An Early Triassic Lystrosaurus Embryo
The genus Lystrosaurus is one of the best-known dicynodonts (Dicynodontia) from the latest Permian to Early Triassic, with an exceptionally abundant fossil record in the Karoo Basin of South Africa. Yet among these numerous specimens, no individual had been confidently identified as an embryo. Among the smallest Lystrosaurus specimens currently known, one specimen, NMQR 3636, preserves the entire animal tightly curled into an ovoid shape. Following high-resolution computed tomography and synchrotron X-ray imaging, a recent study interpreted it as a possible Lystrosaurus embryo that died while still inside the egg.

NMQR 3636 was discovered in 2008 on a farm in the Free State Province of South Africa. The specimen comes from strata dating to the Induan stage of the Early Triassic. It preserves an almost complete but partly disarticulated skeleton, including the skull, mandible, most of the vertebral column, some ribs, the forelimbs, pelvis, and left femur. Its basal skull length is only 3.45 cm.
Body size alone cannot determine whether a fossil represents an embryo, so the researchers compared its skeletal development in detail with that of two other small specimens, BP/1/4011 and BP/1/9332. Both BP/1/4011 and BP/1/9332 contain small unerupted tusk buds within the maxillary alveoli, whereas the corresponding alveoli in NMQR 3636 are empty. An ossified mesethmoid bone can also be identified in the other two specimens, while no mesethmoid is preserved in NMQR 3636. The researchers suggested that it may not yet have ossified, although it could also have been lost during fossilization.

Differences at the back of the skull likewise reflect different developmental stages. In Lystrosaurus, the cranial bones progressively co-ossified during growth, with the occipital bones among the last to do so. In BP/1/9332, the occipital bones remain in their normal anatomical positions and are articulated, indicating a more advanced stage of development. By contrast, several occipital and basicranial bones in NMQR 3636 remain loose and displaced. The limbs show the same pattern. The epiphyses of the humerus and femur in NMQR 3636 are incompletely ossified, while the carpals of the hand are still nearly spherical because of their early stage of ossification.
In NMQR 3636, the sacral vertebrae and the ribs connecting them to the pelvis had not yet formed a stable sacral structure, and several of these elements are separated from one another. The vertebral centra and neural arches also remain unfused. In BP/1/9332, by comparison, the pelvis is preserved in full anatomical articulation, and six sacral vertebrae with their associated ribs form a substantially more stable connection with the pelvic girdle. These differences indicate that NMQR 3636 represents the least developmentally advanced of the three specimens.
The left and right halves of the Lystrosaurus mandible meet at the midline to form the mandibular symphysis, but this structure had not yet fully closed in NMQR 3636. Synchrotron images reveal a conspicuous gap between the paired splenial bones, while the suture between the dentaries also remains open. The margins of the gap are smooth and intact, unlike a fracture produced by damage to the fossil, leading the researchers to interpret it as a genuine developmental feature. For comparison, in embryos of the modern African spurred tortoise (Centrochelys sulcata), another beaked tetrapod, the anterior part of the mandible progressively closes during embryonic development.
The vertebral column of NMQR 3636 curves around the outer margin of the body, with the head, trunk, and limbs tightly packed into a space approximately 7.3 cm long and 5.5 cm wide, producing an approximately ovoid outline. No mineralized eggshell was found around the fossil, so the thickness and composition of the Lystrosaurus eggshell cannot be directly determined. The researchers proposed that the egg may have had a soft, leathery shell that failed to leave an easily recognizable fossil trace after burial, although this interpretation remains speculative.


Based on the space occupied by the curled skeleton, the researchers reconstructed the original egg as an ellipsoid with an estimated volume of about 115 cm³. Assuming a density close to that of water, this corresponds to a mass of approximately 115 g. Because the embryonic skeleton is incomplete and the original egg would also have contained yolk and other tissues, this value is regarded as a conservative minimum estimate.
Estimates of adult Lystrosaurus body mass vary considerably, so the researchers compared three published estimates: 8.825, 18.511, and 50 kg. Regardless of which estimate is used, the reconstructed egg is relatively large for the body size of the parent. This led the authors to suggest that Lystrosaurus young underwent a relatively advanced degree of development within the egg and may have become mobile and capable of feeding independently soon after hatching. BP/1/9332 provides some evidence consistent with this interpretation. Although it was still a very young individual, its mandibular symphysis had already closed, its pelvis and limbs were more extensively ossified than those of NMQR 3636, and the skeleton was preserved in a splayed-out posture rather than tightly curled as it would have been inside an egg.
The specimen also allowed the researchers to compare Lystrosaurus with Cynodontia, the synapsid lineage more closely related to mammals. An Early Jurassic specimen of the tritylodontid cynodont Kayentatherium preserves an aggregation of 38 extremely young individuals. Based on the size of these perinates, the eggs of Kayentatherium may have been much smaller relative to adult body size than those of Lystrosaurus. From relatively early-diverging synapsids such as Lystrosaurus to later groups increasingly close to mammals, the amount of development and nutritional investment occurring inside the egg, as well as the ways in which young were supported after hatching, likely underwent substantial evolutionary changes.


Author: Shui-Ye You
Reference:
Benoit J et al. (2026). The first non-mammalian synapsid embryo from the Triassic of South Africa. Plos One.





Comments