Japan's oldest mosasaur snag: a jawbone from a vanished inland sea
A single premaxilla pulled from a Kishiwada quarry 48 years ago has been re-identified as Japan's oldest known mosasaur, and possibly a species new to science.
On a gravel lot in Kishiwada, a coastal city of roughly 370,000 people pressed against Osaka Bay, a worker pulling rubble out of a quarry in 1978 unearthed a jawbone fragment about the length of a child's forearm and put it in a box. Forty-eight years later, a team of Japanese researchers has re-identified that fragment as a mosasaur premaxilla, the front of the upper jaw, and the oldest confirmed mosasaur specimen ever recorded from Japan.
The find, described in a 2026 publication by researchers from Okayama University of Science, Tokyo City University, the Natural History Museum, Kishiwada City and other institutions, sharpens the map of where these marine predators roamed in the late Cretaceous. It also raises a sharper question: whether the animal it came from is one science has already named, or a species new to the record.
The bone's importance is partly arithmetic. Kishiwada sits inland today, on the eastern edge of Osaka Prefecture, but in the late Cretaceous, the area sat beneath a warm, shallow sea that connected what is now central Japan to the broader Pacific margin. Marine reptiles that died there and fossilised in its sediments are therefore a chronological key to the region. A single dated specimen can push a national fossil record back tens of millions of years, which is what this one appears to do for mosasaurs in Japan, the first of its lineage recovered from Cretaceous rocks in the country.
The full anatomical case rests on the premaxilla itself. The fragment preserves the front of the upper jaw, parts of the tooth row, and small features along the bone's edge. Researchers compared it against named mosasaur taxa and concluded it does not cleanly map onto any of them.
What the jawbone actually preserves
A premaxilla is the paired bone at the very tip of the upper jaw, the part that holds the frontmost teeth. In mosasaurs, a family of fully marine lizards related to modern monitor lizards that dominated late Cretaceous seas worldwide, the premaxilla is diagnostic. Its shape, the way the tooth row sits inside it, and the presence or absence of small pits where sensory organs once sat together form a fingerprint that species can be told apart by.
The Kishiwada specimen is incomplete. It is a partial bone, not a skull. That has constrained what can be claimed. The team has not proposed a new species; the publication identifies the specimen to the family Mosasauridae and flags a possible new taxon for future work once more material is recovered. In paleontology, a single premaxilla with unusual features is a flag in the sand, not a formal name, and the paper's wording reflects that caution.
Japan has a long marine fossil record going back to the early Cretaceous, but mosasaurs are a different branch of that record. The country has yielded sea turtles, plesiosaurs, pliosaurs, and abundant shark teeth, with mosasaur teeth turning up sporadically. What makes the Kishiwada find notable is the jawbone context: teeth without skulls cannot fix a species, and a diagnostic cranial element can.
Where Kishiwada sat 85 million years ago
Kishiwada today is a commuter city between Osaka and Wakayama, hemmed in by the bay and the Kongō–Izumi mountains. In the Santonian to Campanian stages of the late Cretaceous, between roughly 86 and 78 million years ago, the region lay under water on the eastern edge of a continental shelf that ran along what is now the Japanese archipelago.
That shallow shelf was rich in ammonites, inoceramid clams, and marine vertebrates, the standard inhabitants of an epicontinental sea. Mosasaurs occupied the upper part of the food chain. The fossil beds near Kishiwada have produced other marine reptile material, including plesiosaur and sea turtle remains, which sets the ecological scene for a predator at the top of the local web.
The same sea-floor sediments that preserved a mosasaur premaxilla also record the dynamic geology of western Japan. The bed sits near the Median Tectonic Line, a major fault that has shuffled Cretaceous and younger rocks across the region. That context matters: it means researchers have to do careful work ruling out transport, reworking, or mixing of fossils from different ages in the same outcrop.
The counter-read on a single bone
Skeptics will note, fairly, that one partial premaxilla cannot anchor a new species on its own. The standard in vertebrate paleontology is comparison across multiple specimens, ideally with skull material and post-cranial elements. A lone jaw fragment is open to alternative explanations: it could represent a known mosasaur whose range in Asia has not been fully sampled, it could be a juvenile whose features differ from adults, or the apparent uniqueness could shrink once more material is found.
That counter-read is worth carrying. The authors have not overreached. They have identified the family, dated the rock, flagged a possible new taxon, and stopped short of naming it. The standing of the claim is therefore proportional to what a premaxilla can actually tell you: a lot about lineage and geography, less about species.
What a new Japanese mosasaur would mean
If future fieldwork in the Kishiwada area recovers more of the animal, a formally named Japanese mosasaur would shift the regional picture in two ways. First, it would push the mosasaur radiation record further into the western Pacific margin at a time when most described species come from North America, Europe, North Africa and the Middle East. The Asian record is patchy by comparison. Second, it would give Japanese paleontology a homegrown large-predator taxon to anchor Cretaceous marine food-web reconstructions that today lean on imported names.
The practical path from here is unglamorous: more screen-washing of the original quarry rubble, more prospecting in the neighbouring Cretaceous exposures in Osaka, Hyōgo and Wakayama Prefectures, and CT scanning of the existing fragment to see features obscured by matrix. The bone has waited 48 years in a cabinet already. It can wait a little longer.
Desk note: this piece leans on the Phys.org item of 20 July 2026 for institutional attribution and the identification itself; we have not added secondary sourcing where the wire had not yet published corroborating coverage.