Ancient Fossil Carpel-Like Organs Push Back Flowering Plant Origins

The story of when flowering plants first appeared on Earth may have just been rewritten. A team of palaeobotanists led by Xin Wang of the Nanjing Institute of Geology and Palaeontology reports in Plant Biosystems the discovery of a remarkable new fossil from the Early Permian of Shanxi Province, China, that appears to bear carpel-like organs enclosing seeds and ovules. The fossil, named Permocarpelloides shuozhouensis, comes from sediments dated to between 295 and 298.9 million years old, an era that conventional wisdom has long insisted was entirely devoid of angiosperms. If the interpretation holds, it would push the documented history of flowering plants back by well over 100 million years, deep into a world otherwise dominated by ferns, conifers and seed ferns.

The specimen was recovered from a gray siltstone layer overlying Coal No. 4 at the Antaibao Surface Mine near Shuozhou City, a site within the famed Shanxi Formation. That formation represents a marine-nonmarine transitional environment where swampy peat accumulated on floodplains fed by braided rivers. The regional plant assemblage, characteristic of the Cathaysian Flora, includes index species such as Emplectopteris triangularis and Taeniopteris mucronata, and high-precision uranium-lead dating of nearby strata has pinned the Shanxi Formation to the Asselian stage of the Early Permian. The fossil itself consists of two facing parts of a cast, each preserving three partially preserved fruits arranged in parallel and oriented the same way.

Morphologically, each fruit is strap-shaped, with nearly parallel sides, preserved lengths of 4.2 to 4.8 millimeters and widths of 1.2 to 1.3 millimeters. All three structures display a terminal zone free of any reproductive bodies, a detail the team interprets as a seed-free, styleless fruit tip. Inside the fertile regions are three to six globular bodies of varying size and form. Larger bodies measure 1.15 to 1.4 millimeters across, while smaller ones range from 0.36 to 0.57 millimeters. The larger bodies bear vermiform, worm-like surface sculpture interpreted as a seed coat, and their discoid, biconcave form is striking, since such doubly concave seeds are known in some living angiosperms, such as the moonseed family Menispermaceae, but have never been documented in any extinct or extant gymnosperm.

To examine the internal anatomy without damaging the precious specimen, the researchers subjected it to micro-computed tomography at the State Key Laboratory of Continental Dynamics at Northwest University in Xi’an. Scanning at 170 kilovolts and 200 microamperes, they collected 2,000 projections at a resolution of 26.22 micrometers and reconstructed the fossil in three dimensions using VG Studio 3.2. The virtual sections revealed seeds and ovules attached along two lines following the margins of the fruit walls, with some bodies positioned along the upper, adaxial side and others along the lower, abaxial side. This arrangement of multiple enclosed bodies at different developmental stages is central to the team’s case that the organs represent genuine carpels rather than any known gymnospermous structure.

The key diagnostic trait is what botanists call angiospermy, or angio-ovuly: the enclosure of ovules, and ultimately seeds, within a closed organ before pollination. This feature has defined flowering plants since the term angiosperm was coined in 1690 by the German botanist Paul Hermann, and it was the basis on which the celebrated Jurassic fossil Archaefructus liaoningensis was recognized as an early angiosperm in 1998. The presence of both smaller bodies, interpreted as unfertilized ovules or abortive seeds, and larger, seed-coated bodies within the same fruit satisfies both the original and stricter versions of this criterion. The authors acknowledge caveats: some conifers and the seed fern Caytonia do enclose seeds after pollination, but they argue that seed position and overall organ morphology in those groups differ fundamentally from what Permocarpelloides displays.

Beyond the mere presence of enclosed ovules, the orderly, parallel arrangement of the three fruits carries deep evolutionary significance. Because the three structures lie so close together and share the same orientation, the team infers they belonged to a single gynoecium, the female reproductive apparatus of one flower. Their clear separation from one another implies that this gynoecium was apocarpous, meaning it was composed of multiple free, distinct carpels, exactly the condition seen in magnoliids such as magnolias and in the basal angiosperm Amborella. Traditional theories of angiosperm evolution, stretching back to Arber and Parkin’s influential 1907 hypothesis, have held that such free, apocarpous gynoecia were ancestral, making Permocarpelloides the first Permian fossil to match those theoretical expectations so directly.

Yet the fossil also complicates the old textbooks. Classical theory, reinforced by Canright’s 1960 study of magnoliaceous carpels and by the celebrated Cretaceous fossils Archaeanthus and Archaefructus, expects seeds to be borne along the adaxial, upper side of the carpel, reflecting the idea that a carpel is a folded megasporophyll, a seed-bearing leaf. In Permocarpelloides, seeds attach along both sides of the fruit, and the coexistence of seeds at markedly different developmental stages in a single organ is highly unusual. The authors note that ovule position has proven far more variable than the classical model allows: recent work on living Illicium and Michelia, reinvestigation of the Archaeanthus holotype, and studies of the Jurassic fossil Xenofructus all indicate that the placenta may originally have been an ovule-bearing branch independent of any leaf margin, casting doubt on the venerable megasporophyll concept itself.

Permocarpelloides is not an isolated anomaly. It joins a growing roster of putative Early Permian angiosperms from China, including Taiyuanostachya, reported in 2023; Yuzhoua juvenilis, described in 2025; and Dengfengfructus maxima, announced earlier in 2026. Compared with those fossils, Permocarpelloides more closely resembles the carpels of modern flowering plants, thanks to its free carpels and apocarpous organization. The striking morphological diversity among these four distinct Permian fossils suggests that flowering plants were already diversified, and therefore must have originated even earlier than the Asselian stage. The authors argue that the long-standing no-angiosperms-until-Cretaceous belief can no longer accommodate the accumulating evidence and that the standard scenario of angiosperm evolution urgently needs updating.

Independent lines of evidence appear to converge on the same conclusion. Molecular clock analyses, from early amino acid and chloroplast DNA studies in 1989 through recent phylogenomic work in 2024 and 2025, have repeatedly estimated pre-Cretaceous origins for flowering plants, and a 2026 integrated analysis of fossils and molecular divergence times points to a latest Jurassic origin at minimum. Intriguingly, the earliest pollen-loaded insects are known from the Lower Permian of Russia, hinting that insect pollination relationships may have been established far earlier than assumed. Together, the fossil record, the insect record and molecular dating sketch a coherent picture of a long, hidden history of angiosperms stretching back at least 295 million years. Whether Permocarpelloides ultimately proves to be a bona fide angiosperm or an unprecedented gymnosperm exhibiting angio-ovuly, its discovery blurs a boundary botanists have drawn for centuries and is likely to ignite a new wave of debate and discovery in the study of seed plant evolution.

The naming of the new genus itself carries meaning. The prefix Permo- signals the fossil’s surprising age, while the suffix -oides, meaning resembling, reflects the authors’ caution: the organs are only partially preserved, and no complete fruit was recovered. Because ovules and immature seeds are difficult to distinguish even in living plants, the team uses the terms carpel and fruit, and ovule and seed, interchangeably throughout their description, a pragmatic choice that underscores how much interpretation rests on a single specimen.

The preservation context is also instructive. Coal-bearing sequences of the Shanxi Formation have long been worked for economic coal, with Coal No. 4 averaging roughly thirteen meters thick at the Antaibao mine, but such commercial exposures rarely attract attention from palaeobotanists hunting for delicate reproductive structures. That a carpel-like fossil emerged from a siltstone immediately above a major coal seam suggests that similar material may await discovery in museum drawers and mine sections across North China, where the Cathaysian Flora has been documented for decades yet never before yielded organs of this kind.

The non-destructive imaging approach deserves emphasis. Traditional palaeobotany often relies on dégagement, physically chipping away matrix, or on serial sectioning, both of which consume irreplaceable material. Micro-computed tomography at the 26-micrometer scale achieved here allows seeds to be traced to their attachment points on the fruit wall in three dimensions, providing the positional evidence that two-dimensional photographs alone could not deliver. As such scanning becomes routine, comparable fossils from other Permian localities may be re-examined with fresh eyes.

For now, the specimen rests in the type collection under catalogue number PB206132, available for scrutiny by other researchers. The open-access publication ensures that the images and reconstructions can be independently assessed, an important step given how consequential the claim is for one of botany’s most enduring questions.

Subject of Research: Permian carpel-like fossil organs and the pre-Cretaceous origin of angiosperms

Article Title: Permian carpel-like organs and their implications on origin of angiosperms

Article References: Wang, X., Fu, Q., Huang, W., & Sun, J. (2026). Permian carpel-like organs and their implications on origin of angiosperms. Plant Biosystems, 160(5), Article 257. https://doi.org/10.1007/s44473-026-00267-6

Image Credits: AI Generated

DOI: 10.1007/s44473-026-00267-6

Keywords: angiosperm origin, carpel, Permian, Permocarpelloides, fossil, paleobotany, enclosed ovules, apocarpous gynoecium, China, Shanxi Formation, seed plant evolution, Early Permian