Systematic Parasitology 37: 81±92, 1997. 81

c 1997 Kluwer Academic Publishers. Printed in the Netherlands.

An alternative classi®cation of trypanorhynch cestodes considering the tentacular armature as being of limited importance

Harry W. Palm Marine Pathology Group, Department of Fisheries Biology, Institut furÈ Meereskunde an der UniversitatÈ Kiel, DusternbrookerÈ Weg 20, D- 24105 Kiel, Germany

Accepted for publication 22nd November, 1996

Abstract

Based on a light- and scanning electron microscopical study of scolex morphology, tentacles and surface structures of 31 trypanorhynch species, an alternative classi®cation of the trypanorhynch cestodes, adults, plerocerci and postlarvae, is presented. The arrangement of the tentacular armature is no longer used as a distinguishing feature for four different superfamilies. Instead, the presence or absence of ciliated pits and prebulbular organs is used to de®ne three superfamilies: Tentacularioidea Poche, 1926; Otobothrioidea Dollfus, 1942; and Eutetrarhynchoidea Guiart, 1927. A total of 12 families are de®ned by the characters: the presence/absence of blastocysts, the number of bothridia and the reduction of the rhyncheal apparatus, together with a new character, complete rows of ten- tacular hooks (homeoacanth and heteroacanth typica) versus rows of hooks partly reduced (heteroacanth atypica and poeciloacanth). Of the 19 families previously accepted, 10 are retained (Eutetrarhynchidae, Gilquiniidae, Lacistorhynchidae, Mixodigmatidae, Otobothriidae, Paranybeliniidae, Pterobothriidae, Shirleyrhynchidae, Sphyri- ocephalidae and Tentaculariidae, all sensu nov.); one family is reinstated (i.e. Aporhynchidae Poche, 1926 sensu nov.) and a new one is added (i.e. Pseudotobothriidae n. fam.). Advantages of this alternative classi®cation of trypanorhynch cestodes are: (i) the resolution of incongruities and questions caused by the use of the tentacular armature to distinguish superfamilies; (ii) the criteria for the establishment of higher taxa, superfamilies and families are clearly de®ned; (iii) with the ®ndings of new species with different character combinations, this system can be enlarged up to 4 superfamilies and 48 families without loosing its stability; and (iv) all existing genera are easily re-assigned to the superfamilies and families.

Introduction identi®cation of many of the c. 200±250 generally accepted species. Wardle& McLeod (1952) stated (and Trypanorhynch cestodes are characterised by the pos- were recently quoted by Campbell & Beveridge, 1994) session of a scolex bearing four eversible tentacles that, taxonomically, the Trypanorhyncha are complex armed with hooks and two or four bothridia, enabling and considered to be the most chaotic and confused these cestodes to move along and attach to the diges- of the tapeworm groups. Carvajal & Campbell (1975) tive tract of their ®nal hosts which are elasmobranchs. saw the general problems within trypanorhynch classi- In their life-cycles, these cestodes use invertebrates as ®cation as follows: ªThe problems of de®ning natural ®rst intermediate hosts, with teleosts and some inver- af®nities of trypanorhynchs are perplexing because of tebrates as second intermediate or paratenic hosts. The our grossly insuf®cient knowledge concerning life his- presence of the metacestodes (postlarvae and plerocer- tories, larval development, host speci®city, etc., and ci) in the musculature of ®shes decrease the commer- because the signi®cance of morphometric characters

cial value of affected stocks (Deardorff et al., 1984). must be weighed by each investigator. [:::]Asmore Despite their world-wide distribution and impor- species are discovered, it appears that the current tance for commercial ®sheries,trypanorhynch cestodes method of trypanorhynch classi®cation is becoming are still a relatively poorly studied group. One reason increasingly unstable because of the variability inher- for this could be their classi®cation which complicates ent in organismsº. Furthermore, within existing try- 82 panorhynch classi®cations, some characters are used Materials and methods as major distinguishing features for different taxa at different systematic levels at one and the same time. Light- and scanning electron microscopical studies of There are only a few speci®c character combinations morphology, tentacles and surface structures of the which are characteristic for a single taxon. scoleces of trypanorhynch cestodes belonging to 18 Campbell & Beveridge (1994) published a new genera (a total of 31 species) resulted in the descrip- classi®cation of trypanorhynch cestodes, in which the tion of four new species (Grillotia kovalevae, Poeciloa- tentacular armature of the postlarvae, plerocerci and canthum oweni, Pseudogrillotia zerbiae, Pseudolacis- adults was used as a major character for four super- torhynchus noodti) and the ®rst SEM studies of 15 families: Homeacanthoidea Dollfus, 1942, Heteracan- previously known species (see Palm, 1995). thoidea Dollfus, 1942, Otobothrioidea Dollfus, 1942 Fresh postlarvae, plerocerci and adults were col- and Poecilacanthoidea Dollfus, 1942. Additionally, in lected from teleosts and elasmobranchs along the their identi®cation key, Campbell & Beveridge (1994) coasts of Brazil, Nigeria and the southern (Gulf of used bothridial form, tentacle length, position of the Mexico) coast of Mississippi, USA. Preserved postlar- pars bothridialis relative to that of the pars bulbosa, vae and plerocerci were obtained from collections in uterus morphology and the occurrence of a blastocyst Brazil, England, Russia and the USA (Palm, 1995). (these precede the tentacular armature in the key), lead- Characteristics of the 31 trypanorhynch genera not ing the user to the different superfamilies. However, examined by Palm (1995) were obtained from the this system has grouped some families together that literature, in the form of the original descriptions, share similar tentacle arrangements, but have few oth- redescriptions and Campbell & Beveridge (1994), with er morphological similarities in common. the latter having priority in cases of discrepancy. During a study of trypanorhynchs from all four superfamilies, two species with a new character com- bination [Poeciloacanthum oweni Palm, 1995 and The current classi®cation Pseudotobothrium dipsacum (Linton, 1897)], ciliat- ed pits on the bothridia together with a poeciloacan- Based on the fundamental work of Dollfus (1942), thous/heteroacanthous typica armature, were found Campbell & Beveridge (1994) grouped 46 genera and (Palm, 1995). Both show a strong morphological rela- 19 families of trypanorhynchs within four superfam- tionship to species of the Otobothriidae Dollfus, 1942 ilies, the latter characterised by the possession of a (Otobothrioidea Dollfus, 1942), although lacking the speci®c kind of tentacular armature. These authors pre- superfamily character,a heteroacanthous atypica arma- served Dollfus' classi®cation scheme and retained his ture. Following the most recent classi®cation of Camp- superfamily concept. The Homeacanthoidea has hooks bell & Beveridge (1994), which uses the tentacular on the tentacles which form quincunxes or continu- armature as a major character at the superfamily lev- ous spiral rows (homeoacanth, after Dollfus, 1942); el, the species have to be placed in two superfami- the Heteracanthoidea is characterised by hooks on the lies Poecilacanthoidea Dollfus, 1942 and Heteracan- tentacles which form half spiral rows with a constant thoidea Dollfus, 1942, neither of which are related to number of hooks (heteroacanth typica). The Otoboth- the family Otobothriidae. rioidea is similar to the latter superfamily but has extra In the present paper,such incongruities arising from hooks or rows of hooks interpolated on the external sur- the most recent Campbell & Beveridge system togeth- face of the tentacle (heteroacanth atypica). Within the er with the usefulness of ciliated pits and prebulbular Poecilacanthoidea, a distinctive hook ®le or ®les form organs as systematic characters are discussed. As a one or more chainettes on the external tentacle surface result, an alternative classi®cation is presented which (poeciloacanth). Finally, using 13 different characters, considers the tentacular armature as being of no impor- Campbell & Beveridge (1994) arranged the families tance at the superfamily level and of limited importance and superfamilies within a cladogram, showing their at the family level. possible phylogenetic relationships. An innovation in the classi®cation of Campbell & Beveridge (1994) is the use of pattern isometry for interpretation of the characteristic armatures. The use of hook arrangements on the tentacle surfaces as the major distinguishing feature for trypanorhynch classi- 83

®cation has the advantage that this character is sim- canthous atypica and poeciloacanthous, within a range ilar for plerocerci (postlarvae) and adults within the of closely related species. same species (Richmond & Caira, 1991). Further- more, hooks are hard structures and do not change Have ciliated pits developed independently or do they their appearance due to different ®xation and prepa- characterise a monophyletic taxon? ration methods. However, despite the precision of the Campbell & Beveridge system, some incongruities are Only a few species of trypanorhynchshave ciliated pits present, and some questions arise. on the borders of the bothridia, thus presenting some evidence for a monophyleticorigin. Based on this char- Different kind of armatures are present within closely acter, Linton (1890) established the genus Otobothri- related species um, and Dollfus (1942) established the family Otoboth- riidae. Using the classi®cation scheme of Campbell & Within one superfamily or family, some closely related Beveridge (1994), which relies on the tentacular arma- species (as far as other morphological characters other ture as a major distinguishing character, species with than the tentacular armature are considered) can have ciliated pits would be separated into all four super- different characteristic armatures. For example, the families (see above), thereby implying an independent superfamily Otobothrioidea sensu Campbell & Bev- development of this organ in several unrelated lineages. eridge (1994) has a heteroacanthous atypica armature Thus, we are confronted with con¯icting evidence. and therefore contains the family Otobothriidae Doll- Most species with ciliated pits have heteroacanthous fus, 1942. This family is additionally characterised by atypica armatures and therefore belong to the Otoboth- the possession of ciliated pits on the bothridial borders rioidea (sensu Campbell & Beveridge, 1994). In con- (Campbell & Beveridge, 1994). Palm (1995) described trast to this, two species Pseudonybelinia odontacan- a new trypanorhynch species with ciliated pits, Poe- tha Dollfus, 1966 and Paranybelinia otobothrioides ciloacanthum oweni Palm, 1995, having a distinct sin- Dollfus, 1966 have homeoacanthous armatures and gle chainette on the external tentacle surface (Figures consequently have to be placed within the Homeacan- 1a±b), and thus exhibiting the features of a poeciloa- thoidea. The armature of Pseudotobothrium dipsacum canth. The species Pseudotobothrium dipsacum (Lin- is heteroacanthous typica (Heteracanthoidea), whereas ton, 1897) (formerly Otobothrium dipsacum Linton, Poeciloacanthum oweni has a poeciloacanthous arma- 1897) has a heteroacanthous typica armature (Figures ture and therefore belongs to the Poecilacanthoidea. 1c±d), without extra hooks or rows of hooks interpo- However, the non-tentacular morphology of the latter lated on the external side of the tentacle (Palm, 1995). species shows a strong relationship to the otobothri- According to the Campbell & Beveridge system, these ids (see above). Therefore, a monophyletic origin of two species have to be placed into two different super- trypanorhynch cestodes with ciliated pits is likely. families, the Poecilacanthoidea and Heteracanthoidea [sensu Campbell & Beveridge (1994)]. However, both Unique position of the Tetrarhynchobothriidae within species show a strong morphological relationship to the Homeoacanthoidea most species of the Otobothriidae (Otobothrioidea), sharing the following suite of characters: the occur- Members of the homeoacanth family Tetrarhyn- rence of a blastocyst, a craspedote scolex with appen- chobothriidae Dollfus, 1969 have several characteris- dix, two patelliform bothridia with horse-shoe-shaped tics shared with those of the Eutetrarhynchidae Guiart, ciliated pits (Figure 2a), a similar form of the bulbs (not 1927, a family belonging to the Heteracanthoidea. P. dipsacum), a similar position of the pars bothridialis Such characters are an elongated scolex with oval both- relative to the pars bulbosa, the origin of the retrac- ridia, long tentacles as well as tentacle sheaths and tor muscles in the anterior part of the bulbs, and the the occurrence of a blastocyst and prebulbular organs. presence of three-®ngered palmate microtriches on the For this reason the group was placed together with bothridia and ®lamentous microtriches only on the rest the Eutetrarhynchidaeby Dollfus (1969). Furthermore, of the scolex (see Palm, 1995). Considering morpho- Schmidt (1986) synonymised Tetrarhynchobothrium logical characters other than the tentacular armature Dollfus, 1969 with Eutetrarhynchus Pintner, 1913. for characterising the Otobothriidae, there are three Beveridge & Campbell (1988) re-assessed the family different armatures, heteroacanthous typica, heteroa- Tetrarhynchobothriidae, discussed in detail its similar- ity to the Eutetrarhynchidae, and placed the family in 84

Figure 1. Trypanorhynch cestodes, scanning electron micrographs: a±b. External tentacle surface of Poeciloacanthum oweni, note the

characteristic single chainette; c±d. External tentacle surface of Pseudotobothrium dipsacum, note the lack of interpolated hooks or rows of  hooks. Abbreviation: C, chainette. Scale-bars:a,b,d,10m; c, 20 m. the suborder Cystidea Guiart, 1927, although it has a Two different kind of armatures on a single tentacle homeoacanth tentacular armature like species of the Acystidea Guiart, 1927. In contrast to this, but consis- In 1982, Dailey & Vogelbein described Mixodigma tent with their classi®cation, the Tetrarhynchobothri- leptaleum Dailey & Vogelbein, 1982 as possessing a idae is allocated to the Homeacanthoidea in Campbell poeciloacanthous basal armature together with a het- & Beveridge (1994), mainly based on the homeoacan- eroacanthous typica armature in the metabasal part of thous armature (see Campbell & Beveridge, 1994, p. the same tentacle. According to the scheme of Doll- 67). This interpretation is still open for discussion. fus (1942), who used only the metabasal armature, M. leptaleum would be considered as having a heteroacan- thous typica armature. However, in the 1994 classi®ca- tion of Campbell & Beveridge, where the armature is 85

Figure 2. Trypanorhynch cestodes, scanning electron micrographs: a. Horse-shoe shaped ciliated pit of Poeciloacanthum oweni; b. External tentacle surface of Otobothrium cysticum, note the hooks marked with arrows, forming a ®le of alternating hooks along the tentacle. Scale-bars:

5 m. interpreted as possessing a chainette of reduced length, entry to the bulbs. Campbell & Beveridge (1994) listed M. leptaleum is unequivocally a poeciloacanth. M. lep- the following families belonging to all four superfam- taleum is not alone in having different hook patterns ilies: Eutetrarhynchidae; Shirleyrhynchidae Camp- on different parts of the tentacles; other examples are bell & Beveridge, 1994; (both Heteracanthoidea); Dasyrhynchus talismani Dollfus, 1935 and Pseudo- Tetrarhynchobothriidae Dollfus, 1969 (Homeacan- grillotia peruviana Escalante & Carvajal, 1984 (see thoidea); Grillotiidae Dollfus, 1969 (Otobothrioidea); Campbell & Beveridge, 1994, p. 67). and Mixodigmatidae Dailey & Vogelbein, 1982 (Poe- M. leptaleum has a strong resemblance to the cilacanthoidea). As a consequence, the prebulbular Heteracanthoidea, especially to the Shirleyrhynchidae organs could not be used as a systematic character Campbell & Beveridge, 1994, with its metabasal arma- for de®ning higher taxa. However, considering the ture, the presence of prebulbular organs, elongate bulbs Tetrarhynchobothriidae to be closely related to the and gland-cells around the retractor muscle, as well Eutetrarhynchidae and the species Mixodigma lep- as its terminal genitalia (see Beveridge & Campbell, taleum to the Shirleyrhynchidae (see above), the enig- 1989). However, the Mixodigmatidae Dailey & Vogel- matic prebulbular organ characterises a large group of bein, 1982 are placed together with the Gymnorhynchi- closely related species. Thus, it is concluded, although dae Dollfus, 1935 within the Poecilacanthoidea in the the function of this organ is still unclear, that pre- cladogram of Campbell & Beveridge (1994). Beside bulbular organs have developed only once within the the distinct chainette on the external surface in the trypanorhynchs. basal part of the tentacles, the two families have few characters in common. Thus, this implication of phy- Do intermediate forms of hook patterns exist? logenetic relationship between these two families on the basis of the poeciloacanthous armature can be con- Some trypanorhynchs are known which have features sidered doubtful. that indicate some relationships between the major groupings (tentacle armature types ± currently at the Have prebulbular organs developed independently or superfamily level), and present a problem for the sys- do they characterise a monophyletic taxon? tematist. An example of such a worm is Otobothrium cysticum (Mayer, 1842). Its tentacular armature is con- Several trypanorhynch cestodes are known which have sidered to be heteroacanthous atypica (Dollfus, 1942; prebulbular organs around the tentacle sheaths at the Campbell & Beveridge, 1994; Palm, 1995). However, 86 in the metabasal part of the tentacle, the last princi- Carvajal & Campbell (1975) came to the conclusion pal hooks on the external tentacle surface could be that the complex oncotaxy is de®nitive for speci®c but interpreted as a ®le of alternating hooks (Figure 2b), not for familial determination. For this reason, a clas- thus ful®lling part of the criteria for a chainette (if si®cation of superfamilies which uses the tentacular all criteria for a chainette were met, the O. cysticum armature as of major importance splits morphological- armature would be poeciloacanthous). As it is thought ly similar species into different families or even super- that a poeciloacanthous armature has evolved from a families. Palm (1995) gave the main reason for this heteroacanthous atypical armature (Campbell & Bev- as a convergent development of tentacular armatures, eridge, 1994), such a longitudinal hook arrangement caused by adaptation of the trypanorhynch tentacular might already represent a precursor of a poeciloacan- apparatus to their sites of attachment in their ®nal hosts. thous hook pattern, with the principal hooks 6 and 6' For example, a homeomorphous and homeoacanthous thought to form a chainette. Thus, O. cysticum can be armature in combination with short tentacles is most interpreted as an intermediate form between heteroa- favourable for attaching to the muscular tissue of the canthous atypica and poeciloacanthous species, sup- stomach wall, while the heteromorphous armature on porting the idea of a relationship between Poeciloacan- long and slender tentacles is perfectly adapted to a thum oweni Palm, 1995 and other otobothriid species connective tissue dominated intestinal wall (see Palm, (see above). 1995). Further support for this functional hypothesis is A further example of a trypanorhynch species, the dominance of worms with the former tentacle type showing a tentacular armature that could be interpreted in cephalopods, where they often occur encapsulated in to be an intermediate form, is a species of Pteroboth- the walls of stomach and caecum, the mesenteries and rium Diesing, 1850. Within this genus, species have a gonads, and in the ventral musculature of the mantle heteroacanthous atypica armature. Some species, such wall (Kinne, 1990). The mobile postlarvae of Tentacu- as P. hawaiense Carvajal, Campbell & Cornford, 1976 laria coryphaenae even occur in the water-®lled man- and P.lintoni (MacCallum, 1916), lack hooklets on the tle cavity, where they exist embedded (Kinne, 1990) external tentacle surface; however, intercalary hooks or attached to the mantle tissue, a site from which they are still present along the tentacle (Campbell & Bev- risk being expelled by turbulence. Cephalopods are eridge, 1996). Using scanning electron microscopy, voracious, versatile predators (Rodhouse & Nigmat- Palm (1995) described two specimens of a Pteroboth- ullin, 1996) and should have as trophic opportunists a rium species which he identi®ed as P. heteracanthum similar access to both types of trypanorhynchs. Diesing, 1850. In the metabasal part of all tentacles Summarising the above, it is concluded that both the hooklets on the external tentacle surface and tentacular armatures are convergent within the try- the intercalary hooks were absent. Only in the basal panorhynchs. Thus, the tentacular armatures charac- part were some hooklets and 1-4 intercalary hooks terise superfamilies in an empirical system of try- present. Thus, considering the general morphology panorhynchs which is unlikely to re¯ect their phyloge- as well as the basal armature, the examined speci- ny. mens clearly belong to a Pterobothrium species with the expected heteroacanthous atypica armature. How- ever, the metabasal part is similar to a heteroacanthous The proposed classi®cation typica armature. In the following section, an alternative classi®cation Comments of trypanorhynch cestodes is presented, which gives the tentacular armature no importance for superfam- These examples outlined above show that the tentacular ily and a reduced importance for family determina- armatures as de®ned by Dollfus (1942) and Campbell tion. Instead, the ciliated pits and prebulbular organs & Beveridge (1994) can split closely related species, are integrated at a higher systematic level, solving as determined by other morphological characters, into the incongruities and unresolved questions discussed different superfamilies. Additionally, the value of the above. A total of six different characters are used to presence of ciliated pits and prebulbular organs in rela- construct this alternative classi®cation; ciliated pits on tion to trypanorhynch classi®cation appears to have the bothridia; prebulbular organs; blastocysts; num- been underestimated. ber of bothridia; rows of hooks on the tentacles; and rhyncheal apparatus. In addition, each of these can 87

Table I. Characters used for an alternative classi®cation of try- Tetrarhynchobothriidae Dollfus, 1969 are added to the panorhynch cestodes. Eutetrarhynchidae. Character Plesiomorphic Apomorphic The third superfamily, Tentacularioidea, contains

trypanorhynchs having neither ciliated pits nor pre-  Ciliated pits on the bothridia absent present bulbular organs. A character providing evidence for Prebulbular organs absent present a monophyletic origin of this superfamily could not Blastocyst absent present

 be demonstrated. Therefore, this superfamily must Number of bothridia 42 Rows of hooks on tentacles complete partly reduced be considered paraphyletic. The Tentacularioidea con- Rhyncheal apparatus present absent tains all remaining genera listed by Campbell & Bev- eridge (1994) and Palm (1995) rearranged in six dif-  Interpretation of character state in agreement with Campbell & ferent families: Tentaculariidae Poche, 1926; Sphyrio- Beveridge (1994). cephalidae Pintner, 1913; Gilquiniidae Dollfus, 1942; Aporhynchidae Poche, 1926; Pterobothriidae Pintner, 1931; and Lacistorhynchidae Guiart, 1927 (Figure 3). be divided into primitive (plesiomorphic) and derived Families (apomorphic) states (Table I), thereby allowing phylo- genetic interpretation of the order. The following characters are used as criteria at the family level: the presence or absence of a blastocyst; Superfamilies the number of bothridia (two or four); the tentacu- lar armature [complete rows of hooks (as exempli- The presence or absence of ciliated pits on the both- ®ed by homeoacanthous and heteroacanthous typica ridia and the presence or absence of prebulbular organs armatures) or rows of hooks partly reduced (as exem- is used to distinguish three superfamilies, the Otoboth- pli®ed by heteroacanthous atypica and poeciloacant- rioidea Dollfus, 1942, Tentacularioidea Poche, 1926 hous armatures)]; and the reduction of the rhyncheal and Eutetrarhynchoidea Guiart, 1927 (Figure 3). apparatus (see Palm, 1995). The penultimate charac- The Otobothrioidea is characterised by ciliated pits ter, rows of hooks complete or rows of hooks partly on the bothridial borders and the absence of prebulbu- reduced, is new and refers to the metabasal armature. lar organs. The presence of ciliated pits is considered Rows of hooks are considered complete if the armature to be a synapomorphic character of this monophylet- consists of only complete spirals or of only principal ic taxon, relative to the absense of this character in rows (respectively, homeoacanthous and heteroacan- all other trypanorhynchs. All species of the Otoboth- thous typica, as de®ned by Dollfus 1942). Rows of riidae Dollfus, 1942 as well as the Paranybeliniidae hooks are considered partly reduced if the armature Schmidt, 1970 are included into the Otobothrioidea. includes any of the following: intercalary hooks, hooks For the species Pseudotobothrium dipsacum (Linton, or rows of hooks interpolated on the external surface 1897), the family Pseudotobothriidae Palm, 1995 is (band of hooks), and a chainette. The common feature added to this superfamily. of these elements on the external tentacular armature is The Eutetrarhynchoidea is characterised by the their interpretation as relicts of former complete spirals absence of ciliated pits on the bothridia and the pos- or principal rows. session of prebulbular organs, as de®ned by Pintner In order to characterise the families, speci®c com- (1880). The presence of prebulbular organs is consid- binations of these characters are used. Of the 19 ered to be a synapomorphic character of this mono- families previously accepted by Campbell & Bev- phyletic taxon, relative to the absense of this character eridge (1994), 10 are retained (Eutetrarhynchidae in all other trypanorhynchs. The Eutetrarhynchoidea Guiart, 1927; Gilquiniidae Dollfus, 1942; Lacis- includes the families Eutetrarhynchidae Guiart, 1927, torhynchidae Guiart, 1927; Mixodigmatidae Dailey Shirleyrhynchidae Campbell & Beveridge, 1994 and & Vogelbein, 1982; Otobothriidae Dollfus, 1942; Mixodigmatidae Dailey & Vogelbein,1982, as de®ned Paranybeliniidae Schmidt, 1970; Pterobothriidae Pint- by Campbell & Beveridge (1994). Additionally the ner, 1931; Shirleyrhynchidae Campbell & Beveridge, three genera, DidymorhynchusBeveridge & Campbell, 1994; Sphyriocephalidae Pintner, 1913; Tentaculari- 1988, Zygorhynchus Beveridge & Campbell, 1988 and idae Poche, 1926 (all sensu nov.). One family is rein- Tetrarhynchobothrium Diesing, 1854, of the former stated (Aporhynchidae Poche, 1926 sensu nov.) and 88 . Alternative classi®cation of trypanorhynch cestodes. Figure 3 89 a new one is established (Pseudotobothriidae Palm, The major difference between this alternative clas- 1995). The following families are considered to be syn- si®cation and the system of Campbell & Beveridge onymous: Hepatoxylidae Dollfus, 1940 with Sphyri- (1994) is that the tentacular armature is not used at the ocephalidae; Tetrarhynchobothriidae Dollfus, 1969 superfamily level, and, at the family level, the arma- with Eutetrarhynchidae; Gymnorhynchidae Dollfus, ture has only a third level of priority. However, in the 1935, Molicolidae Beveridge & Campbell, 1989 and present alternative scheme, the armature is limited to Rhinoptericolidae Carvajal & Campbell, 1975 with being described as rows of hooks complete (homeoa- Pterobothriidae; and Dasyrhynchidae Dollfus, 1935, canthous and heteroacanthous typica) or rows of hooks Grillotiidae Dollfus, 1960, Hornelliellidae Yamaguti, partly reduced (heteroacanthous atypica and poeciloa- 1954 and Mustelicolidae Dollfus, 1969 with Lacis- canthous). Thus, the four established trypanorhynch torhynchidae. armature types are still recognised as being valid, but The proposed new classi®cation of trypanorhynch they are used in a different way, and only at the family cestodes, with de®nitions for superfamilies and fami- level. Additionally, it is still possible to use the tentac- lies, is shown diagrammatically in Figure 3. Some gen- ular armature for subfamily as well as for genera and era are tentatively placed in superfamilies and families species de®nition. within the proposed classi®cation, despite not know- The family Tetrarhynchobothriidae (sensu Camp- ing if they have a blastocyst (for example when the bell & Beveridge, 1994) shows this difference very species description is based on adults). In these cases, clearly: considering the armature to be of major impor- assignment was based upon the general morphology, tance, this family is allocated to the Homeacanthoidea number of bothridia and kind of tentacular armature by Campbell & Beveridge (1994), together with the (see `Future studies' below). other homeoacanth families Sphyriocephalidae, Hepa- toxylidae, Tentaculariidae and Paranybeliniidae. In Comments on the proposed classi®cation contrast to this, in the presented classi®cation the gen- era belonging to this family (sensu Campbell & Bev- The proposed alternative classi®cation of try- eridge, 1994) were assigned to the heteroacanthous panorhynch cestodes allows all known trypanorhynchs typica family Eutetrarhynchidae. to be unequivocally placed on the basis of a de®ned In agreement with Campbell & Beveridge (1994), combination of characters in superfamilies and fam- the presence and absence of blastocysts is no ilies. Using evolutionary systematics, and no longer longer used to distinguish the two suborders Athe- using the tentacular armature as the major character ca and Thecaphora, originally proposed by Guiart for superfamily separation, both the classi®cation and (1927), and suborders within the Trypanorhyncha identi®cation of trypanorhynchs is simpli®ed. Mayr are no longer recognised. Mattis (1986) showed that & Ashlock (1991) pointed out that, for phenetics as the eutetrarhynchid Prochristianella hispida (Linton, well as for cladistics, a large number of characters 1890) (Eutetrarhynchoidea) develops a different type should be used for classi®cation. However, as the most of blastocyst than the otobothriid Poecilancistrum usable characters show an enormous variability within caryophyllum (Diesing, 1850) (Otobothrioidea). Thus, the Trypanorhyncha, only six of them appear to be use- the blastocysts are proposed to be convergent devel- ful for de®nition of higher taxa. Thus, an evolutionary opments within the superfamilies of trypanorhynchs. systematic approach was chosen. Therefore, this character was used at the family level Other strengths of this new classi®cation are that only. incongruities in the existing classi®cations, caused by The characteristics of the genitalia were not used to the use of the tentacular armature as a major character distinguish higher taxa within the trypanorhynchs, as for distinguishing superfamilies, are omitted (Palm, they are characteristic only of adults. However, follow- 1995). Furthermore, with the ®nding of new species ing the Campbell & Beveridge system, the character with new combinations of the chosen six characters or uterus linear/uterus a transverse sac appears to be use- with increasing knowledge concerning misdetermined ful to distinguish between families without blastocyst species (see `Future studies' below), this new classi®- and those with blastocyst within the Tentacularioidea. cation can be enlarged to up to four superfamilies and Similarly, the muscles encircling the tentacle sheaths, 48 families without losing its stability. It is also pos- yet lacking a prebulbular organ as present in several try- sible to determine all the information required using panorhynch genera, were not included in this alterna- light microscopy. tive classi®cation. This was because their relationship 90 to these organs as well as their occurrence within the Thus, within the present classi®cation, the total reduc- different trypanorhynch species could not be clari®ed. tion of tentacular hooks on the external tentacle surface of heteroacanthous atypica armatures is interpreted as Phylogenetic implications a derived character. It has yet to be proven by new ®ndings, or with the The dendrogram constructed from the above proposed help of molecular techniques, that the present interpre- classi®cation (Figure 4), compared with the dendro- tations of phylogenetic relationships are viable. How- gram of Dollfus (1942) and the cladogram of Campbell ever, this alternative classi®cation is considered to be & Beveridge (1994), shows very different phyloge- closer to a natural system than the proposed classi®- netic relationships between the different families. For cations proposed by Dollfus (1942) and Campbell & example, in this proposed classi®cation, the Otobothri- Beveridge (1994),as it no longer uses the hook arrange- oidea is assumed to be a phylogenetically derived try- ments on the tentacles (concluded to be convergent panorhynch group, which is split from all other forms developments) as a major character for distinguishing by the development of ciliated pits. Otobothrioids superfamilies. have a plesiomorphic tentacular armature (Pseudony- belinia odontacantha Dollfus, 1966; Paranybelinia Future studies otobothrioides Dollfus, 1966; Pseudotobothrium dip- sacum (Linton, 1897)) and an apomorphic (Otoboth- Within the present classi®cation, the importance of the rium spp.) tentacular armature. A second phylogeneti- character ªwith or without blastocystº for distinguish- cally derived group is the Eutetrarhynchoidea, having ing different trypanorhynch families causes problems prebulbular organs as de®ned by Pintner (1880). Like if this feature is not known, due to the availability the ciliated pits, these organs are characteristic devel- of adults only or metacestodes (plerocerci) already opments within a single superfamily. removed from their blastocyst. Such trypanorhynchs The development of poeciloacanthous armatures is can, however, be assigned to a family based on other indicated by the present system to have arisen inde- features. In many cases, assignment will be to one fam- pendently several times: all three superfamilies have ily only, but it is possible that a choice of two families species with this type of armature. It is interesting to will occur. In such cases, assignment should be based note that independent evolution of the poeciloacan- upon the general morphology, the number of both- thous feature ªdouble-winged chainettesº has previ- ridia and the kind of tentacular armature. To date there ously been proposed by Beveridge & Campbell (1989) is a strong correlation between the presence/absence based on the examination of oncotaxy and the anato- of blastocysts and the nature of tentacular armature. my of mature segments. Other remarkable similarities However, reassignment to another position within this between the present alternative classi®cation and the classi®cation is possible if increasing knowledge on study of Beveridge & Campbell (1989) are that the the life-cycles of the species provides con¯icting evi- Gymnorhynchidae (synonymous with Pterobothriidae dence. in this classi®cation) may be derived from a common This study shows that it is possible to assign all ancestor with the Gilquiniidae, and the Myxodigmati- known Trypanorhyncha using a classi®cation based on dae may be derived from a common ancestor with six different characters. However, in developing the the Shirleyrhynchidae, including both Shirleyrhynchus present alternative classi®cation, it was apparent that Beveridge & Campbell, 1988 and Cetorhinicola Bev- certain genera and organs remain poorly known.There- eridge & Campbell, 1988 (Figure 4). In their more fore, it is suggested for future studies, that the follow- recent work, however, Campbell & Beveridge (1994) ing topics should be examined by trypanorhynch spe- indicated that these families were more distantly relat- cialists: (i) the comprehensive characterisation of the ed. surface ultrastructure of Grillotia Guiart, 1927, Nybe- Palm (1995, p. 204) demonstrated a reduction-row, linia Poche, 1926 and Otobothrium Linton, 1890; (ii) a possible direction of development of the tentacu- the functional morphology of prebulbular organs; (iii) lar armature within the genus Pterobothrium Diesing, whether or not the muscular rings around the basal 1850. An important consequence of this was that an part of the tentacle sheaths are homologous to prebul- armature with spiniform principal hooks and lacking bular organs; and (iv) the functional morphology of a band of hooks on the external tentacle surface rep- microtriches and other tegumental structures, includ- resents a highly developed armature within the genus. ing ciliated pits and sensory organs. 91

Figure 4. Dendrogram, showing a possible phylogenetic relationship between trypanorhynch families. Characters: 1. Ciliated pits; 2. Prebulbular organs; 3. Blastocyst; 4. Number of bothridia; 5. Reduction of the rhyncheal system; 6. Partial reduction of tentacular hooks. Roman letters I-IV = numbers of derived characters. The length of the branches does not contain cladistic information.

Acknowledgements cha) with descriptions of two new genera, Didymorhynchus and Zygorhynchus. Systematic Parasitology, 12, 3±29. My thanks are extended to Dr T. Mattis, Dr Ch. Nig- Beveridge, I. & Campbell, R.A. (1989) Chimaerarhynchus n. g. and Patellobothrium n. g., two new genera of trypanorhynch cestodes matullin and Mr H. Seilert for their valuable discus- with unique poeciloacanthous armatures, and a reorganisation of sions. I am grateful to Drs I. Beveridge, R.A. Bray, the poeciloacanthous trypanorhynch families. Systematic Para- H.-U. Dahms, S. Lorenzen, S.L. Poynton, R.M. Over- sitology, 14, 209±225. Campbell, R.A. & Beveridge, I. (1994) Order Trypanorhyncha street and anonymous reviewers for their constructive Diesing, 1863. In: Khalil, L.F., Jones, A. & Bray, R.A. (Eds). criticism on the manuscript. Financial support was Keys to the cestode parasites of vertebrates. Wallingford: CAB provided by the German National Research Council International, pp. 50±148. (DFG). Campbell, R.A. & Beveridge, I. (1996) Revision of the family Pter- obothriidae Pintner, 1931 (Cestoda: Trypanorhyncha). Inverte- brate Taxonomy, 10, 617±662. Carvajal, J. & Campbell, R.A. (1975) Rhinoptericola megacantha References gen. et sp. n., representing a new family of trypanorhynch ces- todes from the cownose ray, Rhinoptera bonasus (Mitchill 1815). Beveridge, I. & Campbell, R.A. (1988) A review of the The Journal of Parasitology, 61, 1,023±1,030. Tetrarhynchobothriidae Dollfus, 1969 (Cestoda: Trypanorhyn- 92

Dailey, M.D. & Vogelbein, W. (1982) Mixodigmatidae, a new fam- Mayr, E. & Ashlock, P.D. (1991) Principles of systematic zoology. ily of cestode (Trypanorhyncha) from a deep sea, planktivorous New York: McGraw-Hill, Inc., 475 pp. shark. The Journal of Parasitology, 68, 145±149. Palm, H.W. (1995) Untersuchungen zur Systematik von Deardorff, T., Raybourne, R.B. & Mattis, T.E. (1984) Infections RusselbandwÈ urmernÈ (Cestoda: Trypanorhyncha) aus atlantischen with trypanorhynch plerocerci (Cestoda) in Hawaiian ®shes of Fischen. Berichte aus dem Institut furÈ Meereskunde Kiel, 275, 238 commercial importance.Sea Grant Quarterly, 6, 1±6. pp. Dollfus, R.P. (1942) Etudes critiques sur les Tetrarhynques du Muse- Pintner, T. (1880) Untersuchungen uberÈ den Bau des Band- um de Paris. Archives du Museum National d'Histoire Naturelle, wurmkorpersÈ mit besonderer BerucksichtigungÈ der Tetrabothrien 19, 1±466. und Tetrarhynchen. Arbeiten aus dem Zoologischen Institut der Dollfus, R.P. (1969) De quelques Cestodes Tetrarhynques UniversitatÈ Wien und der Zoologischen Station in Triest, 3, 163± (Het eracantes et Pecilacanthes) recolt es chez des poissons de 242, pl. XIV±XVIII. la Mediterran ee. Vie et Milieu,Serie A, 20, 491±542. Richmond, C. & Caira, J.N. (1991) Morphological investigations Guiart, J. (1927) Classi®cation des Tetrarhynques. Comptes rendus into Floriceps minacanthus (Trypanorhyncha: Lacistorhynchi- de la 50. Session de l'Association Francaise pour l'Avancement dae) with analysis of the systematic utility of scolex microtriches. des Science, Lyon, 1926, pp. 397±401. Systematic Parasitology, 19, 25±32. Kinne, O. (1990) Diseases of marine animals Vol. III. Hamburg: Rodhouse, P.G. & Nigmatullin, Ch.M. (1996) Role as consumers. Biologische Anstalt Helgoland, 696 pp. Philosophical Transactions of the Royal Society of London,Series Linton, E. (1890) Notes on Entozoa of marine ®shes of New England, B, 351, 1,003±1,022. with descriptions of several new species. Part II.Annual Report of Schmidt, G.D. (1986) Handbook of tapeworm identi®cation. Boca the United States Commissioner of Fish and Fisheries for 1887, Raton: CRC Press Inc., 675 pp. pp. 719±900; pl. I±XV. Wardle, R.A. & McLeod, L.A. (1952) The zoology of tapeworms. Mattis, T.E. (1986) Development of two tetrarhynchidean cestodes Minneapolis: University of Minnesota Press, 780 pp. from the northern Gulf of Mexico. Dissertation, University of Southern Mississippi, 171 pp.