Histomorphogenesis of embryos of Upper Jurassic Theropods from Lourinha (Portugal)

de Ricqlès, A., Mateus O., Antunes M. T., & Taquet P. (2001).  Histomorphogenesis of embryos of Upper Jurassic Theropods from Lourinha (Portugal). Comptes Rendus De L Academie Des Sciences Serie Ii Fascicule a-Sciences De La Terre Et Des Planetes. 332, 647-656., Jan


Remains of dinosaurian embryos, hatchlings and early juveniles are currently the subject of increasing interest, as new discoveries and techniques now allow to analyse palaeobiological subjects such as growth and life history strategies of dinosaurs. So far, available ‘embryonic’ material mainly involved Ornithopods and some Theropods of Upper Cretaceous age. We describe here the histology of several bones (vertebrae, limb bones) from the tiny but exceptionally well preserved in ovo remains of Upper Jurassic Theropod dinosaurs from the Paimogo locality near Lourinhã (Portugal). This Jurassic material allows to extend in time and to considerably supplement in great details our knowledge of early phases of growth in diameter and in length of endoskeletal bones of various shape, as well as shape modelling among carnivorous dinosaurs. Endochondral ossification in both short and long bones involves extensive pads of calcified cartilages permeated by marrow buds. We discuss the likely occurrence of genuine cartilage canals in dinosaurs and of an avian-like ‘medullary cartilaginous cone’ in Theropods. Patterns of periosteal ossification suggest high initial growth rates (20 μ m·day−1 or more), at once modulated by precise and locally specific changes in rates of new bone deposition. The resulting very precise shape modelling appears to start early and to involve at once some biomechanical components.



Related External Link

1. Introduction

During the last ten years or so, the value of bone histology to interpret palaeobiological issues such as growth dynamics and life history traits among fossil vertebrates has been widely recognized [4] and [12]. This is especially true for the study of (non-avian) dinosaurs [10][31][32] and [33]. However, such studies have been traditionally hampered by several circumstancial factors, including availability of fossil material, intrinsic bone tissue variability within the skeleton, lack of standardization of techniques and vocabulary [34]. Accordingly, few studies, so far, could benefit from complete growth series [19] or could analyse important factors of histological variability linked to anatomical sampling [18]. On the other hand, the increasing number of discoveries of juvenile dinosaurs and of in ovo material has allowed research to focus on issues of dinosaurs growth and life history strategies, in connection with hotly debated issues such as dinosaur physiology and the close phylogenetic relationships of Theropods to birds  [5][13] and [29]. This has stimulated much needed comparative and experimental studies of modern animals (notably birds) histogenesis, in order to get quantitative values on the relationship between bone tissues typologies and growth rates, as a basis towards extrapolation to fossil material via simple uniformitarian hypotheses [7] and [8].

‘Dinosaur embryos’ have been recognized for a long time [35], but modern research has unveiled in ovo or presumably hatchling material in unprecedented number and diversity [5]. However, most of the material so far studied is of Upper Cretaceous age and belongs to Ornithopods (dryosaurs, hypsilophodontids, hadrosaurids, and others), although some sauropod (titanosaur?), troodontid and oviraptorid and therizinosaur embryonic material have also been described  [1][3][9][14][16][17][28][38] and [39].

According to palaeontological practice, we will use here the term ‘embryos’ in a wide sense [5] to describe fossil skeletal material found in ovo, regardless of the fact that this material, in all likelyness, is not very far from the hatchling period and has far overpassed the actual embryonic stage of ontogenetic development, as defined by embryologists.

The discovery of exquisitely preserved dinosaur embryos in Paimogo, near the city of Lourinhã, north of Lisbon (Portugal) [26] is an event of unprecedented interest for the study of early ontogeny of dinosaurs for several reasons. First the material is of Upper Jurassic age [26] and [27], which extends the temporal range of ‘dinosaurian embryology’ almost 80 million years back. Second, the material belongs to the Theropodan clade, the spectacular bipedal carnivorous dinosaurs now widely regarded as closely linked to the ancestry of birds [13] and [29]. In view of the incredible amount of minute information brought by the histological study of such a material, it is not within the scope of this paper to deal with detailed descriptions. Accordingly, only the most salient aspects need to be pointed out here.

2. Materials and methods

Lourinhã's embryos are most often found as tiny isolated skeletal elements within eggs or in close association with them, within the matrix between the eggs of clutches (or clutch clusters) discovered in 1993 [26]. Published data [26] and [27] and works currently in progress demonstrate that the embryos are Theropods and morphological details allow to put them close to Allosaurus. A smallish-sized allosauroid (perhaps an incompletely grown individual) has been discovered just six kilometres from the Paimogo eggs locality in contemporaneous layers (Sobral unit: Late Kimmeridgian/Tithonian) and described asLourinhanosaurus[23]. A bivariate analysis of the proportions of the embryo's vertebral centra strongly supports affinities with this genus rather than with e.g. Allosaurus. However, for reasons discussed below, it is still safer not to refer the eggs and embryos to this or another taxon at the genus and species ranks.

After anatomical study, measurements and photography, the bone fragments selected for the histological study were embedded in epoxy resin and sub serially sawed on a thin low speed diamond powder disk. The sections were then ground and polished and observed under the compound microscope in ordinary and polarized light [41]. Six undetermined long bone shafts (cross sections), three femora and four vertebrae (longitudinal and cross sections) were used, yielding a total of seventy thin sections (table).

Table. Material and thin sections used for the histological study presented here.Liste du matériel étudié histologiquement et sections (lames minces) présentées dans ce travail.

Undetermined long bone shafts (stylo- or zeugopodials) / Os longs indéterminés (diaphyses: éléments stylo- ou zeugopodiens)

98.1 2 cross-sections / 2 sections transversales
98.2 6 cross-sections / 6 sections
98.3 3 cross-sections / 3 sections
98.4 3 cross-sections / 3 sections
98.5 4 cross-sections / 4 sections
Determined long bones / Os longs déterminés
84.1: right femur proximal half / moitié proximale fémur droit 5 longitudinal sections of epiphysis / 5 sections longitudinales de l'épiphyse 3 cross-sections of shaft / 3 sections transversales de la diaphyse
101.1: right femoral shaft / diaphyse fémur droit 4 cross-sections / 4 sections transversales
110.1: femur proximal region / région proximale fémur 6 longitudinal sections of epiphysis / 6 sections longitudinales de l'épiphyse 3 cross-sections of metaphysis / 3 sections transversales de la métaphyse
Vertebrae / Vertèbres
99.1: dorsal (?) vertebrae / vertèbre (dorsale?) 9 cross-sections of centrum / 9 sections transversales du centrum
99.2: vertebra / vertèbre 6 cross-sections of centrum / 6 sections transversales du centrum
99.3: vertebra / vertèbre 8 cross-sections of centrum / 8 sections transversales du centrum
99.4: caudal (?) vertebra / vertèbre (caudale?) 6 longitudinal sections of centrum / 6 sections longitudinales du centrum

3. Periosteal bone ossification and growth in diameter of the shaft

All shaft cross-sections show a bony cortex quite typical of ‘embryonic bone’ surrounding a more or less well-marked marrow cavity (figure 2.8). The cortex, entirely formed by primary (not reconstructed) bone, is organized in a variable number of thin trabeculae set apart from each others by numerous vascular spaces, so as to form a rather cancellous or spongy tissue – porosity at least 50 % – (figure 2.4). Bone trabeculae are composed of a bone matrix isotropic under polarized light, with rather plump cell lacunae. Deposition of primary osteonal bone is still entirely lacking. The shape, number and organization of the bone trabeculae are quite variable from bone to bone and even on the same section. Some have a very regularly concentrical orientation, inducing a laminar organization [7][8] and [12] of the cortex (figure 2.8), others have a quite irregular pathway, inducing a reticular organization of the tissue [7][8] and [12] (figure 2.5). Lateral drift of the marrow cavity is often obvious, producing an offseting relative to the cortex, itself growing asymetrically to match the lateral drift. Generally, the periphery of the marrow cavity shows clear images of resorption (Howship's lacunae), evidence for its diametral growth by erosion of the innermost cortex, whose vascular spaces open directly in the marrow cavity (figure 2.5). The marrow cavity may be free of bony trabeculae, giving to the bone the tubular structure diagnostic of Theropods [29] (figure 2.8). However, some sections have a well-developed system of cancelous bony trabeculae within the marrow cavity. They are quite distinct from the periosteal bony trabeculae of the surrounding cortex and are the result of the initial endochondral ossification within the shaft of the former cartilaginous anlage (figure 2.3). Indeed, within some trabeculae,globuli ossei close to the tiny remnants of calcified cartilage matrix are still left, supporting their early endochondral origin [12]. On such sections, the place of the Kastchenko's line [12], marking the periphery of the initial cartilaginous anlage, before the onset of periosteal ossification on the shaft, could still be deciphered.

At the periphery of the cortex, the vascular canals also freely open into the sub periosteal space (figure 2.4), and the outermost bone trabeculae form protruding spikes, evidence of new forming bone being actively laid down at the periphery to allow bone diametral growth and compensate for the marrow cavity increase in diameter [7] and [8] (figure 2.5).

4. Endochondral ossification and longitudinal growth

The ends of the long bones are all formed by a well-developed coating of hypertrophied calcified cartilage (Figure 1 and Figure 1). This cartilage contains a great amount of longitudinally oriented tubules, or pipes, which open either in the marrow cavity or at the surface of the cartilaginous ‘epiphysis’ [17] and [20]. There is little doubt that some of those tubular spaces originate from the marrow cavity and worked as erosion bays carving into the cartilage towards the epiphysis. Their walls demonstrate local erosion of the cartilage and deposition of a thin coating of endochondral bone [1][3][19] and [20] (figure 1.3). Towards the diaphysis, widening and fusion of the erosion bays lead to a bony spongiosa, where small islands of cartilage associated to globuli ossei may be observed within the bony trabeculae, evidence of the endochondral origin of the tissue, as also observed on some shaft cross sections ( Figure 2 and Figure 2) – see above. Those metaphyseal regions are finished outwardly by a thin cortex of periosteal origin (figure 2.6) that ends up towards the epiphysis to form the ‘encoche d'ossification’ [12]. The tubules located most distally within the cartilage of the epiphysis might well be genuine cartilage canals [40], ontogenetically independent from the erosion bays, as they are developed among birds [40], rather than merely erosion bays developed from the marrow cavity [20]. The first interpretation is supported here by the widening in diameter of those canals towards the epiphyseal surface ( Figure 1 and Figure 1). However, the other interpretation cannot be ruled out because the periphery of those canals is also involved in cartilage erosion and early endochondral bone deposition (figure 1.6). This clearly indicates that if those ‘pipes’ are indeed cartilage canals of the avian type, they nevertheless have been already colonized by connective-vascular tissues from the marrow at this ontogenetic stage [15][19][20] and [40].

Some metaphyseal regions show a large central ‘well’ filled up by matrix, lined by thin bony trabeculae and sometimes containing small isolated islands of calcified cartilage (Figure 2 and Figure 2). This suggests the occurrence of a ‘cartilaginous medullary cone’ of the avian type [15], mostly formed by uncalcified cartilage (and hence not preserved by fossilization), as already suggested for the Cretaceous Theropod Troodon  [20],[38] and [39].

5. Local-specific characters and shape modelling

Anatomical shape and size differences within the skeleton are all the morphological consequences of histogenetic processes that may be observed or deciphered, even in fossils, by histological examination [34]. In that context, Lourinhã's embryonic material will allow observations of unprecedented precision and detail for Theropods. Only some preliminary comments are in order here. Morphological differences between long (limb) and short (vertebrae) endoskeletal bones seem directly linked (a) to the thickness of the calcified cartilage pads observed in fossils (itself probably correlated to the thickness of the uncalcified growth cartilage not preserved in fossils), (b) to the orientation and number of the intracartilaginous erosion bays and (c) to the relative speed of periosteal versus endochondral ossifications [2][20][22][30][36] and [37] (compare Figure 1 and Figure 1).

In the femur, differentiation of an extensively off-centred articular head set laterally is in part caused by an extensive metaphyseal ‘region of undercutting’ [15], already well-differentiated and apparently very active at the ontogenetic stage available. Similarly, the proximal femoral extremity is already anatomically and histologically quite distinct from the well-developed internal trochanter of the embryos. In the proximal metaphysis of the femur, it is also possible to survey in detail the histological process responsible for the internal trochanter (figure 2.2) and fourth trochanter formations, and for the sequential relocation of the structures as growth takes place [11] and [20]. Finally, some bone trabeculae of the spongiosa within the inner trochanter already show a peculiar spatial organization, forming locally a series of ‘V’ or ‘W’, suggesting together a ‘Warren beam’ structure (figure 2.2). To that extent, it is likely, as already suggested [6] and [21]that the osseous skeleton differentiates very early under biomechanical strains that are integrated at once in its structure.

6. Concluding remarks

The types of periosteal bone tissues forming the cortex of the long bone shafts, and to a lesser extent, the lateral walls of the vertebrae, are quite typical of a primary (i.e. not reconstructed), hypervascularized ‘embryonic’ bone (figure 2.1), and likely to have been laid down at very high speed, probably higher than at least 20 μ m·day−1[7] and [8]. The epiphyseal structures, at least for the long bones, also suggest the possibility of an extremely active growth in length [1][2][3] and [40]. The likely occurrence of genuine cartilage canals and probably also of a ‘medullary cartilaginous cone’ suggests a close relationship with the avian condition of long bone histogenesis [20], in agreement with phylogenetic conclusions based on independent data [13] and [29].

Finally, apart from their exquisite anatomical and histological preservation, one of the most striking aspect of the Theropod embryos of the Jurassic of Lourinhã is the precision of the bone-specific anatomical details and their early differentiation. Those details (e.g. Figure 1Figure 1 and Figure 2) already show, at a much-reduced scale, Theropod-specific character states that had been formerly observed and analysed, on adult skeletons the size of which could vary two orders of magnitude (100 factor) from the embryos. Of course, the histological construction of such anatomical details is fundamentally different in the embryos and adults, but the stability of such morphological characters all along the extended Theropod ‘ontogenic trajectory’ is striking. This suggests that the (genetic) control on shape is at the same time precocious, powerful and permanent, which is not contradictory with the influence of mecanomorphic (epigenetic) factors, the possible influence of which has been stressed above [6] and [21].

In spite of the very detailed morphological set of character states already available at the early ontogenic stage offered by the Lourinhã embryos, the non-availability of a growth series does not allow to know whether those details were retained along the ‘ontogenic trajectory’ according to iso- or allometric relationships, and accordingly to securely deduce adult shapes and proportions from the embryos. A bivariate analysis of vertebral centra among AllosaurusLourinhanosaurus and the embryos demonstrates similar proportions inLourinhanosaurus and the embryos while the proportions differ between them and Allosaurus. On this basis, the embryos could be tentatively referred to the Portuguese genus under the hypothesis of an isometric growth of the vertebral centra. This hypothesis could be amenable to a test if an isometric growth of the centra was indeed demonstrated in a Theropod where a growth series is available, e.g. Allosaurus from the Cleveland Lloyd Quarry. The final systematic status of the embryos will also have to take into account that several taxa of adult Theropods are now known from localities neighbouring the eggs' site  [24] and [25].


We thank Mrs M.-M. Lath and Dr H. Francillon-Vicillot (Paris-7 University, UMR 8570) for their technical assistance with thin sectioning and plates computer processing. This work was supported by Project No. 225 BO «Échanges scientifiques franco-portugais».