Revision of the genus Ceriantheomorphe (Cnidaria, Anthozoa, Ceriantharia) with description of a new species from the Gulf of Mexico and northwestern Atlantic

Abstract The present study presents a revision of the genus Ceriantheomorphe Carlgren, 1931, including redescriptions of the two presently recognized species, Ceriantheomorphe ambonensis (Kwietniewski, 1898) and Ceriantheomorphe brasiliensis (Mello-Leitão, 1919), comb. nov., and a description of the new species Ceriantheomorphe adelitasp. nov.


Introduction
Ceriantharia is a subclass within the cnidarian class Anthozoa, consisting of species commonly known as tube anemones. This taxon has several taxonomic inconsistencies (Stampar et al. 2016) that are understudied, in part because of sampling difficulties (den Hartog 1977;Stampar et al. 2016). Moreover, ecology, behavior and life cycle in most species are poorly known (Nyholm 1943;Stampar et al. 2015Stampar et al. , 2016. Also, most systematic studies are solely based on morphological characters of few specimens (Carlgren 1912;Arai 1965;den Hartog 1977), leading to unreliable terminology (Arai 1965), and contributing to taxonomic uncertainty. A combination of these problems occurs in the genus Ceriantheomorphe Carlgren, 1931, which currently includes only two species: Ceriantheomorphe brasiliensis sensu Carlgren, 1931, and Ceriantheomorphe ambonensis (Kwietniewski 1898) (Carlgren 1931;den Hartog 1977).
The genus Ceriantheomorphe was described by Carlgren (1931) through the description of C. brasiliensis from southeastern Brazil. In this study, Carlgren also proposed that two species, Cerianthus ambonensis Kwietniewski, 1898 described from Ambon, Indonesia and Cerianthus brasiliensis Mello-Leitão, 1919 described from Guanabara Bay (Rio de Janeiro, Brazil) should be reassigned to the genus Ceriantheomorphe. As well, Carlgren (1931) also pointed out that Cerianthus brasiliensis is likely a synonym of Ceriantheomorphe brasiliensis sensu Carlgren 1931. However, assigning Cerianthus ambonensis as "Ceriantheomorphe ambonensis" would have been premature because the simple description made by Kwietniewski (1898) did not include any mention of deposited type material. Additionally, Carlgren (1931) was not able to observe the holotype of Cerianthus brasiliensis described by Mello-Leitão (1919), so his assignment of the species to Ceriantheomorphe must be viewed as tentative.
More than two decades after the description of Ceriantheomorphe, Carlgren and Hedgpeth (1952) reported C. brasiliensis from the Gulf of Mexico. However, the authors suggested that these specimens could possibly be another species due to the disjunct occurrence in relation to South American specimens (Carlgren and Hedgpeth 1952;den Hartog 1977;Spier et al. 2012).
Despite this taxonomic confusion, Ceriantheomorphe brasiliensis had been listed as an endangered species in Brazil for over 10 years (MMA 2004). Furthermore, the tubes built by C. brasiliensis, and Ceriantharia in general, play an important ecological role in providing suitable alternative substrates to many invertebrate groups (e.g., Bryozoa, Crustacea, Anthozoa) (Tiffon 1987;Kim and Huys 2012;Vieira and Stampar 2014). For example, some species, such as the phoronid Phoronis australis Haswell, 1883, are only found in ceriantharian tubes (Stampar et al. 2010). Thus, the survival of P. australis may be related to the management of the cerianthid species that houses them in the southern Atlantic.
This study aims to present a taxonomic review of the genus Ceriantheomorphe including a redescription of the holotype of C. brasiliensis, a redescription of C. ambonensis, and the description of a new species from the North Atlantic.

Specimens
Twenty specimens of Ceriantheomorphe were sampled by SCUBA: sixteen of C. brasiliensis from the South Atlantic, three from the North Atlantic, and one, C. ambonensis, from the Pacific Ocean (Table 1).

Morphological studies
The morphology of all specimens was studied through internal anatomy and cnidome studies, both based on criteria adopted by Carlgren (1931), Arai (1965), den Hartog (1977) and Stampar et al. (2015). All specimens were observed separately. Specimens were longitudinally dissected along the ventral side using surgical scalpels, photographed under an Opticam stereomicroscope, using the OPT HD 3.7 software and a general description of each body region was made. The morphological characters were compared between specimens and descriptions available in the relevant literature (Kwietniewski 1898;Carlgren 1931;Spier et al. 2012).

Phylum
Distribution. Southwestern Atlantic-Brazil (from the State of Espírito Santo (20.5°S) to Rio Grande do Sul (33.7°S) State) and Uruguay (34°S). This species was only observed in shallow waters (1-40 m depth).
Etymology. The specific name "adelita" is an allusion to an important group of women that fought during the Mexican Revolution. Occasionally, they adopted the identities of men to join in combat against the enemy.
Distribution. Gulf of Mexico (Northern Mexico) to North Atlantic (North Carolina, United States of America), shallow waters.

Comparison between congeners
Both Ceriantheomorphe brasiliensis and C. adelita sp. nov. have labial and marginal tentacles disposed in four cycles, whereas C. ambonensis has its tentacles arranged in three cycles. All three species have distinct labial and marginal tentacles arrangements (Table 5). Labial and marginal directive tentacles are present in C. brasiliensis and absent in C. ambonensis. Ceriantheomorphe brasiliensis has only two pairs of mesenteries connected to the siphonoglyph (Fig. 4), while C. adelita sp. nov. and C. ambonensis have three. In C. brasiliensis, the directive mesenteries are longer than P3 and P5 unlike C. adelita sp. nov. and C. ambonensis. Both C. brasiliensis and C. adelita sp. nov. have directive mesenteries longer than betamesenteries (B and b), while in C. ambonensis the opposite happens (Table 5). Protomesenteries (P3) are longer than metamesenteries (m) in C. adelita sp. nov. unlike C. ambonensis and C. brasiliensis. All three species have distinct proportions between metamesenteries (M × m) and betamesenteries (B × b) disposed along the gastrovascular cavity and a distinct size relation between mesenteries directive and metamesenteries (m) ( Table 5).

Taxonomic studies
As a result of the disjunct distribution of specimens identified as Ceriantheomorphe brasiliensis (Mexico+US/Brazil+Uruguay) and the incomplete description of Cerianthus ambonensis made by Kwietniewski (1898) that later was proposed to be reassigned to the genus Ceriantheomorphe (Carlgren 1931), Den Hartog (1977 pointed out the need for a revision of Ceriantharia with special focus in the genus Ceriantheomorphe. Based on morphological characters and biogeographic perspectives (Table 5), we were able to identify two different morphotypes among specimens assigned as Ceriantheomorphe brasiliensis. Specimens from the Gulf of Mexico were recognized as an undescribed species, formally described here as Ceriantheomorphe adelita sp. nov. Some previous studies with C. brasiliensis from the South Atlantic have shown that this species has short-lived planula larvae (unpublished data). This trait could prevent long dispersion due to biogeographic barriers, and thus this species may not be capable of reaching the North Atlantic. This is a different pattern from that reported for Isarachnanthus nocturnus for Isarachnanthus nocturnus, which is able to disperse over long distances due to the presence of long-lived planktonic cerinula larvae (Stampar et al. , 2015. Nonetheless, the maintenance of C. brasiliensis as a single species occurring in both northern and southern hemispheres would require some biogeographic events of which we have no evidence to date. Thus, in addition to the morphology, biogeographical understanding does not support the maintenance of these two populations as a single taxonomic unit. Carlgren and Hedgpeth (1952) argued that there were no differences between morphological characters in specimens from both areas (North and South Atlantic). Longer than P3, P5, betamesenteries (B and b) and metamesenteries (m), except for m of the 2 nd and 3 rd cycles. Shorter than P2, P4 and metamesenteries (M).
Longer than directive mesenteries and betamesenteries (B and b). Shorter than P2 and metamesenteries (M and m).

P4
Longer than directive mesenteries, P3, P5, betamesenteries (B and b) and metamesenteries (m), except for m of the 2 nd cycle. Shorter P2 and metamesenteries (M). We disagree with this assertion as C. adelita sp. nov. has several morphological characters that can distinguish it from other congeners. For instance, (1) marginal tentacles' arrangement, ratio between metamesenteries (M × m) and betamesenteries (B × b), as well as the proportion occupied by protomesenteries (directive mesenteries, P2 and P3) in the gastrovascular cavity contrast with those observed in other Ceriantheomorphe (Table 5); (2) protomesenteries P3 are found connected to the siphonoglyph while the same is not observed in C. brasiliensis; (3) directive mesenteries are shorter than P3, unlike C. brasiliensis; (4) the number of mesenteries connected to the siphonoglyph is distinct in C. brasiliensis and C. adelita sp. nov.; (5) directive mesenteries are longer than betamesenteries (B and b), the same, however, is not observed in C. ambonensis; (6) directive mesenteries are longer than all metamesenteries (m), distinct from the other two species of the genus; (7) protomesenteries (P3) are longer than all betamesenteries (B), while in C. brasiliensis P3 are shorter than betamesenteries (B) of the 1 st and 2 nd cycles; (8) P3 are longer than metamesenteries (m), in contrast to those observed in C. ambonensis and C. brasiliensis; (9) protomesenteries (P4 and P5) are present in C. brasiliensis, while absent in the other species. Some authors have discussed the taxonomic value of mesenteriel organization regarding the assignment and identification of species (Carlgren 1912;Arai 1965;den Hartog 1977). Spier et al. (2012) have reported that C. brasiliensis in southern Brazil has two pairs of mesenteries connected to siphonoglyph. In this study, C. adelita sp. nov. was found to have three pairs. Nevertheless, our results also showed that the two species of Ceriantheomorphe from the Atlantic Ocean have different numbers of mesenteries connected to siphonoglyph, reinforcing the potential taxonomic value of this character.
Our results also demonstrated that the use of ratios (division of the values) between metamesenteries (M × m) and betamesenteries (B × b) for each quartet can be useful to distinguish species of Ceriantheomorphe. In specimens of C. brasiliensis, the ratio between metamesenteries (M × m) and betamesenteries (B × b) ranged from 1.1 to 3.1% and from 1.2 to 3.1%, respectively. In comparison, the ratios observed in C. adelita sp. nov. are from 2.7 to 5.2% between metamesenteries and 3% between betamesenteries, while in C. ambonensis they range from 2.2 to 3.5% between metamesenteries and 4% between betamesenteries.
Similar to the ratio between metamesenteries (M × m) and betamesenteries (B × b), the proportion of protomesenteries found in the gastrovascular cavity was also useful to distinguish Ceriantheomorphe species in our study. While protomesenteries (P3) in C. adelita sp. nov. extend over half of the entire gastrovascular cavity length (56.4%), the ones in C. brasiliensis and C. ambonensis are much shorter (12.2% and 14.2%, respectively). Furthermore, we found differences between species while comparing protomesenteries length (Table 5). In this way, we suggest that mesenteries have a taxonomic value when used comparatively.

Geographic distribution of the genus Ceriantheomorphe
Currently, the genus Ceriantheomorphe has a wide geographic distribution; one species is restricted to the warm temperate northwest Atlantic (Gulf of Mexico and United States of America), another to the warm temperate southwestern Atlantic (southeast and South of Brazil and Uruguay) and C. ambonensis is recorded from tropical Central Indo-Pacific, Sunda Shelf (Indonesia) (Spalding et al. 2007). However, there are no records of the genus Ceriantheomorphe from areas between the Atlantic and Indo-Pacific Oceans, which suggests that the genus has a disjunctive distribution, since there is no evidence of any connection between the extant populations of the valid species (Fig. 12).
Disjunctive distribution patterns are exhibited by some marine invertebrates, even those having a free-swimming phase that would benefit wide dispersal, for instance, the bivalve Macoma balthica Linnaeus, 1758(Luttikhuizen et al. 2003) and the tunicate Ciona intestinalis Linnaeus, 1767(Caputi et al. 2007). However, in the current case, in our opinion the disjunctive distribution of the genus Ceriantheomorphe is evidence of the need of further studies on the genus, especially focused on taxonomy in some under-investigated areas of the Indo-Pacific Ocean.