Malacological Review, 1988, Suppl. 4: 205-240 PHYLOGENETIC RELATIONSHIPS IN THE GASTROPOD FAMILY ARCHITECTONICIDAE, WITH NOTES ON THE FAMILY MATHILDIDAE (ALLOGASTROPODA) Rudiger Bieler •. ABSTRACT A cladistic analysis for 12 genus-group taxa of the Architectonicidae (Architectonica, Philippia, Psilaxis, Basisulcata, Discotectonica, Granosolarium, Solatisonax, Pseudotorinia, Pseudomalaxis, Spirolaxis, Heliacus, and an unnamed new taxon), including most of the Recent and fossil species, was conducted using the computer program packages PAUP and PHYSYS. The family Architectonicidae is well-defined by a suite of synapomorphies (e.g., a crescent-shaped osphradium, the lack of lateral radular teeth, a projection on the inner side of the operculum). Possible outgroups in the 'Heterogastropoda' and Allogastropoda are discussed, and the data available for a species currently grouped in the Mathildidae, Gegania valkyrie, were used for outgroup comparison. Architectonicidae and Mathildidae, forming the superfamily Architectonicoidea, share synapomorphies in anatomy (e.g., two juxtaposed ciliary tracts on the left side of the mantle cavity; fused salivary glands; a longitudinal crest at the ventral surface of the mantle cavity, built up by the anterior arterial vessel, a so-called chordoid tissue and a posterior pedal gland). Radular and opercular characters thought to be synapomorphic for the Architectonicoidea by other authors are discussed and rejected: the prominent peg on the inner side of the architectonicid operculum is lacking in mathildids, and the "five-toothed-taenioglossate" radulae of mathildids and most architectonicids probably evolved independently by loss of the outer marginal or the lateral teeth, respectively. Twenty-six binary and three multistate taxonomic characters of anatomy, radula, operculum, protoconch and teleoconch are selected and discussed. Special emphasis is placed upon sculptural features of the teleoconch for which homology can be inferred from multiple lines of evidence. Additional autapomorphies are listed for architectonicid genera. The analysis resulted in a resolved cladogram that showed three well-supported clades (Architectonica + new taxon + Philippia + Psilaxis + Basisulcata; Discotectonica + Granosolarium + Solatisonax; Pseudotorinia + Pseudomalaxis + Spirolaxis). The position of Heliacus in the cladogram is weakly supported; this genus is mainly defined by unique derived characters and retained symplesiomorphies. Three architectonicid genera known only from fossils (Stellaxis, Nipteraxis, Climacopoma) can be added to the tree on the basis of shell characters alone. The branching sequence of the cladogram supports the hypothesis that in Architectonicidae smooth-shelled groups with oval-shaped opercula and "ptenoglossate-like" radulae evolved later than sculptured forms with circular opercula and "taenioglossate-like" radulae. Thus the "ptenoglossate-like" radula of Architectonica must be interpreted as convergent when compared to mesogastropod families such as Epitoniidae and Janthinidae. Seven characters that display homoplasy are discussed and hypotheses explaining parallelisms are offered. The branching pattern of the cladogram is supported by fossil and ontogenetic evidence. The use of the generic names Architectonica for the subgenera Architectonica s.s. + new taxon, Philippia for Philippia s.s. + Psilaxis, and Pseudomalaxis for Pseudomalaxis s.s. + Spirolaxis is possible because each represent monophyletic groups. Earlier subdivisions into subfamilies are shown not to be based on monophyly and are rejected. INTRODUCTION For more than a century, the systematic position of the genera now comprising the marine gastropod family Architectonicidae Gray, 1850 (= Solariidae Chenu, 1859) has been a matter of dispute. Features unusual for the mesogastropod prosobranchs, with which they were usually grouped, such as a hyperstrophic proto(205) Rudiger Bieler 206 conch, an acrembolic proboscis, a "pigmented mantle organ" and atypical radulae, led to various attempts to group the architectonicids with other proso- or opisthobranch families and to find a system to subdivide the family, which has a wide spectrum of shell, radular and opercular characters. The proposed systems differ greatly from each other, being introduced before sufficient taxonomic data · were available. A number of recent publications (Boss & Merrill, 1984; Haszprunar, 1985b,d,e; Melone & Taviani, 1985; Bieler, 1984a, 1985a,b, 1987) filled many gaps in our knowledge of the group, but have not yet resulted in a satisfactory classification above the generic level. It is the scope of this paper: - to define taxonomic characters based on homology in the family Architectonicidae; - to discuss possible outgroups, and thus the relationships with other families; - to critically evaluate some of the presumed shared character states that have been used to link the Architectonicidae with the family Mathildidae; and - to present a phylogenetic analysis of the relationships within the Architectonicidae that allows the definition of monophyletic units and discrimination between symplesiomorphies (shared ancestral character states), synapomorphies (shared derived character states) and autapomorphies (derived character states of terminal taxa not shared with other groups) in order to facilitate further discussion. Phylogenetic analyses on a "lower" taxonomic level are still fairly scarce in malacology. As a method that has been constantly changing and improved upon over the past years, as philosophical discussions, mathematical algorithms and computer technofogy are developing, a wide variety of different techniques and approaches have been labelled "phylogenetic" or "cladistic" analyses. While most authors nowadays present their data in a testable format, the actual step from the data matrix to the cladogram presented is rarely documented and therefore often not reproducible. (This is also the case when an author actually cites the name of the computer "program" used. Most of the modem "phylogenetic analysis programs" available are extensive packages comprising a variety of algorithms, sometimes ranging from phenetic to cladistic approaches.) In this paper special emphasis is therefore placed on the methods employed. MATERIALS AND ME1HODS The taxonomic data for this study are based partly on previously unpublished personal observations, partly on my recent revisions of the family Architectonicidae (Bieler, 1984a,b, 1985a,b, 1987), and partly on data from other authors, as acknowledged in the text and tables. Sources of material in addition to those mentioned in the acknowledgements (and methods of measurement) are those outlined in my publications cited. The terminology of architectonicid shell sculpture is explained in Figure 1. While the names of architectonicid taxa are those as recently revised, all names of Mathildidae mentioned in this publication are those used by the authors cited in the text. Generic placement and specific identity need verification. [The family is currently under revision (Bieler, in prep.). Following a conservative approach, Mathildidae s.l. is used, .... Architectonicidae Phylogeny 207 including the genus Gegania Jeffreys, 1884.] "Hennig's method" of phylogenetic systematics (Hennig, 1950, 1966), a cladistic analysis that defines monophyletic groups based on internested sets of synapomorphies (shared derived characters), was used in this study. [For an in- ｕｍｒｾＡＺｚＮＲｺｭﾷ＠ SSR ..ｾﾷ＠ .. ｾﾷｵ＠ I ...ｾ＠ ..ｾﾷ＠ penultimate whorl UPR suture body whorl UPR LPR peripheral keel BF II base UC umbilicus FIG. 1. Diagrammatic representation of the pattern of spiral ribs on the Architectonica teleoconch, which can be homologized with spiral ribs in other genera [modified after Bayer (1940: fig. 1) and Bieler (1984b: fig. 1)). Numbers refer to Bayer's (1940) system. Abbreviations in brackets are those used in earlier publications in the German language: SSR = subsutural rib [ssR]; UMR =upper mid-rib [oMFRJ; I.MR= lower mid-rib [uMFR]; MR= mid-ribs (MFR]; UPR =upper peripheral rib [oRRJ; LPR =lower peripheral rib [uRR]; IPR = infraperipheral rib [ipRJ; PR = peripheral ribs [RR]; BF = basal field [BFI]; PUR = proxumbical rib [puR]; UC= umbilical crenae [er]. troduction to the recent discussion of various methodologies see, for instance, von Vaupel Klein (1984).] The number of taxa and characters involved did not allow the construction of cladograms "by hand" without the intuitive weighting of certain characters. Of the several computer programs available for cladistic analysis, two packages, PAUP (Swofford, 1985) and PHYSYS (Mickevich & Farris, 1984), were used. The PAUP program and the WAGNER.S option of the PHYSYS package are Rudiger Bieler 208 based on the "Wagner method" (Kluge & Farris, 1969; Farris, 1970), where the preferred tree is the one of minimal length in a Manhattan metric (which is a function of the range of character states of all taxa in the study), with no a priori restrictions on the nature of permissible character changes. Of the phylogenetic methods, that of Wagner shows greatest stability, and is less sensitive to the amount of homoplasious (parallel/ convergent) evolution than phenetic methods (Mickevich, 1978). Both programs have several major advantages over earlier programs (e.g., WAGNER78, WAGPROC) that also employed "Wagner methods": (1) they are able to handle missing (not applicable or unavailable) data; (2) they can treat "unordered" multistate characters [those for which the transformation series is not known prior to the analysis; PAUP uses an algorithm based on Fitch's (1971) optimization; PHYSYS uses Mickevich's (1982) "Transformation Series Analysis"]; and (3) they can detect the existence of multiple, equally parsimonious character-state reconstructions (incorporating the least number of character- state changes). PHYSYS can handle non-linear (branching) multistate characters, while PAUP requires recoding into several binary characters for this purpose. The data set used in this paper was small enough to allow the use of an exact algorithm (one that will find all equally parsimonious trees). The algorithm used for this purpose, the ''branch-and-bound-method" [as suggested by Hendy & Penny (1982), modified by Swofford (1985) as part of the PAUP package], represents the most efficient approach currently available (see Platnick, 1987). Of the several available rooting algorithms, outgroup-rooting was used: the taxon designated as the outgroup forms a sister group to the in-group taxa; the program "decides" polarities based on the same parsimony criterion being used in constructing the tree (in contrast to earlier methods, e.g., "hand cladistic" and WAGNER78, where polarity decisions were made in a separate step prior to tree construction). The terms monophyly and paraphyly are employed in their usual meaning, i.e., according to the concepts of Hennig (e.g., 1966: fig. 45) and as defined by Farris (1975). Further details are outlined in the sections "Selection of Characters" and "Tree Construction" below. The abbreviation "Fig." refers to figures in this work; "fig." refers to those in other publications. Cited repositories are: AMS LACM NMNZ SMF USNM Australian Museum, Sydney Los Angeles County Museum of Natural History National Museum of New Zealand, Wellington Senckenberg-Museum, Frankfurt/Main National Museum of Natural History, Smithsonian Institution, Washington. ANALYSIS Selection of in-group taxa The architectonicid genus-group taxa are as revised by Bieler (1984b, 1985a,b, 1987). For the purpose of this analysis subgenera as well as genera have been used as operational taxonomic units in the cases of Philippia/ Psilaxis and Pseudomalaxis / Spirolaxis, while Heliacus s.l. was used, comprising all its recognized ) !. Architectonicidae Phylogeny 209 subgenera (see below). Basisulcata, the type species of which was previously included in Philippia (Bieler, 1985a: 336), is here recognized as a genus-group taxon. Genera known only from fossils, e.g., Climacopoma Fischer, 1885 (= Patulaxis Dall, 1892); Dinaxis Dall in Aldrich, 1895; Eosolarium Chavan, 1947; Nipteraxis Cossmann, 1915, and Stellaxis Dall, 1892, were excluded from the analysis for lack of comparative data, but were subsequently assigned positions in the resulting cladogram. Twelve architectonicid genus-group taxa were used in the study, comprising almost all of the approximately 130 Recent species worldwide (Bieler, 1986): 1) Architectonica Roding, 1798 (= Solarium Lamarck, 1799; = Verticillus Jousseaume, 1888, non Moquin-Tandon, 1848). 2) new taxon (Bieler,1987) (see note at end of manuscript). 3) Philippia J. E. Gray, 1847. 4) Psilaxis Woodring, 1928. 5) Basisulcata Melone & Taviani, 1985. 6) Discotectonica Marwick, 1931 (= Acutitectonica Habe, 1961; = Russetia Garrard, 1961 ). 7) Granosolarium Sacco, 1892 (= Solariaxis Dall, 1892; = Claraxis Iredale, 1936). 8) Solatisonax Iredale, 1931. [The type species of the recently described taxa Heliacus (Granoheliacus) Melone & Taviani, 1985, and Heliacus (Redivivus) Melone & Taviani, 1985, appear to belong to Solatisonax, but need further study and are here excluded from the analysis]. 9) Pseudotorinia Sacco, 1892; (= Awarua Mestayer, 1930; = Calodisculus Rehder, 1935; ?= Punjabia Eames, 1952). 10) Pseudomalaxis Fischer, 1885 (= Discosolis Dall, 1892; = Mangonuia Mestayer, 1930). 11) Spirolaxis Monterosato, 1913 (= Paurodiscus Rehder, 1935; = Aguayodiscus Jaume & Borro, 1946). 12) Heliacus Orbigny in Sagra, 1842 (= Torinia J. E. Gray, 1842) [comprising the following subgenera: Heliacus s.s.; Torinista Iredale, 1936 (= Astronacus Woodring, 1959); Grandeliacus Iredale, 1957; Teretropoma Rochebrune, 1881; Gyriscus Tiberi, 1867; new taxon (Bieler, 1987)]. Systematic Position Of The Architectonicidae: "Heterogastropoda" Versus Allogastropoda The family Architectonicidae has received considerable attention in the past, focusing mainly on two problems of higher classification: (a) the position of this group in the Gastropoda, and (b) the composition of the group of related families around Architectonicidae. The family, classically placed in the mesogastropod superfamily Cerithioidea, has been variously raised to superfamilial level, either without (e.g., Korobkov, 1955: 136) or with (e.g., Taylor & Sohl, 1962: 14) the inclusion of other families, usually on the basis of shared radular features or overall shell similarity. Thiele (1925: 113) pointed out the similarities in heterostrophy, and in opercular and radular characters between Architectonicidae and Mathildidae, and suggested a phylogenetic relationship. Taylor & Sohl (1962: 14) added the Omalaxidae to this group, and Golikov & Starobogatov (1975: 223) included the Cyclostremellidae. However, the Omalaxidae, a caenogastropod family of uncertain position known only from the Eocene, should be 210 Rudiger Bieler excluded from this group (see Bieler, 1984b: 65), because it does not share any homologous characters that could indicate close relationship with the other families in question; Cyclostremella, type genus of Cyclostremellidae Moore, 1966, has been shown to belong to the Pyramidellidae (Robertson, 1973). The "special" position of the Architectonicidae in the Gastropoda was already pointed out by MacDonald (1860: 76), in his excellent but poorly known work on gastropod classification. He placed Architectonicidae (as Solariidae, including "Scalariil" = Epitoniidae in the modem sense) and Janthinidae close to the Pyramidellidae, and suggested that this group might "form a remarkable and beautiful link between the Opisthobranchiata and the unisexual Gastropoda." MacDonald thus outlined an early precursor of the "Heterogastropoda," a taxon formally introduced more than a century later by Habe & Kosuge (1966: 101, as an order) and Kosuge (1966: 320, as a suborder), for the families Architectonicidae, Epitoniidae, Janthinidae, Mathildidae and Triphoridae. Habe & Kosuge's taxon has been accepted by a number of authors, comprising some or all families originally included, as a subdivision of either the Proso- or Opisthobranchia (e.g., Shuto, 1969; Cemohorsky, 1972, 1978; Oimo, 1975; Powell, 1979; Boss, 1982, Fretter & Graham, 1982; Kilburn & Rippey, 1982; partly summarized in Haszprunar 1985a: tab.5). [For a discussion of prosobranch and opisthobranch traits in Architectonicidae, see Merrill (1970: 271) and the detailed account by Robertson (1974); for opisthobranch-like characters of shell-matrix and spermiogenesis, see Ghiselin et al. (1967: 13) and Healy (1982: 197), respectively.] Robertson (1985) compared some of the characters of Architectonicoidea, Epitonioidea and Pyramidellidae with those of other prosobranch and opisthobranch families. Of the four characters listed by him, that "support the ideas that, within themselves, the characters are homologous and reflect phylogeny" (1985: 1), only one, the "pigmented mantle organs" of the larvae and postlarvae, is shared by architectonicids with Epitoniidae and Janthinidae. The majority of the characters shared between Architectonicidae and Epitoniidae, as listed by Robertson (1974: 216), can be better explained by convergence due to similar diets [i.e., association with cnidarians: cuticularized esophagus, ptenoglossate-like radula (see below)] or as symplesiomorphy (streptoneury). The same is true for corresponding similarities between Architectonicidae and Janthinidae. The other characters studied by Robertson (1985) (i.e., the longitudinal ciliated strips dorsal and ventral in the mantle cavity of postlarvae, the chalazae connecting egg cocoons, and the hyperstrophically coiled larval shells) are restricted to three families, Architectonicidae, Mathildidae and Pyramidellidae, forming a group within the "assuredly polyphyletic" (Boss, 1982: 996) assemblage that has been called "Heterogastropoda." The "ciliated strips," listed as "probably absent" for Mathildidae by Robertson (1985: fig.2), do occur in that family (Haszprunar, 1985e: 202; pers. observ.). Studying the fine structure of osphradial epithelium, Haszprunar (1985a,b) arrived at an almost identical grouping: Triphoridae and Epitoniidae (as well as Eulimoidea and Cerithiopsoidea which had been included in the heterogastropods by some authors) were shown to have an osphradial fine-structure characteristic of higher Caenogastropoda, while Architectonicidae and Mathildidae displayed a different pattern. Because of the inherent polyphyly of the "Heterogastropoda," Haszprunar ｾﾷ＠ . Architectonicidae Phylogeny 211 (1985c: 25) introduced the new superorder Allogastropoda, to include three Recent families: Architectonicidae and Mathildidae in the Architectonicoidea, and Pyramidellidae in the Pyramidelloidea. This arrangement was based on anatomical characters thought to be synapomorphic, i.e., the existence of a secondary foliobranch gill to the right of the dorsal ciliary tract, the acrembolic proboscis, and the occurrence of spermatophores. He also included five fossil families: a family named "Streptaclidae" (Haszprunar, 1985c: 25) or "Streptaclididae" (1985c: 32) [err. pro Streptacididae Knight, 1931], which he placed in the Pyramidelloidea (but which he later excluded; Haszprunar, 1985e: 211), and the extinct Nerineoidea, comprising four families [Nerineidae, Nerinellidae, Ceritellidae and "ltieridae" (err. pro ltieriidae Cossmann, 1896)]. Of the several names available for higher taxa comprising some or all of these families (see Ponder & Waren, 1988, this volume), the name Allogastropoda Haszprunar, 1985, is here used to describe the clade that is formed by at least the three extant families mentioned, namely Architectonicidae, Mathildidae and Pyramidellidae (see also Bieler, 1985b). [Haszprunar (1988, this volume) changed his concept of Allogastropoda drastically; he now views this group as an "orthophyletic grade," comprising five Recent superfamilies.] .• Selection Of Outgroup The selection of an outgroup for the cladistic analysis of architectonicid genera is problematic, because the choice of closely related groups with comparable characters is severely limited. Because this analysis involves a considerable number of non-shell characters, only a family with Recent representatives is appropriate, as opposed to extinct groups for which no direct anatomical information can be collected. [Haszprunar (1985c), however, inferred for the fossil Nerineoidea a special secondary gill (1985c: 15), an acrembolic proboscis (1985c: 30) and atypical sperm (1985c: 24); the last probably an error for Neritoidea, judging from the reference cited (Giusti & Selmi, 1982).] Of the three Recent families in the Allogastropoda, the Pyramidellidae are often smooth-shelled, highly specialized ectoparasites which lack a radula (Fretter & Graham, 1949; Fretter et al., 1986), and thus offer few characters for comparison with the shell and radular data available for Architectonicidae. Similar problems arise for the Omalogyridae, a family of incertae sedis that has been tentatively grouped close to the Architectonicoidea/Pyramidelloidea complex (Haszprunar, 1985c: 25; Salvini-Plawen & Haszprunar, 1987: 760). The minute omalogyrids lack many of the organs and structures that are developed in architectonicids (Fretter, 1948). Thus, the only closely related family that displays a sufficient number of comparable characters and could qualify as an outgroup on the basis of close phylogenetic relationship appears to be the Mathildidae. This small, deep-water family had already been grouped with Architectonicidae by Thiele (1921: 143, 1925: 113), based on similarities in radular features and the common occurrence of heterostrophy. Additional anatomical evidence for a phylogenetic relationship was published by Climo (1975) and Haszprunar (1985e), who described a number of synapomorphies in anatomy (e.g., a longitudinal crest at the ventral surface of the mantle cavity, built up by the anterior arterial vessel, a so-called chordoid tissue and a posterior pedal gland; fused salivary glands; two juxtaposed ciliary 212 Rudiger Bieler tracts on the left side of the mantle cavity). However, two of the features that were used in the past to combine the two families need re-evaluation: the presumed shared characters in opercula and radulae. Opercula: All architectonicid opercula are characterized by a peg-shaped internal projection (Fig. 2), where the muscle is attached. A close similarity of a rchitectonicid and mathildid opercula has been stressed in the literature (e.g., Thiele, 1925: 113, 1928: 88; Merrill, 1970: 274; Hasz prunar, 1985d: 41), but the existence of this structure has not been d emonstrated for any mathildid species [it is, however, present in other families, e.g., Triphoridae (Marshall, 1983: 4, fig. ld)] . The opercula of nine mathildid species were examined (species-level systematics under revision; Bieler, in prep.), and they all lack an internal projection (for example, see Fig. 3). The common possession of a corneous multispiral operculum hardly qualifies as a synapomorphy. Radulae: Only a few data on mathildid radulae are available. Climo (1975: fig.3) illustrated (see Fig. 4) the radula of Gegania valkyrie Powell, 1971. This radula has a pair of reduced outer marginal teeth (Fig. 4d) which form "an anchor to which the bases of inner marginal and lateral teeth abut" (Climo, 1975: 277). FIGS. 2-3. Opercula of Heliacus and Mathilda. 2. Oblique side view of operculum of Heliacus tlariegatus (Gmelin, 1791), showing prominent peg on body side (SMF 256388). 3. Oblique side view (body side) of opcrculum of Mathilda cf. quadricarinata (13rocchi, 1814) [USNM 131944). The tooth arrangement and cusp configura tion (Fig. 4a-c) implies derivation from a mesogastropod-like taenioglossate form. The radulae of Mathilda quinquelirata Kuroda, 1958 (see Fig. 5), and Eucharilda sinensis (Fischer, 1867) (see Fig. 6) as illustrated by Habe (1958: 56, pis. 2-3) likewise have a smaller latera l tooth and a single, larger, differently-cusped marginal tooth per half row (Fig. 5b-c, 6b-c). Outer marginals h ave not been found. Habe's drawing of the E. sinensis rachidian (Fig. 6a) contradicts his text (1958: 56) which states tha t the rachidian "has nine cusps normally," falling within the range of the o ther species studied to date. Thiele's drawing (1928: fig.10; see Fig. 7a-c) of the radula of Mathilda elegantissima (0.G. Costa, 1861) reveals the same five-teeth-per-row configuration. Climo (1975: 280-281) misinterpreted Thiele's figure as being typically seven-too thed taenioglossate, because he understood the lateral view of the rachidian (Fig. 7a') as the lateral tooth, thereby counting seven teeth per row. Thiele's text (1928: 88) ("die Mittelplatte ... und die beiden Seitenplat ten") a nd fi g ure caption (1928: 87) ("Mittelplatte von oben und von der Seite gesehen"), however, indicate clearly that Thiele observed only five teeth. Merrill (1970: pl. 8 fig.Id) again found five Architectonicidae Phylogeny rachidian 4 ｾ｡＠ 213 lateral inner marginal outer marginal c 5 c - 6 c - c - c - .• 8 FIGS. 4-8. Mathildid radulae. All figures redrawn and newly arranged (original scale kept within each figure). The letter "a" indicates additional view of rachidian: 4. Gegania valkyrie Powell, 1971 [after Climo, 1975: fig.3]; 5. Mathilda quinquelirata Kuroda, 1958 [after Habe, 1958: pl. 3 fig. 14); 6. Eucharilda sinensis (Fischer, 1867) [after Habe, 1958: pl. 2 fig. 11); 7. Mathilda elegantissima (0.G. Costa, 1861) [after Thiele, 1928: fig. 10); 8. Mathilda quadricarinata (Brocchi, 1814) [after Merrill, 1970: pl. 8 fig. ld). 214 Rudiger Bieler teeth per row for Mathilda quadricarinata (Brocchi, 1S14) (see Fig. S). In his drawing he placed the longest tooth (Fig. Sc) closer to the rachidian than the tooth here reproduced in Fig. Sb. A comparison with the laterals and marginals of the other studied species suggest that this is an artifact. Ponder (19S5: fig. 145H) illustrated the radula of Brookesena neozelanica (Suter, 190S), a species provisionally included in the Mathildidae. The family Mathildidae is rather homogeneous in shell characters: all members have a smooth hyperstrophic protoconch, except for Brookesena Finlay, 1927 (based on Mathilda neozelanica Suter, 190S), which has a sculptured homeostrophic protoconch (Ponder, 19S5: fig. 145F). A similar protoconch (and teleoconch) can be found in members of Turritellopsis G.O. Sars, 1S7S. However, Turritellopsis acicula (Stimpson, 1S51), type species of that genus, is characterized by a different radular type, with only three teeth per row (Sars, 1S7S: pl. 7 fig. 2), and should not be included in the Mathildidae. Ponder's (19S5: fig. 145H) SEM photomicrograph of Brookesena neozelanica shows a five-toothed radula similar to those in Figures 5-S, the strong lateral having five finger-shaped cusps and a wide, asymmetrical base forming an outer tongue. An SEM photomicrograph of the radula of Mathilda cf. quinquelirata Kuroda, 1958 (see Fig. 9) shows a gap between the rachidian and laterals and a very strong central cusp on the rachidian. The radula of a specimen identified as Mathilda decorata Hedley, 1903 (by comparison with the holotype, AMS C.16299; type species of Opimilda Iredale, 1929), has a rachidian with numerous (w24) long, filiform cusps of subequal strength, the central ones being longer (Figs. 10-11). As in the other species studied, the inner marginals are longer and bear more denticles than the laterals (Fig. 11). No outer marginals could be traced. Rachidian and laterals are separated by a wide space as in M. cf. quinquelirata. A character not seen in other mathildid species studied is a feather-like extension on the marginals ("F' in Fig. 11), projecting at an oblique angle from shortly below the tip of the "normal" tooth. This structure is highly flexible, but breaks off easily during preparation and is not readily visible in light microscopy under lower magnification. The eight mathildid species (spanning several nominal genera) for which the radulae are known, display much less variation in number and relative size of teeth than previously assumed. The mathildid radula is modified taenioglossate, with the rachidian separated by a gap from a pair of laterals, which are longer than the rachidian and bear a moderate number of cusps, and a pair of multi-cusped or serrated inner marginals, which are longer than the laterals; the outer pair of marginals is either greatly reduced or lost entirely. This "five-toothed" radula of mathildids has been homologized with the five-toothed radula in a number of architectonicid genera such as Heliacus, Pseudotorinia and Philippia. However, the taenioglossate-like radula of Architectonicidae is of a different structure and, probably, origin. The two outer pairs of teeth are of equal length and structure (Fig. 12). No evidence of an additional outer pair of teeth (as in Gegania) was found. The two outer pairs present closely resemble, and are here homologized with, the inner and outer marginals in a "typical" seven-toothed taenioglossate radula [see, for instance, the radula of Cerithidea costata (da Costa, 1778) in Bandel, 1984: 37, pl. 3 fig. 6)). If this fivetoothed radula was derived as in the Mathildidae, by loss of the outer ｾＮ＠ Architectonicidae Phylogeny 215 FIGS. 9-1 4. 9. Radula of Math ilda cf. quinquelirata Kuroda, 1958 [AMS C.146527] (SEM courtesy of Dr. Anders Wa ren). 10. Rachidian and lateral teeth of the radula of Mathilda decorata Hedley, 1903 [AMS C.14868) (SEM). 11. Lateral teeth and inner marginals of the radula of Mathilda decorata Hedley, 1903 [AMS C.14868 ). "F" indica tes feather-like extensions on inner marginal teeth (SEM). 12. Radula of T-leliacus variegatus (Gmelin, 1791) [SMF 256388) (SEM). 13. Radula of Arcliitectonica picta (Philippi, 1849) [LACM 78-65) (SEM). 14. Protoconch of Gegania valkyrie Powell, 1971 [NMNZ M.36712] (SEM). marginals, two evolutionary s teps would be n ecessary: (a ) the reduction of the outermost pair of teeth, and (b) the transformation of the lateral teeth into virtual duplica tes of the inner marginals. It is therefore proposed that the fivetoothed taenioglossa te-like rad ulae of Mathildidae and Architectonicidae evolved independently, probably in parallel from a seven-toothed taenioglos- Rudiger Bieler 216 sate ancestor. While in the Mathildidae the outer marginals were reduced, the Architectonicidae lost the lateral teeth (Fig. 15). "Five-teeth-per-row" is not a synapomorphy of the Architectonicoidea, but rather an autapomorphy of each of the two families. Data other than shell characters are available for only a few mathildid species. The anatomy of only two species, Opimilda maoria Powell, 1940, and Gegania valkyrie Powell, 1971, have been studied in detail (Climo, 1975; Haszprunar, 1985e; pers. observ.). Within the Mathildidae, G. valkyrie displays a - Mathilda rachldian Heliacus lateral inner marginal outer marginal - FIG. 15. Diagrammatic figure showing hypothetical independent origin of "5-toothed taenioglossate-like" radulae in Mathildidae and Architectonicidae. number of ancestral character states in shell, radular and anatomical (especially buccal system) characters and is therefore here used as the outgroup, rather than 0. maoria, which displays many derived character states within mathildids (Bieler, in prep). Selection of characters Characters useful for phylogenetic analyses have to meet certain criteria: they must be based on features thought to be homologous, they must display at least two different character states, and the character states must be known for at Architectonicidae Phylogeny 217 least three of the taxa involved. The character states used for terminal taxa in an analysis on the generic level could be (a) that of the type species, (b) true for some, but not necessarily all, species, or (c) true for all species in the genus. In this study, the last approach was used, with a few exceptions explained in the text. The inclusion of autapomorphies in cladistic analyses has meaning only as long as they provide additional information that can be used to infer branching sequence. This is the case when the autapomorphy is part of a multistate character. However, character states of binary characters that are unique to one of the taxa involved and have the opposite, plesiomorphic state in all others do not supply this information, should not be used, and were therefore eliminated from this analysis. However, these autapomorphies do serve to define terminal taxa and might become useful for a phylogenetic study when additional data become available, and are listed here: Architectonica: presence of a distinct color pattern on head; rachidian absent (some species); very large teleoconch size; color pattern of spiral bands on shell. Philippia: flanking cusps on rachidian as strong as central cusp; on protoconch, callus overlays anal keel and false umbilicus. Psilaxis: opercular peg mushroom-shaped. Discotectonica: umbilical crenae flattened. Granosolarium: umbilical crenae sunken into umbilicus; nodules on lower peripheral rib developed into spines (some species). Solatisonax: upper peripheral rib on early whorls prominent and undulated. Pseudotorinia: protoconch paucispiral (some species). Spirolaxis: shell with open coiling; strap-like jaws. Heliacus: diffuse black body pigment (some species). In the following, the 26 binary and three multistate characters and their states are listed. The code used in the analysis is added in parentheses [Characters 2 and 5 were inferred from Haszprunar (1985d,e), who did not explicitly define size classes]. It should be noted that at this point no statement is made regarding the direction of change in a series of character states (with one exception; see character 15, below); the polarity decision is made by outgroupcomparison during construction of the tree, i.e., during the program run (see "Materials & Methods"). The resulting data matrix is shown in Table 2 (Appendix). ,. Anatomical characters Alimentary tract: 1) buccal mass: developed (0), absent (1) 2) buccal mass: large (0), medium (1), small (2) 3) esophagus: simple (0), with blind sac containing "rod-like structure" (1) 4) esophagus: straight (0), looped (1) 5) proboscis sheath: short (0), long (1) Other anatomical characters: 6) osphradium organization: bilamellar (0), crescent shaped (1) The majority of the information available on architectonicid anatomy (i.e., Bouvier, 1886a,b, 1887; Risbec, 1955; Merrill, 1970; Robertson, 1974; Climo, 1975, 218 Rudiger Bieler Haszprunar, 1985b,d) is useful for comparison between families, but does not enable a comparative study within the Architectonicidae. Detailed information on the anatomy of the architectonicid reproductive system is restricted to a few species of Heliacus and Architectonica (Haszprunar 1985d: 33; pers. observ.). Variation within the taxa, at both the specific and generic levels, is not yet known, and the present data on reproductive strategies show amazing heterogeneity within Architectonicoidea [protandric and simultaneous hermaphroditism in species of Heliacus and Discotectonica (Haszprunar, 1985d: 33, 34, 38), simultaneous hermaphroditism in Gegania, gonochorism in species of Philippia, Granosolarium and Opimilda (Haszprunar, 1985e: 205, 211)]. This probably accounts for conflicting statements in the literature which state that the Architectonicidae are hermaphroditic (Robertson, 1974: 218; Gosliner, 1981: 207) or gonochoristic (Gosliner, 1981: 209; Boss, 1982: 997). As pointed out by Robertson (1985: 16), simultaneous hermaphroditism as a taxonomic character should be used with care in phylogenetic studies, especially because other gastropod families are known in which hermaphroditism and dioecism co-occur in a single genus (e.g., Patella; Webber, 1977: 11). Reproductive data are considered too fragmentary to be included here. Radular characters 7) radula: present (0), lacking (1) 8) no. of flanking cusps on rachidian: ｾ＠ (0), 1(1),0 (2) 9) central cusp on rachidian: strongest (0), as strong as others (1) 10) lateral tooth: present (0), absent (1) 11) no. of ｭ｡ｾｧｩｮｬｳ＠ per half row: 2 (Fig. 12) (0), '?:.7 (Fig. 13) (1) 12) no. of cusps on outer marginal: '?:.3 (0), 2 (1) The radular characters used in this study are derived from the data shown in Table 1 (Appendix). The apparent intraspecific variability reflected in the table is often caused by the different microscopic techniques applied; when possible, data based on SEM studies have been used here. A true radula is lacking in Discotectonica and Granosolarium (Haszprunar, 1985d,e). Instead, these genera have extremely long, toothed, rod-like structures (see Melone, 1975: pl.1 figs. 5-6; Melone & Taviani, 1985: figs. 4-10; Haszprunar, 1985d: fig. 2). The homology and function of this structure are unclear. The total number of transverse rows of radular teeth might be a useful taxonomic character: e.g., with 60-70 rows counted in Architectonica s.s. and only nine in a species of Solatisonax [Boss & Merrill, 1984: 358, 361; for A. nobilis Roding, 1798, and "Heliacus" borealis (Verrill & Smith in Verrill, 1881), respectively]; not enough data were available to use this character in the analysis. Opercular characters 13) opercular peg: absent (Fig. 3) (0), present (Fig. 2) (1) 14) operculum shape: circular (0), oval (1) Only two opercular characters were used. No distinction was made between circular flat and circular high-spired (cone-shaped) opercula, because this character varies greatly within genera [compare, e.g., the closely related Heliacus (Teretropoma) infundibuliformi$ (Gmelin, 1791) and Heliacus (Teretropoma) mighelsi (Philippi, 1853) (see Bieler, 1985b: 104)]. Although esta- Architectonicidae Phylogeny 219 blished in the architectonicid literature, grouping by paucispiral, "moderately multispiral," and multispiral opercula was omitted, because such splitting was judged to be artificial (all architectonicid opercula are originally multispiral and differ in later whorls, as expressed by character 14, circularI oval). The oval shape of mathildid opercula is seen here as convergence, based on a different "ground plan" (initial "open coiling" with peg formation in Architectonicidae, versus "normal growth" of a flat operculum in Mathildidae). To prevent the superficial resemblance of the mathildid operculum from influencing the analysis, this character was coded as "missing data" for the outgroup. Protoconch characters 15) protoconch size: small (0), medium (1), large (2) 16) protoconch coiling: heterostrophic (0), almost paucispiral (1) [Coded as (1) for Pseudomalaxis and Spirolaxis. Also true for some species of Discotectonica and Pseudotorinia (Bieler, 1985a: 241, 1985b: 92).) 17) anal keel: absent (0), developed (1) The protoconch size in Architectonicidae has been shown to be a good taxonomic character at the species level (Robertson, 1970; Bieler, 1984b). On the generic level, the taxa again display different size ranges (Fig. 16), and are here arbitrarily grouped into three size classes (character 15). With reference to heterostrophy, Haszprunar (1985e: 207) stated: "Gegania valkyrie shows only traces of this condition (see SEM photographs in Climo 1975)." Climo's photographs (1975: fig. 1) were taken at angles that do not emphasize the heterostrophic condition. The shell apex of the specimen for which Haszprunar (1985e) studied the anatomy clearly shows heterostrophy (Fig. 14), the axes of coiling being different in protoconch and teleoconch, and the early protoconch whorls pointing towards the base of the teleoconch. Teleoconch characters 18) basal keel, formed by a basal field rib (Fig. 18): absent (0), present (1) 19) basal keel forming lower peripheral shell keel: no (0), yes (1) 20) infraperipheral rib forming lower peripheral shell keel (Fig. 19): no (0), yes (1) 21) 3 narrow, well-defined outer basal field ribs (Fig. 20): absent (0), present (1) 22) 5 basal field ribs, expanding towards umbilicus: absent (0), present (1) 23) outer basal field: sculptured (0), smooth (1) 24) concave zones flanking peripheral keel (Fig. 21): absent (0), present : (1) 25) peripheral ribs: sculptured (0), smooth (1) 26) mid-rib area with many secondary ribs: no (0), yes (1) 27) areas of fused ribs smooth and glossy: no (0), yes (1) 28) umbilical crenae: prominent (0), reduced (1) 29) teleoconch coiling: orthostrophic (0), planispiral (1) Several teleoconch characters that have been used in the past to classify architectonicids have proven to be unreliable: The adult size can vary greatly within genera (especially in Heliacus, Granosolarium and Pseudotorinia). The general shell shape can be very misleading when not based on homologous struc- Rudiger Bieler 220 tures ["Heliacus (Claraxis)" sensu Garrard (1977: 554) for instance, mainly based on having "one sharp peripheral keel," included 7 Recent species from 4 genera: Granosolarium, Pseudotorinia, Solatisonax and Architectonica). However, shell shape based on homologous sculptural elements is a very good SMALL MEDIUM LARGE Architectonics new taxon Philippia Psilaxis Basisulcata Heliacus Discotecton lea Granosolarium Solatisonax Pseudotorinia Pseudomalaxis Spirolaxis 0.75 1.20mm FIG. 16. Protoconch size ranges of architectonicid genus-group taxa (chosen size classes used as character 15). taxonomic character in this family. Using the homology criteria of position, ontogenetic and fossil evidence, and special morphological qualities such as color pattern and sculpture, a number of main spiral ribs and fields on the teleoconch have been homologized within the Architectonicidae (Bieler, 1984b: 457, fig. 1, 1985b: 96, figs. 1-7; see Figs. 1 and 17). (In some smooth [Philippia) or highly Architectonicidae Phylogeny 221 granulated [Heliacus (Gyriscus)] groups this "ground pattern" is only traceable in the early whorls, and sometimes the fossil record reveals the placement of ribs "missing" in Recent species [e.g., the Pliocene Solarium bicinctum Cantraine, 1842, displays a weak upper peripheral rib, lacking in Recent species of Philippia s.s.; Bieler 1985a: 236)). SSR ｾｐｒ＠ UC UC SSR d SSR ｾＭＯ＠ UPR ｾ＠ LPR UC UC FIG. 17. Placement of homologous ribs in various architectonicid genera. Apertural view; arrow shows point of attachment of next whorl, intrasubgeneric variation indicated by dotted lines. a) Architectonica s.s., b) Philippia s.s., c) Heliacus s.s., d) Pseudomalaxis s.s. 19 FIGS. 18-21. Shape of body whorl and placement of homologous ribs in Architectonicidae. Filled circles, from top to bottom: UPR, LPR, IPR; open circles: additional ribs formed by BF. 18. Character 18; 19. Character 20; 20. Character 21; 21. Character 24. Other characters Fragmentary data on additional characters are available [such as known food requirements: Architectonica - Actiniaria (e.g., Bandel, 1984: 63), Psilaxis - Scleractinia (see Robertson et al., 1970: 60), Heliacus - Zoanthidea (e.g., Robertson, 1967: 246), Gegania - Antipatharia (see Climo, 1975: 278), but could not be utilized in this study. Rudiger Bieler 222 Tree Construction Coding of multistate characters: The multistate characters 2 and 8 (buccal mass size: large/medium/small; number of flanking cusps on rachidian: 2+ /1/0) were coded for different runs in the following way: hypothesis for transformation series (a) ordered (PAUP, PHYSYS): 0-1-2 or 2-1-0 0-1-2 or 0-2-1or2-1-0 or 2-0-1or1-0-2or1-2-0 (b) unordered linear (PAUP): (c) branching (PHYSYS): O< ｾ＠ The multistate character 15 (protoconch size) was always coded as ordered linear (small-medium-large or large-medium-small). The two programs, PAUP and PHYSYS/WAGNER.S, constructed identical cladograms. The analyses resulted in pairs of equally parsimonious trees that were congruent except for the relative positions of Philippia and Psilaxis. The shortest trees (length = 40) were found when the character states of characters 2 and 8 were arranged in "logical" sequence (large-medium-small and many-onenone, respectively), rather than when postulating independent derivations (''branching") of stages 1 and 2 in these characters (length = 42, identical tree topology). The two most parsimonious trees as illustrated in Fig. 22 will be discussed. DISCUSSION The two resulting, equally parsimonious trees differ in the arrangement of Architectonica, Psilaxis and Philippia (Fig. 22a,b). The difference is caused by a hypothetical parallelism in character 15 in one tree (Fig. 22a) and a reversal in character 9 in the other (Fig. 22b). The former hypothesis is here preferred, since parallel evolution in character 15 (protoconch size), a character that displays homoplasy also in other parts of the tree, seems much more likely to have occurred than a reversal from a specialized rachidian organization (character 9). The following discussion is based on this cladogram (Fig. 22a). The tree is fully resolved (no polychotomies), but the various branches (indicated by letters in Fig. 22) are not equally well supported by the present data. Branches A, H and K are well supported by suites of anatomical and radular characters, indicating monophyly of Architectonicidae, Architectonica +new taxon, and Discotectonica + Granosolarium, respectively. Branch Fis supported by a radular character (8, reduced number of flanking cusps on rachidian) as well as by opercular and shell characters. Branches C, E, G and L are each defined by a number of shell characters. Branches Band Dare defined by a single shell character each, one of which (character 15 in B) displays parallelism elsewhere in the tree. The placement of Heliacus is therefore problematic: besides these two characters and the synapomorphies of the family (branch A), it shares with other genera only two additional homoplasious characters (17 and 20, discussed below). FIG. 22a,b (Opposite page). Most parsimonious cladograms of Recent architectonicld genus-group taxa, based on data matrix shown in Table 2 (Appendix). Number to the left of arrow =character, number to the right =character state; right arrow = gain of apomorphic character state; left arrow = reversal; double line = homoplasy; letters = branches as referred to in the text Length =40, consistency index = 0.80. 223 Architectonicidae Phylogeny OI ::c Architectonlca II I\) ｾｎ＠ N--N new taxon u' ' ' N N Q Q ｾ＠ w N c... ,, Psllaxls IO ' ::n).l:J:m Ａｾ＠ Phlllppla ｾ＠ c ｾ＠ Baslsulcata m II N Hellacus ｾ｝＠ CID ｾ＠ .!.!. •!!. ｾ＠ Dlscotectonlca .... U1 W _. y Ay y ... o-- m lU )> Granosolarlum ｾ＠ .!!. -'' ,,,,,, -----ｾ｡ｇｉｕｬｎ＠ Solatlsonax Pseudotorlnla n ｾｳ＠ II f u; ,... Pseudomalaxls II !!!! Splrolaxls OUTGROUP m N N Archltectonlca II I\) aｎｾｃｏ＠ '''' --ON u; ' new taxon N Q ｾｇ＠ y II Psilaxis y ｾ＠ Phlllppla 224 Rudiger Bieler Homoplasious characters Seven of the 29 characters have "less-than-ideal" character state distributions in the cladogram (i.e., a consistency index <1): Character 5 (proboscis sheath length): The reversal to a short proboscis sheath length in branch K, comprising Discotectonica and Granosolarium, may be due to the overall modification of the alimentary tract, including the replacement of the radula by a rod-like structure, and the loss of a buccal cavity. Character 8 (number of flanking cusps on rachidian): The total reduction of the flanking cusps on the rachidian is the final step in the modification from a "taenioglossate" rachidian to a "ptenoglossate" tooth (the first step appears in branch F), and might have occurred in parallel in Architectonica and Basisulcata, possibly due to similar adaptations as cnidarian predators (see below). Character 14 (operculum shape): The change from circular- to oval-shaped opercula occurs twice, in branch F and again, as an autapomorphy, in Discotectonica. Discotectonica and the taxa defined by branch F comprise the species with largest shell size in the family. Whenever architectonicids reach a critical size limit, there has to be compromise between the optimal body volume per whorl and the ability to move a large shell in the sand (where most members of the family spend at least their resting periods). Large architectonicids of various genera develop a more or less sharp peripheral shell keel that allows easier burrowing in the sand. This results in a triangularly shaped aperture and, in tum, a compatible (i.e., oval) operculum. Character 15 (protoconch size): The character states "protoconch medium sized" and "protoconch large" both show parallelism in the cladogram. This character, with its arbitrarily chosen size classes, is probably the weakest in the analysis. The development of very large protoconchs in several groups, however, might be explained by a directly selective pressure on the larvae at settlement [Robertson et al. (1970: 63), proposed an adaptive advantage to large protoconchs in Psilaxis that precludes their being swallowed by the polyps of their host corals]. Character 17 (anal keel): An anal keel [as defined by Robertson, 1964: 3, fig. 10) on the larval shell is present in all Recent Philippia and Psilaxis species (but lacking in some fossil members attributed to these genera) and is more or less well-developed in a few species of Architectonica, of the new taxon, and of Heliacus. This, the occurrence of such a keel in the Paleocene architectonicid genus Dinaxis (see Bieler, 1985a: 239, pl. 2 fig. 10), and a similar (homologous?) structure in Gegania valkyrie (Fig. 14), indicate that this feature is not very reliable as a taxonomic character and is probably older than indicated by the cladogram. Character 20 (lower peripheral shell keel formed by IPR): The infraperipheral rib (IPR) serves as the lower peripheral shell keel in Philippia and Heliacus. Both groups have relatively high-spired shells with rounded whorls, that facilitate movement on and between polyps in cnidarian colonies. Similar shell shape, and resulting identical placement of the IPR, is here interpreted as convergent. Character 21 (3 narrow outer basal field ribs): Architectonicidae Phylogeny 225 The parallel occurrence of this character in branch C (leading to Pseudotorinia, Pseudomalaxis and Spirolaxis) and as an autapomorphy of Granosolarium is most difficult to interpret. In its complexity (number, position and sharp definition of the ribs) it was initially thought to be a good taxonomic character uniting Pseudotorinia, Pseudomalaxis and Granosolarium. The entirely different alimentary tract (Haszprunar, 1985e) however, groups Granosolarium with Discotectonica (branch K). Placement of extinct taxa A number of architectonicid genus-group taxa are known only from fossils. Based on proto- and teleoconch characters alone, three of them can be placed in the present cladogram (Fig. 23): Stellaxis Dall, 1892 (described from the Eocene): Members of this genus have a more or less smooth, cone-shaped teleoconch (Bieler, 1985a: 240, pl. 2 fig. 11, pl. 3 fig. 15), that is very similar to the organization found in Architectonica, Psilaxis and Basisulcata. The small- to medium-sized protoconch has no anal keel. Nipteraxis Cossmann, 1915 (described from the Eocene): This group is very close to Heliacus in shell characters, except for the organization of the peripheral keel [the upper peripheral rib (UPR) is much heavier and the infraperipheral rib (IPR) is much less developed; Bieler, 1985b: 94, 96 fig.2, pl.1 fig. 2]. Climacopoma Fischer, 1885 (described from the Eocene): The characters comprised by branches B and E in Fig. 22 group this genus with Discotectonica, Granosolarium and Solatisonax (Bieler 1985a: 244, pl. 4 figs. 16-18). Two monotypic genera, Eosolarium Chavan, 1947 [described from the Upper Cretaceous (Campanium)], and Dinaxis Dall in Aldrich, 1895 (described from the Paleocene) could not be placed in the tree. Eosolarium lacks sufficient information on protoconch characters and teleoconch shape due to the poor condition of the type material (Bieler, 1985b: 106, pl. 4 fig. 15). The unique shape and sculpture of the juvenile type specimen of Dinaxis did not allow classification with any of the Recent taxa (Bieler, 1985a: 239, pl. 2 fig. 10). Evidence from the fossil record Merrill (1970: 273, fig.1) proposed a tree including nine architectonicid genusgroup taxa here considered valid, arranged by similarity of the taxa and, based on the fossil record, by their appearance in time (Fig. 24). The branching sequence of this tree is remarkably similar to that of the cladogram (Fig. 22). Merrill's time scale could, with two exceptions, be directly applied to the cladogram (Discotectonica, as Acutitectonica, was grouped with Architectonica and Philippia by Merrill, because it was believed to have a similar radula in addition to the rod-like structure; Solatisonax appears relatively late in the time scale because some species now included were then thought to belong to Heliacus). Thus, the overall branching pattern of the cladogram is supported by fossil evidence, with the groups having sculptured shells, circular opercula and "taenioglossate-like" radulae in the Recent forms evolving earlier than those with smooth shells, oval opercula and "ptenoglossate-like" radulae. Evidence from ontogeny In some cases the direction of character state change can be directly inferred Rudiger Bieler 226 ...> Architectonic a new taxon n )I 'i\rchitectonicinae" n I» Psilaxis Philippia -------- )f "Philippiinae" > "Heliacinae'' t Stellaxis Basisulcata - - - - - - - - - - · tNipteraxis Heliacus Discotectonica Granosolarium m -----· t Climacopoma Solatlsonax .. Pseudotorlnia n Pseudomalaxis r- ｾ＠ H "Pseudomalaxlnae" Spirolaxls ii )C enFIG. 23. Modified cladogram of Architectonicidae, including fossil genus-group taxa. Dotted lines indicate tentative placement of extinct taxa; thick branches mark wellsupported clades; subfamily names of authors in quotation marks. Architectonicidae Phylogeny 227 from ontogenetic evidence: all oval-shaped opercula in this family are circular in their early whorls (character 14), and the direction from sculptured to smooth states in teleoconch characters 23, 25 and 27 can be demonstrated in juvenile shells of Architectonica and Philippia, where traces of the presumed ancestral character states can often be seen on the first teleoconch whorl. ｾ＠ ｾ＠ (,) (,) -- 'E -:c0 (,) Cl> ... (,) "' ｾ＠ a. ｾ＠ :c (I) ';;( ｾ＠ ｾ＠ ｾ＠ (,) c Cl> (I) (,) c 0 :s = 0 a; 0 en :c :s Q ::::! ·;;c ｾ＠ :c 0... a. :s Q) "' Q (I) 0 'O Q) (I) ｾ＠ ｾ＠ E 0 'O (,) .! ｾ＠ ｾ＠ - (I) 0 w I( :5 I( ｾ＠ (I) 0 .! ｾ＠ ·r:0 (I) ｾ＠ "' ｾ＠ "' ｾ＠ I 1, RECENT '' MIOCENE I I / ｾＧ＠ / '-...... ｾ＠ / ｾ＠ EOCENE PALEOCENE ＢＧＭＮｾ＠ ｾ＠ ｾ＠ CRETACEOUS JURASSIC TRIASSIC "'-... FIG. 24. Dendrogram of the Architectonicoidea, arranged by sequence of appearance in time (based on fossil evidence). Vertical scale compressed and only periods significant to Architectonicoidea evolution shown within each era. Time dimensions not proportional [Revised after Merrill (1970: pl. 42): manuscript name, non-architectonicid genus and subgenus Heliacus (Gyriscus) omitted; for pairs of nominal taxa here regarded as synonyms (e.g., Spirolaxis-Paurodiscus, Pseudotorinia-Awarua) only earliest placement used; distinction in generic and subgeneric levels omitted; arrangement of taxa modified to facilitate easier comparison with cladogram in Fig. 22). Comparison with earlier ·classifications It was in a way unfortunate that of all the architectonicid genera, Architectonica has the most derived radula in the family. Because Architectonica is the type genus of this family, the Architectonicidae were subsequently considered to be ptenoglossate by various authors. Fretter & Graham (1982: 364) for instance, see a subgroup within the "Heterogastropoda": Architectonicidae, Epitonioidea 228 Rudiger Bieler and Eulimoidea" ... whose radular teeth are numerous and homodont." In order to explain the occurrence of radulae with fewer teeth per row in the family, a reduction had to be postulated: Climo (1975: 281) interpreted the reduction as a reflection of the smaller adult size, and Golikov & Starobogatov (1975: 222) placed the Architectonicidae and Mathildidae in a suborder Gymnoglossa Gray, 1853, as members who "retain their radula (which shows a gradually decreasing number of teeth)." The ptenoglossate-like radula in Architectonica is clearly a derived condition and most likely evolved as an adaptation to its specialized feeding on cnidarians. Similarity with the radulae of Epitoniidae and Janthinidae can only I>e interpreted as convergence (see also Thiele, 1928: 73). The "typical" architectonicid radula is "five-toothed taenioglossate-like" with two pairs of almost identical marginals as illustrated in Fig. 12. Due to the high degree of variability displayed in shell, opercular, and especially radular characters in this family, the Architectonicidae have been variously split into two families [Solariidae and Toriniidae: Troschel (1875), Schepman (1909); Architectonicidae and Heliacidae: Cotton & Godfrey (1933), Rippingale & McMichael (1961); Architectonicidae and Mangonuidae: Iredale (1936)), and some suggested further splitting: "The Architectonicidae and Toriniidae, though, are rather heterogeneous as to their radulae and may have to be divided into a number of smaller families" (Golikov & Starobogatov, 1975: 223)." Others (e.g., Marshall, 1887; Powell, 1965; Abbott, 1974) used two or more subfamilial divisions. At least the following eight family-group names have been proposed: for "Architectonica-group": - Architectonicidae Gray, 1850 (as "Architectomidae," based on Architectonica Roding, 1798). - Solariidae Chenu, 1859 [based on Solarium Lamarck, 1799 (= Architectonica)]. for "Philippia-group": - Philippiinae Boss, 1982 (based on Philippia Gray, 1847). for "Heliacus-group": - Toriniidae Troschel, 1875 [as family, ''Toriniacea," based on Torinia Gray, 1842 (= Heliacus Orbigny, 1842)). -Teretropomidae Rochebrune, 1881 (based on Teretropoma Rochebrune, 1881 [= Heliacus (Teretropoma)], as a family of land snails}. - Heliacidae Cotton & Godfrey, 1933 (based on Heliacus). for "Pseudomalaxis-group": - Mangonuidae Iredale, 1936 [based on Mangonuia Mestayer, 1930 (=Pseudomalaxis Fischer, 1885)]. - Pseudomalaxinae Garrard, 1977 (based on Pseudomalaxis). Three recent revisions used subfamilies as categories in the Architectonicidae (Table 3, Appendix). Garrard (1977), in his revision of Australian architectonicids, used mainly shell shape and opercula as taxonomic characters; Boss & Merrill (1984) exclusively used characters of radulae and jaws; and Melone & Taviani (1985) used shell, opercular and radular characters. None of the three taxonomic arrangements agrees with the data presented here. For instance, the "rod-like structure," a highly derived character shared by species of the genera Discotectonica and Granosolarium (indicated by"*" in Table 3), indicates a very close re- Architectonicidae Phylogeny 229 lationship of the two. In Garrard's system the two genera were separately placed in subfamilies Architectonicinae and Heliacinae; in Boss & Merrill's classification at least one of them (Discotectonica; Granosolarium was not covered) is placed in the Philippiinae. The results of this analysis (Fig. 23) do not support splitting of the family into previously defined subfamilies. Three wellsupported clades (and Heliacus) can be distinguished within the family, and all are incongruent with the established subdivisions. Pairs of taxa that form monophyletic groups (Architectonica/new taxon, Philippia/ Psilaxis, Pseudomalaxis/ Spirolaxis) can be treated as genera (Architectonica, Philippia, Pseudomalaxis, respectively), but the subfamilies of authors (Table 3) are largely polyphyletic assemblages (the "Heliacinae" are usually defined by shared ancestral character states). CONCLUSIONS Two basic problems in earlier classifications of the Architectonicidae have been encountered: (1) the misinterpretation of the type genus as being "typical" for the family, and (2) the fact that no distinction was made between features that can be used to describe a taxon and characters that define monophyletic units (i.e., syn- and autapomorphies). The family Architectonicidae is a monophyletic group that is defined by a number of synapomorphies in anatomical, radular and opercular characters. Splitting the group into several families is not justified. Architectonicidae and Mathildidae form the superfamily Architectonicoidea, which is defined by a suite of anatomical synapomorphies and the common possession of a heterostrophic protoconch (a symplesiomorphy). Opercular and radular characters, however, are not synapomorphic, with the "five-toothed-taenioglossate-like" radulae of Architectonicidae and Mathildidae having evolved independently by parallel reduction of non-homologous teeth. The most parsimonious cladogram (the one that involves the fewest number of "ad hoc hypotheses" of homoplasy) of the architectonicid genus-group taxa based on anatomical, radular, opercular and shell characters shows three distinct, well-supported clades (Architectonica + new taxon + Philippia + Psilaxis + Basisulcata; Discotectonica + Granosolarium + Solatisonax; Pseudotorinia + Pseudomalaxis + Spirolaxis). The position of Heliacus in the tree is weakly supported (Heliacus is mainly defined by unique derived characters and retained symplesiomorphies). The relative branching sequence in the cladogram is supported by fossil and ontogenetic evidence. Three fossil taxa (Stellaxis, Nipteraxis, Climacopoma) can be added to the tree, based on shell characters alone. Architectonicids with smooth shells, oval-shaped opercula and "ptenoglossatelike" radulae apparently evolved later than those with sculptured shells, circular opercula and "taenioglossate-like" radulae. The use of the "ptenoglossate-like" radula as a character to combine Architectonicidae with mesogastropod families such as Epitoniidae and Janthinidae is not justified; the similar radular type must be explained by convergence. In order to derive a classification from the cladogram it is necessary to assign absolute rank to the monophyletic groups inferred: the 12 architectonicid taxa Rudiger Bieler 230 that have been tenned "genus-group taxa" throughout the study can either all be interpreted as separate genera, or, several of them if they form monophyletic groups, can be treated as subgenera of a single genus (e.g., Architectonica, Philippia, Pseudomalaxis). The published subdivisions of the Architectonicidae into subfamilies are not congruent with the groups found in this study; they were not based on monophyly. For a small family such as this, the need for or advantage in using and naming subfamily units is doubtful. ACKNOWLEDGEMENTS Most data on Architectonicidae used in this study were gathered for a Ph.D. thesis (University of Hamburg, Federal Republic of Germany) under Prof. Dr. Otto Kraus, whose advice and support are gratefully acknowledged. That part of the work was funded by a stipend of the "Studienstiftung des Deutschen Volkes" Bonn, Federal Republic of Germany. Additional studies on Architectoniddae and Mathildidae were made during Smithsonian Postdoctoral Fellowships at the National Museum of Natural History, Washington, and at the Smithsonian Marine Station at Link Port, Florida, U.S.A. Material from 59 collections was studied (see acknowledgements in my publications cited). Space forbids me to thank again all the helpful individuals and institutions who made this study possible. I am indebted to Dr. Gerhard Haszprunar (University of Vienna, Austria) for supplying serial sections of mathildids, to Ms. Paula M. Mikkelsen (Indian River Coastal Zone Museum, Ft. Pierce), Dr. James F. McLean (Los Angeles County Museum), Mr. Bruce Marshall (National Museum of New Zealand, Wellington), and Dr. John D. Taylor [British Museum (Natural History), London] for supplying additional alcohol-preserved architectonicid specimens, to Ms. Elana Benamy (Academy of Natural Sciences of Philadelphia) and Mr. Ian Loch (Australian Museum, Sydney) for the loan of additional architectonicid and mathildid specimens, to Dr. Anders Waren (Swedish Museum of Natural History, Stockholm) for providing previously unpublished SEM photomicrographs of mathildid radulae, and to Mr. Gary Rosenberg (Harvard University, Cambridge) for helping with literature searches. The PAUP program (version 2.40) was run on an IBM-AT personal computer, PHYSYS was run on an UNIVAC mainframe computer with the assistance of Dr. Mary F. Mickevich (USNM). Scanning electron micrographs.of gold/palladium-coated specimens were taken on a Zeiss NOVASCAN-30 machine at the Smithsonian Marine Station at Link Port, with the assistance of Ms. Julianne Piraino. I thank Dr. Kenneth J. Boss (Museum of Comparative Zoology, Harvard University, Cambridge), Dr. Jonathan A. Coddington (USNM), Dr. M.G. "Jerry" Harasewych (USNM) and Ms. Paula M. Mikkelsen (IRCZM) for critical comments on this manuscript. LITERATURE CITED ABBOTT, R.T. 1974. American Seashells. The Marine Mollusca of the Atlantic and Pacific Coasts of North America [2nd ed.]. Van Nostrand Reinhold Co., New York. 663 pp. ALDRICH, T.H. 1895. Descriptions of two new Eocene Solariidae from Alabama. The Nautilus, 9: 1- 2 BANDEL, I<. 1984. The radulae of Caribbean and other Mesogastropoda and Neogastropoda. Zoologische Verhandelingen, 214: 1-188. BAYER, C. 1940. Catalogue of the Solariidae in the Rijksmuseum van Natuurlijke Historie. I. Solarium s.s. Zoologische Mededelingen, 22: 223-256. BIELER, R 1984a. Die Gattungen der Architectonicidae (Gastropoda: "Heterogastropoda"). Allgemeines und Tell 1: Pseudomalaxis. Archi'o ftlr Molluskenkunde, 115: 53-103. BIELER, R. 1984b. Morphometrische Analyse der Architectonica maxima-Gruppe im lndo-Pazifik (Mollusca: Gastropoda: Architectoniddae). Verhandlungen des naturwissenschaftlichen Vereins in Hamburg, (NF) 27: 453-492. BIELER, R. 1985a. Die Gattungen der Architectonicidae (Gastropoda: "Heterogastropoda''). Tell 2: Architectonica, Philippia, Dinaxis, Stellaxis, Discotectonica, Solatisonax, Climacopoma, Granosolarium. Archiv ftlr Molluskenkunde, 115: 231-265. BIELER, R. 1985b. Die Gattungen der Architectonicidae (Gastropoda: Allogastropoda). Tell 3: Pseudotorinia, Nipteraxis, Heliacus, Eosolarium. Archiv ftlr Molluskenkunde, 116: 89-117. BIELER, R. 1986. Revision of genera and lndo-Pacific species in the family Architectonicidae (Abstract, American Malacological Union 51st Annual Meeting). American Malacological Bulletin, 4: 108-109. BIELER, R. 1987. Die Gattungen der Architectonicidae (Gastropoda: Allogastropoda). Tell 4: Architectonicidae Phylogeny 231 Heliacus (Pyrgoheliacus) n.subgen. und Architectonica (Adelphotectonica) n.subgen. Archi'o far Molluskenkunde, 117: 203-215. BOSS, K.J. 1982. Mollusca [and] Oassification of Mollusca. In: Parker, S.P. (Ed.), Synopsis and Classification of LiTJing Organisms. McGraw-Hill, New York; Vol.1: 945-1166, Vol.2: 1092-1096. BOSS, K.J. &: MERRILL, AS. 1984. Radular configuration and the taxonomic hierarchy in the Architectonicidae (Gastropoda). Occasional Papers on Mollusks, 4(66): 349-411. BOUVIER, E.L 1886a. Contributions a l'etude des Prosobranches ptenoglosses. Bulletin de la Sociitt malacologique de France, 3: 77-130. BOUVIER, E.L. 1886b. Observations anatomiques relatives aux Solarides et aux Janthinides. Bulletin de la Sociitt philomatique de Paris, (7)10: 151-156. BOUVIER, E.L 1887. Systeme nerveux, morphologie generale et classification des Gasteropodes prosobranches. Annales des Sciences naturelles (Zoologie), (7)3: 1-510. CERNOHORSKY, W.O. 1972. Marine Shells of the Pacific, II. Pacific Publ., Sydney. 411 pp. CERNOHORSKY, W.O. 1978. Tropical Pacific Marine Shells. Pacific Publ., Sydney. 352 pp. CHAVAN, A. 1947. La faune campanienne du Mont des Oliviers d'apres les materiaux VignaJ-Masse. Journal de Conchyliologie, 87: 125-197. CHENU, J.C. 1859. Manuel de Conchyliologie et de Paltontologie conchyliologique, 1: vii+508 pp. V. Masson, Paris. CLIMO, F.M. 1975. The anatomy of Gegania ?Jalkyrie Powell (Mollusca: Heterogastropoda: Mathildidae) with notes on other heterogastropods. Journal of the Royal Society of New Zealand, 5: 275288. COSSMANN, M. 1896. Essais de Paltoconchologie comparie, 2: 179 pp. Paris. COSSMANN, M. 1915. Essais de Paltoconchologie comparie, 10: 292 pp. Paris. COITON, B.C. &: GODFREY, F.K. 1933. South Australian shells (including descriptions of new genus and species), Part VII. South Australian Naturalist, 14: 72-108. DALL, W.H. 1892. Contributions to the Tertiary fauna of Florida, with especial reference to the Miocene Silex-beds of Tampa and the Pliocene beds of the Caloosahatchie River, II. - Streptodont and other gastropods, concluded. Transactions of the Wagner Free Institute of Science, 3: 201-458. EAMES, F.E. 1952. A contribution to the study of the Eocene in Western Pakistan and Western India. C. The description of the Scaphopoda and Gastropoda from standard sections in the Rakhi Nala and Zinda Pir areas of the western Punjab and in the Kohat District. Philosophical Transactions of the Royal Society of London, (B)236(631): 1-168. FARRIS, J.S. 1970. Methods for computing Wagner trees. Systematic Zoology, 19: 83-92. FARRIS, J.S. 1975. Formal definitions of paraphyly and polyphyly. Systematic Zoology, 23(4): 548554. FINLAY, H.J. 1927. A further commentary on New Zealand molluscan systematics. Transactions and Proceedings of the New Zealand Institute, 57: 320-485. FISCHER, P. 1885. Manuel de Conchyliologie et de PaUontologie conchyliologique ou Histoire naturelle des Mollusques TJivants et fossiles (suivi d'un appendice sur les Brachiopodes par D.P. Oehlert). (8): 689-784 (1885). F. Savy, Paris. FITCH, W.M. 1971. Toward defining the course of evolution: minimum change for a specified tree topology. Systematic Zoology, 20: 406-416. FRETIER, V. 1948. The structure and life history of some minute prosobranchs of rock pools: Skeneopsis planorbis (Fabricius), Omalogyra atomus (Philippi), Rissoella diaphana (Alder) and Rissoella opalina (Jeffreys). Journal of the Marine Biological Association of the United Kingdom, 27: 597632. FRETTER, V. &: GRAHAM, A. 1949. The structure and mode of life of the Pyramidellidae, parasitic opisthobranchs. Journal of the Marine Biological Association of the United Kingdom, 28: 493-532. FRETIER, V. &: GRAHAM, A. 1982. The prosobranch molluscs of Britain and Denmark. Part 7 'Heterogastropoda' (Cerithiopsacea, Triforacea, Epitoniacea, Eulimacea). Journal of Molluscan Studies, Supplement 11: 363-434. FRETTER, V., GRAHAM, A. &: ANDREWS, E.B. 1986. The prosobranch molluscs of Britain and Denmark. Part 9 - Pyramidellacea. Journal of Molluscan Studies, Supplement 16: 557-649. GARRARD, T.A. 1961. Mollusca collected by M.V. "Challenge" off the east coast of Australia. Journal of the Malacological Society of Australia, 1(5): 2-37. GARRARD, T.A. 1977. A revision of Australian Architectonicidae (Gastropoda: Mollusca). Records of the Australian Museum, 31: 506-584. GHISELIN, M.T., DEGENS, E.T., SPENCER, D.W. &: PARKER, R.H. 1967. A phylogenetic survey of molluscan shell matrix proteins. Breviora, 262: 1-35. GIUSTI, F. &: SELMI, M.G. 1982. The atypical sperm in the prosobranch molluscs. (Proceedings of the 7th International Malacological Congress), Malacologia, 22: 171-181. GOLIKOV, A.N. &: STAROBOGATOV, Y.I. 1975. Systematics of prosobranch gastropods. Malacologia, 15: 185-232. GOSLINER, T.M. 1981. Origins and relationships of primitive members of the Opisthobranchia 232 Rudiger Bieler (Mollusca: Gastropoda). Biological Journal of the Linnean Society of London, 16: 197-225. GRAY, J.E. 1842. Synopsis of the Contents of the British Museum [44.ed.]. Mollusca: 48-92. British Museum. London. GRAY, J.E. 1847. List of genera of Recent Mollusca, their synonyma and types. Proceedings of the Zoological Society of London, 15: 129-219. GRAY, J.E. 1850. [Explanation of plates and list of genera]. 124 pp. In: Gray, M.E., Figures of Mollusams Animals, 4: iv+219 pp. London. GRAY, J.E. 1853. On the divisions of ctenobrancheous gasteropodous Mollusca into larger groups and families. Proceedings of the Zoological Society of London, 1853: 32-44. HABE, T. 1943. On the radulae of Japanese marine gastropods (1). Venus, 13: 68-76. HABE, T. 1958. On the radulae of Japanese marine gastropods (4). Venus, 20: 43-60. HABE, T. 1961. Coloured lllustrations of the Shells of Japan (II). Hoikusha, Osaka. ix+183+42 pp. HABE, T. & KOSUGE, S. 1966. Shells of the World in Colour. Volll. The Tropical Pacific. Hoikusha, Osaka. vii+193 pp. HASZPRUNAR, G. 1985a. The fine morphology of the osphradial sense organs of the Mollusca. I. Gastropoda, Prosobranchia. Philosophical Transactions of the Royal Society of London, B307: 457-496. HASZPRUNAR, G. 1985b. The fine morphology of the osphradial sense organs of the Mollusca. II. Allogastropoda (Architectonicidae, Pyramidellidae). Philosophical Transactions of the Royal Society of London, B307: 497-505. HASZPRUNAR, G. 1985c. The Heterobranchia - a new concept of the phylogeny of the higher Gastropoda. Zeitschrift filr ZDologische Systematik und Eoolutions-Forschung, 23: 15-37. HASZPRUNAR, G. 1985d. Zur Anatomie und systematischen Stellung der Architectonicidae (Mollusca, Allogastropoda). Zoologica Scripta, 14: 25-43. HASZPRUNAR, G. 1985e. On the anatomy and systematic position of the Mathildidae (Mollusca, Allogastropoda). Zoologica Scripta, 14: 201-213. HASZPRUNAR, G. 1988. A preliminary phylogenetic analysis of the streptoneurous gastropods. In: Ponder, W.F. (Ed.) Prosobranch phylogeny: Proceedings of a Symposium held at the 9th International Malacological Congress, Edinburgh, 1986. Malacological Ret1iew, Supplement, 4: 7-16. (1his volume). HEALY, J.M. 1982. Ultrastructure of spermiogenesis of Philippia (Psilaxis) orytropis, with special reference to the taxonomic position of the Architectonicldae (Gastropoda). Zoomorphology, 101: 197-214. HEDLEY, C. 1903. Scientific results of the trawling expedition of H.M.C.S. "Thetis" off the coast of New South Wales in February and March 1896. Mollusca II, Scaphopoda and Gastropoda. Memoirs of the Australian Museum, 4: 327-402. HENDY, M.D. & PENNY, D. 1982. Branch and bound algorithms to determine minimal evolutionary trees. Mathematical Biosciences, 59: 277-290. HENNIG, W. 1950. Grundzage einer Theorie thr phylogenetischen Systematik. Deutsches Entomologisches Institut, Bonn-Friedrichshagen/ Deutscher Zentralverlag, Berlin. il+370 pp. HENNIG, W. 1966. Phylogenetic Systematics. University of Illinois Press, Urbana. 263 pp. IREDALE, T. 1929. Mollusca from the continental shelf of Eastern Australia. No.2. Records of the Australian Museum, 17: 157-189. IREDALE, T. 1931. Australian molluscan notes, No.1. Records of the Australian Museum, 18: 201-235. IREDALE, T. 1936. Australian molluscan notes, No.2. Records of the Australian Museum, 19: 267-340. IREDALE, T. 1957. An exciting find. Proceedings of the Royal Society of New South Wales, 1955-56: 124-125. JAUME, M.L. & BORRO, P. 1946. Novedades en Moluscos marinos Cubanos. Ret1ista de la Sociedad malacol6gica Carlos de la Torre, 4(1): 13-22. JOUSSEAUME, F. 1888. Description des mollusques recueillis par M. le Dr. Faurot dans la Mer Rouge et le Golfe d'Aden. Mbnoires th la Socilti z.oologique de France, 1: 165-223. KILBURN, R.N. & RIPPEY, E. 1982. Seashells of Southern Africa. Macmillan South Africa. Johannesburg. xi+249 pp. KLUGE, A.G. & FARRIS, J.S. 1969. Quantitative phyletics and the evolution of anurans. Systematic Zoology, 18: 1-32. KNIGHT, J.B. 1931. The gastropods of the St. Louis, Missouri, Pennsylvanian Outlier: Aclisina and Streptacis. Journal of Paleontology, 5: 1-15. KOROBKOV, I.A. 1955. Sprat1ochnik i metodicheskoe posobie po tretichnym mollyuskam. Bryukhonogie. Gostoptekhizdat, Leningradskoe otd. Leningrad. 795 pp. KOSUGE, S. 1966. The family Triphoridae and its systematic position. Malacologia, 4: 297-324. KURODA, T. & HABE, T. 1954. Notes on three remarkable species of Japanese gastropods. Venus, 18: 79-84. LAMARCK, J.B.P.A. DE MONET DE. 1799. Prodrome d'une nouvelle classification des Coquilles, comprenant une redaction appropriee des caracteres genertques, et l'etabllssement d'un grand nombre Architectonicidae Phylogeny . . 233 de genres nouveaux. Memoires de la Socitte d'Histoire Naturelle de Paris, 1: 63-91. MacDONALD, J.D. 1860. Further observations on the metamorphosis of Gasteropoda, and the affinities of certain genera, with an attempted natural distribution of the principal families of the order. Transactions of the Linnean Society of London, 23: 69-81. MARSHALL, B. 1983. A revision of the Recent Triphoridae of Southern Australia. Records of the Australian Museum, Supplement, 2: 1-119. MARSHALL, W.B. 1887. Monograph of the family Solariidae. In: Tryon, G.W., Manual of Conchology; Structural and Systematic, with Illustrations of the Species, 9: 3-32. Philadelphia. MARWICK, J. 1931. The Tertiary Mollusca of the Gisbome District. New Zealand Department of Scientific and Industrial Research, Geological SuTVey Branch, Palaeontological Bulletin, 13: v, 1177. MELONE, G. 1974. Note su alcuni Architectonicidae (Gastropoda, Prosobranchia) del Mediterraneo. Quaderni della Civica Stazione Idrobiologica di Milano, 5: 23-38. MELONE, G. 1975. Considerazioni sistematiche su un Architectonicide giapponese: Acutitectonica acutissima (G.B. Sowerby, 1914) (Gastropoda, Prosobranchia). Conchiglie, 11: 165-174. MELONE, G. & TAVIANI, M. 1985 ["1984"]. Revisione delle Architectonicidae del Mediterraneo. (Atti del Simposio, Bologna 24-26.IX.1982), Lavori della Societa Malacologica ltaliana, 21: 149192. MERRILL, A.S. 1970. The Family Architectonicidae (Gastropoda: Mollusca) in the Western and Eastern Atlantic. Ph.D. thesis, University of Delaware. 338 pp. [University Microfilms, Inc. Ann Arbor, Michigan, No.71-6444). MESTAYER, M.K. 1930. Notes on New Zealand Mollusca. No.5. Transactions and Proceedings of the New Zealand Institute, 61: 144-146. MICKEVICH, M.F. 1978. Taxonomic congruence. Systematic Zoology, 27: 143-158. MICKEVICH, M.F. 1982. Transformation series analysis. Systematic Zoology, 31: 461-478. MICKEVICH, M.F. & FARRIS, J.S. 1984. PHYSYS (Documentation and mainframe FORTRAN program package). MOORE, D.R. 1966. The Cyclostremellidae, a new family of prosobranch mollusks. Bulletin of Marine Science, 16: 480-484. MONTEROSATO, T.A. 1913. Note on the genus Pseudomalaxis, Fischer, and descriptions of a new species and sub-genus. Proceedings of the Malacological Society of London, 10: 362-363. MOQUIN-TANDON, A. 1848. Observations sur les machoires des Helices de la France. Memoires de l'Academie Royale des Sciences de Toulouse, (3)4: 371-381. ORBIGNY, A. d'. 1841-1846. Mollusques. In: Ramon de la Sagra, Histoire Physique, Politique et Naturelle de l'fte de Cuba. 1: 1-208 (1841), 209-264 (1842), 2: 1-112 (1842), 113-380 ["1846" (1853?)]. PLATNICK, N.I. 1987. An empirical comparison of microcomputer parsimony programs. Cladistics, 3: 121-144. PONDER, W.F. 1985. A review of the genera of the Rissoidae (Mollusca: Mesogastropoda: Rissoacea). Records of the Australian Museum, Supplement 4: 1-221. PONDER, W.F. & WAREN, A. 1988. Oassification of the Caenogastropoda and Heterostropha - A list of the family-group names and higher taxa. In: Ponder, W.F. (Ed.) Prosobranch phylogeny: Proceedings of a Symposium held at the 9th International Malacological Congress, Edinburgh, 1986. Malacological Review, Supplement, 4: 288-328. (This volume). POWELL, A.W.B. 1965. New Zealand molluscan systematics with descriptions of new species: Part 5. Records of the Auckland Institute and Museum, 6: 161-168. POWELL, A.W.B. 1979. New Zealand Mollusca; marine, land and freshwater shells. Collins, Auckland. xiv+500 pp. REHDER, H.A. 1935. New Caribbean marine shells. The Nautilus, 48: 127-130. RIPPINGALE, O.H. & McMICHAEL, D.F. 1961. Queensland and Great Barrier Reef shells. Jacaranda Press, Brisbane. Index+210 pp. RISBEC, J. 1955. Considerations sur l'anatomie comparee et classification des gasteropodes prosobranches. Journal de Conchyliologie, 95: 45-82. ROBERTSON, R. 1964. Dispersal and wastage of larval Philippia krebsii (Gastropoda: Architectonicidae) in the North Atlantic. Proceedings of the Academy of Natural Sciences of Philadelphia, 116(1): 1-27. ROBERTSON, R. 1967. Heliacus (Gastropoda: Architectoniddae) symbiotic with Zoanthiniaria (Coelenterata). Science, 156(3772): 246-248. ROBERTSON, R. 1970. Systematics of Indo-Pacific Philippia (Psilaxis), architectonicid gastropods with eggs and young in the umbilicus. Pacific Science, 24: 66-83. ROBERTSON, R. 1973. Cyclostremella: a planispiral pyramidellid. The Nautilus, 87: 88. ROBERTSON, R. 1974 ["1973"]. The biology of the Architectonicidae, gastropods combining prosobranch and opisthobranch traits. (Proceedings of the 4th European Malacological Congress), Malacologia, 14: 215-220. ROBERTSON, R. 1985. Four characters and the higher category systematics of gastropods. American 234 Rudiger Bieler Malacological Bulletin, Special Edition no. 1: 1-22. ROBERTSON, R., SCHELTEMA, R.S. & ADAMS, F.W. 1970. The feeding, larval dispersal, and metamorphosis of Philippia {Gastropoda: Architectonicidae). Pacific Science, 24: 55-65. ROCHEBRUNE, A.T. DE. 1881. Sur un type nouveau de la famille des Cyclostomaceae. Bulletin de la Socittt philomathique de Paris, (7)5: 108-115. [RODING, P.F.] 1798. Museum Boltenianum sit1e catalogus cimeliorum e tn'bus regnis naturae (...}pars secunda continens conchylia si'f1e testacea univalvia, bivalvia & multivalvia. J.C. Trapp, Hamburg. viii+199 pp. SACCO, F. 1892. I molluschi dei terreni terziarii del Piemonte e della Liguria, 12 (Pyramidellidae (fine), Ringiculidae, Solariidae e Scalariidae]. C.Clausen, Torino. 86 pp. SALVINl-PLAWEN, L. von & HASZPRUNAR, G. 1987. The Vetigastropoda and the systematics of streptoneurous Gastropoda (Mollusca). Journal of Zoology, London, 211: 747-710. SARS, G.O. 1878. Bidrag til kundskaben om Norges arktiske Fauna. I. Mollusca regionis arcticae norvegiae. Oversigt over de i Norges arktiske Region forekommende Bleddyr. Christiania. xiil+466 pp. SCHEPMAN, M.M. 1909. The Prosobranchia of the Siboga Expedition. Part II. Taenioglossa and Ptenoglossa. Resultats Siboga-Expeditie, Monograph 49(1b): 109-231. SHUTO, T. 1969. Neogene gastropods from Panay Island, the Philippines. Contributions to the geology and palaeontology of Southeast Asia, LXVIII. Memoirs of the Faculty of Science, Kyushu University, Series D, Geology, 19: 1-250. SWOFFORD, D.L 1985. PAUP- Phylogenetic Analysis Using Parsimony, Version 2.4 {Documentation and IBM PC-compatible FORTRAN program package). Illinois Natural History Survey. TAYLOR, D.W. & SOHL, N.F. 1962. An outline of gastropod classification. Malacologia, 1(1): 7-32. THIELE, J. 1921. Zur Systematik der Mollusken. Archiv fiJ.r Molluskenkunde, 53:140-163. THIELE, J. 1925. Gastropoda der Deutschen Tiefsee-Expedition. II.Tell. Wissenschaftliche Ergebnisse der Deutschen Tie/see-Expedition au/ dem Damp/er "Valdivia" 1898-1899, 17: 35-282 (1348). THIELE, J. 1928. Ober ptenoglosse Schnecken. Zeitschrift fiJ.r wissenschaftliche Zoologie, 132: 73-94. THIELE, J. 1929. Handbuch der systematischen Weichtierkunde. G. Fischer. Jena. 1(1): 1-376. TIBERI, N. 1867. Diagnose du nouveau genre mediterraneen Gyriscus. Journal de Conchyliologie, 15: 303. TROSCHEL, F.H. 1861. Ueber die systematische Stellung der Gattung Solarium. Archiv fiJ.r Naturgeschichte, 27: 91-99. TROSCHEL, F.H. 1875. Das Gebiss der Schnecken, zur Begrllndung einer natflrlichen Classification. Nicolaische Verlagsbuchhandlung. Berlin. 1(4): 133-180. VAUPEL KLEIN, J.C. von, 1984. A primer of a phylogenetic approach to the taxonomy of the genus Euchirella {Copepoda, Calanoida). Crustaceana, Supplement 9: 1-194. VERRILL, A.E. 1881 ["1880"]. Notice of recent additions to the marine Invertebrata, of the northeastern coast of America, with descriptions of new genera and species and critical remarks on others. Part II. - Mollusca, with notes on Annelida, Echinodermata, etc., collected by the United States Fish Commission. Proceedings of the United States National Museum, 3: 356-405. WEBBER, H.H. 1977. Gastropoda: Prosobranchia. Pages 1-97, In: Giese, A.C. & Pearse, J.S. (Eds.), Reproduction of Marine Int1ertebrates, Vol. IV, Molluscs: Gastropods and Cephalopods. Academic Press. New York. 369 pp. WOODRING, W.P. 1928. Miocene mollusks from Bowden, Jamaica. Part II: Gastropods and discussion of results. Carnegie Institution of Washington Publication, 385: vii, 1-564. WOODRING, W.P. 1959. Geology and paleontology of Canal Zone and adjoining parts of Panama. Description of Tertiary mollusks (Gastropods: Vermetidae to Thaididae). United States Geological Survey, Professional Paper, 306-B: iii, 147-239. RUDIGER BIELER Smithsonian Marine Station at Link Port, Florida, 33450 U.S.A. This paper is Contn'bution No.188 of the Smithsonian Marine Station at Link Port, Florida. Note added in proof: The "unnamed new taxon" cited in text and illustrations has meanwhile been described as Architectonica (Adelphotectonica) Bieler, 1987, the subgenus "Heliacus (new taxon)" as Heliacus (Pyrgoheliacus) Bieler, 1987. )... APPENDIX "'1 (") -·--· -· ;:t" TABLE 1. Available data on architectonidd radulae (D =drawing, L =light photomicrograph, S = SEM photomicrograph; "I" marks longest cusp[s] on rachidian; ..... indicates type species of genus-group taxon). ｾ＠ (") 0 species studied radular formula no.of teeth per row no.of cusps rachidian no.of cusps inner marginal no.of cusps outer marginal ;s reference [note 1) (") ;::i.. $::> ｾ＠ Architectonica s.s. maxima 14-0-14 ｾ＠ 28 2 (Philippi, 1849) ;:t" Boss & Merrill, 1984: 358, pl.63 fig.2 (S), pl.64 figs.1-3 (S) [as A. perspectiva(note 2)). ｾ＠ 0 OQ ｾ＠ ? perspectiva (Linne, 1758) 14-0-14 28 • perspectiva 14-0-14 pie ta 1-2 2 Troschel, 1861: 95, pl.5 figs.4-9 (D); 1875: 156, pl.15 figs.4b-f (D); Thiele, 1929: 185 [as Solarium perspectivum (note 3)). 28 2 Boss & Merrill, 1984: 358, pl.56 fig.1 (D) [not pls.63-64). 14-0-14 28 2 This study (S). [?] 12-14 [?] nobilis 7-?-7 14? ? nobilis 8-0-8 16 n.sp. 7-1-7 15 7-1-7 15 (Philippi, 1849) nobilis ROding, 1798 1 2-3 ｾ＠ ;s Bandel, 1984: 62, fig.96 (D); pl.2 fig.4 (S). 2 Boss & Merrill, 1984: 358, pl.56 fig.2, pl.63 fig.1, pl.65 figs.1-2 (S). 1 2 This study (S). 1 2 2 Boss & Merrill, 1984: 359, pl.66 figs.1-3 (S) [as A. d. laevigata (note 4)). 1+1+1 2 2 Climo, 1975: 281, fig.4B (D). Architectonica n.subgen. • reevei (Hanley, 1862) "' (>.) <.n TABLE 1 (cont.) species studied radular fonnula no.of teeth per row no.of cusps rac:hidian hybrida (Linne, 1758) 2-1-2 5 1+1+1 3 3 Melone, 1974: 24, pl.2 figs.1-3 (5). hybrida 2-1-2 5 1+1+1 3 2 Boss & Merrill, 1984: 356, pl50, figs.1-2 (5 by Melone); pl.65 fig.3 (5). hybrida 2-1-2 5 1+1+1 3-4 3-4 • lutea (Lamarck, 1822) 2-1-2 5 1+1+1 3 2 Boss & Merrill, 1984: 359, pl57 fig.1 (D). ? 5? 2-3 2-3 2-3 Thiele, 1925: 113 [79), pl.46 (34], fig.18(0). • krebsii (Morch, 1875) 2-1-2 5 1+1+1 5 3 Boss & Merrill, 1984: 359, pl57 fig.2 (D). oxytropis A. Adams, 1855 2-1-2 5 1+1+1 3 4-5 ? radiata 2-1-2 5 2+1+2 4 5 Thiele, 1928: 87, fig.8 (D); 1929: 184, fig.169 (D) [as Ph. hybrida]. radiata 2-1-2? 5? ? ? ? Robertson, 1970: 9, figs.9B-O (0). radiatn 2-1-2 5 1+1+1 5-6 5-6 no.of cusps inner marginal no.of cusps outer marginal ｾ＠ reference [note 1) Philippia s.s. °' sp. Merrill, 1970: 31, 204, pl.13 fig.2 (D) [as Ph. n.sp.]; Melone & Taviani, 1985: 165, figs.27-28 (S). Philippia (Psilaxis) (Roding, 1798) Climo, 1975: 282, fig.4A (0). This study (5). Basisulcata • lepida (Bayer, 1942) ｾＭ ｾ＠ i::: 2-1-2 5 1 7 7 Melone, 1974: 25, pl5 figs.1-2 (5) [as Acutitectonica mediterranea]; Boss & Merrill, 1984: 356, pl.53 figs.1-2 (5 by Melone) [as Acutitectonica hpida]. ｾ＠ ｾ＠ ""l -·ｾ＠ b:J ""l TABLE 1 (cont) species studied lepida ? lepida radular fonnula no.of teeth per row no.of cusps rachidian 2-1-2 5 1 2-1-2 no.of cusps inner marginal 6-7 no.of cusps outer marginal 6-7 ). reference [note 1) "'t ｾ＠ -·:S-0 --· Melone & Taviani, 19S5: 161, figs.20-21 (5). ｾ＠ n:. 0 5 1 SS SS Solatisonax ;::t Boss & Merrill, 1984: 360, pl.5S fig.1 (0) [as Acutitectonica disca]. -· ｾ＠ ｾ＠ ｾ＠ n:. bannocki 2-1-2 5 (Melone & Taviani, 19SO) borealis 25-30+1+25-30 4-5 4-5 Melone & Taviani, 19S5: 15S, figs.15-16 (5). 2-1-2 5 -12+1+-12 several several 2-1-2 5 -S+I+-S 5-6 5-6 2-1-2? 5? -14+1+-14 7 8 cylindricus 2-1-2 5 13-14+1+ 13-14 -5 -5 Boss & Merrill, 19S4: 361, pl.59 fig.1 (0). trochoides 2-1-2 5 -12+1+-12 6-10? 6-10? Thiele, 1925: 113 (79), pl.46 (34) fig.16(0). 2-1-2 5 >20+1+>20 9-10 10-13 This study (5). bisulcatus 2-1-2 5 6-S+l+6-S 6-S 6-8 ? implexus (Mighels, 1845) 2-1-2 5 >20+1+>20 9 9 • implexus [note 6) 2-1-2 (Verrill & Smith, 1SS1) • injussa '"a :S-0 ｾ＠ Iredale, 1931 Boss & Merrill, 1984: 361, pl.60 fig.2 (0) [as Heliacus (Solatisonax) borealis]. 0 OQ n:. ;::t ｾ＠ This study (5). Heliacus s.s. • cylindricus (Gmelln, 1791) Torinia cylindracea] (Deshayes, 1830) 'f1ariegatus Troschel, 1S75: 157, figs.7a-c (0) [as (Gmelin, 1791) Heliacus (Torinista) Boss & Merrill, 1984: 362, pl.61 fig.1 (D) [as Heliacus (Pseudotorinia) bisulcatus]. Habe, 1943: 75, pl.4 fig.5 (D) [as Heliacus dorsuosus (note 5)). 5 -17+1+-17 N 8-10 8-10 This study (5). (J.) '-1 TABLE 1 (cont) species studied sterkii (Pilsbry & Vanatta, 1908) radular fonnuJa no.of teeth per row no.of cusps rachidian 2-1-2 5 15-17+1+15-17 7-9 7-9 This study (5). 2-1-2 5 9+1+9 [note 8) 8-10 8-10 This study (5). 2-1-2 5 18-20+1+18-20 9-10 6 no. of atsps inner marginal no.of cusps outer marginal ｾ＠ reference [note 1) 00 Heliacus (Grandeliacus) • stramineus [note 71 (Gmelin, 1791) suboariegatus (Orbigny, 1852) Melone, 1974: 25, pl.4 figs.1-2 (5); Melone & Taviani, 1985: 167, figs.31-32 (5) [as H. fallaciosus]. suboariega tus 2-1-2 5 18-22+1+18-22 9-10 9-10 infundibuliformis (Gmelin, 1791) 2-1-2 5 12-14+1+12-14 -7 -7 This study (5). • perrieri [note 9) 2-1-2 5 9-10+1+9-10 S7 'S.7 Boss & Merrill, 1984: 361, pl.59 fig.2 (D) Heliacus (Teretropoma) Boss & Merrill, 1984: 357, pl.52 figs.1-2 (5 by Melone) [as H. fallaciosus]. [as Heliacus (Heliacus) perrieril. Heliacus (Gyriscus) • jeffreysianus (Tiberi, 1867) 2-1-2 5 10-12+1+10-12 SS 'S.8 Boss & Merrill, 1984: 361, pL60 fig.1 (D). jeffreysianus 2-1-2 5 12+1+12 4-8 4-8 Melone & Taviani, 1985: 173, fig.41 CD from Merrill, 1970). Pseudotorinia [note 10) architae 2-1-2 (0.G. Costa, 1841) architae 5 13-15+1+13-15 6 6 Melone, 1974: 24, pl.3 figs.1-2 (5) [as Heliacus architae]. 2-1-2 ｾ＠ ｾ＠ 5 -8+1+-8 2-3 2-3 Boss & Merrill, 1984: 362, pl.61 fig.2 (D) [as Heliacus (Pseudotorinia) architae]. ｾＭ ｾ＠ "'t b:J ｾﾷ＠ ｾ＠ "'t . TABLE 1 (cont.} species studied architae radular formula no.of teeth per row no.ofrusps rachidian 2-1-2 5 15+1+15 no.ofrusps inner marginal 6 no.ofrusps outer marginal 6 ).. reference [note 1) ""t n;:s- -·--· -· nｾ＠ Melone & Taviani, 1985: 170, figs.36-37 (S) [as Heliacus (Torinista} architae]. 0 ;:t Pseudomalaxis s.s. n • zanclaeus zanclaeus (Philippi, 1844) 2-1-2 zanclaeus zanclaeus 2-1-2 zanclaeus meridionalis (Hedley, 1903) 2-1-2 5 5 5 S20+1+S20 18-20+1+18-20 20+1+20 4 5-6 6 4 5-6 9 ｾ＠ Boss & Merrill, 1984: 362, pl.62 fig.1 (D) [as Ps. nobilis]. Melone & Taviani, 1985: 181, figs.56-59 (S). Kuroda & Habe, 1954: 82, fig.3 (D) [as Mangonuia solaris]. ｾ＠ ｾ＠ "'a ｾ＠ ｾ＠ 0 O<) ｾ＠ ｾ＠ zanclaeus meridionlllis 2-1-2 5 -15+1+-15 4 4 This study (S). • centrifuga Monterosato, 1890 2-1-2 5 4+1+4 2 2 Boss & Merrill, 1984: 362, pl.62 fig.2 {D) [as Spirolaxis centrifugal. centrifuga 2-1-2 5 4-5+1+4-5 2? 2? Melone & Taviani, 1985: 185, fig.64 (M) [asSpirolaxis centrifugus1. Pseudomalaxis (Spirolaxis) Note 1: Note 2: Note 3: Note 4: Reproductions of all cited figures published before 1984 can be found in Boss & Merrill (1984), except those of Kuroda & Habe (1954). Based on USNM 747000, juvenile specimen ('Didi). Most likely Architectonica perspecti'Oa, but could have included A. trochlearis (Hinds, 1844). Based on USNM 747441 ('Didi); common but unnamed species, figured as Architectonica sp.III aff. grandiosa Iredale, 1931, by Bieler (1984b: 475, pl.2 fig.IX). Note 5: Solarium dorsuosum Hinds, 1844, is considered a nomen dubium (see Bieler, 1985b: 99); HeUacus dorsuosus of authors is H. implexus (Mighels, 1845) in most cases. Note 6: The type species, Torinista popula Iredale, 1936, is junior subjective synonym of H. implexus (Mighels, 1845} (see Bieler, 1985b: 99). Note 7: The type species, Grantkliacus mortensenae Iredale, 1957, is junior subjective synonym of H. stramineus (Gmelin, 1791) (see Bieler, 1985b: 100). Note 8: In this species the central cusp on the rachidian is not longer/stronger than the flanlcing cusps. Note 9: Heliacus perrieri is here regarded as a subspecies of H. infundibulifonnis (Gmelin, 1791). Note 10: Pseudotorinia is based on a fossil type species, Solarium obtusum Bronn, 1831. ;:t Rudiger Bieler 240 TABLE 2. Data matrix used in cladistic analysis ("?"=not known, "X" =not applicable; both treated as "missing data" in analysis). Characters: 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 Architectonica 0 0 0 0 0 1 1 0 0 0 0 0 0 newtaxon Philippia Psilaxis Basisulcata Discoteeton ica Granosolarium Solatisonax Pseudotorinia Pseudomalaxis Spirolaxis Heliacus OUTGROUP 2 0 1 2 0 1 ? 0 ? 1 0 1 ? 0 ? x1 1 x1 1 1 0 1 1 0 1 1 0 1 1 0 1 1 0 1 0 0 0 1 1 0 1 1 0 ? ? 0 1 1 0 ? ? 0 0 1 1 ? ? 1 1 1 0 1 1 0 1 ? 0 1 1 0 1 1 0 0 0 0 2 1 1 1 2 0 0 1 1 0 1 1 1 1 1 1 1 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 xxxxx xxxxx 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 0 0 0 0 0 0 x x 2 2 1 2 1 1 1 1 0 1 0 1 0 0 0 0 0 0 0 0 0 0 1 1 0 0 1 1 1 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 1 1 1 0 0 1 1 1 1 1 0 0 0 0 0 0 1 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 1 1 1 1 1 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 1 1 0 0 TABLE 3. Cassification of the family Architectonicidae in recent publications ("•" marks taxa with "rod-like structure" instead of radula). Garrard (1977): ARCHITECTONICNAE Boss &: Merrill (1984): ARCHITECTONICINAE Melone &: Taviani (1985): ARCHITECTONICNAE Architectonica (Architectonica) •Architectonica (Discotectonica) Architectonica (Solatisonax) Philippia (Philippia) Philippia (Psilaxis) Architectonica Architectonica Solatisonax Basisulcata PHILIPPIINAE PHILIPPIINAE Philippia (Philippia) Philippia (Psilaxis) •Acutitectonica [note6] Philippia PSEUDOMALAXINAE PSEUDOMALAXINAE Pseudomalaxis (Pseudomalaxis) [note 1] Pseudomalaxis (Spirolaxis) Pseudomalaxis Spirolaxis HELIACNAE Heliacus (Heliacus) Heliacus (Torinista) Heliacus (Awarua) •Heliacus (Claraxis) [note2] [note3] [note4] [note SJ HELIACINAE HELIACINAE Heliacus [note 7) Pseudomalaxis Spirolaxis A warua [note 4) Heliacus Heliacus Heliacus Heliacus Heliacus (Heliacus) (Torinista) [note 8) (Gyriscus) (Granoheliacus) (Redivivus) Note 1: partially included species of Spirolaxis. Note 2: partially included species of H. (Torinista), H. (Teretropoma) and H. (Grandeliacus). Note 3: partially included species of H. (Teretropoma) and Pseudotorinia. Note 4: junior subjective synonym of Pseudotorinia. Note 5: junior subjective synonym of Granosolarium, but partially included species of Pseudotorinia, Architectonica and Solatisonax. Note 6: junior subjective synonym of Discotectonica, but partially included species of Basisulcata. Note 7: partially included species of Pseudotorinia and Solatisonax. Note 8: here based on species of Pseudotorinia. Symposium on Prosobranch Phylogeny, Ninth International Malacological Congress