Species names and metaphyly: a case study in Discodorididae (Mollusca, Gastropoda, Euthyneura, Nudibranchia, Doridina)

Species names and metaphyly: a case study in Discodorididae (Mollusca, Gastropoda, Euthyneura, Nudibranchia, Doridina) Blackwell Publishing, Ltd.

BENOÎT DAYRAT & TERRENCE M. GOSLINER

Accepted: 25 August 2004 doi:10.1111/j.1463-6409.2005.00178.x

Dayrat, B. & Gosliner, T. M. (2005). Species names and metaphyly: a case study in Discodorididae (Mollusca, Gastropoda, Euthyneura, Nudibranchia, Doridina) — Zoologica Scripta, 34, 199–224. Absence of resolution in phylogenetic trees, or metaphyly, is a common phenomenon. It mainly results from the fact that each data set has its own limit and can hardly be expected to reconstruct alone an entire hierarchy. Because metaphyly helps point out which regions of a tree merit further investigation, one should not avoid metaphyly but rather should try to detect it by addressing carefully node reliability. In this paper we explore the implication of metaphyly for species names. We present a phylogenetic analysis of Discodorididae (Mollusca, Gastropoda, Nudibranchia, Doridina), with an emphasis on relationships among species of Discodoris and its traditionally close ‘allies’ such as Peltodoris and Anisodoris. We demonstrate that some species must be transferred to different discodoridid subclades with which they share synapomorphies, but that many species form a metaphyletic group at the base of Discodorididae, and therefore cannot be placed in any taxon of genus level. We demonstrate that the current International Code of Zoological Nomenclature does not allow taxonomists to handle this situation because it requires selecting a taxon name of genus rank for every species binomial. Then we evaluate the results provided by new forms of species names, both in a rank-based system, such as the current nomenclature, and a rank-free system. All solutions considered would cause radical changes to the ‘spirit’ of the current ICZN (and, by extension, to the other current codes). In a rank-free system of nomenclature, such as the PhyloCode, the best result is obtained with an epithet-based form that does not mention supra-specific relationships. Under this method, official species names would take the form ‘boholiensis Bergh, 1877’, although page numbers and letters can be added for uniqueness purposes. Taxonomists would then be free to add supra-specific taxon names in ‘common’ species names, such as Discodorididae boholiensis Bergh, 1877 or simply Discodorididae boholiensis. Here we wish to stimulate discussion of a problem that we believe merits wide debate: absence of resolution in phylogenetic reconstruction and its impact on species nomenclature. Benoît Dayrat, Department of Invertebrate Zoology and Geology, California Academy of Sciences, 875 Howard Street, San Francisco, CA 94103, USA. E-mail: [email protected] Terrence Gosliner, Department of Invertebrate Zoology and Geology, California Academy of Sciences, 875 Howard Street, San Francisco, CA 94103, USA. E-mail: [email protected]

Introduction Our current binominal species nomenclature requires selecting a taxon name of genus rank as the first part of every species binomial. Indeed, the ICZN (Art. 5.1) states: ‘The scientific name of a species, and not of a taxon of any other rank, is a combination of two names (a binomen), the first being the generic name and the second being the specific name.’ This is a requirement even when poor phylogenetic resolution may prevent us from being able to select a generic name. Selecting a mandatory genus name stops being an issue in a non-Hennigian framework where taxon names may refer to non-monophyletic taxa. Like the vast majority of systematists, we think that taxon names should refer to clades; here

we wish to analyse how the impossibility of selecting a clade name of genus rank can be handled. Our case study is a phylogenetic analysis among species of Discodorididae, a monophyletic taxon of dorid sea slugs (Mollusca, Gastropoda, Euthyneura, Nudibranchia, Doridina), with an emphasis on species currently classified within the genus Discodoris. Discodoridids share several synapomorphies (e.g. Valdés 2002), among which the digitiform oral tentacles and the notched superior lip at the anterior end of the foot can be easily observed externally (Fig. 1J–L). Some generic taxa of Discodorididae have a monophyletic status that is well supported by exclusive synapomorphic characters. For example, Geitodoris includes species sharing

© The Norwegian Academy of Science and Letters 2005 • Zoologica Scripta, 34, 2, March 2005, pp199–224

199

Species names and metaphyly • B. Dayrat & T. Gosliner

Fig. 1 A–L. Illustration of labium notch (character #1) and oral appendages (# 2 and #3). —A. Diaphodoris luteocincta, CASIZ 099131, scale = 1.6 mm. —B. Gymnodoris ceylonica, CASIZ 073060, scale = 1.7 mm. —C. Doris verrucosa, CASIZ 067689, scale = 5 mm. —D. Archidoris pseudoargus, CASIZ 076061, scale = 1.2 mm. —E. Austrodoris kerguelenensis, CASIZ 087312, scale = 1.2 mm. —F. Conualevia marcusi, CASIZ 018372, scale = 3 mm. —G. Aphelodoris antillensis, CASIZ 077289, scale = 2 mm. —H. Tyrinna evelinae, CASIZ 069891, scale = 3 mm. —I. Cadlina luteomarginata, CASIZ 071375, scale = 5 mm. —J. Asteronotus caespitosus, CASIZ 070364, scale = 1.8 mm. —K. Discodoris boholiensis, CASIZ 158768, scale = 5.5 mm. —L. Paradoris indecora, ZMUC (Istria), scale = 2.7 mm.

spatulate outermost teeth (e.g. Schrödl 2000), and Platydoris includes species sharing a very flattened body (e.g. Dorgan et al. 2002). Interestingly, most of those taxa with which systematists do not seem to struggle today were already well recognized by traditional taxonomists, who would use the ‘synapomorphies’ as diagnostic characters. 200

However, the phylogenetic status of several other generic taxa of Discodorididae is unknown and problematic, such as Discodoris, Anisodoris, Peltodoris, or Diaulula. Those taxa do not present clear-cut diagnostic characters that could help their delineation, and their distinction has always been vague and arbitrary. For example, the traditional diagnosis of

Zoologica Scripta, 34, 2, March 2005, pp199–224 • © The Norwegian Academy of Science and Letters 2005

B. Dayrat & T. Gosliner • Species names and metaphyly

Peltodoris is the same as the diagnosis of Discodoris except that Peltodoris species have a smooth labial cuticle, whereas Discodoris species have a labial cuticle armed with jaws (e.g. Valdés 2001, 2002). Whether or not this difference has a phylogenetic significance has never been tested. The traditional diagnosis of those taxa, based on symplesiomorphies or the presence of features that also occur in other taxa, can hardly be considered as a guarantee of monophyly. Taxonomists have used the genus name Discodoris to classify species with a general discodoridid morphology but that could not be placed in any other well-distinguished generic taxon, such as Platydoris. As a result, about 80 species have been originally described in or reallocated to Discodoris, which is by far the richest discodoridid genus in terms of nominal species. In the context of a complete taxonomic revision of Discodoris and its ‘allies’, which will be published as a separate monograph (Dayrat, in prep.), hundreds of specimens have been examined, including all type specimens available, and the taxonomic status of all nominal species addressed. Many species of Discodoris are only known from their original description and have no type material known; many names were found to be junior synonyms, while the number of well-delineated and well-known species is actually low. In the present paper, we test the phylogenetic status of Discodoris and some of its ‘allies’ as well, such as Peltodoris and Anisodoris through a cladistic analysis. Our phylogenetic analysis demonstrates that some species must be transferred to other taxa with which they share synapomorphic features. For example, Discodoris dubia and D. liturata are nested with Paradoris species. However, our result also demonstrates that many species form a metaphyletic group at the base of Discodorididae, and that the genera Discodoris, Peltodoris, Anisodoris, Diaulula do not refer to monophyletic taxa. The Greek root meta (= between, near) was used by Donoghue (1985) to coin the term ‘metaspecies’, i.e. species of which the phylogenetic status is unresolved (see Crisp & Chandler 1996). Here we use the term ‘metaphyletic’ in a general meaning to describe unresolved relationships among taxa, regardless of their taxonomic level (e.g. Gauthier 1986; Mishler & Brandon 1987; Archibald 1994). Considered together, the taxa included in a metaphyletic group could potentially represent a monophyletic, paraphyletic, and even polyphyletic group. Here we refer to the absence of phylogenetic resolution among taxa as ‘metaphyly’. We wish to point out that an exhaustive survey of morphological characters was made, and several characters were described or used in a phylogenetic analysis for the first time, such as the grooved digitiform oral tentacles, the wide holes in the dorsal notum, the muscular wall around the atrium, the folded proximal part of the deferent duct, an additional third labial jaw, the grooved lateral teeth. Therefore, unresolved

relationships are not a reflection of incomplete study but rather of a current lack of characters to more fully resolve some regions of the tree. Interestingly, this may explain why traditional taxonomists have had a very hard time delineating genera and classifying species: the phylogenetic position of many discodoridid species cannot be resolved with morphological data even with a rigorous cladistic analysis. Absence of phylogenetic resolution (or poor resolution) in phylogenetic reconstruction is a common phenomenon. As a matter of fact, published trees that are entirely resolved with all nodes being strongly supported are rare. Most trees present at least some polytomies. Absence of resolution constitutes a limit beyond which phylogenetic hypotheses may not be sorted and broader evolutionary scenarios may not be proposed (e.g. Dayrat & Tillier 2003). We believe that metaphyly should be taken into consideration as part of phylogenetic results: it helps draw a line between what we know and what we do not know, a distinction that is critical to any scientific progress; it helps emphasize which regions of the tree need to be studied more closely. One of the most direct consequences of metaphyly is that it jeopardizes classification. The fact that several species could not be placed in any subclade of Discodorididae has a major nomenclatural implication: it was not possible to place them into a monophyletic taxon of genus rank. For example, Discodoris boholiensis Bergh, 1877, a species that belongs to the metaphyletic group at the base of Discodorididae, should be called Discodorididae boholiensis (Bergh, 1877). However, this is not permitted by the International Code of Zoological Nomenclature (it also is prohibited in the other nomenclatural codes). Here we explore how to handle the fact that no genus name may be selected as the first part of a species binomial, both in the context of the current rank-based Linnaean nomenclature and in a rank-free nomenclatural system, such as the PhyloCode. Our primary goal here is to stimulate discussion on a problem that we believe merits wide debate: metaphyly in phylogenetic reconstruction and its impact on species nomenclature.

Materials and Methods Preparation of specimens Specimens were obtained from several institutions indicated by the following abbreviations: Australian Museum, Sydney (AM), Bernice P. Bishop Museum, Honolulu, Hawaii (BPBM), California Academy of Sciences, San Francisco (CASIZ), Hancock Museum, Newcastle-upon-Tyne (HM), Museum of Comparative Zoology, Cambridge, Massachusetts (MCZ), Museo Nacional de Ciencias Naturales (MNCN), Muséum national d’histoire naturelle, Paris (MNHN), National Museum of Victoria (NMV), Natural History Museum of the Smithsonian Institution, Washington DC (NHMSI), Zoological Museum, University of Copenhagen (ZMUC).

© The Norwegian Academy of Science and Letters 2005 • Zoologica Scripta, 34, 2, March 2005, pp199–224

201

Species names and metaphyly • B. Dayrat & T. Gosliner

Specimens collected by T. M. G. in the intertidal zone or by scuba were all examined alive, relaxed in isotonic MgCl2, fixed in Bouin’s fixative for a few days, and then transferred to ethanol 75%. Pictures of specimens alive were taken for most of the material collected by T. M. G. Internal anatomy was examined under a dissecting microscope and drawn with a camera lucida. Hard parts prepared for scanning electron microscopy (SEM), such as radulae, jaws, spines, stylets and so on, were cleaned in NaOH if necessary, rinsed in distilled water, sputter-coated with gold-palladium, and examined in a LEO SEM (CAS SEM-laboratory). Soft parts, such as mantle tissue, rhinophores and penial papillae, were dehydrated in ethanol and critical point dried before coating. Character coding The data matrix (49 terminals and 29 characters) is provided in Appendix 1. Characters are described in Appendix 2. Primary homology between similar structures was always assumed, except when those structures were placed at different positions or could be found together in the same individual. For example, a primary homology was assumed between the star-shaped gill opening found in Platydoris argo (Linnaeus, 1767) and Asteronotus caespitosus (van Hasselt, 1824). On the contrary, no primary homology was assumed between regular tubercles and some long conical papillae such as those found in some species of Taringa because the dorsal notum of those species bears both papillae and regular tubercles. Character coding was established from re-examination of type material, original descriptions, and observation of new specimens. Secondary descriptions based on specimens other than the type material were also used, albeit with much caution, in order to avoid problems of misidentification. Character coding obtained from the literature was always verified except in a few cases where material was not available. All species will be fully described in a lengthier monograph (Dayrat, in prep.). However, new characters introduced here are illustrated in order to avoid confusion (Figs 4 –7). Taxon sampling Our taxon sampling (see Appendix 3) depended on our primary goal, i.e. to establish the position of some anatomically well-known Discodoris species and their relationships with other discodoridid species, in particular Anisodoris and Peltodoris species. In total, 10 species currently classified in Discodoris were sampled: D. boholiensis Bergh, 1877, D. concinna (Alder & Hancock, 1864), D. dubia Bergh, 1904, D. evelinae Marcus, 1955, D. lilacina (Gould, 1852), D. liturata Bergh, 1905, D. maculosa Bergh, 1884, D. rosi Ortea, 1979, D. rubra Bergh, 1905, and D. schmeltziana Bergh, 1877. Four Peltodoris species were sampled: P. atromaculata Bergh, 1889, P. mauritiana Bergh, 1889, P. fellowsi Kay & Young, 1969 and P. lancei Millen & Bertsch, 2000. 202

Three Anisodoris species were sampled: A. nobilis (MacFarland, 1905), A. lentiginosa Millen, 1982, and A. punctuolata (d’Orbigny, 1837). The latter, which represents the type species of Anisodoris, was reallocated to Diaulula by Valdés & Muniaín (2001). Tayuva ketos Marcus & Marcus, 1967 and Sebadoris nubilosa (Pease, 1871), species of two monotypic genera, were also sampled. Additional species were sampled in Asteronotus, Halgerda, Hoplodoris, Carminodoris, Thordisa, Paradoris, Geitodoris, Diaulula, Rostanga, Taringa, Platydoris, and Jorunna. Taxon sampling was established following the ‘domain of definition’ of characters (Dayrat & Tillier 2000, 2002): taxa should be sampled if they bear characters introduced in the analysis and if they increase the morphological diversity in the data matrix. For example, species previously described in the genus Platydoris needed to be sampled because they have grooved digitiform oral tentacles similar to the tentacles observed in Discodoris dubia Bergh, 1904 and Paradoris indecora (Bergh, 1881); however, increasing the number of Platydoris species would have led to redundancy and we limited our sampling to only two of them, Platydoris ellioti (Alder & Hancock, 1864) and Platydoris argo (Linnaeus, 1767). Eight outgroups (see Appendix 3) were sampled among chromodorids and doridids, two cryptobranch groups external to Discodorididae (Valdés 2002). Cladistic analyses The data matrix was edited in MacClade 4.0 (Maddison & Maddison 2000). Heuristic tree searches were performed in PAUP* 4.0b (Swofford 2000) on a Power Mac G4, and using random stepwise-addition sequence (100 replicates) and tree-bisection-reconnection (TBR) branch-swapping options. Characters were all treated as unordered, undirected, and unweighted. Support of clades was assessed by Bremer support analyses (Bremer 1994) with the same settings as in the general tree searches. Character transformations in the strict consensus tree were viewed with MacClade.

Results Hierarchical pattern of relationships in the strict consensus tree The result of the phylogenetic reconstruction is illustrated in the strict consensus tree of 2012 equally parsimonious trees of 51 steps (Fig. 2).

Clade A. This clade corresponds to the ingroup, Discodorididae, and was recently proposed by Valdés (2002). It is confirmed here, although two characters (#2 and #24) supporting its monophyly were coded differently. It includes, in total, 17 taxa of which the relationships are unresolved: five clades B, C, D, E, and J, and 12 species previously currently classified within the genera Discodoris, Peltodoris, Anisodoris, Archidoris, Sebadoris, Halgerda, and Asteronotus.

Zoologica Scripta, 34, 2, March 2005, pp199–224 • © The Norwegian Academy of Science and Letters 2005

B. Dayrat & T. Gosliner • Species names and metaphyly

Fig. 2 Strict consensus tree of 2012 equally most parsimonious trees of 51 steps. Letters A to N indicate the 14 clades that constitute the hierarchical pattern resulting from the analysis. Terminals are species designated by their current Linnaean names. Numbers between parentheses are the Bremer support values.

Clade B. This new clade is supported by two synapomorphic characters described here for the first time: the presence of a muscular wall around the atrium (#20), and a wide folded proximal deferent duct (#26). It includes Discodoris lilacina, D. maculosa, and Tayuva ketos. Clade C. This clade is supported by two characters traditionally used by taxonomists to classify species in the genus Thordisa: the papillae on the notum (#6) and the pectinate outermost lateral teeth (#16). It includes two species previously included within the genus Thordisa. Clade D. This clade is supported by three synapomorphies, one of which is the presence of large and flat tubercles on the notum (#5). It includes one species previously included in the

genus Hoplodoris and another in the genus Carminodoris. Fahey & Gosliner (2003) recently showed that several species of these two genera are inter-nested. Clade E. This new clade is supported by a synapomorphic character described here for the first time: the presence of wide mantle holes in the notum (#10). It includes the clades F and I. Clade F. This new clade is supported by several characters described here for the first time, such as the grooved digitiform oral tentacles (#3), the presence of a third ventral jaw (#11, character state 2), the elongated radula (#12), the orientation of the radular rows at an angle > 45° from the rachidian axis (#13), and the grooved lateral teeth (#14). It includes clade G and Discodoris liturata.

© The Norwegian Academy of Science and Letters 2005 • Zoologica Scripta, 34, 2, March 2005, pp199–224

203

Species names and metaphyly • B. Dayrat & T. Gosliner

Clade G. This new clade is supported by the presence of large and flat tubercles on the notum (#5), a character traditionally used by taxonomist to delineate the genus Paradoris. It includes clade H and Discodoris dubia. Clade H. This clade is supported by the presence of indecoralike stylet sacs (#24). It includes three species previously classified in the genus Paradoris. Clade I. This clade is supported by the presence of spatulate marginal teeth (#15), a character traditionally used by taxonomists (e.g. Schrödl 2000; Muniaín 2001) to delineate the genus Geitodoris. It only includes species previously included within this genus. Clade J. This clade corresponds to what Valdés & Gosliner (2001) called the ‘caryophyllidia-bearing dorids.’ Its monophyly is only supported by two characters: the ciliated apex (#7) and the crown of spicules surrounding the caryophyllidia (#8). It includes four clades K, L, M and N, and the species Peltodoris lancei, P. fellowsi, Diaulula sandiegensis, and Anisodoris punctuolata. Clade K. This clade is supported by a character traditionally used by taxonomists (e.g. Rudman & Avern 1989) to recognize the genus Rostanga: the elongate, slender lateral teeth (#17). It includes two species previously included within the genus Rostanga and Discodoris rosi. Clade L. This clade is supported by a character traditionally used by taxonomists (e.g. Marcus 1955; Marcus & Marcus 1967; Ortea & Martinez 1992) to recognize the genus Taringa: the cuticularized penis (#24). It includes two species previously included within Taringa. Clade M. This clade is supported by two characters used by taxonomists (e.g. Dorgan et al. 2002) to recognize the genus Platydoris: the flattened body shape (#4) and star-shaped gills sheath (#9). The presence of grooved digitiform oral tentacles (#3) constitutes a new synapomorphy for this clade. It includes two species previously included within Platydoris. Clade N. This clade is supported by a character that taxonomists (e.g. Cervera et al. 1986) traditionally use to recognize the genus Jorunna, i.e. the tomentosa-like stylet sac (#21). It includes two species previously included within Jorunna. In addition to these clades, Note that a particular clade found in the majority-rule consensus tree might need attention in the future: in 73% of the equally most parsimonious trees, A. caespitosus, C. grandiflora and H. estreleyado were included in a monophyletic group supported by the presence of a hollow spine in the accessory gland. 204

Metaphyletic relationships in the strict consensus tree There are two metaphyletic groups in the strict consensus tree. Twelve species form a metaphyletic group at the base of Discodorididae and cannot be placed in any subclade of it (Figs 2, 3). Five of these were previously placed in the genus Discodoris (boholiensis, concinna, evelinae, schmeltziana, and rubra), two in the genus Peltodoris (atromaculata and mauritiana), and two in the genus Anisodoris (nobilis and lentiginosa). This metaphyletic group also includes the unique valid species of the genus Sebadoris (S. nubilosa), Asteronotus caespitosus, and Halgerda dalanghita. The second metaphyletic group, at the base of clade J, the ‘caryophyllidia-bearing dorids’, includes four species (Figs 2, 3): Peltodoris fellowsi and P. lancei, Anisodoris punctuolata, and Diaulula sandiegensis. Metaphyly has two major implications: (1) the phylogenetic status of Discodoris, Peltodoris, Anisodoris and Diaulula cannot be addressed with our current morphological data; (2) no genus name can be selected as the first part of the binomials of all the species that belong to these two metaphyletic groups. Clade support Node support, which is evaluated owing to Bremer values (Bremer 1994), is universally low. The most robust node in the consensus tree is clade F (Bremer support of 5). Two clades (clade A and M) are supported with a support value of 3; clades B, C, D, H, I and L with a value of 2; and clades E, G, J, K and N with a value of 1.

Discussion Resolution Addressing the level of resolution requires discussing the topology of a tree and the robustness of the nodes. However, there is no clear-cut borderline separating robust and nonrobust clades. Comparing Bremer support values only makes sense for a given data set and its resulting tree: these values are relative, not absolute. For example, a value of 1 may be seen as low if values vary from 1 to 40 for a given data set, and ‘less low’ if, for another data set, values only vary from 1 to 5, as found here. Bremer support values provide the differential support of each node in a particular strict consensus tree. They only allow us to say that clade F (value of 5) is more supported by the data than clades E (value of 1). However, low Bremer support values indicate that some nodes are unstable (i.e. less stable than other nodes in the same tree). The fact that some clades obtained in the strict consensus tree are less strongly supported than some other nodes does not mean that they are not important. Rather, they constitute hypotheses that will need to be tested in the future with additional data. The two clades E and J, supported with a value of 1, merit discussion in detail here because of the important synapomorphies supporting them.

Zoologica Scripta, 34, 2, March 2005, pp199–224 • © The Norwegian Academy of Science and Letters 2005

B. Dayrat & T. Gosliner • Species names and metaphyly

Fig. 3 Strict consensus tree of 2012 equally most parsimonious trees of 51 steps. Letters A to N indicate the 14 clades that constitute the hierarchical pattern resulting from the analysis. Species are designated by binominal names. These names could be (see Discussion, and Table 1: (1) binominal names under our current code of nomenclature, slightly modified for allowing taxonomists to use a supra-generic taxon name as the first part of a species name when no generic taxon name could be found; (2) common names of species named with Lanham’s method, under the current rank-based system; (3) common names of species named with Lanham’s method (slightly modified according to Dayrat et al. 2004), under a rank-free nomenclatural system such as the PhyloCode. Numbers above branches indicate character transformations supporting the nodes; numbers after the column indicate character states (see data matrix, Appendices 1 and 2). Bold numbers indicate characters for which there is a single transformation in the tree; regular numbers indicate characters for which there is more than one transformation in the tree.

Even though clade E is only supported by a Bremer support value of 1, the presence of wide holes in the notum constitutes an important synapomorphy. Indeed, those holes are accompanied by the presence of mantle glands. Acidic secretions (which are exceptional in opisthobranchs) were only mentioned for species of Geitodoris (see, Avila 1995). Note that the only non-Geitodoris species in which acidic secretions were mentioned is Discodoris palma, but this is most likely to be due to a misidentification because the specimens identified as

palma by Thompson (1975) have spatulate outermost lateral teeth, which is a diagnostic character of Geitodoris species. Our phylogenetic result suggests that wide holes in the notum are homologous. However, this homology is weak and should be investigated with help from thorough comparative chemical and histological studies; the mantle secretions in Paradoris indecora, the only species of Paradoris for which the mantle secretions have been studied so far, do not seem to be acidic (Avila 1995).

© The Norwegian Academy of Science and Letters 2005 • Zoologica Scripta, 34, 2, March 2005, pp199–224

205

Species names and metaphyly • B. Dayrat & T. Gosliner

Fig. 4 A–F. Illustration of characters (see Appendix 2). —A. Dorsal notum bearing a long conical papilla (#6) and caryophyllidia (#7, #8), Taringa sp., CASIZ 078406, scale = 100 µm. —B. Detail of the dorsal notum bearing a caryophyllidium (#7, #8), cilia (#8), and small holes (#10), pulchra MacFarland, 1905, CASIZ 069842, scale = 30 µm. —C. Detail of rhinophoral lamellae bearing ciliae (#8) and small holes (#10), boholiensis Bergh, 1877, CASIZ 158768, scale = 40 µm. —D. Detail of the dorsal notum bearing cilia (#8), small and wide holes (#10), heathi MacFarland, 1905, CASIZ 082091, scale = 20 µm. —E. Detail of the dorsal notum bearing cilia (#8), small and wide holes (#10), mavis Marcus & Marcus, 1967, CASIZ 072930, Scale = 30 µm. —F. Detail of the dorsal notum bearing cilia (#8), small and wide holes (#10), liturata Bergh, 1905, CASIZ 097595, Scale = 40 µm.

In the present analysis, monophyly of clade J (‘caryophyllidiabearing dorids’) is probably weakened by the fact that many characters (e.g. #3, #9, #11, #17, #21, #28) vary among all discodoridids, whether bearing caryophyllidia or not. The 206

monophyly of the caryophyllidia-bearing dorids was initially proposed by Valdés & Gosliner (2001) and then tested again by Valdés (2002) with a larger genus-level taxon sampling. Valdés (2002) stated that he had confirmed the monophyly of

Zoologica Scripta, 34, 2, March 2005, pp199–224 • © The Norwegian Academy of Science and Letters 2005

B. Dayrat & T. Gosliner • Species names and metaphyly

Fig. 5 A–C. Illustration of characters (see Appendix 2). —A. Left outermost lateral teeth (#14), liturata Bergh, 1905, CASIZ 097595, scale = 30 µm. —B. Right outermost lateral teeth (#14), liturata Bergh, 1905, CASIZ 097595, scale = 20 µm. —C. Ventral view of the tooth groove of a lateral

tooth (#14), indecora Bergh, 1884, ZMUC (Istria), scale = 20 µm.

Fig. 6 A–D. Illustration of characters (see Appendix 2). —A. Middle lateral teeth (#14), indecora Bergh, 1884, ZMUC (Istria), scale = 20 µm. —B. Outermost lateral pectinate teeth (#16), bimaculata Lance, 1966; CASIZ 070575, scale = 10 µm. —C. Outermost lateral spatulate teeth (#15), heathi MacFarland, 1905, CASIZ 069140, scale = 40 µm. —D. Lateral hair-like elongate and slender teeth (#17), rosi Ortea, 1979, CASIZ 166874, scale = 20 µm.

© The Norwegian Academy of Science and Letters 2005 • Zoologica Scripta, 34, 2, March 2005, pp199–224

207

Species names and metaphyly • B. Dayrat & T. Gosliner

Fig. 7 A, B. Illustration of characters #19, # 20 and #26. —A. Lateral view of the digestive system of nubilosa Pease, 1871; CASIZ 074254, scale = 10 mm. —B. Reproductive system of maculosa Bergh, 1884, MNCN 15.05/801, scale = 2.5 mm. Abbreviations: a, anus; amp, ampulla; bc, bursa copulatrix; d, deferens duct; dg, digestive gland; e, oesophagus; fd, folded proximal part of the deferent duct; fgm, female gland mass; i, intestine; mva, muscular wall of the atrium; pr1, prostate part 1; pr2, prostate part 2; rs, receptaculum seminis; sy, syrinx; st; stomach; vd, vaginal duct.

‘caryophyllidia-bearing dorids.’ However, the monophyly of ‘caryophyllidia-bearing dorids’ he obtained in his consensus tree is actually due to an error of coding in his data matrix: according to his own data, his character #65 was erroneously coded as ‘1’ (prostate divided) instead of ‘0’ (prostate not divided) for Atagema, Alloiodoris, Nophodoris and Thorybopus. The monophyly of caryophyllidia-bearing dorids obtained by Valdés actually collapses when the coding of his character #65 is corrected for these four terminals (the rest of the data matrix being unchanged). This demonstrates that the monophyly of caryophyllydiabearing dorids must not be taken for granted. It certainly is an interesting hypothesis, but weakly supported by our current morphological knowledge. In that regard, we wish to point out that all caryophyllidia are not strictly identical (e.g. Marcus 1955; Foale & Willan 1987; Valdés & Gosliner 2001; Dorgan et al. 2002). Some are pedunculate; some are flat with the crown of spicules laying on the notum; some bear spicules that are free from the base of the peduncle up to the apex; some bear spicules embedded within the peduncle for part of their length. The phylogenetic significance of those variations, which is beyond the present analysis, will certainly need to be tested in the future. Also, the apex of regular tubercles (i.e. non-caryophyllidia) may also be ciliated, exactly like caryophyllidia (see character #8, Appendix 1). However, the infra-specific variation of this character is difficult to address because of the influence of preservation on the morphology of the notum. This feature, phylogenetically critical, will need to be further studied. 208

Considering node support values may increase absence of resolution Overall, low node support values tend to decrease the level of resolution obtained in the strict consensus tree. In particular, not considering clade J (‘caryophyllidia-bearing dorids’) creates one single metaphyletic group at the base of the Discodorididae, instead of a metaphyletic group at the base of Discodorididae and another at the base of ‘caryophyllidiabearing dorids’ (Figs 2, 3). Our positive phylogenetic knowledge concerning those 16 species is limited to the fact that they belong to the clade Discodorididae (Figs 2, 3). No ‘generic’ level affinities can be proposed for any of these species, with the exceptions of dalanghita Fahey & Gosliner, 1999 and caespitosus van Hasselt, 1824. Indeed, these two species would certainly be part of a subclade of Discodorididae if other species previously described within Halgerda (for dalanghita) and Asteronotus (for caespitosus) had been sampled. This is the reason why we do not consider that our result questions their supra-specific relationships, despite their position at the base of the Discodorididae in our phylogenetic tree (Fig. 3). However, no genus name can be selected for the 14 other species. We shall discuss how metaphyly can be handled, both in a rank-based Linnaean system and a rank-free system. Evolutionary history of characters Character transformations were mapped on to the strict consensus tree because that is the only tree that provides the components common to all the equally parsimonious cladograms (e.g. Anderberg & Tehler 1990). Unambiguous

Zoologica Scripta, 34, 2, March 2005, pp199–224 • © The Norwegian Academy of Science and Letters 2005

B. Dayrat & T. Gosliner • Species names and metaphyly

character transformations are all presented in Fig. 3. Node support values mainly affect the evolutionary history of characters #7, #8, and #10. Indeed, the presence of a ciliated apex on tubercles (#7) and of a crown of spicules (#8) only originate from a single event if clade J (‘caryophyllidia-bearing dorids’) is accepted. If the monophyly of clade J is not accepted, then multiple independent events are required for characters #7 and #8. Wide holes in the notum (#10) only originate from a single event if clade E is accepted. If the monophyly of clade E is not accepted, then two independent transformations are necessary to explain the presence of wide holes in Geitodoris and Paradoris species. Several hypotheses of primary homology between structures were rejected by the cladistic test of our analysis (characters #3, #6, #16, #17, #18, #19, #22, #28). For example, pectinate outermost lateral teeth found in Taringa aivica are morphologically similar to ‘pectinate’ teeth found in Thordisa species. Therefore, all pectinate teeth had been considered a priori primary homologous (character #16) and a single character state was considered. However, the analysis suggested that pectinate teeth originated from two independent events at the level of Discodorididae with our taxon sampling. Note that, although character #16 presents some homoplasy at the level of all Discodorididae, the independent acquisition of pectinate teeth constitutes locally a synapomorphy for Thordisa. Further studies will need to address in greater depth the homology and the history of pectination(s) of the outermost lateral teeth. The same comment is valid for characters #3, #6, #17, #18, #19, #22, and #28. Character transformations of characters #11 and #21 are highly uncertain if one considers the entire Discodorididae. The number of times paired jaws (#11) were acquired and lost cannot be determined. It seems that accessory glands (#21) were acquired and lost independently several times too. Note that the accessory gland could provide interesting phylogenetic information locally. More detailed studies are needed to help separate those vestibular glands into several primary homologous structures instead of one general structure. In particular, histological studies would probably be helpful. Classification A classification is a series of names representing the hierarchical pattern of relationships observed in a cladogram (Hennig 1966). Names are already available for most of the clades found here. Also, not all clades need to be named and considered as taxa, especially unstable clades.

combination D. ketos based on the argument that the anatomy of T. ketos does not differ from Discodoris species. Actually, T. ketos differs greatly from most Discodoris species by the presence of a muscular wall around the atrium (#20) and a folded proximal deferent duct (#26), two characters that Valdés (2002) did not consider. According to our result, T. ketos should not be reallocated to Discodoris, but two Discodoris species should actually be transferred to Tayuva. Clade C. Corresponds to the genus-level taxon Thordisa. Clade D. Corresponds to the genus-level taxon Carminodoris. Fahey & Gosliner (2003) recently reallocated all Carminodoris species to the genus Hoplodoris, based on the argument that the type species, H. desmoparypha Bergh, 1880, was a synonym of a Carminodoris species, C. grandiflora (Pease, 1860). However, this statement is erroneous; H. desmoparypha is anatomically distinct (see e.g. Bergh 1880) from all the specimens observed by Fahey & Gosliner (2003) because the accessory vestibular spine is straight in grandiflora and curved in desmoparypha (Dayrat, in prep.). Therefore, instead of reallocating some Carminodoris species to Hoplodoris, it is actually some Hoplodoris species that need to be transferred to Carminodoris. Clade E. Constitutes an important hypothesis that will need to be tested. However, we consider that it is not necessary to create a new name for an unstable clade. Clade F. Corresponds to the genus-level taxon Paradoris, to which Discodoris liturata and D. dubia are reallocated. Clade G. Does not need to be named. Clade H. Does not need to be named. Clade I. Corresponds to the genus-level taxon Geitodoris. Clade J. Constitutes a hypothesis that will need to be tested in the future. However, we consider that it is not necessary to create a new name for an unstable clade. Discodoridids that present caryophyllidia on the dorsal notum can continue to be colloquially called ‘caryophyllidia-bearing dorids’. Clade K. Corresponds to the genus-level taxon Rostanga, to which Discodoris rosi is reallocated.

Clade A. Corresponds to the Discodorididae (see Valdés 2002).

Clade L. Corresponds to the genus-level taxon Taringa.

Clade B. Can be named Tayuva since it includes the only species of the genus, T. ketos, and two species of Discodoris, D. lilacina and D. maculosa. Valdés (2002) proposed the new

Clade M. Corresponds to the genus-level taxon Platydoris. Clade N. Corresponds to the genus-level taxon Jorunna.

© The Norwegian Academy of Science and Letters 2005 • Zoologica Scripta, 34, 2, March 2005, pp199–224

209

Species names and metaphyly • B. Dayrat & T. Gosliner

Handling metaphyly in the current rank-based code of nomenclature: a binominal solution Under the current code of nomenclature, there are three main solutions to naming species at the base of Discodorididae. The first is based on the concept of the type species. The only clade that includes Discodoris boholiensis Bergh, 1877, the type species of Discodoris, is the ingroup (clade A) that actually includes all discodoridids. Therefore, Discodoris must include all discodoridid species if the taxon name is to correspond to a clade. Consequently, the name Discodoris should be selected as the first part of the binominal of all discodoridid species. Of course, this solution is highly unsatisfactory because it would lead us to lose a great deal of phylogenetic information concerning relationships among discodoridids. The second solution is to continue to use former names without changing them: Discodoris boholiensis, Anisodoris lentiginosa, Peltodoris atromaculata, Diaulula sandiegensis, and so on. However, this solution presents a major drawback: none of those names would reflect phylogenetic reality. More importantly, retaining those names conveys misinformation about supra-specific relationships. For example, Discodoris rosi is a member of Rostanga (clade K) while D. boholiensis is a member of Discodorididae (clade A). The third is to select the name Discodoris for all species included in the metaphyletic group at the base of Discodorididae. However, this presents a major drawback: Discodoris would then not correspond to a clade. A non-monophyletic taxon could be indicated by an asterisk to its name: names would then take the form Discodoris* boholiensis Bergh, 1877. Overall, these three solutions are not satisfactory because they tend to erase all positive phylogenetic knowledge (first solution), to convey information about supra-specific affinities that we know is wrong (second), or to use taxon names that do not refer to clades (third). The reason why metaphyly cannot be handled in the current rank-based Linnaean system is not related to the fact that species names must be binominal, but rather to the fact that the first part of binomials must be a generic name. Another solution would be to modify the current code. An article could be added to the current code, allowing taxonomists to select a taxon name of supra-generic level when generic relationships are unknown. For example, this would allow taxonomists to change the name Discodoris boholiensis Bergh, 1877 to Discodorididae boholiensis (Bergh, 1877). Possible names involving a supra-generic taxon name as the first part of binomials are provided in Fig. 3. Although names resulting from such new combinations are binominal, they are in complete contradiction with the current code that requires the first part of binomials to be generic names. This solution also presents a drawback: it potentially increases the chance of secondary homonymy. Indeed, many 210

species with the same epithet belong to Discodorididae, such as Peltodoris mauritiana and Hoplodoris mauritiana, Discodoris dubia and Thordisa dubia, or Discodoris rubra, Diaulula rubra, Halgerda rubra, Rostanga rubra, and Platydoris rubra. Several of these names would not be homonymous: Discodorididae mauritiana would be distinct from Hoplodoris mauritiana; Paradoris dubia would still be distinct from Thordisa dubia. A possible solution to handle secondary homonymy would be to add the genus name where the species had been described originally: Discodorididae (Discodoris) rubra would then be distinguished from Discodorididae (Diaulula) rubra. Other solutions to address secondary homonymy could certainly be proposed. Handling metaphyly in the current rank-based code of nomenclature: Michener’s uninominal method Another solution to the problem of metaphyly would be to name species without having to mention any supra-specific relationships. Species names would then only be used to designate unique species, but not to convey information about species relationships. Michener (1964) and Lanham (1965), whose primary goal was that species names could be permanently stable, first proposed this solution in the mid-1960s. These authors proposed two possible forms of species names. Under both forms, species names are unique, stable, and do not convey information about supra-specific relationships: (1) names are binominal in form but uninominal in function, and uniqueness is guaranteed, as in the binominal nomenclature, by the combination of a species epithet and a genus-level taxon name (Michener 1964); (2) names are uninominal (‘epithet-based’) in form and function, and uniqueness is guaranteed by an additional tag attached to the species epithet (Lanham 1965). Following Michener’s (1964) method (Table 1), the name Discodoris boholiensis would simply become Discodoris-boholiensis. A dash would indicate that the binominal name is just binominal in form but would function as a uninominal name in which the first part could never be changed. The role of the genus-level taxon name Discodoris is then to guarantee uniqueness of the stable name Discodoris-boholiensis, not to indicate that Discodorisboholiensis belongs to the taxon Discodoris. Michener’s form of species names would still oblige taxonomists to select a genus-level taxon name for all species binomials. It therefore handles metaphyly as poorly as the current Linnaean binominal nomenclature, and also conveys potential misinformation about supra-specific relationships. Handling metaphyly in the current rank-based code of nomenclature: Lanham’s uninominal method Under Lanham’s (1965) method for species names (Table 1), the species name Discodoris boholiensis would become boholiensis Bergh, 1877: 519. This name is permanently stable (because

Zoologica Scripta, 34, 2, March 2005, pp199–224 • © The Norwegian Academy of Science and Letters 2005

B. Dayrat & T. Gosliner • Species names and metaphyly

Table 1 Names of two species following different methods, in a rank-based or rank-free system: our current binominal Linnaean nomenclature,

a binominal but not strictly Linnaean nomenclature (the first part of the binomen can be a name of supra-generic rank), Michener’s (1964) nomenclature, Lanham’s (1965) nomenclature (with both official and ‘common’ names), and 13 methods A through M proposed by Cantino et al. (1999). The actual supra-specific relationships of those two species (see Fig. 3) can only be accurately represented by the modified binominal nomenclature and by Lanham’s common names, under both a rank-based system and a rank-free nomenclature. METHODS

Current rank-based system Binominal and strictly Linnaean Binominal but not strictly Linnaean (proposed here) Michener Lanham (official) Lanham (common)

Rank-free system A B, C D E F G H I J K L

M Lanham (official) Lanham (common)

Discodoris dubia (only combination proposed so far) Paradoris dubia (new combination proposed here)

Montereina nobilis, Anisodoris nobilis, Peltodoris nobilis (three combinations proposed so far)

Paradoris dubia (Bergh, 1904) Discodoris-dubia dubia Bergh, 1904 Paradoris dubia Bergh, 1904; Paradoris dubia Bergh; Paradoris dubia

Discodorididae nobilis (MacFarland, 1905) Montereina-nobilis, Anisodoris-nobilis, Peltodoris-nobilis nobilis MacFarland, 1905 Discodorididae nobilis MacFarland, 1905 Discodorididae nobilis MacFarland; Discodorididae nobilis, etc.

Discodoris dubia discodoris dubia, Discodoris-dubia, Discodoris.dubia Discodoris dubia, Discodoris-dubia, Discodoris.dubia discodoris-dubia, discodoris.dubia Discodoris.dubia Discodorisdubia discodorisdubia Discodoris dubia discodoris dubia, Discodoris-dubia, Discodoris.dubia, etc. dubia, dubia2, dubia.2, Discodoris dubia2, Discodoris/dubia2 (Discodoris) dubia2 dubia238719, dubia.238719, Discodoris dubia238719, Discodoris/dubia238719 (Discodoris) dubia238719, etc. dubia, dubia238719, Discodoris dubia, Discodoris/dubia (Discodoris) dubia, etc. dubia Bergh, 1904 Paradoris dubia Bergh, 1904; Paradoris dubia Bergh; Paradoris dubia

Montereina nobilis, Anisodoris nobilis, Peltodoris nobilis montereina nobilis, Anisodoris-nobilis, Peltodoris.nobilis Montereina nobilis, Anisodoris-nobilis, Peltodoris.nobilis anisodoris-nobilis, peltodoris.nobilis Peltodoris.nobilis Montereinanobilis, Anisodorisnobilis, Peltodorisnobilis montereinanobilis, anisodorisnobilis, peltodorisnobilis Montereina nobilis, Anisodoris nobilis, Peltodoris nobilis, Discodoris nobilis, etc. montereina nobilis, Anisodoris-nobilis, Peltodoris.nobilis, etc. nobilis, nobilis2, Montereina nobilis2, Anisodoris/lancei2 (Peltodoris) lancei2, etc. nobilis156303, nobilis. 156303, Montereina nobilis156303, Anisodoris/nobilis156303 (Peltodoris) nobilis156303, etc. nobilis, nobilis156303, Montereina nobilis, Anisodoris/nobilis (Peltodoris) nobilis

it does not include any mention of supra-specific relationships). The uniqueness of this name is guaranteed by the combination of the specific epithet, the author’s name, the date of publication of the species name, and the page number where the name first appeared (Dayrat et al. 2004). Contrary to what Lanham wrote, his method for species names is not strictly ‘uninominal’ because the specific epithet must be attached to some additional information in order to guarantee uniqueness. Also, Lanham’s method may be slightly modified to guarantee uniqueness in all cases (Dayrat et al. 2004). Because species names, following Lanham’s method, do not convey any information about supra-specific relationships, the method handles metaphyly ideally. For communication purposes one might add a clade name in front of the specific epithet (Dayrat et al. 2004); the official name of Discodoris boholiensis would become boholiensis Bergh, 1877,

nobilis MacFarland, 1905 Discodorididae nobilis MacFarland, 1905 Discodorididae nobilis MacFarland Discodorididae nobilis, etc.

although names such as Discodorididae boholiensis Bergh, 1877, or Discodorididae boholiensis could also be used. These names, however, would then be equivalent to common names: they would not guarantee uniqueness, but would represent our current phylogenetic knowledge (i.e. boholiensis Bergh, 1877 belongs to the clade Discodorididae). Homonymy, as we face it under the current nomenclature, ceases to be a problem using Lanham’s method: two species with the same epithet can be classified in the same supraspecific taxon, because supra-specific taxon names do not help guarantee the uniqueness of species names. For example, Diaulula rubra would become rubra Bergh, 1905: 120, and Discodoris rubra would become rubra Bergh, 1905: 104. If ‘common’ names are identical and confusing, then one can still refer to these species by adding the author’s name, and, if necessary, by adding between parentheses the genus name where species had been described originally.

© The Norwegian Academy of Science and Letters 2005 • Zoologica Scripta, 34, 2, March 2005, pp199–224

211

Species names and metaphyly • B. Dayrat & T. Gosliner

Lanham’s method for species names has, as far as we know, never been considered since it was proposed 40 years ago. The main reason is probably that taxonomists never wanted to abandon the well-accepted binominal nomenclature according to which the uniqueness of species names is guaranteed by a combination of a generic name and a specific epithet. Also, the implementation of Lanham’s method in the current code would probably require taxonomists to change many articles in the ICZN: for example, how would the concept of type species apply? On clades and ranks: towards a rank-free nomenclatural system Hierarchy and ranking must not be confused. The representation of the hierarchy of a cladogram within a classification does not require ranks; it just requires hierarchy. The differences between hierarchy and ranking are profound. Hierarchy reflects the history of successive branching events; clades are hypotheses of real (i.e. historical = biological) entities. Ranks are arbitrary, subjective, and have no biological meaning, which is one of the reasons why the proponents of the PhyloCode (Cantino & De Queiroz 2003; www.ohiou.edu/ phylocode/) favour the use of a rank-free nomenclature in which only two kinds of taxa are distinguished: species and clades, the latter being monophyletic groups of species. We agree that a rank-free nomenclatural system is great progress. One may argue that the fact that genus names are mandatory in the current binominal species nomenclature is not an issue. For example, Discodoris dubia Bergh, 1904 is shown here to be a member of a clade including species previously classified in the genus Paradoris (Fig. 3; Table 1). In the context of the traditional nomenclature, the species name of Discodoris dubia would simply become Paradoris dubia (Bergh, 1904). This name appropriately indicates that the species dubia Bergh, 1904 belongs to the taxon Paradoris, and not to the taxon Discodoris. However, the reason why a binominal name does not fundamentally jeopardize the application of the first principle of phylogenetic systematics (all taxon names must refer to clades) is that, in this particular case, the name Paradoris refers to a clade. The fact that the taxon name is assigned to the genus rank is not important: dubia can be reallocated from Discodoris to Paradoris because Paradoris refers to a clade, not because it relates to a genus-level taxon. In addition, there is no biological criterion to assign a given clade to the genus rank. Even if we know, for example, that the name Paradoris might define a clade, there is no available biological comprehensive criterion to help assign this taxon to the genus rank. Ranking clade names is arbitrary, and therefore biologically meaningless. The upcoming implementation of the rank-free PhyloCode presents a major problem for binominal species names. Indeed, the current Linnaean nomenclature is incompatible 212

with a system in which the genus rank is no longer mandatory. ‘The Linnaean binomial nomenclature is logically incompatible with the phylogenetic nomenclature of de Queiroz and Gauthier: the former is based on the concept of genus, thus making this rank mandatory, while the latter is based on phylogenetic definitions and requires the abandonment of mandatory ranks. Thus, if species are to receive names under phylogenetic nomenclature, a different method must be devised to name them’ (Cantino et al. 1999: 790). This incompatibility is especially problematic because the current draft of the PhyloCode does not deal with species names: it was decided at the First International Meeting of Phylogenetic Nomenclature, Paris, July 2004, that a species section should be added to the PhyloCode (Laurin & Cantino, 2004). Cantino et al. (1999) proposed 13 possible forms for species names, which they compared for various criteria. However, Lanham’s method best fits a rank-free nomenclature (Dayrat 2004; Dayrat et al. 2004) and is superior to the methods proposed by Cantino and co-authors because it guarantees continuity with the existing taxonomic literature. It allows users of species names to rapidly recognize what species is designated by a Linnaean name or its converted epithet-based name. Indeed, one can only be sure that Montereina nobilis, Anisodoris nobilis and Peltodoris nobilis are three names of the same species if one knows that the full names are Montereina nobilis MacFarland, 1905, Anisodoris nobilis (MacFarland, 1905), and Peltodoris nobilis (MacFarland, 1905), and that MacFarland has named only one species with this specific epithet in that year. This is why Lanham suggested that species names become epithet-based while preserving the combination between the species epithet, the author’s name and the date of publication. This is because this combination is needed to ascertain whether names designate distinct entities or the same nomenclatural entity. Dayrat et al. (2004) have compared Lanham’s method to those proposed by Cantino et al. (1999) for criteria such as continuity, stability, degree of ambiguity, ease of pronunciation, and the potential not to convey misinformation about relationships. Here we compare in greater depth how Lanham’s method and those other methods could handle metaphyly. Handling metaphyly in a rank-free system: the superiority of Lanham’s method Cantino et al. (1999) looked for names that could be unique, stable through time, and that would not convey information (or misinformation) about supra-specific relationships. Interestingly, they found the same two main classes of forms of species names that had been proposed earlier in the 1960s (Table 1), although neither Michener (1964) nor Lanham (1965) had suggested that ranks be abandoned. In the first class of methods for naming species, names are binominal in form but uninomial in function (methods A–J).

Zoologica Scripta, 34, 2, March 2005, pp199–224 • © The Norwegian Academy of Science and Letters 2005

B. Dayrat & T. Gosliner • Species names and metaphyly

In methods A, B, D−H, the first part of the name can never be changed: Anisodoris nobilis (e.g. methods A and D), anisodoris nobilis (e.g. method B), Anisodoris-nobilis (e.g. method C), and so on, are binominal in form but uninominal in function. In methods C, I and J, the first part of the name may change under certain conditions, i.e. when the name refers to a clade defined under the PhyloCode and which the species is known not to belong to. In the second class of methods, names are uninominal (‘epithet-based’) in form and function, and an additional tag attached to the specific epithet guarantees uniqueness (methods K−M). In methods L and M, a registration number would help distinguish vulgaris #230363 from vulgaris #2071304. Methods A to J proposed by Cantino and coauthors handle metaphyly as poorly as Michener’s (1964) method because, for the sake of stability through time, combinations between generic names and specific epithets become permanently fixed when Linnaean binomials are converted (Table 1). For example, Anisodoris nobilis (MacFarland, 1905) would become Anisodoris-nobilis when we know that Anisodoris is phylogenetically meaningless. This prevents us from communicating that nobilis MacFarland, 1905 actually belongs to Discodorididae and that Anisodoris is a phylogenetically meaningless taxon. A major drawback of methods A to J (Cantino et al. 1999; Dayrat et al. 2004) is that they do convey misinformation about supra-specific relationships (Table 1). Indeed, Discodoris rosi Ortea, 1979 would become Discodoris-rosi despite the fact that we know that it is a member of the clade Rostanga, and therefore Discodoris-rosi would be phylogenetically closer to Rostanga-pulchra than to Discodoris-boholiensis. This is a serious drawback compared to the useful flexibility of the Linnaean nomenclature. Also, most species are designated by multiple names in the literature. Taxonomists may agree on a valid species name, but they also may not. What criterion is going to help select the Linnaean generic name that will be converted or selected as the permanent taxonomic address? Will the name Montereina nobilis MacFarland, 1905 be Montereinanobilis, Anisodoris-nobilis, or Peltodoris-nobilis? Should it be the original genus name, the most common one, or the last one published? All these solutions are equally arbitrary. This is particularly striking in this case where the three generic names refer to taxa that are phylogenetically meaningless (Figs 2, 3). Methods C, I, and J also convey such misinformation about supra-specific relationships when the first part of a binomial is a name that has not been defined under the PhyloCode. For example, because the name Discodoris does not define a clade under the PhyloCode, methods C, I and J would not require that it be changed into another name when species names would be converted into a new rank-free system. These methods would transform the Linnaean binomial Discodoris rosi Ortea, 1979 into Discodoris rosi (method C), Discodoris-rosi (method I), or discodoris rosi (method J), i.e. the exact same

names as with methods A, B, and D−H. As a result, readers may believe that Discodoris.rosi and Discodoris.dubia are closely related when they are not (Figs 2, 3). Although they are epithet-based and should not convey misinformation about supra-specific relationships, methods K to M may actually convey such misinformation because they may retain phylogenetically meaningless genus names as ‘taxonomic addresses’ (Table 1). In such cases, methods C and I−M would handle metaphyly as poorly as the current binominal nomenclature. However, the main argument against such methods is that they break the continuity with the information that is necessary to trace names throughout the abundant taxonomic literature (Dayrat et al. 2004). Although it was proposed in a rank-based context, Lanham’s method best fits a rank-free nomenclature (Dayrat et al. 2004). Since there is no mandatory genus rank in the PhyloCode, it makes no sense to require the use of genus names. Lanham’s method is globally the same in a rank-based or rank-free system (Table 1). In particular, exactly as in a rank-based system, common names could be used to help communicate our knowledge of supra-specific relationships: the official name of Montereina nobilis MacFarland, 1905 would be nobilis MacFarland, 1905, although nothing should prevent us from using the common names Discodorididae nobilis MacFarland, 1905, or simply Discodorididae nobilis. Overall, Lanham’s method (once modified for uniqueness purposes) handles perfectly the absence of phylogenetic resolution because: (1) species names do not include any mention of supra-specific relationships; and (2) the mention of supraspecific relationships is optional.

Conclusion Absence of resolution (metaphyly) in phylogenetic reconstruction is not surprising. It is mainly the result of the fact that each data set has its own limit and cannot help reconstruct an entire tree. For example, we have known for some time that different molecular markers provide phylogenetic signal at different levels depending on their substitution rate. Morphology may provide informative characters for reconstructing nodes at any depth in the trees, but not for reconstructing all nodes. Metaphyly should not be seen as an indication that a phylogenetic analysis has failed, but rather as an indication that one needs additional data to resolve some regions of a tree. Besides, absence of resolution can also be due to the fact that some dichotomies have originated in a relatively short time and therefore cannot be reconstructed easily. In any case, metaphyly should by no means be denied: it constitutes highly valuable information by pointing out the questions that still need to be addressed. Therefore, absence of resolution should be considered as an integral part of the results. It can be easily detected: (1) by using a strict consensus tree when numerous equally most

© The Norwegian Academy of Science and Letters 2005 • Zoologica Scripta, 34, 2, March 2005, pp199–224

213

Species names and metaphyly • B. Dayrat & T. Gosliner

parsimonious trees are obtained, instead of an arbitrarily selected most parsimonious tree (e.g. Anderberg & Tehler 1990); (2) by addressing carefully node reliability, i.e. not taking into account the nodes that are poorly or not supported. Metaphyly may prevent taxonomists from selecting a mandatory generic name (that refers to a clade) for every species binomial. Here we have compared how new forms of species names, other than in a strictly Linnaean binominal form, could handle this issue. Interestingly, keeping or abandoning our current rank-based system does not really influence how well a form of species names handles metaphyly. Both in a rank-free and a rank-based system, methods that permanently fix Linnaean combinations handle metaphyly as poorly as the current Linnaean nomenclature. Therefore, none of them is favoured here. Under the current rank-based system, two possible new forms of species names could potentially help handle metaphyly. One solution requires the addition of an article to the current Code stating that the first part of a species binomial can be a taxon name of supra-generic rank when no name of genus rank could be found. The second is the method proposed by Lanham (1965), who argued that species names should not include any information about supraspecific relationships. Lanham’s solution, once slightly modified for uniqueness purposes (Dayrat et al. 2004), presents a great advantage: it keeps the critical information that taxonomists need to search names throughout the abundant taxonomic literature. Lanham’s solution also retains nomenclatural stability as phylogenetic resolution increases. However, both solutions would constitute a radical change to the ICZN. Lanham’s form of species name, which was originally proposed in the context of a rank-based system, also perfectly fits a rank-free nomenclature such as the PhyloCode (Dayrat 2004; Dayrat et al. 2004). Interestingly, it was agreed at the First International Phylogenetic Nomenclature Meeting, July 6– 9, 2004, that a new ‘species code’ compatible with the PhyloCode should be written and based on this form (Laurin & Cantino, 2004). Therefore, an epithet-based form is likely to become available under the rank-free nomenclatural system of the PhyloCode before it can be introduced in the traditional codes. This by no means excludes the possibility that it could also be implemented in the traditional codes in future. Finally, we would like to encourage dialogue among systematists and users of species names, to propose solutions that will transcend different codes of nomenclature, provide nomenclatural stability and recognize positive advances in understanding phylogenetic relationships.

Acknowledgements We thank all the persons who collected the specimens studied in the present paper: M. J. Adams, Dave Behrens, R. F. Bolland, Philippe Bouchet, J. Brum, Gonçalo Calado, Yolanda 214

Camacho-García, Roger Clarke, P. Fiene, Gary McDonald, Scott Johnson, Alison Kay, Frank MacFarland, W. Farmer, S. Gigglinger, G. E. MacGintie, Serge Gofas, L. Harris, Ian Loch, Pierre Lozouet, Sandra Millen, K. A. Miller, Mike Miller, M. P. Morse, J. Ortea, Marion Patton, Gustav Paulay, Peter Pickens, Cory Pittman, Bertrand Richer de Forges, Bill Rudman, M. Schrödl, J. Sloan, Franz Steiner, J. Templado, C. Todd, Bob Van Syoc, F. W. Weymouth, and all the members of the different ‘Musorstom’ expeditions (co-organized by the IRD and the MNHN) as well as the Montrouzier expedition (IRD and MNHN, New Caledonia, 1993). We also are very grateful to the staff from the various museums for helping us with loans of material: H. Barlow, Ian Loch and Bill Rudman (Australian Museum, Sydney), Adam Baldinger (Museum of Comparative Zoology, Cambridge, MA), Oscar Soriano and Pepe Templado (Museo Nacional de Ciencias Naturales, Madrid), Virginie Heros and Philippe Bouchet (Muséum national d’histoire naturelle, Paris), Chris Rowley (National Museum of Victoria), Yolanda Villacampa, Tyjuana Nickens and Jerry Harasewych (Natural History Museum of the Smithsonian Institution, Washington DC), Kathe Jensen and Ole Tendal ( Zoological Museum of the University of Copenhagen). We also thank Liz Kools and Bob Van Syoc from the California Academy of Sciences. We are indebted to many systematists who have reacted — sometimes quite strongly — to propositions of new forms of species names. In particular, we wish to thank the members of the Systematics Discussion Group and the Nudibranch Lab of the California Academy of Sciences, the malacologists who attended the 69th meeting of the American Malacological Society, in Ann Arbor, Michigan in 2003, and all participants of the First International Meeting of Phylogenetic Nomenclature, in Paris, July 2004. We are especially grateful to Ken Angielczyk, Phil Cantino, Michael Manuel, Mónica Medina, Brent Mishler, Peter Roopnarine, Kevin de Queiroz, Chris Schander, and Amélie Scheltema for engaging discussions about species names. Two reviewers commented on the manuscript. This work was completed with support from the California Academy of Sciences, and the Partnership for Enhancing Expertise in Taxonomy program of the National Science Foundation (PEET DEB-9978155).

References Alder, J. & Hancock, A. (1846). Notices of some new and rare British species of naked Mollusca. Annals and Magazine of Natural History, 18, 289–294. Alder, J. & Hancock, A. (1864). Notice of a collection of nudibranchiate Mollusca made in India by Walter Elliot Esq., with descriptions of several new genera and species. Transactions of the Zoological Society of London, 5, 113 –147. Allan, J. (1947). Australian Shells. Melbourne: Georgian House. Anderberg, A. & Tehler, A. (1990). Consensus trees, a necessity in taxonomic practice. Cladistics, 6, 399– 402.

Zoologica Scripta, 34, 2, March 2005, pp199–224 • © The Norwegian Academy of Science and Letters 2005

B. Dayrat & T. Gosliner • Species names and metaphyly

Angas, G. (1864). Description d’espèces nouvelles appartenant à plusieurs genres de mollusques nudibranches des environs de PortJackson (Nouvelle-Galles du Sud), accompagnée de dessins faits d’après nature. Journal de Conchyliologie, 12, 43–70. Archibald, J. D. (1994). Metataxon concepts and assessing possible ancestry using phylogenetic systematics. Systematic Biology, 43, 27–40. Avila, C. (1995). Natural products of opisthobranch molluscs: a biological review. Oceanography & Marine Biology: an Annual Review, 33, 487–559. Behrens, D. W. (1981). A color variation in Chromodoris macfarlandi (Nudibranchia: Doridacea). Opisthobranch Newsletter, 13, 5. Behrens, D. W. (1991). Pacific Coast Nudibranchs: a Guide to the Opisthobranchs, Alaska to Baja California. Monterey, California: Sea Challengers. Behrens, D. W. & Henderson, R. (1981). Two new cryptobranch dorid nudibranchs from California. Veliger, 24, 120 –128. Behrens, D. W. & Henderson, R. (1982). Taringa aivica timia Marcus & Marcus, 1967 (Nudibranchia: Doridacea) in California. Veliger, 24, 197–199. Bergh, L. S. R. (1876). Malacologische Untersuchungen. In C. G. Semper (Ed.) Reisen im Archipel der Philippinen, Zweiter Theil, Wissenschaftliche Resultate, Band 2, Theil 2, Heft 10 (pp. 377– 427). Wiesbaden: Kreidel’s-Verlag. Bergh, L. S. R. (1877). Malacologische Untersuchungen. In C. G. Semper (Ed.) Reisen im Archipel der Philippinen, Zweiter Theil, Wissenschaftliche Resultate, Band 2, Theil 2, Heft 12, (pp. 495–546). Wiesbaden: Kreidel’s-Verlag. Bergh, L. S. R. (1878). Malacologische Untersuchungen. In C. G. Semper (Ed.) Reisen im Archipel der Philippinen, Zweiter Theil, Wissenschaftliche Resultate, Band 2, Theil 2, Heft 14 (pp. 603– 645, I–L). Wiesbaden: Kreidel’s-Verlag. Bergh, L. S. R. (1879). Neue Chromodoriden. Malakozoologische Blätter Neue Folge, 1, 87–116. Bergh, L. S. R. (1880). Über die Gättung. Peltodoris. Mittheilungen aus der Zoologischen Station zu Neapel Zugleich ein Repertorium für Mittelmeerkunde, 2, 222–232. Bergh, L. S. R. (1881). Malacologische Untersuchungen. In C. G. Semper (Ed.) Reisen im Archipel der Philippinen, Zweiter Theil, Wissenschaftliche Resultate, Band 2, Theil 4, Heft 2, (pp. 79–128). Wiesbaden: Kreidel’s-Verlag. Bergh, L. S. R. (1884a). Malacologische Untersuchungen. In C. G. Semper (Ed.) Reisen im Archipel der Philippinen, Zweiter Theil, Wissenschaftliche Resultate, Band 2, Theil 3, Heft 15, (pp. 647–754). Wiesbaden: Kreidel’s-Verlag. Bergh, L. S. R. (1884b). Report on the Nudibranchia. Challenger Reports, Zoology, 10, 1–151. Bergh, L. S. R. (1889). Malacologische Untersuchungen. In C. G. Semper (Ed.) Reisen im Archipel der Philippinen, Zweiter Theil, Wissenschaftliche Resultate, Band 2, Theil 3, Heft 16 (pp. 815– 872). Wiesbaden: Kreidel’s-Verlag. Bergh, L. S. R. (1898). Die Opisthobranchier der Sammlung Plate. Zoologische Jahrbücher supplement, 4, 481–582. Bergh, L. S. R. (1904). Malacologische Untersuchungen. In C. G. Semper (Ed.) Reisen im Archipel der Philippinen, Zweiter Theil, Wissenschaftliche Resultate, Band 9, Theil 1, Lief. 1 (pp. 1– 56). Wiesbaden: Kreidel’s-Verlag. Bergh, L. S. R. (1905). Die Opisthobranchiata der Siboga-Expedition. In M. Weber (Ed.) Uitkomsten op Zoologisch, Botanisch, Oceanographisch en Geologisch Gebied, Vol. 50 (pp. 1–248). Leiden: E. J. Brill.

Bergh, L. S. R. (1907). The Opisthobranchiata of South Africa. Transactions of the South African Philosophical Society, 17, 1–144. Bertsch, H. (1978). The Chromodoridinae nudibranchs from the Pacific coast of America. Part II. The genus Chromodoris. Veliger, 20, 307– 327. Bremer, K. (1994). Branch support and tree stability. Cladistics, 10, 295 – 304. Cantino, P. D., Bryant, H. N., de Queiroz, K., Donoghue, M. J., Eriksson, T., Hillis, D. M. & Lee, M. S. Y. (1999). Species names in phylogenetic nomenclature. Systematic Biology, 48, 790 – 807. Cantino, P. D. & de Queiroz, K. (2003). PhyloCode. http:// www.ohiou.edu /phylocode/. Cattaneo-Vietti, R., Chemello, R. & Giannuzzi-Savelli, R. Atlas of Mediterranean nudibranchs. Rome: La Conchiglia. Cervera, J. L., García Gomez, J. C. & García, F. J. (1986). Il genere Jorunna Bergh, 1876 (Mollusca: Gastropoda: Nudibranchia) nel litorale iberico. Lavori Societa Italiana di Malacologia, 22, 111–132. Cervera, J. L., García, J. C. & García, F. J. (1985). Redescription of Geitodoris planata (Alder & Hancock, 1846) (Gastropoda: Nudibranchia). Journal of Molluscan Studies, 51, 198 – 204. Chia, F. S. & Koss, R. (1978). Development and metamorphosis of planktotrophic larvae of Rostanga pulchra (Mollusca: Nudibranchia). Marine Biology, 46, 109–119. Chia, F. S. & Koss, R. (1982). Fine structure of the larval rhinophores of the nudibranch, Rostanga pulchra, with emphasis on the sensory receptor cells. Cell and Tissue Research, 225, 235–248. Chia, F. S. & Koss, R. (1983). Fine structure of the larval eyes of Rostanga pulchra (Mollusca, Opisthobranchia, Nudibranchia). Zoomorphology, 102, 1–10. Chia, F. S. & Koss, R. (1984). Fine structure of the cephalic sensory organs in the larva of the nudibranch Rostanga pulchra (Mollusca, Opisthobranchia, Nudibranchia). Zoomorphology, 104, 131–139. Chia, F. S., Koss, R. & Bickell, L. R. (1981). Fine structural study of the statocysts in Veliger larva of the nudibranch, Rostanga pulchra. Cell and Tissue Research, 214, 67–80. Cockerell, T. D. A. (1901). Three new nudibranchs from California. Journal of Malacology, 8, 85–87. Cockerell, T. D. A. (1902). Three new species of Chromodoris. Nautilus, 16, 19–21. Collier, C. L. & Farmer, W. M. (1964). Additions to the nudibranch fauna of the east Pacific and the Gulf of California. Transactions of the San Diego Society for Natural History, 13, 377–396. Cooper, J. G. (1863). Some new genera and species of California Mollusca. Proceedings of the California Academy of Natural Sciences, 2, 202–207. Crisp, M. D. & Chandler, G. T. (1996). Paraphyletic species. Telopea, 6, 813–846. Cuvier, G. (1804). Mémoire sur le genre Doris. Annales du Muséum National d’Histoire Naturelle, Paris, 4, 447– 473. Dayrat, B. (2004). Selecting a form of species names in the PhyloCode. In M. Laurin (Ed.) First International Phylogenetic Nomenclature Meeting, Abstracts 3. Paris: Muséum national d’histoire naturelle. Dayrat, B., Schander, C. & Angielczyk, K. D. (2004). Suggestions for a new species nomenclature. Taxon, 53, 485– 491. Dayrat, B. & Tillier, S. (2000). Taxon sampling, character sampling and systematics: how gradist presuppositions created additional ganglia in gastropod euthyneuran taxa. Zoological Journal of the Linnean Society, 129, 403– 418.

© The Norwegian Academy of Science and Letters 2005 • Zoologica Scripta, 34, 2, March 2005, pp199–224

215

Species names and metaphyly • B. Dayrat & T. Gosliner

Dayrat, B. & Tillier, S. (2002). Evolutionary relationships of euthyneuran gastropods (Mollusca: a cladistic re-evaluation of morphological characters. Zoological Journal of the Linnean Society, 135, 403–470. Dayrat, B. & Tillier, S. (2003). Limits and goals of phylogenetics: the euthyneuran gastropods. In C. Lydeard & D. Lindberg (Eds) Molecular Systematics and Phylogeograph of Mollusca (pp. 161–184). Washington, DC: Smithsonian Press. Donoghue, M. J. (1985). A critique of the biological species concept and recommandation for a phylogenetic alternative. Bryologist, 88, 172–181. Dorgan, K. M., Valdés, A. & Gosliner, T. M. (2002). Phylogenetic systematics of the genus Platydoris (Mollusca, Nudibranchia, Doridoidea) with descriptions of six new species. Zoologica Scripta, 31, 271–319. Edmunds, M. (1971). Opisthobranchiate Mollusca from Tanzania (suborder Doridacea). Zoological Journal of the Linnean Society, 59, 339–396. Edmunds, M. (1981). Opisthobranchiate Mollusca from Ghana: Chromodorididae. Zoological Journal of the Linnean Society, 72, 175–201. Eliot, C. N. E. (1905a). On the Doris planata of Alder & Hancock. Proceedings of the Malacological Society of London, 6, 180 –181. Eliot, C. N. E. (1905b). Note on Geitodoris planata (Alder & Hancock). Proceedings of the Malacological Society of London, 6, 186 – 187. Fahey, S. J. & Gosliner, T. M. (1999). Description of three new species of Halgerda from the Western Indian Ocean with a redescription of Halgerda formosa, Bergh 1880. Proceedings of the California Academy of Sciences, 51, 365–383. Fahey, S. J. & Gosliner, T. M. (2003). Mistaken identities: on the discodorididae genera Hoplodoris and Carminodoris Bergh (Opisthobranchia, Nudibranchia). Proceedings of the California Academy of Sciences, 54, 169 – 208. Foale, S. J. & Willan, R. C. (1987). Scanning and transmission electron microscope study of specialized mantle structures in the dorid nudibranchs (Gastropoda: Opisthobranchia: Anthobranchia). Marine Biology, 95, 547–558. Gauthier, J. (1986). Saurischian monophyly and the origin of birds. In K. Padian (Ed.) The Origin of Birds and the Evolution of Flight ( pp. 1 – 56). San Francisco: California Academy of Sciences. Gohar, H. A. F. & Soliman, G. N. (1967). The biology and development of Asteronotus cespitosus (van Hasselt) (Gastropoda, Nudibranchia). Publications of the Marine Biology Station, Al-Ghardaqa, Egypt, 14, 177–195. Gosliner, T. M. (1987). Nudibranchs of Southern Africa, a Guide to Opisthobranch Molluscs of Southern Africa. Monterey: Sea Challengers. Gosliner, T. M. & Behrens, D. W. (1998). Two new discodorid nudibranchs from the western Pacific with a redescription of Doris luteola Kelaart, 1858. Proceedings of the California Academy of Sciences, 50, 279–293. Gould, A. A. (1852). United States exploring expedition during the years 1838, 1839, 1840, 1841, 1842, under the command of Charles Wilkes, U.S. N, XII. Mollusca and shells. Haefelfinger, H. R. (1961). Beiträge zur Kenntniss von Peltodoris atromaculata Bergh 1880 (Mollusca, Opisthobranchia). Revue Suisse de Zoologie, 68, 331–343.

216

Hamann, J. C. (1992). A warm water Atlantic synonymy, Aphelodoris antillensis equals Chromodoris bistellata (Opisthobranchia, Gastropoda). Veliger, 35, 215–221. Hasselt, J. C. (1824). Extrait d’une lettre du Dr. J. C. van Hasselt au Prof. van Swinderen sur des mollusques de Java. Bulletin des Sciences Naturelle et de Géologie, 3, 237–245. Hennig, W. (1966). Phylogenetic Systematics. Urbana, Illinois: University of Illinois Press. Hickman, C. S. (1981). Evolution and function of asymmetry in the archaeogastropod radula. Veliger, 23, 189–194. Kay, E. A. & Young, D. K. (1969). The Doridacea (Opisthobranchia; Mollusca) of the Hawaiian Islands. Pacific Science, 23, 172–231. Lance, J. R. (1966). New distributional records of some northeastern Pacific opisthobranchiata (Mollusca: Gastropoda) with descriptions of two new species. Veliger, 9, 69– 81. Lanham, U. (1965). Uninominal nomenclature. Systematic Zoology, 14, 144. Laurin, M. & Cantino, P. D. (2004). First International Phylogenetic Nomenclature meeting: a report. Zoologica Scripta, 33, 475–479. Linnaeus, C. (1758). Systema Naturae (10th edn). Holmiae: L. Salvii. Linnaeus, C. (1767). Systema Naturae (12th edn). Holmiae: L. Salvii. MacFarland, F. M. (1905). A preliminary account of the Dorididae of Monterey Bay, California. Proceedings of the Biological Society of Washington, 18, 35–54. MacFarland, F. M. (1906). Opisthobranchiate Mollusca from Monterey Bay, California, and vicinity. Bulletin of the United States Bureau of Fisheries, 25, 109–151. MacFarland, F. M. (1966). Studies of opisthobranchiate mollusks of the Pacific coast of North America. Memoirs of the California Academy of Sciences, 6, 1–546. Maddison, W. P. & Maddison, D. R. (2000). MacClade, Version 4. [Computer Software and Manual]. Sunderland, MA: Sinauer Associates. Marcus, E. (1955). Opisthobranchia from Brazil. Boletim da Faculdade de Filosofia, Ciencias e Letras, Universidade de São Paulo, Zoology, 207, 89–261. Marcus, E. (1958). On western Atlantic opisthobranchiate gastropods. American Museum Novitates, 1906, 1–82. Marcus, E. (1959). Reports from the Lund University Chile Expedition 1948 –49. Lamellariacea und Opisthobranchia. Lunds Universitets Arsskrift, N. F. Adv., 2 (55), 1–133. Marcus, E. (1970). Opisthobranchs from northern Brazil. Bulletin of Marine Science, 20, 922–951. Marcus, E. (1976). Marine euthyneuran gastropods from Brazil (3). Studies on Neotropical Fauna and Environment, 11, 5–23. Marcus, E. & Hughes, H. P. I. (1974). Opisthobranch mollusks from Barbados. Bulletin of Marine Science, 24, 498–532. Marcus, E. & Marcus, E. (1959). Opisthobranchia aus dem Roten Meer und von den Malediven. Akademie der Wissenschaften und der Literatur. Abhandlungen der Mathematisch-Naturwissenschaftlichen Klasse, 12, 873–934. Marcus, E. & Marcus, E. (1963). Opisthobranchs from the Lesser Antilles. Studies on the Fauna of Curaçao and other Caribbean Islands, 19, 1–76. Marcus, E. & Marcus, E. (1967). American opisthobranch mollusks Part I, Tropical American opisthobranchs, Part II, Opisthobranchs from the Gulf of California. Studies in Tropical Oceanography, Miami, 6, 1–256. Marcus, E. & Marcus, E. (1970). Opisthobranchs from Curaçao and faunistically related regions. Studies on the Fauna of Curaçao and other Caribbean Islands, 33, 1–129.

Zoologica Scripta, 34, 2, March 2005, pp199–224 • © The Norwegian Academy of Science and Letters 2005

B. Dayrat & T. Gosliner • Species names and metaphyly

Marshall, J. G. & Willan, R. C. (1999). Nudibranchs of Heron Island, Great Barrier Reef. A Survey of the Opisthobranchia (Sea Slugs) of Heron and Wistari Reefs. Leiden: Backhuys Publishers. McDonald, G. R. (1983). A review of the nudibranchs of the California coast. Malacologia, 24, 114 – 276. Michener, C. D. (1964). The possible use of uninominal nomenclature to increase the stability of names in biology. Systematic Zoology, 13, 182–190. Millen, S. V. (1982). A new species of dorid nuribranch (Opisthobranchia: Mollusca) belonging to the genus Anisdoris. Canadian Journal of Zoology, 60, 2694 – 2705. Millen, S. V. & Bertsch, H. (2000). Three new species of dorid nudibranchs from southern California, USA, and the Baja California peninsula, Mexico. Veliger, 43, 354 –366. Mishler, B. D. & Brandon, R. N. (1987). Individuality, pluralism, and the phylogenetic concept. Biology and Philosophy, 2, 397– 414. Muniaín, C. (2001). Taxonomical and ecological aspects of the nudibranch Geitodoris patagonica Odhner, 1926 (Opisthobranchia, Doridina) from Argentina. Bollettino Malacologico, 37, 171–176. O’Donoghue, C. H. (1927). Notes on the nudibranchiate Mollusca from the Vancouver Island region. V. Two new species and one new record. Transactions of the Royal Canadian Institute, 16, 1– 12. Odhner, N. H. (1926). Die Opisthobranchien. Further Zoological Results of the Swedish Antarctic Expedition 1901–03 under the direction of Dr. Otto Nordenskjold, 2, 1–100. d’Orbigny, A. (1837). [1835–1846]. Voyage dans l’Amérique Meridionale (le Brésil, la republique orientale de l’Uruguay, la République argentine, la Patagonie, la république du Chili, la république de Bolivia, la république de Perou) executé pendant les années 1826, 1827, 1828, 1829, 1830, 1831, 1832, et 1833. Paris: Pitois-Levrault. Ortea, J. A. (1979). Deux nouveaux Doridiens (Mollusca, Nudibranchiata) de la côte nord d’Espagne. Bulletin du Museum National d’Histoire Naturelle, Série 4, Section a, Zoologie Biologie et Ecologie Animales, 1, 575–584. Ortea, J. A. (1990). El género Geitodoris Bergh, 1891 (Mollusca, Nudibranchia) en las islas Canarias. Revista de la Academia Canaria de Ciencias, 2, 99–120. Ortea, J. A. (1995). Estudio de las especies atlánticas de Paradoris Bergh, 1884 (Mollusca: Nudibranchia: Discodorididae) recolectadas en las Islas Canarias. Avicennia, 3, 5–27. Ortea, J. A. & Martinez, E. (1992). Descripción de una nueva especie del género Carminodoris Bergh, 1889 (Mollusca: Opisthobranchia: Nudibranchia) del piso batial del norte de España. Graellsia, 48, 185–188. Ortea, J. A. & Urgorri, V. (1979). Sobre la presencia de Dendrodoris racemosa Pruvot-Fol, 1951 y Discodoris rosi Ortea, 1977 (Gastropoda: Nudibranchia) en Galicia. Trabajos Compostelanos de Biología, 8, 71–78. Pease, W. H. (1860). Descriptions of new species of Mollusca from the Sandwich Islands. Proceedings of the Zoological Society of London, 28, 18–36. Pease, W. H. (1871). Descriptions of nudibranchiate Mollusca inhabiting Polynesia. American. Journal of Conchology, 6, 299– 305. Perrone, A. S. (1990). Una nuova specie de nudibranchi doridiani, Peltodoris sordii nov. sp., dalla biocenosi a Anadiomene stellata, Geodia cydonium e Holothuria impatiens (Opisthobranchia: Nudibranchia). Bollettino Malacologico, 25, 295–300.

Potts, G. W. (1981). Anatomy of respiratory structures in the dorid nudibranchs, Onchidoris bilamellata and Archidoris pseudoargus, with details of the epidermal glands. Journal of the Marine Biology Association of the United Kingdom, 61, 959–982. Pruvot-Fol, A. (1951). Etudes des nudibranches de la Méditerranée. Archives de Zoologie Expérimentale et Générale, 88, 1–80. Rapp, W. L. (1827). Über das Molluskengeschlecht Doris und Beschreibung einiger neuen Arten desselben. Nova Acta Academiae Caesareae Leopoldino-Carolinae Germanicae Naturae Curiosum Halia Saxonum, 13, 516 –522. Rose, R. A. (1986). Direct development in Rostanga arbutus (Angas) (Mollusca: Nudibranchia) and the effects of temperature and salinity on embryos reared in the laboratory. Journal of the Malacological Society of Australia, 7, 141–154. Rudman, W. B. (1984). The Chromodorididae (Opisthobranchia: Mollusca) of the Indo-West Pacific: a review of the genera. Zoological Journal of the Linnean Society, 81, 115–273. Rudman, W. B. & Avern, G. J. (1989). The genus Rostanga Bergh, 1879 (Nudibranchia: Dorididae) in the Indo-West Pacific. Zoological Journal of the Linnean Society, 96, 281–338. Schmekel, R. L. (1968). Ascoglossa, Notaspidea und Nudibranchia im Litoral des Golfes von Neapel. Revue Suisse de Zoologie, 75, 103– 155. Schmekel, R. L. & Portmann, A. (1982). Opisthobranchia des Mittelmeeres, Nudibranchia und Saccoglossa. Monografia della Stazione Zoologica di Napoli, 40, 1–410. Schrödl, M. (2000). Revision of dorid Nudibranchia collected during the French Cape Horn Expedition in 1882–1883, with discussion of the genus Geitodoris Bergh, 1891. Veliger, 43, 197–209. Schrödl, M. & Millen, S. V. (2001). Revision of the nudibranch gastropod genus Tyrinna Bergh, 1898 (Doridoidea: Chromodorididae). Journal of Natural History, 35, 1143–1171. Soliman, G. N. (1980). On the dorid nudibranch Sebadoris crosslandi (Eliot) from the northwestern Red sea. Journal of Molluscan Studies, 46, 227–238. Swofford, D. L. (2000). PAUP* — Phylogenetic Analysis Using Parsimony, Version 4.0. [Computer Software and Manual]. Washington DC: Smithsonian Institution. Templado, J. (1984). Moluscos de las praderas de Posidonia oceanica en las costas del Cabo de Palos (Murcia). Investigacion Pesquera, 48, 509– 526. Thompson, T. E. (1966). Development and life history of Archidoris pseudoargus. Malacologia, 5, 83–84. Thompson, T. E. (1975). Dorid nudibranchs from eastern Australia (Gastropoda, Opisthobranchia). Journal of Zoology, 176, 477–517. Thompson, T. E. (1976). British opisthobranch molluscs. Mollusca: Gastropoda, Vol. 1. London: Ray Society. Thompson, T. E. (1980). Jamaican opisthobranch molluscs II. Journal of Molluscan Studies, 46, 74 –99. Thompson, T. E. & Brown, G. H. (1984). Biology of Opisthobranch Mollusks, Vol. 2. London: Ray Society. Valdés, A. (2001). Deep-sea cryptobranch dorid nudibranchs (Mollusca, Opisthobranchia) from the Tropical West Pacific, with descriptions of two new genera and eighteen new species. Malacologia, 43, 237–313. Valdés, A. (2002). A phylogenetic analysis and systematic revision of the cryptobranch dorids (Mollusca, Nudibranchia, Anthobranchia). Zoological Journal of the Linnean Society, 136, 535– 636.

© The Norwegian Academy of Science and Letters 2005 • Zoologica Scripta, 34, 2, March 2005, pp199–224

217

Species names and metaphyly • B. Dayrat & T. Gosliner

Valdés, A. & Gosliner, T. M. (2001). Systematics and phylogeny of the caryophyllidia-bearing dorids (Mollusca, Nudibranchia), with descriptions of a new genus and four new species from IndoPacific deep waters. Zoological Journal of the Linnean Society, 133, 103–198.

Valdés, A. & Muniaín, C. (2001). Revision and taxonomic reassessment of magellanic species assigned to Anisodoris Bergh, 1898 (Nudibranchia: Doridoidea). Journal of Molluscan Studies, 68, 345–351. Wägele, H. (1990). Revision of the genus Austrodoris Odhner, 1926 (Gastropoda, Opisthobranchia). Journal of Molluscan Studies, 56, 163–180.

Appendix I Data matrix. 1

2

3

4

5

6

7

8

9

1 0

1 1

1 2

1 3

1 4

1 5

1 6

1 7

1 8

1 9

2 0

2 1

2 2

2 3

2 4

2 5

2 6

2 7

2 8

2 9

Outgroups antillensis evelinae kerguelenensis luteomarginata macfarlandi marcusi pseudoargus verrucosa

0 0 0 0 0 0 0 0

0 0/1 0 0 0 0 2 0

? 0/1 ? ? ? ? ? ?

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 1 0 1 1 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 1 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

1 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

Ingroup boholiensis concinna dubia evelinae lilacina liturata maculosa rosi rubra schmeltziana ketos atromaculata fellowsi lancei mauritiana lentiginosa nobilis nubilosa indecora araneosa mulciber caespitosus dalanghita bimaculata filix grandiflora estrelyado capensis heathi mavis planata sandiegensis punctuolata aivica fanabensis ellioti argo tomentosa pardus arbutus pulchra

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0

0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 2 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0

0 0 0 0 0 0 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1

0 0 0 0 0 0 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0

0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0

1 1 2 1 1 2 1 1 1 1 1 0 0 0 0 0 0 0 2 — 2 0 0 0 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 1 1

0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 — 0 — 0 0 0 0 0 0

0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 — — — — 0 0 0 0 0

0 0 1 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 1 1 0 0 0 0/1 0 0

0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 0 0 0 0 0 0

0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 1 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 1 1 0 0 1 1 1 1 0 0 0 0 1 1 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

0 0 0 0 1 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 2 1 0 0 0 0

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

CHARACTERS TERMINALS

218

Zoologica Scripta, 34, 2, March 2005, pp199–224 • © The Norwegian Academy of Science and Letters 2005

B. Dayrat & T. Gosliner • Species names and metaphyly

Appendix II Characters and character states. #1. Anterior margin of the foot (Fig. 1). The anterior margin of the foot is bilabiate in all the taxa taken into account. The upper lip can be notched at the median line. The notch is absent (0) or present (1). #2. Oral appendages (Fig. 1). The mouth area of Doridina bears very diverse appendages that are not easy to formalize in characters because of the plasticity of their position and shape. Here we recognize two main groups of appendages. The first appendages are expansions of mouth lips (Fig. 1); their wide basis is continuous with the tegument of the mouth lips; they are placed on each lateral side of the mouth and their orientation is more or less transversal; they can be grooved or not. The second appendages are oral tentacles; their basis is narrow and clearly distinct from the tegument of the mouth lips; they are placed between the foot and the inferior mouth lip; they can be digitiform or flattened, grooved or not. Oral tentacles are found in all Discodorididae. Valdés (2002) assumed a primary homology between the velar tentacles of some phanerobranchs such as Polycera and Crimora and the oral tentacles observed in Discodorididae. This primary homology is questioned here because velar tentacles actually originate from the margin of the dorsal notum, not from the mouth area. On the contrary, we consider that all oral appendages found in phanerobranchs are mouth lip expansions similar to those found in the outgroup taxa (Fig. 1). The oral appendages are mouth lip expansions (0) or independent oral tentacles (1). The mouth lips can also be naked (2). #3. Oral tentacle groove (Fig. 1). Oral tentacles can be grooved (0) or not (1). #4. Body shape. The body can be more or less elevated or depressed (0) or significantly flattened (1). #5. Tubercles (Fig. 4). The dorsal notum of Discodorididae is usually covered by tubercles. The general shape of tubercles of the common type (more or less conical, round or pointed) and the abundance of protruding spicules seem to vary intraspecifically, especially due to state of preservation. However, some tubercles have a very distinct shape that cannot be mistaken. They are treated here as distinct apomorphic states. The notum can be finely granular with low and well-spaced tubercles, as found, for example, in indecora Bergh, 1881. The notum can also be covered by large flat tubercles, such as in estrelyado Gosliner & Behrens, 1998. Tubercles are more or less conical, round or pointed (0), low and well-spaced (1), or large and flat (2). #6. Long conical papillae (Fig. 4). The notum can bear long conical papillae. A primary homology is assumed between long papillae found in bimaculata Lance, 1966 and aivica Marcus & Marcus, 1967. Long conical papillae are absent (0) or present (1). #7. Spicule arrangement (Fig. 4). Some tubercles called ‘caryophyllidia’ are highly specialized structures (e.g. Foale & Willan 1987; Valdés & Gosliner 2001). They are characterized by a crown of spicules that surrounds a ciliated apical knob (Fig. 4A,B). However, all caryophyllidia are not strictly identical (e.g. Marcus 1955; Foale & Willan 1987; Valdés & Gosliner 2001; Dorgan et al. 2002). They can be pedunculate or flat with their crown of spicules laying on the notum. The spicules can be free from the base of the peduncle up to the apex, or embedded within the peduncle for part of their length. Testing the phylogenetic significance of those variations is beyond our analysis. Spicules can protrude on the surface of the tubercles without any particular order, or be arranged in a regular crown of usually less than 10 spicules. The crown of spicules is absent (0) or present (1). #8. Cilia (Fig. 4). The notum of Discodorididae is covered with cilia which are grouped in small tufts (e.g. Potts 1981; Foale & Willan 1987; Valdés & Gosliner 2001). The abundance of these tufts varies from one species to another. It may also vary intraspecifically probably due to preservation. Tufts of cilia can be present on the entire epidermis, including the lamellae of the rhinophores. Apical ciliated areas are unambiguously found in all caryophyllidia. We observed similar ciliated areas at the apex of regular tubercles, but they are usually not present in all the tubercles of an individual specimen. Even though we consider that apical ciliated areas are only present in caryophyllidia, futher studies are needed for this feature because of its critical phylogenetic importance. Apical ciliated knobs are absent (0) or present (1). #9. Gills sheath. It is most of the time impossible to decide whether or not the gills opening is surrounded by an elevated sheath, especially in preserved specimen. In some species, however, the sheath is formed by large triangular expansions of the notum separating and covering the gills. This arrangement results in what is usually called a star-shaped gill opening, such as in Platydoris argo (Linnaeus, 1767), and Asteronotus caespitosus (van Hasselt, 1824). The star-shaped gill opening is absent (0) or present (1). #10. Mantle holes (Fig. 4). Two distinct kinds of holes were found in the external tegument. First, small holes which never exceed 20 µm in diameter, usually below 10 µm. Potts (1981) showed that these are openings of subepithelial glandular cells characterized by a highly granular cytoplasm. Their abundance and distribution on the tegument varies considerably among the specimens we observed, also among specimens of the same species. Sometimes they are particularly numerous and are found on the entire surface of the body, including the rhinophores (Fig. 4C); they can also be very rare, as in P. ellioti and P. argo. They were found in all the specimens observed. Their absence, which is never complete, might be due to preservation. Holes of the second type are much wider (Fig. 4D−F). Their diameter always exceeds 20 µm and is usually around 50 µm. Unlike the small holes, which can only be seen in electron microscopy, the wide holes can be seen with an ocular scope. When present, they are present only on the dorsal surface of the notum. They are usually ringed by a thick margin, although the presence of this feature can vary among holes of the same individual. Several authors suggested that those holes are openings of mantle secretory glands (e.g. Odhner 1926; Marcus 1976; Ortea 1990, 1995; Schrödl 2000). We observed hard little whitish balls obstructing those holes in several preserved specimens, which supports the existence of a secretory activity. Further observations on living specimens and histological studies are needed. Note that no wide holes were found in chromodorid species with mantle glands (e.g. Rudman 1984), which indicates that they correspond to one particular kind of secretory gland analogous to chromodorid glands. Wide holes in the dorsal notum are absent (0) or present (1). #11. Labial cuticle. The labial cuticle is smooth or covered by jaws. The latter, when present, are constituted by: (1) a pair of sickle-shaped or rectangular symmetrical plates, or (2) one pair of symmetrical plates and a third additional unpaired plate. The paired jaws can be either rectangular, wider than long, or sickle-shaped and facing each other. The limit between pointed and rounded rodlets is actually difficult to appreciate because jaws, which are used to crush food, are often destroyed. However, some rodlets are clearly pointed and this character may provide some phylogenetic information after careful re-investigation. The cuticle is smooth (0), covered by one pair of symmetrical jaws (1), covered by one pair of symmetrical jaws and an additional unpaired jaw (2). #12. Shape of the radula. Most radulae are more or less square or rectangular. Their length never equals more than three times their width, at least in Discodorididae, which implies that the number of rows is either about the same as the number of teeth per row or inferior. Some radulae are significantly elongate. Their length equals at least three time their width, which implies that the number of rows equals at least twice the number of teeth. Radulae are rectangular or square (0) or significantly elongate (1). #13. Orientation of rows. In most radulae, the rows of teeth are more or less perpendicular to the rachidian axis. In some species, the rows are at an angle that significantly exceeds 45°, which might be related to the narrowness of the radula. This character is unknown for several terminals. The rows are more or less perpendicular to the rachidian axis (0) or at an angle exceeding 45° (1). #14. Tooth groove and radular asymmetry (Figs 5, 6A). The hook of the hamate teeth can be smooth or grooved on its external edge. When a groove is present, all the teeth are grooved. The upper lip of the groove can be significantly expanded. This expansion, however, was only observed on the left side of the radula of D. liturata and results in an unusual radular asymmetry (Fig. 5A, B). Note that this asymmetry is not due to an artifact of observation of the radula at different angles and that it was observed in all specimens dissected. Hickman (1981) showed that asymmetrical rhipidoglossate radulae independently appeared in several lineages of basal prosobranch gastropods. The asymmetry she described is due to a skewing of rows. The asymmetry observed here is due to different shapes of teeth on the left and on the right sides, which is unique. In fact, that constitutes, as far as we know,

© The Norwegian Academy of Science and Letters 2005 • Zoologica Scripta, 34, 2, March 2005, pp199–224

219

Species names and metaphyly • B. Dayrat & T. Gosliner

Appendix II Continued one of the rare cases of physical radular asymmetry for all gastropods. Interestingly, asymmetry always seems to be accompanied by a reduction of the number of teeth per row. Finally, note that some ‘grooved’ lateral teeth are also present in some other non-dorid opisthobranchs, such as Berthella and Tritonia. However, these grooves are anatomically different from those described here. Moreover, subsequent phylogenetic analyses clearly demonstrate they are independently derived in dorids. The external lateral tooth groove is absent (0), present (1), present with an expansion of the dorsal lip (2). #15. Spatulate outermost lateral teeth (Fig. 6C). Outermost lateral teeth can be all independent and simply hamate. They can also be more or less fused and spatulate. Outermost lateral spatulate teeth are absent (0) or present (1). #16. Pectinate outermost lateral teeth (Fig. 6B). Three to four outermost teeth can be simply hamate or pectinate. By ‘pectinate teeth’, we designate all teeth having a comb-shaped tip and assume their primary homology. Outermost lateral pectinate teeth are absent (0) or present (1). #17. Slender lateral teeth (Fig. 6D). Lateral teeth can be simply hamate. They can also be slender and elongate, in which case the radula (or part of it) looks like a hair. Either all the lateral teeth or only some outermost teeth are slender and elongate. Slender and elongate lateral teeth are absent (0) or present (1). #18. Lateral denticles. Lateral teeth can be smooth or can bear flattened denticles on the external surface of the hook. The number of lateral teeth bearing those denticles and the number of denticles per tooth vary considerably. However, since all those denticles have a similar shape and are all placed on the external surface of the hook, there is no reason to not assume their primary homology. Lateral flattened denticles are absent (0) or present (1). #19. Intestinal loop (Fig. 7A). The intestine can be dorsal and straight from the stomach to the anus. It can also form a long loop on the lateral right side and the ventral surface of the digestive gland. The intestinal loop is absent (0) or present (1). #20. Muscular wall around the atrium (Fig. 7B). The atrium can be free or protected by a muscular wall. This structure was first described by Marcus & Marcus (1967, 1970). The atrium can be entirely embedded by the muscular wall or only be protected on one side. The shape and the position of the muscular wall vary among individuals of the same species. The muscular wall is absent (0) or present (1). #21. Distal accessory glands. By ‘distal accessory glands’ we designate glands present in the distal portion of the reproductive system and opening into the vestibule or the distal portion of the vaginal duct. In practice, distinguishing the distal part of the vaginal duct from the vestibule is somewhat difficult, especially when specimens are small. Therefore we prefer using the term ‘distal accessory gland’ to ‘vestibular gland.’ Valdés (2002) distinguishes vestibular glands opening into the distal end of the female gland mass from accessory glands opening into the atrium, and he considers both types as a priori analogous. Vestibular glands have been called ‘vestibular’ precisely because they open into the vestibule, which is the same anatomical part as the atrium (see Rudman 1984). Therefore, a primary homology is assumed here between all distal accessory glands. Accessory glands which host spines and stylets are treated as separate characters because they are anatomically different and can be found in addition to unarmed accessory glands. Accessory glands are absent (0) or present (1). #22. Accessory gland with a hollow spine. Accessory glands can be empty or can contain a hollow spine. The latter is straight or curved, with a narrow or wide basis. Hollow spines end in a hole and most probably help convey the glandular secretion. #23. Stylet sac of the type ‘tomentosa.’ The stylet sac found in J. tomentosa is easily recognizable by its two portions: a proximal, tubular and glandular portion, and a distal portion containing the stylet. The latter is needle-like, solid, straight or curved. The stylet-sac of the type ‘tomentosa’ is absent (0) or present (1). #24. Stylet sac of the type ‘indecora.’ The stylet sac found in P. indecora is easily recognizable by its unique distal portion hosting a needle-like, solid and straight stylet. There is always only one stylet per sac. Stylet sacs of the type ‘indecora’ are absent (0) or present (1). When present, their number can vary (see indecora in Appendix 3). #25. Prostate. The prostate can be tubular and homogeneous (0), globular and homogeneous (1) or flattened and divided in two parts (2). #26. Wide folded proximal part of the deferent duct (Fig. 7A). The proximal part of the deferent duct can be much wider than the distal part and folded internally. This character was first described by Bergh (1884a) in Discodoris maculosa. Since then, this interesting character has largely been overlooked, with the exception of Schmekel & Portmann (1982) who drew it for the reproductive system of D. maculosa. The proximal part of the deferent duct can be similar to the distal part (0) or much wider and folded internally (1). #27. Penis. Penial papillae and penial hooks are not considered here as characters because their extremely plastic variation among dorids, which would imply a sampling of all known species if one wanted to be rigorous. However, we recognize here the presence of a penial cuticle as found in Taringa aivica Marcus & Marcus, 1967 as a distinct character. A hard cuticle is absent (0) or present (1). #28. Vaginal surface. The surface of the distal vaginal duct is usually smooth (0). It can also bear hooks (1) or a be covered by a cuticle (2). #29. Blood glands. The blood glands are located in the anterior part of the body, on the cephalic ganglia. They receive blood from the heart conveyed by the main dorsal blood vessel. There is a single large blood gland (0), covering entirely the cerebral ganglia, or two separate blood glands (1) anterior and posterior to the cerebral ganglia.

Appendix III List of species sampled in the phylogenetic analysis. Species names follow Lanham’s method, slightly modified (Dayrat et al. 2004). Page number is only added after the date of publication when an author has given the same species epithet to two different species in the same year. Material examined, distribution, remarks concerning character coding, diagnostic characters are provided. The literature that helped build the data matrix is indicated between parentheses. Outgroups antillensis Bergh, 1879. Originally described within the genus Aphelodoris. Material examined: two syntypes, Caribbean Sea, Saint Thomas 1859, two specimens 14 and 17 mm preserved length, leg. Riise, ZMUC GAS 2014 and GAS-2016; one syntype, Caribbean Sea, Virgin Islands, 23 June 1858, one specimen 13 mm preserved length, leg. Riise, ZMUC GAS 2015; Caribbean Sea, Cayman Island, Grand Cayman Island, South of West Bay, near Soto’s Reef, 8 May 1991, one specimen 18 mm preserved length, leg. Jeff Hamann, CASIZ 077289, Voucher (Hamann 1992); Caribbean Sea, Quintana Roo, Cozumel, 31 December 1977, two specimens 12 and 13 mm preserved length, leg. T. M. Gosliner, CASIZ 075104. Remarks: (#26) the prostate, not divided in two parts, is more or less spherical and covered by large nodules. Distribution: Caribbean Sea. Autapomorphy: Despite its wide variation, the colour pattern of the dorsal notum seems to be characteristic, with brown, yellow and white dots that can form blotches (Hamann 1992) (Bergh 1879; Marcus & Marcus 1963; Thompson 1980; Hamann 1992; Valdés 2002).

220

Zoologica Scripta, 34, 2, March 2005, pp199–224 • © The Norwegian Academy of Science and Letters 2005

B. Dayrat & T. Gosliner • Species names and metaphyly

Appendix III Continued evelinae Marcus, 1958: 18. Originally described within the genus Cadlina and transferred to Tyrinna by Rudman (1984). The page number ‘18’ is part of the species name because Marcus named another species with the same specific epithet in that year: evelinae Marcus, 1958: 53, originally placed in the genus Corambe. Material examined: Mexico, Nayarít, Matanchen, near San Blas, January 1975, six specimens 8–13 mm preserved length, leg. Gary McDonald, CASIZ 069891. Remark: (#2) it is difficult to determine whether the oral appendages are mouth lip expansions or oral tentacles (Fig. 1H). Distribution: Western Atlantic and eastern Pacific, from Mexico to Peru. Autapomorphy: whitish notum, with golden or orange-red spots on the entire notum (Marcus 1958; Collier & Farmer 1964; Thompson 1980; Edmunds 1981; Rudman 1984; Schrödl & Millen 2001). kerguelenensis Bergh, 1884. Originally described within the genus Archidoris, also placed within the genus Austrodoris by Odhner (1926). Material examined: Antarctica, West side of McMurdo Sound, New Harbor, North-west of Explorer’s Cove, 20 m depth, 17 December 1985, one specimen 89 mm preserved length, leg. K. A. Miller, CASIZ 087312. Distribution: Antarctica. Remark: a penis is usually considered absent, but the evaginable muscular portion of the deferent duct may be considered as a penial papilla. (Bergh 1884b; Odhner 1926; Wägele 1990; Valdés 2002). luteomarginata MacFarland, 1966. Originally described within the genus Cadlina. Material examined: California, Channel Islands, Santa Rosa Island, south side, Bee Rock, 10–15 m depth, 23 October 1986, one specimen 29 mm preserved length, leg. Robert Van Syoc, CASIZ 071375. Distribution: North-eastern Pacific from Alaska to Mexico. Autapomorphy: narrow marginal band of yellow along the edge of the notum (MacFarland 1966; Rudman 1984). macfarlandi Cockerell, 1901. Originally described within the genus Chromodoris, also placed in Glossodoris by O’Donoghue (1927) and MacFarland (1966). Material examined: California, Orange County, Corona del Mar, 25 June 1948, one specimen 18 mm preserved length, leg. G. E. MacGinitie, CASIZ 067089, Voucher (Bertsch 1978). Distribution: Eastern Pacific, from Monterey, California to Baja California. Autapomorphy: brilliant reddish-violet colour of the notum, with three longitudinal yellow lines (one median, two starting from the rhinophores) (Cockerell 1901, 1902; MacFarland 1966; Marcus & Marcus 1967; Behrens 1981). marcusi Collier & Farmer, 1964. Originally described within the genus Conualevia. Material examined: paratype, Mexico, Gulf of California, Baja California, Isla Angel de la Guarda, Puerto Refugio, 29°33′ N, 113°35′ W, 25 March 1963, one specimen 18 mm preserved length, leg. J. Sloan, CASIZ 018372. Remark: according to Valdés (2002), the reproductive system is strictly serial, whereas Collier & Farmer (1964) described a semiserial reproductive system characterized by an ‘arrangement in X’, with the duct of the receptaculum seminis and the duct of the bursa copulatrix converging at the same point; the observation of the paratype dissected by Valdés (2002) confirmed the description provided by Collier & Farmer (1964). Distribution: Baja California. Autapomorphy: smooth rhinophores are shared by two species described by Collier & Farmer (1964) in the genus Conualevia; the differences between those two species might need to be reevaluated (Collier & Farmer 1964; Valdés 2002). pseudoargus Rapp, 1827. Originally described within the genus Doris, also placed within the genus Archidoris (e.g. Thompson 1976; Schmekel & Portmann 1982). Material examined: British Isles, England, North Yorkshire County, Robin Hood’s Bay, April 1976, one specimen 38 mm preserved length, leg. L. Harris and C. Todd, CASIZ 076061, Voucher (Valdés 2002); Italy, Naples, 1902– 03, one specimen 46 mm preserved length, leg. Frank M. MacFarland, CASIZ 081871. Remarks: (#17) outermost tooth can be denticulate on its edge, which is not seen as a pectination. Distribution: Eastern Atlantic, from Norway to Portugal, Mediterranean Sea. Autapomorphy: the colour pattern of the notum is highly variable, but the gigantic penis is a very distinct organ (Thompson 1966; Thompson 1976; Thompson & Brown 1984; Potts 1981; Schmekel & Portmann 1982; Valdés 2002). verrucosa Linnaeus, 1758. Originally described within the genus Doris. Material examined: Italy, Tyrrhenian Sea, Naples, 23 March 1903, one specimen 22 mm preserved length, leg. Frank M. MacFarland, CASIZ 067689. Distribution: Mediterranean Sea, Western and Eastern Atlantic. Autapomorphy: grey or yellowish dorsal notum covered by very large rounded tubercles; the largest ones are narrow at the basis (Schmekel 1968; Thompson & Brown 1984; Schmekel & Portmann 1982; Valdés 2002). Ingroup terminals aivica Marcus & Marcus, 1967. Originally described in Taringa. Material examined: Seven syntypes (aivica timia Marcus & Marcus 1967), Mexico, Pacific coasts, Sonora, Puerto Penasco, Norse Beach, seven specimens from 15 to 27 mm preserved length, leg. P. Pickens, August 1964 and January 1965, NMNH 753569, NMNH 753565, NMNH 753648. An additional specimen was dissected from the Galapagos islands. It obviously belongs to the clade Taringa, most likely to a new taxon given the morphology of its penial cuticle (Dayrat, in prep.: Pacific Ocean, Galapagos Islands, Isla Santa Cruz, Las Bachas, 9 September 1991, one specimen 19 mm preserved length, leg. T. M. Gosliner, CASIZ 078406). Distribution: Panama (Pacific side), Mexico (Pacific side), California. Autapomophy: shape of the penis cuticle (Marcus & Marcus 1967; Behrens & Henderson 1982). araneosa Valdés, 2001. Originally described in Paradoris. Material examined: holotype, Pacific Ocean, South of New Caledonia, 452–454 m depth, 23°03′ S, 166°59′ E, one specimen 14 mm preserved length, 19 March 1994, leg. P. Bouchet and B. Richer de Forges, MNHN; paratype, Pacific Ocean, South of New Caledonia, 314–364 m depth, 21°43′ N, 166°37′ E, one specimen 13 mm preserved length, 19 March 1994, leg. B. Richer de Forges, Expedition Halipro 1, CASIZ 121099. Distribution: off New Caledonia (from 314 to 464 m depth). Autapomophy: according to Valdés (2001), one vestibular gland and two stylet sacs, although this character is very suspicious because it shows a great (and overlooked) variation. arbutus Angas, 1864. Originally described within Doris and transferred to Rostanga by Allan (1947). Material examined: Australia, New South Wales, North side of Hasting Point, 28 September 1986, three specimens from 6 to 9 mm preserved length, leg. T. M. Gosliner, CASIZ 071681. Distribution: Eastern Australia. Autapomophies: marginal teeth multifid and first lateral with 13–14 denticles (Thompson 1975; Foale & Willan 1987; Rose 1986; Rudman & Avern 1989). argo Linnaeus, 1767. Originally described within Doris, and transferred to Platydoris by Bergh (1877). Material examined: Spain, Murcia Province, Cabo de Palos, August 1980, two specimens 68 and 80 mm preserved length, leg. J. Templado, CASIZ 115217, Voucher (Valdés & Gosliner 2001). Distribution: Mediterranean Sea, Eastern Atlantic Ocean (from Portugal to Cape Verde Islands). Autapomophy: notum colour dark red with some diffuse brown markings (Schmekel & Portmann 1982; Thompson & Brown 1984; Valdés & Gosliner 2001; Dorgan et al. 2002). atromaculata Bergh, 1880. Originally described within the genus Peltodoris. Material examined: holotype, Mediterranean Sea, Italy, Naples (Zoological Station), May 1880, one specimen 34 mm preserved length, leg. Dr Dohrn ZMUC GAS-2054. Atlantic Ocean, Azores, Ilha Saõ Miguel, East of Caloura, 20 July 1988, one specimen 60 mm preserved length, leg. T. Gosliner and J. Brum, CASIZ 072584. Distribution: Eastern Atlantic Ocean, Mediterranean Sea. Autapomorphy: white ground colour of the notum with chocolate-brown patches (Bergh 1880; Haefelfinger 1961; Thompson 1976; Schmekel & Portmann 1982; Valdés 2002). bimaculata Lance, 1966. Originally described in Thordisa. Material examined: holotype, California, San Diego County, La Jolla, Windnsea Reef, 32°52′ N, 117°15′ W, 31 May 1965, one specimen 18 mm preserved length, leg. J. R. Lance, CASIZ 018669; paratype, California, San Diego County, La Jolla, Windnsea Reef, 32°52′ N, 117°15′ W, 31 May 1965, one specimen 19 mm preserved length, leg. J. R. Lance, CASIZ 018670; California, San Diego County, Escondido Canyon, 12 m depth, 3 April 1974, one specimen 28 mm preserved length, leg. Marion Patton, CASIZ 070575. Distribution: California and Mexico (Baja California). Autapomorphy: ground colour whitish to yellow-orange, with two large blotches of browns spots on the midline of the notum between the rhinophores and the gills (Lance 1966). boholiensis Bergh, 1877. Originally described within the genus Discodoris. Material examined: three syntypes, Philippines, Aibukit, 1861, two specimens 70 and 42 mm preserved length, leg. Semper & Philippines, Bohol, 1863, one specimen, 42 mm preserved length, leg. Semper, ZMUC GAS-2122; Indonesia, Sulawesi, Celebes Sea, Manado, Bunaken Island, 9 m depth, 16 May 1990, one specimen 14 mm preserved length, leg. P. Fiene, CASIZ 087129; Philippines, Luzon, Batangas Region, Maricaban Island, Devil’s Point, 28 m depth, 17 March 1994, one specimen 16 mm preserved length, leg. T. M. Gosliner, CASIZ 097404; Philippines, Luzon, Batangas Region, Maricaban Island, Bethlehem, 8 May 2001,

© The Norwegian Academy of Science and Letters 2005 • Zoologica Scripta, 34, 2, March 2005, pp199–224

221

Species names and metaphyly • B. Dayrat & T. Gosliner

Appendix III Continued three specimens 46 – 57 mm preserved length, leg. M. J. Adams, CASIZ 158768; Madagascar, Ile Sainte Marie, La Crique Hotel, 8 April 1990, one specimen 8 mm preserved length, leg. T. M. Gosliner, CASIZ 073543. Distribution: Indo-Pacific. Autapomorphy: the colour pattern of the notum (black and white blotches and lines) associated with the very fragile undulating margins; trifurcated denticle on the internal edge of the innermost tooth (Bergh 1877; Thompson 1975; Valdés 2002).

caespitosus van Hasselt, 1824. Originally described in Doris, transferred to the genus Asteronotus by Bergh (1905). Material examined: Madagascar, Mora Mora, two specimens 55 and 65 mm preserved length, 5 April 1989, leg. T. Gosliner, CASIZ 070364. Distribution: Indo-Pacific. Autapomophy: leathery notum and u-shaped receptaculum seminis (Bergh 1905; Edmunds 1971; Thompson 1975; Gohar & Soliman 1967; Kay & Young 1969; Valdés & Gosliner 2001). capensis Bergh, 1907. Originally described within the genus Geitodoris. Material examined: South Africa, Cape Peninsula (Atlantic coast), Kommetjie, 1984, one specimen 42 mm preserved length, leg. T. M. Gosliner, SAM A35368, voucher Gosliner (1987). Species only known from two specimens. Distribution: South Africa (Cape Peninsula). Autapomorphy: orange ground colour of the notum with brown spots (Bergh 1907). concinna Alder & Hancock, 1864. Originally described in Doris and transferred to Discodoris by Bergh (1877). Material examined: Indian Ocean, Seychelles, Aldabra Atoll, Lagon between Passe Femme and Passe du Bois, 19 March 1986, one specimen 45 mm preserved length, leg. T. Gosliner, CASIZ 074193; Djibouti, 1904, two specimens 65 and 75 mm preserved length, leg. Gravier, MNHN; Australia, Queensland, Magnetic Island, Cockle Bay, 19°10.6′ S, 146°49.4′ E, five specimens from 45 to 85 mm preserved length, leg. I. Loch, AM C415872; New Caledonia, Touho, 20°46.5′ S, 165°14′ E, station 1252, 1–4 m depth, September 1993, eight specimens from 30 to 80 mm preserved length, leg. Montrouzier expedition, MNHN; Japan, Ryukyu Islands, Okinawa, 2.5 km SW of Onna Village, S end of Onna Flats, 26°28.4′ N, 127°50.8′ E, 30 April 1987, one specimen 75 mm preserved length, leg. R.F. Bolland, CASIZ 070011; Pacific Ocean, Marshall Islands, Enewetak Island, lagoon side of island, 4 m depth, 14 September 1983, 4 specimens from 45 to 55 mm preserved length, leg. Scott Johnson, CASIZ 121232. Distribution: Indo-Pacific. Autapomorphies: thick muscular wall beneath the notum embedding the entire body wall, peculiar loop of the fertilization duct found in all specimens. This species can be easily confused with nubilosa Pease, 1871. However, concinna and nubilosa have different tubercles on the notum and also a different penis: a small conical papilla in concinna and a long coiled and armed penis in nubilosa. dahlanghita Fahey & Gosliner, 1999. Originally described in Halgerda. Material examined: paratypes, Philippines, Luzon, Batangas Province, Maricaban Island, Bethlehem, 15 m depth, two specimens 28 and 29 mm preserved length, 24 April 1997, leg. T. M. Gosliner, CASIZ 110373. Distribution: from South Africa to Papua New Guinea. Autapomophy: orange ground colour with many small dark brown dots (Fahey & Gosliner (1999). dubia Bergh, 1904. Originally described in Discodoris. Material examined: holotype, Australia, NW coast of Tasmania, one specimen 15 mm preserved length, leg. Miss Bodder, ZMUC GAS-2063; Australia, Victoria, off Mordialloc, 14 December 1958, one specimen 22 mm preserved length, NMV F20111; Australia, South Australia, Eyre Peninsula, Louth Bay, Point Warna, 34°32′ S, 135°56′ E, 11 February 1985, 12 specimens from 9 to 30 mm preserved length, leg. W. B. Rudman and I. Loch, AM C145110. Distribution: south-eastern Australia, Tasmania. Autapomorphy: the dorsal colour pattern (greyish-creamish with black dots) is also found in other Paradoris species, such as mulciber. However, the independent wide white hermaphroditic gland is an autapomorphy of this species (Bergh 1904). ellioti Alder & Hancock, 1864. Originally described in Doris, and transferred to Platydoris by Bergh (1878). Material examined: syntypes (designated here), Madras, three specimens 23, 37 and 39 mm preserved length, leg. Sir Walter Eliot, Hancock Museum (01/09/11, no. 13); Papua New Guinea, North-west of Madang, Laing Island, Hansa Bay, June 1992, one specimen 68 mm preserved length, leg. T. M. Gosliner, CASIZ 086531, Voucher (Dorgan et al. 2002). Distribution: India, New Caledonia and Papua New Guinea. Autapomophy: dorsal colouration and mottling, and ventral pigmentation (Dorgan et al. 2002). estrelyado Gosliner & Behrens, 1998. Originally described within Hoplodoris. Material examined: paratype, Philippines, Luzon, Batangas Province, Maricaban Island, Devil’s Point, 13 m depth, one specimen 40 mm preserved length, 23 May 1993, leg. M. Miller, CASIZ 088113. Distribution: Western Australia, Vietnam, Indonesia, Philippine Islands, Marshall Islands. Remark: the stylet of the distal accessory gland is hollow (#22), contrary to what the original description suggested (Gosliner & Behrens 1998). Autapomophy: the ‘friedeggs’ colour pattern of the notum (Gosliner & Behrens 1998; Fahey & Gosliner 2003). evelinae Marcus, 1955. Originally described within Discodoris. Material examined: Panama, Caribbean Sea, Canal Zone, Galeta Island, 19 August 1974, one specimen 30 mm preserved length, leg. Hans Bertsch, CASIZ 010580; Bahamas, Grand Bahama Island, 40 km East of Freeport, Gold Rock Creek, April 1984, two specimens 35 and 40 mm preserved length, leg. J. N. Worsfold, CASIZ 072292; Brazil, Bahia, Itapoa Farrol, 30 July 1975, two specimens 30 and 35 mm preserved length, leg. M. P. Morse, MCZ 288294. Distribution: Caribbean (Brazil to Panama). Autapomorphy: externally, evelinae can be easily confused with ketos, which has the same distribution, but evelinae has a penis covered with spines that is diagnostic (Marcus 1955; Marcus & Marcus 1967; Marcus & Hughes 1974). fanabensis Ortea & Martinez, 1992. Originally described in Taringa. Distribution: Canary Islands (Tenerife). Autapomophy: colour of the notum dark violet with opaque white areas (Ortea & Martinez 1992). fellowsi Kay & Young, 1969. Originally described within the genus Peltodoris. Material examined: paratype, Hawaii, Oahu, Pupuekea, 10 m depth, July 1966, one specimen 36 mm preserved length, leg. Alison Kay, BPBM 8922; Hawaii, Maui, Black Rock, 4 m depth, 6 May 2003, one specimen 30 mm preserved length, leg. Cory Pittman, CASIZ 166765. Distribution: Hawaiian Islands, New Caledonia, and Henderson Island. Autapomorphy: colour pattern of the notum homogeneously white with black gills and rhinophores (Kay & Young 1969). filix Pruvot-Fol, 1951. Originally described in Thordisa. Material examined: Italy, one specimen 15 mm preserved length, 1902–03, leg. F. M. MacFarland, CASIZ 081885. Species known from many specimens. Distribution: Mediterranean Sea. Autapomorphy: vestibular gland extremely long and coiled in concentric circles (Pruvot-Fol 1951; Schmekel & Portmann 1982). grandiflora Pease, 1860. Originally described within Doris, transferred to Carminodoris by Kay & Young (1969). Material examined: Reunion Island, Saint-Gilles Les Bains, one specimen 44 mm preserved length, 20 April 1989, leg. T. M. Gosliner, CASIZ 070388, Voucher (Fahey & Gosliner 2003). Distribution: Indo-Pacific. Remark: the spine of the vestibular gland is hollow (Fahey & Gosliner 2003). Autapomophy: colour pattern of the notum (Fahey & Gosliner 2003). heathi MacFarland, 1905. Originally described within Discodoris and transferred to the genus Geitodoris by Behrens (1991). Material examined: holotype, California, Monterey Bay, Pacific Groove, June 1897, one specimen 30 mm preserved length, leg. F. M. MacFarland, NHMSI 181282; Alaska, Alaska Peninsula, Shelikof Strait, Kukak Bay, 21 July 1927, two specimens 20 mm and 30 mm preserved length, leg. F. W. Weymouth, CASIZ 031836; California, Monterey County, Asilomar, California, 1 December 1971, one specimen 24 mm preserved length, leg. G. McDonald, CASIZ 069140; Channel Islands, San Miguel Island, Tyler Bight, California, 10 m depth, 15 July 1987, one specimen 22 mm preserved length, leg. R. Van Syoc, CASIZ 082091; Ensenada, Baja California, Mexico, 20 April 1963, one specimen 26 mm preserved length, leg. W. Farmer, CASIZ 113415. Distribution: from Alaska to Baja California; the Alaska Peninsula (this study) constitutes the northernmost locality. Autapomorphy: Colour pattern of the dorsal notum, with a sprinkling of minute black specks and white branchial gills (MacFarland 1905, 1906, 1966). indecora Bergh, 1881. Originally described in Discodoris, transferred to Paradoris by Templado (1984). Material examined: holotype, Mediterranean Sea, Italy, Triest, March 1880, one specimen 18 mm preserved length, leg. Dr Graeffe, ZMUC GAS-2017; Mediterranean Sea, Istria, Premantura, 2 August 1968, nine specimens 7–11 mm preserved length, leg. H. Lemche, ZMUC; Portugal, Sagres, 16 m depth, two specimens 35 mm preserved length (previously dissected by Ortea), 13 May 1988, leg. J. Ortea, MNHN; Distribution: Mediterranean Sea, Canary Islands, Eastern Atlantic (Portugal). Contrary to what Valdés (2002) stated, the type material is not lost, although the holotype held by the Zoological

222

Zoologica Scripta, 34, 2, March 2005, pp199–224 • © The Norwegian Academy of Science and Letters 2005

B. Dayrat & T. Gosliner • Species names and metaphyly

Appendix III Continued Museum, Copenhagen, is said to be ‘almost certainly the holotype of indecora.’ Distribution: Mediterranean Sea, Canary Islands. Autapomorphy: the number of vestibular glands and stylets varies and therefore can hardly be used as an autapomorphy (Bergh 1881, 1884a; Schmekel & Portmann 1982; Perrone, 1990; Ortea 1995).

ketos Marcus & Marcus, 1967. Originally described in Tayuva, recently reallocated to Discodoris by Valdés (2002). Material examined: lectotype, Mexico, Sonora, Puerto Penasco, Morse Beach, 2 November 1963, one specimen 30 mm preserved length, leg. Peter Pickens, USNM 678409; Pacific Ocean, Galapagos Islands, Isla Baltra, 10 September 1991, one specimen 18 mm preserved length, leg. T. M. Gosliner, CASIZ 078386; Mexico, Nayarit, Sayulita, 24 January 1975, two specimens 31 and 23 mm preserved length, leg. Gary McDonald, CASIZ 069113. Distribution: Eastern Pacific (Panamic Province), its record from Galapagos Islands is new. Autapomorphy: morphology does not discriminate ketos from maculosa and lilacina, mainly due to character variation among individuals within each species and among species. However, they are considered as different entities because of their geographical distribution (Indo-West-Pacific for lilacina, and Mediterranean Sea and eastern European Atlantic for maculosa) (Marcus & Marcus 1967, 1970). lancei Millen & Bertsch, 2000. Originally described within the genus Peltodoris. Material examined: paratype, Mexico, Baja California, Gulf of California, Bahía de Los Angeles, 29°02′56′′ N, 113°32′29′′ W, 28 June 2000, one specimen 60 mm preserved length, leg. Mike D. Miller, CASIZ 118572. Remark: (#7, #8) caryophyllidia, not described by Millen & Bertsch (2000), are unquestionably present. Distribution: Baja California. Autapomorphy: colour of the notum homogeneously orange-brown to reddish-orange (Millen & Bertsch 2000). lentiginosa Millen, 1982. Originally described within the genus Anisodoris. Material examined: holotype, Canada, British Columbia, Vancouver Island, Barkley Sound, Dixon Island, 12 m depth, 7 July 1977, one specimen 90 mm length (alive) and 70 mm preserved length, leg. Sandra Millen, CASIZ 024099; Canada, British Columbia, Bamfield Marine Station, 29 April 2000, two specimens 42 and 75 mm preserved length, leg. Mike Blanc, CASIZ 168031. Distribution: southern British Columbia, Canada. Autapomorphy: penial papilla covered with rows of platelike spines (Millen, 1982). lilacina Gould, 1852. Originally described within the genus Doris and transferred to Discodoris in 1999 by Rudman (www.seaslugforum.net). The type material is lost. Material examined: Madagascar, Mora Mora, opposite North Pass, near shore, 7 m depth, 31 March 1990, one specimen 25 mm preserved length, leg. T. M. Gosliner, CASIZ 073241; Hawaii, Kauai, South side of Island, Poipu Beach Park, 27 August 1992, one specimen 26 mm preserved length, leg. T.M. Gosliner, CASIZ 086958; Philippines, Luzon, Batangas Province, Maricaban Island, Sepok Wall, 15 m depth, 12 May 2001, one specimen 28 mm preserved length, leg. A. Valdés, CASIZ 158266; CASIZ 069969, Japan, Ryukyu Islands, Okinawa, 1.3 km ENE of Maeki-zaki, Seragaki Beach, 26°30.4′ N, 127°52.6′ E, 3 m depth, 29 April 1989, one specimen 26 mm preserved length, leg. R. F. Bolland, CASIZ 069969; Indonesia, Sulawesi, 8 km NW of Ujung Pandang, Samalona Island, 2 m depth, 10 October 1982, one specimen 21 mm leg. Antonio J. Ferreira, CASIZ 068283. Distribution: Indo-West-Pacific. Autapomorphy: morphology does not discriminate lilacina from maculosa and ketos, mainly due to character variation among individuals within each species and among species. However, they are considered as different entities because of their geographical distribution (Panamic province of eastern Pacific for ketos, and Mediterranean Sea and eastern European Atlantic for maculosa). Because of problems of identification, no previous anatomical description was used here. liturata Bergh, 1905. Originally described within the genus Discodoris. Material examined: Papua New Guinea, Solomon Sea, Louisiade Archipelago, Calvados Chain, Rawa Reef, Yuma Passage, Snake Channel, 3 June 1998, one specimen 17 mm preserved length, leg. Marty Fenton, CASIZ 113658; Indonesia, Sulawesi, Lembeh Strait, Serena Island, 3 November 1993, two specimens 38 and 40 mm preserved length, leg. Pauline Fiene, CASIZ 097595. Remark: The radula of liturata is unique because its formula is asymmetrical (Fig. 5A, B: 52x (20/22-0-14/16)). Also, the shape of the hook of the hamate lateral teeth is not the same on the left side and the right side: on the left side, the superior lip of the lateral groove of the hook is expanded, but not on the right side (#14). We also found this asymmetry in formula and tooth shape in some non-liturata unidentified specimens from Japan too. We think that it is not due to teratology and that the phylogenetic significance of this feature is real. Distribution: Indonesia and Papua New Guinea (this study). Autapomorphy: the mimetic notum of Phyllidiella pustulosa (Cuvier, 1804) is diagnostic of liturata. maculosa Bergh, 1884. Originally described in Discodoris. Material examined: holotype, Mediterranean Sea, Italy, near Naples, May 1882, one specimen 44 mm preserved length, ZMUC GAS-2131; Mediterranean Sea, Spain, Almería, Los Escullos, 30°48′ N, 2°04′ W, 25 January 1990, 1 m depth, one specimen 30 mm preserved length, leg. J. Templado, MNCN 15.05/801; Mediterranean Sea, Spain, Granada, Malaga, Cala Vaca, Cerro Gordo, 12 June 1993, one specimen 38 mm preserved length, leg. J. Templado, MNCN 15.05/17.769. Distribution: Mediterranean Sea, eastern European Atlantic. Autapomorphy: morphology does not discriminate maculosa from lilacina and ketos. However, they are considered as different entities because of their geographical distribution (Panamic province of eastern Pacific for ketos and Indo-West-Pacific for lilacina) (Bergh 1884a). mauritiana Bergh, 1889: 815. Originally described within the genus Peltodoris, often placed in the genus Discodoris (e.g. Marshall & Willan 1999). Material examined: Indian Ocean, Mascarene Islands, Reunion Island, Saint-Gilles les Bains, 20 April 1989, one specimen 60 mm preserved length, leg. T. M. Gosliner, CASIZ 070389; Madagascar, Mora Mora, Barrier Reef, 20 March 1990, one specimen 19 mm preserved length, leg. T. M. Gosliner, CASIZ 073317; Thailand, Andaman Sea, southern end of Phuket Island, Rawaii Beach, 22 March 1985, one specimen 26 mm preserved length, leg. Kit Stewart, CASIZ 073508; New Caledonia, Baie de Touho, 20°46.5′ S, 165°14′ E, 1–4 m depth, September 1993, Stn 1252, two specimens 32 and 38 mm preserved length, leg. Expedition Montrouzier, MNHN; Japan, Ryukyu Islands, Okinawa, 1.3 km ENE of Maekizaki, Seragaki Beach, 26 May 1991, one specimen 37 mm preserved length, leg. R. F. Bolland, CASIZ 079243. Distribution: Indian Ocean (type locality and present study), Thailand (present study), Hong Kong, Australia, New Caledonia (present study), Japan (present study). Autapomorphy: the colour pattern of the notum, with a whitish ground colour bearing indefinite pale grey blotches and a few small black dots, is fully characteristic (e.g. Marshall & Willan 1999). mavis Marcus & Marcus, 1967. Originally described in Discodoris, and transferred to Geitodoris. Material examined: Mexico, Baja California Sur, Gulf of California, Bahia de Los Angeles, Punta Gringa, 4 m depth, 3 October 1984, three specimens 20 mm preserved, leg. T. M. Gosliner, CASIZ 072930. Distribution: Eastern Pacific, Panamic Province. Autapomorphy: According to Marcus and Marcus, the pinkish dorsal colour (Marcus & Marcus 1967). mulciber Marcus, 1970. Originally described in the genus Percunas, and then transferred to Paradoris by Marcus (1976). Material examined: Costa Rica, Atlantic coast, one specimen 25 mm preserved length, leg. Y. Camacho, INBIO 501499. Distribution: Brazil, Costa Rica (Atlantic side). Autapomorphy: mulciber is hardly delineated because characters used in Paradoris taxonomy vary greatly. However, mulciber is the only Paradoris species found with this distribution (Marcus 1970, 1976). nobilis MacFarland, 1905. Originally described in Montereina, transferred to Anisodoris by MacFarland (1906) and more recently in Peltodoris by Valdés (2002). Material examined: Calidornia, Monterey Bay, Monterey County, Pacific Grove, March 1967, three specimens 35–45 mm preserved length, leg. A. Cameron, CASIZ 007098; California, Monterey County, off Pont Sur, 36°16′21′′ N, 121°57′09′′ W, 16 September 1988, one specimen 38 mm preserved length, leg. T. M. Gosliner, CASIZ 087212. Distribution: Northern Pacific, from Alaska to Mexico. Autapomorphy: Notum thickly tuberculate and orange mottled with irregular blotches of dark (MacFarland 1905, 1906, 1966). nubilosa Pease, 1871. Originally described in Doris, transferred to the genus Sebadoris by Marcus & Marcus (1959). Material examined: Indian Ocean, Seychelles, Mahe Island, 12 March 1986, one specimen 68 mm preserved length, leg. T. Gosliner, CASIZ 074202; Indian Ocean, Seychelles, Aldabra, Island, Passe Femme, 24 April 1984, one specimen 70 mm preserved length, leg. T. M. Gosliner, CASIZ 074252; Pacific Ocean, French Polynesia, Austral Islands, Rurutu Island, South of Avera, leg. G. Paulay, one specimen 65 mm preserved length, CASIZ 071727. Distribution: Indo-Pacific. Autapomophies: numerous pointed tubercles of unequal size and tightly close to each other, penis with two rows of small protuberances (Pease 1871; Marcus & Marcus 1959; Kay & Young 1969; Edmunds 1971; Soliman 1980).

© The Norwegian Academy of Science and Letters 2005 • Zoologica Scripta, 34, 2, March 2005, pp199–224

223

Species names and metaphyly • B. Dayrat & T. Gosliner

Appendix III Continued pardus Behrens & Henderson, 1981. Originally described within Jorunna. Material examined: Holotype, California, Channel Islands, Anacapa Island, Cat Rock, one specimen 45 mm preserved length, 25 October 1979, leg. David W. Behrens, CASIZ 015862; California, Channel Islands, Anacapa Island, Frenchy’s Cove, 20 m depth, four specimens from 33 to 39 mm preserved length, 9 September 1998, leg. David W. Behrens, CASIZ 115285. Distribution: California, Baja California (western coast). Autapomophy: ground colour yellowish with leopard-like black spots (Behrens & Henderson 1981). planata Alder & Hancock, 1846. Originally described within Doris and transferred to the genus Geitodoris by Eliot (1905b). Material examined: Mediterranean Sea, Italy, Naples, 1902– 03, one specimen 13 mm preserved, leg. F. M. MacFarland, CASIZ 081885. Distribution: Eastern Atlantic Ocean, Mediterranean Sea. Remark: Valdés (2002) provided a description of planata that is fully based on syntypes of complanata. Here we decided not to confuse the descriptions of specimens from the western (complanata) and eastern (planata) sides of the Atlantic Ocean (Alder & Hancock 1846; Eliot 1905a,b; Ortea 1990; Cervera et al. 1985). pulchra MacFarland, 1905. Originally described within Rostanga. Material examined: holotype, California, Monterey Bay, one specimen 15 mm preserved length, leg. F. M. MacFarland, NHMSI 181292; California, Los Angeles County, Malibu, Paradize Cove, two specimens 15 mm preserved length, 29 October 1971, leg. Shane Anderson, CASIZ 069842; California, Montery County, Asilomar, one specimen 11 mm length, 16 January 1973, leg. Gary McDonald, CASIZ 072041. Distribution: Pacific Coast, from Alaska to Chile. Autapomophy: brush-like marginal teeth divided into 2–7 denticles, and first lateral with 4–11 denticles (MacFarland 1905, 1966; Chia & Koss 1978, 1982, 1983, 1984; Chia et al. 1981). punctuolata d’Orbigny, 1837. Originally described in Doris, reallocated into Anisodoris by Bergh (1898) and more recently into Diaulula by Valdés & Gosliner (2001). Material examined: Chile, Lacuay Peninsula, Chiloe Island, 10 m depth, one specimen 54 mm preserved length, leg. S. Millen, M. Schrödl, and S. Gigglinger, CASIZ 118012, voucher (Valdés & Gosliner 2001). Remark: some radular characters could not be observed because the radula was removed from the material available. Distribution: Atlantic and Pacific Magellanic Province. Autapomorphy: colour pattern of the notum homogeneously yellowish (Bergh 1898; Marcus 1959; Valdés & Gosliner 2001). rosi Ortea, 1979. Originally described within Discodoris. Material examined: Portugal, Atlantic coast, Setubal District, Outao, 38°29.36′ N, 08°58.48′ W, 7 May 2002, one specimen 9 mm preserved length, leg. Goncalo Calado, CASIZ 166874. Gibraltar Strait, Ceuta, Punta Almina, 25–40 m depth, 35°54.1′ N, 05°16.5′ E, May 1986, one specimen 7 mm preserved length, leg. Bouchet, Gofas, Lozouet, MNHN. Distribution: Mediterranean Sea (Cattaneo-Vietti et al. 1990), Atlantic Coasts of Spain (Ortea 1979) and Portugal (present study), and most probably South Africa. Remark: the vestibular gland mentioned by Ortea (1979) was not found. Autapomophies: the hook is slender and elongate in all lateral teeth; red ground dorsal colour with a few brown rings; the exact same colour pattern was observed in a single specimen called Anisodoris sp. 2 by Gosliner (1987) (Ortea 1979; Ortea & Urgorri 1979). rubra Bergh, 1905: 104. Bergh (1905) also gave the epithet rubra to a species which he placed in the genus Diaulula, which is why the epithet-based name must include a page number. Originally described within the genus Discodoris. Material examined: Papua New Guinea, north coast, near Madang, Pig Island, Barracuda Point, 15 m depth, 13 January 1988, two specimens 62 and 32 mm preserved length, leg. T. M. Gosliner, CASIZ 071238; Thailand, Gulf of Thailand, off Sattahip, Ko-I-Lao, 3 February 1972, one specimen 58 mm preserved length, leg. Franz B. Steiner, CASIZ 081904; Japan, Ryukyu Islands, Okinawa, Seragaki Beach, 26°30.4′ N, 127°52.6′ E, 17 July 1992, one specimen 68 mm preserved length, leg. R. F. Bolland, CASIZ 087907. Distribution: Indonesia, Papua New Guinea, Thailand, Japan, Hawaii. Autapomorphy: red ground colour of the notum, with some tubercles bearing diagnostic black dots at the apex (see Bergh 1905). sandiegensis Cooper, 1863. Originally described within Doris and transferred to the genus Diaulula by Bergh (1880). Material examined: California, Monterey Bay, Monterey County, Pacific Grove, 38 specimens from 16 to 65 mm preserved length (longest specimen fully dissected), leg. Frank M. MacFarland, CASIZ 068279; Alaska, Aleutian Islands, Kanaga Pass, 51°47′ N, 177°39′ W, one specimen 50 mm preserved length, 4 July 1994, leg. Roger N. Clark, Expedition F/V ‘Vesteraalen’, CASIZ 106537. Distribution: north-eastern Pacific Coast, from Alaska to Baja California, as well as Russian and Japan northern Pacific waters. Autapomophy: ground colour yellowish with a few dark brown rings of varying size and position, in general arranged in two longitudinal rows on each side of the median line (Cooper 1863; Bergh 1880; MacFarland 1905, 1966; McDonald 1983; Valdés & Gosliner 2001). schmeltziana Bergh, 1877. Originally described within the genus Discodoris. Material examined: Tanzania, Zanzibar, Ras Nungwi, 6 November 1994, one specimen 58 mm preserved length, leg. Connie Boone, CASIZ 099355; Indonesia, Sulawesi, Lembeh Strait, Pulau Kecil, 6 m depth, 6 November 1993, one specimen 25 mm preserved length, leg. Pauline Fiene, CASIZ 097591; Philippines, Luzon Island, Batangas Province, Balayan Bay, Seafari Beach, 23 April 1997, one specimen 57 mm preserved length, leg. T. M. Gosliner, CASIZ 110381; Japan, Ryuku Islands, Okinawa, 26°17.8′ N, 127°54.3′ W, 18 May 1989, leg. Robert F. Bolland, CASIZ 070019. Distribution: Pacific Ocean (Society Islands), Japan (this study), Indonesia and Philippines (this study), Eastern Africa and Indian Ocean (this study), South Africa (Gosliner 1987). Autapomorphy: white and red indefinite blotches on the notum with a few dark blue larger tubercles (see ‘Discodoris sp.2’ in Gosliner 1987) (Bergh 1877). tomentosa Cuvier, 1804. Originally described within Doris, transferred to Jorunna by Bergh (1876). Material examined: Spain, Asturias, Ovinana, two specimens 30 mm preserved length, August 1979, leg. J. Ortea, CASIZ 115215, Voucher (Valdés & Gosliner 2001); South Africa, Cape Province, Atlantic Coast, Eland’s Bay, two specimens 17 and 18 mm preserved length, 17 February 1980, leg. T. M. Gosliner. CASIZ 073939. Distribution: Eastern Atlantic Ocean (from the Faeroes to South Africa, Mediterranean Sea). Autapomophy: colour pattern of the notum (Camacho-Garcia, pers. com.) (Cervera et al. 1986; Valdés & Gosliner 2001).

224

Zoologica Scripta, 34, 2, March 2005, pp199–224 • © The Norwegian Academy of Science and Letters 2005