TWO NEW SPECIES OF SEPIA LINNAEUS, 1758 (CEPHALOPODA: SEPIIDAE) FROM SOUTH AFRICAN WATERS

A new species of cuttlefish: Sepia barosei sp. nov. is described from South Africa. Its present distribution is limited to the eastern Agulhas Bank (depth range 119–184 m), but this was based on a small material. It is recognised by prominently dorsal eyes and four turrets (papillae on top of each other) dorsally on the head between the eyes. On the dorsal mantle: four large, prominent patches, two mid anteriorly and two mid posteriorly; smaller turrets, patches and/or tubercles and warts dorsally; skin between these structures smooth and shiny. Ends of I pair of arms without suckers. Cuttlebone lightly calcified, thin and fragile. Second new species, Sepia roeleveldi sp. nov. is also described (distribution: from Hondeklip to Port Alfred, depth range 181–497 m), and is recognised by small, flat warts very densely and uniformly covering dorsal mantle, head and arms. Two prominent tubercles covered by warts on dorsal head; one similar tubercle in the middle of dorsal mantle (sometimes missing). Ends of I pair of arms without suckers. Cuttlebone thin and fragile, thinly calcified. Description of these new small cuttlefish has led to proposing a new subgenus Digitosepia with four species. Both holotypes and some paratypes are deposited in the South African Institute of Aquatic Biodiversity (SAIAB) and other paratypes in the Iziko Museum (SAMC) and the Natural History Museum (NHMUK), London. key words: small cuttlefish; new species; new subgenus; Digitosepia; South African waters; Sepia


INTRODUCTION
Previous contribution (Lipinski & LesLie 2018) concerning small size Sepiidae from the southern Africa, has briefly outlined various problems with more than one hundred species of the genus Sepia Linnaues, 1758. These problems were already well covered by Adam & Rees (1966) in their monograph, and then by subsequent authors; small sepiids add to these problems. Steenstrup (1875) was the first to recognise a peculiarity of small sepiid (Sepia typica (Steenstrup, 1875)) for which he created a separate genus, Hemisepius. However, subsequent additions to small sepiids made this genus unsustainable without providing much ideas of how small sepiids should fit into the overall classification. The hope was expressed (Roeleveld & LiLtved 1985) that more new species and more material of species already de-scribed (including Sepia typica, original small sepiid) would ultimately clarify this situation. However, so far, each additional species multiplied questions instead of providing answers (which is apparent in Lipinski & LesLie (2018) contribution); remaining hope was in describing at once several small sepiid species. Fortunately, intense and long-term sampling in southern African waters conducted mainly by RS Africana (Sea Fisheries, Cape Town) and R/V Dr Fridtjof Nansen (Institute of Marine Research, Bergen and Food and Agriculture Organisation, Rome) has accumulated large collections of undescribed species. Among them are small sepiids. Two of them are being described now. With this material at least some sketchy explanations are possible and a road ahead can be charted.

MATERIAL AND METHODS
Methods of this study were already outlined by Lipinski & LesLie (2018) and reader is being referred to that work. Some details however were repeated here for clarity. Some changes and clarifications were also suggested and followed.
Specimens for this study were collected using bottom trawls during the course of demersal research surveys conducted off the west and south coasts of South Africa by the research vessels RS Africana and R/V Dr Fridtjof Nansen (Fig. 1). Details of bottom trawl gear, trawling, sorting and processing of the catch, and references thereof are given in axeLsen & Johnsen (2014), although note that they erroneously record the codend mesh size of the gear deployed by the RS Africana, the correct mesh sizes are 112 mm codend with 35 mm small mesh liner.
Measurements and counts (see Table 1 for definitions) follow Roeleveld (1972), Roeleveld & LiLtved (1985), and Lu & Reid (1997) and were taken from preserved specimens. The following changes to the terms and definitions given in Table 1 of Lipinski & LesLie (2018) were made: 1. There is a vague difference between CIRC and TrRC; therefore, TrRC is renamed LORC and defined as a number of suckers counted at midline along the length (Tcl) of the club.
2. CES was imprecisely defined and not easy to follow. It is defined now as a total number of suckers along the border of the whole club. 3. Widest level of the cuttlebone (WL) is added to the measurements, defined as the greatest width of the cuttlebone (perpendicular to L). 4. TIP1AL is defined as length of the suckerless tip of I pair of arms. Dorsal (ML) and ventral (MLv) mantle length were measured to the nearest mm below using slide callipers. Fin length (FL) was measured by placing a thread along the base of the fin from the anterior edge and marking the position of the posterior end of the fin, the length of the thread was then measured on a metal ruler. All other measurements were taken using dividers or a graticule in a stereo dissecting microscope at 10× magnification. Sucker diameters were measured at 40× magnification. Weights were taken from preserved specimens. Maturities were determined using the scale described by Lipinski & underhiLL (1995).
Bodies of six species of small sepiids in southern African waters have various complex skin patterns that are important for correct identification of the species. Lipinski & LesLie (2018)   was made. Only two attempts to extract cuttlebone were made, resulting in considerable damage to this extremely fragile structure. Present material was accumulated during demersal surveys. These surveys are labour intensive and there is no time for immediate dissections. Therefore, most specimens are preserved initially in 10% buffered formalin with ample space in the jar, and after variable time (few weeks to few months from the moment of preservation) are transferred to 4% formalin, and then after 24h in water, gradually transferred to 70% ethyl alcohol. This procedure is sufficient to preserve most structures (even the cuttlebones of large Sepiidae) but insufficient for small sepiids. Some of their cuttlebones are completely de-calcified, however, this is extremely variable for an unknown reason. There are well preserved cuttlebones of small sepiids after months in formalin. The best way to avoid problems is to preserve fresh specimens in weak ethyl alcohol (<50%) only for cuttlebone dissection. This implies however that there must be enough specimens to preserve in formalin for morphological and ecological research.
Diagnosis. Sepiids with a substantially modified cuttlebone. Modifications: inner cone reduced to thin band, fused completely to the outer cone (flat); striae moderately or strongly convex. Tips of I p. arms completely devoid of suckers; they may be thin, whip-like, or thick, finger-like with a considerable variation between. Type species: Sepia (Digitosepia subgen. nov.) robsoni (Massy, 1927) Sepia (Digitosepia) barosei sp. nov.  Diagnosis. Cuttlebone lightly calcified, thin and fragile; anterior part (~25% of length) triangular, rounded; posterior part broad, oval, rounded posteriorly; spine absent; last septum elevated; no mid-dor-sal longitudinal ridge; phragmocone well defined but very thin, striae borderline slightly convex, other striae convex; inner cone thin, extends anteriorly into striated zone, ventral part reduced; outer cone very broad, width decreasing anteriorly into upper limbs that end close to anterior tip of cuttlebone. Head with four distinct and large turrets dorsally between the eyes (proximal pair larger). Smaller turrets and tubercles anteriorly to the dorsal, large eyes. Dorsal mantle usually with two large patches in the anterior half of the mantle, on each side of the midline. Mantle covered with irregularly placed tubercles and few warts; skin between clusters, tubercles and warts smooth, shiny. Tips of I pair of arms devoid of suckers. Description. Small species; only available male (mature) 17 mm ML, females (maturity III-V) 15-24 mm ML (Table 2). Mantle rather elongated, oval (rounder in smaller animals), dorso-anterior margin wide Ʌ-shape (Figs 2-11), ventro-anterior margin emarginated (inverted trapezoid) in both sexes (Figs 4-7). Ventral margins of mantle with distinct keels (Figs 4-5). Fins narrow, ending well before anterior mantle margin (FIa 19-23%; Figs 6-7); there is small gap between fins posteriorly. Colour of dorsal mantle and head of preserved specimens variable, dark grey or reddish brown (Figs 2-11). Skin covered dorsally by various structures (papillae, tubercles and warts); smooth, shiny between structures (Figs 2-11). Skin sculpture species-specific, although quite variable.
Head width equal to mantle opening width, appears elongated; eyes very large, dorsal (as opposed to lateral or dorso-lateral eyes of most sepiids); neck wide (Figs 2-11). Tentacle pouch large and deep. Buccal membrane without suckers. Four very prominent, long and complicated turrets dorsally on head between eyes, proximal pair larger (Figs 10-11); transverse row of three tubercles between the eyes in occipital region; 2-3 transverse tubercle rows dorsally on head between the eyes and arm bases (some rows may be difficult to see). Three to four small tur- rets immediately anterior to the eyes. One to three small tubercles at the bases of arms I-III. Three-four small tubercles in a row on each side of the head (Figs 6-7, 10-11). Two large patches dorsally on each side of the midline in about one-third of mantle (anteriorly) and two others (smaller) in about one-third posteriorly (Figs 2, 8-9). Anterior part of the dorsal mantle covered with tubercles in a regular pattern: near the margin, nine small tubercles in a convex row, next convex row comprised of five larger tubercles (median largest), and the last convex row comprised of eight tubercles of variable size. Posterior part of the mantle also sparsely covered in rather irregularly placed tubercles and a few warts (Figs 2, 8-9). Arms robust, stout, variable in length (Figs 12-13), relatively short, strong membrane joining pair I proximally for about 20-22% of arm length, membrane becoming gradually shallower between subsequent arms, absent ventrally between arms IV. Protective membranes well developed, fleshy . Suckers globose, biserial on all arms; sucker rows widely spaced in both sexes . Sucker rings on club and arms smooth (Fig. 14). Tips of arms of I pair devoid of suckers; these tips are quite variable and may be of different length on each arm (range 12-25% of I pair arm length) (Figs 15-16 Then there are 4 pairs of enlarged suckers with fine dentitions; hectocotylus tip with biserial small suckers (Fig. 18); fleshy transverse folds running between small marginal sucker pairs (Fig. 17). Right ventral arm also modified (Fig. 19), sucker arrangement from base to tip: eleven pairs of normal biserial suckers (not enlarged); 4 rows of enlarged biserial suckers; 6-10 rows of tiny biserial suckers to tip. Tentacular stalk moderately long (90-133% ML), club moderate (12-21% ML) (Figs 20-21) with subequal small suckers in 15-16 transverse rows of 6-8 suckers each. Protective membranes very narrow, well separated. Natatory membrane well developed, continuing along tentacular stalk for about 1.5 of club length. Club appears not as a continuation of a tentacular stalk, but perpendicular to it (Fig. 20).
Beaks small, fragile, of typical sepiid proportions. Upper beak : rostrum blunt, relatively long, slightly hooked, length greater than width, rostrum angle very little curved; hood high above crest posteriorly; jaw edge straight, jaw angle nearly 90°; lateral wall posterior edge strongly curved; only rostrum and hood dark. Lower beak (Figs 24-26): rostrum short, blunt, jaw angle rounded, broad, >90°, cutting edge wavy; hood low on crest; crest slightly curved; crest and lower edge of lateral wall at broad angle; posterior edge of lateral wall rounded; only rostrum, anterior part of hood and anterior part of shoulders dark.
Radula simple homodont but lost in preparation and cannot be adequately illustrated.
Spermatophores (Fig. 14) do not appear to have species-specific characteristics (illustrated for the re- Locking cartilages: funnel component semi-oval, with internal margin almost straight (Fig. 27), groove not very deep, simple, without additional median cleft; mantle component simple, not prominent (Fig. 28).
Funnel with valve (Fig. 29). Funnel organ well defined: dorsal part with weak anterior ridge and papilla; arms relatively long, thick and fleshy; ventral part simple, elongated oval (Fig. 30).
Cuttlebone: lightly calcified, thin and extremely fragile; anterior 25% triangular and rounded; broad, oval and round posteriorly (Figs 31-35); spine absent. Last septum elevated, stronger calcified than others. Phragmocone well defined, striae borderline clearly convex. Inner cone thin, anteriorly extends  --not applicable * -not possible to measure/estimate *number -approximate into striated zone, ventral portion reduced. Outer cone extremely broad, very thin, its width decreasing anteriorly into upper limbs that end close to anterior tip of cuttlebone. This description is preliminary due to the serious damage to the available material during preparation; however, it allows to appreciate unique character of this structure. Remarks. In the small size range investigated this small species is sexually mature at 17 mm ML (both sexes). Juveniles have not been recorded. The holotype (Figs 2-7) and at least two of the paratypes are mature. Sepia barosei sp. nov. is distinguished from all other small sepiids described so far by the array of the following characters. The absence of ventral pores (present in S. typica), tips of I pair of arms devoid of suckers, which is easily seen despite the small size of the species (normal suckers to the end in S. pulchra Roeleveld et Liltved, 1985, S. shazae Lipinski et Leslie, 2018, S. dubia Adam et Rees, 1966, dorsal mantle and head full of various structures (smooth in S. robsoni (Massy, 1927)), dorsal skin neither covered in dense tubercles with no other structures (as in S. faurei Roeleveld, 1972) nor dense warts with no other structures except two large tubercles on the dorsal head and one on mantle (as in S. roeleveldi sp. nov.).
It is perceived that each of these species differs rather profoundly from all others by the cuttlebone characteristics. However, this needs to be investi-gated, described and illustrated by separate studies. Cuttlebones of all small sepiids are very difficult to dissect without damaging them because of their great fragility. Even if successful, this yields usually limited information because calcium carbonate component of these cuttlebones gets easily dissolved or damaged in most preservative solutions (and, it is suspected, may get damaged shortly after death of the animal in any conditions). The only solution is to dissect the cuttlebone immediately after capture, which is seldom possible. This is why cuttlebone descriptions provided in this paper must be taken with caution as preliminary. The picture of the cuttlebone of S. typica (Figs 36-37) is provided for comparison with S. barosei sp. nov. Interpretation of this comparison is that S. typica still has a proper phragmocone and almost straight striae of which first is the thickest, whereas in S. barosei the striae are much reduced (although the available material is seriously damaged). Distribution. S. barosei is probably endemic to South Africa, it is known only from an eastern Agulhas Bank on the bottom, depth range 119-184 m. Holotype of S. barosei was found in the same trawl as an individual of S. dubia, described by Lipinski & LesLie (2018 Diagnosis. Cuttlebone reduced: striae of the remnants of the phragmocone are extremely convex and appear mixed with layers of growth of the dorsal shield. Cuttlebone broad, especially outer cone; anterior oval; posterior spine absent. Inner cone distinct but completely fused (including limbs) with outer cone; together with posterior outer cone forms shallow spoon-shaped depression (shallow pocket). There is evidence (Fig. 70) of small posterior wings of outer cone.
Dorsal mantle, head and arms covered densely and completely with small flat warts, giving a scalelike appearance. Two long conical tubercles (also covered with these warts) on the head between and anteriorly to dorso-lateral eyes. Tips of I pair of arms devoid of suckers. Description. Small but relatively heavy, known at present from 17 females, ML 25-36 mm (mean 31 mm) and 8 males, ML 23-31 mm (mean 26 mm) (Figs 38-46, Table 3). Overall habitus: globose, compact, round, dorsal skin uniformly warty. Mantle broad oval, length to width ratio 1.2-1.4; antero-dorsal margin variable: flat or even concave, or wide "W" type; antero-ventral margin emarginated inverted trapezoid in both sexes, variation not detected . Fins robust and wide (up to 13% ML) ending well before anterior mantle margin (FIa 13-18%; Figs 44, 45); small gap posteriorly between fins. Colour of dorsal mantle, head and arms in preservative solution uniform, reddish brown. Dorsal mantle, head and arms very densely covered with small, flat warts, giving them a "scaly" appearance (Figs 38-41, 46, 72). However, these warts do not resemble scales at all (Figs 46, 72). In some individuals these warts are present on marginal parts of ventral mantle. Ventral mantle with distinct keels (Figs 42-44); in some cuttlefishes fleshy wrinkled tissue, on margins and posteriorly, forms an adhesive pad. There may be one low but broad tubercle in the middle of dorsal mantle (Fig. 41); it is missing in some specimens. Head robust, neck usually only slightly narrower than head (but see Fig. 40); there are two long, conical tubercles (also covered by warts; Figs 41, 46) anteriorly to the dorso-lateral eyes (towards the midline). Two similar smaller tubercles (Figs 41, 46) may be present in some specimens in line behind the first pair (posterior to the eyes).
Arms II-III pair are mostly subequal in length, arms IV are longest and I are shortest (based upon 25 individuals, eight of which were measured). Arms are relatively short, fleshy; keels especially prominent in ventral (IV) arms; on other arms their presence is variable (Figs 38-49) and depends upon the state of preservation and possibly other factors. Protective membranes thick, fleshy: suckers biserial (in males, they may appear tetraserial on arms II-III (Figs 48-49) because of their zigzag distribution. Arms connected by membrane; slightly less than 1/3 of arm length between arms I and I-II; strongest between arms II-III; no web between arms III-IV and IV. Suckers on arms relatively large; sucker rings with no teeth (Fig. 50). Tips of arms of I pair devoid of suckers; these tips are quite variable and may be of different length on each arm (range 4-17 % of arm length) (Figs 51-54 (Figs 48-50). Left ventral arm hectocotylised (Fig. 55): there are 13 pairs of medium-sized suckers in marginal rows. First three aboral edge suckers in a straight line and then in regular zigzag pattern. Oral edge suckers of same size, in a straight line. Then there are 7 pairs of enlarged suckers with long chitinous tubes (except proximal three suckers) instead of rings; hectocotylus tip with 3 pairs of smaller suckers; fleshy transverse folds running between smaller marginal sucker pairs. Right ventral arm also modified (Fig. 56), sucker arrangement from base to tip: thirteen pairs of enlarged biserial suckers (they appear tetraserial due to their zigzag distribution) with normal chitinous rings; then 10-11 rows of enlarged biserial suckers with tubular chitinous rings, gradually becoming smaller. Tentacular stalk long (171-212% ML). Club relatively large (16-21% ML), uniform with 13-18 rows of 4-5 suckers (Fig. 57). Suckers minute, chitinous rings mostly smooth but tiny indentation seen on some (Fig. 50). Protective membranes narrow. Natatory membrane well developed, continues along the tentacular stalk for about ½ of club length. Upper beak (Figs 58-59): rostrum blunt, relatively short, not hooked, length equal to width, rostrum angle well defined; hood long, distal tip far from crest, jaw edge straight, jaw angle <90°; lateral wall strongly curved. Almost all beak dark except distal lateral wall.
Lower beak (Figs 60-61): rostrum short, blunt, lacking distinct rostrum angle; hood low on crest, slightly curved; crest slightly curved, not indented, not parallel to lower edge of the lateral wall (proximally further apart than distally); lateral wall with curved and angled posterior margin, with fold developed well on upper part of lateral wall. All beak dark except wings and lower and distal part of lateral wall. Radula homodont, with seven teeth per row (Fig.  62). No marginal plates detected. Rachidian teeth low, broad, triangular, fairly symmetrical. First and second laterals similar, simple, symmetrical. Heels small, compact. Marginal teeth not uniform: most of them sharp but some are blunt, moderately curved, not indented. Spermatophores (Fig. 50) not species-distinct, but they are illustrated to aid possible future studies.
Funnel with valve (Fig. 65). Distinct vertical ribs under the valve and on the sides (Fig. 65). Funnel organ with dorsal component well defined, bearing anterior ridge and papilla, limbs short: ventral component well defined, oval, compact (Fig. 66).
Cuttlebone much reduced: only thinly calcified dorsal shield remains with extremely convex layers of deposition and growth. Anterior is tapering but rounded, whole appearance oval and broad. There is no spine. Inner cone completely fused with outer cone. Outer cone very broad, thin, with small posterior wings and shallow pocket (Figs 67-71). Middorsally, there is indistinct ridge. Small posterior wings present (Fig. 70). S. roeleveldi sp. nov. is distinguished from all other small sepiids by the flat, very dense warts all over the dorsal body and the two high tubercles on the dorsal head (also covered by these warts).
Mature male (ML 25 mm, SAIAB209558b) shows peculiar sac-like structure storing spermatophores inside, which is anchored on the left anterior part of the nuchal cartilage and well visible between head and mantle (Fig. 72).
Spermatophores are stored on a buccal membrane in fertilized female (Fig. 73). Distribution. Currently known from the large area of South African seas, from Port Nolloth on the west coast, to Port Alfred on the south coast (Fig. 1). Depth range is wide: 181-497 m, average depth (arithmetic mean) 312 m. Therefore, it is a deep shelf and slope bottom-dwelling species. Etymology. This species is named to honour Ms Martina Adriana Compagno Roeleveld (see aLLcock et al. 2015), sepiid specialist, cephalopod biologist, friend of many years, and discoverer in her collection at SAMS (together with D. N. Khromov) of the species named now after her.   There are now eight such species, and more are expected to be described. They are all mature (both sexes) at ML not larger than 40 mm, and this is considered here as "small Sepia".

DISCUSSION
The small sepiids status and their systematic position (or lack thereof) were discussed at length by Lipinski & LesLie (2018); an interested reader is referred to that paper. However, the two new species described here, re-description of S. robsoni (Massy, 1927) and S. faurei Roeleveld, 1972 (Lipinski in prep.) and re-evaluation of S. typica (Lipinski & LesLie, in prep.) offer some leads and suggestions that small sepiids should be given a generic status as a group. This new status should be based upon a thorough study of their cuttlebones which are all different from the other sepiids, accompanied by study upon sucker morphology and their arrangement on all arms. However, there is no material at the moment to conduct such a study. Because of fragility of cuttlebones of small sepiids and imminent damage during any sort of preservation, collection of such material presents serious challenges. It is hoped that after collection of suitable material, the following questions may be addressed: 1. Which structures are present on the dorsal shield of the cuttlebone, on its dorsal and ventral side? 2. What is the hierarchy of calcification of dorsal shield, expressed by various species of small sepiids?
Such structural study should be accompanied by genetic and molecular biology research.
However, without waiting for the outcome, certain groupings can now be defined inside this new hypothetical genus of all small sepiids. These groupings, following a tradition of sub-genera in this group, should include subgenus Sepia (S. dubia, S. pulchra and S. shazae), subgenus Hemisepius (S. typica), and the new subgenus Digitosepia (S. robsoni, S. faurei, S. roeleveldi and S. barosei). First two subgenera are well defined and described in various papers. Digitosepia was already recognized by Lu (1998) and listed as his Group 5, which was based on the characteristics provided by Roeleveld (1972, also group 5, p. 269).
This arrangement is temporary as a sub-generic status reflects rather a lack of more specific evolutionary understanding than anything else. Until thorough morphological and molecular biology study is completed, this temporary status quo has to guide students of Sepiidae. This is because there were many failed attempts to revise sepiids, and to create new genera and subgenera. Ample warning is provided by Adam & Rees (1966) in their monograph.