Re-evaluation of the taxonomic identity of Angelica fallax (Apiaceae) and its conservation status
Article information
Abstract
A re-evaluation of the taxonomic identity of Angelica fallax H. Boissieu (Apiaceae), a species endemic to Jejudo Island is presented. A. fallax has mostly been treated as a synonym of A. polymorpha Maxim. or sometimes as neither a species nor a synonym in Korea. A. fallax should be considered a species distinct from A. polymorpha based on a detailed re-examination of the original morphological description of the type materials. It is clearly distinguishable from A. polymorpha due to its recurved rachides, obsolete calyx teeth, four or five conspicuous bractlets, slightly zygomorphic peripheral flowers, and mericarps bearing thick primary ribs with an inflated base. Especially, A. fallax has larger mericarps than A. polymorpha; the mericarps do not overlap with each other in size. Furthermore, the two species showed significant differences in leaf and mericarp epidermal characteristics, including the arrangement and shape of epidermal cells and cuticular ornamentation. The morphological differences strongly suggest that the two species are taxonomically distinct. Thus, it is recommended that A. fallax be resurrected from the synonymy of A. polymorpha. A full description, illustrations, color photographs, taxonomic note, ecology, phenology of the species, and an identification key for all Korean species of Angelica are provided. Additionally, listing A. fallax as Vulnerable (VU D2) on the International Union for the Conservation of Nature and Natural Resources Red List is recommended.
INTRODUCTION
Angelica sensu lato (Apiaceae Lindl., tribe Selineae Spreng., subfamily Apioideae Seem.), with nearly 110 species in the temperate regions of the Northern Hemisphere, is one of the largest genera in the family (Hiroe and Constance, 1958; Pimenov and Leonov, 1993). Since its first description by Linnaeus (1753), it has been recognized as having compound leaves, dorsally compressed mericarps bearing filiform or narrowly winged dorsal ribs, broadly winged marginal ribs, and prominent filiform vittae (Hiroe and Constance, 1958; Pimenov, 1968; Qin et al., 1995; Feng et al., 2009). Various infrageneric classifications have been proposed because of the wide morphological variation in the leaves in response to the environment of many species (Miquel, 1867; Franchet and Savatier, 1873; Yabe, 1902; Hara, 1954; Ohwi, 1965; Liao, 2012a). Additionally, many molecular analyses have been conducted to determine the relationship between the species of Angelica s.l. and related genera (Shneyer et al., 2003; Xue et al., 2007; Feng et al., 2009; Liao et al., 2012b, 2013). Based on morphometric and molecular analyses and immunological, chemical, palynological and cytological studies, some major genera, including Archangelica Wolf, Coelopleurum Ledeb., and Ostericum Hoffm., have been treated as distinct from Angelica (Harborne et al., 1986; Vasil’eva and Pimenov, 1991; Pan et al., 1994; Sheh et al., 1997; Shneyer et al., 2003; Shu and Sheh, 2004; Liao et al., 2013).
Previous researches, however, focused only on species from certain countries, not the entire world, making it difficult to complete an infrageneric classification of Angelica. Recently, a large project was conducted to deal with almost all the species (100 spp.) of Angelica, including 26 species from North America (Liao et al., 2022). Their research concluded that Angelica in North America comprised four subclades: North America, Archangelica, Eurasia Angelica, and a perennial endemic North America (PENA) clade (Liao et al., 2022). Nevertheless, a clear infrageneric classification system for Angelica was not developed.
Despite the studies by Liao et al. (2022), the taxonomy of some species is still misunderstood. In particular, the lack of voucher specimens, misidentifications, or missing names from lists have resulted in taxonomic confusion for numerous species. Therefore, critical and concise studies are necessary to define the taxonomic identities of those species.
Angelica fallax H. Boissieu was first described by the French botanist Boissieu (1912). He compared it to the most similar species, A. koreana Maxim. [=Ostericum grosseserratum (Maxim.) Kitag.], based on the number of bracts (0 or 1 vs. 4–8), inflated leaf sheaths and an obsolete calyx. Additionally, A. fallax was considered to be distinct from another similar species, A. uchiyamae Y. Yabe [=A. polymorpha Maxim.], based on the number of bractlets (4 vs. 6 or 7), relatively more developed marginal wings of the mericarp, asymmetric division of the leaflets, and more deeply toothed leaflet margins (Boissieu, 1912). Boissieu referred to specimens that had been sent from Jejudo Island (Quelpaert), South Korea, and collected by two French missionaries, Emile Joseph Taquet and Urbain Jean Faurie, in the early 1900s. To date, A. fallax is considered endemic to Jejudo Island, South Korea, since it has not been reported from any other region.
The distribution of A. fallax on the Korean Peninsula has been debated and the reports are inconsistent (Nakai, 1952; Lee, 1980; Lee, 1996; Chang et al., 2014; Park et al., 2018; Korea National Arboretum, 2022; National Institute of Biological Resources, 2023) (Table 1). Only Nakai listed A. fallax as growing wild in Korea (1952). In 1979, Kitagawa reduced A. fallax to a variety of A. polymorpha, as A. polymorpha Maxim. var. fallax (H. Boissieu) Kitag. Lee (1996) concurred with Kitagawa's treatment. Since then, various Korean studies have treated A. fallax is synonymous with A. polymorpha (Chang et al., 2014; Chung et al., 2017; Park et al., 2018; Korea National Arboretum, 2022; National Institute of Biological Resources, 2023). Since those studies, the status of A. fallax has not been questioned and it is regarded as conspecific with A. polymorpha.
Taxonomic history of Angelica fallax in literature.
The aim of this study was to reexamine A. fallax and to elucidate its taxonomic status. Voucher specimens were collected from the type locality of A. fallax and compared with the original description and with type specimens. Morphological and micromorphological comparisons of A. fallax and A. polymorpha were also conducted. Based on our study, a full description, illustrations, color photographs, taxonomic note, ecology and phenology of A. fallax, and an identification key are presented to separate A. fallax from its congeners.
MATERIALS AND METHODS
The type specimens were examined by visiting virtual herbaria, including the herbarium of the Conservatory and Botanical Garden of the City of Geneva (G, https://www.cjbg.ch/), Royal Botanic Gardens Kew (K, https://apps.kew.org/herbcat/navigator.do) and the Museum national d’Histoire naturelle (P, https://science.mnhn.fr/institution/mnhn/collection/p/item/search). Additional voucher specimens from CBU, KB, KH, KIOM, and SNU were also studied. We also conducted a field survey in Seogwipo-si, Jejudo Island, the type locality of A. fallax, to collect samples and to obtain 28 individuals from the two natural populations. After the morphological examination, all plant samples were deposited in CBU and KIOM as dry specimens.
The dried mericarps of two species, A. fallax and A. polymorpha, were observed with a high-resolution stereomicroscope (Z16 APO, Leica, Wetzlar, Germany) and imaged captured with Dhyana (400DC, Tucsen, Fuzhou, China) using Mosaic ver. 2.4 software to observe and compare the morphological characteristics of the two species.
Before observing the leaf and mericarp epidermal micromorphology, all dried samples were examined under a stereomicroscope (S9D, Leica) to select fully mature leaves and mericarps. The dried materials were rehydrated overnight using the wetting agent Agepon (Agepon: distilled water, 1:200) (Agfa Gevaert, Leverkusen, Germany). The rehydrated materials were then dehydrated using an ethanol series (70, 80, 90, and 100% ethanol) at room temperature for 15 min in each concentration. All dehydrated materials were immersed in 100% isoamyl acetate for 15 min, after which the were immersed in liquid carbon dioxide (CO2) for critical-point drying (HCP-2, Hitachi, Tokyo, Japan). The dried materials were then mounted on aluminum stubs with double-adhesive tape, and the stubs were coated with platinum (Pt) for 20 seconds using an ion sputtering device (Q150T ES Plus, Quorum, UK). The samples were examined using a field-emission scanning electron microscope (Ultra Plus, Carl Zeiss, Oberkochen, Germany) at an accelerating voltage of 3 kV and a working distance of 4.5–5.5 mm.
Cell types were classified based on cell shape, anticlinal wall (primary sculpture) and fine relief of the cuticle sculpture (secondary sculpture) (Barthlott, 1981).
RESULTS AND DISCUSSION
A detailed examination of the type specimens and original description, together with the voucher specimens of A. fallax and A. polymorpha revealed that A. fallax was clearly distinguishable based on several morphological features (Figs. 1, 2, Table 2) and micromorphological characteristics (Figs. 3, 4, Table 3).
Type specimens of Angelica fallax and A. polymorpha. A. Lectotype of A. fallax (U. Faurie 1825, [P], barcode P00602145). B. Syntype of A. fallax (T. Taquet s.n. [P], barcode P03237836). C. Syntype of A. fallax (T. Taquet 4796 [P], barcode P03237838). D. Isolectotype of A. polymorpha (C. J. Maximowicz s.n. [P], barcode P00752989). E. Isolectotype of A. polymorpha (C. J. Maximowicz s.n. [P], barcode P00752990). F. Syntype of A. polymorpha (C. J. Maximowicz s.n. [K], barcode K001097187).
Morphology of mericarps of Angelica fallax (Kim and Kim, 231, KIOM) and A. polymorpha (Kim and Suh, 20231201-001, KIOM). A. Outer view of A. fallax. B. Outer view of A. polymorpha. C. Inner view of A. fallax. D. Inner view of A. polymorpha. E. Cross section of A. fallax. F. Cross section of A. polymorpha.
Comparison of major morphological characteristics of Angelica fallax and A. polymorpha.
Scanning electron micrographs of leaflet epidermis of Angelica fallax and A. polymorpha. A. Abaxial surface of A. fallax. B. Adaxial surface of A. fallax. C. Abaxial surface of A. polymorpha. D. Adaxial surface of A. polymorpha.
Scanning electron micrographs of mericarp epidermis of Angelica fallax and A. polymorpha. A. Overview of mericarp of A. fallax. B. Surface of furrow (between the ribs) of A. fallax. C. Surface of marginal wing of A. fallax. D. Overview of mericarp of A. polymorpha. E. Surface of furrow (between the ribs) of A. polymorpha. F. Surface of marginal wing of A. polymorpha.
Micromorphological characteristics of leaflets and mericarps of Angelica fallax and A. polymorpha.
Angelica fallax has geniculate leaves and smaller inflorescences due to the fewer rays per compound umbel [8–14 vs. 20–30(–40)] and number of flowers per umbellet (20–24 vs. 28–40) (Table 2). The bractlets were also a distinct feature of A. fallax, which has four or five dimorphic bractlets (subulate with entire margins and leaf-like with pinnately lobed margins); A. polymorpha has five to ten monomorphic bractlets (linear to filiform with entire margins) (Table 2). The morphological traits of the flowers were also notable. The peripheral flowers of A. fallax were zygomorphic, whereas A. polymorpha had only actinomorphic flowers (Table 2).
The most noteworthy differences between A. fallax and A. polymorpha were in the mericarps (Fig. 2). The mericarps of A. fallax were 4.9–5.3 × 3.0–3.1 mm, while those of A. polymorpha were 2.9–3.5 × 2.2–2.5 mm (Fig. 2, Table 2). The dorsal ribs of the mericarps of A. fallax were narrowly winged (wings 0.2–0.3 mm wide) with an inflated base (Fig. 2, Table 2) while the marginal ribs were broadly winged (wings 0.8–1.3 mm wide) with an inflated base (Fig. 2, Table 2). In comparison, the mericarps of A. polymorpha had filiform dorsal ribs (0.1–0.15 mm wide) and broadly winged marginal ribs (wings 0.7–0.8 mm wide), which were scarious and lacked an inflated base (Fig. 2, Table 2).
The morphological characteristics listed above were well described in the original description (Boissieu, 1912). Boissieu mentioned that the shape of the rachides, the number and shape of the bractlets, calyx teeth and traits of the wings of the mericarps were the main characteristics of A. fallax that distinguish it from other closely related species. He further inferred that A. fallax was common in Seogwipo-si, Jejudo Island based on other voucher specimens collected by the French missionaries in the early 1900s.
From a micromorphological perspective, both species shared most of the leaf epidermal characteristics, such as irregular epidermal cell arrangement, undulating anticlinal cell walls on the abaxial surface, anomocytic hypostomatic leaves, and simple multi-cellular trichomes (Table 3). However, the leaf epidermal cells on the adaxial surface differed. A. fallax has straight to curved anticlinal cell walls with a striate and convex surface whereas A. polymorpha had undulating anticlinal cell walls and striate and slightly convex periclinal cell walls (Fig. 3, Table 3).
In the Apiaceae, carpological characteristics, including the mericarp surface, are of significant taxonomic importance (Uruşak and Kizilarslan, 2013; Kalsoom et al., 2019; Ostroumova, 2018). Furthermore, characteristics of the epidermal cells on the mericarp surface, especially when viewed using a scanning electron microscope, play a vital role in identification and species delimitation (Ostroumova, 2018; Forycka and Morozowska, 2020; Akturk et al., 2022; Kim et al., 2022). Our study verified that the characteristics of the epidermis of the mericarps, such as the epidermal cell shape (polygonal vs. elongated fusiform) and periclinal cell wall (convex vs. striated and flat) (Fig. 4, Table 3), could clearly distinguish A. fallax and A. polymorpha. The micromorphological evidences suggest that A. fallax is a distinct species.
During a field survey of natural habitats in Seogwipo-si, we found two populations of 100 and 200 individuals each. Peculiarly, while Jejudo Island has little surface water because of its volcanic nature, two habitats commonly had surface water all year, which comes naturally from underground near the coast. The annual average water temperature in the habitats is 15–17°C, which is cool in the summer and warm in the winter (Park, 2010). Further field surveys in similar ecological environments are necessary to identify additional populations.
After a close examination of the habitats, it is inferred the floods from heavy rains in both habitats damaged some individuals. We discovered several stems thickened in the lower part to almost 3 cm in diameter that had been cut due to the water current. All individuals bearing inflorescences had a main stem 5–8 mm in diameter. It was therefore assumed that A. fallax would be much more robust, on average, if floods caused no damage. The populations of A. fallax are well adapted to the unique environment despite severe stress each year and the loss of individuals.
According to Lee et al. (2015a), the average annual rainfall in Seogwipo-si has increased by approximately 1.45 times over the past 50 years. In response, the average direct runoff from Gangjeongcheon Stream, Seogwipo-si, is anticipated to increase significantly by 2070 (Lee et al., 2015b). In particular, the average direct runoff from Gangjeongcheon Stream for the next 80 years (from 2015 to 2100) was predicted to increase by approximately 22% compared to the present (Lee et al., 2015a, 2015b), which means that flood risks in the area will increase. The habitats of A. fallax will thus need protection to prevent reduction of the populations.
In accordance with the verification of A. fallax as a species distinct from A. polymorpha and endemic to Korea, it is necessary to immediately plan conservation strategies for it. Different conservation strategies for endemic taxa may have different implications, depending on taxonomic conclusions and species concepts and delimitation. For Spiraea insularis (Nakai) H. Shin, Y. D. Kim & S. H. Oh, it was not necessary to assess its conservation priority over other threatened species because it was synonymized with S. chamaedryfolia, which is not endemic to Korea (Oh et al., 2010). However, after S. insularis was recognized as distinct and endemic, it was listed as critically endangered (CR) and was eventually protected by the government (Oh et al., 2010; National Institute of Biological Resources, 2012; Chung et al., 2017). Thus, clarifying the taxonomic status of endemic plants is important for understanding the implications for conservation biology of endemic species. Our study indicates that A. fallax should be recognized as an endemic to Korea (Chung et al., 2023).
Angelica fallax is restricted to Jejudo Island, where the total number of individuals is less than 300. Moreover, natural habitats are in danger of shrinking due to flooding driven brought about by increased precipitation caused by climate change. Currently, A. fallax is no included in the list of endemic plants on the Korean Peninsula because it is treated as a synonym of A. polymorpha (Chang et al., 2014; Chung et al., 2017; Park et al., 2018; Korea National Arboretum, 2022; National Institute of Biological Resources, 2023). We therefore recommend that A. fallax be removed from synonymy under A. polymorpha and treated as a distinct species and evaluated under the Red List guidelines. Based on the number of mature individuals and population size, we consider A. fallax to be Vulnerable D2 according to the IUCN Red List (IUCN Standards and Petitions Committee, 2024).
TAXONOMIC TREATMENT
Angelica fallax H. Boissieu, Bull. Soc. Bot. France 59: 199, 1912 (Figs. 5, 6); A. polymorpha H. Boissieu var. fallax (H. Boissieu) Kitag., Neolin. Fl. Manshur. 476, 1979.— TYPE: KOREA. Jejudo Island, near Hongno [Seogwipo-si], towards waterfalls, 2 Jul 1907, U. Faurie 1825 (lectotype: P!, [barcode P00602145] designated by Pimenov, 2017: 110; isolectotypes: G!, [barcode G00359369], KYO [not seen]); Jejudo Island, 8 Apr 1908, T. Taquet 4649 (syntype: not located); 25 Jun 1908, T. Taquet s.n. (syntype: P!, [barcode P03237836]); 11 Apr 1911, T. Taquet 4796 (syntypes: P!, [barcode P03237838], G!, [barcode G00359370])
Illustrations of Angelica fallax. A. Habit in flowering. B. Infructescence. C. Outer view of mericarp. D. Inner view of mericarp. E. Cross section of mericarp. F, G. Cross section of stem. H. Root. I. Bisexual flower. J. Flower after stamens caducous. K. Flower in early stage. L. Inflorescence.
Photographs of Angelica fallax. A. Habit. B, C. Inflorescence. D. Infructescence. E. Stylopodium with style (in early stage). F. Stylopodium with style (in fruiting stage). G. Cross section of mericarp. H. Inflorescence after anthesis.
Korean name: Je-Ju-Sa-Yak-Chae (제주사약채).
Herbs, perennial, hermaphroditic, 0.5–1 m tall. Root a taproot, 1–3 cm in diam., branched. Rhizome short, erect, woody. Stem green, dark green, or dark purple, erect, much branched, 0.3–0.8(–2) cm in diam., longitudinally grooved, solid, glabrous. Leaves alternate, basal and cauline, long petiolate; petiole green, sheathing at base, glabrous; leaf sheath slightly inflated, glabrous, adaxial surface pale green, abaxial surface green or purple; blade triangular in outline, 2- or 3-ternate-pinnate, upper surface green to dark green, lower surface pale green, both surfaces pubescent; rachides geniculate; leaflets triangular to pentagonal, apex acute, base truncate to cuneate, margins irregularly biserrate. Basal leaves caducous. Lower cauline leaves: petiole 8.5–20 cm long, glabrous; blade 21–30 × 25–32 cm, both surfaces sparsely pubescent with short simple hairs, lower surface sometimes glabrous, hair cells translucent or purple; lateral leaflets triangular to pentagonal, 2.5–3 × 2.3–2.4 cm, shallowly 3-lobed or unlobed, base truncate to cuneate, apex acute, petiolule 0.4–1 cm long; terminal leaflets rhombic or rhombic to triangular, 2.7–4.4 × 2.7–5 cm, deeply 3-lobed, apex acute, base truncate, petiolule 1.2–2.2 cm long. Inflorescences terminal and lateral, with 10–14 compound umbels, more or less round-topped, 7–12 cm in diam.; umbellets hermaphroditic, 20- to 24-flowered, 1.2–2.5 cm in diam.; peduncle 15–20 cm long, dark green to dark purple, sparsely pubescent with short simple hairs in middle, densely pubescent with transparent or dark purple simple hairs in upper part; rays 8–14, spreading to ascending, 1.8–5.5 cm long, unequal in length, adaxial surface densely pubescent with simple hairs, abaxial surface sparsely pubescent with simple hairs; bracts absent; pedicels 0.5–1 cm long, unequal in length, densely pubescent with short simple hairs; bractlets 4 or 5, subulate, 0.4–1.5(–2.7) cm × ca. 0.5 mm, apex acute, often pinnately lobed, both surfaces sparsely pubescent with short simple hairs. Flowers bisexual, ca. 4 mm in diam., stamens caducous, inner flowers actinomorphic, peripheral ones slightly zygomorphic; calyx teeth obsolete; petals 5, white, apex and veins abaxially usually purple-tinged, obovate, 1.5–2 × 1–1.5 mm, those on peripheral flowers of umbellets slightly enlarged, 2.2–3.5 × 1.2–1.5 mm, apex 2-lobed; stamens 5, alternating with petals; filaments filiform, 1.5–2.5 mm long; anthers 2-locular, white to whitish yellow, subellipsoid, ca. 0.5 mm long; pistil 1, 2-carpellate; ovary inferior, syncarpous, 2-locular, glabrous, sometimes short simple hairs at apex; stylopodium low-conical; styles 2, free, ascending to spreading, ca. 1 mm long. Fruit a dry schizocarp composed of 2 mericarps, ellipsoid; carpophore 5–6 mm long, 2-cleft nearly to ca. 1/2 its length; mericarps dorsally compressed, 4–6 × ca. 2 mm (except wings), glabrous; dorsal ribs 3, narrowly winged, 0.2–0.3 mm wide, thick; marginal ribs broadly winged, 0.8–1.3 mm wide, base inflated; secondary ribs absent; vittae 6, 1 per vallecula, 2 on commissure; commissure ca. 2 mm wide. Seed sub-pentagonal in cross-section; face plane.
Flowering: November to January.
Fruiting: January to February.
Distribution and habitat: Angelica fallax occurs along the coast of Seogwipo-si; rooting in shallow spring water flowing toward the ocean.
Two populations were identified during the field survey. The straight line distance between them was approximately 8.5 kilometers. The environment of both populations is wet throughout the year and subject to occasional overflows due to heavy rainfall. One population (Gangjeongcheon Stream) comprised approximately 100 individuals, and the other (Geomeunyeo Coast) comprised approximately 200 individuals. Angelica fallax occurs with such common species as Arundinella hirta (Thunb.) Tanaka, Miscanthus sinensis Andersson, Ranunculus japonicus Thunb., Corydalis heterocarpa Siebold & Zucc., Boehmeria pannosa Nakai & Satake ex Oka, Microlepia strigose (Thunb.) C. Presl and Pseudocyclosorus subochthodes (Ching) Ching. Seogwipo-si is the warmest region in Korea, with a 16.9°C annual mean temperature, and is the highest region in rainfall, with 1,989.6 mm average annual precipitation (KMA National Climate Data Center, 2023).
Besides the features that separate Angelica fallax from A. polymorpha mentioned above, the two species differ in sub-pentagonal endosperm in cross-section (A. fallax) vs. oblong endosperm in cross-section (A. polymorpha). Their flowering and fruiting seasons also differ; A. fallax blooms from November to January and fruits from January to February, while A. polymorpha blooms from August to October and fruits from September to November.
Additional specimens examined: KOREA. Jejudo Island (Jeju-do Prov.): Seogwipo-si, Donghong-dong, Geomeunyeo Coast, 9 Jan 2023, C. S. Kim and K. Kim 231 (2 sheets), 232, 233 (3 sheets), 234, 235, 236, 237 (KIOM); Gangjeong-dong, downstream of Gangjeongcheon Stream, 20 Dec 2017, H. M. Lee 171, 172, 173, 174, 175, 176 (KIOM); 21 Nov 2018, G. S. Kim and H.-J. Suh 328, 329 (KIOM); 3 Nov 2019, K. Kim 191, 192, 193 (KIOM); 2 Dec 2019, K. Kim 194, 195, 196, 197 (KIOM); Gangjeong Bridge, 33°14′01.8″N, 126°29′11.3″E, elev. 77 m, 26 Sep 2013, S.-H. Park, C.-W. Lee, J.-G. Hong, S.-J. Ji, P133558 (KH!).
Updated key to the species of Angelica in Korea based on Flora of Korea (Park et al., 2018)
1. Leaves 1- or 2-pinnate ・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・ A. cartilaginomarginata
1. Leaves ternate or ternate-pinnate.
2. Leaf sheaths abaxially densely pubescent with short simple hairs ・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・ A. anomala
2. Leaf sheaths glabrous.
3. Leaf rachis and petiolules geniculate.
4. Rays per compound umbel more than 20; bractlets linear to filiform, entire; all flowers actinomorphic; ovary puberulent or hispidulous.
5. Leaflets irregularly apiculate-serrate, abaxially sparsely to moderately pubescent with short simple hairs along veins; rays (25–)40–60; vittae 4 on commissure ・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・ A. genuflexa
5. Leaflets regularly apiculate-serrate, abaxially glabrous; rays (16–)20–25(–42); vittae 2 on commissure ・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・ A. reflexa
4. Rays per compound umbel less than 20; bractlets subulate or pinnately lobed; central flowers actinomorphic and peripheral flowers zygomorphic (outer petals enlarged); ovary glabrous ・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・ A. fallax
3. Leaf rachis and petiolules straight.
6. Petals purple to dark purple, rarely white.
7. Compound umbels round-topped, nearly spherical; petals dark purple ・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・ A. gigas
7. Compound umbels flat-topped, nearly hemispherical; petals purple, rarely white ・・・・・・・・・・・・・・・・・・・・・・・・・・・・ A. decursiva
6. Petals white or sometimes pinkish.
8. Outer petals on peripheral flowers of umbellets enlarged; seeds pentagonal or semicircular in cross section.
9. Stems dark purple or purplish green, solid; pedicels glabrous; calyx teeth small, triangular; dorsal ribs of mericarp filiform, not winged; seeds pentagonal in cross section ・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・ A. acutiloba
9. Stems green, hollow; pedicels pubescent; calyx teeth obsolete; dorsal ribs of mericarp narrowly winged; seeds semicircular in cross section ・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・ A. czernaevia
8. Outer petals on peripheral flowers of umbellets not enlarged, all petals in umbellets equal in size; seeds oblong in cross section.
10. Seaside plants; sap yellowish white; abaxial surface of leaves shiny; dorsal ribs of mericarp winged ・・・・・・・・ ・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・ A. japonica
10. Inland plants; sap colorless or inconspicuous; abaxial surface of leaves not shiny; dorsal ribs of mericarp filiform, not winged.
11. Basal leaves withered at anthesis; lobes of leaflets incised, margins irregularly serrate; calyx teeth small, narrowly triangular ・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・ A. polymorpha
11. Basal leaves persistent; lobes of leaflets not incised, margins regularly serrate; calyx teeth obsolete.
12. Stems dark green or dark purple; leaf sheaths slightly inflated; anthers dark purple; vittae 8–12 ・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・・ A. amurensis
12. Stems pale green; leaf sheaths strongly inflated; anthers pale yellow; vittae 6 ・・・・・・・・・ A. dahurica
Acknowledgements
This study was supported by a National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT) (grant number RS-2023-00208589) to J.-H. Song. We thank Eun Sook Jung and Gyu Seok Kim for providing information on natural habitats and for collecting additional plant samples. We also express our gratitude to Mi Yeon Ji for providing us with elaborate illustrations. We sincerely appreciate the comments and corrections of the anonymous reviewers, which have improved our manuscript.
Notes
CONFLICTS OF INTERST
The authors declare that there are no conflicts of interest.