For more than a century, the powdery mildew that attacks roses was treated as a single, well-known enemy. A sweeping new study from China shows that assumption was wrong. Researchers examined 112 herbarium specimens, including freshly collected material from 23 Chinese provinces and historical type specimens, and combined detailed microscopy with DNA sequencing of the ITS, 28S, and IGS ribosomal regions.
Their conclusion: what has long been called Podosphaera pannosa is actually a complex of five distinct powdery mildew species across two genera, each tied to different rose lineages. The findings, published in IMA Fungus, fundamentally revise the taxonomy of rose powdery mildews and carry direct implications for disease management in one of the world's most valuable horticultural crops. Rosa is among the most economically significant plant genera in global horticulture, and China is one of its great centers of diversity, home to 86 recorded species.
Native Chinese roses such as Rosa chinensis and Rosa odorata supplied many of the agronomic and ornamental traits that shaped modern cultivars, and edible rose cultivation in the country stretches back more than 2,000 years, concentrated in Shandong, Gansu, and Yunnan provinces. But rose production is persistently threatened by powdery mildew, an obligate fungal pathogen that thrives on living tissue. In China's cut-flower sector, which accounts for over 80 percent of the nation's floricultural output, greenhouse powdery mildew incidence reaches 60 to 90 percent, with yield losses exceeding 30 percent.
Edible and medicinal rose production suffers economic losses of 30 to 40 percent. Despite this toll, the true diversity of the pathogens responsible remained murky. Powdery mildews as a group comprise roughly 1,000 species in 17 genera, infecting around 10,000 flowering plant species worldwide.
Four species and one variety across three genera had been reported on roses in China, and that assemblage covered nearly all rose powdery mildews known globally. The study's authors trace the confusion back to Salmon's 1900 monograph, which noted the erroneous assumption that a fungus on a given host in one part of the world must be the same species as the fungus on that host elsewhere. That assumption, combined with the rarity of the sexual fruiting bodies needed for confident identification, allowed one name to obscure a whole species complex.
They supplemented these with 14 type or representative specimens borrowed from the Chinese national herbarium and one German specimen from the Martin-Luther-Universität Herbarium, which they designated as the neotype of Podosphaera pannosa. Because the original type material of the species, described from Germany in 1819, was never preserved, the neotype anchors the name to a physical, sequenced reference. Genomic DNA was extracted from conidia, mycelia, or chasmothecia using the Chelex-100 method, and three genetic regions were amplified and sequenced to build a concatenated phylogenetic matrix.
The phylogenetic results were striking. In an analysis of Podosphaera section Sphaerotheca based on 254 sequences, the rose-infecting Podosphaera sequences resolved into three distinct, well-supported clades. The ex-neotype sequences of Podosphaera pannosa and the ex-type sequences of Sphaerotheca rosae fell into separate clades, providing definitive molecular evidence that Sphaerotheca rosae, long dismissed as a synonym of P. pannosa, is in fact a distinct species.
The team formally reinstated it as Podosphaera rosae comb. nov. Meanwhile, sequences from mildew collected on Rosa xanthina formed yet another separate, well-supported clade, prompting the description of an entirely new species, Podosphaera rosae-xanthinae sp. nov., so far known only from that host. The study also settled a long-running debate at the genus level.
Medusosphaera rosae, originally described in 1962 as the type species of its own genus, had been transferred to Erysiphe in 2000 on the grounds that its chasmothecial appendage morphology did not justify generic separation, but that move had never been tested with molecular data. Sequences successfully retrieved from specimens of this fungus now place it firmly within the Erysiphe clade, confirming the synonymy for the first time with phylogenetic evidence. Morphological features, including lobed hyphal appressoria and conidiophores bearing singly formed conidia, support the placement.
Interestingly, sequences from the three examined specimens clustered according to their host plants, hinting at previously unrecognized diversity within Erysiphe rosae that may represent cryptic formae, though the limited material available prevented further resolution. A third taxonomic cleanup concerned Erysiphe simulans. Two varieties had been described under this species, but earlier work had already flagged them as dubious, noting indistinct morphological boundaries and a lack of genetic differences.
The researchers re-examined a reference specimen from Japan and the holotype of E. simulans var. rosae-rubi from China, confirming that the two are highly similar in all characteristics, with the differences in appendage number described in 1979 overlapping and indistinct. Although sequencing failed on those two historical specimens, sequences from eleven newly collected specimens showed minimal genetic variation across the species. The varieties were therefore dropped, and E. simulans stands as a single, well-defined species.
Beyond taxonomy, the study revealed a clear ecological architecture. Analyzing 104 Chinese specimens with confirmed host identifications, the researchers calculated the relative frequency of each mildew species across rose sections. Within Podosphaera, P. pannosa showed the broadest host range, infecting seven rose species with primary hosts in sections Chinenses and Synstylae, notably R. chinensis and R. multiflora.
P. rosae, with the second broadest range, mainly infects section Rosa, with R. rugosa and R. davurica as primary hosts. P. rosae-xanthinae is restricted to R. xanthina in section Pimpinellifoliae. Among the Erysiphe species, E. simulans is associated chiefly with sections Chinenses and Synstylae, while E. rosae favors sections Pimpinellifoliae and Rosa.
Each pathogen species thus maps onto a distinct host lineage, a non-random pattern that strongly suggests co-evolution as a primary driver of speciation, consistent with co-evolutionary patterns documented in other powdery mildew systems such as Golovinomyces on Asteraceae. Distribution analysis reinforced this evolutionary story. Using precise coordinates for 110 Chinese specimens mapped in ArcGIS, the researchers found that P. pannosa and P. rosae distributions closely match the ranges of their respective primary hosts.
E. simulans presents a fascinating exception: its range, concentrated in southwestern China and hemmed in by the Hengduan and Qinling Mountains, is considerably smaller than that of its primary host in section Synstylae. The authors speculate that the fungus originated near the center of origin of its host lineage but failed to keep pace with the host's subsequent radiation, constrained by mountain barriers, though records in Hubei and Liaoning suggest it may now be expanding eastward alongside its hosts. The overall picture is a bimodal evolutionary history: ancient co-speciation at lower taxonomic levels, overlaid by recent, human-mediated host jumps.
P. pannosa, after all, infects plants from six other genera across five families, a phylogenetically discontinuous host list that implies repeated host jumps, likely accelerated by the global trade of ornamental roses and cut-flower cultivars in the Anthropocene. The practical consequences are immediate. Disease-resistance breeding, pathogen screening, and control strategies for rose powdery mildew have long been built around a single species name that actually concealed at least five pathogens.
The authors argue that future research and management must extend beyond P. pannosa to consider P. rosae, P. rosae-xanthinae, E. simulans, and E. rosae, tailoring resistant cultivars to locally dominant species and specific ecological regions. Residual subclades within the P. pannosa clade hint at further cryptic diversity awaiting discovery, and the team calls for multi-gene and genomic analyses with global sampling to fully resolve the genetic and population-level structure of these fungi. What seemed like one of the most familiar plant diseases in horticulture turns out to be a far richer evolutionary story, with roses revealed as a key host lineage in the diversification of the entire powdery mildew family.
Subject of Research: Taxonomic revision and co-evolution of powdery mildew fungi infecting Rosa in China
Article Title: Revision of powdery mildews ( Ascomycota , Erysiphaceae ) on Rosa in China: unexpected taxonomic complexity with phytopathological implications
Article References: Jin, D.-N., Wang, S.-B., Guan, L.-P., Wu, X.-L., Feng, J., Liu, L., Bradshaw, M., Braun, U., Yang, J.-H., Yu, S.-R., Li, Y., & Liu, S.-Y. (2026). Revision of powdery mildews (Ascomycota, Erysiphaceae) on Rosa in China: unexpected taxonomic complexity with phytopathological implications. IMA Fungus, 17, Article e184484. https://doi.org/10.3897/imafungus.17.184484
Image Credits: AI Generated
DOI: 10.3897/imafungus.17.184484
Keywords: powdery mildew, Rosa, Erysiphaceae, Podosphaera pannosa, Podosphaera rosae, Erysiphe, phylogenetics, co-evolution, taxonomy, plant pathology, China, new species
News Source: Drew Townsend. (October 20, 2026). Research report: Revision of powdery mildews ( Ascomycota , Erysiphaceae ) on Rosa in China: unexpected taxonomic complexity with phytopathological implications. Scienmag.
