Caymanostella, BELYAEV, 1974
|
publication ID |
https://doi.org/ 10.1093/zoolinnean/zlab060 |
|
DOI |
https://doi.org/10.5281/zenodo.5799397 |
|
persistent identifier |
https://treatment.plazi.org/id/039687A5-DD44-9210-3565-35D7AFC6FBDA |
|
treatment provided by |
Plazi |
|
scientific name |
Caymanostella |
| status |
|
CAYMANOSTELLA CF. SPINIMARGINATA BELYAEV, 1974 View in CoL
Material examined
KuramBio II expedition, RV Sonne, cruise SO250, St. 9, 19.08.2016, 43° 48.439’– 43° 47.643’ N; 151° 44.351’– 151° 44.513’ E, depth 5101–5134 m, slope of the Kuril-Kamchatka Trench. Three specimens that were collected from sunken wood fragments. Senckenberg Research Institute and Natural History Museum, Frankfurt, field ID KB2 121.
Description
Specimen № 1 (voucher CsSO250-9-1) ( Fig. 2G, H View Figure 2 ). Body subpentagonal, one ray is distinctly longer than the others and bent to the oral side. Colour in life and in alcohol is white. R = с. 5.3–6.7 mm (the longest ray is 8 mm), r = c. 5–5.7 mm.
Abactinal plates are fan-shaped. They imbricate uniformly towards the centre of the disc. Abactinal plates are similar in size, only five enlarged primary inter-radial plates are distinguishable. One of the primary inter-radial plates bears a simple madreporite, which has the appearance of several branching grooves surrounded by a few nodules ( Fig. 3D View Figure 3 ). Carinal rows are not discernible. At the tip of each arm, there is a terminal plate with a central opening. An unpaired terminal tube foot, possibly a sensory organ, is visible inside three of the five openings. Anal opening is central in position.
Marginal plates are arranged in two parallel rows along the body margin, peripheral to the abactinal plates. These plates are narrow and oriented perpendicularly or slightly obliquely to the body margin. There are 11 superomarginal and 11 inferomarginal plates along each side of the arm. The inter-radial-most superomarginal plates (two per inter-radius) are conspicuously larger in size compared with the adjacent superomarginals, and have notches for the gonopores on the upper margins ( Fig. 2G View Figure 2 ). More distally superomarginal plates gradually decrease in size towards the arm tip. Each inferomarginal plate bears on its outer edge two to three elongated fringe spines. These spines are 0.38–0.6 mm long and finely serrated. They are slightly compressed laterally, club-shaped in aboral view and dragonfly wingshaped, with an aborally pointing ‘shoulder’, in lateral view ( Fig. 4I View Figure 4 ). In total, there are 23–26 inferomarginal fringe spines along each side of the arm, from the interradius to the arm tip.
Abactinal and superomarginal plates, as well as the aboral sides of the inferomarginal plates, are uniformly covered by slightly capitate spinelets that have thorny terminal parts and short bases (hereafter ‘abactinal spinelets’) ( Fig. 4P View Figure 4 ). The spinelets are 0.12–0.30 mm in size. The largest abactinal spinelets (0.35–0.36 mm) are found around the terminal plate openings, anus and close to the distal margin of inferomarginal plates ( Fig. 4R View Figure 4 ). These enlarged abactinal spinelets are more capitate and more thorny than typical abactinal spinelets.
There are 12 pairs of adambulacral plates in each arm, including the longest arm. With the exception of the first (most proximal) plate, adambulacral plates are elongate and bar-like, extending from the furrow to the inferomarginal plates. The first adambulacral plate is the shortest and borders the actinal chamber, whereas the second plate extends alongside the chamber. The second and third plates are the longest. The subsequent plates decrease in size towards the arm tip. Adambulacral plates and the oral sides of the inferomarginal plates are uniformly covered by tapering, serrate spines. These spines are slightly appressed to the plate surface and point towards the body margin ( Figs 2H View Figure 2 , 4M, N View Figure 4 ). Adambulacral plates bear up to 12 spines (three spines on the first, shortest plate). Each plate bears one furrow spine. The second spine is usually placed further from the furrow, and the subsequent spines are arranged across the plate in two alternating rows. Adambulacral spines are 0.37–0.50 mm long.
Five heart-shaped actinal chambers covered by a thin semitransparent membrane (‘actinal membrane’, following Rowe, 1989) occur inter-radially on the oral side, one in each inter-radius. Each actinal chamber is separated into two parts by a septum. Oval gonad lobes (approximately 0.7–1.0 mm long, measured in dried specimen) are visible inside some of the chambers through the actinal membrane. A preliminary histological study of the gonads showed that they contain sperm cells at early developmental stages (A. L. Drozdov, pers. comm.).
The oral opening is large, slightly distorted, elongated, approximately 5 mm long and 3.5 mm wide. The stomach was partly everted in freshly caught specimen. Paired oral plates are broader than long, without a median keel. Each plate bears three (rarely two) marginal spines, situated on the oral edge and at the corner, and one suboral spine. All oral spines are similar to adambulacral spines in size but are more robust. Like the adambulacral spines, they are appressed to the plate surface and point towards the body margin. Suckered tube feet are in two rows, 13 pairs along each arm.
Specimen № 2 (voucher CsSO250-9-2) ( Fig. 5C, D View Figure 5 ). Body is about subpentagonal in shape, compressed laterally and bent in the oral–aboral direction. Colour in life and in alcohol is white. R = c. 5.3–9.5 mm, r = c. 4.5–6.5 mm. The general appearance and plate arrangement are similar to those of specimen №1. Abactinal spinelets are 0.14–0.24 mm long. There are 27–30 inferomarginal fringe spines from the interradius to the arm tip. The spines are 0.45–0.80 mm long, two (three) on each plate. In lateral view, these spines vary in shape from dragonfly wing-shaped to slightly curved ( Fig. 4J, K, L View Figure 4 ). There are 13–14 pairs of adambulacral plates in each arm. Each plate bears up to 12 spines that are perpendicular to the plate surface or slightly appressed towards the body margin, 0.35– 0.53 mm long ( Fig. 4O View Figure 4 ). Two (sometimes three) spines closest to the furrow (including the furrow spine) are arranged in a row transversely to the furrow edge, and the others follow in two alternating rows.
Actinal chambers are covered by thin semitransparent membrane through which lobate gonads are visible in some interradii; and in other interradii, globules approximately 0.4 mm in diameter can be observed.
The oral opening is distorted and elongated, and is approximately 5 mm long. Stomach is not everted. Paired oral plates are slightly broader than long, without median keel. Each oral plate bears two to three marginal spines and one or two (rarely three) suboral spines. Suckered tube feet are in two rows, with 14–15 pairs along each arm.
Specimen № 3 (voucher CsSO250-9-3) ( Fig. 5A, B View Figure 5 ). Body is subpentagonal in shape. It is bent in the oral–aboral direction. R = c. 5.3–8.1 mm, r = c. 4.5–5.1 mm. The general appearance is similar to that of specimens №1 and №2. Abactinal spinelets are 0.14–0.22 mm long (0.23–0.29 mm around terminal openings). Anal opening is not visible.
There are 25–30 inferomarginal fringe spines from the inter-radius to the arm tip. These spines are 0.43– 0.6 mm long, two on each plate. There are 14–15 pairs of adambulacral plates per ray. Each adambulacral plate bears up to 11 spines. Spines of the adambulacral plates are perpendicular to the plate surface and 0.3–0.41 mm long.
Actinalchambersarecoveredbythinsemitransparent membrane through which gonad lobes or oval bodies (0.3–0.4 mm in size) are visible. The oral opening is large, circular and approximately 3.4 mm in diameter. Stomach is partly everted. Paired oral plates are slightly broader than long, without median keel. Each oral plate bears two to three (rarely one) marginal and one to three suboral spines. Suckered tube feet are in two rows, 15–16 pairs along each arm.
Remarks
Specimens from the KuramBio collection are morphologically most similar to Caymanostella spinimarginata ( Table 4 View Table 4 ; Fig. 2E–F View Figure 2 ). In both typical C. spinimarginata from the Cayman Trench and specimens from the deep-sea North-West Pacific, the abactinal skeleton consists of uniformly imbricating fanlike plates of similar size, among which five slightly larger primary inter-radial plates are distinguishable, one in each inter-radius. The carinal row is not discernible, several radial rows of uniform plates extend towards the tip of the arm. Such abactinal skeleton arrangement is also characteristic of C. phorcynis , as seen in fig. 5 of Rowe, 1989 (although Rowe mentioned the presence of a carinal row). Caymanostella admiranda and C. madagascarensis have a different type of abactinal plate arrangement. In these two species, abactinal plates are less numerous, polygonal rather than fan-shaped in outline and irregularly imbricating. One distinguishable carinal row of plates extends along the mid-radial line of each arm ( Fig. 2A–D View Figure 2 ).
The new specimens are also consistent with the typical C. spinimarginata in having gonopores that are placed in notches on the upper margins of the proximal-most superomarginal plates and not piercing as in C. admiranda or C. phorcynis . The position of the gonopores is unknown in C. madagascarensis . In the holotype (the only known specimen of this species) gonopores were not detected during current examination, as well as by Belyaev & Litvinova (1991). Based on the small size of the C. madagascarensis specimen, the gonopores have likely not yet been developed. Rounded upper margins of the proximalmost superomarginal plates indicate that the gonopores are more likely piercing in adult specimens of C. madagascarensis .
The new Caymanostella specimens and C. spinimarginata are similar in the appearance of the abactinal spinelets, which are slightly capitate and consist of base, short stem and expanded crown ( Fig. 4F–H, P, R View Figure 4 ). The other three Caymanostella species are covered dorsally by granulation, not by spinelets. At high magnification, granules of C. admiranda appear as dome-shaped spiky bodies ( Fig. 6A, B View Figure 6 ). Dome-shaped abactinal granules were described by Rowe (1989) for C. phorcynis but were not illustrated in detail. Granules of C. madagascarensis markedly differ from those of other congeners: they are tiny, almost spherical and are pierced with holes ( Fig. 6D, E View Figure 6 ).
Inferomarginal fringe spines of the KuramBio II specimens are similar to those of most Caymanostella species (with the exception of C. madagascarensis ): slightly compressed laterally and more or less wing-shaped in lateral view ( Figs 4A–C, I–L View Figure 4 , 6C View Figure 6 ). Caymanostella madagascarensis has fringe spines of a highly specific shape. They are flattened in the aboral–oral direction, crown-shaped at the tips and bear longitudinal ridges ( Fig. 6F View Figure 6 ).
The appearance of the adambulacral plates and spines in the new Caymanostella specimens is typical for all species of the genus. However, the maximum number of spines per adambulacral plate differs among species and among specimens of different sizes. New Caymanostella specimens, as well as C. spinimarginata and C. phorcynis , bear ten or more spines on the longest plates in the largest specimens, in contrast to C. admiranda and C. madagascarensis , which bear a maximum of four and five spines, respectively ( Table 4 View Table 4 ). The low number of spines recorded in C. madagascarensis (the only known specimen has R = 2.5 mm) is probably related to its small size. By comparison, a similarly small C. spinimarginata specimen (paratype №2, R = 2.1–2.6 mm) has a maximum of five spines per plate. In contrast, the low number of adambulacral spines recorded in C. admiranda cannot be explained by size differences. The holotype of C. admiranda ( R = 3.35 mm) has a maximum of four spines per adambulacral plate, whereas C. spinimarginata specimen of a similar size (holotype, R = 3.5–3.7 mm) has up to eight spines.
Despite their morphological similarity, KuramBio II specimens show some differences from typical C. spinimarginata : larger body size, more complex madreporite, larger size of the inferomarginal fringe spines and adambulacral spines ( Table 5 View Table 5 ). The size of typical (not enlarged) abactinal spinelets is also larger. Most likely, the differences in the size of the adambulacral and marginal fringe spines are related to the size of the specimens. Among eight C. spinimarginata- type specimens, larger ones are characterized by larger inferomarginal and adambulacral spines. At the same time, abactinal spinelets are nearly equal in length in all eight specimens. Therefore, the larger abactinal spinelets of the KuramBio II specimens are unlikely to reflect their larger body size.
The madreporite of Caymanostella has not yet been described in detail. Rowe (1989) only noted the presence of the madreporite in C. spinimarginata and C. phorcynis without providing any further details. However, heindicatedthatthefamilyCaymanostellidae is characterized by a ‘small but typical madreporite’ in contrast to Smith’s (1988) opinion that the madreporite in Caymanostella forms a simple pore (from: Gale, 2011: 76). Examination of all Caymanostella specimens at our disposal shows the presence of a more or less simple madreporite that differs in shape among different species and specimens of different sizes. In each case, the madreporite was located on one of the primary inter-radial plates. In C. admiranda (both holotype and paratype examined), the madreporite has the appearance of a small knob pierced with a pore ( Fig. 3B View Figure 3 ). The madreporite of C. spinimarginata (examined in paratype №3) and C. madagascarensis (examined in holotype) has an appearance of a slitlike opening oriented perpendicular to the body margin ( Fig. 3A, C View Figure 3 ). The KuramBio II specimen №1 has a more complex madreporite shaped as several branching grooves, surrounded by a few nodules ( Fig. 3D View Figure 3 ). In other examined Caymanostella specimens, the madreporite was not visible because of the dense cover of abactinal spinelets or a distorted body shape. The observed difference in the madreporite appearance between C. spinimarginata and the new specimens can be explained by age variability. In the smaller C. spinimarginata specimen, the madreporite is at its early stage of development, whereas the KuramBio II specimen №1, due to its larger size, demonstrates the later stage. A similar pattern of madreporite development, from slit-like early madreporite to more complex and comprising several grooves, was described for Asterias rubens Linnaeus, 1758 ( Gondolf, 2002) . Based on Rowe’s opinion (1989), and the present study, it can be suggested that C. spinimarginata (including the new material), C. madagascarensis and C. phorcynis have slit-like madreporites in small specimens and typical but simplified madreporites in large specimens. Caymanostella admiranda differs significantly from congeners in this character. Smith and Rowe possibly described madreporites from different Caymanostella species or from specimens of different sizes.
The analysis performed above shows that the new specimens most likely represent morphological variation of C. spinimarginata and are thus considered here as C. cf. spinimarginata . However, molecular data on typical C. spinimarginata are needed to confirm that they belong to the same species.
MOLECULAR PHYLOGENETIC ANALYSES
Convergence of the MCMC runs of the BI analysis is confirmed by the usual set of diagnostics. The average standard deviation of split frequencies is lower than 0.01, and the PSRF (potential scale reduction factor) is close to 1 for all parameters. High swap acceptance rates (0.28–0.63) and large (4641–38 316) effective sample size ( ESS) values corroborate the efficiency of the analysis.
Bayesian inference and ML analyses resulted in trees with largely congruent topologies (BI tree shown in Fig. 7 View Figure 7 ). The most striking difference is that ML analysis usually provides lower statistical support compared with BI for the same clade. On both phylogenetic trees, the clade, upholding the taxa belonging to the order Velatida together with the genus Xyloplax (order Peripodida ), is recovered with strong to moderate (depending on the analysis) statistical support ( Fig. 7 View Figure 7 ). Within this clade, Xyloplax is distinguished as a sister-taxon to the family Korethrasteridae (posterior probability = 0.8; bootstrap support = 48%). The sisterclade to the Velatida + Peripodida clade upholds two major groups. The first group represents the wellsupported, monophyletic group of Forcipulatacea . The forcipulatacean clade contains several groupings of the paraphyletic order Forcipulatida along with the well-upheld clade, comprising the members of Brisingida ( Fig. 7 View Figure 7 ). The second major group has strong to moderate (depending on the analysis) support and comprises taxa belonging to the superorders Valvatacea and Spinulosacea. Valvataceans are represented by several clades of the paraphyletic order Valvatida , along with the Paxillosida + Notomyotida clade. All spinulosaceans are grouped into a well-supported Spinulosida clade. Caymanostella is recovered within one of the valvatidan clades as a sister-taxon to the ophidiasterid Leiaster Peters, 1852 ( Fig. 7 View Figure 7 ). However, monophyly of this grouping remains questionable: despite the strong support in the BI analysis (posterior probability = 0.98), this clade is weakly supported by ML (bootstrap support = 41%).
At the same time, Bayes’ factor ( BF) tests strongly support the Caymanostellidae + Ophidiasteridae grouping within the Valvatacea + Spinulosacea clade, as indicated from BI and ML trees ( Table 3 View Table 3 : H 0). Alternative topological hypotheses that nested caymanostellids within other major asteroid groups or placed caymanostellids as a sister-taxon to the valvatidan family Asterinidae are rejected with 2lnBF ranging from 20.86 to 171.34 ( Table 3 View Table 3 ).
The AU test shows that unconstrained ML phylogenetic tree has a significantly (P <0.05) better likelihood than the Caymanostellidae + Forcipulatacea , Caymanostellidae + Spinulosida and Caymanostellidae + Asterinidae constrained ML trees ( Table 3 View Table 3 ). Hypotheses H 1 and H 3 that concerned the groupings Caymanostellidae + Velatida + Peripodida and Caymanostellidae + Paxillosida + Notomyotida , result in lower likelihood values but are not rejected by the AU test (P = 0.175 and P = 0.337, respectively).
| RV |
Collection of Leptospira Strains |
| R |
Departamento de Geologia, Universidad de Chile |
| MCMC |
Museo de Historia Natural de la Ciudad de Mexico |
| ESS |
University of Duisburg Essen |
| ML |
Musee de Lectoure |
No known copyright restrictions apply. See Agosti, D., Egloff, W., 2009. Taxonomic information exchange and copyright: the Plazi approach. BMC Research Notes 2009, 2:53 for further explanation.