Afrolittorina knysnaensis (F. Krauss in Philippi, 1847)

Reid, DG & Williams, ST, 2004, The Subfamily Littorininae (Gastropoda: Littorinidae) in the Temperate Southern Hemisphere: The Genera Nodilittorina, Austrolittorina and Afrolittorina, Records of the Australian Museum 56, pp. 75-122 : 106-110

Afrolittorina knysnaensis (Krauss, in Philippi, 1847)

Figs. 17C,D, 19–21

Litorina knysnaensis Krauss , in Philippi, 1847: 196, Litorina pl. 4, fig. 4 (Caput Bona Spei ad regionem fluminis Knysna [Cape of Good Hope near Knysna River, South Africa]; lectotype ( Janus, 1961) Stuttgart Museum ZI 0050273 (formerly MT 108; fig. 19A; seen) plus 30 paralectotypes ZI 0050274; 20 paralectotypes MCZ 154114; possible paralectotypes Senckenberg Museum uncatalogued (Knysna Fluss) and SMF 314706/7 (Algoa Bay) (Herbert & Warén, 1999); additional paralectotypes may be in Nationaal Natuurhistorisch Museum,Leiden (van Bruggen, 1992). Krauss, 1848: 102. Weinkauff, 1882: 71–72, pl. 9, figs. 10–11.

Littorina (Melarhaphe) knysnaensis .—Adams & Adams, 1854: 314 (as Melaraphe ). Janus, 1961: 6 (as Melaraphe ).

Littorina knysnaensis .— Reeve, 1857: Littorina sp. 75, pl. 14, fig. 75a,b. Troschel, 1858: 134 (radula) (as knysnaniensis). Sowerby, 1892:36 (in part; includes A. africana ). Bartsch, 1915:120. Turton, 1932: 132, pl. 28, fig. 955. Kilburn & Rippey, 1982: 50, pl. 10, fig. 2. Potter & Schleyer, 1991: pl. 2.1 (radula).

Littorina (Melarhaphe) africana var. knysnaensis .— Tryon, 1887: 249, pl. 44, figs. 65, 68 (as Melaraphe ).

Littorina (Austrolittorina) knysnaensis .— Rosewater, 1970: 423, 478– 479, pl. 364, figs. 17–23, pl. 365, fig. D (penis), pl. 369 (distribution). Kilburn, 1972: 402.

Nodilittorina (Nodilittorina) knysnaensis .—Bandel & Kadolsky, 1982: 37–38.

Nodilittorina (Austrolittorina) knysnaensis .— Reid, 2002 a: 154.

Afrolittorina knysnaensis .— Williams et al., 2003.

Littorina picea Reeve, 1857 : Littorina sp. 83, pl. 15, fig. 83 (no locality; type locality designated as South Africa ( Rosewater, 1970); holotype BMNH 1968320, seen).

Litorina africana .— Weinkauff, 1883: 218 (not Krauss, in Philippi, 1847; in part; includes A. africana ).

Litorina decollata .— Weinkauff, 1883: 219 (not Krauss, in Philippi, 1847 = A. africana ; in part; includes A. africana ).

Littorina africana tryphena Bartsch, 1915: 120 , pl. 38, fig. 6 (Port Alfred, South Africa; holotype USNM 187091, seen; paratype USNM 664353, not seen [ Rosewater, 1970]).

Littorina tryphena .— Turton, 1932: 132.

Littorina rietensis Turton, 1932: 131 , pl. 28, fig. 948 ( Port Alfred , South Africa; holotype OUM, seen).

Littorina mauritiana acuta .— Turton, 1932: 131, pl. 28, fig. 950 (not Litorina acuta Menke, 1843 = A. unifasciata ).

Littorina kowiensis Turton, 1932: 132 , pl. 28, figs. 956 ( Port Alfred , South Africa; 5 syntypes OUM, seen).

Littorina africana pica “Reeve ” Turton, 1932: 133 (unjustified emendation of L. picea Reeve, 1857 ).

Littorina indistincta Turton, 1932: 133 , pl. 28, fig. 959 ( Port Alfred , South Africa; 2 syntypes OUM, seen).

Littorina punctata .—Penrith & Kensley, 1970a: 191–239 (not Turbo punctatus Gmelin, 1791 = Echinolittorina punctata ). Penrith & Kensley, 1970b: 243–268 (in part, includes E. punctata , see Kilburn, 1972; personal observation of voucher material in SAM).

Taxonomic history. Although recognized as a distinct species by most authors throughout its history, there has been a recurring tendency to synonymize it with the sympatric A. africana ( Weinkauff, 1883; Tryon, 1887; Sowerby, 1892; Hughes, 1979; see Taxonomic History of A. africana ).

Reeve (1857) named a small, unlocalized shell of this species as Littorina picea ; this was erroneously synonymized with Litorina decollata (= A. africana ) by Weinkauff (1883), but correctly identified as L. knysnaensis by Tryon (1887). There has been confusion about the identification of the holotype of this taxon. Reeve (1857) referred to a single shell, but Rosewater (1970) incorrectly selected a “ lectotype ” from among “ 3 syntypes ”. An original label records that the two additional shells were added in 1877 (BMNH 1877.4.30.1, obtained from Sowerby). Rosewater (1970: 479, pl. 364, figs. 19–20) selected and figured the best specimen as “ lectotype ”, but this is not in fact Reeve’s holotype. One of the remaining two shells is a small, eroded, specimen of A. knysnaensis that is blackish brown above the periphery, paler near the suture and shows traces of the characteristic speckled pattern only on the base. This shell exactly matches Reeve’s original figure in its pattern of erosion and growth lines, and corresponds with his description of a plain black shell with a slightly rugose surface; it is clearly the holotype.

Various unusual forms of this species have been named. Littorina africana tryphena Bartsch, 1915 was based on two beach-worn shells of unusually high-spired, pupoidal shape; similar specimens have been found to contain trematode parasites (see Remarks below). Turton (1932) named several small or beach-worn shells: Littorina rietensis is a rare form with carinate ribs; L. kowiensis is brown without the normal speckled pattern; L. indistincta was based on small brown juveniles about 2 mm tall that are probably of this species.

Bandel & Kadolsky (1982) doubtfully listed Turbo dispar Montagu, 1815 , described from Poole, England, in the synonymy of A. knysnaensis , but recommended that it should be considered a forgotten name. Nevertheless, this can be recognized as Melarhaphe neritoides from the original figure and description (“obsoletely striated in a spiral direction” Montagu, 1815; not “spiral furrows”, Bandel & Kadolsky, 1982), as confirmed by a specimen from Montagu’s collection in the Royal Albert Memorial Museum, Exeter, England (no. 4202).

Schniebs (2000) identified the holotype of Phasianella tesselata Anton, 1838 as Littorina knysnaensis , but examination has shown this to be Echinolittorina punctata s.st.

Diagnosis. Shell moderately large, high-turbinate, angled at periphery, usually sculptured with fine spiral striae; pale, with fine red-brown speckling and broad bark brown band above periphery. Penis with rounded, slightly swollen filament, slight grey subepithelial glandular tissue distally, crimped lips of sperm groove; mamilliform gland on base, but no glandular disc. Pallial oviduct with three consecutive loops of egg groove, in albumen gland, capsule gland and in terminal portion of oviduct.

Material examined. 71 lots (7 AMS, 41 BMNH, 1 IRSNB, 8 ZMA, 1 MNHNP, 7 NM, 6 SAM), including 22 penes, 3 sperm samples, 11 pallial oviducts, 5 radulae.

Shell ( Fig. 19). Mature shell height 3.8–19.2 mm. Shape high-turbinate (H/B = 1.26–1.61; SH = 1.29–1.92; exceptionally tall-spired specimens contain trematode parasites, H/B = 1.98, SH = 2.36; some eroded shells can appear almost globular, H/B = 1.10, SH = 1.41); spire outline slightly concave at apex; whorls moderately rounded, suture distinct, periphery angled, sharply so in juveniles, but sometimes rounded; moderately solid. Columella pillar slightly concave; columella broad, slightly excavated; eroded parietal area absent. Sculpture of numerous fine spiral elements (usually not clearly differentiated into primary grooves, ribs and microstriae); on penultimate whorl 10–13 threads visible above suture; on last whorl numerous closely spaced flat striae over entire whorl, of which 20–27 may be raised as threads; periphery marked by an enlarged riblet; occasionally almost smooth with only faint impressed spiral lines and microstriae; growth lines weak; spire often eroded. Protoconch rarely preserved, 0.34–0.43 mm diameter, 3 whorls. Colour variable, from pale grey to almost black; most commonly grey to cream ground colour, with fine red-brown speckling, palest on base, but darkening to form a broad, almost solid, brown to black band at and above periphery; palest shells grey to white, with only faint brown speckling or fine spiral lines, rarely unpatterned; darkest shells black with white speckling on base; aperture dark brown with basal white band; first 2 whorls of teleoconch brown.

Animal. Head and sides of foot grey to black; tentacles with fine black transverse lines (sometimes faint or absent). Opercular ratio 0.51–0.62. Penis ( Fig. 20A–G): filament about 0.5 total length of penis, wrinkled and therefore not clearly differentiated from base, rounded and slightly expanded distally, slightly swollen, small patch of grey subepithelial glandular tissue may be visible distally; sperm groove open to filament tip, lips of groove crimped near tip; single mamilliform gland (rarely 2) borne on lateral branch of base; glandular disc absent; instead the epithelium around the papilla of the mamilliform gland is tall and glandular; base may be slightly pigmented. Euspermatozoa length unknown; paraspermatozoa ( Fig. 20I) oval, 14–21 µm long, containing large spherical granules and single elongate, straight, blunt-ended rod-body 24–48 µm long projecting from cell. Pallial oviduct ( Fig. 20H; not as illustrated by Gosliner, 1981) with simple loop of albumen gland, followed by large, almost circular loop of capsule gland, of which a small distal part may be differentiated as reddish translucent capsule gland; additional simple loop of glandular material between capsule gland and terminal portion of pallial oviduct; small flexure of egg groove just before opening to mantle cavity; large copulatory bursa separates in a posterior position and continues back to overly capsule gland.

Spawn and development. Not recorded; pelagic spawn and planktotrophic development predicted from egg size (87 µm, McQuaid, 1981a), from dimensions of protoconch and large capsule gland ( Reid, 1989) and from low genetic differentiation on a geographical scale (Grant & Lang, 1991); continuous recruitment recorded at False Bay ( McQuaid, 1981a).

Radula ( Fig. 17C,D). Relative radular length 3.2–5.7 (5.59, s.e. 0.27; Potter & Schleyer, 1991). Rachidian: length/width 1.00–1.27; major cusp elongate to rectangular, bluntly rounded at tip. Lateral and inner marginal: major cusps large, elongate to rectangular, truncate at tip. Outer marginal: 7–9 cusps.

Habitat. On the rocky coasts of Namibia and South Africa (south of the Transkei) this is a dominant species on the upper shore. It is abundant across a wide zone from the littoral fringe down to the barnacle belt, extending to mean low water of neap tides ( Stephenson et al., 1937; Bokenham et al., 1938; Bright, 1938; Stephenson et al., 1940; Penrith & Kensley, 1970a,b, as Littorina punctata ; Kensley & Penrith, 1980; McQuaid, 1981a, 1992). The vertical range is less wide on shores of moderate exposure, and the species is absent from sheltered and muddy bays (Penrith & Kensley, 1970a, as Littorina punctata ). Settlement occurs mainly at the top of the shore, to avoid heavy wave action, although some settlement also takes place in crevices at lower levels; adults move downwards as they grow, to reach more abundant food ( McQuaid, 1981b). Where both species occur, A. knysnaensis and A. africana show broad overlap of their vertical ranges ( Hughes, 1979), although at East London the former has a higher vertical limit (Eyre et al., 1938). The temperature relations of these two species differ, A. knysnaensis attaining higher tissue temperatures when attached by mucus in full sun, as a consequence of its darker coloration; it therefore occurs mainly in shaded microhabitats (McQuaid & Scherman, 1988).

Range ( Fig. 21). Namibia and South Africa. On the Atlantic coast the northernmost samples examined are from Honolulu (20°36'S SAM A31185 ) and Rocky Point (18°59'S SAM A31184) (as also recorded by Kilburn, 1972). On the Natal coast the northernmost records are one specimen from Umhlali (29°20'S 31°14'E, Stephenson, 1947), one shell from Amanzimtoti (30°03'S 30°53'E, BMNH) and another from Umkomaas (30°12'S 30°48'E, NM 8604; also Kilburn, 1972), but the species is relatively uncommon (in comparison with the sympatric A. africana ) to the north of Port St Johns (31°38'S 29°33'E; Hughes, 1979, also Stephenson, 1944).

Remarks. Although they have been synonymized by several authors, the shells of A. knysnaensis and A. africana are almost always immediately separable by their coloration. That of A. africana is blue-grey, sometimes with faint fine tessellation of yellow-brown; that of A. knysnaensis is densely and minutely speckled with red-brown, with a broad blackish peripheral band. Occasional pale or unpatterned shells of A. knysnaensis are more difficult to identify, but the periphery is usually more strongly angled than in A. africana . The operculum is more tightly wound in A. knysnaensis (i.e. higher opercular ratio). Anatomically, the penial filament of A. knysnaensis is more wrinkled, with a crimped margin to the distal part of the sperm groove, the grey colour of the subepithelial glandular tissue within the filament is slight or absent, and the rod-bodies within the paraspermatozoa are much longer and project from the cells. No anatomical difference has been detected between the females of the two species. Both Rosewater (1970) and Hughes (1979) have noted that the margin of the mantle is pigmented in A. africana and not in A. knysnaensis , but this character is inconsistent and both pigmented and unpigmented states can be found frequently in both species.

Distribution maps give the impression of a broad overlap between the ranges of A. knysnaensis and A. africana , from False Bay to the vicinity of Durban. However, in a survey of the distribution of these species on the South African coast, Hughes (1979) found that A. knysnaensis was dominant at all sites as far north as East London, both species were common at Port St Johns, and only A. africana was common in Natal. Afrolittorina knysnaensis occurs alone on the Atlantic coast of southern Africa, under the influence of the cold Benguela Current, and is clearly a cool-temperate species, whereas A. africana shows a warmtemperate and subtropical distribution ( Stephenson, 1944). The temperature relations of these two species have been studied by McQuaid & Scherman (1988), who showed that body temperatures of brown-shelled A. knysnaensis were 2–3°C higher than those of white-shelled A. africana , when both were attached by mucus to open rock surfaces on the shore in full sunlight (there was no difference when attached by the foot, because conductive heat gain overrode the effect of decreased radiant absorption by the paler shell). Behavioural observations also showed that A. knysnaensis was less tolerant of insolation than A. africana , for the former tended to occur in damp, shady pits in the rock rather than on open surfaces, at least when subjected to long periods of emersion during neap tides. These authors suggested that the lower tolerance of A. knysnaensis to heating by sunlight might explain its replacement by A. africana on the subtropical east coast of South Africa. There is an interesting parallel with the two species of Austrolittorina in New Zealand, where the brown A. cincta is more common in the south and the blue-white A. antipodum in the north. In Australia, Afrolittorina praetermissa has a patterned or brown shell and is a cool-water species, whereas Austrolittorina unifasciata with a white shell extends along warm-temperate coasts. The pattern is repeated within Austrolittorina araucana on the west coast of South America, where entirely brown populations occur in the south and mixed brown and white populations in the north ( Reid, 2002 a). Climatic selection might account for all these trends. The importance of pale shells as a means of temperature regulation in intertidal gastropods has been reviewed by Vermeij (1973) and has also been recorded in some tropical littorinid species ( Markel, 1971). McQuaid & Scherman (1988) found no evidence of visual selection on shell colour, but Hughes (1979) noted a correlation between the abundance of Afrolittorina africana and pale rocks, implying that visual selection might also be involved.

Two other littorinid species have been recorded from the temperate coast of southern Africa. Littorina saxatilis (Olivi, 1792) occurs at two sites, Langebaan and Knysna, among marsh grass and on stones in sheltered lagoons. Neither of the southern African Afrolittorina species can be found in this habitat. The shells of these populations of L. saxatilis are smooth, thin-walled, with well-rounded whorls and are variously coloured in shades of yellow and brown, usually with a coarsely marbled or tessellated pattern.Anatomically they are quite distinct and no confusion should occur (see Reid, 1996, for review). These colonies are apparently the basis for the records of Littorina punctata in South Africa by Barnard (1963; see Kilburn, 1972). Rosewater (1970) accepted Barnard’s records of L. punctata from South Africa and added others from Port Elizabeth; these may have been based on misidentification of A. knysnaensis , for E. punctata does not occur south of Namibia ( Kilburn, 1972). It has recently been suggested that Echinolittorina punctata consists of a complex of three species ( Reid, 2002 b), of which the one in Namibia should be known as Echinolittorina pulchella (Dunker, 1845) . Kilburn (1972) stated that the range of this species abuts that of A. knysnaensis at Rocky Point in Namibia, where both species occur together. Furthermore, he suggested that the presence of apparent intermediates (in shell pattern) to the south implied hybridization. It is true that shells of E. pulchella and of A. knysnaensis can be similar, but it is now clear that there are significant anatomical differences between the two. The pallial oviduct of E. pulchella contains only a single loop of the egg groove, in the albumen gland, and the penis shows a glandular disc in addition to the mamilliform gland. The colour pattern of E. pulchella is of relatively coarse tessellation of brown to black spots, aligned in oblique series on a pale ground, often appearing as a black shell with large white spots. In contrast, the shell of A. knysnaensis has a finely speckled brown pattern; this is visible on the base even in those that are almost black above the periphery, as is frequent in samples from Namibia.

Hughes (1979) mentioned and illustrated unusually elongated shells from Arniston ( Fig. 19F). Some of these have been dissected and found to contain trematode parasites, which were presumably responsible for the distortion. Kilburn (1972) reported that the peripheral keel was more well developed in samples from wave-exposed shores, whereas it was said to be absent in estuarine shells, which were also found to be larger in size. The largest examples seen in the present study were from the Atlantic coast.