Platyhelminthes)

Proc. Helminthol. Soc. Wash. 52(1), 1985, pp. 1-20

The Phylogeny of the Cercomeria Brooks, 1982 (Platyhelminthes)

DANIEL R. BROOKS, RICHARD T. O'GRADY, AND DAVID R. GLEN Department of Zoology, University of British Columbia, 2075 Westbrook Mall, Vancouver, B.C. V6T 2A9, Canada

ABSTRACT: A new classification of the parasitic platyhelminths is presented. It is derived by phylogenetic analysis of 39 morphological characters drawn from 19 putative homologous series. The construction and interpretation of phylogenetic trees using Hennigian phylogenetic systematics is discussed briefly to provide a reference point for the conclusions drawn. Parasitic platyhelminths having, at some time in ontogeny, a posterior adhesive organ formed by an expansion of the parenchyma into, minimally, an external pad, form the subphylum Cercomeria. Three superclasses are recognized within the Cercomeria: the Temnocephalidea, which is the most plesiomorphic of the three; the Udonellidea; and the Cercomeridea. Within the Cercomeridea, two classes are recognized: the Trematoda and the Cercomeromorphae. The Trematoda contains two subclasses, the Aspidocotylea and the Digenea, and the Cercomeromorphae contains the Monogenea, Gyrocotylidea, Amphilinidea, and Eucestoda. The Monogenea is the sister-group of the latter three, which together form the Cestodaria. The Gyrocotylidea is the sister group of the Amphilinidea and Eucestoda, which together form the Cestoidea. Several of the character complexes examined are discussed in detail, and all are listed. The number of hooks on the larval cercomer is concluded to be of little help in analyzing phylogenetic relationships among the cercomeromorphs. Other characters are more informative and provide additional support for the phylogeny proposed herein. These include the relative positions of the genital openings, and the structure of the anterior and posterior parts of the nervous system. New homologies are proposed for posterior adhesive organs and anterior body invaginations. The complex life cycles of digeneans and eucestodes are concluded to differ from each other in manner of origin. Eucestodes exhibit terminal addition of ontogenetic stages and nonterminal addition of an invertebrate host. Digeneans exhibit nonterminal addition of ontogenetic stages and terminal addition of a vertebrate host. Com- parison of the classification presented with eight previous classifications shows that it provides the best fit to the data considered.

In the past 75 years there have been at least posed that parasitic platyhelminths possessing a eight proposals for the phylogenetic relationships posterior adhesive organ be included in a sub- and classification of the parasitic platyhelminths. phylum, the Cercomeria. This study extends and Two of the earliest, by Sinitsin (1911) and Fuhr- supports that hypothesis. mann (1928, 1931), postulated a dichotomy between those with a gut and those without a gut. Materials and Methods The remaining six studies (Spengel, 1905; Jani- To provide a reference point for some of the con- cki, 1920; Bychowsky, 1937, 1957; Llewellyn, clusions we will draw, we present a precis of the theory 1965; Price, 1967;Malmberg, 1974) viewed those and methodology of Hennigian phylogenetic system- with a gut as either paraphyletic or polyphyletic. atics. See Wiley (198la) and Brooks et al. (1984) for a more complete discussion. This precis is followed by No general agreement has emerged. The present a list of the characters we analyzed, as well as a brief study is an application of Hennigian phyloge- discussion of diagnoses and keys. netic systematics (Hennig, 1950, 1966) to the Phylogenetic systematics problem. Organisms from different species may resemble each Previous attempts at phylogenetic analysis of other by possessing similar traits that either have been certain parasitic platyhelminths (Brooks, 1977, inherited from a common ancestor or have evolved 1978a, b, 1981a; Brooks and Overstreet, 1978; independently. The first types of traits are homologies, Brooks et al., 198 la, b; Brooks and Caira, 1982) the second types are analogies. In a system produced by evolution, only homologies indicate phylogenetic have been hampered by the lack of well-corrob- (i.e., genealogical) relationships; thus, only homologies orated outgroups (see the next section). This has can be used for a phylogenetic classification. Given that made character analysis difficult. Before attempt- homologues covary in greater numbers than do ana- ing more analyses of particular groups, we con- logues, such a classification will be the most efficient sidered it necessary to construct an hypothesis information storage and retrieval system for systematics (see Farris, 1979; Brooks, 1981b). of the higher level relationships among the platy- Phylogenetic systematics avoids two types of ad hoc helminths. Brooks (1982) provided a brief dis- assumptions that can be introduced into classificatory cussion of the results presented here, and pro- studies. The first is the use of a unique trait to highlight 1

Copyright © 2011, The Helminthological Society of Washington PROCEEDINGS OF THE HELMINTHOLOGICAL SOCIETY a particular taxon, with the concomitant grouping of Phylogenetic classifications are based on the discov- the remaining taxa. Unique traits do set a taxon off by ery of appropriate levels of generality for homologous itself, but it does not follow that the remaining taxa traits, and the recognition of groupings supported by not so highlighted form an evolutionary group. The synapomorphic traits. A number of protocols have been second type of ad hoc assumption concerns explana- advanced for determining the plesiomorphy and apo- tions of ambiguities in the data. Ambiguities are caused morphy of traits in a study group (see, e.g., Stevens, by homoplasy, or parallel and convergent evolution, 1980). To date all have been found either to be capable which is not recognized as such beforehand. This is of giving incorrect estimates (such as the principle of caused by the independent acquisition or loss of a trait "common equals primitive" of Estabrook, 1971,1978, by two or more species. Homoplasious characters allow and Crisci and Stuessy, 1980), or to be special cases of more than one classification to be inferred from the the more general method of "outgroup comparisons." same data set. They will be inconsistent with the phy- Outgroup comparisons are based on the concept that logenetic classification that explains their true origins, a trait found in at least one member of the study group and consistent with at least one classification based on and in a taxon outside the group (the outgroup—a close particular parallel or convergent traits. For example, relative of the group being analyzed) is plesiomorphic. in a group of organisms in which a trait is primitively Such a trait is considered to have evolved prior to the absent, then evolves, and then is lost by some mem- existence of the ancestral species from which the study bers, those lacking the trait could all be classified to- group evolved. By contrast apomorphic traits are those gether, or they could be classified separately to recog- found only in some members of the study group. Be- nize the primitive and derived conditions. cause outgroups can themselves evolve, it is sometimes Phylogenetic systematics recognizes that both of the necessary to use more than one outgroup to confirm above problems require additional data for their res- the plesiomorphy of a trait (Wiley, 198 la). olution. It eliminates the first type of ad hoc reasoning by grouping taxa only by shared derived characters. Homologous series of platy helminth Unique derived characters may be reported in a taxon's characters examined diagnosis, but they are not used to make groupings. Nineteen kinds of larval and adult platyhelminth The second type of ad hoc reasoning is minimized by traits are given below. Each is considered to be a po- considering the largest subset of the data that covaries tential series of homologues produced through evolu- in a single pattern to consist of homologous traits, and tionary transformation. Each component of a series is the exceptions to that pattern to indicate true homo- a character. It is these characters (e.g., bifurcate gut), plasy. Phylogenetic systematics thus neither denies the rather than the series (e.g., intestine), that individual existence of homoplasy, nor assumes its rarity. Its only taxa display. Homologous series may consist of two assumption is that these false indicators of phylogeny characters (e.g., the presence or absence of locomotor do not themselves covary in a pattern better supported cilia) or more (e.g., the various types of posterior ad- than that of the homologues. hesive organs). The derived state of a two-state series Three levels of homologous traits may be discerned. will be a shared derived character for a single mono- Some may be found in all members of a group being phyletic group. A multistate character is produced when classified. These help establish that group's identity, the derived character of a two-state series undergoes but give no clues to the relationships of its members. further evolution. This may produce either further Others may be present in all individuals of one member structural modification or loss (vs. primitive absence) taxon, but absent in the rest. These establish the iden- of the character. The result is an internesting of mono- tity of single taxa, but provide no clues to relationships phyletic groups (decreasing inclusiveness of taxa), each with other group members. Finally, there may be ho- of which inherited a particular character of the series. mologues shared by two or more of the taxa in the The discovery of this internesting is made possible by study group. These indicate particular relationships determining the relative primitiveness and derivation within the group. Traits that are general to the group of the characters. Outgroup comparisons help to es- being studied are plesiomorphies. Those found only tablish the direction, or polarization, of this transfor- within part of the study group and not in any taxa mation. outside the group are apomorphies. Shared plesiomor- a. LOCOMOTOR CILIA: Among platyhelminths adult phies are symplesiomorphies. Unique apomorphies are dalyelloid rhabdocoels exhibit restricted locomotor cil- autapomorphies. Shared apomorphies are synapomor- ia, whereas adults of the Temnocephalidea, Udonel- phies. Two aspects of phylogenetic systematics make lidea, and the trematode and cercomeromorphan groups this nomenclature necessary. First, Hennig rejected (sensu nobis) lack locomotor cilia altogether (see Wil- "idealized morphology" and its stress on the search liams, 1981). The secondary loss of cilia in adult cer- for archetypal forms, and postulated that any species comerians is further corroborated by the presence of is a composite of ancestral and derived traits. Second, ciliated larvae in some members of all groups. traits produced as novelties at one time can become b. VAGINA: In dalyelloids and temnocephalideans a generalized traits of a descendant group. Plesiomorphy single duct extends from the ovary. In some taxa this and apomorphy therefore refer to relative primitive- duct joins the uterus, and in others it opens directly ness and derivation, depending on the level of gener- into the genital atrium. Udonellideans, trematodes ality of the investigation. For example, hair is a syn- (sensu nobis) and cercomeromorphs (sensu nobis) pos- apomorphy (shared derived trait) for mammals when sess two ducts extending from the ovary. One of these discussing tetrapod evolution, but a symplesiomorphy connects with the uterus, contains the ootype region, (shared ancestral trait) when discussing the evolution and receives the vitelline ducts. The other duct may of rodents. open to the exterior by means of a separate pore or a

Copyright © 2011, The Helminthological Society of Washington OF WASHINGTON, VOLUME 52, NUMBER 1, JANUARY 1985 common atrium. Following traditional terminology, Table 1. The four basic modifications to the posterior we call the first duct the oviduct (Hyman [1951] called adhesive organ in the platyhelminth taxa studied. Ad- it the ovovitelline duct), and the second the vagina. hesive secretions are produced by glandule-epidermal We do not believe that this vaginal duct has a coun- modifications, whereas the remaining three traits in- terpart among other turbellarians. Udonellideans, gy- volve parenchymal modifications as well. The expan- rocotylideans, amphilinideans, and eucestodes possess sion of the parenchyma into an external "pad" at the well-developed vaginae. Digeneans and aspidocoty- posterior end of the body is considered to mark the leans have a small duct, called the Laurer's canal, ex- first appearance of a cercomer. This organ is relatively tending from the oviduct region. This usually opens to unmodified in the Temnocephalidea and Udonellidea, the dorsal surface, but can end blindly in either taxon and augmented with muscularization and/or hooks in (almost always in aspidocotyleans), or connect with the the more derived taxa. excretory system (in some aspidocotyleans). Monoge- neans have paired lateral vaginal openings, with ap- Pa- parent secondary loss of vaginae in many groups. Adhe- renchy- c. OVARY AND TESTES NUMBER: Dalyelloids have sin- sive mal gle ovaries and paired testes. Among various groups secre- expan- Mus- of parasitic platyhelminths, testes number has either tions sion cles Hooks decreased to one, or increased to many. The dalyelloid condition is nevertheless found in some temnoceph- Dalyelloidea alideans, digeneans, aspidocotyleans, and monoge- Temnocephalidea neans. Because of the widespread occurrence of changes Udonellidea in testes number, we have not used any of the derived Aspidocotylea states in our analysis. Digenea d. EXCRETORY SYSTEM: Paired lateral excretory ducts Monogenea joining posteriorly into a vesicle comprise an addi- Gyrocotylidea tional trait linking dalyelloids with the rest of the para- Amphilinidea sitic platyhelminths. Eucestoda e. PHARYNX: The doliiform pharynx (barrel-shaped, see Hyman, 1951) found in dalyelloid rhabdocoels occurs in temnocephalideans, udonellideans, digeneans, aspidocotyleans, and monogeneans. It has generally tions exist in the taxa studied (see Table 1). Dalyelloids been considered that there is no pharynx in gyrocot- (e.g., Dalyellia; see Hyman, 1951) have only glandular ylideans, amphilinideans and eucestodes (but see Dis- secretions (see also the discussion for behavioral ob- cussion). servations). Temnocephalideans and udonellideans f. COPULATORY STYLET: Some form of sclerotized have a parenchymal expansion as well (Williams, 1981). copulatory apparatus is part of the male genitalia of In digeneans this expansion has become muscularized dalyelloids, temnocephalideans, udonellideans, and to form a sucker that is midventral in many species, monogeneans. Digeneans, aspidocotyleans, gyrocoty- but posterior in a number of apparently primitive lideans, amphilinideans, and eucestodes lack such groups. Aspidocotyleans possess a posterior ventral structures. sucker early in ontogeny that becomes a rugose or g. INTESTINE: Dalyelloids have a saccate gut, as do loculate disk in the adults. The adhesive organ reaches the majority of turbellarians. This trait is found in its maximum complexity in monogeneans, which pos- temnocephalideans, udonellideans, and most aspido- sess hooks in addition to the three earlier modifica- cotyleans. Digeneans and monogeneans, for the most tions. The early larval stages of gyrocotylideans, am- part, possess bifurcate guts. Gyrocotylideans, amphi- philinideans, and eucestodes possess a posterior linideans, and eucestodes lack a gut. expansion bearing hooks, but apparently lacking ad- h. MEHLIS' GLAND: This gland is lacking in dalyel- hesive secretions. The hooks may persist in a disar- loids, at least as a centralized glandular structure. Al- rayed configuration near the dorsal pore of the rosette though it is not unambiguously clear that temnoceph- funnel in gyrocotylidean adults (e.g., Gyrocotyle urna; alideans possess this gland (see Williams, 1981), all see Lynch, 1945) and at the rounded posterior end of other groups in this study have been shown to have it. the body in amphilinidean adults (e.g., Austramphilina i. POSTERIOR ADHESIVE ORGAN: Posterior attachment elongata; see Rohde and Georgi, 1983). structures in the platyhelminths consist of adhesive j. TENTACLES: Temnocephalideans possess tentacles glandular secretions, suckers, and hooks (see Hyman, on their anterior ends, although in some cases these 1951). The first primarily involve tegumental modi- may be very small and possibly secondarily reduced fications, and the other two involve extensive paren- (Williams, 1981). Some digenean (e.g., bucephalids) chymal modifications as well. Adhesion can thus be and eucestode (trypanorhynchs) groups possess "ten- achieved by chemical action, vacuum principle, and tacles" as well, but these are neither common to all mechanical embedding, respectively. There appear to digeneans and eucestodes, nor of the same structure as be four basic types of morphological modifications to those in temnocephalideans. the posterior holdfast organ: (1) glandule-epidermal k. GENITAL PORES LOCATION: Turbellarians possess secretions, (2) expansion of the parenchyma into an genital pores at the posterior end of the body. With the external "pad" of some sort, (3) development of suck- exception of amphilinideans the rest of the taxa in this ers by muscularization of the pad, and (4) the presence study generally have all genital pores in or near the of hooks. Different combinations of these modifica- anterior half of the body, or proglottid (in eucestodes).

Amphilinideans have the male pore and the vagina in s. LARVAL HOOKS ON THE POSTERIOR ADHESIVE OR- the posterior half, and the uterine pore in the anterior GAN: Monogeneans have 12-16 hooks, gyrocotylideans half (see Fig. 3). 10 equal-sized hooks, amphilinideans six large and four 1. GENITAL PORES ASSOCIATION: In dalyelloids and small hooks, and eucestodes six equal-sized hooks. The many other turbellarians the genital openings occur other groups have no such hooks. together, often in a common genital atrium. In the digeneans and monogeneans the vagina is separate from Method of character analysis the male and uterine openings. The latter two may open Dalyelloids were chosen as the putative outgroup to a common genital atrium and exit through a com- because of the characters they shared with some mem- mon genital pore. In the aspidocotyleans the vaginal bers of the study group. This choice agrees with Kar- opening is lacking altogether, but the relative positions ling's (1974) cladistic analysis of the Turbellaria. Our of the male opening and uterine pore are as in dige- cladogram was initially constructed by Hennigian ar- neans. Gyrocotylideans have all three genital pores sep- gumentation (Hennig, 1966; see Wiley, 198la), and arated but in close proximity in the anterior half of the then checked with WAGNER analysis from the PHY- body. Amphilinideans have the uterine pore in the SYS computer program of Drs. J. S. Farris and M. F. anterior half of the body, and the vagina and male pore Mickevich. Those character series that were entirely separate in the posterior end of the body. Eucestodes within the study group, and thus not amenable to po- possess closely associated male and vaginal openings larization by a taxonomic outgroup (the dalyelloids), (see Fig. 3). were polarized by the functional outgroup method of m. ORAL SUCKER: Most digeneans, aspidocotyleans, Watrous and Wheeler (1981); Farris' (1982) expansion and monogeneans have oral suckers. Gyrocotylideans, of this method was taken into consideration. amphilinideans, and eucestodes possess, at some pe- riod in ontogeny, invaginations at the anterior end that Diagnoses and keys we interpret as vestigial mouths and associated struc- A phylogenetic analysis produces a cladogram, or a tures (see Discussion). tree diagram, depicting the inferred genealogical rela- n. LIFE CYCLE PATTERNS: All of the platyhelminths tionships of taxa. From this tree, a verbal classification considered here are associated with a host of some sort. may be constructed. The main purpose of this form of Temnocephalideans, udonellideans, and aspidocoty- classification is to describe the topology of the tree and leans are associated primarily with invertebrates; al- the internested monophyletic groups it contains. This though some aspidocotyleans occur in vertebrates, these isomorphism produces an efficient information storage hosts are not necessary for completion of the life cycle and retrieval system. Such a classification may be writ- (see Rohde, 1971). Monogeans, gyrocotylideans, and ten as a Linnaean hierarchy, and when augmented with amphilinideans are associated primarily with verte- character information a diagnosis of sorts is produced. brates, although at least one species of amphilinidean But it is not intended to be a traditional, inclusive has an intermediate arthropod host (see Rohde and diagnosis of the traits of each taxon. It is instead a Georgi, 1983). Both digeneans and eucestodes have, description of the characters used in the analysis and with few exceptions, life cycles involving both an in- the hierarchical level at which they are postulated to vertebrate and a vertebrate host. For reasons eluci- be synapomorphic. The diagnosis encompasses char- dated in the discussion we consider the digenean life acters from homologous series that are two-state (the cycle to involve secondary addition of a vertebrate derived state remains a synapomorphy) or multistate host, and the eucestode life cycle to involve secondary (derived state I evolves into derived state II, and thus addition of an invertebrate host. becomes synapomorphic at one level and symplesio- o. NERVOUS SYSTEM: Monogeneans, gyrocotylideans, morphic at another). Hull (1979) and Wiley (198la) amphilinideans, and eucestodes have doubled nervous have discussed these properties extensively. For ex- commissures at the anterior and posterior ends of the ample, a traditional "taxon" diagnosis of the Cercom- body (see Tower, 1900; Watson, 1911; Lynch, 1945; eria would include "gut saccate, bifurcate, or absent." Rohde, 1968, 1975; Allison, 1980; Fairweather and As we show in our results, a cladistic diagnosis sepa- Threadgold, 1983). All other taxa in the study group rates the plesiomorphy and apomorphy in this state- have single anterior and posterior commissures, as do ment and diagnoses the Cercomeria as "gut saccate" all rhabdocoels (Bullock and Horridge, 1965). because that is concluded to be the primitive gut con- p. INVAGINATION OF THE POSTERIOR ADHESIVE ORGAN: dition. The other two gut characters are introduced at Gyrocotylideans, amphilinideans, and eucestodes ex- the less inclusive taxonomic levels where they are pos- hibit partial to complete invagination of the posterior tulated to have evolved. adhesive organ when it occurs during ontogeny. This Cladistic diagnoses are also often inappropriate for trait appears to be unique to them. immediate use as a key. The important attributes of a q. POLYZOIC BODY: Eucestodes are the only taxon in key are clarity and convenience to allow rapid iden- the study group that may possess a polyzoic body. tification of specimens. It is in classifications, not keys, r. OSMOREGULATORY SYSTEM: Eucestodes are the only that phylogenetic relationships are proposed. Some- members of the study group whose larvae have pro- times a key can be both succinct and natural (i.e., it tonephridia in the posterior half, rather than anterior follows the classificatory groupings). But this will be so half, of the body. In addition, whereas adult gyro- only when the character series involved are all two- cotylideans, amphilinideans, and eucestodes possess a state (e.g., if a saccate gut were primitively present and reticulate osmoregulatory system, their early ontoge- always present in the Cercomeria). However, the ex- netic stages have paired lateral ducts joining in a pos- istence of multistate character series (i.e., continuing terior vesicle. This latter condition is characteristic of evolution) usually makes an artificial key (i.e., no evo- the rest of the study group throughout ontogeny. lutionary connotations) preferable.

Results Class Trematoda Rudolphi, 1808 Figure 1 gives the cladogram representing the DIAGNOSIS: Cercomeridea with dorsal vagina best supported phylogenetic hypothesis for the a Laurer's canal (15), cercomer a sucker (16), and cercomerians, based on 39 characters drawn from copulatory stylet lost (17). the 19 putative homologous series listed above. For reasons given in the discussion, some of the Subclass Aspidocotylea Monticelli, 1892 characters and series were not used. The same DIAGNOSIS: Trematoda with vaginal opening cladogram is obtained whether one uses dalyel- lost (18) and ventral sucker modified in adults loids as the outgroup or calculates the most par- into ventral adhesive disk (19). simonious tree possible regardless of the outgroup (see Farris, 1979). We will show later that Subclass Digenea Van Beneden, 1858 the use of acoel turbellarians as an outgroup for DIAGNOSIS: Trematoda with bifurcate gut (20) some of the members of our study group gives and complex life cycle with vertebrate host sec- a poorer fit to the data than does our use of ondarily acquired (21). rhabdocoels. Numbers accompanying the slash marks on each branch of the cladogram refer to Class Cercomeromorphae Bychowsky, 1937 characters that are postulated to be synapomorphic at the level of the branch where they occur. DIAGNOSIS: Cercomeridea with armed cercom- The numbered characters are identified in the er (22), doubled cerebral commissures (23), and classification. doubled posterior commissures (24).

Subclass Monogenea Car us, 1863 Classification and cladistic diagnosis for DIAGNOSIS: Cercomeromorphae with paired the subphylum Cercomeria Brooks, 1982 lateral vaginae (25), bifurcate gut (26), and 12- Subphylum Cercomeria Brooks, 1982 16 hooks on larval cercomer. DIAGNOSIS: Rhabdocoelous platyhelminths Subclass Cestodaria Monticelli, 1891 with no vagina (1), single ovary and paired testes (2), paired lateral excretory vesicles (3), doliiform DIAGNOSIS: Cercomeromorphae with osmo- pharynx (4), saccate gut (5), copulatory stylet (6), regulatory system becoming reticulate in later no locomotor cilia in adults (7), Mehlis' gland ontogeny (27), no intestine (28), posterior body (8), and posterior adhesive organ formed by an invagination (29), no copulatory stylet (30), cer- expansion of the parenchyma into an external comer reduced in size and partially or totally pad, called a cercomer (9). All associated with at invaginated (31), male genital pore not proxi- least one host type. mate to uterine opening (32), and vestigial oral structures (33).

Superclass Temnocephalidea Benham, 1901 Infrasubclass Gyrocotylidea Poche, 1926 DIAGNOSIS: Cercomeria with anterior tentacles DIAGNOSIS: Cestodaria with rosette at poste- (10). Ectoparasites or commensals of inverte- rior end (34) and ten equal-sized hooks on larval brates. cercomer.

Superclass Udonellidea Ivanov, 1952 Infrasubclass Cestoidea Rudolphi, 1808 DIAGNOSIS: Cercomeria with genital pores in DIAGNOSIS: Cestodaria with vagina and male anterior half of body (11) and vagina (12). Ec- genital pore proximate (35) and cercomer totally toparasites of arthropods. invaginated (36). Superorder Amphilinidea Poche, 1922 Superclass Cercomeridea taxon novum DIAGNOSIS: Cestoidea with uterine pore in an- DIAGNOSIS: Cercomeria with genital pores in terior half of body and genital pores in posterior anterior half of body (11), vagina (12), male gen- half (37) and with six large and four small hooks ital pore and uterus proximate (13), and oral on larval cercomer. At least one species known sucker (14). Ecto- and endoparasites of verte- with complex life cycle with invertebrate host brates, invertebrates, or both. secondarily acquired.

Figure 1. Cladogram representing the hypothesized phylogenetic relationships of the parasitic platyhelminth taxa examined. Characters are denoted by numbered slash marks and identified in the diagnoses in the text. Each slash mark postulates a synapomorphy, or evolutionary novelty, common to all the taxa above that branch. In some cases subsequent evolution has modified the character, creating an homologous, or transformation, series. There are 41 postulated changes for 39 characters indicating two cases of parallel evolution: loss of copulatory stylet (#17 and #30) and acquisition of bifurcate gut (#20 and #26). The taxa are identified as having life cycles that are direct in an invertebrate host (DI) or vertebrate host (D V) or complex with both an invertebrate and vertebrate host (CB) (see text for comments on the life cycles of amphilinideans).

Superorder Eucestoda Southwell, 1930 late osmoregulatory system develop- DIAGNOSIS: Cestoidea with polyzoic body (38), ing later in ontogeny complex life cycles with invertebrate host sec- Subclass Cestodaria ... 3 ondarily acquired (39), protonephridia in pos- 3a. Cestodaria in which the posterior ad- terior half of larva (40), cercomer lost during hesive organ of the adult forms a ro- ontogeny (41), and six hooks on larval cercomer. sette; larval cercomer with ten equal- sized hooks; parasitic in Holocephali Infrasubclass Gyrocotylidea Artificial Key for the Subphylum b. Cestodaria in which the cercomer of the Cercomeria Brooks, 1982 adult is totally invaginated; vaginal and la. Platyhelminths with double nervous male genital pores are proximate commissures at the anterior and pos- Infrasubclass Cestoidea ... 4 terior ends of body; larvae with hooks 4a. Cestoidea in which the larval cercomer on the posterior adhesive organ has six large and four small hooks; Class Cercomeromorphae ... 2 uterine pore is in the anterior half of b. Platyhelminths with single nervous the body, the genital pores are in the commissures at the anterior and pos- posterior half; parasitic in fish and tur- terior ends of body; larvae without tles Superorder Amphilinidea hooks on the posterior adhesive organ b. Cestoidea in which the cercomer is lost 5 in ontogeny; larval cercomer has six 2a. Cercomeromorphae in which the pos- hooks; adults usually with polyzoic terior adhesive organ is an opisthap- body; protonephridia in anterior half tor, bearing suckers and hooks; larval of the larvae; complex life cycles opisthaptor armed with 12-16 hooks; Superorder Eucestoda direct life cycles; usually ectoparasites 5a. Copulatory stylet absent; modified va- of fish Subclass Monogenea gina forming a Laurer's canal; oral b. Cercomeromorphae with an invagina- sucker usually present tion at the posterior end of the body Class Trematoda ... 6 in the adult; gut absent; pharynx ab- b. Copulatory stylet present; oral sucker sent; copulatory stylet absent; reticu- absent 7

6a. Trematodes in which the posterior ad- to the armed posterior adhesive organ found in hesive disk is modified to form a large all larval and some adult cercomeromorphs, the ventral adhesive disc; saccate gut pres- term ventral sucker or ventral adhesive disk was ent; direct life cycles; host usually an restri cted to the adhesive organs of trematodes, invertebrate Subclass Aspidocotylea and the term posterior adhesive disk was restrict- b. Trematodes in which the posterior ad- ed to the structure found in udonellideans and hesive organ is a sucker, which can be temnocephalideans. If the posterior adhesive or- terminal, midventral, or secondarily gan of each taxon is given a different name, this lost; gut present, usually bifurcate; can obscure the question of homology, whereas complex life cycles; definitive host at the same time reinforcing arguments for con- usually a vertebrate Subclass Digenea vergent evolution of holdfast organs. But if all 7a. Vagina lacking; genital pores in the pos- of the posterior holdfast organs of the platyhel- terior half of body; anterior tentacles minths are derived from a common ancestor, present (may be secondarily reduced) there then exists a multistate homologous series, Superclass Temnocephalidea each of whose characters are synapomorphic at b. Vagina present; genital pores in the an- certain levels of the cladogram. If this is the case, terior half of body; tentacles lacking then all such organs should have the same name, Superclass Udonellidea with appropriate modifiers for further derived states. Character interpretations build cladograms, Discussion and cladograms can be used to interpret other Character evolution characters. We can justify our cladogram in Fig- The majority of the characters used in this ure 1 by noting that even if all characters relating analysis have been used previously by other to posterior adhesive organs were removed (nos. workers studying the parasitic platyhelminths. 9, 16, 19, 22, 29, 31, 34, 36, and 41), the clas- This was done partly to facilitate comparisons sification would still be fully supported. We can with earlier classifications (discussed later), and justify our interpretation of the homology of pos- partly to offer a phylogenetic hypothesis for fur- terior adhesive organs, and our resulting broad- ther comparative work on newer characters. The ening; of the cercomer appellation, by referring study by Jamieson and Daddow (1982) on the to the four basic series of character development ultrastructure of platyhelminth spermatozoa ap- in that organ, discussed earlier and summarized pears to be compatible with our analysis, al- in Table 1. Muscularization and hooks appear though more work is required (see Rohde, 1971, to be modifications of the phylogenetically ear- 1980). Observations on platyhelminth excretory lier posterior parenchymal expansion. We there- systems by Rohde (1980) and Rohde and Georgi fore suggest that this expansion be called a cer- (1983) offer another synapomorphy for the Cer- comer, that modifiers be used to describe further comeridea (flame cell weir apparatus), the Trem- derived states (loculate, sucker, armed, etc.), and atoda (lamellated walls in the protonephridial that all of the taxa postulated to have inherited ducts), and the Cestoidea (microvilli in the pro- it or its modifications be united into a mono- tonephridial ducts). phyletic group called the Cercomeria. We must comment on several of the characters Some recent discussions of cercomeromorph we have used, in light of their arrangement on relationships (Llewellyn, 1965;Malmberg, 1974) the cladogram. First, we have broadened the use have placed emphasis on the number and struc- of the term cercomer. This is not simply an a ture of hooks on the larval cercomer. There are priori decision, but an indication of the homol- four major characters: 12-16 hooks in mono- ogy that we postulate to exist among the posterior geneans, ten equal-sized hooks in gyrocotyli- adhesive organs of the taxa studied. There are deans, six large and four small hooks in am- thus two aspects to the matter: the hypotheses philinideans, and six hooks in eucestodes. An of homology drawn from structural studies, and initial interpretation from our cladogram would the terms that are used to refer to those struc- be that of a linear homologous series, from larg- tures. The former must have precedence over the est number to smallest, indicating an evolution- latter. Standard usage of the nomenclature would ary trend towards reduction in number of hooks be maintained if the term cercomer was restricted (Fig. 2a). Llewellyn (1965) proposed another se-

'12 - 16 (d) 10 > 6 >6 Figure 2. Four possible evolutionary transformations for the number of hooks on the larval cercomer in the Cercomeromorphae. All four series fit equally lemnocepholideci well on the phylogenetic tree in Figure 1 (see text for discussion). lies (Fig. 2b), suggesting that the four small hooks on the amphilinidean cercomer were secondarily evolved. Malmberg (1974) proposed a series op- Figure 3. Diagrammatic representation of postu- posite to that of Figure 2a, proceeding from lated evolutionary transformations of the relative po- smallest to largest number of hooks. We did not sitions of the genital pores in the Cercomeria. With the exception of the Temnocephalidea, the three pores use the larval hooks as characters because we are the male pore, the vagina, and the uterine pore (see found that a number of different series fit the character series k and 1 in Materials and Methods). classification equally well (Fig. 2). That is, they Temnocephalideans possess a common genital atrium all give equally parsimonious interpretations of into which open a male canal, a uterine canal, and a canal from the ovary. This last canal functions as both the data without the need for postulating parallel an oviduct and a vagina—functions that become sepa- or convergent evolution. Mickevich's (1982) rated in the other taxa with the evolution of a vaginal Transformation Series Analysis would recognize canal. The anterior end of the diagrammatic bodies is cercomer hook number as a "trivial" character. at top. Transformations: (1) pores relocate in anterior, This does not mean that it is of no phylogenetic uterine and male pores meet in atrium, vagina ends in parenchyma, opens to dorsal surface, or joins excretory significance, only that current analytic tech- system; (2') uterine and male pores remain together, niques allow too many evolutionary sequences vagina bifurcates; (2) pores open separately; (3') male to be considered equally likely. The data cannot pore and vagina relocate in posterior; (3) male pore be used to choose between hypotheses. and vagina meet in atrium, uterine pore remains separate. One might consider, however, the series in Fig- ures 2b and 2d to be less likely because they require that the four small hooks of amphilini- illustrated the putative homologous series in Fig- deans be derived from a different embryonic ure 3. source than the other hooks when no empirical Previous studies of cercomerian relationships support for this exists. In fact, studies by Malm- that attempted to link the cercomeromorphs as berg (1974) suggest the opposite, that all larval a group were restricted to one character: the armed hooks are homologous. But this is still an open cercomer. However, for at least some monoge- question, pending comparative developmental neans (Rohde, 1968, 1975), gyrocotylideans studies. Regardless of the resolution of this am- (Watson, 1911; Allison, 1980), and eucestodes biguity, there would still be more than one equal- (Tower, 1900;FairweatherandThreadgold, 1983) ly likely series. This character appears at present enough is known of the anatomy of the nervous to be of little help in studies of platyhelminth system to give two additional characters. Dige- phylogenetic relationships beyond the recogni- neans, aspidocotyleans, and all groups of rhab- tion that those flatworms possessing an armed docoels for which we could obtain data (see Bul- cercomer form a monophyletic group. lock and Horridge, 1965) possess a central On the other hand, we found that the various nervous system comprised of two bilateral lon- characters pertaining to the genital openings in gitudinal trunks with an anterior (cerebral) and cercomerians provided useful evidence not fully posterior commissure. The cercomeromorphs utilized previously. To highlight this, we have differ in their possession of doubled cerebral and

Copyright © 2011, The Helminthological Society of Washington OF WASHINGTON, VOLUME 52, NUMBER 1, JANUARY 1985 posterior commissures. Some eucestodes may dence that a gut never existed in the ancestors even have a third anterior commissure. It was of gyrocotylideans, amphilinideans, and euces- the discovery of doubled posterior commissures todes (we discuss other authors' hypotheses of in Gyrocotyle that led Watson (1911) to postulate an acoel ancestry for the Cercomeromorphae in that the eucestode scolex was derived from the a later section). Our interpretation of secondary posterior end of the body. This theory is falsified loss can be supported by two lines of evidence. by the observation that in gyrocotylideans the First, those taxa lacking a gut are placed as highly rosette develops at the same body end bearing derived groups by synapomorphic characters the larval cercomer, whereas in eucestodes (e.g., other than those involving the condition of the the cysticercus of Hymenolepis; see Alicata and gut. All of the cercomerian taxa postulated to be Chang, 1939) the scolex and cercomer develop plesiomorphic to the Cestodaria possess a gut. at opposite ends of the body. Some ambiguity in Second, it may not be the case that cestodarians character analysis of the nervous system re- lack any vestige whatsoever of a digestive system mains, particularly about the condition in am- (as opposed to a gut). By this we do not refer to philinideans, and the differences between ring the question of the existence of entoderm or en- and bridge commissures (see Lynch, 1945). toderm precursor cells in eucestode development Our analysis also offers new interpretations of (see Mackiewicz, 1981, and references therein). the evolution of the gut in platyhelminths. We If these are present, there is simply a confirma- have considered three characters: saccate gut, bi- tion of the symplesiomorphic trait of the pres- furcate gut, and the lack of a gut. Our cladogram ence of a gut (with the derived character no. 28 in Figure 1 postulates that the bifurcate condition in Fig. 1 becoming "gut components present in is convergent, having evolved separately in the ontogeny" instead of "no gut"). But given our Digenea and Monogenea. The same number of first line of evidence above, this a secondary point character steps (see the Comparisons with Other whose perceived importance to phylogenetic Classifications section) results if it is postulated study arises from assumptions of the necessity that a bifurcate gut arose once as a synapomor- for recapitulatory ontogeny (as well as assump- phy for the Cercomeridea, then reverted to a tions about the formation of germ layers during saccate condition in the Aspidocotylea. This al- cestodarian embryogenesis—see below). We re- ternate interpretation must be considered, es- fer instead to the question of structures associ- pecially in view of the existence of aspidocoty- ated with a gut in the more plesiomorphic flat- leans (e.g., Zonocotyle) with bifurcate guts, and worm taxa. Trematodes and monogeneans digeneans (e.g., Haplosplanchnidae) and mono- exhibit, as do rhabdocoels, a form of ectolecithal geneans (e.g., Bothitrematidae) with saccate guts. embryogenesis (see Hyman, 1951; Rees, 1940). These anomalies may be atavistic traits—char- There is no folding of germ layers to form body acters from an earlier phylogenetic position in a cavities. Instead, a part of the ectoderm grows homologue series. The likelihood of this is great- inward, then hollows out and forms a mouth er when the organisms are one of the most ple- (which may become surrounded by an oral suck- siomorphic taxa in their clade. The more derived er) and pharynx invagination, and parts of the a taxon is, the more likely it becomes that an parenchyma hollow out and form the intestine. apparent atavism is actually a further derived The embryogenesis of the Cestodaria is not so state with superficial similarity to the plesio- easily categorized, mainly because of changes in morphic state. A cladistic analysis of the Digenea the relative sizes of the blastomeres. The basic to the family level by Brooks et al. (1985) suggests pattern is nevertheless ectolecithal or hemiec- that haplosplanchnids are a relatively derived tolecithal (see Douglas, 1963, and references taxon and therefore nonatavistic in their gut con- therein). dition. Similar studies are necessary for the As- Cestodarian embryos clearly do not develop pidocotylea and Monogenea before any further into adults possessing an intestine formed by an conclusions can be made. internal parenchymal cavitation (the larvae of The third gut character we have examined is the Cyclophyllidea possess another cavity, called that of the lack of a gut. We have suggested that the primary lacuna [see Freeman, 1973]). They this condition, which occurs in the Cestodaria do, however, display, at some stage in ontogeny, (sensu nobis), is best interpreted as a derived an anterior invagination of the ectoderm with state within the Cercomeria, rather than evi- varying degrees of muscularization and paren-

Copyright © 2011, The Helminthological Society of Washington 10 PROCEEDINGS OF THE HELMINTHOLOGICAL SOCIETY chymal modification. In gyrocotylideans and Adaptationist explanations require either that most amphilinideans this persists throughout life. functional equivalents be nonhomologous (e.g., We have observed living Gyrocotyle in the spiral birds' wings and insects' wings are adaptations valve of ratfish using this structure to take in for flying but are not homologues; monogenean host gut contents. If functional definitions are cercomers and eucestode scolices are adaptations used, the absence of an intestine precludes calling for holding onto the host but are not homo- the anterior invagination a mouth. In gyrocot- logues), or that parts of an homologous series ylideans and amphilinideans it is termed an an- have different functions (e.g., shrew forelimbs are terior invagination, and in eucestodes it is called adaptations for burrowing, bat forelimbs are ad- an apical sucker. But, as with existing terminol- aptations for flying, yet the two are homologues; ogy for posterior holdfast organs, the nomencla- monogenean cercomers are adaptations for hold- ture may be obscuring homologies. All of the ing onto the definitive host, eucestode cercomers Cercomeria possess an anterior invagination of are adaptations for penetrating the intermediate the ectoderm at some time in ontogeny, and in host, yet the two are homologues). It is undoubt- all groups except the Cestodaria it becomes a edly the structure of the central nervous system functional pharynx and mouth when it connects in eucestodes that allows the functioning of the with an intestine formed by a parenchymal cav- rostellum and four suckers, or of the four bo- ity. Given these observations, we postulate that thridia, but that does not mean that the structure the anterior invagination of cestodarians is a ves- is an adaptation for that function (i.e., that it was tigial mouth, and is therefore evidence of the ever the focus of selection). primitive presence of a complete digestive sys- Gould and Vrba (1982) proposed the term ex- tem. aptation to refer to structures that arose evolu- tionarily prior to the time they were coopted by Adaptive significance of characters natural selection for a particular function. This Under current evolutionary theory, percep- term was offered as a refinement of the more tions of the adaptive significance of characters general concept of "pre-adaptation." We suggest will affect hypotheses of their origin and their that there is another way to interpret such char- relationship to similar characters in other taxa. acter evolution. Some traits may evolve without In the cercomerian nervous system, the doubled any functional difference from the ancestral trait, commissures can be examined. Although it is i.e., the two would be functional equivalents. relatively easy to envisage plausible functional However, these new traits might allow other evo- explanations of what these structures do today, lutionary changes to take place, which in turn we suggest that there is no adaptive significance could have major evolutionary consequences. For in their evolution, that is, the reasons for their example, dalyelloid rhabdocoels kept alive in our appearance in the first place. Digeneans and as- laboratory adhere to a substrate by flattening the pidocotyleans have oral suckers and pharynges posterior end of their body and applying their that operate with a single cerebral commissure, adhesive secretions (an indication that posterior so the second commissure in monogeneans can- adhesive behavior was present in the outgroup not be an adaptation for oral function. The two before complex posterior adhesive organs evolved cerebral commissures are intimately involved in in the study group). With innervation capable of the development of the eucestode scolex, yet the functioning in such a manner, many structural doubled structure seems to have evolved before modifications of the posterior end could act as either the loss of the gut or the modifications of holdfasts. Similarly, the occurrence of doubled the mouth and anterior end into a holdfast organ. anterior nerve commissures would provide the We interpret this to imply that the doubled com- innervation basis for the subsequent evolution missures first evolved without affecting the func- of scolices. tioning of the body ends of ancestral cercomer- Our alternate interpretation can also be ap- omorphs. That is, the single commissures in plied to explanations of the evolution of tegu- trematodes and the doubled commissures in cer- mental structures in the Cercomeria. Finger-like comeromorphs are not only characters in the projections of the tegumental surface are espe- same homologous series, but functional equiv- cially well developed in those parasitic platy- alents (different structures performing the same helminths lacking a gut. This has led to expla- function) as well. nations that such structures evolved in order to

Copyright © 2011, The Helminthological Society of Washington OF WASHINGTON, VOLUME 52, NUMBER 1, JANUARY 1985 11 produce an alternative nutrient absorptive sur- Cercomeria (see also O'Grady, 1985). The life face. But these projections have been found, to cycle criteria for the nine taxa can be mapped one degree or another, in temnocephalideans, onto the cladogram in Figure 1 as direct life cycles trematodes, monogeneans, gyrocotylideans, am- involving invertebrate or vertebrate hosts, or philinideans, and eucestodes (e.g., Lee, 1966, complex life cycles involving both invertebrates 1972; Williams, 1981; Rohde and Georgi, 1983). and vertebrates. A direct cycle in an invertebrate They have, in fact, been reported to be present is concluded to be plesiomorphic because it is in all platyhelminths (Bedini and Papi, 1974). the life cycle of the most plesiomorphic taxa. Lee called all such projections microvilli. Other This conclusion is reached by both direct in- workers (see Jarecka et al., 1981, and references spection of the taxa and Farris optimization (Far- therein) have adopted Rothman's (1963) dis- ris, 1970) of the nodal values. The questions asked tinction between those projections containing are (1) in what manner did a direct cycle in a only cytoplasm (microvilli) and those with elec- vertebrate arise, and (2) are the complex cycles tron-dense caps as well (microtriches). The latter of digeneans and eucestodes (as well as some are especially prevalent around the scolex and species of amphilinideans) an example of con- neck in eucestode adults. Jarecka et al. (1981) vergent evolution? Certainly the life cycles in the demonstrated that although microtriches are Digenea and Eucestoda appear functionally sim- present on the necks and scolices of procercoids, ilar, with larval stages developing in inverte- cercoscolices, cysticercoids and cysticerci, there brates and the adults in vertebrates. In each case, are microvilli (sensu Rothman) on the "tails" of a dissemination of life stages results. Differences procercoids, cercoscolices and cysticercoids, the arise, however, when the morphological traits cyst wall of cysticercoids, and the bladder wall and sister-group relationships are examined. of cysticerci. An intermediate type of projection In eucestodes, it is the larvae that bear the occurs in gyrocotylideans (Lyons, 1969) and am- closest resemblance to the adults of the sister- philinideans (Rohde and Georgi, 1983). Devel- groups, the Amphilinidea, Gyrocotylidea, and opmental studies (e.g., MacKinnon and Burt, Monogenea. The armed cercomer, for example, 1984) on eucestodes indicate that microtriches synapomorphic for the Cercomeromorphae, is are derived from microvilli. Other studies sug- present in only the early developmental stages gest that rostellar hooks are derived from mi- of eucestodes. These include the hexacanth, pro- crotriches (Mount, 1970). cercoid, cysticercoid, and exceptionally the cys- Malmberg (1974) (see Fig. 7) noted that teg- ticercus (see Jarecka, 1975, Fig. 4). The charac- umental similarities could be the result of (1) ters representative of most eucestodes, such as a convergent evolution in digeneans and cercom- scolex and polyzoic body, are best interpreted as eromorphs, with parallel evolution within the adult characters added to the end of the ancestral Cercomeromorphae (a postulate that is more developmental sequence by terminal addition. parsimoniously interpreted to be a synapomor- Furthermore, whereas adults of other members phy); or (2) the result of a common ancestor for of the Cercomeromorphae develop in inverte- digeneans and cercomeromorphs. Rohde and brate hosts, the comparable (i.e., of the same Georgi (1983) considered the projections in am- phylogenetic origin) developmental stage in eu- philinideans to be an adaptation to food absorp- cestodes—the larva—develops in a more recent- tion and of "no great phylogenetic significance" ly acquired invertebrate host. Because this rel- (but see Rohde, 1980). We suggest that not only atively newer host harbors an intermediate are digitiform tegumental projections in the Cer- developmental stage, its means of acquisition is comeria homologous, but that their functional one of intercalation, or nonterminal addition to utility in cestodarians is a consequence of their the ancestral direct life cycle in vertebrates. Rohde primitive occurrence in platyhelminths with a and Georgi's (1983) discovery that Austram- gut, which allowed the survival of descendants philina elongata develops through an interme- lacking a gut. No postulates of convergent or diate arthropod host indicates that the life cycle parallel evolution are necessary. changes hypothesized above may actually be synapomorphic for the Cestoidea. The study by The evolution of life cycles Janicki (1928) on Amphilina foliacea indicates Our classification allows some hypotheses to this as well (see Rohde and Georgi, 1983). be formed about the origin of life cycles in the The data in Figure 1 suggest that the evolution

Copyright © 2011, The Helminthological Society of Washington 12 PROCEEDINGS OF THE HELMINTHOLOGICAL SOCIETY of the life cycle in digeneans occurred in a reverse have two meanings. One is functional, the other manner to that of eucestodes. Adult digeneans is historical. Our study has shown that the func- most closely resemble the adults of their sister- tional usage may obscure evolutionary differ- groups, the Aspidocotylea, Udonellidea, and ences. The historical usage, however, does not Temnocephalidea. For example, these groups obscure the functionality of the hosts. Thus, we possess an unarmed cercomer (as defined ear- would call the plesiomorphic host for digeneans lier), synapomorphic for the Cercomeria. The (the mollusc) the functional intermediate host and only nonadult digenean stages to have this struc- the plesiomorphic host for eucestodes (the ver- ture are the cercariae and metacercariae. The lar- tebrate) the functional definitive host. This re- val stages associated with the complex life cycle— tains information about both the history and the miracidia, sporocysts and rediae—lack this char- function of the taxa involved. acter and do not appear to have comparable developmental stages in the most closely related taxa. This suggests an intercalation of develop- Comparisons with other classifications mental stages by nonterminal addition. With re- Previous classifications for these platyhel- spect to the hosts, adults of dalyelloids, temno- minths include those by Spengel (1905), Janicki cephalideans, udonellideans, and aspidocotyleans (1920), Fuhrmann (1928, 1931), Bychowsky develop in association with invertebrates (some (1937, 1957; see also Beklemishev, 1969; Du- aspidocotyleans occur with vertebrates). In di- binina, 1974), Llewellyn (1965), Price (1967), and geneans, this plesiomorphic host group contains Malmberg (1974). For reasons we will discuss the relatively new larval stages, whereas the ple- later, it is difficult to make comparisons with the siomorphic developmental stage, the adult, is work of Spengel and Janicki. The classifications found in the more recently acquired vertebrate of the remaining authors are presented in Figures host. This acquisition is best interpreted as a 4-9, rendered in a form comparable with our terminal addition to the ancestral life cycle. results. Numbers accompanying the slash marks Our analysis therefore suggests that the com- refer to the characters from our cladogram in plex life cycles of digeneans and eucestodes Figure 1, optimized with Farris' (1970) method evolved by different means. Eucestodes exhibit in order to obtain the best fit possible. In some terminal addition of ontogenetic stages and non- cases this resulted in hypotheses of evolution and terminal addition of an invertebrate host. Di- inclusion of groups which the original author did geneans exhibit nonterminal addition of onto- not suggest, but with which a better fit to the genetic stages and terminal addition of a stated hypotheses could be obtained. We have vertebrate host. Vertebrates also appear to have thus attempted to ensure that differences result been colonized by the ancestral cercomero- from the classifications themselves and not from morphs. differential treatment of the data. At the same A number of general conclusions are implied time we recognize that by adding data to an ear- by this hypothesis. First, as with morphological lier classification we are not examining that pro- characters, Hennigian analysis allows life cycle posal exactly as its originator formulated it. For components to be treated as composites of ple- systematics, however, these concerns are siomorphic and apomorphic states, inherited and superseded by the necessity to treat every clas- modified at different times. It is not necessary to sification as an hypothesis subject to testing with search for the "archetypal" digenean or euces- new information. tode. Second, the term "complex life cycle" is Classifications can be assessed in two ways. too broad to be applied to both the Digenea and The first is the efficiency with which the data are Eucestoda when evolutionary conclusions are to described. This amounts to postulating the least be drawn from the comparison. Third, the terms amount of homoplasy or minimizing the type 2 "intermediate host" and "definitive host" may ad hoc assumptions discussed earlier. Homopla-

Figures 4-9. Cladograms depicting previous classifications of the parasitic platyhelminths, and the postulations of character evolution they require. Numbered slash marks refer to characters in Figure 1, "X" indicates an evolutionary reversal. Branches ending in a node (rather than a terminal taxon) that bear no slash marks propose a grouping that has no character support.

Figures 4-6. 4. Fuhrmann (1928, 1931). 5. Bychowsky (1937, 1957). 6. Llewellyn (1965).

8 PRICE (1967)

9 PRICE (1967)

Figures 7-9. 7. Malmberg (1974). 8. Price (1967) (phylogenetic tree). 9. Price (1967) (classification).

Copyright © 2011, The Helminthological Society of Washington OF WASHINGTON, VOLUME 52, NUMBER 1, JANUARY 1985 15 sy shows up in two ways in the classifications. Table 2. Comparisons of previous classifications of These are: (1) postulations of parallel or conver- parasitic platyhelminths with the present study. Three criteria are used to evaluate the fit of the classifications gent evolution, tabulated by counting the num- to the data set of 39 characters in Figure 1: the effi- ber of duplicated numbers on the trees; and (2) ciency of character representation (number of steps), postulations of evolutionary reversals, or the sec- the degree to which the classification is supported by ondary loss of a trait, tabulated by counting the ambiguous characters [the consistency index, or CI, of number of characters denoted by an "X" on the Kluge and Farris (1969)—see text], and the number of unsupported groupings (number of empty branches). A trees. The fewer of either of these, the better the good fit to the data is indicated by a high CI and a low summary of the data. Efficiency can also be mea- number of steps and empty branches. Two of the four sured by the consistency index (CI) of Kluge and empty branches in Llewellyn's classification come from Farris (1969). This is calculated by dividing the the postulation of paraphyly in the Eucestoda. minimum numbers of steps needed to represent the data (i.e., each character evolves once) by the No. of No. of empty actual number of steps required to support a par- Author steps CI branches ticular classification. The closer the CI is to 1.0, the better the fit of the data to the tree. Our Fuhrmann, 1928, 1931 46 0.85 2 Bychowsky, 1937, 1957 55 0.71 2 classification, for example, has a minimum num- Llewellyn, 1965 59 0.66 4 ber of 39 steps (39 characters), and an actual Price, 1967 (tree) 43 0.91 0 number of 41 steps, giving a CI of 39/41, or 0.95. Price, 1967 The departure from maximum efficiency comes (classification) 50 0.78 2 Malmberg, 1974 50 0.78 2 from the postulation of two cases of homoplasy: Present study 41 0.95 0 the loss of a copulatory stylet in the Trematoda and Cestodaria, and the development of a bifurcate gut in the Digenea and Monogenea. logenetic tree may be suspected of postulating The second way in which classifications can artificial, nonevolutionary groups. Parentheti- be assessed is by examining them for groupings cally, the extent to which such artificial groups of taxa for which there are no distinguishing have been defined by functional criteria is the characters; that is, the branch uniting those taxa extent to which departures from a strictly phy- on the tree has no character support (slash marks logenetic classification have been justified by ref- in Figs. 4-9). This can result from a type 1 ad erence to adaptive scenarios. hoc assumption, discussed earlier, in which taxa Table 2 summarizes the step length, CI, and of secondary importance in a study group are empty branch statistics for the classifications we grouped together by default because they do not examined. Those by Fuhrmann (1928, 1931), possess a structure that has been used to highlight Llewellyn (1965), Bychowsky (1937, 1957), and the taxa of primary interest (see Price, 1967, on Malmberg (1974) appear to have made unnec- the Acercomeromorphae). An unsupported essary postulates of convergent evolution. Such grouping can also be included in a classification empirically unjustified departures from the data for heuristic, rather than empirical, reasons (see are difficult to detect when they can be consid- Inglis, 1983, on the Aschelmintha). Clearly, it is ered to support perceptions of the adaptive plas- not helpful for taxonomic groupings to be devoid ticity (i.e., capacity for homoplasious evolution) of diagnostic features; nevertheless, a number of of parasites (e.g., Price, 1980). higher taxa have been defined by what they are The classification by Malmberg (1974) is es- not. For example, "reptiles" are amniotes with- pecially interesting because it is the only explicit, out bird-like or mammal-like features; the rep- rather than anecdotal, attempt to support the tiles as a group have no distinguishing traits. In- hypothesis that cercomeromorphs are more spection of a phylogenetic tree for the Tetrapoda closely related to acoel turbellarians than they shows that reptiles are not an evolutionary group are to trematodes. A review of this hypothesis is (Wiley, 1981b), that is, not all reptiles are each beyond the scope of this paper; it has gained others' closest relatives. Birds and crocodiles (the recent support from Malmberg (1974), Logachev Archosauria) are more closely related to each and Sokolova (1975), Freeman (1982). and other than they are to any other group of organ- Mackiewicz (1981,1982). Its basic argument cites isms. Portions of classifications, therefore, that the absence of a digestive system in cestodarians have empty branches when represented as a phy- (sensu nobis). Malmberg (1974) stated:

"If the absence of mouth, pharynx and intes- Turbellaria, considered trematodes and cerco- tine in the ontogeny of cestodes, the amphi- meromorphs to be derived from rhabdocoels. linideans and the gyrocotylideans implies that Also, platyhelminths may not be primitively gut- these body parts were never evolved here, then less at all. The Cnidaria, Ctenophora, Gnathos- these groups cannot have originated from tomulida, and Nemertinea, as well as most pla- rhabdocoelan creatures." tyhelminths, have intestines. Even some acoels, such as Nemertoderma, have intestines (see Kar- Freeman (1982) gave the most succinct state- ment of the evidence: ling, 1967). If the cercomeromorphs are grouped with acoels, on the basis of their gutless condition "Incidentally not even the suggestion of en- alone, there are still some gutless platyhelminths doderm let alone a tube-like gut in the ontog- excluded from that grouping. Some dalyelloid eny of any present-day cestode (e.g. see Lo- rhabdocoel (Fecampiidae) parasites of crusta- gachev and Sokolova, 1975), suggests that it ceans, such as Kronborgia amphipodicola, have never had such a gut.. .." no mouth, pharynx, or intestine (Christensen and Kanneworff, 1964). Levenseniella (Monarrhe- We are dissatisfied with this type of reasoning. nos) capitanea is a microphallid digenean that It is not possible to decide, by reference to a lacks a pharynx and has only a few fibrous in- character's absence alone, whether that condi- testinal tissues (Overstreet and Perry, 1972). tion is due to primitive absence or subsequent Clearly, one does not classify these taxa with the loss (i.e., whether it is a primitive or derived cercomeromorphs and acoels. Such a decision character). Other characters must be examined. would require the postulation of a very large We suggest that the perceived necessity for all amount of homoplasious evolution of rhabdo- ancestral characters to remain in a descendant's coel-like and digenean-like traits. Our hypothesis ontogeny comes from perceptions of the ubiquity of the phylogeny of the cercomeria is an attempt of recapitulatory development and evolution by to maintain consistency in the method of infer- nothing but terminal addition of developmental ence of genealogical relationships, namely, to ap- stages. The necessity of corroborative characters, ply parsimony considerations to character anal- however, may be a moot point in this case, in ysis at every level of generality in the study group. light of our earlier suggestion that the anterior We do not believe that previous classifications invaginations of cestodarians may be vestigial can offer adequate justification for the departures mouths and pharynges. from parsimony that some of their groupings re- Malmberg (1974) recognized the problems that quire. secondary loss and expectations of recapitulatory The classifications by Spengel (1905) and Ja- development could create for phylogenetic stud- nicki (1920) are too incomplete to be fully com- ies, but his proposed solution—the primitive ab- pared with ours. We have attempted an analysis sence of a gut in the cercomeromorphs—creates of Fuhrmann's work even though it is only slight- problems of another sort. Let us assume for the ly more complete. Spengel and Janicki consid- sake of argument that anterior invaginations in ered four taxa: rhabdocoels, digeneans, mono- cestodarians are convergent traits having nothing geneans, and eucestodes. Spengel's placement of to do with a gut. There are still 38 other char- these four corresponds not only to ours, but also acters to be explained. Confirmation of the prim- to almost every other classification we have con- itively gutless nature of cestodarians could come sidered. Our analysis of Bychowsky's classifica- from those other characters analyzed indepen- tion, which is an expansion of Spengel's, shows dently. But, as we have shown, the most efficient that subsequent taxa and data were interpreted and least ad hoc interpretation of those other in a suboptimal manner. characters supports relationships with the trem- Janicki's classification was connected with his atodes, udonellideans, temnocephalideans, and cercomer theory. He proposed that monoge- rhabdocoels—not the acoels. Malmberg's scheme neans arose from rhabdocoels, then gave rise to (Fig. 7) requires 9 cases of convergent evolution digeneans, which gave rise to eucestodes. This to account for the structural similarities among interpretation was supported by the conclusion the parasitic platyhelminths that we include in that (1) the cercomer of procercoids, the cyst and the Cercomeria. tail of cysticercoids, and the bladder of cysticerci Karling (1974), in a cladistic analysis of the are homologues; (2) these are homologous with

Copyright © 2011, The Helminthological Society of Washington OF WASHINGTON, VOLUME 52, NUMBER 1, JANUARY 1985 17 the opisthaptor of monogeneans; and (3) these phae. This grouping may have some "pragmatic" are homologous with the tail of digenean cer- utility for the recognition of flatworms whose cariae. Janicki termed these cercomer-homo- larvae do not possess an armed cercomer, but logue-bearing taxa the "Cercomeromorphae." given the phylogenetic tree whose relationships Bychowsky removed the Digenea from this the classification is intended to communicate, this grouping and retained the term to denote those trait cannot be considered to be an indicator of platyhelminths with an armed cercomer. We have monophyly, as can an armed cercomer in the used the term in Bychowsky's sense but altered Cercomeromorphae. the inferred relationships of the taxa involved (cf. Figs. 1 and 5). Jarecka et al. (1981) drew from Conclusions earlier work by Jarecka (e.g., 1975) and suggested Based primarily upon characters previously that the presence of microvilli on the posterior used to classify parasitic platyhelminths, we have body expansions of eucestode larvae provided derived a classification that best fits the data. It support for Janicki's postulates of cercomer ho- embodies elements common to most earlier pro- mology with the exception of the Digenea. Free- posals, and some aspects of each previous clas- man (1973) noted the importance of distinguish- sification agree with ours. In total, however, nei- ing between the cercomer and the midbody of ther our classification nor its justifications have eucestode larvae when examining such homol- been presented before. Some characters used pre- ogy. We have offered evidence for extending the viously are concluded to be too ambiguous for series of cercomer homologues first to the trem- phylogenetic inference. Others, notably genital atodes (on the basis of the ventral adhesive disk, pores, posterior adhesive organs, and anterior rather than the cercarial tail)—to form the Cer- invaginations are each united into homologous comeridea, and then to the Temnocephalidea— series. Revised terminology is proposed for the to form the Cercomeria. latter two character series. The sequence of origin A third problematic classification is that by for some traits previously taken to be of special Price (1967), who presented both a phylogenetic taxonomic importance is considered to be in- tree and a classification (see Table 2). The tree consistent with explanations that they evolved is not resolved at its base and is ambiguous as as adaptations for a particular function. The clas- to the monophyly of the Cercomeria. Price's dis- sification indicates that the plesiomorphic life cussion, however, suggests that monophyly was cycle of the Cercomeria is direct and in an in- being postulated. If we assume this, his tree can vertebrate. Vertebrates appear to have been col- be represented by Figure 8. It differs from ours onized, twice: by the Digenea and by the ancestral in two ways. First, it places Udonella in a tri- cercomeromorphs. Invertebrates, especially ar- chotomy (three branches arising from one node) thropods, appear to have been recolonized by with trematodes and cercomeromorphs. Second, the Eucestoda or possibly the Cestoidea. The it does not distinguish between dalyelloid and functional terms of "complex life cycle," "inter- temnocephalid turbellarians. This creates a mediate host," and "definitive host" obscure problem for the mapping of characters 7, 8, 9, phylog;enetic relationships. The postulate that and 10 (see Fig. 1). We think it best, and fairest, cestodarians and monogeneans are descended to minimize the number of steps by maintaining from acoel turbellarians is rejected on the basis Price's Turbellaria category, and simply note that that it derives its support from unjustified as- not all of its members possess characters 7, 8, sumptions of convergent evolution. The present and 9, or 10. New problems arise when the phy- study also offers new character analyses that con- logenetic inferences of Price's classification are flict with this hypothesis. The evidence presented made explicit by converting it into a tree (Fig. herein links monogeneans and cestodarians with 9). The result gives a poorer fit to the data than trematodes and these three groups with dalyel- does his original phylogenetic tree. The classifi- loid rhabdocoels. cation is therefore inconsistent with the original tree (see Wiley, 198 Ib). This means that the evo- Acknowledgments lutionary relationships it implies are not the same Funds for this study were provided through a as those given by the tree from which it was grant (A7696) to DRB from the Natural Sciences constructed. This is caused by the creation of a and Engineering Research Council of Canada type 1 ad hoc grouping—the Acercomeromor- (NSERC). ROG thanks NSERC for postgraduate

Copyright © 2011, The Helminthological Society of Washington 18 PROCEEDINGS OF THE HELMINTHOLOGICAL SOCIETY support. DRG thanks the University of British minthes: Cercomeria) with comments on their Columbia for fellowship support. Susan Bandoni adaptive radiation. Can. J. Zool. (In Press) offered useful comments on character analysis. —, and R. M. Overstreet. 1978. The family Lio- lopidae (Digenea) including a new genus and two Ms. Maggie Hampong prepared the diagrams. new species from crocodilians. Int. J. Parasitol. 8: We have profited from discussions with various 267-273. colleagues and a number of referees. Errors of -, T. B. Thorson, and M. A. Mayes. 1981b. interpretation rest solely with us. Fresh-water stingrays (Potamotrygonidae) and their helminth parasites: testing hypotheses of evolution and coevolution. Pages 147-178 in V. A. Funk and D. R. Brooks, eds. Advances in Cladistics. Literature Cited Vol. 1. New York Botanical Garden, New York. 250 pp. Alicata, J. E., and E. Chang. 1939. The life history Bullock, T. H., and G. A. Horridge. 1965. Structure of Hymenolepis exigua. J. Parasitol. 25:121-129. and Function in the Nervous System of Inverte- Allison, F. R. 1980. Sensory receptors of the rosette brates. W. H. Freeman and Co., San Francisco. organ of Gyrocotyle rugosa. Int. J. Parasitol. 10: 1719pp. 341-353. Bychowsky, B. E. 1937. [Ontogenesis and phyloge- Bedini, C., and F. Papi. 1974. Fine structure of the netic interrelations of parasitic flatworms.] News turbellarian epidermis. Pages 108-147 />zN. J. Riser of the Acad. Sci., USSR. Dept. Math. Nat. Sci. 4: and M. P. Morse, eds. Biology of the Turbellaria. 1354-1383. McGraw-Hill, New York. 530 pp. . 1957. Monogenetic trematodes—their sys- Beklemishev, W. N. 1969. Principles of the Com- tematics and phylogeny. American ed. (1961), W. parative Anatomy of Invertebrates. (English trans- J. Hargis, Jr., ed. Graphic Arts Press, Inc., Wash- lation of 1964 Russian edition, by Z. Kabata.) ington, D.C. 627 pp. Oliver and Boyd, University Press, Aberdeen. Vol. Christensen, A. M., and B. Kanneworff. 1964. Kron- 1, 490 pp.; Vol. 2, 529 pp. borgia amphipodicola gen. et sp. nov., a dioecious Brooks, D. R. 1977. Evolutionary history of some turbellarian parasitizing ampeliscid amphipods. plagiorchid trematodes of anurans. Syst. Zool. 26: Ophelia 1:147-166. 277-289. Crisci, J. V., and T. F. Stuessy. 1980. Determining . 1978a. Systematic status of proteocephalid primitive character states for phylogenetic recon- cestodes from reptiles and amphibians in North structions. Syst. Bot. 5:112-135. America with descriptions of three new species. Douglas, L. T. 1963. The development of organ sys- Proc. Helminthol. Soc. Wash. 45:1-28. tems in nematotaeniid cestodes. III. Gametogen- . 1978b. Evolutionary history of the cestode esis and embryonic development in Baerietta diana order Proteocephalidea. Syst. Zool. 27:312-323. and Distoichometra kozloffi. J. Parasitol. 49:530- . 198la. Revision of the Acanthostominae 558. (Digenea: Cryptogonimidae). Zool. J. Linn. Soc. Dubinina, M. N. 1974. [The state and immediate 70:313-382. goals of taxonomy of cestodes (Cestoidea Rud., . 1981b. Classifications as languages of em- 1808).] Parazitologiya 8:281-289. pirical comparative biology. Pages 61-70 in V. A. Estabrook, G. 1971. Some information theoretic op- Funk and D. R. Brooks, eds. Advances in Cladis- timality criteria for general classification. Math. tics. Vol. 1. New York Botanical Garden, New Geol. 3:203-207. York. 250 pp. . 1978. Some concepts for the estimation of . 1982. Higher level classification of parasitic evolutionary relationships in systematic botany. platyhelminths and fundamentals of cestode clas- Syst. Bot. 3:146-158. sification. Pages 189-193 in D. F. Mettrick and S. Fairweather, I., and L. T. Threadgold. 1983. Hy- S. Desser, eds. Parasites—Their World and Ours. menolepis nana: the fine structure of the adult ner- Elsevier Biomedical, Amsterdam. 465 pp. vous system. Parasitology 86:89-103. , and J. N. Caira. 1982. Atrophecaecum lo- Farris, J. S. 1970. Methods for computing Wagner bacetabulare n. sp. (Digenea: Cryptogonimidae: trees. Syst. Zool. 19:83-92. Acanthostominae) with discussion of the generic . 1979. The information content of the phy- status of Paracanthostomum Fischthal and Kuntz, logenetic system. Syst. Zool. 28:483-519. 1965 and Ateuchocephala Coil and Kuntz, 1960. -. 1982. Outgroups and parsimony. Syst. Zool. Proc. Biol. Soc. Wash. 95:223-231. 31:328-334. -, T. R. Platt, and M. H. Pritchard. Freeman, R. S. 1973. Ontogeny of cestodes and its 1984. Principles and Methods of Phylogenetic bearing on their phylogeny and systematics. Pages Systematics: A Cladistics Workbook. Spec. Publ. 481-557 in B. Dawes, ed. Advances in Parasitol- No. 12 Mus. Nat. Hist., Univ. Kansas. 92 pp. ogy. Vol. 11. Academic Press, London. 774 pp. —, M. A. Mayes, and T. B. Thorsen. 198la. . 1982. How did tapeworms get that way? Bull. Systematic review of cestodes infecting freshwater Can. Soc. Zool. 13:5-8. stingrays (Chondrichthyes: Potamotrygonidae) in- Fuhrmann, O. 1928. Zweite Klasse des Cladus Platy- cluding four new species from Venezuela. Proc. helminthes: Trematoda. Pages 1-140 in W. Ku- Helminthol. Soc. Wash. 48:43-64. kenthal and T. Krumbach, eds. Handbuch der —, R. T. O'Grady, and D. R. Glen. 1985. Phy- Zoologie. Vol. 2. Walter de Gruyter, Berlin. 1,392 logenetic analysis of the Digenea (Platyhel- pp.

. 1931. Dritte Klasse des Cladus Platyhel- Lynch, J. E. 1945. Redescription of the species of minthes: Cestoidea. Pages 141-416 in W. Kuken- Gyrocotyle from the ratfish, Hydrolagus colliei (Lay thal and T. Krumbach, eds. Handbuch der Zoo- and Bennett), with notes on the morphology and logie. Vol. 2. Walter de Gruyter, Berlin. 1,392 pp. taixonomy of the genus. J. Parasitol. 31:418-446. Gould, S. J., and E. Vrba. 1982. Exaptation—a miss- Lyons, K. M. 1969. The fine structure of the body ing term in the science of form. Paleobiology 8:4- wall of Gyrocotyle urna. Z. Parasitenkd. 33:95- 15. 109. Hennig, W. 1950. Grundzuge einer Theorie der Phy- Mackiewicz, J. S. 1981. Caryophyllidea (Cestoidea): logenetischen Systematik. Deutsche Zentralverlag, evolution and classification. Pages 139-206 in W. Berlin. 370 pp. H. R. Lumsden, R. Muller, and J. R. Baker, eds. . 1966. Phylogenetic Systematics. Univ. Illi- Advances in Parasitology. Vol. 19. Academic Press, nois Press, Urbana. 263 pp. London. 225 pp. Hull, D. L. 1979. The limits of cladism. Syst. Zool. . 1982. Parasitic platyhelminth evolution and 28:416-440. systematics: perspectives and advances since ICO- Hyman,L.H. 1951. The Invertebrates. Vol. II. Platy- PA IV, 1978. Pages 179-188 in D. F. Mettrick helminthes and Rhynchocoela. McGraw-Hill, New and S. S. Desser, eds. Parasites—Their World and York. 550 pp. Ours. Elsevier Biomedical, Amsterdam. 465 pp. Inglis, W. G. 1983. An outline classification of the MacKinnon, B. M., and M. D. B. Burt. 1984. The Phylum Nematoda. Austral. J. Zool. 31:243-255. development of the tegument and cercomer of the Jamieson, B. G. M., and L. M. Daddow. 1982. The polycephalic larvae (cercoscolices) of Parictero- ultrastructure of the spermatozoon ofNeochasmus taenia paradoxa (Rudolphi, 1802) (Cestoda: Di- sp. (Cryptogonimidae, Digenea, Trematoda) and lepididae) at the ultrastructural level. Parasitology its phylogenetic significance. Int. J. Parasitol. 12: 88:117-130. 547-559. Malmberg, G. 1974. On the larval protonephridial Janicki, C. 1920. Grundlinien einer "Cercomer"— system of Gyrocotyle and the evolution of the Cer- Theorie zur Morphologic der Trematoden und comeromorphae (Platyhelmithes). Zool. Scripta 3: Cestoden. Fortschr. Zschokke. 30:1-22. 65-82. . 1928. Die Lebensgeschichte von Amphilina Mickevich, M. F. 1982. Transformation series anal- foliacea G. Wagen., Parasiten des Wolga-Sterlets, ysis. Syst. Zool. 31:461-478. nach Beobachtungen und Experimenten. Raboty Mount, P.M. 1970. Histogenesis of the rostellar hooks Volzkiaia Biologicheskaia Stantsiia, Sartov 10, no. of Taenia crassiceps (Zeder, 1800) (Cestoda). J. 3:1-134. Parasitol. 56:947-961. Jarecka, L. 1975. Ontogeny and evolution of ces- O'Gra.dy, R. T. 1985. Ontogenetic sequences and the todes. Acta Parasitol. Pol. 23:93-114. phylogenetics of parasitic flatworm life cycles. Cla- , W. Michajlow, and M. D. B. Burt. 1981. distics. (In Press) Comparative ultrastructure of cestode larvae and Overstreet, R. M., and H. M. Perry. 1972. A new Janicki's cercomer theory. Acta Parasitol. Pol. 28: microphallid trematode from the blue crab in the 65-72. northern Gulf of Mexico. Trans. Am. Microsc. Karling, T. G. 1967. Zur Frage von dem systemati- Soc. 91:436-440. schen Wert der Kategorien Archoophora und Price, C. E. 1967. The phylum Platyhelminthes: a Neoophora (Turbellaria). Commentat. Biol. Soc. revised classification. Riv. Parassitol. 28:249-260. Sci. Fenn. 30:1-11. Price, P. W. 1980. Evolutionary Biology of Parasites. . 1974. On the anatomy and affinities of the Princeton Univ. Press, Princeton, New Jersey. turbellarian orders. Pages 1-16 in N. W. Riser and Rees, G. 1940. Germ cell cycle of the digenetic trem- M. P. Morse, eds. Biology of the Turbellaria. atode Parorchis: Part II. Structure of the miraci- McGraw-Hill, New York. 530 pp. duum and germinal development in the larval Kluge, A. G., and J. S. Farris. 1969. Quantitative stages. J. Parasitol. 32:372-391. phyletics and the evolution of anurans. Syst. Zool. Rohde, K. 1968. Das Nervensystem der Gattung 18:1-32. Polystomoides'WaTd, 1917 (Monogenea). Z. Mor- Lee, D. L. 1966. The structure and composition of phol. Okol. Tiere 62:58-76. the helminth cuticle. Pages 187-254, in B. Dawes, ed. Advances in Parasitology. Vol. 4. Academic . 1971. Phylogenetic origin of trematodes. Par- Press, London. 412 pp. asitol. Schriftenr. 21:17-27. . 1972. The structure of the helminth cuticle. . 1975. Fine structure of the Monogenea, es- Pages 347-379 in B. Dawes, ed. Advances in Par- pecially Polystomoides Ward. Pages 1-33 in B. asitology. Vol. 10. Academic Press, London. 408 Dawes, ed. Advances in Parasitology. Vol. 13. Ac- pp. ademic Press, London. 277 pp. Llewellyn, J. 1965. The evolution of parasitic platy- . 1980. Some aspects of the ultrastructure of helminths. Pages 47-78 in A. E. R. Taylor, ed. Gotocotyla secunda and Hexostoma euthynni. An- Evolution of Parasites. 3rd Symposium of the Brit- gew. Parasitol. 21:32-48. ish Society for Parasitology. Blackwell Publ., Ox- -, and M. Georgi. 1983. Structure and devel- ford. opment ofAustramphilina elongata Johnston, 1931 Logachev, E. D., and L. A. Sokolova. 1975. [Some (Cestodaria: Amphilinidea). Int. J. Parasitol. 13: aspects of early development of cestodes related 273-287. to their phylogeny.] Trudy Nauchno-Issled Inst. Rothnian, A. H. 1963. Electron microscopic studies pri Tomskom Universitete 5:145-152. oif tapeworms: the surface structures of Hymeno-

lepis diminuta (Rudolphi, 1819)Blanchard, 1891. group comparison method of character analysis. Trans. Am. Microsc. Soc. 82:22-30. Syst. Zool. 30:1-11. Sinitsin, D. F. 1911. Partenogeneticheskoe pokilenie Watson, E. E. 1911. The genus Gyrocotyle, and its trematod i ego potomstvo v chernomorskikh mol- significance for problems of cestode structure and liuskakh. Mem. Acad. Imp. Sc. St.-Petersb., Cl. phylogeny. Univ. Calif. Publ., Zool. 6:353-468. Phys.-Math. 8. s. 30:1-205. Wiley, E.G. 198 la. Phylogenetics: The Theory and Spengel, J. W. 1905. Die Monozootie der Cestoden. Practice of Phylogenetic Systematics. Wiley-In- Z. Wiss. Zool. 82:252-287. terscience, New York. 439 pp. Stevens, P. 1980. Evolutionary polarity of character . 1981b. Convex groups and consistent clas- states. Ann. Rev. Ecol. Syst. 11:333-358. sifications. Syst. Hot. 6:346-358. Tower, W. L. 1900. Nervous system of the cestode Williams, J. B. 1981. Classification of the Temno- Monezia expansa. Zool. Jahrb. Anat. 13:359-384. cephaloidea (Platyhelminthes). J. Nat. Hist. 15: Watrous, L. E., and Q. D. Wheeler. 1981. The out- 277-299.

INTERNATIONAL CODE OF ZOOLOGICAL NOMENCLATURE New Edition The International Commission on Zoological Nomenclature has indicated that the 3rd edition of the International Code of Zoological Nomenclature will soon be available. The book consists of parallel English and French texts, totaling about 320 pages (216 mm x 318 mm). It will be for sale for 15 pounds, plus 1.50 pounds for postage and handling, through the Publications Sale, British Museum (Natural History), Cromwell Road, London SW7 5BD England. Individuals and organizations may also purchase it for $21.50, postage and handling included, through the: American Association for Zoological Nomenclature Room W-l 15, National Museum of Natural History Washington, D.C. 20560 In order that the Code will be widely available, a prepublication price of $ 18.75, postage and handling included, for individuals is in effect until March 31, 1985. Delivery of the Code will be in spring 1985. Checks drawn on a United States bank or International Money Orders should accompany all U.S. orders! Please make checks payable to American As- sociation for Zoological Nomenclature.