Revision of the deep-water spider crab genus, Scyramathia A. Milne-Edwards, 1880, with the description of a new species from the Mediterranean and notes on Rochinia A. Milne-Edwards, 1875, and Anamathia Smith, 1885 (Crustacea, Decapoda, Brachyura, Epialtidae)

The taxonomy of the deep-water spider crabs of the genus Scyramathia A. Milne-Edwards, 1880, is revised and four extant species are recognised from the Atlantic and western Indian Ocean: S. carpenteri (Norman, in Thomson 1873) (type species), S. umbonata (Stimpson, 1871), S. hertwigi Doflein, in Chun 1900, and S. tenuipes sp. nov. Scyramathia tenuipes sp. nov. from the Mediterreanean is easily distinguished from its congeners by its slender and elongate ambulatory legs. All species are diagnosed and figured. The taxonomy of two allied genera from the Atlantic and Mediterranean, Rochinia A. Milne-Edwards, 1875, and Anamathia Smith, 1885, are also treated and their type species redescribed and figured.


Introduction
The deep-water epialtid spider crab genus, Scyramathia A. Milne-Edwards, 1880, was recently reinstated by Tavares and Santana (2018) who recognised two extant valid species, S. carpenteri (Norman, in Thomson 1873), andS. umbonata (Stimpson, 1871), as well as one fossil taxon, S. boschii (Casadío, Feldmann, Parras & Schweitzer, 2005). Scyramathia had previously been in the synonymy of Rochinia A. Milne-Edwards, 1875, for many years  Remarks. No type specimen was designated when this species was described by A. Milne-Edwards (1875: 86 [footnote]) although various specimens from "la collection du Muséum" (i.e., the Muséum national d'Histoire naturelle (MNHN), Paris) as well as specimens from Hassler expedition were mentioned. There is also a fig-ure of a male specimen provided in the same paper, but the name was incorrectly spelled as "Rachinia gracilipes" in the caption (see A. Milne-Edwards 1875: pl. 18 fig. 1). The male specimen that is found in the MNHN (MNHN-IU-2000-4460 [= MNHN-B4460]) matches the figure provided by A. Milne-Edwards (1875: pl. 18 fig.  1-1d). The same male specimen was used by Tavares (1991: figs 3A, 5B) and was considered a syntype by the author. This was, however, not listed by Tavares and Santana (2018) in their study to redefine Rochinia, when they designated another specimen as the lectotype for this species. It was mentioned by Tavares and Santana (2018) that the type material needs clarification and that the male specimen that was illustrated by A. Milne-Edwards (1875: pl. 18, fig. 1-1d) and deposited in the MNHN was clearly not the holotype as had been identified by Rathbun (1925). Rathbun's (1925) recognition of the "holotype" is incorrect as no type was selected in the original description by A. Milne-Edwards (1875) and all the material must be considered syntypes. Tavares and Santana (2018) designated one male specimen (MCZ 1950) as the lectotype for this species and five other female specimens as the paralectotypes, all collected from the Hassler expedition. Based on the original description by A. Milne-Edwards (1875: 86) and the figure (A. Milne-Edwards 1875: pl. 18 fig. 1-1d), the specimens from MNHN are also part of the type series. The G1 and G2 of the paralectotype male (MNHN-IU-2000-4460 [= MNHN-B4460]) is illustrated here ( Fig. 2A-F). The type locality of this species is Cape Corrientes (= Cabo Corrientes), Argentina, and specimens are found in relatively shallow waters (about 55 m depth) (see also Tavares & Santana, 2018: 221).
Rochinia gracilipes was recorded from the coast of Africa near Gabon by Monod (1956). The male specimen figured (Monod 1956: figs 706-708) was compared with the MNHN type specimen and it does resemble the present concept of R. gracilipes. Monod (1956: 517), however, was not certain of the origin of the specimen and noted that its presence in African waters needs to be confirmed. Monod (1956)'s dubious African record is not included in the synonymy for this species for the time being.
Remarks. The identity of Amathia carpenteri Norman, in Thomson 1873 s. str., is critical to the status of Scyramathia as it is the type species of the genus. Although Tavares and Santana (2018) clarified the identity of the genus and had material of the species, they did not examine the types for this species. The species was named by Thomson in 1873 in his famous book about the exploration of the deep Atlantic by the H.M.S. Porcupine, "The Depths of the Seas" (p. 175: fig. 35; reproduced here as Fig. 3), under the name "Amathia Carpenteri" to honour his colleague Dr. W.B. Carpenter. There was no clear indication of number of specimens examined, with the author only commenting that "Another handsome new species, Amathia carpenteri, Norman ( fig. 35), was common in the sandy chalkmud of the Holtenia ground." (Thomson 1873: 176). The figure shows that the cardiac region of the carapace appeared to be damaged (Thomson 1873: fig. 35; reproduced here as Fig. 3). In the NHM are three specimens that are labelled as syntypes of A. carpenteri, all obtained by the H.M.S. Porcupine. None of these specimens, however, match the measurements of the type figure, which was "once and a half the natural size" (Thomson 1873: fig. 35; reproduced here as Fig. 3). Of these, one male specimen (NHM 1907.8.28.3) agrees relatively well with the figure by Thomson (1873: fig. 35; Fig. 3), and the specimen label mentions that it was figured in Thomson (1873). This specimen, with catalogue number, NHM 1907.8.28.3, was noted by Christiansen (1969: 122) to be the lectotype for this species, and in an unpublished note, was also selected by Isabella Gordon (also unpublished data) to be the lectotype. The same specimen was also mentioned by Rice (1990: 10), noting that it is similar to the figure by Norman, in Thomson (1873: fig. 35). The type locality of this species was stated to be on "sandy chalk of the Holtenia ground". Holtenia Ground is an area between North Scotland and the Faeroe Islands that covers an area of approximately 500 km 2 , with a depth of 820 to 1000 m (Carpenter et al. 1870;Reiswig and Champagne 1995). Specimen NHM1907.8.28.3 is here formally designated as the lectotype of Amathia carpenteri Norman, in Thomson 1873, to stabilise the taxonomy of the species and the genus.
The three type specimens differ markedly from each other in many ways although they are all from Holtenia Ground. The lectotype male (21.6 × 15.9 mm, NHM 1907.8.28.3) selected here differs markedly from the other two paralectotypes in having a smoother carapace, without the distinct raised plate-like structures on the carapace and the branchial regions are also proportionately more inflated (Figs 4A, 7A). The next largest type is a subadult female (13.0 × 8.4 mm, NHM 1911.11.8.377), and its carapace more closely resembles typical S. carpenteri, with the plates more developed. It is almost certainly the same specimen figured by Norman, in Thomson (1873: fig. 35; Fig. 3), with the shape and structures of the spines and plates agreeing very well, althought the left pseudorostral spine of the specimen is now broken (Fig. 7C). The third type specimen is a small and poorly preserved juvenile female (5. 4 × 3.0 mm, NHM 1910.2.4.213) in which the dorsal surface of the carapace as well as the third maxillipeds are covered with setae but without any obvious plates. The surfaces of the third maxillipeds of the other two type specimens are finely granulated or smooth (Figs 4B, 7B, D).
Scyramathia carpenteri exhibits strong variation on the plate-like structures on the carapace of various sized specimens showed variation. D'Udekem d'Acoz (1999) observed that "L'ornementation de la carapace de R. carpenteri est extrêmement variable" but did not mention if it is due to the size of the specimens. Similar sized specimens showed either weak or strong plate-like structures ( Fig. 9A-I). These differences are also seen in the type  specimens where the larger sized male lectotype specimen has weak plate-like structures on the carapace whereas the same structures on the smaller sized female paralectotype specimens are relatively stronger. The variation of the plate-like structures on the carapace are therefore, not always size dependent. Other morphological characters that show variation on S. carpenteri includes the divergence of the pseudorostral spines, from relatively diver- Superficially, S. carpenteri resembles S. umbonata (Stimpson, 1871) from the western Atlantic, which also appears to grow to a much larger size (57 mm maximum carapace length; see Tavares et al. 2015). There are, however, consistent differences, with S. carpenteri, which has relatively longer pseudorostral spines, and the outer mar-gin of the basal antennal article is always straight or only slightly convex. The pseudorostral spines of S. umbonata are always proportionately shorter, even when similar sized specimens are compared, and the outer margin of the basal antennal article is always distinctly convex (cf. Tavares et al. 2015: figs 1, 2, 4, 5). These differences are reliable for specimens of similar sizes. As such both species are here recognised as separate taxa.
It is interesting to note that some studies (A. Milne-Edwards and Bouvier 1899Bouvier , 1900Bouvier 1922Bouvier , 1940 mentioned that S. carpenteri does not occur in the Mediterranean, which is the region where the species, S. tenuipes sp. nov., is now found (see remarks for S. tenuipes sp. nov.). Alphonse Milne- Bouvier (1899, 1900) and Bouvier (1940) had material they referred to "S. carpenteri" from the Azores, and up north of São Jorge, and the Sahara beach. Based on their figures (see A. Milne-Edwards and Bouvier 1899: pl. 1 fig. 4; A. Milne- Edwards and Bouvier 1900: pl. 20 Figs 1, 6;Bouvier 1940: pl. 14 fig. 1), however, the specimens they figured are more similar to S. tenuipes sp. nov., with the ambulatory legs long and slender, and these records are hence referred there for now. The plates on the carapace as figured (see A. Milne-Edwards and Bouvier 1899: pl. 1 fig. 4; A. Milne- Edwards and Bouvier 1900: pl. 20 Figs 1, 6;Bouvier 1940: pl. 14 fig. 1) appear relatively stronger than what has been observed in S. tenuipes sp. nov., but as discussed earlier, this character is known to vary in S. carpenteri s. str. in any case.
Distribution. Scyramathia carpenteri is known from the type locality, "sandy chalk of the Holtenia ground" [= between North Scotland and the Faeroe Islands] (Carpenter et al. 1870;Norman, in Thomson 1873;Reiswig and Champagne 1995), South Iceland (Hansen 1908), Southwest Faeroe Islands, South of Norway, West of British Isles (Clark 1986).

Remarks. The author and year of publication for Scyramathia hertwigi has traditionally been cited as Doflein
The Atlas by Doflein (1904: pl. 27 figs 1-7) has figures that show a series of sizes of the carapace with plates, in particular the horse-shoe shaped epibranchial plate. In this species, the pseudorostral spines are always proportionately shorter than those in S. carpenteri. The differences between S. carpenteri and S. hertwigi were discussed by MacPherson (1983: 35), and we agree with his interpretation. There are also slight variations amongst the specimens from South Atlantic (SMF 23085), which tend to have less distinct plates on the carapace, and slightly shorter and stouter ambulatory legs while a large specimen (AM P34653) from South Africa has distinct plates on carapace, and somewhat longer and slender ambulatory legs. These differences, however, are not substantial and should not be a reason to treat them as different.
The specimens of "Scyramathia carpenteri" from off Cape Bajador, West Africa (= off Boujdour, Western Sahara) reported by Tavares and Santana (2018: 204) may not be that species. This location overlaps with the distribution range of S. tenuipes sp. nov., and it is possible their specimens belong to this species instead. Their material will need to be re-examined.

Scyramathia umbonata (Stimpson, 1871)
Remarks. Scyramathia umbonata was originally described as a species of Scyra by Stimpson (1871) from off Sand Keys in Florida. Stimpson (1871) provided mea-surements for one male specimen but did not indicate if he had more material. It was transferred to Scyramathia by A. Milne-Edwards (1880a), and Rathbun (1894) subsequently referred it to Anamathia. Rathbun (1925) later transferred the species to Rochinia when she synonymised the genera. The type material for the S. umbonata is probably no longer extant (Tavares et al. 2016). Tavares et al. (2016) discussed the taxonomy of S. umbonata at length (see also Tavares and Santana 2018) and synonymised Rochinia confusa Tavares, 1991, under S. umbonata, stating that R. umbonata undergoes drastic changes in morphology as it grows, with more distinct plates being present on larger sized specimens. This is the same pattern as seen in S. carpenteri and S. hertwigi. Tavares and Santana (2018) later transferred it back to Scyramathia. There is no need to redescribe or figure this species as the detailed study and figures of this species by Tavares et al. (2016) and Tavares and Santana (2018) are quite sufficient. For these reasons, a neotype is also not needed for this species for the time being.  1999: 194 (in part); Abelló et al. 2002: 187 (  Diagnosis. Carapace pyriform. Pseudorostral spines straight, almost half of carapace length. Supraorbital eave fused with carapace, with blunt preorbital spine; postorbital lobe fused with hepatic spine. Carapace with plates: 1 hepatic spine, 1 small granule above 1 oblong mesogastric plate, 1 protogastric granule, 1 epibranchial plate, 1 oblong cardiac plate, 1 metabranchial granule, 1 lateral branchial spine, 1 blunt posterior spine (Figs 12, 13A). Antennal flagellum shorter than pseudorostral spines, about half of pseudorostral length. Basal antennal article longer than broad, straight outer margin with blunt roundish distal angle of article (Fig. 13B). Buccal frame with distal angle of buccal frame slightly raised with round edge. Pterygostomial region with 3 or 4 granules on outer margins (Fig. 13B). Chelipeds slender, propodus longer than fingers, slightly carinate margin; carpus with granules; merus smooth, triangular in cross-section. Ambulatory legs slender; merus with distal angle blunt; male P2 merus length 1.0-1.4 times carapace length, female P2 merus length 1.0-1.3 times carapace length, male P2 merus length 17.3-20.5 times width, female P2 merus length 14.8-21.5 times width; male P5 merus length 0.4-0.7 times carapace length, female P5 merus length 0.5-0.6 times carapace length, male P5 merus length 7.1-9.5 times width, female P5 merus length 7.9-8.8 times width (Figs 12, 13A, 14F-J). Male thoracic sternum flat, slightly concave; sternites 3, 4 widest. Male pleon with triangular telson and all 6 somites free; somites 2, 3 widest; surface of somites smooth (Fig. 13B). G1 straight, distal part relatively shorter, sharp tip ( Fig. 15E-H).
Etymology. The term "tenuis", which means thin in Latin, is used for the slender ambulatory legs. The name is used as a noun in apposition.
Remarks. Scyramathia tenuipes sp. nov. is superficially similar to S. carpenteri, but there are several important differences between the two species. Firstly, all the adult specimens of S. tenuipes sp. nov. examined have relatively less distinct plates on the carapace (Figs 12, 13A) (versus adult S. carpenteri usually have more prominent plates on the carapace; Figs 3, 4A, 5A, 7A, C, 9A-I).
More significantly, it also has distinctly longer and more slender ambulatory legs, with the P2 meri length of all the male specimens longer or subequal in length (1.0-1.4 times) to the carapace length, and the P5 meri length 7.1-9.5 times longer than the width (Figs 12, 13A, 14F-J) (versus S. carpenteri has relatively shorter and less slender ambulatory legs, with the P2 meri of all male specimens typically subequal in length (0.7-1.1 times) to the carapace length, and the P5 meri length 4.2-7.3 times the width; Figs 3, 4A, 5A, 14A, B). This character is consistent in all the adult male and female specimens examined. The pterygostomial region of adult S. tenuipes sp. nov. has relatively more distinct granules (Fig. 13B) compared to the more plate-like granules on that of adult S. carpenteri (Figs 4B, 5B, 6C, 7B, D); and the distal part of the G1 of S. tenuipes sp. nov. is always relatively shorter ( Fig. 15E-H) compared to the longer ones of S. carpenteri ( Fig. 8C-F).
An unpublished genetic study by the first author comparing S. tenuipes sp. nov. with S. carpenteri, S. hertwigi, S. umbonata and other Rochinia species shows small but consistent differences that indicate we are dealing with a recent but separate species. A total of five genes, three mitochondrial genes: COI, 12S, 16S, and two nuclear genes: 18S and H3, were used for the molecular analysis in this unpublished work. The cytochrome oxidase I (COI) dendrogram for S. tenuipes sp. nov. and S. carpenteri shows a consistent 0.3% difference between them whereas the differences between the other Scyramathia and Rochinia species ranged from 1.3-10.9 %. Significantly, the phylogenetic tree from the Maximum Likelihood analysis shows an 87% support for S. tenuipes sp. nov. and S. carpenteri as separate clades.
The type locality of S. tenuipes sp. nov. is an interesting area in the Mediterranean. The Alboran Sea is at the western narrow part of Mediterranean that ends at the Strait of Gibraltar. The complex circulation of the waters through this very narrow strait of approximately 14 km, is known to play a key role in regulating the gene flow for a number of benthic species (Palero et al. 2011). The peculiar hydrographic features of the Strait of Gibraltar were established since the end of the Messinian (upper Miocene) about 5.4 million years ago, after a period of very low sea level of 500,000 years where the strait was dry, and the Mediterrannean Sea was completely closed. Between Spain and Morocco, the very narrow strait has several choke points not deeper than 350 m that seriously affects the water currents. The surface Atlantic waters are entering at speeds varying from 1 to 2 knots, whereas the outflow of the Mediterranean waters nearer the sea bottom flows at a speed of 1.1 to 0.9 knots (Millot 2005). The physical characteristics of these two opposite currents are also very different: the Atlantic surface water has a salinity of 36 g/l while the Mediterranean waters are denser at 38 g/l of salt. As a result of such conditions, the Alboran Sea has more influx from the Altantic Ocean than the rest of the Mediterranean basin (Pascual et al. 2016). These peculiar oceanographical conditions could explain why S. tenuipes sp. nov. is most common in the Alboran Sea and not in the rest of Mediterranean or even in the main Atlantic waters adjacent. In the Mediterranean, S. tenuipes sp. nov. is known only from the western part of Mediterranean along the coast of Spain (Marco-Herero et al. 2015;Palero pers. comm.). The un-    Tortonese (1964) on the Mediterranean fish fauna have shown that the fish fauna differs in the different regions in the Mediterranean, for example the scorpionfish, Trachyscorpia cristulata echinata (Köhler, 1896) (Sebastidae), is found only in the western part of Mediterranean, which is also observed by submersible during the Cyanalboran cruise (Zibrowius pers. comm.). It has also been discussed by d'Udekem d'Acoz (1999) that there are crustacean species that showed variations in morphological differences between the Atlantic and Mediterranean populations with unknown, continuous or absence of intermediate populations in between (see d'Udekem d'Acoz 1999: table 1). These morphological differences due to the geographical distribution are linked to the different physical characteristics of the Atlantic and Mediterranean waters (d'Udekem d' Acoz 1999). However, it is likely that these species might be separate ones, as in this case, which requires more study.
There are various records of S. tenuipes sp. nov. as "S. carpenteri", and while A. Milne- Bouvier (1899, 1900) and Bouvier (1922Bouvier ( , 1940 state that S. carpenteri was not known from the Mediterranean, d'Udekem d'Acoz (1999) noted that a Mediterranean specimen was figured by Dieuzeide (1955). The figure of the Mediterranean specimen by Dieuzeide (1955: fig. 9), shows an animal with long and slender ambulatory legs, weak plates on carapace, and relatively straight and slightly divergent pseudorostral spines, diagnostic characters of S. tenuipes sp. nov. (Figs 12, 13A). Subsequently, Dieuzeide (1960) commented that the species was commonly found among the bamboo coral, Isidella elongata (Esper, 1788) (Isididae: Alcyonacea) in the Mediterranean. Some of the specimens of S. tenuipes sp. nov. that were examined in this study were found with unidentified deep-sea anemones attached on their carapace.
Remarks. The genus was first described as Amathia by Roux (1828). The name was changed to Anamathia by Smith (1885), as the earlier name was preoccupied for a bryozoan (Lamouroux 1812). This genus was synonymised by Rathbun (1925) under Rochinia A. Milne-Edwards, 1875, with no detailed explanation. Ng et al. (2008) listed the genus as valid as well but again with no explanation. Tavares and Santana (2018) discussed the matter and considered this genus as valid, listing the morphological differences from Rochinia and Scyramathia. Anamathia hystrix (Stimpson, 1871) was also transferred to this genus by Tavares and Santana (2018) without explanation. Although the external features of this species resemble A. rissoana, R. hystrix has a strong preorbital spine and a pronounced mesogastric spine in the middle of the carapace; and on the basis of this as well as unpublished genetic data, we are of the opinion that it should not be placed in this genus. As such, R. hystrix together with the rest of the Atlantic and East Pacific Rochinia sensu lato species will be dealt with separately by Lee et al. (in prep). Rochinia rissoana : Garth 1958: 283, 289;Clark 1986: 192, 193 (map); d'Udekem d' Acoz 1999: 194;González-Gordillo et al. 2001 Description. Carapace pyriform, covered with layer of setae, smooth when denuded (Figs 16A, 17A). Supraorbital eave fused with carapace with blunt preorbital angle; postorbital lobe cup-like, small (Fig. 17B).  Carapace with 13 spines: 1 hepatic spine, 1 protogastric spine, 1 metagastric spine, 1 lateral epibranchial spine, 1 lateral branchial spine, 1 cardiac spine, 1 metabranchial spine, 1 strong intestinal spine near posterior of carapace (Figs 16A, 17A).

Anamathia rissoana (Roux, 1828)
Antennal flagellum shorter than pseudorostral spines. Basal antennal article longer than broad, narrow; straight outer margin with blunt distal angle. Presence of granule at base of article. Buccal frame covered by third maxilliped, distal angle of buccal frame distinct, highly protruded, forming blunt angle. Pterygostomial region with 2 or 3 granules on outer margin; second granule biggest (Fig. 17B).
Remarks. Roux (1828) described Amathia rissoana from the Mediterranean, and it was noted by Roux (1828) that the rare crab is found 20 metres deep amongst "les algues et les focus". Roux (1828) did not indicate the number of specimens that was examined, but he had at least one male and one female which were figured (Roux 1828: pl. 3; reproduced here as Fig. 16A-C). These specimens are therefore syntypes; but the whereabouts of Roux's specimens are not known. His figures, however, are relatively detailed and leave no doubt regarding the species identity.
The morphology of this species is distinct from species of Rochinia in having distinct carapace spines (Figs 16A, 17A) (versus smooth carapace in Rochinia; Fig. 1A), and buccal frame with highly protruded blunt angle (Fig.  17B) (versus lack of protruded angle on buccal frame in Rochinia; Fig. 1B).
There are studies that report the prezoea and first zoea stage morphology for A. rissoana (Guerao and Abelló 1996), as well as the size of the species at sexual maturity (Mura et al. 2005). This species has been noted to be associated with bamboo coral, Isidella elongata (Esper, 1788) (see Maynou and Cartes 2012;Cartes et al. 2013;Mastrototaro et al. 2017), and black coral, L. glaberrima (see Bo et al. 2015), and typically with the hydroid epibiont, Rosalinda incrustans (Kramp, 1947), on its carapace (Kramp 1947;Vervoort 1966;Mastrototaro et al. 2016). This species was also recently observed to have R. incrustans on its pseudorostral spines and climbing on colonies of I. elongata to catch small prey (Mastrototaro et al. 2017: 217, figs 5j, k, 6j-n).