Ancient Footprints Reveal Tiny Mammal Relatives Walking Jurassic China

In the summer of 2019, researchers exploring sandstone outcrops near two villages in Zizhou County, in northern China’s Shaanxi Province, noticed something small and easy to overlook: delicate, five-toed footprints pressed into slabs of rock that had been quarried and moved by local residents. Those unassuming traces have now turned out to be a scientific first. A new study published in Fossil Record describes the tracks in detail and concludes that they most likely belong to small mammaliaforms — the close relatives of mammals that scurried beneath the feet of dinosaurs some 160 million years ago. The find marks the first record of the ichnogenus Ameghinichnus in Asia and, remarkably, sits alongside what may be one of the earliest pterosaur track fossils ever found in China.

The tracks come from two sites, Wangzhuang and Longweimao, roughly 12 kilometres apart in Dianshi Town, Zizhou County, within the southeastern part of the vast Ordos Basin. The region has long been a treasure trove for vertebrate trace fossils. The very first tetrapod tracks ever reported in China were dinosaur footprints found in Jurassic strata at Dongshanya, Shenmu County, in 1929, and discoveries have accumulated steadily since 2010 across Shaanxi localities including Shenmu, Tongchuan, Shangluo, Jingbian, Ansai and Ganquan. The Zizhou area itself first gained attention when dinosaur tracks were spotted on building stones used by villagers, eventually revealing an assemblage of theropod, basal ornithischian and thyreophoran footprints. The new mammaliaform tracks add an entirely different body size class to that ancient fauna.

The footprints are preserved in the Middle Jurassic Yan’an Formation, one of the most important coal-bearing rock units in the Ordos Basin, a sprawling sedimentary basin covering roughly 250,000 square kilometres atop the western North China Craton. The formation, 200 to 300 metres thick, was deposited under relatively stable tectonic conditions as fluvial and lacustrine-deltaic systems spread across a subsiding plain. Its strata alternate between sandstones, mudstones, shales and coal seams, and its fossils — dominated by ginkgophytes and ferns, along with bivalves, ostracods, fish and pollen — paint a picture of a waterlogged, vegetated landscape. While the exact age of the Zizhou track-bearing horizon has been debated, with some researchers favouring the underlying Early Jurassic Fuxian Formation, pollen samples collected near Wangzhuang Village point firmly to the Middle Jurassic Yan’an Formation.

The key specimen, catalogued as YPC-ZZG1, is a slab from the Wangzhuang site preserving seven tracks, six of which form a trackway of three consecutive pairs of hand-like (manus) and foot-like (pes) prints. All are convex natural casts, pentadactyl and plantigrade, meaning the animals walked with the full sole on the ground. The manus tracks average 4.0 centimetres long and 5.6 centimetres wide, with a broad total digit spread of about 150 degrees and the middle digit, III, projecting furthest. The pes tracks, averaging 3.9 by 4.9 centimetres, show a narrower spread of about 83 degrees, with digits II through IV nearly parallel and pointing forward, and clear impressions of the metatarsal pads behind them. Both sets of prints rotate inward toward the animal’s midline, and the alternating manus-pes arrangement, with the pes set slightly behind and inside the manus, is characteristic of a four-legged walker.

To capture every detail, the team documented the slabs using 3D photogrammetry, stitching together between 29 and 41 overlapping photographs taken with a digital camera under controlled lighting. The resulting digital surface models, processed in software including Agisoft Metashape, Meshlab, CloudCompare and Paraview, yielded false-colour elevation maps and contour images that reveal subtle features invisible to the naked eye — digital pads, bilobate heel notches and, crucially, tapered digit tips ending in distinct claw marks. Measurements followed standard ichnological conventions, and the three-dimensional model has been made publicly available through FigShare, allowing other researchers to re-examine the tracks without handling the fragile original.

Comparing these features with named ichnogenera proved to be a process of elimination. The tracks differ from Brasilichnium and Eotetrapodiscus, which are typically four-toed, and from Malutitetrapodiscus, made by hopping animals. Pooleyichnus, another five-toed candidate, rests on a single poorly preserved type specimen and shows a more widely splayed digit pattern with an elongated third digit. Synaptichnium, generally attributed to archosauriform reptiles, has a long, slender, outwardly rotated pes with a different digit-length sequence. Triassic contenders such as Procolophonichnium, made by small parareptiles, and Capitosauroides, attributed to therocephalian therapsids, also fail to match: the former lacks the long second digit and clear palm impressions of the new tracks, while the latter shows broad, rounded digit tips without claws and strong outward rotation.

Ameghinichnus, first described from the Middle Jurassic La Matilde Formation of Patagonia, Argentina, in 1961, emerged as the closest match. It is characterised by quadrupedal trackways with nearly symmetrical, five-toed manus and pes prints of low length-to-width ratio, arranged alternately along the midline — features the Zizhou tracks share. But the Chinese material is also distinctive. The tracks are substantially larger, roughly four to five times the size of those of the type species A. patagonicus, whose manus prints average only about 8 millimetres long. The digit tips taper with claw marks rather than ending in rounded, bulbous tips, and both manus and pes rotate inward, whereas in A. patagonicus the pes rotates outward and oversteps the manus. Differences from A. manantialensis include much wider digit divergence in the manus, and from A. mirabilis, a lack of the near-identical manus and pes seen in that South African species. Given these discrepancies and the limited material available, the authors provisionally assigned the tracks to Ameghinichnus isp., an undetermined species within the genus, rather than erecting a new ichnospecies.

Who made the tracks remains an open question, and one the study addresses with appropriate caution. Since Casamiquela’s original work, Ameghinichnus has been attributed at various times to marsupials,

Subject of Research: Middle Jurassic mammaliaform and pterosaur trace fossils from the Yan'an Formation in Shaanxi Province, China

Article Title: New Middle Jurassic mammaliaform tracks from northern Shaanxi Province, China

Article References: Xing, L., Lallensack, J. N., Yin, H., Klein, H., Romilio, A., Gao, J., & Qi, Q. (2026). New Middle Jurassic mammaliaform tracks from northern Shaanxi Province, China. Fossil Record, 29(1), 37-56. https://doi.org/10.3897/fr.29.169194

Image Credits: AI Generated

DOI: 10.3897/fr.29.169194

Keywords: Ameghinichnus, mammaliaformes, ichnology, Middle Jurassic, Yan'an Formation, Ordos Basin, pterosaur tracks, Pteraichnus, paleoecology, trace fossils, Shaanxi, China