The origin of turtles has always been a bit of a puzzle for scientists who study the evolution of animals. To this day, where they fit in the tree of life remains a highly debated topic.
The evolutionary relationships of most vertebrate groups are well understood. Thanks to genetic and morphological (anatomical, body shape) data, even animals with highly specialised skeletons can be clearly placed on the animal family tree. Examples include whales or birds.
Turtles, however, have long remained an exception. Genetic studies identify them as relatives of the so-called archosaurs. This is a group that includes modern birds and crocodiles as well as extinct reptiles like dinosaurs and pterosaurs. But the fossil record seemed to tell a different story. Living turtles and their fossil relatives were so specialised that they offered few clues that would link even the oldest turtle fossils to other reptile groups. Or so scientists thought.
Our international team of palaeontologists has now provided a comprehensive reassessment of the turtle’s place in the animal world. Our analysis sheds new light on the relationships among primitive turtles. It confirms that Eunotosaurus africanus, a fossil from South Africa and Malawi, which was presumed to be a “proto-turtle”, is not a direct ancestor of modern turtles. Instead, this animal is very distantly related to modern reptiles, finding its deep root among much older reptilian ancestors that have no modern representatives.
Based on anatomy, the phylogenetic analysis also provides the first robust support from fossil studies for the close relationship between turtles and the archosaur (bird-crocodilian) lineage.
For more than 20 years, genetic data and anatomical data reached different conclusions about the relationships of turtles. Now they agree.
Comparing reptile anatomy
Fifteen researchers from South Africa, the US, UK, France and Germany participated in the study. Their combined expertise included:
-
computed tomography (CT) technology (advanced x-rays)
-
reptilian anatomy and phylogenetics
-
Permo-Triassic stratigraphy (the study of rock layers where fossils are found).
The combination was critical to obtain these groundbreaking results. Collection staff from the Evolutionary Studies Institute, Iziko South African Museum, National Museum, Albany Museum and Council for Geoscience in South Africa were also instrumental in enabling access to the specimens.
The team painstakingly compiled anatomical comparisons across more than 200 fossil reptile species. We hoped to find previously overlooked similarities between early shelled turtles, their shell-less predecessors, and other early reptiles. Comparisons of the bones that frame the brain cavity were particularly important. These couldn’t previously be seen by scientists, but with powerful CT scanning methods their anatomy was laid bare.
Artist: Gabriel Ugueto, Author provided (no reuse)
Particularly surprising was what we learned about Eunotosaurus africanus, a 30cm-long burrowing reptile that lived in southern Africa some 260 million years ago. Previous studies considered it as the oldest known member of the turtle family, or a “proto-turtle”. Its broadened trunk and wide ribs looked something like a turtle shell. We studied almost all of the material of Eunotosaurus available in South African collections to address this idea once again.
Our working group at the Evolutionary Studies Institute studies some of the oldest rock layers from the Karoo Basin of South Africa, where Eunotosaurus is found. If Eunotosaurus was indeed a “proto-turtle”, we’d expect to find the forerunners of living lizards, crocodiles or birds (that is, reptiles) in these same layers. Paradoxically, we’ve found no other close relatives of modern reptiles at all. This made us suspect that even if turtles are ancient relatives of living birds and crocodilians, perhaps Eunotosaurus was no “proto-turtle” at all.
One breakthrough was reconstructing the bones of the braincase (housing the brain and ear) from high-resolution x-ray images of fossil and living reptiles. By peering inside the skull of Eunotosaurus, and comparing its bones with those of undisputed fossil turtles, we could see previously out-of-reach aspects of their anatomy for the first time.
These x-ray scans revealed the very primitive anatomy of Eunotosaurus. For example, it has bones in the back of the skull that were lost in turtles and all living reptiles. Features like a slender ear bone (the stapes) and the hooked fifth toe that are present in many living reptiles and other fossil turtles were completely lacking in Eunotosaurus. In contrast, the braincase of unambiguous fossil turtles, such as Proganochelys quenstedti, shared a suite of characteristics that are found in the ancestors of crocodilians and birds, but absent in Eunotosaurus.
These lines of evidence provides firm anatomical support that turtles are the closest living relatives of archosaurs. When Eunotosaurus was considered a “proto-turtle”, many of these features were considered to have evolved independently in the turtle lineage. Now, we show that turtles share these features with their archosaur relatives because they inherited them from a common ancestor.
Author provided (no reuse)
These new results now place the origin of turtles where it fits better with both fossil and genetic data. When geneticists study living turtles, they compare their DNA to modern birds, crocodiles and lizards to infer evolutionary relationships. Our fossil findings now align with what those genetic comparisons have been suggesting all along: turtles branched off from the same ancestor that gave rise to crocodiles and birds.
Instead of being a living group of relics with ancestors present in the Middle Permian, turtles, like other modern reptiles, diversified and evolved their shell in the Triassic Period, approximately 20 million years after Eunotosaurus was already extinct.
With turtles now firmly placed among their closest living relatives, palaeontologists will need to reassess other long-standing questions about reptile evolution. Advanced imaging techniques like computed tomography should now be applied to other enigmatic fossil groups, potentially clarifying their evolutionary relationships.
Our work highlights the fact that overlooked early reptile fossils, particularly those found in the South African fossil record, may hold the key to understanding reptile relationships.
