INTRODUCTION
Pteridophytes are vascular plants that reproduce via spores rather than flowers or seeds (Smith et al., 2006). They comprise 51 families (PPG I, 2016) and more than 12,000 species distributed worldwide (PPG I, 2016; Kumar et al., 2022; Nitta et al., 2022). Among them, the family Thelypteridaceae includes 30 genera (PPG I, 2016) and approximately 1,200 species (World Flora Online, 2024), most of which are from tropical and subtropical regions (Salino and Semir, 2002). On the Korean Peninsula, Thelypteridaceae includes 6 genera and 21 species (Kim et al., 2015).
Thelypteridaceae is considered a taxonomically complex fern family because the boundaries between genera are ambiguous. This ambiguity stems from frequent interspecific convergence driven by diverse morphological variation and distribution (Almeida et al., 2016; Fawcett and Smith, 2021). Accordingly, this study generally follows the Pteridophyte Phylogeny Group I (PPG I, 2016) classification system, but for Thelypteridaceae it follows the Flora of Korea (Kim et al., 2015).
The genus Cyclosorus consists of approximately 70 species worldwide (Xu et al., 2019), and four species have been reported in Korea: C. acuminatus (Houtt.) Nakai, C. dentatus (Forssk.) Ching, C. interruptus (Willd.) H. Itô, and C. penangianus (Hook.) Copel. (Kim et al., 2015). Morphological characteristics of this genus include 1-pinnate laminae and anastomosing proximal lateral veins (Kim and Sun, 2018; Lee and Lee, 2018). Among these, C. penangianus has been treated as Menisciopsis lineata (Colebr. ex A. Braun) Kovalchuk in some studies (Kovalchuk, 2025; Hassler, 2026); however, it is maintained within Cyclosorus in the Flora of Korea (Kim et al., 2015).
Cyclosorus parasiticus (L.) Farw. has been placed in several genera across taxonomic treatments. Representative treatments include Thelypteris parasitica (L.) Tardieu (Ebihara and Kasetani, 2022) and Christella parasitica (L.) H. Lév. (PPG I, 2016; International Plant Names Index, 2025), whereas previous Korean studies and some taxonomic treatments have adopted Cyclosorus parasiticus (eFloras, 2008; Kim et al., 2015; Lee and Lee, 2025). In molecular phylogenetic studies, the Cyclosoroid clade has been recognized as a monophyletic lineage (Smith and Cranfill, 2002; Smith et al., 2006), and subsequent studies have examined the phylogenetic position of Christella (Almeida et al., 2016; Fawcett et al., 2021).
Cyclosorus parasiticus was originally described under the basionym Polypodium parasiticum L. (Linnaeus, 1753) and was first recorded in Korea by Nakai (1914) as Dryopteris parasitica (L.) Kuntze in Polypodiaceae. Farwell (1931) later proposed the new combination Cyclosorus parasiticus (L.) Farw. By contrast, Chung et al. (1937), in their Botanical Names of Joseon, adopted the treatment reported by Nakai (1914). Ching (1940) proposed a classification that subdivided Polypodiaceae (Moon and Sun, 2008), whereas Holttum (1974, 1976) treated Thelypteridaceae as an independent family and systematically organized the genera and species within it. In Korea, Park (1961) placed this species in Aspidiaceae, and subsequent literature (Park, 1975; Lee, 1980, 2006) continued to apply this classification without a taxonomic review. Korea National Arboretum (2008) placed this species in Dryopteridaceae, but subsequent literature and taxonomic studies report that its identity and distribution in Korea remain unconfirmed (Moon et al., 2014; Kim and Sun, 2018; Lee and Lee, 2018). Lee and Lee (2025) have recently reported this species.
In this study, C. parasiticus, whose identity has long been debated in Korea, was detected in Dongheung-dong, Seogwipo-si, and on Jigwido Island. We aim to clarify its distribution in Korea, examine previously used taxonomic characters across related taxa, and determine molecular phylogenetic relationships with closely related species using the chloroplast rbcL region.
MATERIALS AND METHODS
Sample collection
From 2020 to 2022, specimens were collected to investigate naturally occurring ferns on Jejudo Island. Sampling was conducted primarily in the deciduous broad-leaved forests around the Seongpanak and Gwaneumsa trails of Mt. Hallasan, Gyorae Gotjawal, and Jeolmul Oreum; in the evergreen broad-leaved forests around Hwasun Gotjawal, Seonheul Gotjawal, Seojungcheon, Dongheungcheon, and Hyodoncheon; and in island areas such as Gapado, Marado, and Jigwido.
During the survey period, the discovered ferns were photographed with a Nikon D750 camera (Nikon Co., Japan) before collection to document observable characteristics. For specimen observation, each individual, including the rhizome, was collected, and information on the collection site, time, collector, GPS coordinates, and habitat was recorded. For DNA extraction, approximately 15 cm of the lamina was cut, placed in a collection bag containing silica gel, and stored at room temperature after recording preliminary information. Among the collected samples, voucher specimens of C. parasiticus and related taxa were deposited in the herbarium of the Warm-temperate and Subtropical Forest Research Center (WFRC) in Seogwipo-si, Jeju-do.
Morphological study
The morphological characteristics of C. parasiticus and related taxa were documented from photographs and specimens, with particular attention to the rhizome, sorus position, trichome length, presence or absence of glands, and lateral vein forms. Morphological terminology followed Flora of Korea (Kim et al., 2015). Each species was identified morphologically using Pteridophytes of Korea: Lycophytes & Ferns (Lee and Lee, 2018), Flora of Korea (Kim et al., 2015), The Genera of Vascular Plants of Korea (Kim and Sun, 2018), Illustrated Pteridophytes of Korea (Korea National Arboretum, 2008), The Standard of Ferns and Lycophytes in Japan (Ebihara, 2017), Flora of China (eFloras, 2008), and related studies (Linnaeus, 1753; Holttum, 1974, 1976, 1983; He and Zhang, 2012; Li et al., 2013; Moon et al., 2014; Mazumdar et al., 2017). Morphometric measurements were taken from both fresh and dried specimens. The quantitative traits of total rhizome length, lamina length and width, pinnae width, and trichome length were averaged across seven individuals. Measurements were taken with a 30 cm ruler.
In addition, the morphological characteristics of specimens collected in this study were compared with those of the holotype housed in the Herbarium of the Swedish Museum of Natural History (S). Other specimens identified as C. parasiticus, stored in the herbaria of the National Institute of Biological Resources (KB) and the Korea National Arboretum (KH), were also examined. The specimens examined in this study are summarized in Table 1.
Molecular study
Specimens of C. parasiticus and related taxa were collected on Jejudo Island and used in this study. Dried samples were finely ground with liquid nitrogen, and 50 mg of the ground material was weighed and placed in a microtube. Total DNA was extracted using the yesGTM Plant DNA Extraction Kit (GenesGen, Korea). To amplify the chloroplast rbcL region, polymerase chain reaction (PCR) was performed with primers 1F (5′-TGTCACCACAAACAGAAACT-3′) and 1351R (5′-CTTCACAAGCAGCAGCTAGTTCAGGACTCC-3′) (Liu et al., 2007), along with the middle primers ESRBCL645F (5′-AGAYCGTTTCYTATTYGTAGCAGAAGC-3′) and ES RBCL663R (5′-TACRAATARGAAACGRTCTCTCCAACG-3′) (Schuettpelz and Pryer, 2007).
The PCR mixture contained 2 μL of template DNA, 1 μL of each primer, 5 μL of 10× PCR buffer, 5 μL of dNTP, 4 μL of MgCl2, and 0.5 μL of Taq polymerase (Intron Biotechnology Inc., Seongnam, Korea); the remaining volume was adjusted to 50 μL with distilled water. Amplification was performed on a PCR cycler (Biometra T One 96G, Biometra GmbH, Gottingen, Germany) with the following conditions: pre-denaturation at 95°C for 3 min, denaturation at 95°C for 1 min, annealing at 54°C for 40 s, extension at 72°C for 90 s, and a final extension at 72°C for 7 min, repeated for 35 cycles (Hasebe et al., 1994; Wolf et al., 1999; Ko et al., 2024).
The PCR products were confirmed by electrophoresis on a 1% agarose gel and then sequenced by Macrogen Inc. (Seoul, Korea). The sequences were compared with registered sequences using the Basic Local Alignment Search Tool (BLAST; http://blast.ncbi.nlm.nih.gov/Blast.cgi) for molecular identification. The retrieved sequences were aligned with ClustalW (Thompson et al., 1994) in the Molecular Evolutionary Genetics Analysis software version 12.0 (MEGA 12) (Stecher et al., 2026).
Sequences of C. parasiticus (PX640882, PX640883) and related species, including C. acuminatus (PX640878), C. interruptus (PX640880), C. dentatus (PX640881), and C. penangianus (PX640879), used in this study were deposited in GenBank (https://submit.ncbi.nlm.nih.gov/). To investigate molecular relationships within the genus Cyclosorus, sequences of Pseudocyclosorus subochthodes (Ching) Ching (PP560797), Leptogramma pozoi (Lag.) Ching (MN159435), Phegopteris decursive-pinnata (H. C. Hall) Fée (LC536353), Thelypteris quelpartensis (Christ) Ching (AB575039), and Macrothelypteris viridifrons (Tagawa) Ching (AB575049) were obtained from NCBI (National Center for Biotechnology Information). Sequences were selected for their phylogenetic relevance to Cyclosorus and allied genera, and only those covering the target rbcL region with sufficient length and high sequence quality were included in the analysis. Information pertaining to each sequence is provided in Appendix 1.
The phylogenetic analysis included 17 taxa (Appendix 1). A maximum likelihood (ML) analysis was conducted using IQ-TREE version 3.1.2 (Wong et al., 2026). The best-fit nucleotide substitution model was selected automatically using ModelFinder (Kalyaanamoorthy et al., 2017). Nodal support was assessed using 1,000 ultrafast bootstrap replicates (UFBoot2) (Hoang et al., 2018). Selliguea hastata (Thunb.) Fraser-Jenk. served as the outgroup.
RESULTS
Morphological differences between C. parasiticus and related species
Rhizomes: Cyclosorus parasiticus and C. dentatus have suberect to short-creeping rhizomes, whereas C. penangianus, C. interruptus, and C. acuminatus have longer creeping rhizomes (Table 2).
Fronds: In C. parasiticus and C. dentatus, the upper part of the lamina gradually narrows from approximately two-thirds of the lamina. In C. penangianus and C. interruptus, the lamina abruptly narrows, whereas in C. acuminatus, it either abruptly narrows to a caudate form or narrows gradually. The lower pinnae typically do not reduce in C. parasiticus; however, they are slightly reduced (ca. 1 cm) in some sterile fronds and in individuals from Jigwido Island. In C. dentatus, 2–3 pairs are rapidly reduced in fertile fronds, whereas in sterile fronds, the degree of reduction varies from slight to nearly absent. The lower pinnae of C. acuminatus typically do not reduce, but some individuals show a reduction, with the degree varying among individuals. The lower pinnae of C. interruptus mostly do not reduce, while those of C. penangianus do not reduce at all or very slightly (Table 2).
Trichomes: Cyclosorus parasiticus has densely distributed trichomes ca. 1 mm long on the rachis and on the adaxial surface of the lamina, costae, veins, veinlets, and indusium, whereas C. dentatus has densely distributed trichomes ca. 0.2 mm long on the same parts. Cyclosorus acuminatus has trichomes of irregular length, ca. 0.2–1 mm, on the stipe, and somewhat shorter trichomes on the adaxial surface of the lamina, costae, and veins. These are sparsely distributed on the veinlets, and short trichomes were also present on the indusium. Trichome density varied by region and individual. Cyclosorus interruptus had moderately dense trichomes distributed across the stipe, costae, veins, and veinlets, whereas C. penangianus was generally almost glabrous (Table 2).
Glands: Cyclosorus parasiticus has densely distributed reddish-orange glands on the veins and veinlets, whereas in C. interruptus, these glands are sparsely distributed. C. dentatus, C. acuminatus, and C. penangianus lack such glands (Table 2).
Veins: In C. parasiticus and C. dentatus, one pair of lateral veins anastomoses, and occasionally, one vein connects to the sinus membrane of the incision; the remaining veins are free. In C. acuminatus, one pair anastomoses, and the next 1–2 pairs connect to the sinus membrane. In C. interruptus, one pair anastomoses, whereas in C. penangianus, 4–6 pairs anastomose; the remaining veins are free (Figs. 1, 2, Table 2).
Segment apex shape: C. parasiticus and C. dentatus have rounded or obtuse apices. C. acuminatus and C. interruptus are acuminate, whereas C. penangianus has a dentate margin (Table 2).
Sori and indusia: The sori of C. parasiticus and C. acuminatus are submarginal on the pinnules; C. interruptus is marginal; and C. dentatus and C. penangianus are medial. C. parasiticus, C. acuminatus, C. interruptus, and C. dentatus possess orbicular-reniform indusia, whereas C. penangianus lacks an indusium (Figs. 1, 2, Table 2).
Molecular analysis
The rbcL sequences of C. parasiticus and related taxa from Jejudo Island showed 99.40–100% sequence identity with NCBI accessions, confirming their taxonomic identities. A phylogenetic analysis of 17 species was conducted, including six Cyclosorus species from Jejudo Island and sequences from NCBI (Appendix 1); the resulting phylogenetic tree is shown in Fig. 3. C. parasiticus (PX640882, PX640883, LC484388) formed a clade with 63% bootstrap support, whereas C. dentatus (PX640881, PP559797) and C. acuminatus (PX640878, AB575007) each formed independent clades with bootstrap support of 79% and 96%, respectively. These taxa were grouped with 98% bootstrap support. P. subochthodes (PP560797) was resolved as a sister lineage to this clade with 99% bootstrap support. C. penangianus (PX640879, JN572373) formed an independent clade with 100% bootstrap support and was grouped with the aforementioned taxa with 97% bootstrap support. C. interruptus (PX640880, MT657524) also formed an independent clade with 100% bootstrap support and was resolved as a sister lineage to the remaining Cyclosorus-related taxa with 95% bootstrap support. L. pozoi (MN159435) and T. quelpartensis (AB575039) were resolved as successive sister lineages with bootstrap support of 98% and 95%, respectively. M. viridifrons (AB575049) and P. decursive-pinnata (LC536353) formed a clade with 61% bootstrap support (Fig. 3, Appendix 1).
TAXONOMIC TREATMENT
Cyclosorus parasiticus (L.) Farw., Amer. Midl. Naturalist 12: 259, 1931; Polypodium parasiticum L., Sp. Pl. 2: 1090, 1753; Dryopteris parasitica (L.) Kuntze, Revis. Gen. Pl. 2: 811, 1891; Aspidium parasiticum (L.) Sw., J. Bot. (Schrader) 1800: 25, 1801; Christella parasitica (L.) H. Lév, Fl. Kouy-Tchéou 475, 1915; Thelypteris parasitica (L.) Tardieu, Notul. Syst. (Paris) 7: 75, 1938.—TYPE: CHINA. Guangzhou, without date. P. Osbeck s.n. (holotype, S!, S-P-2653).
Cyclosorus orientalis Ching, Fl. Reipubl. Popularis Sin. 4: 332, 1999.
Korean name: 털별고사리 (Teol-byeol-go-sa-ri).
Plants Perennial, evergreen herbs. Rhizomes Suberect to shortly creeping, scales brown-dark brown, lanceolate, with short filiform projections. Fronds Monomorphic, 19−38 cm tall. Stipe 10−13 cm long, densely hairy (ca. 1 mm), adaxially grooved, sparsely scaly; scales brown, lanceolate, with short filiform projections. Lamina 1-pinnate; narrowly ovate-triangular, densely hairy (ca. 1 mm), length 10−20 cm, width 7−14 cm, soft papery, yellowish green. Pinnae 10−15 pairs; length 6−8.7 cm, width 0.9−1.5 cm, lanceolate, distant to 0.3−1 cm apart, apex gradually narrowed, proximal pinnae not reduced (or only slightly), ascending, proximal 1 or 2 pairs deflexed and sessile. Segments entire; rounded-obtuse. Veins 4−6 pairs, simple; proximal pair anastomosing to form a reticular veinlet, with the subsequent veinlet occasionally extending to the sinus membrane. Reddish-orange glands on the veinlets. Sori abaxial, submarginal. Indusium yellowish-white, orbicular-reniform, entire, hairy (ca. 1 mm).
Distribution: Korea (Jejudo Island); China, India, Indonesia, Japan, Taiwan, Thailand; Polynesian islands, tropical Asia; Kenya, Madagascar, United States.
Habitat: Warm, sunlit crevices in stone walls along roadsides in lowland areas (Seogwipo-si, Dongheung-dong) and in forest clearings on islands (Seogwipo-si, Weimi-ri, Jigwido Island).
Specimens examined. See Appendix 1.
Key to the species of Cyclosorus in Korea
1. Indusium absent; three or more pairs of veinlets anastomosed ································ C. penangianus
-
1. Indusium present; three or fewer pairs of veinlets anastomosed.
-
2. Lamina green; rhizomes long creeping; segment acute at apex.
-
2. Lamina yellowish-green; rhizomes suberect to shortly creeping; segments rounded or obtuse at apex.
4. Reddish-orange glands absent on veinlets; proximal pinnae with 2–3 pairs reduced in fertile fronds; frond with short hairs (ca. 0.2 mm long) ··· C. dentatus
4. Reddish-orange glands present on veinlets; proximal pinnae not reduced in fertile fronds; (or only slightly); frond with long hairs (ca. 1 mm long) ········································································· C. parasiticus
-
DISCUSSION
In Korea, C. parasiticus was first recorded by Nakai (1914). The morphological characteristics of C. parasiticus are as follows: leaf veins are indistinct; the lamina is 60–120 cm long; the rachis and indusium are densely covered with fine hairs; and the pinnae are linear and deeply divided (Nakai, 1914). However, the specimens examined in this study are smaller (19–38 cm) and exhibit distinct leaf veins, findings that are inconsistent with those reported by Nakai (1914). In addition, Nakai’s (1914) record was based on a single voucher specimen (in silvis, 800 m, Sep. 1908, E. J. Taquet 2381), which has not been confirmed. The reported collection site at an elevation of 800 m in a forested area is ecologically inconsistent with the typical lowland habitats of C. parasiticus. Moon (2007) also noted that the occurrence of this species in the deciduous forest zone of Mt. Hallasan is highly unlikely. Therefore, the substantial discrepancy in plant size, together with the ecological inconsistency of the reported habitat, suggests that the Nakai (1914) record contains a misidentification (Table 2).
Park (1961) and Lee (2006) recognized C. parasiticus by its long, creeping rhizome, an abruptly narrowed, pointed lamina apex, and a non-narrowed base, with both surfaces bearing dense, soft hairs. The lobes were narrow and long-elliptic, with rounded apices and wavy serrations along the margins. Park (1975) also described the rhizome as long and creeping underground, with abundant hairs on both surfaces of the lamina, especially on the rachis, and somewhat shorter, lower pinnae. However, previous studies (Park, 1961, 1975; Lee, 2006) did not provide voucher specimens, and the rhizome characteristics differed from those of C. parasiticus confirmed in the present study (Table 2).
A specimen of C. acuminatus collected from Jejudo Island (Seogwipo-si, Jul. 19, 1958, T. B. Lee s.n., KHB1040103) also exhibits rhizome characteristics that differ from those of C. parasiticus. Cyclosorus acuminatus is the most widely distributed species in the Jejudo region. This species is characterized by a long, creeping rhizome, and the lamina apex varies from abrupt to gradual narrowing. Trichome density also varies by region and individual. The rhizome and lamina characteristics described previously (Park, 1961, 1975; Lee, 2006) are similar to those of C. acuminatus (Tables 1, 2).
Korea National Arboretum (2008) described C. parasiticus as having a rhizome that spreads laterally, with the first lateral veins of the segments connected to each other and the second and third lateral veins connected along this extension to form several reticulate veins. In the illustration, the trichomes of the indusium were ca. 0.2 mm long, and no reddish-orange glands were recorded. In addition, the voucher specimen presented in that study (Seogwipo-si, Nov. 20, 2002, B. C. Kim s.n., KHB1048319) consisted only of the lamina without a rhizome, showing a gradually narrowing lamina apex, abruptly reduced lower pinnae, and an obtuse segment. Therefore, considering the description, illustration, and specimen characteristics, this literature is believed to be based on C. dentatus (Tables 1, 2).
Re-examination of specimens identified as C. parasiticus in KB revealed that they were actually C. acuminatus, Thelypteris japonica, or Leptogramma pozoi rather than true C. parasiticus (Table 1).
Recently, Lee and Lee (2025) reported C. parasiticus as a species distributed in Korea. The plant height was 50–100 cm, and the segment apex was obtuse or slightly acute. However, the specimens examined in this study were smaller (19–38 cm) and did not have a rounded to obtuse apex. In addition, the reddish-orange glands on the veins, a key diagnostic character of C. parasiticus, were absent. On Jigwido Island, C. hispidulus, a previously unrecorded taxon in Korea, grows to 43–83 cm, has light green lamina, and is characterized by an obtuse or slightly acute segment apex, the absence of reddish-orange glands, and the presence of capitate hairs on veinlets and dense transparent unicellular acicular hairs on the costa. Therefore, the description by Lee and Lee (2025) appears to be based on C. hispidulus. The light green lamina features are consistent with their habitat and unfurling frond photographs, although the other photographs require additional confirmation (Fig. 1). The characteristics of C. hispidulus and its distribution in Korea will be presented in a subsequent study.
Among the species reported in Korea, C. parasiticus is most similar to C. dentatus. Both species share features such as slightly erect or short-creeping rhizomes and segments with obtuse or rounded apices. However, C. dentatus is mainly observed in flower beds or semi-sunny roadside areas. With regard to fertile fronds, the lowest 2–3 pinnae are reduced, reddish-orange glands are absent from the veinlets, and hairs approximately 0.2 mm long occur on the frond and indusium. In contrast, C. parasiticus grows in crevices of stone walls along roadsides or in forest clearings within island areas; the lower pinnae are not reduced (rarely slightly reduced), hairs approximately 1 mm long occur on the frond and indusium, and reddish-orange glands occur on the veins of the segments (Figs. 1, 2, Table 2). These morphological similarities and differences suggest a close relationship between the two taxa while supporting their classification as distinct species.
This close relationship was also supported by the results of a molecular phylogenetic analysis. In the rbcL-based ML phylogenetic tree, C. parasiticus, C. dentatus, and C. acuminatus formed a strongly supported clade (bootstrap support = 99), indicating a close relationship among these taxa. However, the bootstrap support for the C. parasiticus clade was relatively low (63%), below the generally accepted threshold (70%). This relatively low support is attributable to the limited resolution of the single chloroplast marker (rbcL). Nevertheless, the specimens examined in this case were consistently identified as C. parasiticus based on both morphological and molecular evidence (Fig. 3). Despite low bootstrap support, placing this species in the genus Cyclosorus is sufficiently supported by morphological diagnostic traits (Kim et al., 2015) and existing taxonomic treatment trends (He and Zhang, 2012). In contrast, C. penangianus and C. interruptus were resolved as successive sister lineages outside this clade. Similar phylogenetic relationships have been consistently recovered in previous studies (He and Zhang, 2012; Almeida et al., 2016; Fawcett et al., 2021). In particular, generic relationships among the group to which C. parasiticus belongs and allied genera, including Menisciopsis (C. penangianus) and Pseudocyclosorus, have not been consistently resolved across studies (He and Zhang, 2012; Almeida et al., 2016; Fawcett et al., 2021). However, because this study is based on limited taxon sampling and a single chloroplast marker (rbcL), the current evidence is insufficient to justify generic-level taxonomic changes. The PPG I classification (2016), synthesizing previous molecular phylogenetic studies (Smith and Cranfill, 2002; He and Zhang, 2012; Rothfels et al., 2012; Almeida et al., 2016), recognized Christella as an independent genus. Subsequent phylogenomic analyses revealed that the gene concordance factor within Christella is very low, indicating frequent discordance among gene trees and raising concerns about taxonomic instability within this lineage (Fawcett et al., 2021). Alternatively, a strongly supported Cyclosoroid clade has been recognized (Smith and Cranfill, 2002; Smith et al., 2006), with the report claiming that this clade shares the common cytogenetic characteristic of a basic chromosome number of x = 36 (Smith and Cranfill, 2002; de León et al., 2008). Therefore, considering that transferring this species to the genus Christella is not appropriate at this time, this study provisionally retains the name Cyclosorus parasiticus given its diagnostic morphological characters, nomenclatural stability, and continued use in Korean plant research.
PDF Links
PubReader
ePub Link
Full text via DOI
Download Citation 
Print 
