A study by researchers from the American Museum of Natural History, Stony Brook University, the University of Arizona, and Arcadia University suggests that zelestids were not closely related to modern placental mammals, as many scientists had believed for decades. Instead, the fossils indicate that zelestids belonged to a separate lineage of early mammals.

The new interpretation overturns conclusions based largely on isolated fossilized teeth.

“This discovery shows why fieldwork remains essential to understanding the history of life,” said lead author Andrés Jarombardo, who discovered the specimen as a graduate student during a museum-sponsored expedition in 2004. “It was the thrill of a career to discover a new species so perfectly preserved, one that also solves long-standing scientific questions, and a reminder that the Gobi Desert, famous for its fossils, continues to change science.”

Fossil teeth led scientists in the wrong direction

Zelestids have been known to paleontologists for nearly 40 years, but most discoveries consisted of isolated teeth or incomplete remains. Their teeth were unusually specialized for eating plants compared with the sharp teeth used by many insect-eating Cretaceous mammals.

Those characteristics led some paleontologists to suggest that zelestids represented an unknown group of ungulates. Scientists also debated whether they were placental mammals that lived during the Cretaceous Period—the group that includes humans and most modern mammal species—or whether they belonged to a separate lineage that independently evolved similar features.

Molecular clock studies have supported the possibility that placental mammals arose long before the end of the Cretaceous Period. However, the newly described fossil provides physical evidence pointing in a different direction.

“Molecular models based on living animals can predict the past, but fossils provide the evidence needed to test those predictions,” said study co-author Paul Velasco, a comparative biologist at Arcadia University.

Nearly complete skeleton reveals the animal’s identity

The new species, tamilkan balkarseri, was discovered in the eastern Gobi Desert. It is the most complete gerestid specimen ever found, preserving much of the animal’s skeleton and allowing scientists to reconstruct what members of this mysterious group looked like.

The animal had low, rounded teeth associated with zelestids. It also had long, continuously growing incisors and distinctive cranial features found in zalamudestoids, a group of small, shrew-like insectivorous mammals.

Its limbs provided another important clue.

“The hind legs are long and slender, with distinctive ankles that make them unmistakably zalamudestoid,” said study author Sean Zack, a paleontologist at the University of Arizona.

Taken together, these features led researchers to conclude that zelestids were not placental mammals and were not particularly closely related to them. Instead, they were part of the zalamudestoid group.

Similar teeth may be an example of convergent evolution

The discovery suggests that Cretaceous mammals may not have been as anatomically diverse as scientists had inferred from their teeth alone. The rounded teeth associated with zelestids resemble those of many plant-eating mammals and were likely an example of convergent evolution.

Convergent evolution occurs when unrelated animals independently develop similar traits as they adapt to comparable lifestyles.

Why teeth cannot always reveal an extinct animal’s whole body

The fossil highlights the risks of reconstructing extinct animals from only one part of the skeleton.

“More than 200 years ago, the French naturalist Georges Cuvier famously claimed that scientists could predict the anatomy of an entire animal from a single tooth,” said study co-author Maureen O’Leary, a paleontologist at Stony Brook University and a research fellow at the museum. “Teeth provide some of the most informative information of any available fossil, but this study shows that teeth don’t always tell us about the whole animal.”

The animal measured approximately 6 to 7 inches from head to tail. Its unusually long hind legs gave it a rabbit-like appearance. Paleontologists sometimes informally refer to zelestids as the “Cretaceous rabbits” because they resembled rabbits, even though they are distantly related to modern rabbits.

The Gobi Desert continues to reshape mammalian evolution

The specimen also demonstrates the scientific value of Mongolia’s Gobi Desert. The region is one of the few places where gerestids have been preserved, along with fossil remains found in Kazakhstan, Kyrgyzstan, and Uzbekistan.

“The Gobi is one of the few places where such amazingly complete Cretaceous mammals are routinely recovered,” said co-author Michael Novacek, curator of the museum’s paleontology department. Since 1990, he has co-led an annual expedition to the Gobi in partnership with the Mongolian Academy of Sciences.

“These extraordinary fossils continue to change our understanding of mammalian evolution,” Novacek said.

Researchers also placed the new species within the mammalian evolutionary tree using MorphoBank, a comprehensive research platform for studying evolutionary relationships among mammals. First published in 2013, the database has expanded for more than a decade as scientists have added important fossil species.

It currently provides the most comprehensive framework assembled for analyzing relationships among early mammals.

Study authors and research support

Other authors of the study include Eva Hoffman of Yale University.

Fieldwork and laboratory research in the Gobi Desert were supported in part by the Margaret and Will Hearst Paleontological Research Fund and the Museum’s Frick Institute Endowment. The research was also supported by the National Science Foundation under grant numbers MRI-R2 0959384, EAR 2506729, and EAR 2506727.