Amphibambusa fangchenggangensis sp. nov., Am. yunnanensis sp. nov., and Arecophila viscosa sp. nov. (Xylariales, Cainiaceae) associated with bamboo from southwest China Authors: Zuquan Yao, Shilian Zheng, Qing Yang, Shiping Zou, Kamran Habib, Dong-Qin Dai, Hind A. Al-Shwaiman, Abdallah M. Elgorban, Faten Zubair Filimban, Nalin N. Wijayawardene, Qirui Li Published: 2026 ABSTRACT AbstractDuring an ongoing investigation of bambusicolous fungi within the Xylariomycetidae, three microfungal specimens belonging to the family Cainiaceae were collected from southern China. Based on detailed morphological observations and phylogenetic analyses of combined internal transcribed spacer (ITS) and large subunit ribosomal RNA (LSU) sequence data, these collections were identified as two new species of Amphibambusa, namely Amphibambusa fangchenggangensis and Am. yunnanensis, and one new species of Arecophila, Ar. viscosa. Comprehensive descriptions, micrographs, and a phylogenetic tree of the new species are provided, highlighting bamboo forests as a rich reservoir of fungal diversity. Introduction The family Cainiaceae represents a well-defined lineage within the order Xylariales (Samarakoon et al. 2022; Habib et al. 2025). Members of this family are characterized by asci with a distinctive complex apical ring and ascospores typically bearing longitudinal germ slits (Samarakoon et al. 2016; Hyde et al. 2020). Species of Cainiaceae are primarily saprobic or weakly parasitic on monocotyledons, including grasses, bamboos, and palms (Krug 1978; Habib et al. 2025; Ren et al. 2025). Arecophila is one of the largest genera in Cainiaceae. It was established by Hyde (1996) to accommodate species occurring on monocotyledons and possessing striate or verrucose, brown ascospores, with Ar. gulubiiicola designated as the type species. The genus is characterized by immersed, subglobose to lenticular ascomata; a poorly developed or absent clypeus; and asci with a wedge-shaped apical ring that is J+ in Melzer’s reagent and two-celled, brown ascospores with wall striations that are surrounded by a mucilaginous sheath (Hyde 1996; Kang et al. 1999; Umali et al. 1999). Hyde (1996) initially placed the genus in Amphisphaeriaceae based on its unitunicate asci with a J+ apical ring and brown, two-celled ascospores. Kang et al. (1999) later re-evaluated the genus and transferred it to Cainiaceae. Subsequent molecular phylogenetic studies incorporating large subunit ribosomal RNA (LSU) and small subunit ribosomal RNA (SSU) sequence data demonstrated that Arecophila clusters with Cainia within Xylariales (Smith et al. 2003). Its placement in Cainiaceae has since been consistently supported by internal transcribed spacer (ITS)-based and multi-locus phylogenetic analyses (Jeewon et al. 2003; Senanayake et al. 2015; Hyde et al. 2020; Wijayawardene et al. 2020). The genus Amphibambusa was introduced by Liu et al. (2015), with Am. bambusicola as the type species. Similar to Arecophila, it was initially placed in Amphisphaeriaceae, but later studies transferred it to Cainiaceae based on multi-locus phylogenetic evidence of ITS and LSU sequence data (Senanayake et al. 2015; Hyde et al. 2020). Species of Amphibambusa are predominantly reported from bamboo and are mainly distributed in tropical and subtropical regions of Asia. Many records originated from southern China, particularly Yunnan, Guizhou, and Guangxi provinces, suggesting a distinct geographical occurrence for this genus (Zhou et al. 2024; Liu et al. 2025). During an ongoing investigation of bambusicolous fungi within Xylariomycetidae, several specimens belonging to Cainiaceae were collected from bamboo in southern China. Morphological examination and phylogenetic analyses based on ITS and LSU sequence data revealed that these specimens represent three species new to science, belonging to the genera Amphibambusa and Arecophila. Detailed descriptions, illustrations, and phylogenetic analyses of these new taxa are provided in this study. Materials and methods Sample collection and preservation Samples were collected during field surveys conducted at Mount Foding Nature Reserve in Guizhou Province, Mount Wuliang Nature Reserve in Yunnan Province, and the Shiwandashan National Nature Reserve in the Guangxi Zhuang Autonomous Region, China, from July 2024 to September 2025. Detailed habitat information, including geographical coordinates (latitude, longitude, and altitude), was recorded during field collection. Samples were placed in sealed bags and transported to the laboratory for subsequent morphological examination and isolation of fungal strains. To preserve freshness, specimens were dried using a portable fan dryer. The dried specimens were carefully labeled and stored for further study. Subsequently, the specimens were used for morphological and molecular studies. All specimens were deposited in the Herbarium of Guizhou Medical University (GMB) and the Herbarium of Cryptogams, Kunming Institute of Botany, Chinese Academy of Sciences (KUN-HKAS). Morphological characterization and isolation Macroscopic features, e.g., ostiole and clypeus, of the specimens were examined using an Olympus SZ61 stereomicroscope and photographed with a Canon 700D digital camera. Microscopic morphological characteristics, e.g., ascomata, peridium, paraphyses, asci, and ascospores, were observed using an optical microscope (Nikon Ni) and photographed with an attached Canon 700D digital camera. The amyloid reaction of the apical apparatus was tested using Melzer’s iodine reagent. The dimensions of asci and ascospores were measured using Tarosoft Image Framework (v.0.9.0.7). Images were processed using Adobe Photoshop CS6 (Adobe Systems, USA). DNA extraction, PCR amplification, and sequencing Mycelium was scraped from pure culture plates using a sterilized scalpel, and genomic DNA was extracted according to the manufacturer’s instructions using the BIOMIGA Fungus Genomic DNA Extraction Kit. For specimens in which ascospores failed to germinate, DNA was extracted directly from perithecial contents. DNA samples were stored at –20 °C. The ITS region and the LSU gene were amplified by polymerase chain reaction (PCR) using primers ITS1/ITS4 (White 1990; Gardes and Bruns 1993) and LR0R/LR5 (Vilgalys and Hester 1990), respectively. The PCR reaction mixture had a total volume of 25 μL, containing 9.5 μL double-distilled water, 12.5 μL PCR Master Mix, 1 μL of each forward and reverse primer, and 1 μL template DNA. PCR products were verified by electrophoresis on a 1.5% agarose gel stained with Golden View and subsequently sent to Sangon Biotech Co., Ltd. (Shanghai, China) for sequencing. Isolation attempts were performed using the single-ascospore isolation method described by Long et al. (2019); however, these attempts were unsuccessful, and no pure cultures were obtained. Sequence alignment and phylogenetic analyses All obtained sequences were deposited in GenBank (Table 1). Sequences were compared both pairwise and against available sequences in GenBank using the BLASTn algorithm for accurate identification. Molecular phylogeny was inferred from a combined dataset of ITS and LSU sequences. Reference sequences retrieved from public databases were selected based on recently published literature and BLASTn results showing high sequence similarity. Sequence alignments were performed using the MAFFT v.7.110 online program (https://mafft.cbrc.jp/alignment/server/index.html; Katoh et al. 2019) with default parameters and were manually adjusted where necessary using BioEdit v.7.0.5.3 (Hall 1999). Maximum likelihood (ML) analysis was conducted using RAxML v.8.2.12 under the GTR+G+I substitution model with 1,000 bootstrap (BS) replicates (Stamatakis 2014). Bayesian inference (BI) phylogenetic analysis was performed using MrBayes v.3.2.1 (Ronquist et al. 2012). Posterior probabilities (PP) were estimated using Markov chain Monte Carlo (MCMC) sampling in MrBayes v.3.2.2 (Ronquist et al. 2012). Six simultaneous Markov chains were run for 1,000,000 generations, with trees sampled every 1,000th generation. The resulting phylogenetic tree was visualized using FigTree v.1.4.4 (Rambaut 2018). All phylogenetic analyses were conducted on the CIPRES Science Gateway v.3.3 web portal (Miller et al. 2010). Table 1.Taxa and corresponding GenBank accession numbers of sequences used in the phylogenetic analysis.SpeciesStrainGenBank accession numbersReferences ITS LSU Amphibambusa aureae GMB4550T PQ066508 PQ066514 Zhou et al. (2024) Am. aureae GMB4561 PQ066509 PQ066515 Zhou et al. (2024) Am. aquatica MFLUCC 18-1046T PP584660 PP584659 Manawasinghe et al. (2025) Am. bambusicola MFLU 14-0825T KP744433 KP744474 Liu et al. (2015) Am. bambusicola GMB5602T PQ884689 PQ885401 Liu et al. (2025) Am. bambusicola GMB5608 PQ884690 PQ885402 Liu et al. (2025) Am. cerosissimae GMB5603T PQ884693 PQ885405 Liu et al. (2025) Am. cerosissimae GMB5613 PQ884694 PQ885406 Liu et al. (2025) Am. guangxiensis GMB4910T PV617366 PV617364 Yao et al. (2025) Am. guangxiensis GMB4911 PV617365 PV617363 Yao et al. (2025) Am. hongheensis KUN-HKAS112723T MW892971 MW892969 Jiang et al. (2021) Am. hongheensis KUMCC 20-0334 MW892972 MW892970 Jiang et al. (2021) Am. subbambusicola GMB5606T PQ884691 PQ885403 Liu et al. (2025) Am. subbambusicola GMB5618 PQ884692 PQ885404 Liu et al. (2025) Am. xishuangbannaensis GMB5601T PV644573 PV644575 Ren et al. (2025) Am. xishuangbannaensis GMB5634 PV644574 PV644576 Ren et al. (2025) Am. fangchenggangensis GMB7602T PX932424 PZ106884 This Study Am. fangchenggangensis GMB7702 PX932425 PZ106885 This Study Am. yunnanensis GMB7603T PX932422 PZ106882 This Study Am. yunnanensis GMB7703 PX932423 PZ106883 This Study Arecophila amphibambusina GMB6259T PQ874066 PQ860514 Habib et al. (2025) Ar. amphibambusina GMB6260 PQ874067 PQ860515 Habib et al. (2025) Ar. australis GZUCC0124 MT742125 MT742132 Li et al. (2022) Ar. australis GZUCC0112T MT742126 MT742133 Li et al. (2022) Ar. bambusae HKUCC 4794 N/A AF452038 Umali et al. (1999) Ar. chinensis GMB6217T PQ874007 PQ860452 Habib et al. (2025) Ar. chinensis GMB6218 PQ874008 PQ860453 Habib et al. (2025) Ar. clypeata GZUCC0127 MT742128 MT742135 Li et al. (2022) Ar. clypeata GZUCC0110T MT742129 MT742136 Li et al. (2022) Ar. clypeata MFLU 18-2280 OR224995 OP612558 Zhang et al. (2023) Ar. gaofengensis GMB4541T PQ066512 PQ066516 Zhou et al. (2024) Ar. gaofengensis GMB4559 PQ066513 PQ066517 Zhou et al. (2024) Ar. guizhouensis GMB1138T PQ884695 PQ885407 Liu et al. (2025) Ar. guizhouensis GMB5614 PQ884696 PQ885408 Liu et al. (2025) Ar. miscanthi MFLU 19-2333T NR171235 NG088086 Hyde et al. (2020) Ar. miscanthi FU31025 MK503821 MK503827 Hyde et al. (2020) Ar. miscanthi GZAAS 19-1753 OR224996 OP612557 Zhang et al. (2023) Ar. maolanensis HGUP24-0024T PQ568137 PQ569318 Sun et al. (2025) Ar. muroiana GZUCC0122 MT742127 MT742134 Li et al. (2022) Ar. subguizhouensis GMB5617T PQ884697 PQ885409 Liu et al. (2025) Ar. subguizhouensis GMB5610 PQ884698 PQ885410 Liu et al. (2025) Ar. xishuangbannaensis GMB-W1283T OR995736 OR995743 Han et al. (2024) Ar. xishuangbannaensis ZHKU23-0280 OR995737 OR995744 Han et al. (2024) Ar. yunnanensis HKAS 136906 N/A PQ569319 Sun et al. (2025) Ar. zhaotongensis GMBCC1145T OR995740 OR995747 Han et al. (2024) Ar. zhaotongensis ZHKU23-0260 OR995738 OR995745 Han et al. (2024) Ar. sp.HKUCC 6487 N/A AF452039 Jeewon et al. (2003) Ar. viscosa GMB7601T PX932426 PZ106886 This Study Ar. viscosa GMB7701 PX932427 PZ106887 This Study Atrotorquata lineata Mt25 AF009807 N/A Unpublished Barrmaelia moravica CBS 142769T NR153495 N/A Voglmayr et al. (2018) B. rhamnicola CBS 142772T NR153497 N/A Voglmayr et al. (2018) B. rappazii CBS 142771T NR153496 N/A Voglmayr et al. (2018) Cainia anthoxanthis MFLUCC 15-0539T NR138407 NG070382 Senanayake et al. (2015) C. desmazieri CBS 137.62 MH858124 MH869702 Vu et al. (2019) C. daweishanensis GMB5619T PQ884699 PQ885411 Liu et al. (2025) C. daweishanensis GMB5624 PQ884700 PQ885412 Liu et al. (2025) C. globosa MFLUCC 13-0663T NR171724 KX822123 Unpublished C. graminis CBS 136.62 MH858123 MH869701 Vu et al. (2019) C. graminis LSU0203 MT000398 MT000493 Unpublished C. shilihetanensis GMB5612T PQ884701 PQ885413 Liu et al. (2025) C. shilihetanensis GMB5622 PQ884702 PQ885414 Liu et al. (2025) Endocalyx cinctus JCM 7946 LC228648 LC228704 Unpublished E. cinctus CC-HSC6 PP407858 PP407674 Unpublished E. cinctus NBRC 31306 MZ313191 MZ313152 Unpublished E. metroxyli MFLUCC 15-0723B MT929163 MT929314 Konta et al. (2021) E. metroxyli MFLUCC 15-0723AT NR176745 NG081486 Konta et al. (2021) E. metroxyli MFLUCC 15-0723C N/A MT929315 Konta et al. (2021) E. ptychospermatis UESTCC:23.0130 OR253165 N/A Unpublished E. ptychospermatis SNT334 PP592386 N/A Zhang et al. (2024) E. indumentum JCM 8042 N/A MZ313157 Gregorio et al. (2021) E. indumentum JCM 5171 N/A MZ313161 Gregorio et al. (2021) E. phoenicis ZHKUCC 22-0128 OR164915 N/A Senanayake et al. (2023) E. phoenicis ZHKUCC 22-0135 OR164914 N/A Senanayake et al. (2023) Longiappendispora chromolaenae MFLUCC 17-1485T NR169723 NG068714 Mapook et al. (2020) Notes: Type specimens are marked with T; “N/A” indicates that no sequence is available in GenBank; newly generated sequences are indicated in red. Taxa and corresponding GenBank accession numbers of sequences used in the phylogenetic analysis. Notes: Type specimens are marked with T; “N/A” indicates that no sequence is available in GenBank; newly generated sequences are indicated in red. Results Phylogeny After excluding ambiguously aligned regions and long gaps, the final concatenated dataset comprised 1,400 characters (ITS: 1–517, LSU: 518–1400). Three Barrmaelia species, viz., B. moravica (CBS 142769), B. rhamnicola (CBS 142772), and B. rappazii (CBS 142771), were used as outgroup taxa. The phylogenetic analyses based on ML and BI produced highly congruent tree topologies. The best-scoring ML tree is shown in Fig. 1, with a final optimization likelihood value of -8071.573220. The estimated parameters of the GTR+G+I model were as follows: base frequencies A = 0.251, C = 0.234, G = 0.282, T = 0.233; substitution rates A↔C = 2.010, A↔G = 3.796, A↔T = 2.812, C↔G = 1.717, C↔T = 8.502, and G↔T = 1.000. The gamma distribution shape parameter (α) was 0.540, and the tree length was 1.567. Phylogenetic tree of Arecophila and Amphibambusa with related taxa based on combined ITS and LSU sequences. Nodes with bootstrap support values ≥ 70% (ML) and posterior probabilities (PP) ≥ 0.90 are indicated (ML/PP). Type materials are marked with “T,” and newly described species are highlighted in red. In the phylogram (Fig. 1), Amphibambusa yunnanensis (GMB7603, GMB7703) formed a clade sister to Am. aquatica and Am. bambusicola. Although this relationship received low bootstrap support, the clade remained consistently stable across repeated phylogenetic analyses. The sequences of Am. fangchenggangensis (GMB7602, GMB7702) formed a well-supported (91% BS/1.0 PP) independent lineage sister to the clade comprising Am. aureae and Am. subbambusicola, supporting its recognition as a distinct species. Within the genus Arecophila, Ar. viscosa (GMB7601, GMB7701) formed a distinct clade with Ar. chinensis as its sister species. Taxonomy Amphibambusa fangchenggangensis Z.Q. Yao & Q. R. Li sp. nov. 9077C599-6F4E-5865-9C6C-493103010EBA 862737 Fig. 2 Etymology. The specific epithet ‘fangchenggangensis’ refers to the locality of the type collection, Fangchenggang City, Guangxi Zhuang Autonomous Region, China. Holotype. China • Guangxi Zhuang Autonomous Region, Fangchenggang City, Shiwangdashan National Nature Reserve (21°42'93.19"N, 107°19'16.67"E), altitude 754 m, on dead bamboo culms, 14 August 2024, Z.Q. Yao, 2024SWS100 (GMB7602, holotype); ibid. KUN-HKAS 152927, isotype. Description. Saprobic on decaying bamboo culms, forming black, rounded spots on the host surface. Ascomata 547–875 × 524–817 µm (x̄ = 728.4 × 676.8 µm, n = 15), immersed, solitary to scattered, globose to oval-subglobose, black, ostiolate. Ostiole 73.9–169.2 × 117.4–170.8 µm (x̄ = 126.6 × 146.9 µm, n = 10), centric, periphysate. Peridium 13–28 µm (x̄ = 22.5 µm, n = 10) wide, composed of thick-walled, hyaline to brown cells, textura angularis. Paraphyses 2.5–9.5 µm (x̄ = 4.1 µm) wide, filamentous, aseptate, thin-walled, with hyaline, guttulate cells. Asci 211–274 × 12–23 µm (x̄ = 238.5 × 15.5 µm, n = 25), 8-spored, unitunicate, cylindrical, short-stalked, with a J+, disc-shaped, subapical ring, blues in Melzer’s reagent, measuring 3–5 × 2–4 µm (x̄ = 4.5 × 2.5 µm, n = 10). Ascospores 23.5–35.5 × 6–10 µm (x̄ = 30.9 × 7.9 µm, n = 30), uniseriate, hyaline, fusiform, verrucose, 1-septate, slightly constricted at the septum, tapering towards both ends, with longitudinal striations along their entire length, surrounded by a mucilaginous sheath 9.5–15 µm (x̄ = 12.2 µm, n = 10) thick. Asexual morph: Undetermined. Amphibambusa fangchenggangensis (GMB7602, holotype). A. Bamboo host; B, C. Ascomata on the surface of the host; D, E. Longitudinal sections of ascomata; F. Peridium; G–I. Asci in distilled water; J. Paraphyses; K. A J+ apical apparatus bluing in Melzer’s reagent; L–N. Ascospores (N: in Indian ink). Scale bars: 0.2 mm (B–D); 100 µm (E); 10 µm (F–N). Paratype. China • Guangxi Zhuang Autonomous Region, Fangchenggang city, Shiwangdashan National Nature Reserve (21°42'93.11"N, 107°19'16.68"E), altitude 723 m, on dead bamboo, 14 August 2024, Z.Q. Yao, 2024SWS200 (GMB7702). Notes. In the phylogenetic tree, Amphibambusa fangchenggangensis (GMB7602) clusters with Am. aureae (GMB4550) and Am. subbambusicola (GMB5606), forming a well-supported sister clade (91% BS/1.0 PP; Fig. 1). Morphologically, Am. fangchenggangensis differs from Am. aureae by having larger asci (211–274 × 12–23 µm vs. 90–190 × 9–18 µm) and larger ascospores (23.5–35.5 × 6–10 µm vs. 15–22.5 × 5–7.9 µm) (Zhou et al. 2024; Liu et al. 2025). In addition, the ascospores of Am. fangchenggangensis possess distinct longitudinal striations, whereas those of Am. subbambusicola are smooth. The new species also has a thicker mucilaginous sheath (9.5–15 µm) compared with Am. aureae (2.5–7 µm) and Am. subbambusicola (6–10 µm) (Zhou et al. 2024; Liu et al. 2025). Comparative sequence analyses revealed that Am. fangchenggangensis differs from Am. aureae (GMB4550) and Am. subbambusicola (GMB5606) by 6% (26/431 bp) and 6.6% (31/468 bp) in the ITS region, respectively, and by 1% (8/842 bp) from both taxa in the LSU region. These combined morphological and molecular differences support the recognition of Am. fangchenggangensis as a distinct species. Amphibambusa yunnanensis Z.Q. Yao & Q. R. Li sp. nov. 51DEDC75-2755-5E30-AC2B-DBAF783A25EA 862738 Fig. 3 Etymology. The specific epithet ‘yunnanensis’ refers to Yunnan Province, China, the locality of the type collection. Holotype. China • Yunnan Province, Wuliangshan National Nature Reserve (24°25'33.76"N, 101°23'55.57"E), altitude 1467 m, on dead bamboo culms, 16 September 2025, Z.Q. Yao, 2025WLS39 (GMB7603, holotype); ibid. KUN-HKAS 152930, isotype. Description. Saprobic on decaying bamboo culms, forming black, round spots. Ascomata 953–1340 × 760–1003 µm (x̄ = 1109 × 862 µm, n = 15), immersed, solitary, scattered, globose to subglobose, black, with a central ostiole. Peridium 12.5–28 µm (x̄ = 19.1 µm, n = 10) wide, composed of thick-walled, hyaline to brown cells of textura angularis. Paraphyses 1.5–5.5 µm (x̄ = 2.9 µm) wide, filamentous, aseptate, thin-walled, with hyaline, guttulate cells. Asci 236–290 × 12.5–17.5 µm (x̄ = 269.1 × 15.5 µm, n = 20), 8-spored, unitunicate, cylindrical, short-stalked, with a J+ wedge-shaped apical apparatus 3–4 × 2–3.5 µm (x̄ = 3.5 × 2.5 µm, n = 10). Ascospores 30–40 × 6.5–10 µm (x̄ = 35.5 × 8.2 µm, n = 30), uniseriate, hyaline, fusiform, 1-septate at the center, slightly constricted at the septum, tapering at both ends, with longitudinal striations along their entire length, surrounded by a mucilaginous sheath 1.7–5.5 µm (x̄ = 2.5 µm. n = 10) thick. Asexual morph: Undetermined. Amphibambusa yunnanensis (GMB7603, holotype). A. Bamboo host; B, C. Ascomata on the surface of the host; D, E. Longitudinal sections of ascomata; F. Peridium; G–I. Asci in distilled water; J. Paraphyses; K. A J+ apical apparatus bluing in Melzer’s reagent; L–N. Ascospores. Scale bars: 0.2 mm (B–D); 100 µm (E); 10 µm (F–N). Paratype. China • Yunnan Province, Puer city, Mount Wuliang National Nature Reserve (24°25'33.70"N, 101°23'55.52"E), altitude 1455 m, on dead bamboo culms, 16 September 2025, Z.Q. Yao, 2025WLS121 (GMB7703). Notes. In the phylogenetic tree (Fig. 1), Amphibambusa yunnanensis (GMB7603) is closely related to Am. aquatica (MFLUCC 18-1046) and Am. bambusicola (GMB5602). Morphologically, Am. yunnanensis is distinguished from Am. aquatica by its longer asci (236–290 µm vs. 190–240 µm) and wider ascospores (6.5–10 µm vs. 6–7.5 µm). Furthermore, Am. aquatica occurs on submerged decaying wood in freshwater habitats, whereas Am. yunnanensis was collected from dead bamboo in a terrestrial habitat (Manawasinghe et al. 2025; Liu et al. 2025). Compared to Am. bambusicola, the new species has larger asci (236–290 × 12.5–17.5 µm vs. 179–196 × 14.0–16.7 µm), larger ascospores (30–40 × 6.5–10 µm vs. 29.7–36.9 × 6.7–7.8 µm), and a thinner mucilaginous sheath (1.7–5.5 µm vs. up to 10 µm). Comparative sequence analysis revealed that Am. yunnanensis differs from Am. aquatica (MFLUCC 18-1046) and Am. bambusicola (GMB5602) by 5.4% (26/461 bp and 25/467 bp, respectively) in the ITS region, and by 1.9% (16/840 bp and 16/848 bp, respectively) in the LSU region. Based on these molecular and morphological differences, supported by phylogenetic evidence, Am. yunnanensis is introduced as a new species. Arecophila viscosa Z.Q. Yao, K. Habib & Q. R. Li sp. nov. 75C92679-36BA-5988-8D8B-F1489D8480E8 862739 Fig. 4 Etymology. The specific epithet ‘viscosa’ refers to the viscid sheath surrounding the ascospores, a distinctive morphological characteristic of this species. Holotype. China • Guizhou Province, Fodingshan National Nature Reserve (27°19'41.16"N, 108°8'29.74"E), altitude 695 m, on dead bamboo, 16 September 2025, Z.Q. Yao, 2025FDS20 (GMB7601, holotype); ibid. KUN-HKAS 152928, isotype. Description. Saprobic on dead bamboo culms. Sexual morph: Ascomata 506–789 × 640–795 µm (x̄ = 626 × 710 µm, n = 8), immersed under a black clypeus, solitary, slightly raised and dome-shaped, scattered or aggregated, subglobose to globose, with a central conical papilla visible in vertical section. Ostioles elongated, central, black, periphysate. Peridium 19–30 µm thick (x̄ = 25 µm, n = 10), multi-layered, outer layer composed of brown, thick-walled cells of textura angularis, inner layer hyaline. Paraphyses 2.3–4.3 µm wide (x̄ = 3.2 µm, n = 15), hyaline, unbranched, septate. Asci 172–202 × 8.4–13.2 µm (x̄ = 185.6 × 11.2 µm, n = 20), 8-spored, unitunicate, long-cylindrical, with a long pedicel, apex obtuse, containing a square, J+ apical ring (2.7–3.4 × 1.6–2.4 µm, n = 10). Ascospores 14–19 × 5.9–7.6 µm (x̄ = 16.9 × 5.9 µm, n = 30), overlapping uniseriate, 2-celled, pale yellow to yellowish brown, equilateral ellipsoidal, constricted at the septum, with a longitudinal shallow groove running the entire length of the spore, surrounded by a 2.6–13.6 µm thick mucilaginous sheath, lacking germ slits and appendages. Asexual morph: Undetermined. Arecophila viscosa (GMB7601, holotype). A. Bamboo host; B, C. Ascomata on the surface of the host; D, E. Longitudinal sections of ascomata; F. Peridium; G–I. Asci in distilled water; J. Paraphyses; K. A J+ apical apparatus bluing in Melzer’s reagent; L–N. Ascospores (N: in Indian ink). Scale bars: 0.2 mm (B–D); 100 μm (E); 10 µm (F–N). Paratype. China • Guizhou Province, Tongren city, Fodingshan National Nature Reserve (27°19'41.11"N, 108°8'29.72"E), altitude 690 m, on dead bamboo, 16 September 2025, Z.Q. Yao, 202FDS103 (GMB7701). Notes. In the phylogram (Fig. 1; Arecophila clade 2), Arecophila viscosa (GMB7601) forms a well-supported (96% BS/1.0 PP) sister clade with Ar. chinensis (GMB6217) and also shares morphological similarities, including immersed ascomata, an amyloid ascal apical apparatus, septate paraphyses, and brown, 1-septate ascospores. Arecophila viscosa is distinguished from Ar. chinensis by its smaller ascomata (506–789 × 640–795 µm vs. 973–1338 × 746–966 µm) and shorter asci (172–202 µm vs. 215–240 µm). Furthermore, the ascospores of Ar. viscosa are surrounded by a distinct, thick mucilaginous sheath (2.6–13.6 µm thick), whereas the sheath is absent in Ar. chinensis (Habib et al. 2025). Further, comparative nucleotide analysis revealed that Ar. viscosa differs from Ar. chinensis (GMB6217) by 1.2% (5/426 bp) in the ITS region and 0.1% (1/849 bp) in the LSU region. Compared to Ar. subguizhouensis (clustered in Arecophila clade 4; Fig. 1), which also possesses ascospores with a sheath, Ar. viscosa has smaller ascospores (14.5–18.2 × 5.9–7.6 µm vs. 21–26 × 6–8.5 µm) and a slightly thicker sheath (2.6–13.6 µm vs. 6.5–9 µm). In addition, the ascospores of Ar. viscosa are ellipsoidal with a longitudinal shallow groove, whereas those of Ar. subguizhouensis are fusiform with longitudinal striations (Liu et al. 2025). Based on these molecular and morphological differences, supported by phylogenetic evidence, Ar. viscosa is introduced as a new species. Discussion The phylogenetic position of the genus Arecophila reveals significant complexity (Fig. 1). Multiple studies indicate that Arecophila is polyphyletic, with species scattered across different clades within the family (e.g., Habib et al. 2025; Yao et al. 2025; Ren et al. 2025). Some species occupy more basal positions in the phylogeny, whereas others form clades higher up in the tree (Fig. 1). This scattered arrangement suggests that the current circumscription of the genus does not accurately reflect evolutionary relationships. Such a pattern may indicate either convergent evolution of morphological characters or the need for taxonomic reassessment. A notable example is the description of Arecophila amphibambusina by Habib et al. (2025), which clusters phylogenetically within the genus Amphibambusa rather than with core Arecophila species. This placement contradicts its morphology, as the species exhibits the diagnostic features of the genus Arecophila, particularly brown ascospores. The discordance between molecular phylogeny and morphology raises important systematic questions, suggesting that pigmentation and other morphological traits previously considered taxonomically informative may be homoplasious and insufficient for defining generic boundaries within this group. The basal clade of Arecophila species (Fig. 1; Arecophila clade 4) likely represents a distinct evolutionary lineage that warrants recognition as a new genus. However, these taxa morphologically conform closely to the original generic concept of Arecophila, which is characterized by immersed, subglobose to lenticular ascomata, asci with a wedge-shaped apical ring, and two-celled, brown ascospores with longitudinally striated walls. This discordance between phylogeny and morphology underscores the need for taxonomic revision, potentially involving the segregation of clades into separate genera while retaining or emending Arecophila for a more monophyletic core group. Host association patterns provide additional insight into evolutionary relationships within the genus. Most species of Arecophila have been reported from bamboo hosts, suggesting a degree of host specialization that may have influenced diversification. However, three species deviate from this pattern: Ar. australis occurs on Phragmites australis, Ar. miscanthi on Miscanthus sinensis, and Ar. clypeata on an unidentified gramineous host. These three species form a well-supported monophyletic clade in the phylogenetic analysis. This correlation between host preference and phylogenetic grouping suggests that host shifts within Poaceae may have influenced lineage divergence. The clustering of non-bamboo-associated species into a single clade may indicate an evolutionary transition from bamboo to other grasses. Alternatively, it may represent an independent lineage specialized on non-bamboo hosts. This ecological differentiation further supports the hypothesis that Arecophila, as currently circumscribed, comprises multiple evolutionary lineages that may deserve taxonomic recognition at the generic level. Taken together, evidence from phylogeny, morphology, and host association strongly supports the need for a comprehensive taxonomic revision of the family. Future studies incorporating additional taxa, broader geographic sampling, multi-locus datasets, and genomic analyses will be essential to clarify generic boundaries and better understand evolutionary relationships within the family. Supplementary Material XML Treatment for Amphibambusa fangchenggangensis XML Treatment for Amphibambusa yunnanensis XML Treatment for Arecophila viscosa