Re-evaluation of the eggshell structure of eggs containing dinosaur embryos from the Lower Jurassic of South Africa

The first clutch of Mesozoic eggs containing embryos was collected more than twenty years ago from the Rooidraai locality in Golden Gate Highlands Park, South Africa. The Elliot Formation crops out throughout the area, and although the sediments exposed at Rooidraai were once thought to be Late Triassic in age, they are now regarded as Early Jurassic. The Elliot Formation preserves a rich terrestrial vertebrate fauna, with the remains of the prosauropod dinosaur Massospondylus carinatus found to be relatively common at Rooidraai. The Rooidraai embryos were originally attributed to dinosaurs on the basis of cranial features of one of them. The eggs were later identified as ‘probably prosauropod’, but more interestingly, the eggshell structure was described as sharing characteristics with the eggs of crocodilians. The shells were described as consisting of ‘broadly wedge-shaped and ill-defined units’ and so were thought to be ‘seemingly more similar to crocodilian than to avian eggs’. The reconstruction provided by Grine and Kitching reaffirmed these statements. If the fossil eggs and embryos from Rooidraai are indeed attributable to prosauropod dinosaurs, they would fill an enormous gap in our knowledge of eggs, eggshell structure, and early development in dinosaurs, as ornithopod eggs and embryos are now known only from Late Cretaceous hadrosaurs, theropod eggs with embryos from Late Jurassic and Late Cretaceous avetheropods (sensu Padian et al.), and sauropodomorph eggs with embryos from Late Cretaceous titanosaurs. It is now accepted that fossil eggshell structure has phylogenetic implications. For example, dinosaur eggshell can be distinguished from that of crocodilians by the presence of acicular (radial)-tabular wedges, and theropod eggshell from that of other dinosaurs by the presence of two structural zones. Therefore, the crocodilian-like units reported by Grine and Kitching in probable prosauropod eggshell contrasts with the current understanding of dinosaur eggshell structure and warrants further investigation of the Rooidraai eggs. Here we re-examine the structure of eggshell from the same egg that was described by Grine and Kitching in order to identify any diagnostic characteristics of the Rooidraai eggshell, and to determine whether these features are shared with the eggshells of the major archosaurian groups. The original clutch of six consisted of what in life were probably spherical to subspherical eggs. Eggshell samples were removed from BP/1/5347b (from next to the sample taken by Grine and Kitching) and examined both in thin section and using scanning electron microscopy (SEM). The thickness of the eggshell ranges from 75–120 μm. The microscopic analyses reveal no recognizable shell units that comprise archosaurian eggshell. The irregular blocky and wedge-like structures within the shell (as shown in Fig. 1A,B) may represent recrystallized calcite. X-ray diffraction analysis confirmed that the shell consists of calcite. The subspherical shape of the eggs is shared with those of titanosaurid sauropods and ornithopods among dinosaurs, whereas crocodilian and non-avian theropod eggs tend to be elongate. There appear to be no features, other than the shape and the calcite composition, that are diagnostic of the Rooidraai eggshell. The absence of shell units or zonation in the eggshell, which are present in the eggshell of all known archosaurian reptiles, implies that the Rooidraai eggs have been altered diagenetically. Furthermore, the preserved shell thickness is significantly less than what would be expected for a crocodilian or avian egg of an equivalent volume. Whether this represents Research Letters South African Journal of Science 98, July/August 2002 407