Diagnosis (revised from Carlgren, 1918; revised parts are shown in italics). Actinernidae with cylindrical body, which is distally drawn out into eight distinct lobes, all eight of same size or four larger and four smaller alternating. Column without papillae. No sphincter. Tentacles in at least two cycles without distinct mesogloeal thickenings, all same length or largest at apices of the lobes, numerous. Longitudinal muscles of tentacles ectodermal, radial muscles of oral disc chiefly ectodermal, strong. Two broad siphonoglyphs on actinopharynx. 36 macrocnemes cyclic in arrangement, beyond them weak mesenteries of unequal size in upper part of the body. Retractors weak, parietal muscles weak or rather well developed.

Type species. Synactinernus flavus Carlgren, 1918.

Remarks. The genus Synactinernus was established in the family Actinernidae by Carlgren (1918) monotypically for Synactinernus flavus Carlgren, 1918. Recently, Fautin and den Hartog (2003) argued that Synactinernus flavus, the only species of Synactinernus, cannot be distinguished from Isactinernus quadrilobatus and synonymized Synactinernus with Isactinernus. However, Synactinernus was listed as a valid genus in later studies (Fautin, 2013, 2016). Therefore, the actual status of S. flavus and the validity of Synactinernus remained unclear. Additional specimens are needed to resolve this taxonomical problem, but no new specimen of Synactinernus had been found for a century after the original description of Carlgren (1918). The present study is the first revision of the genus Synactinernus, as we obtained a second specimen of S. flavus and the second species of this genus.

Our morphological analyses of the newly collected specimens of two Synactinernus species and I. quadrilobatus and of the type specimens of I. quadrilobatus showed that tentacles of S. flavus and S. churaumi sp. nov. are simple in shape, apparently lacking thickening of the aboral side (Figs. 2C, G, 3C, 6B, 7C), while those of I. quadrilobatus, both in our collection and type specimens, are apparently thickened at the base of the aboral side (Fig. 6C). This clear difference suggests that Synactinernus species can be distinguished from I. quadrilobatus by their tentacles. In addition, Synactinernus species had a fixed number (36) of macrocnemes (Figs. 3A, 7A), and this prominent feature in the mesentery structure should be a diagnostic characteristic of the genus, while I. quadrilobatus has many indistinguishable mesenteries. In contrast, the alternate arrangement of larger and smaller oral lobes of S. flavus, considered as an important morphological feature of this monotypic genus (Carlgren, 1918, 1949), turned not to be a diagnostic characteristic, because the new species S. churaumi has eight lobes of the same size.

The result of the current molecular phylogenetic analyses that Synactinernus formed a well-supported independent clade demonstrates the morphological deviation from the other genera of Actinernidae including the former belonging genus, Isactinernus.

Synactinernus is endemic to Japan and is distributed around a depth of 300 m in the East China Sea and the Pacific Ocean.

Synactinernus flavus Carlgren, 1918, p. 31, “Gote Island, Kin Shin 137 m” [sic., most likely “Goto Islands, Kyushu” in modern orthography].

Material examined. NSMT-Co 1660: specimen dissected, tissues embedded in paraffin, histological sections (16 slides) prepared, nematocysts prepared (four slides); originally collected on 19 November 2012, the Goto-nada Sea off southeastern Goto Islands, Nagasaki Prefecture (32°23.00′N, 129°03.00′E), at a depth of 304–504 m, using a biological dredge of R/V Nagasaki-Maru (Cruise No. 365), collected by Masanori Nonaka, and fixed by Takuma Fujii and Hiroko Takaoka from the tank at Okinawa Churaumi Aquarium on January 25, 2013. CMNH-ZG 09732: whole specimen attached to a rock; originally collected in January 2018, at Uragami, Wakayama Prefecture, at 100–130 m, by Isao Hirabayashi, and fixed by Takuma Fujii and Kensuke Yanagi from the tank at the Kushimoto Marine Park on May 22, 2018. UUZM 232 holotype: dissected specimen, 17 May 1914, off Goto Islands (33°41′N, 128°50′E [collection information listed on the label]), depth 110 m, collected by Sixten Bock (Fig. 1).

Description. External anatomy. Body cylindrical (Fig. 2A), up to ca. 5–8 cm in length and ca. 2–4 cm in width when alive, and 2–4 cm in length and 1–2 cm in width on specimen. Column surface comparatively smooth, without tenaculi, with numerous discontinuous, wavy wrinkles running in a transverse direction, pale surface with pale orange ectoderm layer. Ectoderm of column without nematocyst batteries, nematocysts very sparsely distributed containing numerous spirocysts. Upper part of column widely expanded and thrown into eight lobes, four larger and four smaller ones (Fig. 2B, C). Upper margin of the lobes fluorescent yellow in color (Fig. 2B). Tentacles simple, all marginal, ca. 2–5 mm in length, no thickenings including their aboral base, pale white in color, and ca. 120–200 in number on oral disk; inner and outer ones alternatingly bared. The tip of tentacles pointed. Pedal disk in aboral end, semitransparent, and mesenterial insertion visible. Oral disk diameter ca. 3–6 cm, same color as tentacles, mouth at center of oral disk, apparently swelled, lip-like.

Internal anatomy. 36 (18 pairs of) macrocnemes (Figs. 3A, 4) on actinopharynx; twelve, including four directives, in the first cycle; eight in the second cycle; 16 in the third cycle. Macrocnemes in the second cycle born in the endocoel of the first cycle mesenteries, an arrangement obeying the rule of mesenterial arrangement of Actinernidae (Fig. 4). Pairs of microcnemes, in the fourth cycle, observed in transverse section (Fig. 3A, arrowhead). Judged by the numbers of tentacles, mesenteries of Synactinernus flavus in the first–sixth cycles. All mesenteries perfect near the mouth, and each tentacle between either exo- or endocoelic. Tentacular longitudinal muscle ectodermal (Fig. 3C); tentacular circular muscle too weak to observe in histological sections (Fig. 3D). Retractor muscles comparatively weak and restricted actinopharynx or filament side (Fig. 3B, E). Muscle processes short, mostly simple or slightly branched, around 30–50 in each muscle pennon. Parietal muscles of macrocnemes weak, distinct with approximately 10 muscle fibers in upper part near actinopharynx (Fig. 3B), but indistinct in lower part near filaments (Fig. 3E). Mesoglea thickest in body wall and actinopharynx (Fig. 3A), reaching 2–3 mm in thickness, and far thicker than the ectoderm and endoderm. However, mesoglea thinner in mesenteries (Fig. 3E), and thinnest in tentacles (Fig. 3C, D). Actinopharynx, with siphonoglyphs on dorsal and ventral sides (Figs. 3A, 4), always connected to actinopharynx, and with 10 longitudinal grooves as deep as siphonoglyphs. Sphincter muscle absent. On the aboral end, basilar muscle absent. Dioecious, matured eggs in gonads in NSMT-1660 (Fig. 3F). Mesenteries in the first and second cycles only fertile (Fig. 3E).

Cnidom. Basitrichs, spirocysts, and microbasic p-mastigophores. See Table 2 and Fig. 5 for size and distribution.

Derivation of new Japanese name. “Kuroba” means the plant clover. The oral disk with four large and foursmall lobes resembles the four-leaf clover, which is considered a symbol of good luck.

Taxonomic remarks. We provide a detailed description of S. flavus based on the holotype and our new specimens in this study. We presented cnidom data of every part of the body (Table 2) and the figures of external features (Fig. 2), which were lacking in Carlgren (1918).

Almost all morphological features of the present specimen (NSMT-Co 1660) corresponded to Carlgren (1918) and the holotype (UUZM 232): cylindrical body (Figs. 2A, 6A); ca. 120–200 tentacles with no thickening even on base (Figs. 2B, C, 3D, 6B); about 10 longitudinal grooves on actinopharynx (Fig. 3A); mesoglea on the actinopharynx are thick (Fig. 3B); ectodermal but weak tentacular muscle (Fig. 3D); numerous spirocysts in column (Fig. 5F, Table 2). The size of spirocysts, 20.9–47.1 µm in length and 4.1–7.6 µm in width in our specimens (Table 2), is similar to that of the syntype, 26–38 µm in length and 3–5 µm in width (Carlgren, 1918).

The number of macrocnemes is fixed to 36 in S. flavus (Fig. 3A), while over 100 in I. quadrilobatus (Fig. 6D; there are the same number of mesenteries next to the actinopharynx), but it is not a strong evidence to divide the species and genus because Fautin and den Hartog (2003) reported a range (18–40) in the number of mesenterial pairs among 81 specimens when the two species were synonymized. However, we confirmed that there are several differences between the two species other than the above-mentioned differences between the two genera; whether as mentioned in remarks of genus Synactinernus, thickening on the base of tentacle is only observed on I. quadrilobatus. At the oral disc, this species has apparently four large and four small lobes (Figs. 2B, C, 6A), but I. quadrilobatus has only four lobes; S. flavus fertilizes only the first and second cycle mesenteries, total 20 in number (Fig. 3E), while I. quadrilobatus has many fertile mesenteries (Fig. 6D). In addition, the size of I. quadrilobatus is approximately twice to three times that of S. flavus (Fig. 6A, C). These several morphological differences of both species are sufficient to discuss that they are independent groups, and our molecular phylogeny analysis strongly supports the division into two species based on the several morphological differences mentioned above; S. flavus and I. quadrilobatus are not closely related on the phylogenetic tree and the difference between them is considered to be at the genus level (Fig. 8; see Discussion for details). In conclusion, S. flavus is not a synonym of I. quadrilobatus and is different at the genus level both in morphology and molecular phylogeny. Therefore, the morphological features that Fautin and den Hartog (2003)describedasintraspecificvariationsinI.quadrilobatus canalsobeusedtodistinguishS.flavusfromI.quadrilobatus.

In the present study, we redescribed the species using our newly collected specimens with the addition of important features like cnidom of every part of the body (Fig. 5), which were lacking in Carlgren (1918), and also photographic information (Figs. 2, 3).

The locality of the holotype of S. flavus, UUZM 232, is described on the label as follows: “Goto Islands 28 miles N. 1/2 Ost fran Shirase fyr 128 50 O.L., 33 41 N.Br.” The locality, “Shirase”, would mean “Shiroze”, an isolated, uninhabited island north of the Goto Islands (33°11′00″N, 128°48′14″E). In Japanese, the Kanji-character of “Shiroze” is sometimes pronounced as “Shirase”. Swedish “fyr” means a lighthouse. The lighthouse on Shiroze was built in December 1904, and was obviously present in 1914, when the holotype of S. flavus was collected. The point, 1/2 sea mile east off Shiroze, is approximately 40–45 km north from Fukue Island (the main island of Goto Islands). These distances and directions completely match “Goto Islands 28 miles N” on the label. The depth, 110 m written on the label, is also a match. Considering the above information, the exact type locality should be here. The latitude and longitude written on the label is far away from “Shiroze”, so it may be mistaken. Our specimens would be the first topotype specimens after the collected type specimen.

Ecologicalremarks. ItwasobservedthatSynactinernus flavus reproduced several times by transverse fission in the tank at Okinawa Churaumi Aquarium. When the bodies were mechanically or thermally stimulated, for example when their basal disc was detached from the substrate, it cracked on the surface of the column, and transverse fission of the body occurred soon thereafter (Fig. 2D). Both halves became independent full-grown individuals. This is the first observation of transverse fission in Actinernoidea, while fission from the basal disc has been observed in Halcuriidae of Actinernoidea (Uchida, 2004).

Material examined. Holotype: NSMT-Co 1661: specimen dissected, tissues embedded in paraffin, histological sections (20 slides) prepared, nematocysts prepared (four slides); originally collected on February 22, 2004, off northwestern Ishigaki Island, Okinawa Prefecture, at a depth of 281–312 m, using LEO, ROV of Okinawa Churaumi Aquarium, by Takuo Higashiji, and fixed by Takuma Fujii and Hiroko Takaoka from the aquarium tank on 25 January 2013. Paratype: CMNH-ZG 09733: originally collected on 5 September 2018, off Onna Village, Okinawa Island, Okinawa Prefecture (26°34.22′N, 127°47.96′E), at a depth of 320 m, using ROV of the Okinawa Churaumi Aquarium by Takuo Higashiji, and fixed by Takato Izumi from the aquarium tank on 10 September 2018 (Fig. 1).

Description. External anatomy. Body cylindrical (Fig. 2E), up to 20–25 cm in length and 15 cm in width when alive. Regarding specimens, 12 cm in length and 10 cm in width on holotype, and 13 cm in length and 9 cm in width on paratype. Column surface smooth, without tenaculi, pale surface with orange or yellow ectoderm layer. Ectoderm of column without nematocyst batteries and nematocysts very sparsely distributed but contain numerous spirocysts. Upper part of column widely expanded and thrown into eight lobes of the same size (Fig. 2F, G). Upper margin of the lobes same as column in color (Fig. 2E). Tentacles simple, all marginal, 5–20 mm in length, no thickenings including their aboral base, pale white in color, and ca. 350–500 in number on oral disk; inner and outer ones alternatingly bared (Fig. 2F). The tip of tentacles pointed. Pedal disk in aboral end, opaque and mesenterial insertion invisible. Oral disk diameter ca. 12–20 cm in living specimen, same color as tentacles or pale orange, Mouth at center of oral disk, exceedingly swollen, lip-like, pale white to yellow (Fig. 2F).

Internal anatomy. 36 (18 pairs of) macrocnemes (Figs. 4, 7A) on actinopharynx; 12, including four directives, in the first cycle; eight in the second cycle; and 16 in the third cycle. Macrocnemes in the second cycle born in the endocoel of the first cycle mesenteries; their arrangement obeying the rule of mesenterial arrangement of Actinernidae (Fig. 4). Mesenteries in younger than the fourth cycle microcnemes, and judged by the numbers of tentacles, mesenteries of Synactinernus churaumi sp. nov. in the first–seventh cycles. All mesenteries perfect near the mouth, and each tentacle between either exo- or endocoelic. Tentacular longitudinal muscle ectodermal (Fig. 7C, D); tentacular circular muscle too weak to observe in histological section; retractor muscles comparatively weak and restricted on the middle of mesenteries (Fig. 7B); or diffused and integrated to parietal muscle (Fig. 7E). Muscle processes short, mostly simple or slightly branched, around 60–90 in each muscle pennon. Parietal muscles of macrocnemes weak, distinct with approximately 10 muscle fibers in upper part near actinopharynx (Fig. 7B) but integrated to retractor muscle near filaments (Fig. 7E). Mesoglea thickest in body wall and actinopharynx (Fig. 7A, B), reaching to 8–10 mm in thickness, and far thicker than ectoderm and endoderm. However, mesoglea thinner in mesenteries (Fig. 7B, E) and thinnest in tentacle (Fig. 7C, D). Actinopharynx, with siphonoglyphs on dorsal and ventral sides (Figs. 7A, 4), always connected to actinopharynx, and with 10 longitudinal grooves as deep as siphonoglyphs. Sphincter muscle absent. On the aboral end, basilar muscle absent. Dioecious, matured testes in NSMT-1661; a variety of stages of maturation in the same testis (Fig. 7F). Mesenteries in the first and second cycles only fertile. The release of eggs and sperm by S. churaumi was observed in the tank at the Okinawa Churaumi Aquarium.

Cnidom. Basitrichs, spirocysts, and microbasic p-mastigophores. See Table 2 and Fig. 5 for size and distribution

Etymology. The species epithet “churaumi” is named after Okinawa Churaumi Aquarium, an aquarium in Okinawa, Japan, which provided us the specimens of this species. “Chura” means beautiful in Okinawan language, and “umi” means sea in Japanese.

Derivation of new Japanese name. Same as the species epithet.

Taxonomic remarks. Synactinernus churaumi sp. nov. is the second species of this genus; this species is clearly distinguished from S. flavus. Comparing the two species in adult individuals which have matured gametes (Figs. 3E, 5E), S. churaumi is approximately 3–5 times larger than S. flavus in body length (S. churaumi is one of the largest species of Actinernidae), and S. churaumi has over 350 tentacles while S. flavus has around half the number. The margin of the oral disk of S. churaumi develops into eight same sized lobes, while that of S. flavus has four larger and four smaller lobes. The molecular phylogenetic tree (Fig. 8) indicated that these two species formed apparently separate clades. These evidences indicate that S. churaumi is not conspecific with S. flavus but is an independent species. In addition, there is possibly a difference in the presence of asexual fission (see Ecological remarks).

The eight-lobed oral disk of S. churaumi sp. nov. resembles that of Actinernus Verrill, 1879. However, S. churaumi does not correspond to the diagnosis of Actinernus of Actinernidae Stephenson, 1922. Actinernus species have mesogleal thickening at the tentacle base, and we confirmed this in both the holotype of Actinernus robustus Hertwig, 1882, the only Actinernus species from the North Pacific, (Natural History Museum, London, BM 89-11-25-30; Fig. 6E, F) and our specimens (CMNH-ZG 09735). However, tentacles of S. churaumi were not thickened at any part. Moreover, mesenterial arrangement of Actinernus is bilateral, and the fourth cycle mesenteries are born in the endocoel of the third cycle (Carlgren, 1918), and Uchida (2007) confirmed this for A. robustus. The mesenterial arrangement of S. churaumi is completely different from that of Actinernus (Fig. 4).

Ecological remarks. Synactinernus churaumi sp. nov. is distributed around depths of 320–350 m off Okinawa Island. This species forms small aggregations on the top of hill-like submarine topography as recorded by the ROV (Fig. 2H). Incidentally, no asexual fission of S. churaumi in the tank at the Okinawa Churaumi Aquarium was observed for 15 years, while S. flavus has self-divided several times over seven years (Fig. 2D).

Phylogenetic analyses

Regarding the phylogenetic tree of 18S rDNA (1623 bp), all Synactinernus specimens significantly formed a clade (ML boot strap value = 79; BI posterior probability = 0.99), which was separated from the other genera of Actinernidae (Fig. 8). The Synactinernus clade was separated into two clades at high support (ML bootstrap value = 100; BI posterior probability = 1.0 for both clades), and they corresponded to respectively S. flavus and S. churaumi specimens. All three sequences of S. flavus were completely identical, and S. churaumi showed only a 2-base-pair difference between the two sequences (this small difference may be intraspecific variation because nuclear 18S marker has higher base pair substitution late than mitochondrial ones in Actiniaria [Daly et al., 2010]). In addition, the sequence of I. quadrilobatus deposited in GenBank (KJ483024) was also identical with S. flavus. On the other hand, our specimen of I. quadrilobatus (NSMT-Co 1662) was nested with S. elegans (KJ483021), the most basal clade in Actinernidae, and completely separated from “I. quadrilobatus” in GenBank. The relationship between Actinernus and Synactinernus was unsolved with unreliable node (ML bootstrap value = 43; BI posterior probability = 0.64).

Regarding 16S (581 bp), there were no differences in the sequences between S. flavus and S. churaumi at all. It has been shown that 16S is less useful than 18S in the phylogenetic analyses of sea anemone for phylogeny in species level in previous research (Daly et al., 2010), and the results of this research support that assertion.