Three vertebrae sat misidentified in a museum collection for more than two decades before anyone realized they belonged to a species science had never named. That is the quiet, unglamorous origin story behind Limnospondylus ajimuensis, a newly recognized genus and species of giant salamander described this month by researchers at Kyoto University, in a paper that turns a filing-cabinet mystery into a genuine addition to the fossil record.

Bones dug up in the 1990s, understood only now

The fossils themselves are not new discoveries in the literal sense. They were excavated between 1995 and 1997 in the Ajimu area of Oita Prefecture, on Japan's Kyushu Island, preserved within the Tsubusugawa Formation, a layer of ancient lake sediment dating to the Pliocene epoch roughly 3.5 million years ago. At the time, researchers working with limited comparative material tentatively filed the bones under Andrias, the genus that includes today's living giant salamanders. That classification stuck for decades, not because anyone was confident in it, but because nobody had the tools or comparative fossil record needed to challenge it.

That changed when a Kyoto University team, led by Dr. Masahiro Noda, pulled the specimens back out of storage at the Lake Biwa Museum and ran them through a micro-computed tomography scanner, capturing internal bone structure that no visual inspection of the surface could ever have revealed. Combined with geometric morphometric analysis, a technique for precisely quantifying shape variation in the vertebrae, the team found a distinct combination of anatomical features that did not match Andrias or any other known giant salamander. The bones, it turned out, belonged to something researchers had simply never named.

Meet the Ajimu giant salamander

The newly described species, Limnospondylus ajimuensis, takes its genus name from the Greek words for lake and vertebra, a nod to both the lake sediments it was found in and the fact that vertebrae are the only physical evidence of its existence so far. Researchers estimate the animal reached roughly 1.1 meters, or about 3.6 feet, in total length, making it a substantial amphibian even if considerably smaller than its living relative Andrias japonicus, the Japanese giant salamander, which can grow up to 1.8 meters.

Limnospondylus belonged to the family Cryptobranchidae, a group of giant, fully aquatic salamanders that first appeared more than 60 million years ago and today survives in just two living genera: Andrias, spread across Japan and China, and Cryptobranchus, represented solely by the hellbender native to the eastern United States. Despite that long evolutionary history, the family's fossil record across East Asia has remained surprisingly sparse, a gap this discovery helps narrow.

A very different Japan, 3.5 million years ago

What makes the Tsubusugawa Formation particularly valuable to researchers is not just the salamander bones themselves but everything found alongside them. The same sediments have also yielded fossils of elephants and crocodiles, animals with no wild population left in Japan today, painting a picture of a region that was considerably warmer and more humid during the Late Pliocene than it is now. Broader excavations across the formation, collectively referred to as the Ajimu fauna, have also turned up rhinos, turtles, and a range of bird and fish species, describing a subtropical lake and wetland ecosystem almost unrecognizable compared to modern Kyushu.

Unlike its living cousin Andrias japonicus, which inhabits fast-flowing rivers, Limnospondylus appears to have lived primarily in still lake environments, according to the researchers, suggesting the broader Cryptobranchidae family may have occupied a wider range of freshwater habitats in the past than its narrower modern range would suggest.

Why one extinction and one survival happened side by side

The study's authors propose that climatic cooling during the transition from the Pliocene into the early Pleistocene, along with the accompanying habitat changes across the region, likely contributed to Limnospondylus's eventual extinction. Its cousin Andrias japonicus, meanwhile, survived that same climatic shift and still inhabits rivers running through the Ajimu region and other parts of Japan today, even as it now faces its own modern threats, including hybridization pressure from introduced Chinese giant salamander populations.

That contrast, one lineage vanishing under ancient climate pressure while its close relative persisted into the present only to face a very different, human-driven threat today, underscores how giant salamanders as a group have spent millions of years navigating environmental upheaval with mixed success.

A hint that more lost species are hiding in plain sight

Perhaps the most significant implication of this discovery has nothing to do with Limnospondylus itself and everything to do with what it suggests about museum collections generally. Cryptobranchid fossils have turned up across the Northern Hemisphere for decades, but a persistent lack of detailed comparative anatomical knowledge has led many of them to be filed under the living genus Andrias by default, essentially a placeholder classification for bones nobody could confidently sort any other way.

Researchers now suspect that a meaningful number of those existing museum specimens may represent additional unrecognized species, misclassified for the same reason Limnospondylus sat unidentified for more than 25 years: not because the fossils were too fragmentary to study, but because nobody had yet applied the right combination of imaging technology and comparative analysis to look closely enough. That possibility turns natural history collections around the world into something closer to an unopened archive than a settled catalog, one where a fresh look with modern tools can still turn up genuinely new branches of the evolutionary tree.

What comes next

The Kyoto University team's work on the Ajimu fauna is ongoing, and the same site that yielded Limnospondylus continues to produce material worth re-examining with current techniques. Given how long these particular bones sat correctly excavated but incorrectly identified, the more immediate lesson for paleontologists may be a simple one: some of the next new species waiting to be described are probably already sitting in a drawer, just waiting for someone to look again.

---

*Sources cited in this article include the peer-reviewed study published July 2026 describing Limnospondylus ajimuensis, and reporting from ScienceDaily, Sci.News, Phys.org, and Earth.com covering research conducted by Dr. Masahiro Noda and colleagues at Kyoto University. All figures reflect reporting available as of July 27, 2026.*