Tyrinna nobilis ( Abraham, 1877 ) Schrödl & Millen, 2001

Schrödl, M. & Millen, S. V., 2001, Revision of the nudibranch gastropod genus Tyrinna Bergh, 1898 (Doridoidea: Chromodorididae), Journal of Natural History 35, pp. 1143-1171 : 1146-1160

Tyrinna nobilis ( Abraham, 1877) comb. nov.

(®gures 1±6)

Doris delicata Abraham, 1877: 259 ±260; SchroÈdl 1996: 22.

Tyrinna nobilis Bergh, 1898: 523 ±526, pl. 30, ®gures 21±29, pl. 32, ®gures 21±24; Marcus 1959: 29 ±33, ®gures 45±53; Rudman 1984: 244 ±246; MuniaõÂn et al. 1996: 265±273; SchroÈdl 1996: 22, pl. 3, ®gure 15; 1997a: 38±42.

Tyrinna pusae Marcus, 1959: 33 ±34, ®gures 54±64; Rudman 1984: 244 ±246; MuniaõÂn et al. 1996: 265±273; SchroÈdl 1996: 22.

Chromodoris juvenca Bergh, 1898: 531 ±532, pl. 31, ®gures 4±11; Odhner 1921: 226.

Glossodoris juvenca: Carcelles and Williamson, 1951: 314 .

Cadlina juvenca: Marcus, 1959: 87 ; SchroÈdl 1996: 21.

Material examined ( tables 1 and 2). Doris delicata : HOLOTYPE, BMNH no. 1995038; Chile. Tyrinna nobilis : HOLOTYPE, ZMB, no. 50750; Calbuco, Chile, collected by Plate. Three specimens, SMNH no. 1544, 1545; Station M 94 (41 ss 46 ¾ 30 ² S, 73 ss 45 ¾ 45 ² W), Strait of Chacao, Chile, 40 m depth, small stones, 4 May 1949, collected by the L.U.C.E. Station M 42 (42 ss 20 ¾ 50 ² S, 73 ss 22 ¾ 00 ² W), Chiloe Island, Chile, 70 m depth, hard substrata, 24 January 1949, collected by the L.U.C.E. Eleven specimens, diOEerent Chilean and Argentinian localities (see table 1), collected by S. Millen, S. Gigglinger, K. Salger, M. SchroÈdl. One entire specimen, MACN no. 33873, Pta. MarqueÂs (45 ss 58 ¾ S, 67 ss 34 ¾ W), labelled:`13 January 1995, 33 mm’. Tyrinna pusae : HOLOTYPE, SMNH no. 3736; Station M 100 (41 ss 49 ¾ 25 ² S, 73 ss 48 ¾ 58 ² W), Chiloe Island, Chile, 10 m depth, small stones with calcareous algae, 5 May 1949, collected by the L.U.C.E. PARATYPE, SMNH no. 3737; Station M 106 (41 ss 48 ¾ 40 ² S, 73 ss 21 ¾ 00 ² W), Gulf of Ancud, Chile, 36 m depth, rough sand with shells, 5 May 1949, collected by the L.U.C.E. Chromodoris juvenca : LECTOTYPE, ZMB no. 50725, Isla de PaÂjaros, northern Chile, collected by L. Plate, under stones; PARALECTOTYPE, ZMB no. 50725, collected together with the lectotype.

External morphology. Living specimens are translucent white. Opaque white lines run along the notal and pedal edges. The rhinophores are marked by median opaque white lines, and opaque white pigment also occurs on the gills. The notum is surrounded by orange to red spots forming two or three irregular submarginal rows. The diameter of the spots decreases from the inner to the outer rows. In the frontal region there is usually only one row of spots. Most of these spots indicate the position of large white subepidermal glands which are shining through the tissue (see ®gures 1A,B). There are no orange to red spots nor subepidermal glands in the central part of the notum. After ®xation the specimens lost their orange and red pigmentation becoming whitish or yellowish. The rhinophores usually became darker.

Crawling specimens have an elongate body shape reaching 46 mm in length, 13 mm in width and 6 mm in height. Stationary specimens have an oval shape. Due to contraction, the largest preserved specimen only reaches 25 mm in length, 10 mm in width and 8 mm in height. The notal surface is slightly folded but bears no macroscopic papillae. The rhinophores and the gill cavity are surrounded by slightly elevated, smooth sheaths. The rhinophores bear up to 22 lamellae. There are six to nine bi- to tripinnate gills which form a semicircle in front of the anal papilla in living specimens. In preserved specimens the anal papilla may appear to be situated in the centre of the gill circle due to contraction. The oral tentacles in living animals are long and digitiform; they have a long, deep longitudinal groove (®gure 1C), giving it an`enrolled’ shape. The groove does not, however, reach the base of the tentacles which is solid and rounded in cross-section. In preserved specimens there may be additional transverse folds on the inner side of the oral tentacles due to contraction. Anteriorly the foot is divided into two lips which are not notched. The foot corners are elongated and slightly tapered. In living specimens the tail is long, narrow and tapered posteriorly. After ®xation the contracted tail may or may not project behind the notum ( table 2).

Mantle structures. Spicules appeared to be absent during external and anatomical examination. However, maceration of mantle tissue and histological sections con®rm the presence of a few calcareous spicules in the deeper layers of the mantle tissue (®gure 1D). These spicules are chunky, somewhat banana- or rice-shaped, and reach 230 m m in length. Cross-sections show that the spicules are hollow with surrounding concentric layers (®gure 2A).

There are two or three submarginal rows of spherical, subepidermal mantle glands (mantle dermal formations, compound glands). The MDFs of the outer band reach 0.3 mm in diameter, the inner MDFs up to 2 mm in the large specimen no. 1 (®gure 2B), in the smaller no. 8 they reach about 0.15 mm and 0.5 mm (®gure 2C), respectively. Both outer and inner MDFs of the serially sectioned specimen no. 8 are surrounded by a thin muscular capsule as was described for Chromodoris britoi (Ortea and PeÂrez, 1983) by Avila and Durfort (1996). The MDFs are ®lled with vacuolar cells, but cell limits are not clearly detectable. Each vacuolar cell appears to contain one large, central, more or less rounded vacuole, the nucleus lies peripherally (®gure 2B). The vacuoles reach 15±30 m m in diameter; some are weakly stained with light bluish content. The vacuoles appear to be densely packed throughout the MDF. In the periphery of the MDFs, between muscular capsule and vacuolar cells, there are cells containing violet staining particles which, in structure and stain, closely resemble epidermal glands. This layer may correspond to the`single cell layer’ found in Chromodoris purpurea (Laurillard, 1831) and Hypselodoris elegans (Cantraine, 1835) by Garcia-Gomez et al. (1991: ®gure 4), but cell limits are not clearly visible.

Within larger, inner MDFs of specimen no. 8 and the similar-sized outer MDFs of specimen no. 1 this layer of violet staining particles is not limited to the periphery, but, in certain positions, muscles and violet staining glandular cells extend in a foldlike pattern towards the interior of the MDFs (®gure 2A), giving the MDFs an internal substructure.The degree of folding increases with the size: within larger MDFs of specimen no. 1 (up to about 2 mm in diameter) there are clusters of vacuolar cells with less densely packed vacuoles which are separated by thick layers of violet staining cells. These layers, as in the smaller MDFs mentioned above, are apparently invaginated folds of the peripheral, violet staining layer (®gure 2B). This situation of smaller MDFs homogeneously ®lled with round vacuolar cells and larger MDFs having vacuolar cells which are successively more clustered by folds of the surrounding cell layer agrees with the drawings of sections of the holotype of T. pusae by Marcus (1959). The MDFs have no natural opening onto the dorsal or ventral mantle surface and expulsion of MDFs has never been observed in living specimens.

Central nervous system ( CNS). The cerebral and pleural ganglia are fused. There is a separate rhinophoral ganglion at the base of the rhinophoral nerve, attached to each cerebral ganglion. The optical nerves are short. The statocysts contain several otoconia. The CNS is covered by the ¯at blood gland which may be separated into an anterior and a posterior lobe.

Digestive system. The lip cuticle bears bi®d rodlets; a few with three cusps also occur (®gure 3). There are 65 to 93 radula rows with a central tooth and 47 to 74 lateral teeth per half row in the recently collected material ( table 2). Teeth are very variable in shape, even within the same radula (for comparison see table 2): in the smaller specimens no. 3, 4 and 5 the rhachidian teeth reach about half the size of the lateral teeth and have a claw-like appearance bearing three to six small and equal sized denticles (®gure 4A). In specimen no. 2 there are small claw-like rhachidian teeth with ®ve to six equal sized denticles in some older rows, but in younger rows the rhachidian teeth are elongated medially and bear pairs of irregularly shaped denticles (®gure 4B). The large specimen no. 6 has more or less claw-like rhachidian teeth in older rows but in younger rows they are irregularly elongated and form a single projecting median cusp (®gure 4D). Specimen no. 7 has elongated rhachidians nearly reaching the size of the ®rst laterals with a paired or a single projecting median cusp (®gure 4E). Specimen no. 1 has rhachidian teeth nearly reaching the size of the lateral teeth. There are ®ve to six denticles on each side of a single projecting median cusp (®gure 4F). The ®rst lateral teeth usually have three to ®ve inner denticles and two to six outer denticles beside a projecting cusp (®gures 4A,B,E). First laterals of specimen no. 1 lack inner denticles while being extremely aberrant in no. 6 (®gure 4C). The second laterals generally have ®ve to six outer denticles beside the principal cusp. The lateral teeths are hook-shaped; the size and number of outer denticles successively decreases towards the outside, especially in the larger specimens ( table 2). In no. 1 and no. 6, denticles are only present in the 10±15 inner-and outermost laterals; in the centre of the half rows no denticles could be detected. The approximately 20 outer laterals generally become more erect; the number of posterolateral denticles increases up to eight (®gures 4G,H). The three to ®ve outermost laterals are small, irregular in shape and bear some denticles (®gure 4H).

The salivary glands are long tubes becoming rather thick posteriorly. They may run along the pharynx laterally or be adjacent to the oesophagus. The oesophagus starts as a wide tube which becomes narrower leading backwards into the digestive gland. It enters the stomach ventrally or postero-ventrally. The stomach lies medially or to the left, embedded within the digestive gland. In the specimens no. 2, 4 and 5 the stomach and the caecum are very small. Specimens no. 1, 6, 7 and 8 have a more swollen stomach, but only the bulbous tip of the caecum is visible at the surface of the digestive gland (®gure 5A). Specimen no. 3 has a large stomach with a wide caecum which are both partly visible dorsally (®gure 5B). The intestinal system of the latter specimen was ®lled with a dark brown mass containing sand granules. The stomach dorsally passes into the wide, longitudinally folded intestine which becomes narrower more distally. The intestine may extend to the anterior border of the digestive gland, and then, with or without convolutions, follow its right edge backwards to the anal papilla, or it may be curved backwards from the stomach to the anus in a narrow loop.

Reproductive system. The reproductive system is similar in all specimens examined; an outline is given in ®gure 6. The gonad covers the digestive gland. The thin hermaphroditic duct passes into a banana-like and somewhat ¯attened ampulla which reaches up to 0.5 mm in diameter. The proximal gonoduct is very short. The oviduct immediately enters into the female gland mass after its junction with the vas deferens. The morphology of the vas deferens agrees with the descriptions of T. nobilis and T. pusae by Marcus (1959): after a short, rather straight portion which is prostatic, the vas deferens distally becomes narrower. This strongly curved portion is sheathed by a muscular layer. More distally the vas deferens passes into the ejaculatory portion as a straight duct closely surrounded by a muscular sheath. The total length of the vas deferens measures from about 12 mm (specimen no. 2) to 40 mm (specimen no. 6). The vas deferens extends considerably into the left body portion in specimens no. 1, 2, 6 whereas it is limited to the right portion in specimens 3, 4 and 7 ( table 2). The vas deferens enters the common vestibule next to the insertion of the vagina, forming a slender unarmed penis as indicated by Marcus (1959). The vas deferens does not open separately as suspected by Rudman (1984) and MuniaõÂn et al. (1996). The vestibular walls are soft and longitudinally folded, without cuticular structures. In specimens no. 6 and 7, parts of the vestibule and the penis are everted, the latter forming a long, ¯agelliform tube (®gure 1C). The vagina is a long and narrow duct (only 50±100 m m in diameter) which is strongly curved distally. This coiled distal portion is surrounded by connective tissue and contracted muscular ®bres, forming a bulb with a ®lamentous surface. There is no branched vestibular gland surrounding the vagina as mentioned by MuniaõÂn et al. (1996). After most of its length, the vagina gives rise to the somewhat wider insemination duct which enters the nidamental duct far distally. More proximally, the vagina divides into the short stalk of the rounded receptaculum seminis and inserts on the oval translucent bursa copulatrix as a short duct. Both allosperm receptacles lie close to each other, their stalks are crossed. The female gland mass is divided into the widely convoluted yellowish portion of the mucus gland, and a whitish portion having a more granular appearance. The nidamental duct is rather long, has a conspicuously swollen bulb and opens separately below the common male and female genital aperture. Histological sections of specimen no. 9 show no traces of glandular structures attached to the atrium or to the vagina. Small, poorly staining tubules within the body wall of specimen no. 9 near the genital aperture may or may not have a glandular function. This structure has not been detected in any other specimen.

Remarks. The specimens examined during this study were collected from a wide geographical range ( table 1) and show considerable diOEerences regarding radular characters. In particular, specimen no. 1, which is the northernmos t Chilean specimen found, is aberrant due to its rhachidian teeth bearing a large projecting median cusp (®gure 4F). In contrast to other specimens, the ®rst laterals of specimen no. 1 do not have inner denticles. Since radular characters are widely used to distinguish chromodoridid species (e.g. Bertsch, 1977; Rudman, 1984), these diOEerences might easily lead to the assumption that the specimens studied belong to two or more diOEerent species. However, only the smaller specimens no. 3, 4, 5 possess rather homogeneous teeth. Specimen no. 2, as well as the holotype of Tyrinna pusae have both claw-like and medially elongated rhachidians with a paired projecting median cusp within their individual radulae (this study; Marcus, 1959). The rhachidians of the Argentinian specimen no. 6 may either have a claw-like appearance or bear a single projecting median cusp. The presence of claw-like rhachidians together with those having a single projecting central cusp was also described for the holotype of Tyrinna nobilis Bergh, 1898 , collected from Chiloe Island ( Bergh, 1898: pl. 32, ®gures 21 and 23) as well as for a specimen of Tyrinna evelinae by Collier and Farmer (1964), indicating a high intra-individual and intraspeci®c variability.

The diOEerent shape of the ®rst lateral within the material studied may also be due to intraspeci®c variation as was suggested for the specimens assigned to T. pusae by Marcus (1959). The lack of well developed inner denticles on all ®rst laterals of specimen no. 1 could also be due to individual aberrations of corresponding odontoblasts. This would be analogous to the absence of outer denticles on the ®rst and on the ®fth left as well as on the third (or fourth) right lateral teeth of specimen no. 6 (®gure 4C). The latter tooth positions are surrounded by normal, denticle-bearing teeth, but lack denticles through all the radular rows. Another explanation is given by deletion and fusion of odontoblasts: since the denticle lacking ®rst laterals of specimen no. 1 closely resemble the second laterals of other specimens examined, both ®rst laterals may have fallen out, being replaced by the second laterals. Since the aberrant, large central tooth somewhat resembles the usual ®rst laterals in size and shape it even seems possible that the rhachidian and one of the ®rst laterals have fallen out in specimen no. 1, with the remaining ®rst lateral occupying a central position. However, there is no clear asymmetry of the central teeth to support this hypothesis (®gure 4F). Due to the broad base and large size of the central tooth of no. 1, it may also be the result of the fusion of central and lateral teeth. Odhner (1926) considered the fusion of central teeth in cadlinids to be very unlikely, but it appears that at least the ®rst two right laterals of no. 6 may be fused into one very broad tooth, and also the following lateral is broad and has a bi®d projecting cusp. There is still little data on the mechanisms leading to aberrant radular teeth but this study and the results of Rudman (1984; 1995) indicate that deformations of the central radula portion are rather common among individuals of several chromodoridid species.

Regarding radular characters, the specimens dissected here ®t well into the known range of variation of the specimens assigned to T. nobilis and T. pusae by Marcus (1959) and MuniaõÂn et al. (1996) (see table 2). The higher number of radular rows and teeth per half row of T. pusae may be due to the larger body size of these specimens. Marcus (1959) saw speci®c diOEerences between T. nobilis possessing denticulate lateral teeth and T. pusae with most laterals lacking denticulation. However, he also noted diOEerent degrees of denticle reduction within his material assigned to Tyrinna nobilis . The material examined during this study shows further transitional stages and no essentially diOEerent traits in the quality and quantity of denticle reduction can be stated ( table 2).

Another distinguishing feature which Marcus (1959) used to separate T. pusae from T. nobilis is the position of the stomach. In T. pusae the stomach is described as completely embedded within the digestive gland ( Marcus, 1959) whereas in T. nobilis the stomach is visible dorsally ( Bergh, 1898; Marcus, 1959). The holotype of T. nobilis was completely eviscerated and thus is not appropriate for reexamination of the digestive system. Re-examinatio n of Marcus’ specimens, however, shows that the stomach is always embedded within the digestive gland, reaching the surface at the insertion of the intestine. In specimens assigned to T. nobilis by Marcus (1959) a well-developed caecum reaches the surface of the digestive gland lying to the right of the intestine insertion. A caecum in the same position also appears to be present in the holotype of T. pusae ; it is not freely visible but covered by some tissue dorsally. The material dissected during this study shows that the size of both stomach and caecum is highly variable; it may be inūenced individually by body size, nutritional stage and arti®cial contraction due to ®xation (®gures 5A,B).

The holotype of T. nobilis is immature ( Bergh, 1898; this study). Although the general genital structure of material assigned to T. nobilis and T. pusae by Marcus (1959) coincided, Marcus (1959) saw speci®c diOEerences between both species concerning the length and the position of the vas deferens. The vas deferens of T. pusae is described as reaching 40 mm in length, and that of the largest specimen of T. nobilis up to 60 mm ( Marcus, 1959). Re-examination of a smaller specimen of T. nobilis det. Marcus indicates that this partly dissected specimen has a much shorter male duct (about 20 mm), although the length of the convoluted and hardened parts could not be measured exactly. The length of the vas deferens in specimens examined during this study ranges between 12 and 40 mm and is not strictly correlated with the body size ( table 2). The vas deferens is highly convoluted as a whole and, additionally, the central duct is more or less looped within the muscular sheath. Thus the vas deferens must be stretched to be measured and the results are hardly comparable; the same must be supposed for Marcus’ data. As seen in the material dissected, even a short vas deferens may extend into the left body portion (specimen no. 2; table 2) as was described for T. nobilis . On the other hand, a longer one may be limited to the right portion (specimen no. 7; table 2) as was proposed for T. pusae .

It is concluded that the specimens studied here and the material assigned to T. nobilis and T. pusae by Bergh (1898) and Marcus (1959) are heterogeneous, especially regarding radular characters, but there are always transitional stages connecting the extremes. Even specimens no. 1 and 6, having the most aberrant central radulae agree well with all other specimens regarding external, reproductive and other digestive features. Therefore, until more information on further specimens is available, all specimens are considered to belong to the same species. Encompassing the range of morphological variation originally described from T. nobilis and T. pusae by Bergh (1898) and Marcus (1959), both species are considered to be conspeci®c. This synonymy is in accordance with the proposal of MuniaõÂn et al. (1996).

The observation of MuniaõÂn et al. (1996) that T. nobilis has a ramifying vestibular gland covering the vagina diOEers from the original descriptions of T. nobilis and T. pusae . During this study no such glandular structures were detected in the reexamined type material of T. nobilis and T. pusae , nor within the eight dissected, recently collected specimens. Dissection of museum material ( MACN 33873) identi- ®ed as T. nobilis by MuniaõÂn et al. (1996) shows the`vestibular gland’ drawn by MuniaõÂn et al. (1996: ®gure 5B) to be the highly coiled distal vagina which is surrounded by a bulb of contracted muscle ®bres and connective tissue.

Another surprising ®nding within specimens examined during this study is the presence of spicules, which had not been detected earlier by Bergh (1898), Marcus (1959) or MuniaõÂn et al. (1996). An explanation may be their low number and deep position within the mantle tissue. As known for other material preserved for a long time, spicules also may have been dissolved (see SchroÈdl, 1997b,c).

The poorly described holotype of Doris delicata Abraham, 1877 ( BMNH, no. 199503 8) collected from the Chilean coast at Chiloe Island ( Abraham, 1877) was externally examined. This entire, yellowish specimen measures 20 mm in length, 9 mm in width and 6 mm in height. It has a narrow foot (2.5 mm in width), a long (3 mm) but contracted tail, and digitiform, longitudinally grooved oral tentacles. The notum is rather smooth, not bearing papillae. Along the notal border there are two to three rows of large, empty subepidermal glands. There are elevated sheaths with smooth edges around the rinophores and the gill cavity. The rhinophores bear about 15 leaves. There are six bi- to tripinnate gills in a semicircle anterior to the anus. This holotype coincides with the specimens examined during this study regarding all external characters and collecting locality, and thus there is no reason to doubt their conspeci®ty. The same is true for the lectotype (and paralectotype) of Chromodoris juvenca which externally agree perfectly with the other specimens studied. The oral tentacles, described as being short-conical in the (lost) specimen from the Juan FernaÂndez Islands, are long, digitiform and longitudinally grooved in the lectotype. Large, marginal MDFs (`white sacs’) shining through the notal tissue were already mentioned in the original species description by Bergh (1898), and also the anatomical features of C. juvenca ®t into the range of variation of the T. nobilis examined (see table 2). Consequently, Tyrinna nobilis Bergh, 1898 , Tyrinna pusae Marcus, 1959 , and Chromodoris juvenca Bergh, 1898 , are all regarded as junior synonyms of Tyrinna delicata ( Abraham, 1877) .

Following the new code of nomenclature, retention of the commonly used name T. nobilis instead of its replacement by the senior but disregarded synonym T. delicata is proposed.