Integrative taxonomic revision of three Camellia species from section Tuberculata (Camellia, Theaceae) by morphological, anatomical, palynological, and molecular evidence.

Abstract

BACKGROUND: The section Tuberculata (Camellia L.) comprises 18 species, forming a monophyletic group with unique “tuberculate-wrinkled fruit pericarp” morphological characteristics. However, the interspecific relationships within this section remain poorly resolved. A notable taxonomic controversy involves Camellia lipingensis, C. zengii, and C. rhytidocarpa, which were previously considered conspecific. On the basis of extensive population surveys conducted in their type localities, we identified significant morphological disparities among these three taxa. To comprehensively clarify their taxonomic status and relationships, we conducted an integrated study incorporating morphology, micromorphology (leaf epidermis and pollen), and molecular systematics (cpDNA and nrDNA ITS). RESULTS: Evidence from morphology, anatomy, palynology, and molecular systematics consistently supports the treatment of Camellia zengii as a heterotypic synonym of C. lipingensis, while confirming the distinct species status of C. rhytidocarpa. Morphological analysis revealed continuous variation in key traits: Leaves lanceolate (6.42–12.50 × 2.16–4.45 cm); floral parts with 6–9 rounded sepals, 3–5 hairy styles, and 2.2–4.1 cm long filaments; fruit subglobose (diameter 2.24–3.18 cm), ovary 3-4-loculed (1 seed per locule). Anatomical and pollen characteristics are conservative: The leaf epidermal stomata are elliptical (39.9–41.2 × 31.4–36.7 μm), with a density of 62–86 per mm², and the pollen is nearly spherical (polar axis 36.7–37.8 μm/equatorial axis 40.3–41.3 μm, P/E ratio 0.87–0.91). Molecular phylogenetic analyses confirmed that Camellia lipingensis and Camellia zengii form a strongly supported monophyletic group (ML/PP = 100/1.00; clade Ⅰ), with Camellia rhytidocarpa forming a separate clade sister to it. The chloroplast genomes of the three taxa are conserved in structure, with consistent chloroplast genome structures (157,029, 157,029, 157,048 bp; GC 37.3%; containing 87 protein-coding genes, 37 tRNA genes, and 8 rRNA genes). CONCLUSIONS: This study conclusively resolved the taxonomic delimitation among these closely related species within sect. Tuberculata. We propose the synonymization of Camellia zengii with C. lipingensis and confirm that C. rhytidocarpa is a distinct species, a conclusion robustly supported by congruent evidence from multiple disciplines. Our research provides a practical framework for reconstructing phylogenies in morphologically complex plant groups, informs conservation prioritization, and contributes to the development of standardized species delimitation protocols.