The Yumidong Cave sits at about 1100 meters above sea level in the Miaoyu Basin at the northern foot of the Wushan Mountains, roughly four kilometers from the famous Longgupo Cave, a site central to Quaternary paleontology in China. The cave is a horizontal karst feature developed in limestone, and its stratigraphic sections interbed sandy clay with limestone breccia, recording alternating warm and cool phases that researchers have correlated with marine isotope stages from MIS 8 down to MIS 1. Dating results and faunal comparisons place the fossil-bearing layers in the Mid-Late Pleistocene. Flying squirrels rank among the most abundant groups in the mammalian fauna collected there, giving paleontologists excellent material for detailed morphological comparison of a group that has often suffered from misidentification in earlier studies.

The largest and most abundant species in the assemblage is Pteromys volans, the Siberian flying squirrel, a Palaearctic cold-adapted glider that today ranges from northeast Europe to the Korean Peninsula and Hokkaido. Specimens are plentiful in the cave, making up 9 percent of small mammalian fossils in layer ②-2 and a striking 46 percent in layer ④. The mandibles are very short, with length close to height, and carry a particularly marked angular process wider than the ascending ramus. The cheek teeth are brachyodont, with P3 small but bearing differentiated cusps, P4 fully molarised, and M3 lacking a metaloph and metaconule. On lower teeth, the mesoconid and entoconid are permanent, the trigonid basin remains open, and m3 is conspicuously elongated. These features distinguish the Yumidong material from other small Pleistocene flying squirrels of south China, such as Hylopetes electilis with its pitted cheek tooth enamel, and from the Early Pleistocene Pteromys huananensis known only from Longgupo Cave.

Trogopterus xanthipes, the complex-toothed flying squirrel endemic to China, is the second species identified. It still survives in the region today, living from temperate subalpine coniferous forest to subtropical broad-leaved forest at altitudes of 1360 to 2750 meters. The Yumidong mandibles are high and strong, with a deep pterygoid muscle fossa and a vascular notch that is extensive. The upper cheek teeth carry a well-developed mesostyle connecting paracone and metacone, an enlarged metaconule, and numerous cristae dividing the anterior and posterior valleys into several enamel loops. Lower molars develop metalophids and folds in the talonid basin, making the occlusal surface more complex than that of p4. The genus is diagnosed by this complex dentition and by size: Trogopterus is significantly larger than its closest relative Belomys, with almost no overlap in cheek tooth measurements.

Belomys pearsonii, the hairy-footed flying squirrel, is the third species. It is widely distributed today in south China, northeastern South Asia and Southeast Asia, but has completely disappeared from the Wushan area, although subfossils were reported from the Neolithic Dahekou site in Chongqing. The mandible and teeth closely resemble Trogopterus xanthipes, but the species is much smaller and more lower-crowned. Several dental differences separate the two: P3 of Belomys sits closer to the protocone of P4; the protocone of P4-M2 is more developed while the hypocone is weaker; and the hypocone and metaloph of M3 are absent, whereas in Trogopterus the hypocone is permanent and connects with the metaconule by a straight metaloph. Within the genus, the Early Pleistocene Belomys parapearsoni is smaller with a lower crown, and the Thai Late Pleistocene Belomys thamkaewi falls within the size range of Chinese fossil B. pearsonii, so the Yumidong material is confidently assigned to B. pearsonii.

The fourth species, Aeretes melanopterus, is the largest flying squirrel in the assemblage and one of the rarest mammals in China today, confined to the northeastern Qinghai-Tibet Plateau, the Hengduan Mountains, the Qinling Mountains and a small population near Beijing. It inhabits coniferous and broad-leaved mixed forest at 2500 to 3000 meters and nests on high trees. The Yumidong mandible is robust with a short diastemal portion, and the cheek teeth are subhypsodont with slightly rugose enamel, the lingual wall of upper teeth higher than the buccal wall. P4 is molariformed and triangular in outline, with a small protocone separated from the developed hypocone by a vertical anterolingual flexus extending to the crown base. The genus differs from Petaurista, whose M1/2 lacks a hypocone, and the Yumidong specimens fall within the size and morphological variation of the living species, though they are smaller than Pleistocene specimens from the Tianyuan Cave in Beijing.

Reading these four species together as environmental indicators produces a clear paleoecological signal. All four can adapt to coniferous and broad-leaved mixed forest and alpine coniferous forest, but Pteromys volans and Aeretes melanopterus are cold-adapted Palaearctic species that have completely vanished from the study area and its vicinity. Both were excavated mainly from layer ②-2 and layer ④, the cool phases correlated with MIS 2 and MIS 4, and neither appears in layers ③ and ⑤, the warm phases of MIS 3 and MIS 5. The conclusion is that during MIS 2 and MIS 4 the area around Yumidong Cave was dominated by subalpine evergreen coniferous forest or coniferous and broad-leaved mixed forest, colder and richer in conifers than today, resembling the vegetation now found on Wushan peaks above 2000 meters. During MIS 3 and MIS 5 the landscape was similar to the present subtropical environment.

The regional extinction of the cold-adapted species follows a pattern the authors trace through the glacial-interglacial cycle. Cold-adapted flying squirrels spread southwards during glacial periods and disappeared from the area during interglacials. They likely occupied the middle zone of the Wushan Mountains during the latest Pleistocene or the Last Glacial Maximum, but when the Holocene Megathermal arrived, their habitats contracted to higher peaks. Those limited ecological spaces could not sustain viable populations, and the squirrels died out regionally. Nearby caves reinforce the picture of vertical zonation controlling faunal divergence: Baotansi Cave at about 820 meters shares the same species combination, while Migong Cave at only 160 to 200 meters records a cool but sparser forest with no flying squirrel fossils at all, showing that altitude differences within the same period produced different habitats across the Three Gorges area.

The study also corrects earlier records that seemed to show cold-adapted flying squirrels surviving into the Holocene. Re-examination of material from the Dahekou Neolithic site in Fuling District shows that specimens originally identified as Trogopterus xanthipes should be Belomys pearsonii, that the records of Petaurista petaurista and Aeretes melanopterus appear to be an undetermined Petaurista species, and that the material called Hylopetes alboniger actually belongs to the subfamily Sciurinae and resembles Tamiops swinhoei. Current evidence therefore cannot demonstrate that Pteromys volans, Trogopterus xanthipes or Aeretes melanopterus persisted from the Last Glacial through the Holocene in the Three Gorges area, sharpening the extinction timeline for the cold-adapted fauna.

With four species of four genera identified from abundant maxillae, mandibles and isolated teeth, Yumidong Cave now stands as one of the richest flying squirrel localities of the Chinese Quaternary, and the detailed morphological descriptions set a new standard for a group long neglected in Pleistocene faunal lists. The assemblage demonstrates how a single cave sequence can track the interplay between climate oscillation, vertical mountain zonation and regional extinction. As ice ages deepened, coniferous forest climbed down the slopes and cold-adapted gliders followed; as warmth returned, the forest retreated upward and the squirrels with it, until the Holocene Megathermal pushed their refugia beyond the limits of survival. The fossil teeth of Yumidong thus preserve not just species but the rhythm of forests rising and falling across the mountains of the Three Gorges.

Subject of Research: Pleistocene flying squirrel fossils from Yumidong Cave, China

Article Title: Fossil flying squirrels ( Petauristinae , Sciuridae , Rodentia ) from the Yumidong Cave in Wushan County, Chongqing, China

Article References: Pang, L.-B., Chen, S.-K., Hu, X., Wu, Y., & Wei, G.-B. (2024). Fossil flying squirrels (Petauristinae, Sciuridae, Rodentia) from the Yumidong Cave in Wushan County, Chongqing, China. Fossil Record, 27(1), 209-219. https://doi.org/10.3897/fr.27.115693

Image Credits: AI Generated

DOI: 10.3897/fr.27.115693

Keywords: flying squirrels, Yumidong Cave, Pleistocene, paleoenvironment, China, Fossil Record, extinction, Petauristinae, Fossil, flying, squirrels, Sciuridae

Bergmann’s rule, formulated in 1847, holds that animals within a species tend to be larger in colder environments. Allen’s rule, proposed in 1877, predicts that appendages such as ears, tails, and limbs become shorter in colder climates to conserve body heat. Both patterns are widely attributed to thermoregulation: a larger, more compact body loses heat more slowly, while protruding appendages radiate warmth. These rules have broad support across birds and mammals, but the new study reveals that flying squirrels, with their unique gliding lifestyle, may play by a different set of biomechanical constraints.

The findings, published in Ecology and Evolution, show that the three species responded in strikingly different ways. The northern flying squirrel, which ranges across boreal forests from Alaska to the Appalachian Mountains, largely followed the overall pattern: body weight and length increased with latitude, consistent with Bergmann’s rule. Yet its ear length showed no relationship with latitude at all, and its tail actually grew longer toward the north, directly contradicting Allen’s rule. The Humboldt’s flying squirrel, found along the Pacific Coast, displayed the opposite of Bergmann’s rule, with body size weakly declining toward higher latitudes, though this trend was not statistically significant. Its ear length, however, did decrease with latitude, matching Allen’s prediction. The southern flying squirrel, occupying deciduous forests of eastern North America, showed almost no latitudinal trends in body size or ear length, but its tail grew significantly longer toward the north.

The most consistent and unexpected pattern across all three species was tail length. In every analysis, tails grew longer at higher latitudes, the precise opposite of what Allen’s rule predicts. The researchers argue that this discordance reflects the powerful influence of aerodynamic function over thermoregulatory adaptation. Thorington and Heaney previously demonstrated that tail length scales positively with body size across gliding squirrels, proposing that larger animals require longer tails to maneuver during flight. Because body size increases with latitude in these species, tail length may simply follow suit, driven by the demands of gliding rather than by cold climates.

Flying squirrels offer a particularly rigorous test of these rules because their patagium, the furry membrane stretching between forelimbs and hindlimbs, makes them obligate gliders. The tail acts as a stabilizer and brake during descent, and its dimensions are tightly linked to flight performance. The study suggests that within specialized locomotor guilds, functionally critical traits may respond more strongly to biomechanical demands than to ambient temperature. This finding reinforces that ecogeographical rules are not universal and that functionally constrained traits can exhibit latitudinal patterns that directly oppose thermoregulatory predictions.

The research also highlights the value of museum collections for testing ecological patterns across vast spatial and temporal scales. Because the specimens span more than 150 years, the morphological variation likely reflects not only spatial differences but also temporal changes driven by climate warming and habitat modification. Previous work has shown that anthropogenic climate change is altering body size and appendage morphology in mammals, and increasing habitat fragmentation has affected the distribution of flying squirrels in recent decades. The authors note that their dataset does not allow robust statistical analysis of temporal trends, but they suggest that investigating climate-driven morphological change over such a vast timescale would provide insight into microevolutionary dynamics under environmental change.

The study’s species-specific results underscore that ecogeographical patterns can vary dramatically even among closely related organisms occupying overlapping ranges. The northern flying squirrel, a boreal specialist, adhered most closely to the overall pattern, while the Pacific Coast Humboldt’s flying squirrel and the eastern southern flying squirrel each displayed unique combinations of adherence and opposition. This variation likely reflects differences in forest structure, gliding ecology, and evolutionary history among the three species. The authors emphasize that interpreting the negative relationship between latitude and body size in Humboldt’s flying squirrel requires caution because of its comparatively smaller sample size and more restricted geographic distribution.

The findings carry broader implications for understanding how animals will respond to rapid climate change. If functionally critical traits such as tail length are shaped more by locomotor demands than by temperature, then warming climates may not drive the morphological shifts that Allen’s rule would predict. Conversely, if body size follows Bergmann’s rule in some species but not others, the consequences of warming for thermoregulation and energy balance will vary among even closely related organisms. The authors argue that investigating both adherence to and exceptions from ecogeographical rules is critical to anticipating organismal responses in changing environments.

The research team included Joseph P. Morton and Cori Lopazanski, with supervision from faculty at Duke University. The work received no specific funding from public, commercial, or nonprofit agencies, with stipend support provided through Duke University. The authors declare no conflicts of interest. The full dataset, including specimen records from all contributing museums, is publicly available through the Duke Digital Repository and the GitHub repository of the lead author, allowing other researchers to build on this work and continue probing the boundaries of ecological rules in one of nature’s most remarkable mammalian lineages.

Subject of Research: Ecogeographical rule adherence in flying squirrels across latitudinal gradients

Article Title: Breaking the Rules: Museum Data Reveal Divergent Support for Classic Ecogeographical Rules in North American Flying Squirrels

Article References: Paalman, T. A., Morton, J. P., Lopazanski, C., Silliman, B. R., & Wolf, A. T. (2026). Breaking the Rules: Museum Data Reveal Divergent Support for Classic Ecogeographical Rules in North American Flying Squirrels. Ecology and Evolution, 16(10), Article e74470. https://doi.org/10.1002/ece3.74470

Image Credits: AI Generated

DOI: 10.1002/ece3.74470

Keywords: flying squirrels, Allen's rule, Bergmann's rule, museum specimens, latitudinal gradients, thermoregulation, gliding biomechanics, tail length, ecogeography, climate change, Breaking, Rules

A new study led by Academician Fuwen Wei at the Institute of Zoology, Chinese Academy of Sciences, and published in Science China Life Sciences, tackles that question head-on with a genome-scale comparison of independently evolved gliding mammals. The team generated chromosome-level reference genomes for two iconic gliders, the red-and-white giant flying squirrel (Petaurista alborufus) and the sugar glider (Petaurus breviceps), and then compared these with genomes from 17 mammalian species in total, spanning three lineages in which gliding evolved independently. The central finding is striking: the different gliding lineages almost never share the same rapidly evolving regulatory regions of DNA. Yet the genes and developmental pathways connected to those distinct regions converge on strikingly similar functions, from limb and membrane development to muscle formation and Wnt signaling. In other words, evolution reached the same destination by different roads.

The technical heart of the study lies in a class of genomic features known as conserved non-coding elements, or CNEs. These are stretches of DNA that do not encode proteins but have remained remarkably unchanged over deep evolutionary time, a pattern that usually signals an essential role in regulating gene expression. Because regulatory changes are widely believed to underpin many morphological innovations, CNEs are a natural place to look for the genetic basis of a new body plan. The researchers scanned these elements across their species panel and flagged those that showed accelerated evolution specifically in gliding lineages, defining them as gliding-specific accelerated conserved non-coding elements, or GACNEs. Acceleration in an otherwise conserved element is a classic signature of recent adaptive change, suggesting that the region may have acquired a modified regulatory function in the lineage where it evolved fastest.

When the team mapped the GACNEs onto the genomes of the three gliding lineages, the results were unambiguous. The accelerated elements were widely scattered, and the overlap between lineages was minimal. The same regulatory regions were generally not altered again and again across the independent origins of gliding. This runs counter to a simple model of convergent evolution in which the same molecular targets are hit repeatedly. Instead, it points to a more flexible architecture of adaptation, in which many different regulatory elements, distributed across the genome, can be recruited into the service of building a similar trait. For a complex structure like the patagium, which involves coordinated changes to skin, limbs, muscles, and possibly even sensory and circadian systems, there may simply be many genetic routes to the same functional endpoint.

If the specific regulatory regions differ, do the genes they influence differ too? The answer, according to the study, is subtle. The particular genes associated with GACNEs were not identical across the gliding lineages, but when the researchers performed functional enrichment analysis, a statistical test of what biological processes a set of genes participates in, the associated functions converged. Genes near the accelerated elements in each lineage were enriched for roles in appendage development, brain development, and the Wnt signaling pathway, a fundamental developmental cascade that patterns embryonic tissues and is central to limb formation. This is the crux of what the authors call regulatory convergence: different pieces of DNA, altered in different lineages, feed into overlapping gene networks and developmental programs, producing similar biological outcomes despite sequence-level divergence.

The study also revealed that GACNEs are not sprinkled randomly across the genome. Instead, they cluster into distinct evolutionary hotspots, genomic neighborhoods where accelerated regulatory elements accumulate. These hotspots occupied different locations in each of the three gliding lineages, but the genes linked to the hotspots once again showed similar functional profiles. Critically, closely related non-gliding species did not display comparable patterns of clustered acceleration, which strengthens the case that these regulatory changes are connected to gliding itself rather than to shared ancestry or general rodent- or marsupial-specific evolution. The hotspot structure suggests that gliding adaptation may involve coordinated regulatory rewiring within particular genomic regions, even when the identity of those regions varies from lineage to lineage.

To move from correlation toward mechanism, the researchers combined three lines of evidence: the genes associated with GACNEs, the genes linked to the evolutionary hotspots, and the transcription factors whose binding motifs are enriched in the accelerated elements. Transcription factors are proteins that bind regulatory DNA and switch genes on or off, so identifying which ones can interact with the accelerated regions offers a window into how those regions might act. This integrative analysis surfaced several candidate regulators of gliding adaptation, including EMX2, ZFHX3, TCF7L2, TLE1, and CTBP2. Some of these names carry real developmental weight. EMX2, for example, has previously been shown to play an important role in wing membrane formation, lending immediate biological plausibility to the candidate list.

Among the newly highlighted candidates, ZFHX3 stands out as a particularly intriguing target for future work. The study suggests it may be involved in several processes relevant to the gliding body plan, including wing membrane development, muscle formation, cartilage development, and circadian regulation. That combination is notable because a functioning glider needs more than just a sheet of skin; it requires strengthened limb bones, remodeled musculature to hold the patagium taut during flight, and, some researchers argue, appropriate timing of activity to match the demands of a nocturnal, aerodynamic lifestyle. A single regulatory hub touching multiple aspects of the phenotype could help explain how a complex adaptive trait is assembled in a coordinated fashion, and ZFHX3 now becomes a priority for experimental validation.

The team did not stop at comparative genomics. They also compared their candidate regulatory regions against previously reported enhancers, the short DNA segments that boost gene transcription in specific tissues. Several of the candidate enhancers associated with gliding overlapped with regulatory regions already known to be active in embryonic forelimb and lateral body tissue, precisely the anatomical territories from which the patagium develops. This overlap provides independent, functional support for the idea that the accelerated elements identified in gliding lineages are not genomic noise but genuine participants in the developmental program that builds the gliding apparatus. It is exactly the kind of cross-validation, linking sequence acceleration to known embryonic expression domains, that gives comparative genomics its explanatory power.

Taken together, the findings reframe how scientists think about repeated evolution. Complex adaptive traits such as gliding do not require evolution to replay the same genetic changes in every lineage. Different regulatory regions, altered independently, can converge on similar genes, networks, and developmental pathways, generating functional similarity from divergent DNA. The regulatory regions and transcription factors identified here remain candidates, and the authors are careful to note that further work combining genomic, epigenomic, and gene expression data with experimental approaches will be needed to establish exactly how these elements influence development. Even so, the study offers a framework that extends well beyond gliding mammals. For any trait that has evolved independently in multiple species, from echolocation to venom, comparing genetic changes across lineages can now reveal when evolution repeatedly tinker with the same molecular machinery, and when it takes entirely different routes to the same biological destination. In the case of gliding, at least, nature has proven itself a resourceful engineer, solving the same problem many times with many different tools.

Subject of Research: Convergent evolution of gliding in mammals through distinct regulatory genomic changes

Article Title: Different DNA, similar functions: Regulatory convergence in gliding mammals

Article References: Different DNA, similar functions: Regulatory convergence in gliding mammals. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: convergent evolution, gliding mammals, conserved non-coding elements, regulatory evolution, patagium, genomics, flying squirrels, sugar glider, Wnt signaling, transcription factors, developmental biology, evolutionary genetics