QH 1 S67X NH leg eg alley? THE VALIDITY OF RAZA RHIZACANTHUS REGAN AND RAZA PULLOPUNCTATA SMITH, BASED ON A STUDY OF THE CLASPER November 1966 November Volume 48 Band Part 20 &Deel ANNALS OF THE SOUTH AFRICAN MUSEUM ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM Cape Town Kaapstad The ANNALS OF THE SOUTH AFRICAN MUSEUM are issued in parts at irregular intervals as material becomes available Obtainable from the South African Museum, P.O. Box 61, Cape Town (Cash with order, post free) Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM word uitgegee in dele op ongereelde tye na beskikbaarheid van stof Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad (Kontant met bestelling, posvry) OUT OF PRINT/UIT DRUK I, (1, 3> 5: 8) 3 (1-2, 5> Goes 5(2, 5> 7-9), 6(1, E=p-1)5 7(1, 3) 8, g(1-2), 10(I-3), i Giy 7> tile) 2i, 24(2), 31(1-3), 44(4). Price of this part/Prys van hierdie deel 40c Printed in South Africa by In Suid-Afrika gedruk deur The Rustica Press Pty., Ltd., Die Rustica-pers Edms., Bpk., Court Road, Wynberg, Cape Courtweg, Wynberg, Kaap THE VALIDITY OF RAJA RHIZACANTHUS REGAN AND RAJA PULLOPUNCTATA SMITH, BASED ON A STUDY OF THE CLASPER By P. A. HuLLey South African Museum, Cape Town (With 8 figures in the text) CONTENTS PAGE Introduction SD Rea ace at DRM aD a oiAue AO The Claspers of Myxopterygia . : -. 4098 Clasper of R. rhizacanthus (i) the external anatomy . : - 499 (ii) the internalanatomy . . . 500 Clasper of R. pullopunctata (i) the external anatomy . : - 505 (ii) the internal anatomy . . . 505 Discussion . ; F P : ; . 508 Acknowledgements. : Sect err see ey TS Summary . : F : : FOS fa TA IVELETENCES Mitsm me MPR tk tacts Bul ie stint INTRODUCTION The present classification of South African skates is more or less tentative, for this difficult genus shows great variability in its morphological characters, which vary with age and/or sex in the individuals within the species. This high intraspecific variability probably accounts for the present inclusion of 95-100 species in the genus Raja. Several South African species have, as yet, proved to be indistinguishable from European species from the corresponding latitudinal belt in the northern hemisphere (Bigelow & Schroeder, 1953), and it would appear that these pairs represent cases of ‘masked bipolarity’ (Ekman, 1953). But in very few cases have actual comparisons of the specimens from the two regions been made, although Norman (1935) has given some brief references. Raa alba, Lacépéde, from the N.E. Atlantic is reported here under the same name, but it appears to be cosmopolitan in distribution, being recorded from Morocco (Pietschmann, 1906), West Africa (Fowler, 1936) and Angola (da Franca, 1959). 497 Ann. S. Afr. Mus. 48 (20), 1966: 497-514, 8 figs. 498 ANNALS OF THE SOUTH AFRICAN MUSEUM Norman expressed some doubt about the identification of R. batis from South Africa but the material he examined included a single stuffed specimen, on which accurate measurements were impossible. This South African species is now thought to be distinct from R. batis Linnaeus, and Smith (1964), on the basis of the shape, the presence of a single median spine and the general markings, has reclassified this skate as R. pullopunctata, although he points out that only juvenile material was examined. Raja rhizacanthus Regan, has been thought to be identical with the Euro- pean thorn-back skate R. clavata Linné, but this has been a matter of speculation. Barnard (1925) and Fowler (1940) have included R. rhizacanthus as a synonym of R. clavata, but Norman regarded the two as distinct species on the basis of spination and the position of the vent. This has been followed by Smith (1961). It is the purpose of this paper to compare and contrast certain morpho- logical characters, in the hope that the validity of these latter two South African species may be determined with greater certainty. Since skates show such great variations in external features, a more critical method of examination needs to be employed. Leigh-Sharpe (1920- 1926) has pointed out that in the genus Raja, individual species can be recog- nized by the morphological structures of their claspers. He has extended this further and has postulated the existence of pseudogenera (termed Alpha-, Beta- and Gamma-raia, etc.), recognizable by the similar presence or absence of the clasper elements. Ishiyama (1958) has shown that in Japanese rajids, although intraspecific variations in the clasper can be recognized, the ‘external and internal structures of the clasper are species specific without exception’ (p. 224). On this basis he has also been able to postulate phyletic, generic and inter-specific relationships, and has concluded that ‘the characters in the male external organ (claspers) of the Japanese rajids seem to give the most reliable basis for the systematics of this group of fish’. (p. 243.) Hence it would appear that since the anatomy of the clasper is valid for the purposes of identification, it should also be valid for the purposes of com- parison of geographically separated species. However, ‘genitalic differences must be evaluated just like other characters’ (Mayr et al., 1953, p. 109), and so differences in the external and measurable characters of the specimens must also be taken into account. THe CLASPERS OR MyxXOPTERYGIA Comprehensive anatomical studies have been carried out on the claspers of European elasmobranchs by Petri (1878), Jungersen (1899) and Huber (1901). Two points of view were held as to the function of the claspers: that they were ‘holders’ (analogous to hands) was originally postulated, but they were later recognized as organs of intromission. Extensive study on the external anatomy and the function of the clasper in the various elasmobranch groups has been made by Leigh-Sharpe, and further work on the function of the organ has been performed by Friedman (1935). RAJA RHIZACANTHUS REGAN AND RAJA PULLOPUNCTATA SMITH 499 The nomenclature used in this paper is mainly in accordance with that of Leigh-Sharpe and Ishiyama, together with that of the above-mentioned authors. THE CLASPER OF R. RAIZACANTHUS REGAN The claspers from seven adult specimens of R. rhizacanthus, trawled off Cape Columbine and in False Bay, were examined. apopyle hypopyle slit signal shield Sw90-OL pseudorhipidion pseudosiphon sentinel A f sentina | 2:0cms | spike Fic. 1. Raja rhizacanthus A, external view of right clasper from dorsal side; B, lateral view of right clasper, gngacd to show structural features of glans. (i) The external anatomy of the clasper (fig. 1 A, B) The clasper of R. rhizacanthus is a moderately long, strongly built structure, the outer surface of which is entirely naked. Proximally the clasper is cylin- drical, but from about 1/3 its length from the distal end, it becomes broader and dorso-ventrally flattened, especially on its ventral surface. It tapers gradually towards its distal end and terminates in a broadly rounded, spatulate 500 ANNALS OF THE SOUTH AFRICAN MUSEUM tip. The ventral surface is pale and the dorsal surface tones with the general coloration of the body. The apopyle (fig. 1 A) is situated dorsally, some little distance behind the vent, and the appendix groove, arising from this point, runs across to the outer margin of the organ, and continues along the distal two-thirds of the lateral edge almost to the tip of the clasper. In some cases, due to the skin stretching across the dorsal terminal 1 cartilage, the groove may be seen from the dorsal side. Proximally the groove is bordered on by the two ‘scroll-like’ dorsal and ventral marginal cartilages, so that the edges of the groove closely approximate to form a tube, running from the apopyle to the hypopyle. Distally, from about half the clasper length, fleshy lips border the groove, and during erection, brought about by the powerful adductor muscles, and by vascular dilation (Friedman), these lips are opened and the structures of the clasper glans protrude (fig. 1 B). Two pouches are present on the dorsal surface of the concavity of the clasper glans (fig. 1 B), the smaller, more proximal slit and the larger, distal pseudosiphon. The median hypopyle is bordered dorsally by the pseudo- rhipidion, a plate-like structure employed as a splash-plate for ejaculating spermatozoa and distinguished from the rhipidion of Leigh-Sharpe in not being fan-shaped or pitted. Posterior to the hypopyle, and at about the same level as the pseudorhipidion, is a fleshy pad, the szgnal, which is capable of rotation about the longitudinal axis of the clasper. The sentinel is strongly developed and in older specimens its knife-like edge is exposed. There is a deep sentina situated distally, its inner border raised into a curved spike. The ventral shield is narrow and not well developed. The lining of the concavity is naked, scale and spine being absent. All these above-mentioned structures are associated with the cartilages of the terminal group. (ii) Lhe internal anatomy of the clasper (figs. 2, 3) The skeleton of the clasper is a continuation, in the median axis, of the basipterygium of the fin, and is composed of two groups of cartilages, the basal group (including the basipterygium) and the terminal group. (a) The basal group (fig. 2) Ishiyama has pointed out, that in Japanese rajids, the basal group is not species specific, and, although White (1937) has proved that the number of proximal segments and other cartilages in the basal group tend to vary in large groups of the elasmobranchs, the basal group may only be used in phyletic considerations among the batoid fishes. Examination of the basal group of R. rhizacanthus seems to support this view, since the same seven cartilages are present as in the Japanese rajids (see Ishiyama, fig. 7). The basal group in R. rhizacanthus consists of the basipterygium, and b, and b, cartilages, the Beta-cartilage, and the axial cartilage, which extends RAJA RHIZACANTHUS REGAN AND RAJA PULLOPUNCTATA SMITH 5OI basipterygium RS WS fin rays . ‘ \™“ b, cartilage beta cartilage dorsal marginal axial l 4-0cms Fic. 2. Raja rhizacanthus. Cartilages of the basal group of the clasper. to the terminal part of the clasper, and is covered by the dorsal and ventral marginal cartilages (fig. 2). The basipterygium is a short, stout structure, to which are attached in a groove on its lateral margin, the last 6 fin rays of the pelvic fin. Distally, the basipterygium is elongated on its dorsal surface to form a protuberance which overlaps the b, cartilage. The proximal end of the b, cartilage is visible from the dorsal side, but the Beta-cartilage, situated dorsally to the b, and b, cartilages, completely obscures the b, cartilage and the distal end of b,. The Beta-cartilage is more or less flat, serving as a region of attachment for the m. flexor internus, but is slightly thickened and upraised at its proximal end, forming a shelf. The distal ends of the b, and Beta-cartilage are superimposed so that both of these elements play a part in the joint with the axial and marginal cartilages. The long axial cartilage is cyclindrical at its proximal end, but it becomes dorso-ventrally flattened distally and assumes a spatulate appearance. It is tightly enclosed by the calcified dorsal and ventral marginal cartilages and is only free at its spatulate tip, although it is visible dorsally. The dorsal marginal cartilage attaches tightly to the axial, starting where the axial is connected to the b,/Beta joint, and runs along the outer lateral margin of the axial for about half its length. The dorsal marginal cartilage is somewhat broadly pointed anteriorly but it becomes expanded distally, its outer lateral edge forming the dorsal edge of the ‘scrolled’ groove of the clasper. The ventral marginal carti- lage, unlike the dorsal marginal, starts posteriorly to the axial and is a flat, tongue-like structure. Its outer lateral margin forms the ventral lip of the 502 ANNALS OF THE SOUTH AFRICAN MUSEUM ‘scroll’ and medially it is attached to the axial, overlying it distally. The ventral marginal cartilage extends farther down the length of the clasper than the dorsal marginal cartilage. Attached distally to the axial and to the marginal cartilages, are the elements which comprise the terminal group of cartilages. (b) The terminal group The elements which comprise the terminal group of cartilages are asso- ciated with the structures of the glans, and may be divided into three types: ventral terminal, dorsal terminal and accessory cartilages. 1. Ventral terminal (figs. 3 A, B, G) The ventral terminal cartilage is attached to a groove in the distal end of the ventral marginal cartilage. It has a J-like shape, being by far the largest cartilage of the terminal group. It is strongly curved medially, and runs around the axial cartilage, so that it can be seen from the dorsal side. The upright portion of the ventral terminal cartilage is fairly well developed and there is a small shelf on its dorsal side, which is seen as the shield of the clasper glans. Distally, the ventral terminal is thick and somewhat pointed. 2. Dorsal terminal 1 (figs. 3 C, G) The dorsal terminal 1 cartilage (‘cover-piece’ of Jungersen) is situated on the dorsal side of the clasper, partially covering the dorsal terminal 2 cartilage, to which it is attached along the length of its longest margin. It is flatly rounded and curved proximally, so that it is elongated on its outer lateral margin. From this point, there is a shelf, running in a curve along the anterior margin, which serves as a point of insertion for the aponeurosis of the m. dilatator, which conse- quently covers the whole area of the cartilage. Distally, the cartilage is flat and its distal end is attached, together with the dorsal terminal 2 cartilage, to the axial cartilage. The dorsal terminal 1 cartilage is associated with the development of the pseudosiphon. 3. Dorsal terminal 2 (figs. 3 D, G) This curved element is inserted between the dorsal terminal 1 and the dorsal marginal/axial cartilages. It is firmly attached to the distal region of the dorsal marginal cartilage, and by means of a ligament (which forms the shelf between the slit and pseudosiphon) it is attached at its distal end to the outer lateral margin of the axial cartilage, at about 2/3 the length of that cartilage from its anterior end. The dorsal terminal 2 cartilage is strongly curved, and raised along the side of longest curvature, to form a flat surface ‘articulating’ with the dorsal terminal 1 cartilage. The dorsal terminal 2 cartilage is asso- ciated with the slit. 4. Accessory cartilages 1 and 2 (figs. 3 E, G) These two cartilages are strongly united and are attached to the distal end of the ventral marginal cartilage. The accessory cartilage 1 is the S-shaped RAJA RHIZACANTHUS REGAN AND RAJA PULLOPUNCTATA SMITH 503 axial dorsal marginal dorsal terminal 1 dorsal terminal 2 accessory 3 SW 0-01 accessory 1 ventral terminal accessory 2 | 2:0cms | E Fic. 3. Raja rhizacanthus. Cartilages of the terminal group of the clasper. A, ventral terminal (ventral view); B, ventral terminal (dorsal view); C, dorsal terminal 1 (dorsal view); D, dorsal terminal 2 (dorsal view); E, accessory cartilages 1 and 2 (dorsal (view); F, accessory cartilages 3 and 4 (ventral view); G, terminal cartilages in situ from dorsal side. element, and its knife-edged protruberance is the sentinel. Accessory cartilage 2 is elongate in the longitudinal axis of the clasper, and its distal end is recurved and hook-like, forming the spike of the clasper glans. 5. Accessory cartilages 3 and 4 (fig. 3 F, G) Accessory cartilages 3 and 4 form the signal of the glans and are situated on the ventral side of the clasper, between the upright portion of the ventral terminal cartilage and the accessories 1 and 2, The smaller proximal cartilage, 504. ANNALS OF THE SOUTH AFRICAN MUSEUM accessory 4 is attached in a groove formed between the accessory cartilages 1 and 2 and the ventral marginal cartilage; accessory 3 is movable on accessory 4 and its distal end is bluntly rounded. Musculature Only a brief mention of the musculature of R. rhzzacanthus needs be made for it is identical with that of R. clavata as given by Jungersen. M. dilatator Proximally this consists of a single muscle mass, but distally it seems to be split into dorsal and ventral muscles. Its origin is on the axial, b, and b, carti- lages, and it inserts as an aponeurosis on the shelf of the dorsal terminal 1 cartilage. The m. compressor sacct, m. adductor, m. flexor (not extensor as given in Junger- sen), and muscles of the fin are the same as for R. clavata. hypopyle rhipidion SUD 0-p pseudosiphon shield spike B Fic. 4. Raja pullopunctata A, external view of right clasper from dorsal side; B, lateral view of right clasper, opened to : show structural features of the glans. RAJA RHIZACANTHUS REGAN AND RAJA PULLOPUNCTATA SMITH 505 THE CLASPER OF R. PULLOPUNCTATA SMITH The claspers from three adult specimens of R. pullopunctata, trawled West of Cape Town and in Algoa Bay were examined. (i) The external anatomy of the clasper (fig. 4 A, B) The clasper of R. pullopunctata is long and stout; its cyclindrical body is dorso-ventrally flattened distally, and slightly truncate, giving a spatulate appearance. The glans is much broader than the rest of the clasper, due to the strong development of the shield, which can be clearly seen from the dorsal side. The claspers are entirely naked, scale and spine being absent. The apopyle (fig. 4 A) opens some distance away from the vent, on the dorsal side. The rhipidion (fig. 4 B) is well developed and fan-shaped and it is attached along the inner ventral wall of the clasper, from the region of the hypopyle to about half-way along the length of the glans. Its lower surface is pitted with cavities of irregular outline, and this spongy tissue is said by Leigh-Sharpe to be erectile and to act as a splash-plate for the ejaculating spermatozoa. R. pullopunctata may be placed in the pseudogenus Gamma-raia, for there is a reduction and crowding of the structural features of the glans. As in all Gamma-raia, the signal is absent, while the shield is very well developed and extends along the whole ventral surface of the glans, from the level of the hypopyle almost to the distal extremity of the clasper. The distal region of the shield is covered by pleated epithelia, but this is so thin at the outer lateral edge, that the knife-like appearance of the shield is not unlike the sentinel of R. rhizacanthus. Both the slit and the pseudosiphon are present in the dorsal wall of the concavity of the glans. In R. pullopunctata the pseudosiphon is larger and more distally placed than the slit, and these two cavities are separated by a prominent cartilaginous shelf. The pseudosiphon is well separated from the hypopyle by the elongated dorsal marginal cartilage. Medially placed is a large, single, tongue-like structure, which examination of the terminal cartilages reveals as the spike (cf. Leigh-Sharpe). Although covered by a thick layer of epithelial tissue, the calcified cartilage can be easily felt. The small accessory 1 cartilage is not greatly developed, and the sentinel is not a recognizable element in the external appearance of the clasper. (ii) The internal anatomy of the clasper (fig. 5) The basipterygium, b, and b, cartilages and the Beta-cartilage are the same as in R. rhizacanthus. The axial cartilage is long, cyclindrical at its proximal end, but narrower and dorso-ventrally flattened distally. The distal end of the axial is more heavily calcified than the proximal region, and is slightly curved outwards, terminating in a bluntly rounded point. The dorsal marginal carti- lage starts more proximally than the ventral marginal cartilage, and runs for about 2/3 the length of the axial, to which it is tightly bonded for its entire length. The dorsal marginal cartilage is convex at its proximal end, but becomes 506 ANNALS OF THE SOUTH AFRICAN MUSEUM axial dorsal marginal ventral marginal dorsal terminal 2 ventral terminal Fic. 5. Raja pullopunctata. Cartilages of the terminal group of the clasper. A, ventral terminal (dorsal view) ; B, ventral terminal (ventral view) ; C, dorsal terminal 1 (dorsal view); D, dorsal terminal 2 (dorsal view); E, accessory 1 (ventral view); F, accessory 2 (dorsal view); G, terminal cartilages in situ from dorsal side, but with dorsal terminal 1 removed. flatter distally and is extended in a small point, forming the shelf between the pseudosiphon and the hypopyle. The ventral marginal cartilage is flat and slightly expanded distally, so that it more or less overlaps the axial cartilage, to which it is united. Its distal end is truncate. 1. Ventral terminal cartilage (figs. 5 A, B, G) As in R. rhizacanthus, the ventral terminal is J-shaped, but its distal end is not so strongly developed. The cartilage is attached to the inner surface of the ‘scroll-like’ ventral marginal cartilage. It is locked to this cartilage by means of RAJA RHIZACANTHUS REGAN AND RAJA PULLOPUNCTATA SMITH 507 a protuberance, situated ventrally about 1/3 of the length of the ventral terminal from its proximal end. The protuberance engages with a corre- sponding point (the minute ‘sentinel’) of the accessory cartilage 1, hooking the two elements together. The ventral terminal cartilage does not extend to the tip of the clasper. The outer lateral edge of this cartilage is well developed and calcified, forming a convex lamina, the shield of the clasper glans. Distally, the apex of the ventral terminal is grooved, through which the distal regions of the axial and dorsal terminal 2 cartilages pass, before becoming dorso- ventrally flattened. The upright limb of the J-like ventral terminal is raised into a ridge on its dorsal surface. 2. Dorsal terminal 1 (fig. 5 C) This cartilage is situated on the dorsal surface of the clasper, and is attached along its outer edge to the dorsal terminal 2 cartilage. The dorsal terminal 1 cartilage is ‘shield-like’ in appearance and its proximal end is elongated into an obtuse point, to which the dorsal bundle of the m. dilatator attaches. The inner distal region of this cartilage is so curved that it wraps around the axial stem and can be seen on the ventral side, where it slightly overlaps the upright limb of the ventral terminal cartilage. Distally, the dorsal terminal I is attached by connective tissue to the axial, dorsal terminal 2 and accessory cartilages. 3. Dorsal terminal 2 (fig. 5 D) The dorsal terminal 2 cartilage is attached to the inner distal end of the dorsal marginal cartilage, so that the point and connective tissue, which forms the shelf between the pseudosiphon and hypopyle, lie behind the dorsal terminal 1. The dorsal terminal 2 cartilage is a long, thin, single cartilaginous element, curving and then re-curving, so that its pointed distal end is attached to the axial cartilage (fig. 5 G). Proximally, the dorsal terminal 2 is strongly calcified and has a short blunt protuberance on its outer lateral margin, while distally calcification is not so marked, the soft cartilage becoming dorso-ventrally flattened. About half-way along its length, on the inner edge, there is a liga- ment (forming the shelf between the pseudosiphon and the slit), which arches and inserts on the axial cartilage about 1/5 the length of this cartilage from its distal end (fig. 5 G). 4. Accessory cartilage 1 (fig. 5 E) This minute V-shaped cartilage is attached to the distal end of the ventral marginal cartilage. The longer of its two arms runs forward along the outer lateral edge of the ventral marginal, while the shorter traverses the distal end of the ventral marginal cartilage. The base of the V is slightly pointed (the minute ‘sentinel’) and is curved inwards, forming the point on which the ventral terminal engages. It must be noted that this cartilage does not manifest itself as an external structural element in the glans. 508 ANNALS OF THE SOUTH AFRICAN MUSEUM 5. Accessory cartilage 2 (fig. 5 F) The accessory cartilage 2 resembles an inverted Y. Its cyclindrical proxi- mal end is attached to the inner distal end of the ventral marginal cartilage. The longer limb of the accessory is tightly bonded to the axial, so that distally these two cartilages resemble a single element. The shorter arm of the cartilage is simple in shape and projects laterally outwards. Its terminal end is slightly upturned and forms the spike of the clasper glans. Musculature The musculature of R. pullopunctata was found to be identical with the description of the musculature of R. clavata given by Jungersen, and like R. batis, the m. dilatator is in two bundles. (a) Dorsal m. dilatator This muscle has its origin on the proximal region of the axial cartilage and its insertion on the dorsal and lateral regions of the dorsal terminal 1 cartilage. The muscle is not strictly dorsal in position but more medio-dorsal, and is much larger than the ventral m. dilatator. (b) Ventral m. dilatator The muscle is entirely separate from its dorsal counterpart and has its origin on the ventral proximal region of the axial cartilage and the distal end of the b, cartilage. It inserts as connective tissue across the surface of the ventral terminal cartilage, although some fibres attach to the tip of the axial cartilage. The m. compressor sacci, m. adductor, m. flexer and the muscles of the fin are the same as for R. clavata. DIscussION Two pairs of claspers from specimens of R. batts Linnaeus, caught $.W. of Ireland, and a pair of claspers from R. clavata Linnaeus, caught at Lowestoft, were examined for purposes of comparison. Although there is a marked difference in the size of the claspers of R. pullopunctata and R. batis, the general external appearance of the claspers is similar. In both, the clasper is long, stout and naked, becoming dorso-ventrally flattened and spatulate at its terminal end. The outer lateral edge of the shield can be seen from the dorsal side in both species, but in R. batis it is more promi- nent. This may be due to differences in the size of the claspers. However, examination of the clasper glans reveals that the structural features in R. pullopunctata and R. batis are different (figs. 4 B, 6). The pseudosiphon and slit are present on the inner dorsal surface of the glans in both species, but their formation is somewhat different, although they are associated with the degree of development of the dorsal terminal cartilages in both cases. In R. pullopunctata there is a marked shelf-like ridge separating RAJA RHIZACANTHUS REGAN AND RAJA PULLOPUNCTATA SMITH 509 shield slit hypopyle rhipidion sentinel we “™ spike pseudesiphon | 4-0cms | Fic. 6. Raja batis. Lateral view of right clasper, opened to show structural features of the glans. the pseudosiphon from the slit; this shelf is found to be wanting in R. batis. The pseudosiphon is similarly situated in the two species, but in R. pullopunctata the pseudosiphon resembles a concavity rather than the simple pouch found in R. batis. The slit is more proximally situated in R. pullopunctata, occurring about half-way along the length of the glans (about one-third the length in R. batis). The hypopyle seems to be more distally placed in R. batis, but this may be due to the heavier development of the shield. The rhipidion is pitted and fan- shaped in both species, occurring along the inner ventral wall of the hypopyle, and extending about half-way along the length of the glans in R. pullopunctata and about two-thirds the length in R. batis. The rhipidion is more compact in R. batis. The prominent shield extends along the whole inner ventral surface of the glans in both species, but it is thinner and covered with pleated epithelia 510 ANNALS OF THE SOUTH AFRICAN MUSEUM along its entire length in R. batts, while in R. pullopunctata the broader shield is only pleated at its distal region. The most noticeable difference in the structure of the claspers is in the presence of the enlarged sentinel in R. bates, which extends distally and fits together with the spike. In R. pullopunctata, the sentinel is reduced and lies under the rhipidion, so that it is not seen when the glans is opened, and these two elements are not expanded terminally as they are in R. batis. As has been pointed out, the visible external characters of the clasper are dorsal terminal 2 axial dorsal terminal 3 Fic. 7. Raya batis. Cartilages of the terminal group. A, dorsal terminal 1 (dorsal view); B, ventral terminal (dorsal view) ; C, accessory 1 (dorsal view); D, accessory 2 (dorsal view); E, dorsal terminal 2 and 3 in position. RAJA RHIZACANTHUS REGAN AND RAJA PULLOPUNCTATA SMITH bel closely associated with the structure and number of the internal cartilages, and therefore differences in the structure of the glans will be shown by differ- ences in the terminal cartilages. In R. pullopunctata there are a total of five cartilages in the terminal group (2 dorsal terminals, 1 ventral terminal and 2 accessory terminals), while in R. bats six terminal cartilages are present (3 dorsal terminals, 1 ventral terminal and 2 accessory terminals) (figs. 5, 7). Although there is a great difference in the size of these elements, their differences in shape and their position may be examined. The dorsal terminal 1 cartilage is similarly situated in both R. batis and R. pullopunctata, occurring on the dorsal side of the clasper, and it is flat and plate-like, being slightly convex distally. In both species, the median edge is bent in such a manner, that the cartilage wraps round the axial stem and may be seen on the ventral side. Distally, the cartilage is pointed in the two species, but in R. pullopunctata, the proximal edge is elongated into a point for the attachment of the aponeurosis of the m. dilatator, while in R. batis (fig. 7), the proximal edge is almost straight and somewhat more convex. The dorsal terminal 2 cartilage is attached to the distal end of the dorsal marginal cartilage in both species, and it is interesting to note that in R. pullopunctata, the dorsal terminal 2 extends to the tip of the axial cartilage, while in R. batis, the crescent-shaped dorsal terminal 2 (fig. 7 D, E) is attached to about half way along the axial and that the dorsal terminal 3 extends to the tip of the clasper, i.e. these two cartilages in R. batis have been reduced to a single cartilage in R. pullopunctata, although there is linking between the dorsal terminal 2 and the axial in both cases. The terminal end of the axial cartilage is similarly curved (laterally outwards) in the two species, and the attachment of the dorsal terminal 2 of R. pullopunctata and the dorsal terminal 3 of R. batis (fig. 7 E) is similar. Although both R. pullopunctata and R. batis possess two accessory cartilages, the accessory cartilage 1 (forming the sentinel) is reduced to a small point in R. pullopunctata. In R. batis this cartilage (fig. 7 C) is well developed and extends down towards the spike, and fits closely to it. The ventral terminal cartilages are similarly situated in both R. batis and R. pullopunctata and these elements are similar in shape. Both possess the ‘hooking’ protuberance on the inner ventral margin, the enlarged convex outer lateral edge and the raised dorsal ridge. However in R. bats (fig. 7 B) the distal end of the convex edge projects slightly from the main body of the limb, and the distal end of the ventral terminal is not so markedly grooved as in R. pullopunctata. It may therefore be concluded that, in the structure of the glans and in the number and arrangement of the internal cartilaginous elements, the claspers of R. pullopunctata and R. batis are different. Besides differences noted in the structure of the claspers of the two species, some differences can also be seen in the external characters. Smith (1964) points out that there is a difference in shape between the two, but examination of later and adult stages of R. pullopunctata reveals that the shape of the disc varies and that the pectorals and anterior margins of the disc become more 512 ANNALS OF THE SOUTH AFRICAN MUSEUM concave, as they do in R. batis. The author is unable to detect any appreciable difference in shape. Examination of specimens of comparable size indicates that the inter- orbital is narrower in R. pullopunctata than in R. batis, and that there is little difference in the comparative size of the eye. The most noticeable difference is the presence of the large, single, median, nuchal spine in R. pullopunctata, which is found both in juveniles and adults (although sometimes worn down). Except about the eyes, the upper surface of the disc is entirely devoid of other large spines, as it is in R. batts. Therefore Smith’s proposal that R. pullopunctata and R. batis are distinct species as based on the general external morphology, is confirmed on the basis of the clasper structure. Recently, two large ‘black-bellied’ skates, lacking signs of a median nuchal spine, were obtained in trawls off Cape Columbine in 250 fathoms. The external characters of the clasper of the mature male were seen to be identical with those of R. batis from European waters. Therefore R. batis Linnaeus also may occur off South Africa; however, further investi- gation on these specimens is necessary. It is interesting to note that a ‘black-bellied’ skate from the Falkland Islands, R. flavirostris Philippi 1892, greatly resembles the South African R. pullopunctata in possessing a single, large, median, nuchal spine. But as yet, clasper material is unavailable for the purposes of comparison. The author was unable to find a single difference, both in the external structure of the glans and in the number and arrangement of the internal cartilaginous elements, between the claspers of R. clavata and R. rhizacanthus. The claspers in these two species are identical. Although Norman points out that the South African species is closely related to the European species, he considers that the two are distinct on the basis of the degree of spination in adults, the position of the vent and the shape of the snout. Preliminary examination of some two hundred specimens of R. rhizacanthus has shown that the shape of the snout varies considerably in local populations, as does the degree of spination, and the position of the vent. Therefore bearing this in mind, and in view of the fact that the claspers are identical, the author agrees with Barnard and Fowler that R. rhizacanthus specimens should be referred to R. clavata and that R. rhizacanthus therefore becomes a synonym of R. clavata. This is further supported by the fact that masked bipolarity is already acknowledged in R. alba, so that it is not unlikely that distribution of this type also occurs in R. clavata. It may be seen from fig. 8 that the distribution of R. clavata ranges from Arctic waters to the tropical waters of the Indian Ocean, having been recorded from greater extremes than R. alba. However, through most of their recorded range both species occur in the same waters. As yet, these two species have never been recorded from the equatorial waters of West Africa, but equatorial submergence probably occurs, allowing gene flow from northern hemisphere to southern hemisphere and vice versa. RAJA RHIZACANTHUS REGAN AND RAJA PULLOPUNCTATA SMITH 513 & CL aw ry ay N A- @ A:.- © 6 A &@ @R. alba ge 49 AR. clavata @t Fic. 8. Map showing the masked bipolar distribution of R. alba and R. clavata. ACKNOWLEDGEMENTS I am grateful to Messrs. M. J. Holden and C. H. Humphries of the Fisheries Laboratory at Lowestoft for the R. clavata and R. batis material and to Mr. S. X. Kannemeyer for assistance in collection of the South African skates. My thanks are due to Dr. N. F. Paterson for checking the manuscript and for many helpful recommendations. The Trustees of the South African Museum acknowledge gratefully the receipt of a grant from the Council for Scientific and Industrial Research towards the cost of publication. 514 ANNALS OF THE SOUTH AFRICAN MUSEUM SUMMARY After detailed studies of the anatomy of the claspers of the males, it is considered that Raja rhizacanthus Regan is synonymous with R. clavata Linnaeus and that R. pullopunctata Smith is distinct from R. batis Linnaeus, but that R. batis Linnaeus may also occur in South African waters. REFERENCES BARNARD, K. H. 1925. A monograph of the marine fishes of South Africa. Ann. S. Afr. Mus. 21: 1-418. . BiceLow, H. B. & SCHROEDER, W. C. 1953. Fishes of the western North Atlantic. Pt. II. Saw- fishes, guitarfishes, .. . Mem. Sears Fdn mar. Res. t (2): 1-588. ExMAN, S. 1953. a 1 TT, 3 9088 01206 599