A phylogenetic analysis of Tubificinae and Limnodriloidinae (Annelida, Clitellata, Tubificidae) using sperm and somatic characters Blackwell Science, Ltd ROBERTO MAROTTA, MARCO FERRAGUTI & CHRISTER ERSÉUS Accepted: 24 April 2002 Marotta, R., Ferraguti, M. & Erséus, C. (2003). A phylogenetic analysis of Tubificinae and Limnodriloidinae (Annelida, Clitellata, Tubificidae) using sperm and somatic characters. — Zoologica Scripta, 32, 255 – 278. The spermatozoa and the sperm aggregates of 13 species belonging to four genera (Smithsonidrilus, Limnodriloides, Thalassodrilides, Doliodrilus) in the tubificid subfamily Limnodriloidinae (Annelida, Oligochaeta) were studied and compared with the spermatozoal patterns already described in the subfamily Tubificinae. Two characters considered exclusive for the Tubificinae were found in the more spermatologically variable Limnodriloidinae: the production of two kinds of spermatozoa, eusperm and parasperm, and the presence, in the spermathecae, of sperm aggregates formed by a combination of the two sperm types. A parsimony analysis was performed on the spermatozoal data of the species examined and compared with that based on the somatic characters of the same species: a critical revision of the already codified eusperm characters was carried out and the ultrastructure of parasperm was used as a new subset of spermatozoal characters. In a total evidence approach, a further parsimony analysis was run using a matrix combining both sets of characters. This analysis suggested that the double sperm line and the sperm aggregates composed of both eusperm and parasperm may well be homologous in tubificines and limnodriloidines. It thus supported the previous notion that Tubificinae and Limnodriloidinae are closely related and indicated that these subfamilies may be sister taxa. R. Marotta, M. Ferraguti, Department of Biology, University of Milano, 26 via Celoria, I-20133, Milano, Italy. E-mail: roberto.marotta@unimi.it C. Erséus, Department of Invertebrate Zoology, Swedish Museum of Natural History, Box 50007, SE-10405 Stockholm, Sweden. Introduction Tubificidae is a large and cosmopolitan group of aquatic oligochaete worms living in freshwater, brackish water and marine habitats. The family comprises about 1000 species at present and is divided into five subfamilies: Rhyacodrilinae Hrabe, 1963, Phallodrilinae Brinkhurst, 1971, Limnodriloidinae Erséus, 1982, Telmatodrilinae Eisen, 1885 and Tubificinae Eisen, 1885 (Erséus 1990). In recent years, however, phylogenetic analyses based on different character types have suggested that this is not a phylogenetic classification: Tubificidae itself and some of its constituent subfamilies are probably paraphyletic groups (Erséus et al. 2000, 2002). Morphological (Erséus 1990; Brinkhurst 1994), spermiological (Ferraguti & Erséus 1999) and molecular data (Erséus et al. 2000, 2002) suggest that the Rhyacodrilinae is a paraphyletic taxon, and that the Naididae, another oligochaete family, is also a tubificid group. The Tubificinae and Limnodriloidinae are spermiologically interesting as they are the only euclitellates known to © The Norwegian Academy of Science and Letters • Zoologica Scripta, 32, 3, May 2003, pp255 – 278 possess a double sperm line (Ferraguti 2000), producing two sperm types that differ in their structure and function — eusperm and parasperm (Healy & Jamieson 1981). Ultrastructural studies on species belonging to five different tubificine genera have shown that their different sperm models are variations on the same structural theme. Moreover, in all tubificines examined, parasperm and eusperm are grouped to form, in their spermathecae, characteristically organized rod-shaped spermatozeugmata (Ferraguti et al. 1989, 2002). Although morphological studies have suggested that the Limnodriloidinae is the sister group of the Tubificinae– Telmatodrilinae assemblage (Erséus 1990; Brinkhurst 1994), there is limited knowledge concerning its sperm ultrastructure. So far, we have spermatological data for only two species, Smithsonidrilus hummelincki (Righi & Kanner 1979) and Thalassodrilides ineri (Righi & Kanner 1979), and some limited information for S. tuber (Erséus 1983) [reviewed in Ferraguti (2000)], but this already suggests a higher 255 Phylogeny of Tubificinae and Limnodriloidinae • R. Marotta et al. spermatological variability inside this subfamily than in the Tubificinae. While S. hummelincki has a double sperm line, which to date is the only documented case of dichotomous spermatogenesis in euclitellates outside the Tubificinae, T. ineri produces only one sperm type. Their spermatozeugmata also differ considerably: T. ineri, considered a highly apomorphic form (Righi & Kanner 1979), has a large, single spermatozeugma in each spermatheca and it is formed by only euspermatozoa which are in close contact with a large proteinaceous structure; in the spermathecae of S. hummelincki, in contrast, there is a mixture of several small spermatozeugmata, each one formed by either eusperm or parasperm (Erséus & Ferraguti 1995). Light microscopical observations of limnodriloidine species belonging to the large genus Limnodriloides revealed a great variation in the shape and organization of the sperm aggregates, from poorly organized sperm bundles to rodshaped arrangements foreshadowing the more complex tubificine spermatozeugmata (Erséus 1982). The phylogenetic meaning of the tubificine-like spermatozeugmata among tubificids is problematic. In his cladistic analysis, Erséus (1990) considered the presence of this type of spermatozeugma as the only synapomorphy of Tubificinae and Limnodriloidinae, but noting that this ‘is an assumption … somewhat weak because the detailed structure of sperm aggregates in various taxa is poorly known’. Brinkhurst (1994), however, regarded the tubificine-like spermatozeugmata as an autapomorphy for the Tubificinae; while the sperm of limnodriloidines were interpreted as ‘bundles’ only. In this study, we compared the sperm morphology and organization of 13 species belonging to four limnodriloidine genera: Limnodriloides, Smithsonidrilus, Thalassodrilides and Doliodrilus; the sperm ultrastructure of 11 of these species is described for the first time. Our aims were to establish whether a double sperm line is present in more than one species of Limnodriloidinae, and to describe the ultrastructure of the various sperm aggregates inside this subfamily, to understand if they are useful as phylogenetic markers and if they are models en route towards the tubificine spermatozeugmata. Using the spermatozoal characters, our purpose was also to improve the cladistic analysis of tubificids, with particular attention to the relationships inside the Limnodriloidinae and between this subfamily and Tubificinae. Within the Euclitellata, the ultrastructure of spermatozoa has proved useful for phylogenetic assessment, not only at higher (Jamieson et al. 1987; Ferraguti & Erséus 1999) but also at lower taxonomic levels, among species of Enchytraeus (Westheide et al. 1991) and Branchiobdellidae (Cardini et al. 2000). On the other hand, due to homoplasy among many sperm characters, the relationships within Tubificidae were poorly resolved in a study by Erséus & Ferraguti (1995). 256 For this reason, we carried out a critical revision of the already codified eusperm characters (Erséus & Ferraguti 1995; Ferraguti & Erséus 1999) and used, for the first time in an analysis of oligochaete phylogeny, the ultrastructural characters of parasperm. These sperm types may be less constrained by the biology of fertilization and thus reveal more of the phylogenetic structure than eusperm. The results of the parsimony analysis based on spermatozoal morphology are compared with a parallel phylogenetic analysis of the same species but based only on morphological characters, and with an analysis performed on a data matrix combining both characters sets (Kluge 1989). Materials and methods Specimen collection and microscopical techniques Specimens of S. luteolus (Erséus, 1983) were collected in subtidal muds and sands in Belize, in April 1993. Material from S. sacculatus (Erséus, 1983), L. tenuiductus Erséus, 1982, L. uniampullatus Erséus, 1982 and L. armatus Erséus, 1982 was collected in subtidal coral reef sand, at Heron Island, Great Barrier Reef, Australia, in April 1994. Individuals of S. hummelincki (Righi & Kanner, 1979), S. westoni Erséus, 1982, L. monothecus Cook, 1974, L. barnardi Cook, 1974, T. gurwitschi (Hrabe, 1971) and T. bruneti Erséus, 1990 were collected from intertidal and subtidal sediments at Lee Stocking Island, near Great Exuma, Bahamas, in April 1999. A new species of Doliodrilus [described separately by Wang & Erséus (2003)] was collected from brackish water sediments in Hainan, southern China, in March 2000. Specimens of Isochaetides arenarius (Michaelsen, 1926) were collected in Lake Baikal, in July 1996. The last mentioned material was provided by P. Martin, all other worms were collected by C. Erséus. The specimens were fixed in a cacodylate-buffered paraformaldehyde–glutaraldehyde mixture in a saturated solution of picric acid (SPAFG; Ermak & Eakin 1976). They were sent in SPAFG to Milan, where they were washed overnight in 0.1 M cacodylate buffer (pH 7.4), postfixed in 1% osmium tetroxide in the same buffer for 2 h, washed in distilled water, en bloc stained for 2 h in the dark in 2% aqueous uranyl acetate, dehydrated in a graded ethanol series and embedded in Spurr’s resin. Sections were cut with an LKB Ultratome III or V or with a Reichert Ultracut E. Thick (about 0.5 µm) sections were cut and stained to localize sperm funnels and spermathecae, and thus mature spermatozoa and spermatozeugmata, using a light microscope; thin sections were stained with lead citrate, carbon coated and observed with a JEOL 100SX transmission electron microscope. Taxa The ingroup and outgroup taxa used for the phylogenetic analysis are listed in Table 1. They cover a range of marine Zoologica Scripta, 32, 3, May 2003, pp255– 278 • © The Norwegian Academy of Science and Letters R. Marotta et al. • Phylogeny of Tubificinae and Limnodriloidinae Table 1 List of ingroup and outgroup tubificid taxa included in the parsimony analyses. The genera are grouped according to their current subfamilial position. Taxon Reference Ingroup taxa Tubificinae Tubifex tubifex Clitellio arenarius Isochaetides arenarius Limnodriloidinae Smithsonidrilus westoni Smithsonidrilus hummelincki Smithsonidrilus luteolus Smithsonidrilus sacculatus Limnodriloides armatus Limnodriloides uniampullatus Limnodriloides tenuiductus Limnodriloides monothecus Limnodriloides barnardi Thalassodrilides ineri Thalassodrilides bruneti Thalassodrilides gurwitschi Doliodrilus sp. Outgroup taxa Rhyacodrilinae Rhizodrilus russus Monopylephorus limosus Phallodrilinae Pectinodrilus molestus Coralliodrilus rugosus and brackish water tubificids belonging to the subfamilies Limnodriloidinae and Tubificinae. The Rhyacodrilinae + Phallodrilinae assemblage emerged as the sister group of Limnodrilinae + Tubificinae in cladistic analyses based on morphological (Erséus 1990), and combined spermatozoal and morphological characters (Erséus & Ferraguti 1995). Thus, the outgroup taxa for rooting the trees (see Table 1) were selected from these other subfamilies, the rhyacodrilines Monopylephorus limosus Hatai, 1898 and Rhizodrilus russus Erséus, 1990 (see Ferraguti et al. 1994) and the phallodrilines Pectinodrilus molestus (Erséus, 1988) and Coralliodrilus rugosus, Erséus, 1990 (see Erséus & Ferraguti 1995). Characters and character states In this study two different sets of characters were used, referred to as the ‘spermatozoal’ and ‘somatic’ characters. The characters are ‘operationally equivalent to putative synapomorphies’, i.e. ‘conjectural, based on similarity’ statements of primary homology (de Pinna 1991). Spermatozoal characters. Twenty-nine spermatozoal characters, all treated as unordered, were considered (Appendix 1 — Fig. 7). Of these, 13 (characters 1–13) refer to euspermatozoa and 13 (characters 14–26) deal with paraspermatozoa. Three other characters pertain to another level of sperm organization, © The Norwegian Academy of Science and Letters • Zoologica Scripta, 32, 3, May 2003, pp255 – 278 Braidotti & Ferraguti (1982); Braidotti et al. (1980); Ferraguti et al. (1989) Ferraguti & Ruprecht (1992); Ferraguti et al. (1994) Ferraguti et al. (2002) Present study Erséus & Ferraguti (1995); present study Present study Present study Present study Present study Present study Present study Present study Ferraguti et al. (1989) Present study Present study Present study Ferraguti et al. (1994) Ferraguti et al. (1994) Erséus & Ferraguti (1995) Erséus & Ferraguti (1995) i.e. the presence/absence of a double sperm line (character 27), the various ways in which the spermatozoa aggregate to form spermatozeugmata (character 28), and the presence/ absence of junctions between the tails of the parasperm within the spermatozeugmata (character 29). Somatic characters. Twenty-nine somatic characters were considered, of which five can be treated as ordered (Appendix 2). Three characters are associated with the morphology of the chaetae, two to the morphology of the oesophagus and one to the presence of coelomocytes around the digestive tube. All the other 23 characters concern the morphology of the reproductive system; five of them deal with the morphology of the spermatheca, the others with the different parts of the male apparatus. Parsimony analysis Partitioned somatic and sperm data analysis. Each set of characters was analysed using PAUP version 4.0b8 for 32-bit Microsoft WINDOWS (Swofford 1998). Strict parsimony analysis was used to determine character congruence as a test to evaluate conjectures of homology, i.e. primary homologies. Accelerated transformation optimization (ACCTRAN) was used for tracing character evolution. The exact branch-and-bound search algorithm was selected, with the following options: furthest addition 257 Phylogeny of Tubificinae and Limnodriloidinae • R. Marotta et al. sequence (ADDSEQ = FURTHEST), save all minimal trees found during the branch-and-bound search (MULTREES = YES) and collapse a branch if its minimum possible length is zero (COLLAPSE = MINBRLEN). Congruence and data support. To evaluate the statistical significance of character incongruence among different data partitions, for the same group of taxa, the incongruence length differences (ILD) test (Mickevich & Farris 1981; Farris et al. 1995) was used. To assess the incongruence between eusperm and parasperm data, the ILD test was performed, culling, prior to the analysis, all taxa with only one sperm line. One thousand ILD replicates, using branch-and-bound searches, were used to estimate the null distribution. Including all taxa, the incongruence was examined between spermatozoal and somatic characters. Five hundred replicates, using heuristic search, were used to estimate the null distributions. To evaluate the support for tree topologies, 500 bootstrap replicates were subjected to a separate heuristic search with tree bisection-reconnection branch swapping, 10 random addition sequences were calculated for each bootstrap replicate (ADDSEQ = RANDOM; NREPS = 10) and majority rule bootstrap trees were calculated (Felsenstein 1985). Combined analysis. The spermatozoal and somatic characters were combined in a single morphological data set (Appendix 3) and were considered, in a first analysis, all unordered. In a second analysis, the spermatozoal character 5 (acrosome vesicle withdrawn) and the somatic characters 35 (peripheral position of spermathecal pores), 40 (entrance of vas deferens into atrium), 45 (position of prostatic pad in atrial wall), 49 (length and shape of atrial ampulla) and 52 (atrial duct sac) were considered as ordered, as, in each of these cases, a particular transformation series, reflecting an order of relative similarity, could be postulated. Results Sperm morphology A great variation in sperm models and organization characterizes the Limnodriloidinae. This variability concerns the presence or absence of a double sperm line, the ways in which the sperm aggregate to form different spermatozeugmata, and the morphology of the individual spermatozoa. Euspermatozoa. The euspermatozoa of all observed species share their general structural plan with other oligochaetes, i.e. they are filamentous cells with a tapering, tubular, terminal acrosome, an extremely elongate nucleus which may be straight, spirally flanged or spiral, a cylindrical sometimes spiral midpiece, interpolated between nucleus and axoneme, with two to six radially adpressed mitochondria, and a long flagellum (Jamieson 1981); they also have a characteristic basal cylinder (Ferraguti 1984a) and two modifications of the central axonemal apparatus referred to as ‘tetragon fibres’ and a ‘prominent central sheath’ (Ferraguti 1984b). For a further understanding of the sperm descriptions, please refer to Appendix 1 (Fig. 7). Smithsonidrilus species. The euspermatozoa of S. hummelincki, S. luteolus (Fig. 1A) and S. westoni (Fig. 1E) have an acrosome formed by a characteristic thin acrosomal tube (length 0.3– 0.5 µm) of uniform thickness and ending, at its posterior extremity, in an internal thickening, a limen ( Jamieson 1978). The anterior portion of the acrosomal tube contains an acrosome vesicle that protrudes anteriorly contacting apically the plasma membrane. From the posterior margin of the acrosomal vesicle, a thin secondary tube starts, and it ends, inside the acrosomal tube, near the limen. An acrosome rod passes through the secondary tube to end in an invagination of the acrosome vesicle, but without protruding from the upper margin of the acrosomal tube (Fig. 1A and E). The nucleus is a long cylinder showing regions with different shapes in the various species: it is apically corkscrew-shaped or twisted, Fig. 1 A–W. Eusperm of six species belonging to the tubificid subfamily Limnodriloidinae. For a correct interpretation of the details, please refer to Appendix 1 (Fig. 7). A–W. A–D. Smithsonidrilus luteolus. —A. Acrosome (× 82 000). —B. The twisted apical portion of the nucleus (× 52 000). —C. The twisted basal nuclear portion ( × 35 000). —D. Cross-sectioned midpiece showing the four mitochondria ( × 67 500). E–G. Smithsonidrilus westoni. —E. Acrosome (× 75 000). —F. The corkscrew-shaped apical portion of the nucleus (× 26 000). —G. The basal portion of the nucleus is straight and the mitochondria are roundish (× 35 000). —H. Cross-sections of three tails of S. luteolus showing both prominent central sheath (top) and tetragon fibres (bottom, left) as modifications of the central axonemal apparatus (× 50 000). I–L. Smithsonidrilus sacculatus. —I. Acrosome (× 82 000) [compare with (A, E)]. —J. Longitudinal section of the loosely corkscrew-shaped nucleus ( × 45 000). —K. Coiled thread-shaped basal portion of the nucleus: the basal cylinder is particularly evident (asterisk) (× 55 000). —L. Cross-section of the tail showing the tetragon fibres (arrowhead) as modification of the central axonemal apparatus ( × 50 000). M–P. Limnodriloides armatus. —M. Acrosome (× 82 000). —N. Apical portion of the nucleus (× 22 500). —O. The straight basal portion of the nucleus and midpiece (× 55 000). —P. Cross-sectioned midpiece showing the three mitochondria and a flagellum with tetragon fibres (× 50 000). Q–S. Limnodriloides tenuiductus. —Q. Acrosome (× 82 000). —R. The apical portion of the nucleus is an elongated coiled thread (× 28 000). —S. The basal portion of the nucleus is a simple coiled thread (× 35 000). T–V. Limnodriloides monothecus. —T. Acrosome (× 82 000). —U. The loosely twisted apical portion of the nucleus (× 27 500). —V. Longitudinal section of the basally straight nucleus (× 55 000). —W. Cross-sections of two tails of L. barnardi showing the tetragon fibres as modifications of the central axonemal apparatus (× 50 000). 258 Zoologica Scripta, 32, 3, May 2003, pp255– 278 • © The Norwegian Academy of Science and Letters R. Marotta et al. • Phylogeny of Tubificinae and Limnodriloidinae © The Norwegian Academy of Science and Letters • Zoologica Scripta, 32, 3, May 2003, pp255 – 278 259 Phylogeny of Tubificinae and Limnodriloidinae • R. Marotta et al. basally twisted or straight, and its diameter decreases considerably from the basis to the convex apex (Fig. 1B,C,F and G). The midpiece has four parallel and roundish mitochondria (Fig. 1D and G) and the axoneme of the tail shows regions with a prominent central sheath and the tetragon fibres in sequence (Fig. 1H). The euspermatozoon of S. sacculatus differs from those of the other three Smithsonidrilus species studied in that its acrosome (Fig. 1I) is formed by a thicker and shorter acrosome tube (length only about 0.2 µm), which shows uniform thickness and with an acrosome vesicle almost completely external to the tube (acrosome vesicle withdrawal about 20% only), and an acrosome rod that runs inside the acrosome invagination and protrudes from the upper margin of the acrosome tube (Fig. 1I). The nucleus ends with a flat apex and does not notably decrease in diameter from the basis to the apex (Fig. 1J and K). The midpiece is formed by five roundish mitochondria and finally the flagellum presents only tetragon fibres (Fig. 1L). Limnodriloides species. The euspermatozoa of L. barnardi, L. monothecus, L. tenuiductus and L. uniampullatus have a straight acrosome with an acrosome tube (length 0.2–0.3 µm) of uniform thickness and ending, at its posterior extremity, in a limen (Fig. 1Q and T). The acrosome tube contains a thin secondary tube, starting from the inner margin of the acrosome vesicle and ending near the limen, and an acrosome rod protuberant from its upper margin (Fig. 1Q and T). The nucleus is a long cylinder showing regions with different shapes: it is apically twisted, coiled thread-shaped or corkscrewshaped, basally straight or coiled thread-shaped, and its diameter is constant from the basis to the slightly convex nuclear apex (Fig. 1R and U). Limnodriloides tenuiductus and L. uniampullatus have an acrosome vesicle situated almost completely outside the acrosomal tube, and a flat nuclear apex (Fig. 1Q). The midpiece has four roundish mitochondria. The central apparatus of the axoneme shows only tetragon fibres (Fig. 1W). The euspermatozoon of L. armatus differs from those of the other Limnodriloides species observed. It has a longer acrosome tube (length about 0.4 µm); the acrosome vesicle is deeply withdrawn inside the acrosome tube, and the acrosome rod does not protrude from the anterior extremity of the acrosome tube (Fig. 1M). The nucleus ends with a convex nuclear apex and its diameter decreases consistently from the basis to the apex (Fig. 1N and O). There are only three roundish mitochondria in the midpiece and the central apparatus of the axoneme shows tetragon fibres and a prominent central sheath in sequence (Fig. 1P). Thalassodrilides and Doliodrilus species. The euspermatozoa of T. bruneti, T. gurwitschi and T. ineri have a straight acrosome with a characteristic long acrosome tube (length about 0.6 µm) that is basally thick, but then abruptly thinner. The tube has no limen and the acrosome vesicle is almost completely outside it. The thin secondary tube starts from the inferior margin of the acrosome vesicle, ends far from the base of the acrosomal tube, and thus there is a wide empty space at the base of the acrosome, the basal chamber. The acrosome rod passes through the secondary tube and enters the invagination of the acrosome vesicle, and at the same time it protrudes from the apical extremity of the acrosome tube (Fig. 2A; the protrusion is not visible in this particular section showing a slightly tangential view). The nucleus is a coiled thread, but basally straight or twisted, and has a slightly convex apex; its diameter is constant from the basis to the apex (Fig. 2B and C). The midpiece contains four roundish mitochondria. Tetragon fibres are present in the central axonemal apparatus (Fig. 2D). The euspermatozoon of Doliodrilus sp. has a very slender acrosome (length/width ratio about 7) formed by an acrosomal Fig. 2 A–R. Eusperm and parasperm in Limnodriloidinae and Tubificinae. For an interpretation of the details, please refer to Appendix 1 (Fig. 7). A–D. The euspermatozoon of Thalassodrilides gurwitschi (Limnodriloidinae). —A. The acrosome with the long acrosome tube basally thick but then (arrow) abruptly thinner (× 60 000). —B. Longitudinal, apical section of the nucleus showing its coiled thread shape (× 40 000). —C. The nucleus is basally straight (× 40 000). —D. Cross-section of the tail showing the tetragon fibres (× 67 500). E–H. The euspermatozoon of Doliodrilus sp. (Limnodriloidinae). —E. The acrosome: the acrosome tube has a helical ridge (arrowhead) (× 60 000). —F. Longitudinal, apical section of the nucleus showing its coiled thread shape × 40 000). —G. The basal portion of nucleus is loosely twisted (× 26 000). —H. Cross-sectioned midpiece showing the four mitochondria (× 80 000). I–L. The euspermatozoon of Isochaetides arenarius (Tubificinae). —I. The acrosome (× 60 000). —J. The nucleus is gently twisted in its apical portion (× 40 000). —K. The basal portion of the nucleus is straight (× 40 000). —L. Cross-sectioned midpiece showing the five mitochondria (× 67 500). M–O. The parasperm of Isochaetides arenarius as an example of tubificine parasperm. —M. The short and straight nucleus has, at the top, a simple acrosome (arrow) (× 22 500). —N. A longitudinal section of the midpiece and the basal portion of the tail shows the reduced basal cylinder (arrow) at the base of the axoneme and the plasma membrane separated from the axoneme for a long tract [compare with (Q), bottom] (× 26 000). —O. The midpiece has three mitochondria; cross-sectioned nuclei show an oval outline. An arrow points to a cross-section of parasperm acrosome (× 40 000). P–R. Limnodriloidine parasperm. —P. Longitudinal section of the parasperm of Smithsonidrilus sacculatus: the nucleus is straight and has no acrosome (× 8500). Inset: apical portion of the nucleus of Limnodriloides armatus (× 35 000). —Q. Longitudinal section of the midpiece and basal portion of the tail of S. luteolus: the basal cylinder is completely absent; the plasma membrane is separated from the axoneme for a short tract (compare with the crosssections of the spermatozeugma in Fig. 3G) (× 26 000). —R. Cross-sections of the midpieces and tails of L. armatus showing two mitochondria and no modifications of the central axonemal apparatus (× 40 000). 260 Zoologica Scripta, 32, 3, May 2003, pp255– 278 • © The Norwegian Academy of Science and Letters R. Marotta et al. • Phylogeny of Tubificinae and Limnodriloidinae © The Norwegian Academy of Science and Letters • Zoologica Scripta, 32, 3, May 2003, pp255 – 278 261 Phylogeny of Tubificinae and Limnodriloidinae • R. Marotta et al. tube that is basally thick, then abruptly thinner with a characteristic helical ridge that completes 1.5 turns; it contains a thin secondary tube and a wide basal chamber (Fig. 2E). The nucleus, apically a coiled thread and basally twisted, decreases considerably in diameter from the basis to the convex apex (Fig. 2F and G). The midpiece contains four roundish mitochondria (Fig. 2H) and only tetragon fibres are present in the central axonemal apparatus. Isochaetides. The euspermatozoon of Isochaetides arenarius (Tubificinae) has been described in detail elsewhere (Ferraguti et al. 2002), and only some pictures will be given here for reference. The slender acrosome (length /width ratio about 5.8) has an acrosome tube that is basally thick then is gradually thinner and shows, at least in the spermatozoa found inside the spermatheca, a deep withdrawal of the acrosome vesicle surrounding a long acrosomal rod (Fig. 2I). The nucleus is a long cylinder, with a corkscrew-shaped apical part, a twisted middle part and a straight basal portion, and its diameter decreases considerably from the basis to the convex apex (Fig. 2I, J and K). The midpiece contains five roundish mitochondria (Fig. 2L) and finally the flagellum presents only tetragon fibres. Paraspermatozoa. Paraspermatozoa were observed, among Limnodriloidinae, only in Smithsonidrilus and Limnodriloides species. The differences between euspermatozoa and paraspermatozoa were consistent in all nine species of these genera examined. In all of these, the paraspermatozoa are devoid of an acrosome (Fig. 2P inset), and their short (length 5–10 µm) and straight nuclei are elongated cone-shaped and regularly condensed (Fig. 2P). The mitochondria are less numerous (Fig. 2R). Finally, the tails lack a basal cylinder, and the plasma membrane is often characteristically separated from the axoneme (Fig. 2Q). In all Limnodriloides species observed, the paraspermatozoa have basally concave nuclei, each with a nearly circular outline (cross-section); and in all except L. armatus there are a midpiece with two straight and roundish mitochondria (Fig. 2R) and a tail with tetragon fibres; the paraspermatozoa of L. armatus have no modifications of the central axonemal apparatus. In all the Smithsonidrilus species examined, the paraspermatozoa have basally convex nuclei with an irregular outline in cross-section, except the paraspermatozoa of S. westoni and S. sacculatus, which have, respectively, basally concave nuclei or nuclei with an almost circular outline. The midpiece contains three (two in S. luteolus) straight and roundish mitochondria in all species except S. sacculatus in which the mitochondria are spiral and elongate. The tails have no modifications of the central axonemal apparatus, except those of S. luteolus which has tetragon fibres. The paraspermatozoon of Isochaetides arenarius (Tubificinae) has been described in detail elsewhere (Ferraguti et al. 2002), thus only some pictures will be given here for reference. The acrosome, 0.5 µm long, has no rod, no secondary tube, and its acrosome vesicle lies completely external to the tube (Fig. 2M). The cone-shaped nucleus is short (length 3.7 µm), never condensed completely, and with an oval outline in cross-section (Fig. 2M and O). The midpiece contains three mitochondria, the overall volume of which being about two times that of the mitochondria in the eusperm (Fig. 2O). The tail, showing a reduced basal cylinder inside the basal body of the flagellum, has no modification of the central axonemal apparatus, and the plasma membrane is characteristically separated from the axoneme (Fig. 2N and O). Spermatozeugmata. Inside each of the four different genera studied, a great variety in the shape and the structure of sperm aggregates in the spermathecae was observed. Thalassodrilides bruneti and T. gurwitschi are obvious exceptions as they have no spermathecae. Based on their structure, these different spermatozeugmata can be grouped into at least three classes: 1 Thalassodrilides ineri and Doliodrilus sp. have spermatozeugmata formed only by euspermatozoa, which are packed together and arranged in parallel (Fig. 3C). In T. ineri, the single spermatozeugma is connected to a peculiar proteinaceous structure (Righi & Kanner 1979; Ferraguti et al. 1989), whereas in the spermathecae of Doliodrilus sp., the few spermatozeugmata are seahorse-shaped (Wang & Erséus 2003). 2 In the Smithsonidrilus species studied, paraspermatozoa and euspermatozoa form separate spermatozeugmata inside the spermathecae. Fig. 3 A–G. Tubificine (A, B) and limnodriloidine (C–G) spermatozeugmata. —A. Cross-sectioned spermatozeugma of Isochaetides arenarius, with the axial cylinder (a) formed by parallel eusperm and the cortex (c) formed by tightly packed parasperm (× 2200). The arrow indicates helliptical cross-sections of parasperm nuclei. —B. Cell junctions connect the parasperm tails in the cortex of Isochaetides arenarius (× 45 000). —C. Cross-section of the large spermatozeugma formed by eusperm inside the spermatheca of Doliodrilus sp. (× 2200). —D. Cross-sectioned spermatozeugma of Limnodriloides monothecus, formed by both eusperm and parasperm. Note the large axial cylinder (a) and the narrow cortex (c), formed by a few rows of parasperm tails (× 5500). The arrow indicates round cross-sections of parasperm nuclei [compare with (A)]. —E. There are no visible junctions connecting the plasma membranes of neighbouring parasperm tails inside the cortex of L. monothecus (× 45 000) [compare with (B)]. —F. Cross-section of part of the large and poorly organized eusperm aggregate of Smithsonidrilus luteolus with its characteristic whirlpool-like arrangement (× 3000). —G. Cross-sectioned spermatozeugma of S. luteolus formed by parasperm only (× 7000) [compare with (A, D)]. 262 Zoologica Scripta, 32, 3, May 2003, pp255– 278 • © The Norwegian Academy of Science and Letters R. Marotta et al. • Phylogeny of Tubificinae and Limnodriloidinae © The Norwegian Academy of Science and Letters • Zoologica Scripta, 32, 3, May 2003, pp255 – 278 263 Phylogeny of Tubificinae and Limnodriloidinae • R. Marotta et al. Fig. 4 A, B. Trees obtained with a parsimony analysis of a single data set. —A. Strict consensus of 56 most parsimonious trees based on the somatic data alone (all unordered). —B. Most parsimonious tree resulting from cladistic analysis of the spermatozoal characters only (all unordered). The numbers above the internodes are bootstrap values. Branch lengths are proportional to amount of change. Particularly in S. hummelincki and S. sacculatus, the paraspermatozoa form small and irregular bundles and their shape is slightly modified in comparison with those observed at the sperm funnels, whereas the euspermatozoa form larger and more organized spermatozeugmata. In S. luteolus, the paraspermatozoa are organized in spermatozeugmata formed by a central axis of nuclei surrounded by a large cylinder of tails, and the latter with plasma membranes largely separated from the axonemes (Fig. 3G). These parasperm aggregates, which superficially resemble the tubificine spermatozeugmata but consist of only one sperm type, are distributed among less organized eusperm aggregates. The latter are spermatozeugmata with a characteristic whorl-like arrangement of euspermatozoa (Fig. 3F). We found only empty spermathecae in S. westoni. 3 In all the Limnodriloides species studied, there is only one kind of spermatozeugmata, and it is formed by a combination of the two sperm types. These spermatozeugmata vary in size and shape, from spindle to club-shaped, and from stout to slender (Erséus 1982), but they have the same general organization. As for tubificines (Fig. 3A), it is possible to recognize a large ‘axial cylinder’, formed by parallely 264 arranged euspermatozoa, surrounded by a narrower ‘cortex’, formed by a few rows of helically arranged paraspermatozoa (Fig. 3D). However, in contrast to the situation in many tubificines (Fig. 3B), there are no visible junctions connecting the plasma membranes of neighbouring parasperm tails inside the cortex (Fig. 3E). 4 Tubificine spermatozeugmata (here, for comparison in Fig. 3A and B, Isochaetides arenarius) have large spermatozeugmata formed by eusperm surrounded by parasperm. Parasperm tails are connected by prominent cell junctions. Analysis of individual data sets Somatic data. The branch-and-bound search on the somatic data alone (all characters unordered) yielded 56 equally most parsimonious trees (MPTs), all with 66 steps and a consistency index (CI) of 0.576, and a retention index (RI) of 0.728. Ordering some of the somatic characters (see above), the analysis resulted in 15 MPTs, also with 66 steps with CI = 0.576 and RI = 0.733. The strict consensus trees with all characters unordered is shown in Fig. 4A. A large basal polytomy does not allow any resolution at the subfamily level, among the Phallodrilinae, Tubificinae and Limnodriloidinae. Zoologica Scripta, 32, 3, May 2003, pp255– 278 • © The Norwegian Academy of Science and Letters R. Marotta et al. • Phylogeny of Tubificinae and Limnodriloidinae Table 2 Results of the incongruence length differences (ILD) tests for comparison among two data partitions. Comparison N°MPT L CI P-value for ILD test Eusperm–parasperm characters (1–13 vs. 14–26)* Sperm–somatic characters (1–29 vs. 30–58) 1 1 44 145 0.636 0.52 0.334 0.018 *Eight characters are uninformative due to the deletion of some taxa that have no parasperm. L, length; CI, consistency index. Inside this last subfamily, only the three Thalassodrilides species with Doliodrilus sp. and a few species belonging to Smithsonidrilus and Limnodriloides form distinct groups in all trees. Inside the Tubificinae, Isochaetides arenarius and Tubifex tubifex (Müller, 1774) are always sister species, but the position of Clitellio arenarius (Müller, 1776) is ambiguous inside the large basal polytomy. The majority rule bootstrap tree is similar to the strict consensus tree: the majority of monophyla present in all MPTs have reasonable to good support. The main difference is the lack of bootstrap support for the three Thalassodrilides species and Doliodrilus sp. Spermatozoal data. The parsimony analysis of the 29 sperm characters resulted in a single MPT, with 67 steps and CI = 0.567, RI = 0.758 (Fig. 4B). A similar tree was obtained using only the eusperm characters. Inside the outgroups Pectinodrilus molestus with Coralliodrilus rugosus (Phallodrilinae) group on their own. The Tubificinae, at the basis of one of the two ingroup clades, is paraphyletic. The monophylum formed by the three Thalassodrilides species and Doliodrilus sp., being the sister group of all other ingroup taxa, makes the Limnodriloidinae paraphyletic. Inside this last subfamily, the genera Smithsonidrilus and Limnodriloides, but not Thalassodrilides, are polyphyletic. The bootstrap analysis gave notable support only to a few monophyla, those formed by the three Thalassodrilides species and Doliodrilus sp., L. uniampullatus and L. tenuiductus and, to a lesser degree, Pectinodrilus molestus and Coralliodrilus rugosus, respectively. Congruence. Only one of the two comparisons between the different data sets yielded statistically significant results at the P < 0.05 level (Table 2). In fact, sperm and somatic characters were significantly incongruent (P-value = 0.018), whereas the incongruence between eusperm and parasperm data was not statistically significant (P-value = 0.334). Combined data analysis Spermatozoal and somatic data. Somatic and spermatozoal characters (all unordered), used together, produced a single MPT, with 144 steps and CI = 0.528, RI = 0.695 (Fig. 5). As in the tree based on only the sperm characters, Pectinodrilus molestus groups with Coralliodrilus rugosus (Phallodrilinae) outside the basis of the tubificine–limnodriloidine ingroup. The Limnodriloidinae stays paraphyletic: the clade formed by Doliodrilus sp. and the monophyletic Thalassodrilides is the © The Norwegian Academy of Science and Letters • Zoologica Scripta, 32, 3, May 2003, pp255 – 278 sister group of the assemblage formed by the Tubificinae and the other limnodriloidine taxa. The Tubificinae is a monophyletic group: Clitellio arenarius emerges as the sister taxon to Isochaetides arenarius + Tubifex tubifex. Inside the Limnodriloidinae, Limnodriloides is paraphyletic; L. armatus is sister to a group formed by the monophyletic Smithsonidrilus and the other more apomorphic Limnodriloides species. The majority rule bootstrap tree is more resolved than that obtained with the sperm characters only: there are more smaller branches that ‘survive’ the bootstrapping. However, the sister group position of the Phallodrilinae with respect to the ingroup is not resolved; there is no bootstrap support for the group Tubificinae + Limnodriloidinae. The position of the Tubificinae vis-à-vis Limnodriloidinae is also unresolved. Ordering some of the somatic and spermatozoal characters (see above), the analysis resulted in two MPTs, each with 146 steps, CI = 0.521, RI = 0.694. One tree is the same as that obtained with all characters unordered. The second MPT differs significantly: the Limnodriloidinae is monophyletic, i.e. the Tubificinae clade is sister to the Limnodriloidinae, while the three Thalassodrilides species and Doliodrilus sp. form a sister group to the rest of the limnodriloidines (Fig. 6). However, monophyly of the Limnodriloidinae is not supported by the bootstrap analysis. Discussion The double sperm line Among the five tubificid subfamilies the presence of a double sperm line is now established not only in the Tubificinae (Ferraguti 2000: Table 1), but also in the subfamily Limnodriloidinae: the species belonging to the genera Smithsonidrilus and Limnodriloides produce a dichotomous sperm line. The parasperm of limnodriloidine species are more similar to each other than the eusperm, as in the Tubificinae. In both subfamilies, the differences between parasperm and eusperm are associated with the same parts of the cell; in parasperm the various acrosome structures are reduced, the nuclei are smaller and simpler, the number of mitochondria is lower and there are differences in the axoneme and in the plasma membrane around it. However, there are some differences between the parasperm of Tubificinae and Limnodriloidinae. In the limnodriloidine parasperm, the acrosome is absent (Fig. 2P inset), whereas in tubificines this structure is vestigial (Fig. 2M). Furthermore, the nucleus is longer (about 10 µm compared 265 Phylogeny of Tubificinae and Limnodriloidinae • R. Marotta et al. Fig. 5 The single most parsimonious tree resulting from cladistic analysis of combined sperm and somatic characters (all unordered). The numbers above the internodes are boot-strap values. Branch lengths are proportional to amount of change. with 3 µm) and fully condensed in Limnodriloidinae (Fig. 2P). The parasperm mitochondria are less numerous as in Tubificinae (Fig. 2O and R), but generally in Limnodriloidinae they have the same volume as that of eusperm. A structure similar to the basal cylinder (Ferraguti 1984a), which is present, albeit reduced, in the basal bodies of tubificine parasperm (Fig. 2N), is completely absent in the parasperm of limnodriloidines (Fig. 2Q). Finally, in some limnodriloidine species, the plasma membrane is not separated from the parasperm axoneme, whereas in other species it is separated only for a short tract (Fig. 2Q). Tubificid parasperm show some morphological similarities with other parasperm models, both in annelids and other invertebrates. Among Polychaeta, for example, the parasperm of different Protodrilus species (von Nordheim 1989) and of Siboglinum ekmani, a perviate pogonophoran (Franzén 1973), as compared with the eusperm, have a reduced acrosome, a simpler and not fully condensed nucleus and a simpler midpiece with larger mitochondria. Even if ‘the “invention” of dichotomous sperm lines has been an independent phenomenon [ … ] which occurred many times probably as an answer to different evolutionary pressures’ (Ferraguti 2000: 163) in many invertebrates groups ( Jamieson 1987), the present analyses of combined somatic 266 and spermatozoal characters (see Figs 5 and 6) support the idea that in tubificids the double sperm line has originated only once, in a common ancestor of the Tubificinae, and the genera Smithsonidrilus and Limnodriloides; i.e. the double sperm line of tubificines and limnodriloidines may well be homologous. In one of the trees resulting from the parsimony analysis with some characters ordered (Fig. 6), the double sperm line appears to be secondarily lost at the basis of the Thalassodrilides + Doliodrilus clade. Although it is difficult to envisage a reversal to a single sperm line from such a complex model of dichotomous spermatogenesis, a molecular analysis of 18S rDNA suggested that the same three Thalassodrilides species plus a species of Doliodrilus (D. tener) form a terminal branch close to some species of Smithsonidrilus (Erséus et al. 2002). Moreover, the secondary loss of a multiple sperm line is a phenomenon not so unusual among the invertebrates. Inside the Gastropoda (Mollusca) there is sperm polymorphism in many groups belonging to various orders of Prosobranchia (Healy & Jamieson 1981; Giusti & Selmi 1982). Inside Prosobranchia, parasperm models diversified in the Caenogasteropoda, but they have completely disappeared in the more apomorphic Heterobranchia (Buckland-Nicks 1998). Also, in different species of stalk-eyed flies (Diopsidae), sperm dimorphism is present as an ancestral character associated with functional spermathecae, and the secondary reduction or loss of sperm storage function by spermathecae appears to coincide with the loss of sperm dimorphism (Presgraves et al. 1999). Spermatozeugmata In all species of Tubificinae studied to date, with the exception of Aulodrilus and related forms (Brinkhurst 1990), there are typical spermatozeugmata, each formed by two sperm types (Fig. 3A) (Ferraguti 2000). Within the Limnodriloidinae, this architecture is present only in Limnodriloides, whereas all examined species of the genus Smithsonidrilus produce separate spermatozeugmata of two kinds, each formed by one sperm type. The spermatozeugmata of the genus Limnodriloides, in fact, although more variable in shape and dimension (Erséus 1982), share the same general architecture with those of Tubificinae; both are composed of two different layers: an internal axial cylinder containing eusperm and an external cortex composed of parasperm, the latter packed together (Braidotti & Ferraguti 1982; Ferraguti 1998). However, the cortex of the Limnodriloides spermatozeugma is formed by fewer rows of parasperm tails than in tubificines, resulting in a thinner sheath around the axial cylinder, and the Limnodriloides parasperm tails are not connected by a junctional complex. Inside the tubificine spermatozeugmata, with one known exception for Clitellio arenarius (Ferraguti & Ruprecht 1992), there is a complex pattern of junctions inside a wider cortex (Fig. 3A and B) Zoologica Scripta, 32, 3, May 2003, pp255– 278 • © The Norwegian Academy of Science and Letters R. Marotta et al. • Phylogeny of Tubificinae and Limnodriloidinae Fig. 6 One of the two most parsimonious trees resulting from cladistic analysis of combined sperm and somatic characters (characters 5, 35, 40, 45, 49, 52 ordered). The numbers on the left of the internodes are bootstrap values. Branch lengths are proportional to amount of change. (Braidotti et al. 1980; Ferraguti 1998; Ferraguti et al. 2002). It is interesting to note that the junctions among parasperm tails are more complex and differentiated in the tubificine freshwater genera (Tubifex, Heterochaeta, Isochaetides) than in © The Norwegian Academy of Science and Letters • Zoologica Scripta, 32, 3, May 2003, pp255 – 278 the marine genus Tubificoides (Ferraguti et al. 1989, 2002). Therefore, the absence of a junctional complex in the spermatozeugmata of the genus Limnodriloides, which is exclusively marine, is perhaps not surprising. 267 Phylogeny of Tubificinae and Limnodriloidinae • R. Marotta et al. Within the other tubificid subfamilies Rhyacodrilinae and Phallodrilinae, several types of eusperm bundle have been found (Ferraguti et al. 1994; Erséus & Ferraguti 1995). They differ from the spermatozeugmata present in Doliodrilus sp. and T. ineri in being less structured. The morphology of the spermatozeugmata formed by eusperm only inside the spermathecae of Doliodrilus sp. and T. ineri differs in many aspects in these two species. The parsimony analysis indicates that these spermatozeugmata may even have arisen independently. On the other hand, in spite of the differences in morphological details, the analysis suggests that the spermatozeugmata composed of both eusperm and parasperm may be homologous in the Tubificinae and Limnodriloides. Furthermore, our analysis interprets the simpler spermatozeugmata observed in Smithsonidrilus as the result of an apomorphic secondary transformation of the tubificine-like spermatozeugmata in Limnodriloides. Character analysis To discuss the patterns of sperm and somatic characters, under a total evidence approach, the phylogram in Fig. 6 has been selected. In it, as in previous phylogenetic hypotheses (Erséus 1990; Brinkhurst 1994), the Tubificinae and Limnodriloidinae are sister groups. However, as noted above, this tree is only one of two equally parsimonious trees resulting from the analysis, and we do not judge whether it should be regarded as the ‘preferred’ hypothesis of the two. Spermatological character patterns. The presence of a thin secondary tube inside the acrosome (character 6) is the sperm character that defines the Tubificinae and Limnodriloidinae clade in relation to the outgroup. The presence of a double sperm line (character 27) is another character state common to this clade, but with a reversal for the terminal group comprising the Doliodrilus and Thalassodrilides. Autapomorphic spermatological character states for the subfamily Tubificinae refer only to parasperm, while the tubificine eusperm remain plesiomorphic. Autapomorphies are the presence of a vestigial acrosome (character 14), the presence of a shorter nucleus (character 15) and the presence of a basal cylinder inside the proximal region of the axoneme (character 26). A nearly oval nuclear outline (character 18) and the presence of a junctional complex among the parasperm tails in the spermatozeugmata (character 29) are synapomorphies for Tubifex tubifex and Isochaetides arenarius. The presence, in the eusperm, of an acrosomal tube of uniform thickness and basally with limen (character 2) is the only synapomorphic state for all Limnodriloidinae except the Thalassodrilides + Doliodrilus clade which possesses an autapomorphic basally thick, then abruptly thinner acrosome tube, and lacking a limen (character 2). An acrosomal tube of 268 uniform thickness and with a limen seems to have evolved convergently in Rhizodrilus russus. Our analysis suggests that the genus Limnodriloides is either paraphyletic (as in the example tree shown in Fig. 6) or polyphyletic. The L. armatus eusperm is different from that of the other Limnodriloides species, instead being similar to tubificine eusperm. Moreover, in L. armatus, the parasperm belong to the ‘limnodriloidine type’, but they have areas of uncondensed chromatin in their nuclei (character 19), and a plasma membrane expanded for a long tract around the axoneme (character 24). Most characters of the eusperm of the Thalassodrilides and Doliodrilus species are plesiomorphic; the synapomorphies shared with the ‘typical Limnodriloides’ (i.e. excluding L. armatus) and S. sacculatus are an acrosome with a protuberant acrosome rod (character 4) and a vesicle external to the tube (character 5). The ordering of character 5 (acrosome vesicle withdrawal), unlike that imposed on the somatic characters (see above), is not reflected in the tree of Fig. 6. Character 5 changed in its evolution in a discontinuous way, from completely inside the acrosome tube to completely outside it, at the basis of the ‘typical Limnodriloides’ and Doliodrilus + Thalassodrilides clade. Autapomorphies for the ‘typical Limnodriloides’ + Smithsonidrilus clade are the presence in the eusperm of a stout (character 1) and short acrosome (character 3). A stout acrosome occurs also as the result of convergence, in Rhizodrilus russus. No autapomorphism defines the ‘typical Limnodriloides’ clade; but the presence of parasperm with a filiform apical shape of the nucleus is synapomorphic for L. tenuiductus and L. uniampullatus. There are two synapomorphic character states for all species of Smithsonidrilus except S. sacculatus: the acrosome vesicle is only partially withdrawn inside the acrosome tube (character 5) and there are parasperm with an irregular nuclear outline in cross-section (character 18); the plasma membrane expanded around the nuclear apex of the parasperm (character 20) characterizes only the sister species S. hummelincki and S. westoni. Many spermatological characters show homoplasy (convergence or reversal to ancestral states) as is evident from the low CI value of the MPTs in Fig. 6, and this may be a common feature of sperm in lower level tubificid phylogenies (Erséus & Ferraguti 1995). Moreover, as also shown from the high RI value, most spermatozoal features show a low level of homoplasy in their evolution: they are homoplasic characters that give support for tree topology or, following de Pinna (1991), secondary homology with a low level of generality. Only seven characters show a high level of homoplasy in their evolution: four of them refer to eusperm, three to parasperm. The eusperm characters are those dealing with the shape of the nucleus (characters 7, 8, 9) and the number of mitochondria Zoologica Scripta, 32, 3, May 2003, pp255– 278 • © The Norwegian Academy of Science and Letters R. Marotta et al. • Phylogeny of Tubificinae and Limnodriloidinae (character 10), while the parasperm ones are the presence of transformed nuclei in the spermatozeugma (character 21), the number of mitochondria (character 22) and the presence of a basal body with dense material (character 23). The trend towards an increased withdrawal of the acrosome vesicle in the evolution of the oligochaete sperm types ( Jamieson et al. 1987) was not found in this analysis. Among the oligochaetes, inside the ‘more evolved’ tubificids, a completely withdrawn vesicle (inside the acrosome tube) is the plesiomorphic condition, whereas a more or less external vesicle is the apomorphic state characterizing all limnodriloidine species. Somatic character patterns. The presence/absence of coelomocytes (character 37) and the way in which the vas deferens enters into the atrium (character 40) are the somatic characters that identify the Phallodrilinae + Tubificinae + Limnodriloidinae clade, when the tree is rooted at the Rhyacodrilinae. The position of spermathecal pores in the middle part of the spermathecal segment (character 34) seems to be the only character state common to Tubificinae and Limnodriloidinae, excluding the Thalassodrilides + Doliodrilus clade. A pendant penis located centrally inside a well-developed copulatory sac (character 53) characterizes the subfamily Tubificinae, and the presence of an ‘ectoprostate’ gland (character 41) is the synapomorphic state for Tubifex tubifex and Isochaetides arenarius. The presence of an atrium bearing a lobed, broadly attached prostate gland derived from peritoneal cells (Gustavsson 2002) is the synapomorphy for all members of the Limnodriloidinae, and this gland communicates with a prostatic pad inside the atrial ampulla (character 43). The presence of paired oesophageal diverticula in segment IX (character 39) is a state common to all Limnodriloidinae analysed here except the Thalassodrilides + Doliodrilus clade. The presence of a barrel-shaped, dilated portion of the oesophagus (Gustavsson & Erséus 1999) is the only autapomorphy for the Thalassodrilides + Doliodrilus group (character 38), while the presence of a well-developed and strongly muscular atrial duct sac unites T. gurwitschi and T. ineri (character 52). As was the case with the sperm characters, L. armatus also differs in many somatic characters from the other Limnodriloides species (the ‘typical Limnodriloides’ group). No somatic synapomorphy supports the ‘typical Limnodriloides’ + Smithsonidrilus clade, but the subapical entrance of the vas deferens on the dorsal surface of the atrium (character 40) and the presence of a pseudopenial papilla (character 54) are autapomorphies for the ‘typical Limnodriloides’ group. Dilatation of the inner end of the atrium (character 47) is the only state shared by all Smithsonidrilus species excluding S. sacculatus. Only three somatic characters show a high level of homoplasy in their evolution: the shape of the teeth of bifid chaetae © The Norwegian Academy of Science and Letters • Zoologica Scripta, 32, 3, May 2003, pp255 – 278 (character 30), the absence/presence of modified spermathecal chaetae (character 32) and the size/shape ratio of the prostatic pad (character 44). Many others show a low level of homoplasy. The position of spermathecal pores (character 35), the atrial ciliation (character 46), the presence of a copulatory organ associated with an ectodermal invagination at male pores (character 53) and the presence of an ‘exterior’ spermathecal vestibule (character 58) are homoplasic because there is no uniform character state in the outgroup species. Congruence test and parsimony analysis. The inclusion of parasperm characters in our data set, even if not all taxa have a double sperm line, proved useful in the phylogenetic analysis. Thus, the presence of parasperm constitutes an important character shared by Tubificinae and Limnodriloidinae, and the ultrastructure of these cells contributes to a better discrimination of one subfamily from the other. Besides, as shown from the ILD test (Table 2), parasperm characters are not significantly incongruent with those of eusperm ones. This suggests the possibility that the two sperm types carry information of a similar phylogenetic structure. On the other hand, the ILD test shows significant incongruence between sperm and somatic characters (Table 2). This may indicate that both data sets contain some internal incongruence (homoplasy) and that the phylogenetic signal (internal congruence or synapomorphies) is, in one particular data set, stronger for some groups, while the signal is stronger for other groups in the other data set. The result of the combined data set, in a total evidence approach, gives more phylogenetic information than any of the separate data sets. The sperm tree (Fig. 4B), which differs from the somatic one (consensus tree in Fig. 4A) by being more resolved, contains only three well-supported clades: the phylogenetic structure in the sperm data is thus weak. The somatic tree is different from the sperm tree and also the patterns of bootstrap support are different. There are characters in the somatic data which give good support to groups not so strongly supported in the sperm tree. For instance, the somatic tree has a bootstrap support of 80 for a large group, comprising all species except Rhizodrilus russus and Monopylephorus limosus; this group exists in the sperm tree too, but without bootstrap support. In the somatic tree, there is strong support (94) for L. tenuiductus + L. monothecus + L. uniampullatus, which is a group that is ‘indicated’ in the sperm tree too, but with S. sacculatus inserted. Smithsonidrilus hummelincki and S. westoni, which in the sperm tree do not form a group, are located adjacent to each other at the base of a large clade in this tree; in the somatic trees the two species comprise a strongly supported clade (97). In contrast, the sperm tree contains support for groups that are not supported by the somatic data: the group formed by Doliodrilus sp. and the three Thalassodrilides 269 Phylogeny of Tubificinae and Limnodriloidinae • R. Marotta et al. species is supported by the sperm characters but is not supported (although it exists) in the somatic tree. In the tree based on the combined data with some characters ordered (Fig. 6), the number of groups supported by bootstrap values over 50 (values between 54 and 93) has greatly increased. A group comprising all species except Rhizodrilus russus and Monopylephorus limosus is well supported. The group Pectinodrilus molestus + Coralliodrilus rugosus is well supported (89), an additive effect of the already existing but less supported groups by the somatic and sperm characters alone (see Fig. 4A and B). Moreover, there is a ‘trade off’ in the group formed by L. uniampullatus + L. tenuiductus + L. monothecus, which in the combined tree only has a bootstrap value of 78, while the support was 94 using the somatic data alone. The latter fact somehow had to be balanced by the ‘interference’ of S. sacculatus in the sperm data (see above). In conclusion, even if the two data sets are significantly incongruent, they both contribute to a more resolved and more strongly supported phylogenetic hypothesis when combined. To conclude, our analysis, based on combined somatic and spermatological characters, supports the notion indicated previously (Erséus 1990; Brinkhurst 1994; Erséus & Ferraguti 1995; Ferraguti & Erséus 1999) that Tubificinae and Limnodriloidinae are closely related, possibly sister groups (Fig. 6). However, this result is neither strongly supported by the data (there is no bootstrap support >50 for a clade comprising the two subfamilies), nor based on more than a few selected species. There are about 150 species of Tubificinae (rough estimation), and over 100 species of Limnodriloidinae known (C. Erséus, unpublished observation), and we cannot be certain that our study has covered more than a fraction of the total character variation in the taxa. Nevertheless, our analysis has shown that sperm ultrastructure exhibits great similarities in tubificine and Limnodriloides spermatozeugmata, and that sperm characters may be important to broaden the basis of evidence in phylogenetic reconstruction. Acknowledgements This research has been supported by a grant (Molecular evolution and markers for phylogenesis and adaptation) from MURST (Rome) to MF. We are indebted to P. Martin (Institut Royal des Sciences Naturelles de Belgique, Belgium) for providing material of Isochaetides arenarius. The third author is grateful to K. Ruetzler (National Museum of Natural History, Washington, DC), and the directors and staff of Carrie Bow Cay Field Laboratory (Belize), Heron Island Research Station (Queensland, Australia), and Caribbean Marine Research Center (Lee Stocking Island, Great Exuma, Bahamas), for providing excellent facilities for field work; to friends and colleagues who assisted in the collection of oligochaetes (in particular, N. Dubilier, O. Giere, Z. Gong, and H. Wang), and to the Swedish Research Council (and the former 270 Swedish Natural Science Research Council), and the Perry Foundation, Inc. (operator of Caribbean Marine Research Center) for financial support. References Braidotti, P. & Ferraguti, M. (1982). Two sperm types in Tubifex tubifex spermatozeugmata. Journal of Morphology, 171, 123–136. Braidotti, P., Ferraguti, M. & Fleming, T. P. (1980). Cell junctions between spermatozoa flagella within the spermatozeugmata of Tubifex tubifex (Annelida, Oligochaeta). Journal of Ultrastructure Research, 73, 299 –309. Brinkhurst, R. O. (1990). A phylogenetic analysis of Tubificinae. Canadian Journal of Zoology, 69, 3 9 2 –397. Brinkhurst, R. O. (1994). Evolutionary relationships among the Clitellata: an update. Megadrilogica, 5, 109 –112. Buckland-Nicks, J. (1998). Prosobranch parasperm: sterile germ cells that promote paternity? Micron, 29, 267–280. Cardini, A., Ferraguti, M. & Gelder, S. R. (2000). A phylogenetic assessment of the branchiobdellidan family Branchiobdellidae (Annelida, Clitellata) using spermatological and somatic characters. Zoologica Scripta, 29, 347 – 366. Ermak, T. H. & Eakin, R. M. (1976). Fine structure of the cerebral and pygidial ocelli in Chone ecaudata (Polychaeta: Sabellidae). Journal of Ultrastructural Research, 54, 243 – 260. Erséus, C. (1982). Taxonomic revision of the marine genus Limnodriloides (Oligochaeta: Tubificidae). Verhandlungen des Naturwissenschaflichen Vereins in Hamburg (Neue Folge), 25, 207 –277. Erséus, C. (1983). Taxonomic studies of the marine genus Marcusaedrilus Righi & Kanner (Oligochaeta, Tubificidae), with description of seven new species from the Caribbean area and Australia. Zoologica Scripta, 12, 25– 36. Erséus, C. (1990). Cladistic analysis of the subfamilies within the Tubificidae (Oligochaeta). Zoologica Scripta, 19, 57– 63. Erséus, C. & Ferraguti, M. (1995). The use of spermatozoal ultrastructure in phylogenetic studies of Tubificidae. In B. G. M. Jamieson, J. Ausio & J. L. Justine (Eds) Advances in Spermatozoal Phylogeny and Taxonomy (pp. 189–201). Paris: Mémoires du Muséum National d’Histoire Naturelle. Erséus, C., Källersjö, M., Ekman, M. & Hovmöller, R. (2002). 18S rDNA phylogeny of Tubificidae and its constituent taxa (Clitellata): dismissal of the Naididae. Molecular Phylogenetics and Evolution, 22, 414–422. Erséus, C., Prestegaard, T. & Källersjö, M. (2000). Phylogenetic analysis of Tubificidae (Annelida, Clitellata) based on 18S rDNA sequences. Molecular Phylogenetics and Evolution, 15, 381– 389. Farris, J. S., Källersjö, M., Kluge, A. G. & Bult, C. (1995). Testing significance of incongruence. Cladistics, 10, 315 –319. Felsenstein, J. (1985). Confidence limits on phylogenies: an approach using the bootstrap. Evolution, 39, 783 –791. Ferraguti, M. (1984a). Slanted centriole and transient anchoring apparatus during the spermiogenesis of an oligochaete (Annelida). Biologie Cellulaire, 52, 175 –180. Ferraguti, M. (1984b). The comparative ultrastructure of sperm flagella central sheath in Clitellata reveals a new autapomorphy of the group. Zoologica Scripta, 13, 201 – 207. Ferraguti, M. (1998). The double spermatogenesis in Annelida and Pogonophora. In K. W. Wolf (Ed) Double Spermatogenesis. The Theory and a Collection of Experiments. Reading: Harwood Academic Publishers. Zoologica Scripta, 32, 3, May 2003, pp255– 278 • © The Norwegian Academy of Science and Letters R. Marotta et al. • Phylogeny of Tubificinae and Limnodriloidinae Ferraguti, M. (2000). Euclitellata. In K. G. Adyiodi & R. G. Adyiodi (Eds) Reproductive Biology of Invertebrates, Vol. IX. Part B: Progress in Male Gamete Ultrastructure and Phylogeny (pp. 165–222). Chichester: John Wiley. Ferraguti, M. & Erséus, C. (1999). Sperm types and their use for a phylogenetic analysis of aquatic clitellates. Hydrobiologia, 402, 225–237. Ferraguti, M., Grassi, G. & Erséus, C. (1989). Different models of tubificid spermatozeugmata. Hydrobiologia, 180, 73 – 82. Ferraguti, M., Marotta, R. & Martin, P. (2002). The double sperm line in Isochaetides (Annelida, Clitellata, Tubificidae). Tissue and Cell, 34, 305–314. Ferraguti, M. & Ruprecht, D. (1992). The double sperm line in the tubificid Clitellio arenarius. Bollettino di Zoologia, 59, 349 –362. Ferraguti, M., Ruprecht, D., Erséus, C. & Giere, O. (1994). An ultrastructural overview of tubificid spermatozoa. Hydrobiologia, 278, 165–178. Franzén, Å. (1973). The spermatozoon of Siboglinum (Pogonophora). Acta Zoologica, 54, 179 –192. Giusti, F. & Selmi, M. G. (1982). The atypical sperm in prosobranch molluscs. Malacologia, 22, 171–181. Gustavsson, L. (2002). The development of the genital ducts in Limnodriloidinae (Tubificidae, Clitellata). Journal of Zoology, 257, 83– 91. Gustavsson, L. & Erséus, C. (1997). Morphogenesis of the genital ducts and spermathecae of Clitellio arenarius, Heterochaeta costata, Tubificoides benedii (Tubificidae) and Stylaria lacustris (Naididae) (Annelida, Oligochaeta). Acta Zoologica, 78, 9 – 31. Gustavsson, L. & Erséus, C. (1999). Morphology and phylogenetic implications of oesophageal modifications in the Limnodriloidinae (Oligochaeta, Tubificidae). Journal of Zoology, 248, 467– 482. Healy, J. M. & Jamieson, B. G. M. (1981). An ultrastructural examination of developing and mature paraspermatozoa in Pyrazus ebeninus (Mollusca, Gastropoda, Potamididae). Zoomorphology, 112, 101–119. Jamieson, B. G. M. (1978). A comparison of spermiogenesis and spermatozoal ultrastructure in megascolecid and lumbriculid earthworms (Oligochaeta: Annelida). Australian Journal of Zoology, 26, 225–240. © The Norwegian Academy of Science and Letters • Zoologica Scripta, 32, 3, May 2003, pp255 – 278 Jamieson, B. G. M. (1981). The Ultrastructure of the Oligochaeta. London: Academic Press. Jamieson, B. G. M. (1987). A biological classification of sperm types, with special references to annelids and molluscs, and an example of spermiocladistics. In H. Mohri (Ed.) New Horizons in Sperm Cell Research (pp. 311–332). New York: Japan Scientific Society Press, Gordon and Breach Scientific Publications. Jamieson, B. G. M., Erséus, C. & Ferraguti, M. (1987). Parsimony analysis of the phylogeny of some Oligochaeta (Annelida) using spematozoal ultrastructure. Cladistics, 3, 145–155. Kluge, A. G. (1989). A concern for evidence and a phylogenetic hypothesis of relationships among epicrates (Boidae, Serpentes). Systematic Zoology, 38, 7– 25. Mickevich, M. F. & Farris, J. S. (1981). The implications of congruence in Menidia. Systematic Zoology, 30, 351– 370. von Nordheim, H. (1989). Vergleichende Ultrastrukturuntersuchungen der Eu- und Paraspermien von 13 Protodrilus-Arten (Polychaeta, Annelida) und ihre taxonomische und phylogenetische Bedeutung. Helgoländer Meersuntersuchungen, 43, 113 –156. de Pinna, M. C. C. (1991). Concepts and tests of homology in the cladistic paradigm. Cladistics, 7, 367– 394. Presgraves, D. C., Baker, R. H. & Wilkinson, G. S. (1999). Coevolution of sperm and female reproductive tract morphology in stalk-eyed flies. Proceedings of the Royal Society of London, 266, 1041– 1047. Righi, G. & Kanner, E. (1979). Marine Oligochaeta (Tubificidae and Enchytraeidae) from the Caribbean Sea. Studies of the Fauna of Curacao and Other Caribbean Islands, 58, 44–68. Swofford, D. L. (1998). PAUP: Phylogenetic Analysis using Parsimony, Version 4. Sunderland, MA: Sinauer Associates. Wang, H. & Erséus, C. (2003). New species of Doliodrilus and other Limnodriloidinae (Oligochaeta, Tubificidae) from Hainan and other parts of the northwest Pacific Ocean. Journal of Natural History (in press). Westheide, W., Purschke, G. & Middendorf, K. (1991). Spermatozoal ultrastructure of the taxon Enchytraeus (Annelida, Oligochaeta) and its significance for species discrimination and identification. Zeitschrift fur Systematik und Evolutionforschung, 29, 323–342. 271 Phylogeny of Tubificinae and Limnodriloidinae • R. Marotta et al. Appendix 1 The 29 sperm characters and character states (Fig. 7). Fig. 7 The 29 sperm characters and character states. 272 Zoologica Scripta, 32, 3, May 2003, pp255– 278 • © The Norwegian Academy of Science and Letters R. Marotta et al. • Phylogeny of Tubificinae and Limnodriloidinae Fig. 7 Continued. © The Norwegian Academy of Science and Letters • Zoologica Scripta, 32, 3, May 2003, pp255 – 278 273 Phylogeny of Tubificinae and Limnodriloidinae • R. Marotta et al. Fig. 7 Continued. 274 Zoologica Scripta, 32, 3, May 2003, pp255– 278 • © The Norwegian Academy of Science and Letters R. Marotta et al. • Phylogeny of Tubificinae and Limnodriloidinae Fig. 7 Continued. © The Norwegian Academy of Science and Letters • Zoologica Scripta, 32, 3, May 2003, pp255 – 278 275 Phylogeny of Tubificinae and Limnodriloidinae • R. Marotta et al. Appendix 2 The 29 somatic characters and character states. 30 Teeth of bifid chaetae: upper tooth about as long as, or shorter than, lower tooth (0); upper tooth somewhat (up to 1.5 times) longer than lower tooth (1). 31 Penial chaetae, i.e. modified chaetae associated with male pores: absent or bifid (0); present, with single-pointed tips (1). 32 Modified spermathecal chaetae, i.e. slender and ectally grooved, single-pointed chaetae associated with spermathecal pores: absent (0); present, at least in some individuals (1). 33 Number and position of male pores: paired, and lateroventral or ventral (0); unpaired, midventral (1). In most cases, state 1 refers to a median bursa that receives two atrial ducts at its inner end. 34 Position of spermathecal pores within spermathecal segment (in Tubificidae, generally segment X): in anterior part of segment (0); in middle part of segment (1). 35 Peripheral position of spermathecal pores: midventral (0); lateroventral or lateral (1); dorsal (2). (This character may be treated as ordered.) 36 Number of spermathecae: two (0); one (1); none (2). 37 Coelomocytes: few or absent (0); numerous and conspicuous (1). 38 Barrel-shaped, dilated portion of oesophagus, with a reticulate blood plexus: absent, i.e. oesophagus is a simple unmodified tube, or modified in some other way (0); present, extending at least through most of segment IX (1). See Gustavsson & Erséus (1999). 39 Paired oesophageal diverticula: absent, i.e. oesophagus is a simple unmodified tube, or modified in some other way (0); present in segment IX (1). See Gustavsson & Erséus (1999). 40 Entrance of vas deferens into atrium: clearly subapical, on ventral surface of atrium (or on anterior surface when atrium is erect) (0); more or less on apex of (inner end of) atrium (1); clearly subapical, on dorsal surface of atrium (2). (This character may be treated as ordered.) 41 ‘Ectoprostate’, stalked prostate gland derived from ectodermal (inner) epithelium of atrium: absent (0); present (1). So far, an ectodermal origin of prostates is only known for members of the Tubificinae (see Gustavsson & Erséus 1997). 42 Granulation of inner epithelium in atrial duct: none or partial only (0); inner epithelium densely granulated throughout length of atrial duct (1). This character is applicable to Limnodriloidinae only; granulation of the atrial epithelium also occurs in some of the other taxa, but a priori statements about which parts of their atria correspond to the limnodriloidine ‘atrial ducts’ are difficult to make. Thus, for species outside the Limnodriloidinae, this character is coded as ‘?’. 43 ‘Mesoprostate’, prostate gland derived from peritoneal cells: absent, i.e. neither diffuse nor lobed gland present (0); diffuse gland present, i.e. a more or less continuous layer of prostatic cells covering most of atrial surface (1); each atrium bearing lobed, broadly attached, prostate gland, com276 municating with a distinct ‘prostatic pad’, a granulated body of mesodermal origin, embedded in ventral wall of atrial ampulla (2). Non-additive character. It has been demonstrated that the diffuse prostates of one naidid (Stylaria lacustris) and two rhyacodriline tubificids (including Monopylephorus rubroniveus) are mesodermal, derived from peritoneal cells (Gustavsson & Erséus 1997, 1999). As a deduction from these studies, a mesodermal origin of the similarly diffuse prostates of Monopylephorus limosus and Rhizodrilus russus is assumed here. Gustavsson (2002) recently showed that the prostate glands and prostatic pads of Limnodriloidinae are also derived from peritoneal cells. State 2 of this character is fully congruent with the differentiation of the atrium into an ental ampulla and an ectal duct, a feature shared by all members of the Limnodriloidinae (Erséus 1982: 267). 44 Size/shape of prostatic pad: small and round to oval, and unless atrial ampulla is short (see state 0 of character 49), restricted to only a part of ampulla (0); elongated, and extending through a greater part of ampulla (1). For taxa lacking prostatic pads, this character is coded as ‘?’. 45 Position of prostatic pad in atrial wall: pad somewhat external on ampulla, i.e. it forces outside of atrial wall to bulge considerably (0); pad more confined to wall of atrial ampulla, i.e. it neither forces wall to bulge considerably nor extends markedly into the interior of ampulla (1); pad extending inwards, markedly reducing lumen of atrial ampulla (2). (This character may be treated as ordered.) For taxa lacking prostatic pads, this character is coded as ‘?’. 46 Atrial ciliation: lining epithelium of atrium with few or no cilia (0); lining of ental part of atrial ampulla bearing cilia (1); atrium conspicuously ciliated throughout (2). 47 Dilatation of inner end of atrium: absent (0); present, i.e. inner end of atrium wider, and also more thin walled, than other parts of atrium (1). 48 Curvature of inner end of atrium: absent, i.e. apical end of atrium not markedly curved (0); present, i.e. innermost end of atrium abruptly curved towards ventral side of worm (1). 49 Length and shape of atrial ampulla: short, more or less spherical (0); of medium length, oval to cylindrical, and without clearly prolonged ental part (1); long, cylindrical, with ental end markedly prolonged in relation to extension of prostatic pad (2). (This character may be treated as ordered.) For taxa without differentiated atrial ampullae (and prostatic pads), this character is coded as ‘?’. 50 Coiling of ectal end of atrium: ectal end of atrium short or moderately elongated, but not forming coils (0); much prolonged and coiled (1). 51 Unpaired ectal end of atria: absent, atria paired throughout (0); ectal ends of atria merged into an unpaired duct leading to (unpaired) male pore (1). 52 Atrial duct sac, i.e. a mesodermal sac enclosing ectal part of atrium, making latter act as an eversible pseudopenis: Zoologica Scripta, 32, 3, May 2003, pp255– 278 • © The Norwegian Academy of Science and Letters R. Marotta et al. • Phylogeny of Tubificinae and Limnodriloidinae absent (0); present, but thin and simple, not markedly muscular (1); well developed and strongly muscular, with great complexity, also involving an invagination of the body wall to take part in the pseudopenis (2). (This character may be treated as ordered.) 53 Copulatory organ associated with an ectodermal invagination at male pore: absent, i.e. atrium opening directly to exterior on body wall (0); present as a more or less developed invagination (copulatory sac), into which atrium opens, but no pendent penis developed (1); present as a pendent penis located centrally inside a well-developed copulatory (penial) sac (2). State 1 here may appear to overlap with state 2 of character 52, but the latter feature is interpreted as a derived condition of the pseudopenis associated with the atrial duct sac, i.e. not as homologous with the copulatory organs of character 53. 54 Pseudopenial papilla, i.e. distinct papilla formed in lateral wall of copulatory sac and bearing opening of atrium: absent © The Norwegian Academy of Science and Letters • Zoologica Scripta, 32, 3, May 2003, pp255 – 278 (0); present (1). This character is only applicable for taxa with state 1 of character 53 (see above). For all other taxa, this character is coded as ‘?’. 55 Folds in wall of copulatory sac: absent (0); present (1). For taxa without copulatory sacs (i.e. without states 1 or 2 of character 35), this character is coded as ‘?’. 56 Copulatory sac musculature: indistinct (0); well developed, present as distinct layer enclosing copulatory sac (1). For taxa without copulatory sacs (i.e. without states 1 or 2 of character 53), this character is coded as ‘?’. 57 Spermathecal duct: distinct, more or less cylindrical and well set off from spermathecal ampulla (0); not distinct, i.e. short and not well set off from ampulla, and if distinguishable, generally triangular in shape (1). 58 ‘Exterior’ spermathecal vestibule, i.e. a secondary invagination at spermathecal pore, forming a distinct chamber immediately outside spermathecal duct proper, and typically this chamber is wider than the duct proper: absent (0); present (1). 277 Data matrix. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 Zoologica Scripta, 32, 3, May 2003, pp255– 278 • © The Norwegian Academy of Science and Letters R. russus M. limosus P. molestus C. rugosus T. tubifex C. arenarius I. arenarius Doliodrilus sp. T. bruneti T. gurwitschi T. ineri L. armatus L. uniampullatus L. tenuiductus L. monothecus L. barnardi S. sacculatus S. hummelincki S. luteolus S. westoni 0 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 2 2 2 2 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 ? 1 1 1 1 0 0 0 0 0 1 0 1 0 0 0 0 0 0 ? ? 1 1 1 0 1 1 ? 1 1 0 0 0 2 2 2 2 2 2 ? ? 0 0 0 2 0 0 ? ? 0 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 3 3 3 2 2 3 1 1 0 0 2 1 1 3 3 3 3 2 3 0 1 1 1 1 1 1 1 1 1 1 1 0 0 1 1 0 1 1 1 0 1 1 1 0 1 1 1 0 0 0 1 0 0 0 1 0 1 1 1 0 1 0 0 0 0 0 0 0 0 0 1 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 1 1 1 2 1 1 ? 1 1 2 1 1 1 1 1 2 2 2 1 1 1 1 0 0 0 0 ? 0 0 0 0 0 1 1 0 1 1 1 ? ? ? ? 0 0 0 ? ? ? ? 1 1 1 1 1 1 1 1 1 ? ? ? ? 1 1 1 ? ? ? ? 0 0 0 0 0 0 0 0 0 ? ? ? ? 1 1 1 ? ? ? ? 1 0 0 1 1 1 1 1 1 ? ? ? ? 0 0 0 ? ? ? ? 0 0 0 0 0 1 1 1 0 ? ? ? ? 2 0 2 ? ? ? ? 0 0 0 0 0 0 1 1 1 ? ? ? ? 1 1 1 ? ? ? ? 1 0 0 0 0 0 0 0 1 ? ? ? ? 0 0 0 ? ? ? ? 0 0 0 0 0 0 1 0 1 ? ? ? ? 1 1 0 ? ? ? ? 0 0 0 0 0 0 0 0 0 ? ? ? ? 0 1 1 ? ? ? ? 0 0 0 0 0 1 1 0 1 ? ? ? ? 0 ? ? ? ? ? ? 1 0 0 ? 0 0 1 0 1 ? ? ? ? 1 1 1 ? ? ? ? 1 0 0 2 2 0 2 2 1 ? ? ? ? 1 0 0 ? ? ? ? 0 1 1 1 1 0 0 1 0 ? ? ? ? 1 1 1 ? ? ? ? 0 0 0 ? 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 3 3 3 1 0 0 1 3 3 3 3 3 2 2 2 ? ? ? ? ? 1 0 1 ? ? ? ? 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 0 1 1 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 0 0 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 0 1 1 0 0 0 0 1 1 0 1 0 0 0 0 1 1 ? 0 ? ? 0 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 ? 1 ? ? 1 0 2 2 2 1 0 0 1 0 0 0 0 0 0 0 0 0 2 2 0 0 1 0 1 0 1 1 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 1 1 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 1 1 1 1 1 1 0 ? 1 1 1 1 1 1 1 1 1 1 2 2 2 2 1 1 1 1 0 0 ? 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 ? ? ? ? ? ? ? 0 0 1 1 0 0 0 0 0 0 0 1 0 1 1 0 0 0 0 0 2 2 2 2 2 2 2 2 2 2 2 2 2 ? ? ? ? ? ? ? 0 0 0 1 1 0 0 0 0 0 0 0 0 ? ? ? ? ? ? ? 0 1 1 0 2 1 1 1 0 1 0 1 0 0 ? 2 1 0 2 ? 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 1 1 1 ? ? 0 ? 0 0 ? 1 ? 0 0 0 ? 0 0 0 0 1 1 1 ? ? ? ? ? ? ? 2 1 1 1 1 1 1 1 1 0 2 2 2 1 ? 0 0 0 0 ? 0 0 0 0 0 1 1 1 0 0 1 0 1 0 ? 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 ? 0 2 0 ? 0 1 2 2 0 ? 0 0 0 0 0 0 0 1 1 1 0 ? 2 2 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 ? 0 ? ? ? ? ? ? 0 1 1 1 1 0 0 0 0 1 1 1 ? 0 0 ? ? ? ? ? 0 0 0 0 0 0 1 1 1 1 1 1 ? 0 0 ? ? ? ? ? 1 0 0 0 0 1 1 0 1 0 0 0 0 0 0 ? 0 ? ? 0 0 1 1 1 0 0 0 0 0 1 1 1 0 0 0 ? 0 ? ? 1 1 0 0 0 0 1 1 1 1 Phylogeny of Tubificinae and Limnodriloidinae • R. Marotta et al. 278 Appendix 3