Helonias, Tanaka, 2019

Tanaka, Noriyuki, 2019, Taxonomy, evolution and phylogeography of the genus Helonias (Melanthiaceae) revisited, Phytotaxa 390 (1), pp. 448-450 : 448-450

publication ID

https://doi.org/ 10.11646/phytotaxa.390.1.1

persistent identifier

https://treatment.plazi.org/id/03E5713A-FF99-9129-4099-FC07D6C0310F

treatment provided by

Felipe

scientific name

Helonias
status

 

12. Helonias breviscapa (Maxim.) Tanaka (1998a: 111) View in CoL ( Figs. 27 View FIGURE 27 , 28 View FIGURE 28 ).

Heloniopsis breviscapa Maximowicz (1867: 436) View in CoL . ≡ Heloniopsis japonica (Thunb.) Maxim. var. breviscapa (Maxim.) Honda (1938: 1678) View in CoL . ≡ Hexonix breviscapa (Maxim.) Wang & Tang (1949: 113) View in CoL . ≡ Heloniopsis orientalis (Thunb.) Tanaka var. breviscapa (Maxim.) Ohwi View in CoL (July 1953 b: 68). ≡ Heloniopsis orientalis (Thunb.) Tanaka subsp. breviscapa (Maxim.) Kitamura & Murata (1966: 72) View in CoL . Type (lectotype, designated here):― JAPAN. Kiusiu, in m. ignivomo Wunzen [Kyûshû, Nagasaki Pref., Mt. Unzen], medio Majo, 1863, C.J. Maximowicz s.n. (K-000901111*!). Isolectotypes: NY-00319713*–5*!, P-00730557*–9*!, US-00650432*!

=? Scilla japonica Thunberg View in CoL (in Murray 1784 May–June: 329; August 1784: 137; 1802: t. 4). ≡ Hexonix japonica (Thunb.) Rafinesque (1837: 13) View in CoL . Type:― JAPAN. Thunberg (Holotype: UPS-THUNB 8323*!). Japanese name:―Kochô (‘Kotjo’ in Thunberg 1784); Shôjôbakama ( Itô 1829, Iinuma 1861).

= Sugerokia nipponica Ohwi (1930: 566) View in CoL . ≡ Heloniopsis nipponica (Ohwi) Nemoto (1936: 1058) View in CoL . Type:― JAPAN. Honshû. Prov. Yamato [Nara Pref.], Yoshino, between Kashiwagi and Obagamine, 19 May 1928, J. Ohwi (Holotype: KYO, photograph in TNS!). Japanese name:―Yamato-shôjôbakama ( Nemoto 1936); Shirobana-shôjôbakama ( Hiyama 1939: 79).

= Heloniopsis japonica (Thunb.) Maxim. var. flavida Nakai (1933: 243) View in CoL . Japanese name:―Kibana-shôjôbakama ( Nakai 1933). ≡ Heloniopsis breviscapa Maxim. var. flavida (Nakai) Hara (1947: 149) View in CoL . ≡ Hexonix breviscapa (Maxim.) F.T.Wang & Tang var. flavida (Nakai) Wang & Tang (1949: 113) . ≡ Heloniopsis orientalis (Thunb.) Tanaka var. flavida (Nakai) Ohwi View in CoL (July 1953 b: 68. Type:― JAPAN. Honshû. Prov. Yamato [Nara Pref.]. Not designated. – Heloniopsis japonica (Thunb.) Maxim. var. grandiflora (Franch. & Sav.) Nakai (1933: 243) View in CoL , p.p., excl. syn.

= Heloniopsis japonica Maxim. var. yakusimensis Masamune (1934: 551) View in CoL . Japanese name:―Hime-shôjôbakama ( Masamune 1934). ≡ Heloniopsis yakusimensis (Masam.) Honda (1938: 1679) View in CoL . ≡ Sugerokia yakusimensis (Masam.) Koidzumi (1939: 53) View in CoL . ≡ Heloniopsis breviscapa Maxim. var. yakusimensis (Masam.) Hara (1947: 148) View in CoL . ≡ Hexonix breviscapa (Maxim.) F.T.Wang & Tang var. yakusimensis (Masam.) Wang & Tang (1949: 113) . ≡ Heloniopsis orientalis (Thunb.) Tanaka var. yakusimensis (Masam.) Ohwi View in CoL (July 1953 b: 68). ≡ Heloniopsis yakusimensis (Masam.) Masamune (1969: 11) View in CoL , isonym. Type:― JAPAN. Kyûshû. Kagoshima Pref.: Isl. Yakushima, ca. 1500 m, G. Masamune (from the protologue). The following specimen is possibly holotype, as it coincides well with the original description: Isl. Yakushima, 6 April 1927, G. Masamune (TI!).

= Heloniopsis japonica (Thunb.) Maxim. var. albiflora Honda (1938: 1678) View in CoL . ≡ Heloniopsis breviscapa Maxim. var. albiflora (Honda) Hara (1947: 149) View in CoL . Type:― JAPAN. Honshû. Prov. Musasi [Tokyo Pref.], Sekido, [3 April] 1938, T. Sakai s.n. (Holotype: TI!). Paratypes cited in the additional specimens examined below. Japanese name:―Sirobana-shôjôbakama ( Honda 1938).

= Heloniopsis japonica (Thunb.) Maxim. var. tesselata Nakai ex Honda (1938: 1679) . Type:― JAPAN. Honshû. Prov. Awa [Chiba Pref.], Mt. Kiyozumi, April 1937, T. Nakai s.n. (Holotype: TI!). Japanese name:―Kiyozumi-shôjôbakama ( Honda 1938).

Heloniopsis pauciflora View in CoL auct. non A.Gray: Miquel (1867: 146), as to specimens from Kiusiu by Maximowicz, ‘ Helionopsis ’.

Sugerokia japonica Miquel (1870: 52) View in CoL , p.p., as to synonyms? Scilla japonica View in CoL , Heloniopsis breviscapa Maxim. View in CoL , and specimens from Mt. Wunzen, Kiusiu [distributed under the name Heloniopsis pauciflora View in CoL by Maximowicz with the date of 1865].

Heloniopsis japonica View in CoL auct. non Maxim.: Matsumura (1905: 197), p.p.; Honda (1938: 1677), p.p.

Sugerokia japonica View in CoL auct. non Miquel: Koidzumi (1930: 95).

Japanese name:―Kochô-shôjôbakama (Tanaka 1998, based on ‘Kojyo’ for Scilla japonica View in CoL in Thunberg 1784: 137), Shirobana-shôjôbakama (e.g. Iwasaki 1824; Makino 1891, under Heloniopsis japonica Maxim. View in CoL , p.p.?; Koidzumi 1930), Tsukushi-shôjôbakama ( Honda 1938), Yakushima-shôjôbakama ( Honda 1938).

Description:―Rhizome cylindrical, subannulate with many close scars, to 5 cm long, to 1.5 cm in diam., pale brown. Roots filiform, to 1.5 mm in diam., white, some showing signs of contraction. Leaves hysteranthous, main fresh rosette leaves 7–15, persistent usually for fully 1 year (sometimes for ca. 2 years), spatulate or narrowly oblanceolate, 3.5–21.5 cm long, to 3.5 cm wide, often minutely undulate at margin, apex with apiculus 0.5–1 mm long, light (often yellowish) green, chartaceous (thin to moderately thick in texture), young fresh leaves somewhat glossy, mature leaves almost dull. Stem usually ca. 2–17 cm long at anthesis, becoming elongate toward fruiting stage; peduncle to 37 cm long in fruit; scale-like leaves 5–11 (excl. basal ones), linear-lanceolate, narrowly lanceolate, narrowly elliptic-oblong, to ca. 3.8 cm long, apex acute or acuminate; inflorescence compactly racemose or sub-umbellate, rachis to 1.7 cm long at anthesis, to 8 cm long in fruit; pedicels terete, 2–7 mm long at anthesis, to 23 mm long in fruit, 6-ribbed, ebracteate. Flowers 1–11, usually nodding at mid anthesis, funnelform, 1–1.9 cm across, scent or odorless. Tepals 6, obconically expanded, sometimes slightly recurved distally, 5–7(–9)-veined, white, often tinged (purplish) pink proximally, turning creamy yellow or sometimes pink with progression of anthesis, eventually green after anthesis, (oblong-)oblanceolate or spatulate, 8.3–14 mm long, 1.8–5.2 mm wide, apex rounded to subacute; abaxial basal part scarcely to moderately gibbous; adaxially shallowly to moderately cannaliculate toward base, proximal submarginal portions winged (lamellate) inward; wings extending over proximal ca. 1/4 of tepal, connate proximally to those of adjoining tepals for 0.25–1 mm in height, and also adnate to opposite filament for 0.7–1.8 mm in height, forming a sheathing nectary; surface of the nectary positioned nearly as high as base of ovary (or top of receptacle); base of tepal affixed to receptacle (apex of pedicel) for ca. 1.5 mm in length. Stamens 6, equaling or exceeding tepals; filaments filiform, 5–12.5(–15) mm long, 0.4–0.6 mm wide, nearly straight, scarcely dilated but slightly complanate toward base, white or pale purplish pink; anthers affixed to filament on adaxial portion close to base of connective (i.e. [sub]dorsi-basifixed), bilocular, (dark) purple, narrowly ovate or oblong-ovate, (1.5–)2–3.5(–4.5) mm long, base sagittate, apex scarcely confluent between thecae, sometimes minutely apiculate; pollen whitish. Pistil 1, 9–19.5 mm long, exceeding stamens and tepals; ovary globose, slightly trilobed, 2–4 mm long, 2–5 mm across, apex emarginate-concave, (pale yellowish) green or dark purplish pink, subdistal portion of carpel divergently protruding in fruit; ovules numerous, multiseriate (ca. 8-seriate) on central axile placentae; style terete, slightly narrowed proximally, straight or sometimes upcurved distally in nodding flowers, 6.7–13(–18.5) mm long, white or pale purplish pink; stigma discoid or depressed capitate, sometimes obtusely trigonous, or slightly trilobed to tricleft, 1.0–2.5(–3.3) mm across, white or pale purple. Capsules antrorse, broadly obpyramidal, tripartite, subtriangular in frontal aspect, obcordate or broadly obdeltoid in lateral aspect, 5–8 mm long, 1–1.5 cm across; lobes mitriform (or ovoid-pyriform), ascending, strongly divergent. Seeds with white or pale brownish testa, linear-fusiform, usually falcate, 4.8–6.7 mm long, 0.3–0.5 mm wide, proximally sublinear, distally subulate; body of seed oblong-fusiform, 1.4–1.7 mm long, 0.3–0.4 mm wide, brown.

Additional specimens examined (see also Tanaka 1998a):― JAPAN. Honshû. Chûbu Distr. Shizuoka Pref.: Mt. Higaneyama, 10 April 1920, fl., Ôsaka-fu Joshi Senmon-gakkô (MAK-137590). Chûgoku Distr. Yamaguchi Pref.: Tokuyama-shi, Nakasu-kita, Datoko, 11 April 1971, fl., H. Manabe no. 3 (TNS); Abu-gun, Tokusa, Mt. Nodôyama, ca. 900 m, N. Miake no. 5 (TNS-285022). Kantô Distr. Chiba Pref.: Awa-gun, Mt. Kiyosumi, May 1938, fl., S. Asano (TNS-59232). Kanagawa Pref.: prov. Sagami, Mt. Hakone, [May] 1910, fl. yellow, G. Koizumi s.n. (paratype of Heloniopsis japonica var. albiflora, TI !); Mt. Tanzawa, April 1960, fr., D. Yoshikawa (TNS-140984). Tokyo Pref.: Tama-shi, Sekido 7 April 1952, fl., U. Mizushima (TI, TNS-02633). Kinki Distr. Hyôgo Pref.: Kôbe, Mt. Rokkô, 11 April 1948, fl., C. Kataoka (TNS-132386). Kyôto Pref.: prov. Yamashiro, Mt. Hiei, [29 March] 1911, fl.-albi, T. Nakai s.n. (paratype of Heloniopsis japonica var. albiflora, TI !). Mie Pref.: Ise, Rokusôdani, 24 March 1961, fl., S. Okuyama 13565 (TNS-261521). Nara Pref.: Yoshino-gun, Mt. Mikamiyama, May 1888, fr., H. Sakurai (TNS-3836). Ôsaka Pref.: Mt. Kongôsan, 25 April 1953, fl., M. Togashi 720 (BM-001118052, TNS-105921); Mt. Kongôsan, 5 May 1935, fr., Z. Yoshino (TI). Wakayama Pref.: Mt. Kôyasan, Yatate, 20 Mar., G. Murata 22166 (TNS-305431). Kyûshû. Fukuoka Pref.: Nogouchi-keikoku, 27 March 1932, fl., H. Arao (TNS-255640). Kagoshima Pref.: Mt. Takakuma, 20 March 1944, fl., Y. Satake & S. Okuyama (TNS-79936–79938); Isl. Sakurajima, April 1904, fl., Y. Yoshise (MAK-137580). Kumamoto Pref.: Mt. Aso, Akamizu, 500 m, 2 April 1954, fr., Y. Shimada (TNS-248995); Kami-mashikigun, Yabe-machi, 280 m, 1 April 1994, T. Kawasaki 1059 (TNS-605322). Miyazaki Pref.: Sakatani-mura, Inotani, 22 Mai 1942, fr., S. Hattori 760 (TNS-307327); Kawaminami-mura, 10 March 1950, fl., M. Nagasawa (TNS-98301). Nagasaki Pref.: Simabara in montibus, bot. Japonicus s.n. (LE-01011595*); Mt. Unzendake, 29 March 1912, fl., Z. Tashiro (TNS-29965). Ôita Pref.: Mt. Kujyû, 24 April 1908, Nakamura 698 (TNS-79925). Saga Pref.: Ogi-gun, Mt. Hikodake, 22 April 1980, fr., T. Hashimoto (TI). Shikoku. Ehime Pref.: Shingû-mura, March 1889, fl., T. Makino (TNS-3838). Kôchi Pref. : Kami-gun, mononobe-mura, 3 March 1976, fl., K. Suzuki (MAK-270211); Nagaoka-gun, Ôtoyo-mura, Tatekawa, 5 May 1893, T. Makino (MAK-137567). Tokushima Pref.: Mt. Kôtsusan, 30 April 1903, J. Nikai (TNS-49296); Naka-gun, Aioi-chô, 300 m, 22 April 1983, fr., M. Takahashi & K. Ueda 1522 (TI); Mima-gun, Tsuzuro, 28 March 1940, fl., T. Yoshinaga (TI).

Distribution:― Japan: Honshû (southern Kantô District and westward to Chûgoku District), Shikoku, Kyûshû ( Fig. 29-B View FIGURE29 ).

Habitat:―Shady or semishady moist rocks or slopes usually along or near streams in forests at elevations ca. 150–900 m.

Conservation status:―The species is comparatively widespread, and assessed as LC according to the criteria set out in IUCN (2001). However, local populations near human habitation may be more or less subject to disturbance and threats.

Flowering:―March–April.

Ripening:―May.

Remarks:―Plants of Helonias breviscapa here circumscribed have often been treated as conspecific with H. orientalis (including its synonymous taxa) or as infraspecific taxa of the latter ( Miquel 1870, Ohwi 1953a, b, 1965, Kitamura 1964, Kitamura & Murata 1966). The two species may look somewhat alike, but H. breviscapa differs from H. orientalis by the thinner, less glossy, somewhat lighter (yellowish) green leaves usually minutely undulate at the margin ( Fig. 27G View FIGURE 27 ), usually smaller flowers, and white tepals of which the proximal submarginal wings are relatively shorter and the nectary is positioned nearly as high as the ovary base ( Fig. 28B, C View FIGURE 28 ; Tanaka 1997b, 1998a). Compared with H. orientalis , H. breviscapa is usually in warmer regions with less snowfall in winter in Japan. Their geographic ranges overlap partly ( Fig. 29-O, -B View FIGURE29 ) ( Tanaka 1997e, 1998a; Kawano et al. 2007). They were recognized as distinct by Maximowicz (1867), Hara (1947), and Tanaka (1998a). A few local forms of Helonias breviscapa have hitherto been recorded. For example, plants from Kyûshû ( Figs. 27A View FIGURE 27 , 28A, B View FIGURE 28 ), which were recognized as Heloniopsis breviscapa var. breviscapa ( Hara 1947) , usually have comparatively broad oblanceolate tepals which are abaxially often tinged (pinkish) purple and scarcely gibbous at the basal portion. Those from Isl. Yakushima, southern Kyûshû, which were distinguished as H. breviscapa var. yakusimensis ( Hara 1947) , are usually somewhat smaller in habit and have 1 or a few flowers and slightly thicker leaves. Those from the Kinki District, central Honshû ( Figs. 27C, D View FIGURE 27 , 28D View FIGURE 28 ), named H. breviscapa var. flavida ( Hara 1947) , have white, slightly longer, narrowly oblong-oblanceolate tepals with a slightly gibbous base. When some typical representatives of these varieties are compared, they may look somewhat distinct from one another. However, they are rather indistinct in delimitation due to a gradational variation between them when many samples from different localities are compared ( Tanaka 1998a). For instance, plants from eastern Shikoku (Tokushima Pref.; Figs. 27B View FIGURE 27 , 28C View FIGURE 28 ) have tepals scarcely inflated at the base and often tinged with purple toward the base like those from Kyûshû. Some plants from the Kinki District of Honshû also have tepals tinged purplish proximally and scarcely gibbous at the base ( Figs. 27D View FIGURE 27 , 28D View FIGURE 28 ). Fuse et al. (2004) stated that the leaves of H. breviscapa var. flavida are somewhat thin and spotted with reddish brown in winter, while those of H. breviscapa var. breviscapa are slightly thicker and spotted with brown to blackish purple in winter. In my observation, however, this distinction is not necessarily constant between them. For instance, under a cultivated condition in Tokyo, the leaves of plants from Kyûshû ( H. breviscapa var. breviscapa ) often turn reddish brown in winter. Those from different localities (Kyûshû, Shikoku, and the Kinki District of Honshû) often turn similarly brownish, irrespective of the original localities ( Fig. 27A–D View FIGURE 27 ). Such coloration of the leaves appears to reflect coldness in winter, and may change to some extent in response to the fluctuation of temperature and habitat conditions. It seems more reasonable to regard these varieties ( breviscapa , yakusimensis , flavida) as local forms (races) not distinctly delimited from one another. I have noticed that the flowers of some plants from Kyûshû (Fukuoka and Miyazaki Pref.) and Shikoku (Tokushima Pref.) are fragrant, while those from the Kinki District (Nara, Mie, and Wakayama Pref.) are odorless. This intraspecific variation in floral fragrance needs a further survey. Despite these variations, Helonias breviscapa as a whole appears to retain a fairly high unity, possessing a unique combination of features as noted earlier. It is noteworthy that Helonias breviscapa ( Fig. 27G View FIGURE 27 ) and H. koreana ( Fig. 23H View FIGURE 23 ) share somewhat thin, dull or subdull leaf blades usually with minutely undulate margins. They usually occur in regions which are drier and less snowy in winter than the habitats of H. orientalis . Helonias breviscapa occurs in warmer regions than H. koreana . The relationship between the species of H. ser. Heloniopsis and the geographic and climatological features of their ranges are dicussed in Part III-5. Tanaka (1998a) once assigned Scilla japonica to Helonias breviscapa on the basis of the description by Thunberg (1784) (especially of floral color: corolla albo purpurea ) and the scapes figured in Thunberg (1802). Very recently (July 2018) I had an opportunity to examine the image of the scapes of the type material (Thunberg s.n., UPS-THUNB 8323*). Unexpectedly, the scapes looked fairly similar to those of H. orientalis . Further studies are necessary to precisely identify them. Scilla japonica is tentatively placed in the above synonymy. Nakai (1933) described the perianth of Heloniopsis japonica var. flavida as greenish yellow at anthesis, but did not cite any specimen. White tepals of Helonias breviscapa turn yellow or yellowish green soon after the peak anthesis, and moreover, the floral parts are persistent even at the fruiting stage. So the flowers after anthesis are often taken as those at anthesis with yellow or yellowish green tepals. At present there is no means to examine Nakai’s taxon as to the flowering stage and floral color on a specimen basis. A few herbarium specimens of Helonias breviscapa record the flowers as yellow, e.g., Kanagawa Pref., Hakone, May 1910, fl. yellow, G. Koizumi s.n. (paratype of Heloniopsis japonica var. albiflora, TI !); Mie Pref., Suzuka Mts, Gozaisho-dake, 20 Apr. 1941, fl. yellow, Morimoto no. 5 (TI!). The flowers appear to be past the flowering stage, implying the possibility that they were white at anthesis. Doi (1983: 135) recorded a yellow-flowered form of Helonias breviscapa as “ Heloniopsis orientalis C. Tanaka subsp. breviscapa Kitam. f. satsumensis Doi ” based on a specimen from southern Kyûshû, Japan (Kagoshima Pref., Ijûin-chô, April 1938, Toshio Shin, currently preserved at HIRO, the image seen). However, the name did not accompany a description nor a diagnosis in Latin, hence it was invalidly published. He noted that the flowers are yellow from the beginning of flowering. Such a yellow-flowered form has not been reported from anywhere since then. Studies based on fresh living plants seem necessary for a more precise understanding of this form.

Part III. Evolution and phylogeography

Evolutionary aspects of various characters (Part I) and taxa (Part II), and the historical background of the current distribution of Helonias (Part II) are discussed.

1. Pseudoduality in generation of flowering stems in vernal species, and variation in dormancy of leaf and flower buds

Pseudodual generation of flowering stems in vernal species, and its origin:―In mature plants of vernal species, such as Helonias bullata , H. thibetica and H. orientalis , two kinds of buds are produced in a basal rosette ( Fig. 1A, B View FIGURE 1 ; Part I-2). One is a flower bud (inflorescence bud) that develops into a flowering stem (sexual reproductive organ; F in Fig. 1 View FIGURE 1 ); the other is a leaf bud that develops a basal rosette of leaves (vegetative organs; L in Fig. 1 View FIGURE 1 ). Both types of buds may be regarded as of the same generation, since they develop in the same growing season ( Fig. 1C, D View FIGURE 1 ). Contemporaneous development of two kinds of buds in Helonias is unusual. Many related perennials, such as Veratrum L. and Anticlea Kunth ( Melanthiaceae - Melanthieae; Dahlgren 1985), produce only one kind of bud that develops into vegetative and sexual reproductive organs consecutively. In these plants, development of vegetative organs precedes development of sexual organs. They complete their yearly growth cycle within a single growing season (from spring through autumn). Helonias differs from Veratrum and allied genera in having leaves that persist after a single season, but their mode of development appears basically similar. As in Veratrum , the two organs of Helonias develop successively. That is, in temperate vernal species of Helonias , a leaf bud sprouts in the growing season (spring to summer), and around the time when the development of most fresh leaves is nearly completed, the flower bud starts to form (usually from July to September). Unlike in Veratrum and its allies, the flower bud of Helonias is formed, but does not sprout. Instead, it remains dormant until the coming spring (F in Figs. 1A, B View FIGURE 1 , 4 View FIGURE 4 ). The vegetative and the sexual organ of Helonias thus develop successively as in Veratrum , but their growth periods are largely separated by a cold season inserted in the middle of the yearly growth cycle. A flower bud initially forms in one year and the fresh flowering stem develops in the next (F in Fig. 1C, D View FIGURE 1 ). Rosulate leaves sprouted in the preceding year (R in Fig. 1 View FIGURE 1 ) are regarded as of the same generation, since they originate within a single growth cycle, then develop sequentially. Two kinds of buds sprouting nearly concurrently in the spring from a rosette (F and L in Fig. 1C, D View FIGURE 1 ), mentioned above, are regarded as two generations, or two consecutive generations. That is, the flower bud (F in Fig. 1A, B View FIGURE 1 ) is of the same generation as the rosulate leaves sprouted in the preceding year (R in Fig. 1C, D View FIGURE 1 ); the leaf bud (L in Fig. 1A, B View FIGURE 1 ) to sprout in the current growing season is of the next generation. Thus the flower buds (flowering stems) superficially look dual in generation. Supposedly, ancestral Helonias grew in warmer regions with no severe cold season, and the two kinds of organs developed nearly consecutively as in Veratrum . Then the climate became cooler and was accompanied by a cold season or, alternatively, the plants extended their range to cooler regions with severe winters. Through the course of evolution, the ancestors eventually acquired a gapped, prolonged yearly growth cycle in which the development of the sexual organs became separated seasonally from development of vegetative organs due to the insertion of a cold season. This yearly growth cycle naturally requires a longer span than a non-gapped growth cycle as seen in Veratrum . It takes a span over two years (or two growing seasons with a cold season in between) when estimated from the time of sprouting of a leaf bud (in spring) to the death of a flowering stem in the following year (usually in late spring to early summer). If estimated from the time of initial formation of a leaf bud, usually from summer to early autumn in the year preceding sprouting, then an entire growth cycle spans over three years (or three growing seasons with two cold seasons in between) ( Fig. 2 View FIGURE 2 ). Chamaelirium (including Chionographis ), which is closely allied to Helonias ( Tanaka 1997e, 2017a, b), also has such a gapped, prolonged yearly growth cycle and a flowering stem pseudodual in generation, suggesting that Chamaelirium and Helonias are sister groups sharing a common ancestor.

Reduced dormancy in leaf and flower buds of Asian subtropical species:―In Helonias kawanoi and H. jinpingensis in the subtropics, anthesis is usually in late summer to autumn. In H. kawanoi , the flower bud sprouts shortly after most fresh leaves nearly complete their development ( Figs. 2 View FIGURE 2 , 16A, D. E View FIGURE 16 ). That is, the two organs develop nearly consecutively. This mode of development differs from other temperate vernal species in which the sprouting of a flower bud is suppressed until the spring (F in Figs. 1A, B View FIGURE 1 , 4 View FIGURE 4 ). The yearly growth cycle of H. kawanoi and H. jinpingensis was probably secondarily acquired as a result of adaptation to a warm subtropical climate. Under such a climate with no severe cold season, dormancy of the flower bud may have been reduced or lost. In temperate vernal species of Helonias , the leaf bud is similar to the flower bud, initially formed from summer to early autumn in the central portion of a rosette, then remaining dormant until spring (L in Figs. 1A, B View FIGURE 1 , 4 View FIGURE 4 ). In H. kawanoi , H. leucantha and H. umbellata , however, the leaf bud often begins to sprout in autumn and winter ( Figs. 16A, D, E View FIGURE 16 for H. kawanoi , 18F for H. leucantha ), although growth appears more or less suppressed under the cooler conditions ( Fig. 2 View FIGURE 2 ). In these species the dormancy of the leaf bud appears to have been reduced. The basic mode of development of ancestral Helonias (or of the Helonias and Chamaelirium lineage) might have been similar to that of the subtropical species. The phenological process of a yearly entire growth cycle in seven species of Helonias is schematized in Fig. 2 View FIGURE 2 . Ancestral Helonias is presumed to have existed somewhere in the northern circumpolar region in the Paleogene when temperatures became predominantly lower toward the end of the period (for a geological and climatological aspect, see: Chaney 1947, Tanai 1967, 1972, 1988, 1992, Uemura 1995, Ogasawara & Uemura 2006). The historical background of the migration or extension of the ancestral populations of Helonias is discussed below (section 5).

2. Trends in the evolution of characters and structures

Evaluation of character states from an evolutionary perspective:―In the course of evolution, simple structures of plants often become more complex by fusion of similar or dissimilar parts or by developing additional structures. In general, with an increase of such complexity, the level of structural organization or specialization becomes higher. It is often difficult to determine evolutionary directions in phenotypic characters, but in characters showing stepwise structural changes, it seems possible to draw fairly reasonable inferences on directions. In Helonias , for instance, the tepals and the filaments of several species (e.g., H. sect. Helonias , H. subsect. Ypsilandra ) are free and the nectaries are more or less shallow, while those of some other species (H. ser. Heloniopsis ) are adnate to form saccate or tubular nectaries (Part I-8). These are judged to be derived, more advanced structures (i.e. apomorphies). As for the pistil, the degree of connation among three carpels to form a single columnar style varies among taxa (Part I-8). It is evident that the carpels least connate among them are most primitive in state, and those entirely connate along their length are most advanced (Part III-2-3). Evolution in the reverse direction (or retrogressive evolution), from a higher level of organization or specialization to a lower level, is highly improbable in these structures of Helonias . Inferences from these structures therefore appear to be applicable to the determination of evolutionary directions of characters (Part III-2) and to the elucidation of the process of phyletic diversification in this genus (Part III-3, -4). In inferring evolutionary trends of characters, the corresponding character states in other related genera, such as Chamaelirium , Xerophyllum and Veratrum , were also taken into account (e.g. floral color, nectar-secretion in tepals, morphology of ovaries, capsules and seeds). Table 10 shows evolutionary trends of several characters thus inferred.

Inflorescence:― Helonias has different types of inflorescence (Part I-7). Helonias sect. Helonias and H. subsect. Ypsilandra (H. sect. Heloniopsis ) have a raceme, while H. subsect. Heloniopsis has a raceme, subumbel or an umbel. From a structural point of view, umbels appear more advanced or specialized than racemes. Flowers considered to be primitive, such as those of H. sect. Helonias , are borne in racemes, while flowers regarded as advanced, such as those of H. ser. Heloniopsis , form umbels . In related genera such as Chamaelirium , Xerophyllum and Veratrum , racemes or panicles are common ( Britton & Brown 1896, Tamura 1998, Utech 2002). Considering these facts, it is highly probable that the inflorescence of Helonias was originally a raceme, but evolved to become a subumbel or an umbel through abbreviation of the rachis. With this change, the arrangement of flowers on the rachis also changed from spiral (in a raceme; e.g. Fig. 13C View FIGURE 13 ) to subverticillate (in an umbel; e.g. Fig. 23E View FIGURE 23 ). In H. sect. Helonias and H. subsect. Ypsilandra , the nectary at the base of the tepal is shallow and the nectar is exposed (Part I-8), so even insects with short mouthparts (trophi) could suck the nectar. In H. ser. Heloniopsis (H. subsect. Heloniopsi s), the nectary is saccate or tubular (Part I- 8), suggesting that the flowers of H. ser. Heloniopsis are adapted to pollinators that can extract nectar from such deeper nectaries. The evolutionary shift from raceme to umbel may also be related to the kinds and habits of pollinators. This shift is supposed to have improved the efficiency in pollination. Whether the inflorescence is a raceme or an umbel, flowers in aggregation (e.g. Figs. 4A–E View FIGURE 4 , 13C View FIGURE 13 , 25A–D View FIGURE 25 ) must have been more effective in attracting insects. This state is likely to have increased the opportunity for pollination. The ebracteate inflorescence of Helonias (e.g. Figs. 13C View FIGURE 13 , 18D View FIGURE 18 , 23G View FIGURE 23 ) is also shared by the closely allied Chamaelirium ( Tanaka 2017b) , so it must have been the original state. The bracteate inflorescences of H. jinpingensis and H. kawanoi ( Fig. 16C View FIGURE 16 ) appear to be secondarily derived ( Table 10).

Pistil:―In Helonias , the three carpels are connate to form a columnar style the relative length of which differs among the species (Part I-8; Tanaka 1997c, e). The style is shortest in H. bullata ( Fig. 5D View FIGURE 5 , Table 6; Tanaka 1997a). In H. thibetica (H. subsect. Ypsilandra ) and H. subsect. Heloniopsis the style is well developed and the pistil often exceeds the stamens (Part I-8; e.g. Figs. 13B, C View FIGURE 13 , 16B View FIGURE 16 , 27A–D View FIGURE 27 , 28A View FIGURE 28 , Table 6 for H. thibetica ). In Helonias , pistils with a relatively short style (columnar part) and long stigmatic segments, such as those of H. bullata ( Fig. 5D View FIGURE 5 ) and H. yunnanensis var. yunnanensis ( Fig. 7B View FIGURE 7 ), are regarded as more primitive. This state apparently shifted evolutionarily toward a relatively long style and short stigmatic segments. A pistil with a long style and discoid or (sub)capitate stigma occurs in H. thibetica (e.g. Figs. 3A–C View FIGURE 3 , 14A View FIGURE 14 ) and subsect. Heloniopsis (e.g. Figs. 3D, E View FIGURE 3 , 20D View FIGURE 20 , 22A View FIGURE 22 , 24F View FIGURE 24 , 28A View FIGURE 28 ; Tanaka 1997c, e), and is regarded as the most advanced state. The evolution of the pistil in Helonias thus appears to be directional and anagenetic.

Anthers:―The anthers of Helonias vary in structure, shape, size, and in position relative to the stigmas (Part I-8). The anthers of H. bullata (H. sect. Helonias ) and H. subsect. Ypsilandra are unilocular and basifixed (e.g. Figs. 5B View FIGURE 5 , 14E View FIGURE 14 ). They expose their pollen not only laterally but also apically ( Fig. 2 View FIGURE 2 in Tanaka 1997a), hence they may readily smear pollen on insects foraging for pollen and/or nectar. In H. subsect. Heloniopsis the filament tends to be located closer to the opposing tepal ( Figs. 19B–E View FIGURE 19 , 22A–E View FIGURE 22 , 28B, C, E View FIGURE 28 ; Tanaka 1997b, c, e). The anthers are virtually bilocular, extrorse, and often comparatively long ( Figs. 22F View FIGURE 22 , 24F View FIGURE 24 , 28F View FIGURE 28 ; Tanaka 1997c), accordingly their pollen is likely to be smeared on insects trying to extract nectar from the base of the tepals. They appear to be more adapted to insects seeking nectar ( Tanaka 1997e), although insects seeking only pollen, such as flies, also visit and pollinate the flowers, as reported by Takahashi (1988). During the course of evolution the anthers of Helonias appear to have changed from unilocular to bilocular ( Tanaka 1997c). Almost perfectly bilocular anthers are produced in H. breviscapa ( Fig. 28F View FIGURE 28 ; Tanaka 1997c) and may represent the ultimate evolutionary outcome. It is also noteworthy that the anthers of H. ser. Heloniopsis , especially of Helonias koreana ( Fig. 24F View FIGURE 24 ) and H. orientalis , are (sub)dorsi-basifixed, i.e. the anther is affixed to the filament adaxially slightly above the connective base (Part I-8). This state is obviously apomorphic.

Floral structure:―Flowers of Helonias are highy diversified especially in structure (Part I-8). For instance, the tepals are almost flat or only slightly canaliculate in H. sect. Helonias and H. subsect. Ypsilandra , while they are more prominently grooved especially in H. ser. Heloniopsis . Nectaries are nearly flat or only slightly concave in H. sect. Helonias and H. subsect. Ypsilandra , they are markedly saccate in H. ser. Heloniopsis (Part I-8). In H. sect. Helonias , H. subsect. Ypsilandra and H. ser. Umbellatae , the filaments and tepals are distinct or almost so. They are basally fused in H. ser. Heloniopsis (Part I-8). From a structural point of view, the floral evolution of Helonias appears to have advanced stepwise from the state in H. sect. Helonias ( Figs. 4 View FIGURE 4 , 5 View FIGURE 5 ) and H. subsect. Ypsilandra (e.g. Fig. 7 View FIGURE 7 ), through H. ser. Umbellatae (e.g. Figs. 17 View FIGURE 17 , 19 View FIGURE 19 , 20 View FIGURE 20 , 22 View FIGURE 22 ), and lastly to H. ser. Heloniopsis (e.g. Figs. 24 View FIGURE 24 , 26 View FIGURE 26 , 28 View FIGURE 28 ; Tanaka 1997e). The flower of H. ser. Heloniopsis , which is made up of a single pistil and six surrounding subunits each consisting of a stamen and the opposing tepal that jointly form a saccate or tubular nectary, is viewed as an integrated apparatus for achieving efficient pollination through the behavior of certain insects ( Tanaka 1997e).

Ties with insects:―The flowers of many species of Helonias are fragrant and the tepals are nectariferous and often pink, bluish, violet or white (Part I-8). It is obvious that these features effectively attract insects foraging for nectar and/or pollen (Part I-11, Takahashi 1988). The flowers of Helonias are basically similar to one another, so they must have been coevolving with the insects since their origin. In this connection, it is notable that the flowers of H. leucantha (H. ser. Umbellatae ), H. breviscapa , H. koreana , and H. orientalis (H. ser. Heloniopsis ) are often odorless or nearly so, perhaps indicating that fragrance is not always necessary for attracting some kinds of pollinators, which may discern the flowers by other sensory means, such as vision. In the most closely allied Chamaelirium the flowers have no nectaries ( Tanaka 2017b). It is therefore obvious that close ties with insects foraging for nectar have been a principal factor for the evolution of Helonias ( Fig. 30 View FIGURE 30 , List 2).

Traits securing seed production under poor availability of pollen vectors:― Helonias orientalis blooms in early to mid spring. Although infrequently visited by insects ( Takahashi 1988, as Heloniopsis ), the flowers produce numerous seeds in the field ( Kawano & Masuda 1980). The following three traits of this species appear to cooperate effectively to secure seed production under poor availability of pollen vectors: high self-compatibility (known also in H. bullata and H. thibetica ; Part I-12), long flowering duration (known also in many species such as H. bullata , H. thibetica and H. subsect. Heloniopsis ; Part I-8), and widely open, funnel-shaped flowers that are structurally highly accessible for a wide variety of insects (known also in H. bullata ; Part I-11). Judging from the high similarity in sexual reproductive characters (e.g. self-compatibility, floral morphology, flowering phenology, pollination ecology) among the species, these three traits appear to have been cooperating as essential, effective sexual means to survive over the course of evolution in Helonias .

Evolution of structurally advanced flowers of Helonias ser. Heloniopsis coupled with a reduction in flower number and an increase in ovule number:―Flowers of Helonias ser. Heloniopsis appear to be the most advanced in Helonias . For instance, the tepals are proximally adnate to the opposing filament to form a saccate or tubular nectary ( Utech & Kawano 1981 for Heloniopsis orientalis ; Tanaka 1997b, e) and also connate to adjoining ones ( Figs. 24C–E View FIGURE 24 , 26A, C View FIGURE 26 , 28E View FIGURE 28 ). The inflorescence comprises a comparatively small number of flowers (up to 14; Figs. 23 View FIGURE 23 , 25 View FIGURE 25 , 27 View FIGURE 27 ) each bearing numerous ovules (e.g. to 189 per locule in H. orientalis ( Table 1); Utech 1978 for H. orientalis ; Tanaka 1997c, d). Such advanced floral structures appear to have evolved as a result of adaptation to insects foraging mainly for nectar. Presumably, the flowers improved pollination efficiency through restructuring and thereby enabled a reduction in flower number and an increase in ovule number in individual flowers. Utech (1978) reported that the total number of ovules per inflorescence is, on average, similar between Helonias bullata and Heloniopsis orientalis , although the number of flowers per inflorescence and the number of ovules per locule differ significantly. From my observations, I believe that H. bullata is the more primitive species and H. orientalis is one of the most advanced in the genus. The two species nearly represent the two extremes of phyletic evolution ( Fig. 30 View FIGURE 30 , List 2). Similar data come from studies on plastid DNA sequences ( Kim et al. 2016). The evolutionary trend toward H. orientalis , which produces comparatively fewer flowers and more numerous ovules, appears to be well reflected in this data. Floral evolution in Helonia s thus appears to have been directional and anagenetic. In both Helonias bullata and H. orientalis , seedling establishment in habitats is equally very low (Part I-15; Kawano & Masuda 1980, Sutter 1984), and asexual reproduction by rhizomic division is fairly limited. Accordingly, to survive, they need to produce numerous seeds enough to cover the low rates of seedling and rhizomic recruitment. This necessity of producing numerous seeds needed for survival may account for the similarity in the total number of ovules per inflorescence between the two species ( Utech 1978) mentioned above. Presumably, in H. bullata bearing comparatively few ovules (per locule or ovary; Table 1), production of a necessary number of ovules for survival was achieved by producing many flowers (per scape), while in H. orientalis having comparatively few flowers (per scape), it was attained by producing numerous ovules in individual flowers. Since the ovules similar in total number are produced by fewer flowers, the pollination efficiency in individual flowers of H. orientalis is regarded as higher than that of H. bullata . An increase in ovule number per locule may have resulted in producing narrower seeds, because a locular space available for fertilized ovules to develop is likely to be more or less limited. Divergence in seed width (Part I-10) must have occurred at an early stage of evolution in Helonias , as it is seen between the two sections H. sect. Helonias ( H. bullata ) and H. sect. Heloniopsis (including H. orientalis ).

3. Phylogenetic divergence and relationships

Evolutionary facets of the respective taxa (Part II) are outlined mainly on the basis of comparative studies of character states (Parts I, III-2).

Helonias :― Helonias is most closely allied to Chamaelirium in North America and eastern Asia ( Tanaka 1997e, 2017b). They share many similar features; for instance, the persistent oblanceolate or spatulate leaves in a basal rosette, a flowering stem with the growth period usually markedly separated from development of the rosette in the preceding year (Part III-1; F and R in Fig. 1C, D View FIGURE 1 ; Fig. 2 View FIGURE 2 ), peduncles with scale-like leaves (e.g. Figs. 1C View FIGURE 1 , 21A, B View FIGURE 21 ), ebracteate pedicels (e.g. Figs. 12A View FIGURE 12 , 13C View FIGURE 13 ), and carpels connate dorsally to adjoining ones along their ventral sutures (e.g. 26C–E). Helonias as a genus is monophyletic ( Fig. 30 View FIGURE 30 , List 2), consisting of two sections Helonias and Heloniopsis (Part II) .

Sect. Helonias :― Helonias bullata has three styles distinct nearly to the base ( Fig. 5D View FIGURE 5 ; Tanaka 1997a), suggesting that it is the most primitive member of the genus, at least in style characteristics. Helonias bullata also has fewer ovules (per locule or ovary) and slightly larger (broader) seeds than its congeners ( Fig. 5F View FIGURE 5 , Table 1; Tanaka 1997c). The surface of the nectary is nearly flat. These character states may also be primitive in Helonias . Helonias sect. Helonias is regarded as paraphyletic ( Tanaka 1997e).

Sect. Heloniopsis :―In Helonias sect. Heloniopsis the three styles (carpels) are consistently connate to one another partially or entirely along their length to form a single column (e.g. Figs. 9A–C View FIGURE 9 , 12 View FIGURE 12 , 14A View FIGURE 14 , 22A View FIGURE 22 ). This character state is regarded as more advanced than in H. sect. Helonias . Helonias sect. Heloniopsis is monophyletic ( Fig. 30 View FIGURE 30 , List 2), and all members are in eastern Asia (Figs. 15, 29, 31, 32). This section is composed of two subsections: H. subsect. Ypsilandra and H. subsect. Heloniopsis (Part II) . The former subsection is considered to be more primitive (Part II; Tanaka 1997e) and occurs west of the range of the latter ( Fig. 32 View FIGURE 32 ). The species of H. subsect. Heloniopsis are either insular or coastal ( Figs. 29 View FIGURE29 , 31-3 View FIGURE 31 , -4, 32-3, -4).

Subsect. Ypsilandra :― Helonias subsection Ypsilandra appears more advanced in several characters than H. sect. Helonias . For instance, the styles (columnar part) are longer ( Figs. 9A–C View FIGURE 9 , 10E View FIGURE 10 , 12 View FIGURE 12 , 14A View FIGURE 14 , Table 6), the stigmatic segments are relatively shorter ( Figs. 7B View FIGURE 7 , 9A–C View FIGURE 9 , 10E View FIGURE 10 , Table 6), and the stigmatic portion is sometimes simple and unlobed ( Figs. 3A–C View FIGURE 3 , 13A–C View FIGURE 13 ), the anthers are unilocular and not septate between the thecae ( Fig. 14E View FIGURE 14 ), and the seeds are significantly more numerous and narrower ( Figs. 7G View FIGURE 7 , 9D View FIGURE 9 , Table 1). In H. subsect. Ypsilandra , Helonias yunnanensis var. yunnanensis is regarded as the most primitive taxon, since it has the shortest style and a deeply trifurcate stigma ( Fig. 7B View FIGURE 7 , Table 6). All other members of the section are regarded as descendants from the ancestral lineage of this variety ( Fig. 30 View FIGURE 30 , List 2). Helonias jinpingensis with a short style and a somewhat deeply trifurcate stigma also appears to be primitive. Helonias alpina , with longer styles and slightly shorter stigmatic segments ( Fig. 10E View FIGURE 10 ) than H. jinpingensis , may be slightly more advanced. Helonias parviflora ( Figs. 11 View FIGURE 11 , 12 View FIGURE 12 ), with a very slightly trilobed (or nearly trigonous) stigma, is regarded as more advanced, at least in pistil characters. Helonias thibetica ( Figs. 13 View FIGURE 13 , 14 View FIGURE 14 ) with a relatively long slender style and a small capitate or discoid stigma ( Figs. 3A–C View FIGURE 3 , 14A View FIGURE 14 , Table 6) is the most advanced in pistil characteristics in this subsection. In H. subsect. Ypsilandra , the flowers have shifted from nearly adichogamous (in species other than H. thibetica ) to protogynous (in H. thibetica ) (Part I-13). This subsection is regarded as paraphyletic ( Tanaka 1997e), and presumed evolutionary relationships between the members are shown in Fig. 30 View FIGURE 30 (List 2). From plastid DNA sequence analyses, Fuse & Tamura (2016) suggested that Helonias s.str. (H. sect. Helonias in the present paper), Ypsilandra (H. subsect. Ypsilandra ) and Heloniopsis (H. subsect. Heloniopsis ) are all monophyletic. Heloniopsis is monophyletic as shown in Fig. 30 View FIGURE 30 and in Tanaka (1997e), but the monophyly of Helonias and Ypsilandra is not supported by the present study, because many characteristics of Heloniopsis , such as long united styles, discoid or subdiscoid stigmas, and filaments liberated from the ovary base (List 2-14), cannot have been derived if they did not pass through the precursory stages as possessed by Helonias and Ypsilandra . It is highly improbable that similar characteristics have developed independently in Ypsilandra and Heloniopsis as an outcome of parallel evolution, and the ancestral lineage of Heloniopsis , which linked Helonias and Heloniopsis directly and had intermediate character states between them, was entirely extinct, whereas species of Ypsilandra are extant. This discordance in results between morphology and DNA sequence analyses probably stems from the difference in research objects (whole characters of organisms manifested as the total sum of all genic activities involved vs. selected fractions of DNA molecules) and methods employed (clustering objects by synapomorphies vs. degree of similarity among DNA sequences).

Subsect. Heloniopsis :―Several character states of Helonias subsect. Heloniopsis are regarded as more advanced than those of H. subsect. Ypsilandra . For instance, the tepals are usually more markedly canaliculate toward the base ( Figs. 19D, E View FIGURE 19 , 20A, B View FIGURE 20 , 22D, E View FIGURE 22 , 24D, E View FIGURE 24 , 26A View FIGURE 26 , 28E View FIGURE 28 ), the nectaries are usually more prominently saccate ( Figs. 22D, E View FIGURE 22 , 24A–E View FIGURE 24 , 26A–C View FIGURE 26 ), the stamens tend to be located closer to the tepals ( Figs. 19B–E View FIGURE 19 , 22B–E View FIGURE 22 , 24B–E View FIGURE 24 , 26A–C View FIGURE 26 , 28B, C, E View FIGURE 28 ; Tanaka 1997e – Fig. 2 View FIGURE 2 ), and the anthers are virtually bilocular, extrorse, and often longer ( Figs. 20C View FIGURE 20 , 24F View FIGURE 24 , 28F View FIGURE 28 ). These character states are considered to be apomorphies, and this subsection is regarded as monophyletic in sharing them ( Fig. 30 View FIGURE 30 , List 2). Helonias thibetica (H. subsect. Ypsilandra ) has protogynous flowers, tepals slightly canaliculate toward the base, a pistil exserted beyond the tepals, and a capitate or discoid stigma. These features approach those of H. subsect. Heloniopsis (Part II) , suggesting that H. subsect. Heloniopsis originated from the ancestry of H. thibetica ( Fig. 30 View FIGURE 30 , List 2; Tanaka 1997c, d, e). Helonias subsect. Heloniopsis consists of two series: H. ser. Umbellatae and H. ser. Heloniopsis (Part II) .

Ser. Umbellatae :―Several character states of Helonias subsect. Heloniopsis are regarded as more advanced than those of H. subsect. Ypsilandra , as stated in the preceding section. In addition, the umbels, sub-umbels or somewhat compact racemes in species of H. ser. Umbellatae (Part I-7, II-2) also appear to be more advanced character states. The apex of the leaves of H. ser. Umbellatae tends to be slightly more prominently apiculate than in its congeners (apiculus of H. ser. Umbellatae to ca. 2 mm long vs. to ca. 1.3 mm long in other species), probably suggesting that species in this series originated from the same ancestry (List 2-14). Of the three species of H. ser. Umbellatae , H. kawanoi appears to be the most primitive in some floral characters. For instance, the inner filaments of H. kawanoi are close to the base of the ovary ( Fig. 17E View FIGURE 17 ), thereby approaching H. subsect. Ypsilandra , where the inner filaments are adnate to the base of the ovary ( Fig. 14C View FIGURE 14 ; Tanaka 1997a). Further, the tepals of H. kawanoi are shallowly canaliculate proximally and the nectary is least concave. The autumnal flowering is also indicative of an old origin of this species. These facts suggest that H. kawanoi diverged earlier than H. umbellata and H. leucantha from the ancestral lineage of H. ser. Umbellatae ( Fig. 30 View FIGURE 30 , List 2). Helonias kawanoi ( Figs. 16 View FIGURE 16 , 17 View FIGURE 17 ) is a dwarf plant with small comparatively thin leaves, short slender flowering stems, bracteate pedicels, and solitary or a few small autumnal flowers. All these character states are considered to be apomorphies ( Fig. 30 View FIGURE 30 , List 2). Supposedly, H. kawanoi has acquired many of its unique features as a consequence of adaptation to an ecologically unusual location. Being small, the plants can not only utilize microhabitats for their life space, but also colonize on such unusual locations as steep cliffs and banks along streams that may overflow after heavy rainfall. Often growing among mosses, H. kawanoi may be viewed as an element of the moss layer community in the vertical stratification of vegetation. Helonias leucantha ( Figs. 18–20 View FIGURE 18 View FIGURE 19 View FIGURE 20 ) is normally larger than H. umbellata ( Figs. 21 View FIGURE 21 , 22 View FIGURE 22 ), and usually has larger leaves and slightly larger, pendulous, campanulate flowers that bloom from (late December) January through mid March. All of these character states are viewed as apomorphies. Helonias leucantha grows on shady moist rocky slopes near streams. It appears to share an immediate ancestry with H. umbellata ( Fig. 30 View FIGURE 30 , List 2). Helonias umbellata usually grows on moist rocky slopes in shade or semi-shade at low to high elevations of Taiwan. It usually flowers from (late January) February through March, except on high peaks. Helonias umbellata , endemic to Taiwan, is geographically isolated from two Ryukyuan species H. leucantha and H. kawanoi ( Fig. 29 View FIGURE29 ; Tanaka 1997e, 1998a). The Ryukyuan species overlap in distribution on four small islands; Tokunoshima, Okinawa, Ishigaki and Iriomote ( Tanaka 1998a) , but there are no reports of their co-occurrence in the same habitats on these islands, thereby implying that their habitat preferences subtly differ. Moreover, their flowering periods scarcely overlap ( Tanaka 1997d). The three species of H. ser. Umbellatae therefore appear to be isolated from one another, at least by one or more of the following; geography, ecology, sexual reproduction. Flowers of both Helonias umbellata (e.g. Figs. 21A–C View FIGURE 21 , 22A View FIGURE 22 ) and H. kawanoi (e.g. Figs. 16A, B View FIGURE 16 , 17C View FIGURE 17 ) are similar in shape (both obconic), but differ in size. Flowers of H. umbellata and H. leucantha (e.g. Figs. 18B, D View FIGURE 18 , 19A View FIGURE 19 ) differ in shape, degree of fragrance, and usually in size ( Table 8). Their pollinators are also likely to differ to some extent in kind and/or visiting frequency.

Ser. Heloniopsis :― Helonias ser. Heloniopsis comprises Helonias koreana , H. orientalis , and H. breviscapa (see Part II). They share subumbels or umbels of comparatively few flowers ( Figs. 23A–G View FIGURE 23 , 25A–E View FIGURE 25 , 27A–D View FIGURE 27 ), tepals with submarginal wings (lamellae) by which they are connate basally to adjoining tepals and adnate to opposing filaments (e.g. Figs. 24C–E View FIGURE 24 , 26A View FIGURE 26 , 28E View FIGURE 28 ), and sheath-like (tubular) nectaries formed by the adnation (e.g. Figs. 24A–E View FIGURE 24 , 26A–C View FIGURE 26 , 28E View FIGURE 28 ). All of these structures are obviously apomorphies. Helonias ser. Heloniopsis is therefore a monophyletic group ( Fig. 30 View FIGURE 30 ; Tanaka 1997e). In H. ser. Umbellatae , the proximal submarginal portions of the tepal are raised or ridged, but not winged ( Figs. 17D View FIGURE 17 , 19D, E View FIGURE 19 , 20A, B View FIGURE 20 , 22D, E View FIGURE 22 ). Further, unlike H. ser. Heloniopsis , they are not connate between adjoining tepals in H. kawanoi (e.g. Fig. 17D View FIGURE 17 ) and H. leucantha (e.g. Fig. 19E View FIGURE 19 ). In H. umbellata , however, they are occasionally minutely connate basally between adjoining tepals ( Fig. 22D View FIGURE 22 ). In all three species of H. ser. Umbellatae , there is no adnation between the tepals and their opposing filaments ( Figs. 17D, E View FIGURE 17 , 19D, E View FIGURE 19 , 22B, D, E View FIGURE 22 ; Tanaka 1997b, e, 1998a). There is no doubt that the proximal submarginal wings of the tepals of H. ser. Heloniopsis are derived from the ridged or raised submarginal portion of the tepal of H. ser. Umbellatae . As aforementioned, like H. ser. Heloniopsis , H. umbellata occasionally has proximal ridges minutely connate basally to those of adjoining tepals ( Fig. 22D View FIGURE 22 ). The floral structure of H. ser. Heloniopsis therefore appears to have originated from the one as seen in H. umbellata ( Tanaka 1997b, e). Evolution in the reverse direction, i.e. from H. ser. Heloniopsis to H. ser. Umbellatae , is unlikely. It is also improbable that the floral structure exactly the same as that of H. ser. Heloniopsis developed independently in a different lineage as a consequence of parallel evolution. Taking all these improbabilities into account, the cladogram based on some fractions of plastid DNA sequences presented by Fuse & Tamura (2003, as Heloniopsis ) does not appear to reflect the actual phylogeny (cladogenesis). In their analyses, two species of H. ser. Heloniopsis ( Helonias orientalis and H. breviscapa here circumscribed) cluster with three species of H. ser. Umbellatae (here circumscribed), but not with the Korean species ( Heloniopsis koreana and H. tubiflora in their paper) of the same H. series Heloniopsis . Their observations on the floral structure also do not necessarily coincide with mine (see also related remarks under H. koreana in Part II of present paper). Results of my observations on phenotypic characters thus markedly contradict those of their analyses on DNA sequences and floral morphology. As stated above, in my data, Helonias koreana is phylogenetically close to the rest of species of H. ser. Heloniopsis ( Fig. 30 View FIGURE 30 ), and in data published by Fuse & Tamura (2004), Korean species deviate significantly from the remaining (Japanese and Taiwanese) species of H. subsect. Heloniopsis . This incongruence might suggest that the degree of divergence in DNA sequence is not necessarily correlated with phylogenetic distance. Obviously, this issue needs further studies. In Helonias koreana (including Heloniopsis tubiflora ), the adnate portion between a tepal and the opposite filament is nearly as high (long) as or slightly higher (longer) than the connate portion between adjoining tepals ( Figs. 24C, D View FIGURE 24 ), while in H. orientalis and H. breviscapa the adnate portion is consistently higher (longer) than the connate portion ( Figs. 26A View FIGURE 26 , 28E View FIGURE 28 ). In this respect, I regard the character state of H. orientalis and H. breviscapa to be more advanced than in H. koreana . This view is reflected in constructing the phylogram in Fig. 30 View FIGURE 30 (List 2) in which H. orientalis and H. breviscapa are clustered as sister species. Helonias ser. Heloniopsis appears most closely allied to Helonias umbellata of H. ser. Umbellatae ( Tanaka 1997e) , since the tepals of the latter are sometimes basally connate ( Fig. 22D View FIGURE 22 ) like those of H. ser. Heloniopsis . Helonias ser. Heloniopsis and H. umbellata are hence likely to be monophyletic, sharing a common ancestry ( Fig. 30 View FIGURE 30 , List 2). Two species of H. ser. Umbellatae , H. kawanoi and H. leucantha , are somewhat specialized and do not appear to be directly related to H. ser. Heloniopsis . Helonias umbellata , on the contrary, appears less specialized except for some characters (e.g. umbellate inflorescence) and even occurs at high elevations on Taiwan, implying that it can survive a cooler climate. The ancestor of H. umbellata may have evolved the lineage of ser. Heloniopsis while staying somewhere in or around the current range (such as Japan and Korea) of ser. Heloniopsis as a consequence of northward migration. The three species of H. ser. Heloniopsis appear to have individual preferences for climatic conditions. They appear to have phyletically diverged as a result of adaptation to different climatic types covering maritime regions of the Far East. For related remarks see section 5.

4. Process of phyletic diversification

A phylogenetic tree or phylogram (phenogram) based on cladistic analyses of phenotypic characters is shown in Fig. 30 View FIGURE 30 . It is revised from Tanaka (1997e) and was constructed from hypotheses based on observations in the present study.

Attempts to construct phylograms (cladograms) from analyses of DNA sequences are valuable, but we need to be cautious especially in applying the results to classifications of organisms. DNA fractions used for phylogenetic analysis are generally limited in portion and sampling and hence the results may fluctuate to some degree. For instance, if samplings from highly variable species are strongly biased, the resultant phylogram inferred may more or less deviate in topology from the true one. Further, as stated earlier (Part III-3), analyses based only on the degree of similarity between DNA sequences often override paraphyletic lineages recognizable by phenotypical cladism, and hence it seems difficult for them to faithfully trace the true process of phylogenetic diversification. To obtain a phylogram closer to the actual phylogenesis, data from other aspects (particularly phenotypic characters) need to be taken into account.

In this study, in constructing the phylogenetic tree, the taxa are clustered by synapomorphies inferred from comparative surveys of character states (Part I, III-2). The character states regarded as apomorphies are enumerated in List 2 under the respective lineages or clades that correspond to those in Fig. 30 View FIGURE 30 . Synapomorphies for lineage 1 rest on the presumption that Chamaelirium (including Chionographis ) is the sister lineage of Helonias , which is also supported by plastid DNA sequence data ( Fuse & Tamura 2000, Givnish et al. 2016, Kim et al. 2016).

LIST 2. Character states regarded as apomorphies, which are presumed to have been acquired by each lineage in Fig. 30 View FIGURE 30 , are enumerated below under the respective lineage numbers corresponding to those in Fig. 30. View FIGURE 30

1) Immediate (closest) ancestral lineage common to lineages 2 and 3 (vs. lineage of Chamaelirium ).

1. Anthers become distinctly unilocular, losing adaxial septum (furrow) between thecae.

2. Apex of ovary becomes emarginate-depressed.

3. Carpels of ripe fruit become abaxially protruding subdistally.

4. Three styles become connate basally to form very short column.

2) Lineage of Helonias bullata .

1. Tepals become pink.

3) Immediate ancestral lineage common to lineages 4 and 5.

1. Plants become somewhat smaller, as reflected in leaf size and scape height.

2. Pedicels become slightly shorter.

3. Stamens become slightly shorter.

4. Anthers become latrorse (except apical part which is antrorsely dehiscent).

5. Abaxial septum (furrow) between thecae as in H. bullata becomes lost (i.e. anthers become perfectly unilocular with no septum between thecae).

6. Style (columnar part) becomes slightly longer.

7. Ovules per locule become more numerous.

8. Seeds (with testa) become more slender (linear-fusiform).

4) Lineage of Helonias yunnanensis var. yunnanensis .

1. Whole body (habit) often becomes reduced in size, as reflected in leaf size, scape height, and flower size.

2. Stamens usually become further shortened.

5) Immediate ancestral lineage common to lineages 6 and 7.

1. Style (columnar part) becomes slightly longer.

6) Lineage of Helonias jinpingensis .

1. Pedicels become relatively slightly longer.

2. Pedicels become bracteate at base.

3. Flowering in late summer to early autumn.

7) Immediate ancestral lineage common to lineages 8, 9 and 10.

1. Style becomes relatively slightly longer.

2. Stigmatic lobing becomes relatively slightly reduced.

8) Lineage of Helonias yunnanensi s var. mesostyla .

1. Apomorphy not yet specified.

9) Lineage of Helonias alpina .

1. Tepals become slightly longer.

2. Tepals become narrowly oblong(-oblanceolate).

3. Style becomes slightly longer.

4. Stigmatic lobing slightly reduced.

5. Filaments become longer.

10) Immediate ancestral lineage common to lineages 11 and 12.

1. Stigma becomes only slightly trilobed.

11) Lineage of Helonias parviflora .

1. Flowers become slightly smaller, as reflected in tepals.

12) Immediate ancestral lineage common to lineages 13 and 14.

1. Pedicels become longer.

2. Flowers become protogynous.

3. Flowers tend to become slightly larger.

4. Stigma and stamens usually become exserted beyond tepals.

5. Stigma becomes discoid.

13) Lineage of Helonias thibetica .

1. Stigma becomes slightly smaller.

2. Stigma usually becomes (sub)capitate.

14) Immediate ancestral lineage common to lineages 15 and 16.

1. Apiculus at leaf apex becomes (relatively) slightly more prominent.

2. Racemes tend to become slightly more compact at anthesis.

3. Inner filaments become liberated from base of ovary.

4. Anthers become adaxially affixed to filament (i.e. adaxially basifixed).

5. Anthers become extrorse.

6. Anthers become virtually bilocular with traces of apical confluence between thecae.

15) Lineage of Helonias kawanoi .

1. Whole body (habit) becomes significantly smaller, as reflected in rhizome, leaves, height of scape, cauline scaly leaves, flowers. 2. Leaves become somewhat thinner (in texture).

3. Main parallel veins of leaves become slightly raised adaxially.

4. Flowering in late summer to autumn (or dormancy in development of flowering stem becomes almost lost).

5. Pedicels become bracteate.

6. Flowers (per scape) become fewer.

7. Inflorescence becomes usually (sub)umbellate.

8. Ovary becomes sessile.

9. Seeds (incl. testa) become smaller (shorter).

16) Immediate ancestral lineage common to lineages 17 and 18.

1. Tepals become more distinctly canaliculate toward base.

2. Proximal submarginal portions of tepal become more prominently raised.

3. Nectary at base of tepal becomes more prominently concave.

4. Filaments tend to become slightly closer basally to opposing tepal (rather than to base of ovary).

17) Lineage of Helonias leucantha .

1. Entire body (habit) becomes slightly larger, as reflected in leaves.

2. Flowering in (late December) January through early March.

3. Flowers become campanulate with tepals recurved distally.

4. Flowers become less fragrant (nearly odorless).

5. Tepals usually become narrowly oblong-elliptic.

6. Ovary sometimes becomes (nearly) sessile (due to abbreviation of short gynophore or gynophore-like floral axis).

18) Immediate ancestral lineage common to lineages 20 and 21.

1. Racemes become slightly more compact.

2. Proximal submarginal ridges of tepal become more closely juxtaposed basally with those of adjoining tepals.

19) Lineage of Helonias umbellata .

1. Inflorescence usually becomes (sub)umbellate.

2. Proximal submarginal ridges of tepal sometimes become minutely connate basally to those of adjoining tepals.

3. Ovary sometimes becomes (almost) sessile (due to abbreviation of short gynophore or gynophore-like floral axis).

20) Immediate ancestral lineage common to lineages 22 and 23.

1.?Apiculus at leaf apex becomes moderate.

2. Flowers (per scape) tend to become fewer.

3. Flowers tend to become less fragrant or odorless.

4. Proximal submarginal ridges of tepal become lamellate (winged) and directed inward.

5. Proximal submarginal wings (lamellae) of tepal become connate basally to those of adjoining tepals.

6. Proximal submarginal wings of tepal become adnate proximally to opposing filament, forming sheathing nectary.

7. Anthers often become affixed (to filament) slightly above base of connective (i.e. (sub)dorsi-basifixed).

8. Anthers tend to become slightly longer.

9. Traces of apical confluence between thecae tend to become slightly less distinct.

10. Ovary becomes perfectly sessile.

21) Lineage of Helonias koreana .

1. Leaf blades become slightly thinner (in texture).

2. Leaf blades become nearly dull.

3. Filaments become slightly dilated toward base.

4. Filaments become more complanate and often subancipital (two-edged or narrowly winged) toward base.

22) Immediate ancestral lineage common to lineages 23 and 24

1. Basal adnation between proximal submarginal wings of tepal and opposing filament becomes more advanced (viz. adnate portion becomes longer).

23) Lineage of Helonias orientalis .

1. Leaves become somewhat firmer (thicker) in texture.

2. Proximal submarginal wings of tepal become relatively longer.

3. Basal adnation between proximal submarginal wings of tepal and opposing filament becomes slightly more advanced (viz. adnate portion becomes longer).

24) Lineage of Helonias breviscapa .

1. Basal abaxial gibbosity of tepal often becomes less conspicuous.

2. Nectary at adaxial base of tepal becomes relatively higher in position (surface of nectary becomes nearly as high as base of ovary or top of receptacle).

3. Traces of apical confluence between thecae often become indistinct.

5. Historical background of the current geographic ranges

The distributions of the sections of Helonias are remarkably disjunct. Helonias bullata is restricted to eastern North America ( Figs. 6 View FIGURE 6 , 31 View FIGURE 31 ), while H. sect. Heloniopsis is restricted to eastern Asia ( Figs. 31 View FIGURE 31 , 32 View FIGURE 32 ). The main factor behind this large scale disjunction may have been drastic climatic change. According to paleobotanical surveys, the so-called Arcto-Tertiary Geoflora, which comprises many species of deciduous broad-leaved trees, once extensively covered northern circumpolar regions ( Chaney 1947, Tanai 1967, 1972, 1988, 1992, Uemura 1995, Ogasawara & Uemura 2006). The flora began to migrate southwards in response to lowering temperatures toward the end of the Eocene and early Oligocene. Ancestral plants of Helonias , which are presumed to have been members of the Geoflora, may also have migrated southward and in the process may have separated into two groups, one reaching Asia and the other North America ( Fig. 31 View FIGURE 31 ) ( Tanaka 1997e). Helonias bullata ( Figs. 4 View FIGURE 4 , 5 View FIGURE 5 ) of North America appears to be the most primitive, since its three styles tend to be largely free ( Fig. 5D View FIGURE 5 , Part III-3). This remarkable disjunction in the range of Helonias appears to have originated at an early stage of diversification ( Fig. 30 View FIGURE 30 ; Tanaka 1997e) before the existence of many of the current species of Helonias (most species of H. sect. Heloniopsis such as H. parviflora , H. thibetica , H. umbellata , and H. orientalis ).

Of the Asian members of Helonias (H. sect. Heloniopsis ), H. yunnanensis var. yunnanensis ( Figs. 7 View FIGURE 7 , 15-Yy) appears to be the most primitive ( Fig. 30 View FIGURE 30 ), since it has the shortest style ( Fig. 7B View FIGURE 7 , Part III-3; Tanaka 1997c, e). Presumably, the ancestral plants of this Helonias yunnanensis migrated from area at high latitudes (encircled by broken line in Fig. 31 View FIGURE 31 ) to somewhere in or around the Tibetan Plateau and the Himalaya of today ( Figs. 31 View FIGURE 31 , 32 View FIGURE 32 ). After having settled there, they probably underwent gradual uplift caused by the subduction of the so-called Indian Plate (Indo-Australian Plate) under the Eurasian Plate. The ancestors of H. yunnanensis var. yunnanensis must have gradually adapted to alpine or subalpine environments created by the orogenesis of that region. This variety is now confined to the highlands between ca. 2700 and 4300 m (Fig. 15-Yy). Plants in such harsh alpine situations have inevitably become smaller and mostly correspond to Ypsilandra yunnanensis var. micrantha (or var. himalaica ). Helonias alpina ( Figs. 10 View FIGURE 10 , 15-A; elev. 3962–4267 m) may also have differentiated in the highlands. Some ancestors of H. yunnanensis var. yunnanensis in the eastern part of the plateau may have evolved species such as H. jinpingensis now on the border between southern Yunnan and northern Vietnam (Fig. 15-J) and H. parviflora in Guizhou ( China) (Fig. 15-P). The ancestor of H. parviflora is likely to have branched off the lineage of H. thibetica which is now widespread over the eastern stretch or periphery of the Himalaya and Tibetan Plateau (Figs. 15-T, 30, 32). Ancestral plants of H. ser. Umbellatae (subsect. Heloniopsis) , whose descendants are now in Taiwan and the Nansei Islands (including the Ryukyus) in Japan ( Figs. 29 View FIGURE29 -Ka, -L, -U, 31, 32), may have originated from the ancestor of H. thibetica (Figs. 15-T, 30) which was once in the Far East maritime regions (including today’s southern China, Taiwan, the Nansei Islands, and lands subsided under the Taiwan Strait and East China Sea).

Thus, the geographic range of ancestral Asian Helonias after settlement in or around the present Tibetan Plateau and the Himalayan regions is presumed to have extended eastward over time and through evolutionary progression ( Figs. 30 View FIGURE 30 , 32-B, -C View FIGURE 32 ; Tanaka 1997e). Li (1944) recognized 14 phytogeographic regions in China. If his regional divisions are adopted, a presumable approximate course of range extension of ancestral Asian Helonias (H. subsect. Ypsilandra and probably H. ser. Umbellatae in part) would be:―Tibetan Highland Region (14) → Sino-Himalayan Region (4) → Southwestern China Plateau Region (3) → Southern China Maritime Region (1) (number in parentheses assigned by Li, 1944).

During warmer geological periods the ancestors of Helonias ser. Umbellatae (or of H. umbellata ; Figs. 29 View FIGURE29 , 31-3 View FIGURE 31 , 32-3 View FIGURE 32 ) are supposed to have extended their range to somewhere in or around today’s Japan and Korea ( Fig. 32-D View FIGURE 32 ). After the climate cooled again, some may have migrated south, while others may have remained to form the ancestral lineage of H. ser. Heloniopsis ( Figs. 29 View FIGURE29 , 30 View FIGURE 30 , 31-4 View FIGURE 31 , 32-4 View FIGURE 32 ).

The approximate course of migration or extension of geographic range of Helonias traced from this study is outlined in Fig. 33 View FIGURE 33 . Infrageneric taxa, such as H. subsect. Ypsilandra and H. ser. Umbellatae , have their own geographical ranges disjunct from one another ( Figs. 29 View FIGURE29 , 31 View FIGURE 31 , 32 View FIGURE 32 ), and the largest scale disjunction between the two sections ( Helonias and Heloniopsis ) appears to reflect not only the most drastic climatic change in the evolutionary history of Helonias but also the deepest diversification in the genus.

Fuse & Tamura (2004) suggested that Heloniopsis as a remnant of the Arcto-Tertiary flora has migrated from the north to the present range (via Sakhalin in case of species distributed in Japan and Taiwan). However, no remarks were made on the migration route of Ypsilandra ( Helonias subsect. Ypsilandra ) and the geological age and palaeogeological states of eastern Asia when the migration had occurred. It is hard to accept their view, considering the process of phylogenetic progression in Helonias s.lat. as traced from the present study ( Figs. 30–33 View FIGURE 30 View FIGURE 31 View FIGURE 32 View FIGURE 33 , List 2).

The three species of Helonias ser. Heloniopsis , H. koreana , H. orientalis and H. breviscapa , appear to be adapted to different climates. Helonias orientalis ( Fig. 29-O View FIGURE29 ) in Japan and southern Sakhalin, Russia, tends to be more abundant in snowy regions, such as along the coast of Honshû facing the Sea of Japan (see Japan Meteorological Agency 1971, 1972, for climatological data). Helonias breviscapa ( Fig. 29-B View FIGURE29 ), which is confined to southwestern Japan mainly facing the Pacific, appears to prefer a somewhat warmer climate with less snowfall in winter ( Tanaka 1998a). Helonias koreana ( Fig. 29 View FIGURE29 -Ko), endemic to the Korean Peninsula, appears to be adapted to a bitter cold, drier, and less snowy winter (see Central Meteorological Observatory, Tokyo 1929).

Plants, such as Helonias orientalis and H. koreana , that occur in very cold regions may be protected from coldness and dryness by being covered by snow. The three different types of climate appear to be closely related to the geographic features of the respective regions. It is generally accepted that the following three factors are prerequisite for creating such different climatic types: a) high atmospheric pressure develops over the Siberian region in winter, resulting in strong seasonal wind toward Japan; b) the Sea of Japan between the eastern Asian mainland and the Japanese archipelago provides moisture to the dry air mass blowing from the Siberian region; c) the high mountains in Honshû, Japan obstruct the northwesterly wind, creating two markedly contrasting climatic types in winter ( Uemura 2011).

The three species of Helonias ser. Heloniopsis may be closely related to such geographic and climatic features. These features appear to have become remarkable especially from the mid-Pleistocene onward (e.g. maps 30- 27–30 in Minato et al. 1965 for the process of palaeogeographical changes in eastern Asia). If the timing of establishment of these features (or the above three factors) becomes more precise, the approximate time of origin of the three species may also become clearer.

Because of fluctuating temperatures in the Neogene period of the Tertiary and in the Quaternary ( Minato et al. 1965, Tanai 1972, 1992, Uemura 1995, Ogasawara & Uemura 2006), some north-southward migration of vegetation, including Helonias , must have occurred. The occurrence of H. orientalis in southern Sakhalin, Russia ( Miyabe & Kudo 1932, Sugawara 1937, 1939), which is at the northernmost limit of its range ( Fig. 29 View FIGURE29 ), may have resulted from such migrations during the Quaternary. Despite such migrations, the relative geographic ranges of the species of Helonias may not have changed extensively.

Helonias are small hemicryptophytes with basal rosulate leaves spreading on the ground. They prefer moist or wet situations, such as swamps, meadows or steep rocky slopes along streams in forests, that are not necessarily so heavily covered by larger plants. In general, these habitats are more exposed than closed, dense forest floors, and are hence more likely to soon dry if water becomes scarce. Ample precipitation may therefore be required to maintain stable populations.

The main factors for the eastward migration in Asia of Helonias sect. Heloniopsis ( Fig. 32-B–D View FIGURE 32 ) may have been ample precipitation and moderate temperatures.Asian Helonias appears to be more abundant and diversified in regions blessed with such climatic factors, e.g., maritime regions of the Far East including Japan and Taiwan, and the Himalaya and Tibetan Plateau ( Fig. 32 View FIGURE 32 , Table 10; Yan 2002 for climatological data for China). Eastern North America, where H. bullata occurs (e.g. Figs. 6 View FIGURE 6 , 31 View FIGURE 31 ), resembles eastern Asia in climate, especially in precipitation and temperature (e.g. plates 20, 24, 26 in Good 1974).

Kingdom

Plantae

Phylum

Tracheophyta

Class

Liliopsida

Order

Liliales

Family

Melanthiaceae

Loc

Helonias

Tanaka, Noriyuki 2019
2019
Loc

Heloniopsis japonica (Thunb.) Maxim. var. albiflora

Hara, H. 1947: )
Honda, M. 1938: )
1938
Loc

Heloniopsis japonica (Thunb.) Maxim. var. tesselata Nakai ex

Honda, M. 1938: )
1938
Loc

Heloniopsis japonica Maxim. var. yakusimensis

Masamune, G. 1969: )
Wang F. T. & Tang, T. 1949: )
Hara, H. 1947: )
Koidzumi, G. 1939: )
Honda, M. 1938: )
Masamune, G. 1934: )
1934
Loc

Heloniopsis japonica (Thunb.) Maxim. var. flavida

Wang F. T. & Tang, T. 1949: )
Hara, H. 1947: )
Nakai, T. 1933: )
Nakai, T. 1933: )
1933
Loc

Sugerokia nipponica

Hiyama, K. 1939: 79
Nemoto, K. 1936: )
Ohwi, J. 1930: )
1930
Loc

Sugerokia japonica

Koidzumi, G. 1930: 95
1930
Loc

Heloniopsis japonica

Honda, M. 1938: 1677
Matsumura, J. 1905: 197
1905
Loc

Sugerokia japonica

Miquel, F. A. W. 1870: )
1870
Loc

Heloniopsis breviscapa

Kitamura, S. & Murata, G. 1966: )
Wang F. T. & Tang, T. 1949: )
Honda, M. 1938: )
Maximowicz, C. J. 1867: )
1867
Darwin Core Archive (for parent article) View in SIBiLS Plain XML RDF