High paroccipital process, broad and anteriorly oblique”—The paroccipital process is similar in
“Broad and flat occipital condyle”—A relatively high morphological diversity exists among the occipital condyles of non-hadrosaurid styracosternans, but condyle shape was never considered as a diagnostic feature. The condyle is as broad as that of
“Basipterygoid processes directed mainly laterally”—This feature is observed also in
“Large upper temporal fenestrae, very divergent rostrally”—The upper temporal fenestrae are comparatively large and their main axis is laterocranially oriented also in
“Dentary that is deep anteriorly and low posteriorly”—
“Dorsal margin of the dentary bearing a long diastema anteriorly”—According to
“Well-developed retroarticular process [of the mandible]”—This process has the same development in other taxa, for example
“Teeth of
“Vertebral count: 11 [cervicals]—17 [dorsals]—6 [sacrals] 40 [caudals]” —The count of 11 cervicals is not diagnostic as it is the same in
“Relatively short tail”—This apparent shortness is due to the fact that the caudal segment of the vertebral column is incomplete in both the holotype and the paratype. Consequently it is not a diagnostic feature.
“Extremely long neural spines of the dorsal vertebrae”—This is clearly a potentially diagnostic feature, but needs to be quantified because
“Long and straight ischium with a foot-like distal expansion”—The ischium of
“Very proximal obturator process” and “very narrow obturator gutter” [of ischium]—These are related features because it is the position of the process that makes the obturator gutter narrow. The obturator gutter is a broad embayment in
“Very elongated/slender [
“Very deep and very developed prepubic blade”—Compared to the prepubic portion of the pubes of
“Straight pubic rod, much shorter than ischium and with widened distal extremity”—The “pubic rod” (=posterior pubic ramus or pubis s.s.) is straight and shorter than the ischium in many styracosternans (
“Slender ilium”—The ilia of other styracosternans are similarly slender, for example those of
“Preacetabular process accounting for half the total length of the ilium”—The preacetabular process of
“Convex dorsal margin of the ilium with a hint of an anti-trochanter”—In the Venice specimen the dorsal margin is actually straight, while it is convex in the holotype. Both conditions occur in a sample of
“Shallow acetabular cavity”—The acetabular notch is not shallower in
“Shallow post-acetabular notch”—The notch is shallow also in
“Ascending process of the astragalus placed posteriorly instead of anteriorly”—Actually,
“Tridactyl foot”—Actually, no complete foot of
“Phalangeal formula [of the pes] 0-3-4-5-0”—This is just a hypothesis because no complete foot of
“Humerus long and nearly straight”—The humerus/femur length ratio is 0.60 in
“Tiny hand bearing a spur-like and small fifth metacarpal that is not laterally directed”—Actually, the spur-like bone is not the metacarpal V but the ungual phalanx of manus digit I. According to
“Phalangeal formula [of the manus] 1-3-3-3-3 or 4”—This is just a hypothetical statement because no complete manus of
MSNVE 3714 is a partial skeleton; the missing elements were replaced by plaster copies, with the exception of the hyoid apparatus, the atlas and the cervical ribs, which are missing. Most of the original bones have also been partly reconstructed and restored (
The reconstructed parts are in red.
In this section, only the skeletal elements of MSNVE 3714 that add new information with respect to the description of the osteology of
The axial skeleton of MSNVE 3714 is composed of 76 vertebrae, but 10 caudals are completely reconstructed with plaster, so only 66 are actually preserved (
MSNVE 3714, the cervical series (A); the dorsal series (B); the sacrum (C); and the caudal series (D). Holotype, the cervical series (E); the dorsal series (F); the sacrum (G); and the caudal series (H). Numbers are progressive within each series. White vertebrae are those of the mount that are totally reconstructed. E–G are redrawn from
The last (11) cervical (A); the first dorsal (B). Both are in left lateral view. Reconstructed parts are in dark gray colour. Abbreviations: dia, diapophysis; ncs, neurocentral suture; ns, neural spine; par, parapophysis; poz, postzygapophysis; prz, prezygapophysis. Scale bar equals 10 cm.
Axis in cranial (A), right lateral (B), left lateral (C), and caudal view (D); cervical vertebra 3 in cranial (E), right lateral (F), left lateral (G), and caudal (H) views; cervical vertebra 11 in cranial (I), right lateral (J), left lateral (K), and caudal (L) views. Abbreviations: dia, diapophysis; od, odontoid process; ns, neural spine; par, parapophysis; poz, postzygapophysis; prz, prezygapophysis. Scale bar equals 10 cm.
The neural spine of the axis (
(A) left side; (B) right side. The parts of the neural spines that have been reconstructed or just covered by resin are highlighted in white. Reconstructed parts of the centrum, transverse processes, zygapophyses and pedicels of the neural arch are not highlighted. Numbers are progressive. Scale bar equals 10 cm.
Dorsal vertebra 1 in right lateral (A), caudal (B), and ventral view (C); dorsal vertebra 9 in left lateral (D), cranial (E), and ventral view (F) views; dorsal vertebra 17 in left lateral (G), caudal (H) and ventral (I) views. Abbreviations: bpl, ‘bump’ of the prespinal lamina; kl, keel; par, parapophysis. Scale bars equal 10 cm.
The peduncles of the neural arches, parapophyses, transverse processes and relative diapophyses are all reconstructed in the Venice specimen, presumably taking as reference for proportions and morphology those of the holotype. The neural spines, which are the most important feature of those skeletal elements, are also partly restored (
The centra of dorsals 1 and 2 are opisthocoelous. From dorsal vertebra 3 onwards, they become slightly amphicoelous to amphiplatyan. Their length ranges from 87 (vertebra 12) to 112 mm; they are slightly longer than high but dorsal 17 is more elongated than the others;
The morphology of the tall neural spines changes along the vertebral column (
The basal part of the spine in vertebrae 5, 7, 8, and 16–17 shows a cranial bump that is made by the cranially expanded prespinal lamina (
According to
Some other inconsistencies regarding the dorsal vertebrae are found in
Left lateral view. Scale bars equal 10 cm.
Vertebra 6 in caudal (A) and right lateral (B) views; vertebra 10 in cranial (C), right lateral (D), left lateral (E) and ventral (F) views; vertebra 21 in left lateral (G), caudal (H) and ventral (I) views; vertebra 35 in left lateral (J) and caudal (K) views; haemapophysis 8 in left lateral (L) and cranial (M) views. Abbreviations: af, articular facets of the haemapophysis, afh, articular facet for the haemapophysis; bpl, bump of the prespinal lamina; ns, neural spine; pla, pleurapophysis; poz, postzygapophysis; prz, prezygapophysis; vld, ventral longitudinal depression of the centrum. Scale bar equals 10 cm.
There are 20 proximal and 12 middle caudal vertebrae (17 including the five that are totally reconstructed). The last preserved caudal (caudal 38) seems to be a distal element (but see below).
The holotype also preserves 33 caudals. According to
In MSNVE 3714, centra are slightly amphicoelous in the proximal and first middle caudals to become amphiplatyan caudally. The caudal surface of proximal and middle caudals is more squared than the rounded cranial one because of the presence of the raised facets for the haemapophysis (
The lateral surface of the centrum near its articular facets is rough, with longitudinal grooves in some proximal and all middle caudal vertebrae (caudals 11–25;
Rough surface in vertebral centrum 24 (A); neurocentral suture in vertebra 8 (B); vertebra 10 (C); and vertebra 21 (D). Vertebrae are figured in left lateral view. Arrows point to the grooved surface in A and to the neurocentral suture in B–D. Scale bar equals 5 cm in A and 10 cm in B–D.
The pleurapophyses of the proximal caudals are flattened dorsoventrally and scarcely project laterally (
Vertebra 18 (A); vertebra 20 (B); vertebra 21 (C). They are shown in dorsal view. Abbreviations: ns, neural spine; pla, pleurapophysis; prz, prezygapophysis. Scale bar equals 5 cm.
The neural spines are mostly spatulate in lateral view, with a slight craniocaudal apical expansion (
The first haemapophysis is located between caudals 3 and 4, while the last one occurs between vertebrae 28 and 29, but it is artificial like the two vertebrae. There are two articular facets per pedicel in the chevrons of the first proximal caudal vertebrae because each pedicel articulates on two centra. The dorsoventral length of the haemapophyses tends to decrease caudally, but chevron 14 is shorter than chevron 15, both preserved distally. Possibly chevron 14 is in the wrong position and should be placed in a more distal position. The spine of haemapophysis 1 is straight, but those of the following elements up to haemapophysis 5 are arched, while the following show a variable degree of curvature from nearly straight to slightly arched. As with the caudal neural spines, the morphology and sloping of the chevrons of the Venice specimen are less regular and harmonic than in the holotype. This is probably a consequence of the breakage of the long and thin spines and subsequent restoration.
In both basal and derived iguanodontians, chevrons 1–3 (up to chevron 6 in some taxa) taper distally (e.g.,
Left coracoid in caudal (A) and dorsomedial (B) views; drawings of the left coracoid in caudal and dorsomedial views with the reconstructed parts evidenced in dark gray colour (C). Left scapula in cranial (D) and lateral (E) views; drawing of the left scapula in lateral view with the reconstructed parts evidenced in dark gray colour (F). Left sternal plate in dorsomedial (G) and ventrolateral (H) views; drawings of the left sternal plate in dorsomedial and ventrolateral views with the reconstructed parts evidenced in dark gray colour (I). Abbreviations: ap, acromion process; cof, coracoid foramen; clp, caudolateral process (‘handle’); crmp, craniomedial plate; cvp, caudoventral process; cos, coracoid sutural surface; df, deltoid fossa; dr, deltoid ridge; gl, glenoid; scl, scapular labrum; scs, scapular sutural surface; sp, broken sternal process. Scale bar equals 10 cm.
Left humerus in caudal (A), cranial (B), proximal (C) and distal (D) views; left radius-ulna in caudal (E), cranial (F), medial (G), lateral (H), proximal (I) and distal (J) views. Abbreviations: ch, caput humeri; dpc, deltopectoral crest; gtb, greater tuberosity; itb, inner tuberosity; lf, lateral flange; mf, medial flange; ol, olecranon; ra, radius; rc, radial condyle; u, ulna; uc, ulnar condyle. Scale bar equals 10 cm.
Left (A) and right manus (B). (C) and (D) are the drawings of the left and right manus, respectively. Totally reconstructed parts are evidenced in dark gray colour and those just covered by a film of resin are pale gray coloured; also minor reconstructed portions are pale gray. The reconstruction of the articulated digits II–V is reported in E. Abbreviations: int, intermedium; mc II–V, metacarpals II–V; mk, medial distal knob on the metacarpal V; ph II–V, phalanges of manus digits II–V (the last phalanx of each digit is the ungual); rad, radiale; uln, ulnare. Elements without abbreviation in A and B are reconstructed. Scale bar of A and B equals 10 cm.
The orientation of the manus in the following description (that integrates the description by
Metacarpal II (85 mm long) is straight, relatively slender (it is constricted in the middle) and with expanded extremities. Its proximal end is expanded craniopalmarily and mediolaterally. Its proximal surface is flat (it is convex in
Metacarpal III is expanded mediolaterally at both extremities and has a straight shaft. Its proximal surface is gently convex (it is strongly convex for
Metacarpal IV is also expanded at both extremities and has a straight shaft. The outline of its proximal surface is elliptical (the bone is craniopalmarily compressed) and is convex, while it is described as strongly convex and with a triangular outline by
Metacarpal V is stout and hourglass-shaped in craniopalmar view. Its distal extremity is wider than the proximal one. Its proximal surface has a rounded outline and is shallowly concave in the middle. Its palmar surface is slightly concave, whereas the cranial surface is convex. The distal end is expanded palmarily, whereas the proximal end is expanded cranially. A small knob occurs near the distal extremity of the metacarpal at its palmomedial corner. The distal surface has a sub-rectangular outline, with a shallowly concave central part (it is convex according to
Metacarpals II–IV would be separated by spatii interossei if appressed. This and the absence of articular facets between metacarpals III and IV suggests that metacarpals II–IV did not form a compact block in
Phalanx II-1 (right hand) is hourglass-shaped in palmar and dorsal views, with proximal and distal extremities that are mediolaterally expanded. Its distal end is more dorsopalmarily expanded than the proximal end. Its proximal surface has a sub-circular outline and is flat; its distal surface has a sub-triangular outline and is also flat. Its dorsal surface is covered by a film of resin. This phalanx is slenderer than phalanx II-1 of the paratype figured by
Phalanx III-1 (preserved in both hands) is a stout and hourglass-shaped element in palmar and dorsal views. Its proximal surface is flat with a sub-elliptical outline. Distally, it bears two rounded condyles that are separated by a broad and shallow intercondylar groove. The right phalanx III-1 of the paratype figured in
Phalanx IV-1 (right hand) is hourglass-shaped in palmar and dorsal views and its distal extremity is more mediolaterally expanded than the proximal one. Its proximal surface has a sub-elliptical outline and is flat. Its sub-triangular distal condyles are separated by a shallow intercondylar groove that is narrower than that of phalanx III-1. This phalanx is slenderer than phalanx IV-1 of the paratype figured by
Phalanx V-1 (preserved in both hands) is the most robust phalanx of the manus. It is hourglass-shaped in palmar and dorsal views. Its proximal extremity is more dorsopalmarily expanded than the distal one. Its proximal surface has a bean-like outline and a shallowly concave center. Its palmar surface is slightly concave and presents some thin longitudinal grooves and wrinkles (which are more evident in the right element than in the left one), possibly for muscular insertion or evidence of a cartilage cover (like the rough surface of the lateral side of the caudal vertebrae near their articular faces). Its distal surface has also a bean-like outline and is slightly convex. The phalanx V-1 figured in
Left ilium in medial (A); lateral (B); dorsal (D); and ventral (E) views. Left pubis in lateral (F) and medial (G) views. Left ischium in medial (I) and lateral (J; upside-down) views. C, H and K are the drawings of the elements of figures A–B, F–G and I–J, respectively, with the reconstructed parts evidenced in dark gray colour. Abbreviations: ac, acetabulum; bs, brevis shelf; ilp, iliac peduncle of ischium and pubis; isf, distal ‘foot’ of the ischium; isp, ischial peduncle of ilium and pubis; ms, medial shelf; no, notch; obn, obturator notch; obp, obturator process; pop, postacetabular process; por, posterior pubic ramus (pubis s.s.); ppb, prepubic process; ppn, neck of the prepubic process; prp, preacetabular process; pup, pubic peduncle of ilium and ischium; sap, supracetabular process; sf, facet for the articulation with sacrum. Scale bar in A–B, F–G and I–J equals 10 cm.
Right femur in medial (A), lateral (C), cranial (D), caudal (E), proximal (F) and distal (G) views; B is a particular of the fourth trochanter. The left tibia and fibula in medial (H), lateral (I), cranial (J), caudal (K), proximal (L) and distal (M) views. L is the mirrored proximal view of the right tibia-fibula because the proximal part of the left tibia-fibula is poorly preserved and badly mounted. Astragalus in cranial (N) and caudal (O) views. Abbreviations: as, astragalus; ca, calcaneum; caap, caudal ascending process of the astragalus; caig, caudal intercondylar groove; cf, caput femoris (femoral head); cmc, proximal caudomedial condyle of tibia; cnc, cnemial crest; cod, condylid; crap, cranial ascending process of the astragalus; crig, cranial intercondylar groove; ftr, fourth trochanter; gtr, greater trochanter; hn, neck of the femoral head; lc, proximal lateral condyle of tibia; lcd, distal lateral condyle of femur; lm, lateral malleolus; ltr, lesser trochanter; mcd, distal medial condyle of femur; mm, medial malleolus. Scale bar equals 10 cm in A–M.
Left pes in dorsocranial view (A) and drawing with the totally reconstructed parts evidenced in dark gray colour and minor reconstructed portions in pale gray (B). Ungual phalanx of digit III in dorsal (C) and medial (D) views. Abbreviations: mtII–IV, metatarsals II–IV; ng, nail groove; pphII–IV, phalanges of pedal digits II–IV (the last phalanx is always the ungual one). Elements without abbreviation are reconstructed. Scale bar in A equals 10 cm.
Because of this confused original description, all of the phalanges are re-described as they are mounted in MSNVE 3714.
Phalanx II-1 is stout, gently arched laterally and hourglass-shaped in dorsoplantar view. Its distal end is more expanded lateromedially than the proximal end. Its proximal surface has a quadrangular outline and is flat; its distal end is divided into two condyles that are separated by a wide and shallow intercondylar groove. Phalanx II-2 is much smaller than phalanx II-1 and has a quadrangular outline in dorsoplantar view. Its proximal surface has a sub-triangular outline; the distal end bears two scarcely developed articular condyles.
Phalanx III-1 is large, stout and as long as it is wide. Its proximal surface has a sub-elliptical outline and is slightly concave. Its distal end bears two scarcely developed articular condyles. Phalanx III-2 is disc-like, and much proximodistally shorter than lateromedially wide. Both proximal and distal surfaces have a sub-triangular outline; the proximal one is slightly convex, whereas the distal one is concave. Phalanx III-3 has a shape similar to that of phalanx III-2, but it is transversely narrower. The long phalanx III-4 (the ungual phalanx, 70 mm long) is spade-like, dorsoplantarily flattened and slightly arched (
Phalanx IV-1 is hourglass-shaped in dorsoplantar view and resembles phalanx II-1. Its proximal end is more dorsoplantarily expanded than the distal end. Its proximal surface has a sub-rectangular outline. The distal end bears two condyles that are separated by a wide and shallow intercondylar groove. The ungual phalanx IV-5 is gently arched plantarily and has a squared outline in dorsoplantar view, but its distal portion is missing and it was possibly like ungual phalanx III.
There is no agreement on the phylogenetic relationships of
Using this new information from the Venice specimen, we performed a cladistics analysis to reassess the phylogenetic position of
We used the data matrix of
Adam consensus tree obtained from the revised matrix of
Using the data matrix from
Revision of the phylogenetic analyses by
Panoramic view under lambda filter (the outer surface of the bone is at the top of the figure) (A); detail of the progressive transition between the compacta and the medullary cavity, characterized by erosional cavities (B); detail of the fibrolamellar bone and longitudinal vascularization forming the primary bone (C); detail of the microstructure of the outermost cortex showing the absence of an EFS (D); no secondary osteons are observed in the inner cortex. Green arrows point to the LAGs. Abbreviations: lb, lamellar bone; mc, medullary cavity; ps, periosteal surface; vc, vascular canals; wb, woven bone. Scale bars equal 10 mm in A and 1 mm in B–D.
The left humerus was sampled on its caudal side at mid-shaft. The cylindrical core sample was 23 mm thick and 14 mm in diameter (
The right femur was sampled on the craniolateral side at mid-diaphysis. The cylindrical core sample was 21 mm thick and 14 mm in diameter (
Vascularization is irregularly organized and a clear orientation is not evident. Primary vascular canals are still open, although infilling of lamellar bone is present. Secondary osteons (Haversian systems) cannot be identified in the thin section. No LAGs or annuli can be observed and there is no EFS (
Panoramic view under lambda filter (the outer surface of the bone is at the top of the figure) (A); the gradual transition between the compacta and the medullary cavity made of CCB and erosional cavities (B); detail of the plexiform vascularization and fibrolamellar bone forming the primary bone and the absence of zonation and LAGs within the compacta (C); detail of the microstructure of the outermost cortex showing the absence of an EFS (D). Abbreviations: ec, erosional cavities; lb, lamellar bone; ps, periosteal surface; vc, vascular canals; wb, woven bone. Scale bars equal 10 mm in A and 1 mm in B–D.
The right tibia was sampled craniolaterally in the diaphysis, slightly below mid-shaft. The cylindrical core sample was 26 mm thick and 14 mm in diameter (
Panoramic view under lambda filter (the outer surface of the bone is at the top of the figure) (A); gradual transition between the compacta and the medullary cavity made of CCB and erosional cavities (B); detail of the deeper cortex, showing zonation of the primary bone, irregular vascularization and fibrolamellar bone (C); detail of the outermost cortex showing the absence of an EFS (D); note the remodeling in the inner compacta. Abbreviations: lb, lamellar bone; mc, medullary cavity; ps, periosteal surface; so, secondary osteons; vc, vascular canals; wb, woven bone. Green arrows point to the LAGS. Scale bars equal 10 mm in A and 1 mm in B–D.
The neural spine of dorsal vertebra 14 was cross-sectioned within the basal third, in the middle, and within the apical third (
Panoramic view of the cranial half of the section (A); Haversian systems in the innermost cortex (B); transition between the compacta and the medullary cortex with erosional cavities and remodeling (C); detail of the outer cortex showing absence of an EFS and outermost LAGs (D). Abbreviations: lb, lamellar bone; ps, periosteal surface; so, secondary osteons; tb, trabeculae; vc, vascular canals; wb, woven bone. Green arrows point to the LAGS. Scale bars equal 10 mm in A, 1 mm in B and C, and 500 µm in D.
Panoramic view of the cranial half of the section (A); Haversian systems in the inner most cortex and endosteal bone (B); gradual transition between the compacta and the medullary cortex with erosional cavities and marked remodeling of the primary bone (C); detail of the outer cortex showing absence of an EFS and zonation of the primary bone (D). Abbreviations: eb, endosteal bone; ec, erosional cavities; ps, periosteal surface; so, secondary osteons; tb, trabeculae; vc, vascular canals; wb, woven bone. Green arrows point to the LAGS. Scale bars equal 10 mm in A and 1 mm in B–D.
Panoramic view of the cranial half of the section; note how the compact cortex becomes thinner trending through the top of the neural spine (A); detail of the Haversian systems in the inner most cortex and endosteal bone (B); gradual transition between the compacta and the medullary cortex (C; note the erosional cavities and deep remodeling of the primary bone); detail of the outer cortex with outermost LAGs but without an EFS (D). Abbreviations: eb, endosteal bone; ec, erosional cavities; lb, lamellar bone; ps, periosteal surface; so, secondary osteons; wb, woven bone. Green arrows point to the LAGS. Scale bars equal 10 mm in A and 1 mm in B–D.
The transverse cross-section of the proximal part of dorsal rib 15 has an oval outline. The cortex of the lateral side is heavily eroded; its maximum thickness is 17 mm (
Panoramic view of the thin section of the dorsal rib (A); transition between the outer cortex and the medullary cavity (B; note erosional cavities and deep remodeling of the primary bone); Haversian systems in the inner cortex (C); detail of the outermost cortex showing zonation and the absence of an EFS (D). Abbreviations: ec, erosional cavities; ps, periosteal surface; so, secondary osteons; vc, vascular canals; wb, woven bone. Green arrows point to the LAGS. Scale bars equal 10 mm in A, 1 mm in B and D, and 500 µm in C.
The Venice specimen undoubtedly includes the paratype of
Some incongruity exists between MSNVE 3714 and the mapped paratype skeleton. As anticipated above, a maximum of 15 dorsal vertebrae are identifiable in the map of the paratype (
In the map, only three neural spines can be seen in the region of the sacrum (the drawing of a distal fourth was marked as it were drawn in error; see
The number of caudal vertebrae (33) correspond to the map of the paratype and to MSNVE 3714, only when assuming that the element identified as a phalanx in the handwritten note on the map was a further centrum (which it probably was because of size and position of the bone and its alignment with the vertebral segment) and also considering in the count the vertebra that was not mapped (see
Only the pedicels of the first two chevrons are drawn in the field map of the paratype, between caudals 3 and 4 and 4 and 5, respectively (
Only one element is identified as a coracoid by the handwritten notes in the field map of the paratype. It is not specified whether it a left or a right one, but the paratype preserved the right one according to
Both radius-ulna pairs are preserved in MSNVE 3714. The map of the paratype shows one radius-ulna pair and a splint of bone that is identified as a second ulna (
Only two rounded elements are labeled as “carpals” in the field map of the paratype, while the Venice specimen has three presumed carpals. However, it is possible that the smaller element of the carpus is the fragment of bone (bone n. 104; see
Eleven bones at maximum can be referred to the manus in the map. They include three or four metacarpals and two or three ungual phalanges, one phalanx occurring 2.5 m away from the others (
Only one skeletal element can be an ilium in the paratype field map (
MSNVE 3714 preserves both pubes, although they are incomplete distally. Only one element is labeled as “pubis” in the paratype field map; it preserves most of the long and thin posterior pubic ramus (
Only one element is labeled as “ischium” in the paratype field map (
According to the field map, only the left femur was preserved in the paratype. No femur is documented in sheet 2. MSNVE 3714 also preserves the right femur, which, therefore, comes from another source. It cannot be the referred femur GDF 302 because the cranial intercondylar groove is completely encircled in that specimen (
The field map of the paratype shows only one tibia and fibula, plausibly the left one because of its position with respect to the left femur. MSNVE 3714 preserves both tibiae. A tibia and at least part of a fibula occur in the map on sheet 2 and might be the other crural elements of the Venice specimen.
The field map of the paratype shows elements only from the left pes (
Recapitulating, the field map of the paratype does not show the following elements that occur in MSNVE 3714: two dorsal vertebrae, the first two chevrons, the left coracoid, one ilium (probably the left one), one pubis, at least one ischium, the right femur and the right tibia and fibula. Therefore, MSNVE 3714 is most probably a composite, although it is mostly composed by the paratype of
The left femur and the right coracoid of the paratype were replaced in MSNVE 3714 with plaster copies. One original ungual phalanx of manus digit I (the spike-like ungual of the pollex) was replaced with a plaster copy. The left femur was sent back to the MNBH with the holotype: it can be seen in a photo of the skeleton exhibited at the MNBH that is available in the web (
Mature individuals of
The holotype was estimated to be 7 m long (
Holotype (A) and MSNVE 3714 (B). Redrawn and modified from
The ontogenetic stage of MSNVE 3714 was established by observation of macroscopic evidence of osteological immaturity in the skeletal elements and by their osteohistological features. The neurocentral sutures in proximal and middle caudal vertebrae up to caudal 25 (
The increasing organization of vascular canals toward the outer surface, the presence of Haversian systems, the decreasing spacing between LAGs and the absence of an EFS observed in the thin sections of tibia, neural spine and dorsal rib suggest that these skeletal elements belonged to a sub-adult individual (
The conclusions are that an over 6.5 metres-long individual of
In crocodiles, the obliteration of the neurocentral suture during ontogeny starts in the tail and ends in the neck (posterior-anterior sequence of neurocentral closure;
Most of the differences between MSNVE 3714 and the holotype are minor differences that can be ascribed to intraspecific variability (individual or ontogenetic variation), as also reported in a statistically significant sample of
Other differences are probably caused by mistakes in the preparation or assemblage of the skeletal elements in both specimens. These mistakes include the morphology of the carpus (which is possibly made of fragments of other bones in MSNVE 3714) and the morphology of neural spines and chevrons.
The differences in relative height of the spines of the sacral vertebrae could also be ascribed to individual variability or even sexual dimorphism. However, the steps in the ‘sail’ outline in correspondence of the passage dorsal vertebrae-sacrum and sacrum-caudal vertebrae in the holotype are puzzling and could be an artifact.
The presence of five or six more proximal caudals in MSNVE 3714 than in the holotype (four to five more than in
The metacarpal V is a stout element in relatively primitive iguanodontians such as
Possibly, the ischia of the Venice specimen are longer than those of the holotype because they are not from the paratype, but from a larger individual.
The osteohistology of many ornithopod taxa has been studied, including
As expected,
Faster growth is phylogenetically coincident with the taxonomical diversification of the derived iguanodontians and their increase in body size (
Among the Dinosauria, hyperelongation of the neural spines reaches its maximum in the dorsal vertebrae of the theropod
In
In iguanodontian ornithopods, the dorsal neural spines are usually much shorter than those of the theropod
In
The only work dealing also with the function of iguanodontian ‘sails’ is that by
The Venice specimen of
The Venice specimen shows several differences with the holotype that probably reflect intraspecific variability. Other differences are caused by mistakes in the identification, preparation and reconstruction of the skeletal elements.
The new information from the Venice specimen and the much higher number of iguanodontian taxa known today with respect to the mid-seventies allows emending of the diagnosis of
The recoding of the character states in the matrices by
Based on histological analysis, (the first performed on
Natural History Museum, London (formerly British Museum (Natural History)), UK
Commisariat à l’
Comitato Nazionale delle Ricerche, Rome, Italy
Centro Studi e Ricerche Ligabue, Venice, Italy
field acronym used by French palaeontologists for fossils found in the Gadoufaoua locality
Geologisches und Palaontologisches Institut der Universität Tübingen, currently the Institut für Geowissenschaften, Tübingen, Germany
Institut Royal des Sciences Naturelles de Belgique
Musée National Boubou-Hama, Niamey, Niger
Muséum National d’Historie Naturelle, Paris, France
Museo di Storia Naturale di Venezia (Natural History Museum of Venice), Venice, Italy
Royal Tyrrell Museum of Palaeontology, Drumheller, Canada
We thank Mauro Bon, responsible for Research and Scientific Divulgation of the MNSVE, for allowing us access to the specimen under his care; Barbara Favaretto and the staff of the MSNVE for assistance during the study the specimen. Mauro Bon also provided access to the historical information on MSNVE 3714. We acknowledge Philippe Taquet (MNHN) for providing historical information on the paratype and the French expeditions in Gadoufaoua, and Ronan Allain (MNHN) for sending us a copy of the field map of the paratype and the permission to use it. The Fondazione Giancarlo Ligabue (formerly Centro Studi e Ricerche Ligabue) gave us additional information about the Italian expeditions in Gadoufaoua during the 1970s. Karen Poole (Stony Brook University, New York) and Andrew McDonald (St. Louis Science Center, St. Louis) shared information on the codings of
The authors declare there are no competing interests.
The following information was supplied regarding data availability:
The raw data is included in the manuscript in the Results section.