NEW HOST AND OCEAN RECORDS FOR DRIOCEPHALUS CEREBRINOXIUS ( SPHYRIIDAE , SIPHONOSTOMATOIDA ) AND A RECONSIDERATION OF PHYLOGENY WITHIN SPHYRIIDAE

Sphyriidae Wilson, 1919 (Siphonostomatoida, Copepoda) contains 9 genera and 35 species (Ho et al. 2003, Turner et al. 2003, Boxshall 2004) whose transformed adult females are highly modified mesoparasites (sensu Kabata 1979) of fishes. A member of Sphyriidae, Driocephalus cerebrinoxius (Diebakate, Raibaut et Kabata, 1997) was originally placed in Thamnocephalus (see Diebakate et al. 1997). However, subsequent discovery that Thamnocephalus Diebakate, Raibaut et Kabata, 1997 was a junior homonym of Thamnocephalus Packard, 1877, a genus comprised of fairy shrimp (Anostraca, Branchiopoda), required the erection of Driocephalus Raibaut, 1999 to accommodate T. cerebrinoxius Diebakate, Raibaut et Kabata, 1997. Driocephalus cerebrinoxius is an unusual copepod, even amongst sphyriids (Sphyriidae), because it infects the nervous system of its host. Transformed adult females attach in and about the olfactory lobe with the posterior of the body trailing free within the lumen of the olfactory sac (Diebakate et al. 1997). We herein present the second report of D. cerebrinoxius based on collections representing new host and ocean records. We also present results of a phylogenetic analysis of relationships within Sphyriidae. ACTA ICHTHYOLOGICA ET PISCATORIA (2006) 36 (1): 19


MATERIALS AND METHODS
Seven star-spotted smooth hounds, Mustelus manazo Bleeker, 1854 (Triakidae, Carcharhiniformes); four females and three males, 532-995 mm total length, 479-3780 g body weight) were captured between July 1994 and January 1995 in Tokyo Bay, central Japan, using a trawl fishing at a depth of 20-50 m.Sharks were transported on ice to the laboratory where copepods were collected from the olfactory sacs and fixed in 70% ethanol.For study, copepods were cleared and stained in a solution of lactic acid and lignin pink before being examined under a stereomicroscope.Illustrations were made with the aid of a drawing tube.Body features were measured by marking point to point lengths on paper using a drawing tube and then measuring these marks, using a stage micrometer.Body measurements are reported as x ± s x , general copepod terminology follows Kabata (1979), dorsal-ventral convention for copepods follows Diebakate et al. (1997), and host systematics follows Compagno (1999).Copepod voucher specimens have been deposited as follows: 3 specimens from 3 hosts deposited in the National Museum of Natural History, Smithsonian Institution, Washington, D.C. (USNM 1088537, 1088538, 1088539), and 4 specimens from 3 hosts deposited in the National Science Museum (Tokyo, Japan) (NSMT-Cr 16826, 16827, 16828).The remaining 2 specimens are in the collection of the senior author.

RESULTS
All seven examined sharks were infected with Driocephalus cerebrinoxius; five sharks infected with one copepod each, two sharks infected with two copepods each.The copepods match the description of D. cerebrinoxius well, although they are not fully intact, i.e., each lacks most or all of the cephalothorax.As noted by Diebakate et al. (1997), it is difficult to remove these mesoparasites from the host.In this case, it is likely that specimens were decapitated when they were removed from the olfactory sacs, such that the firmly embedded portions of the cephalothorax remained in situ.The body (Fig. 1A-C) is comprised of three general regions: cephalothorax, thoracic neck (neck), and genito-abdominal complex (trunk).The base of the cephalothorax is present on several specimens, as indicated by an enlarged area at the anterior of the neck (Fig. 1A-C).In addition, one specimen (see Fig. 1D) possesses remnants of the branching cephalothoracic dendrites that are unique to D. cerebrinoxius as well as two small hemispherical swellings that were interpreted by Diebakate et al. (1997) as possibly being reduced appendages (first or second antennae).The neck of all specimens is cylindrical and thin, 0.44 ± 0.02 mm wide (n = 9), 6.07 ± 0.82 mm long (n = 5).Trunk (Fig. 1A-C) 3.72 ± 0.24 mm wide (n = 7) at midpoint, 5.77 ± 0.43 mm long (n = 7) not including posterior processes, roughly orbicular, slightly dorsoventrally depressed, dorsal aspect with convex contour (Fig. 1A, B), ventral surface with shallow concave contour (Fig. 1C), anterolateral margins rounded, lateral and some dorsal margins appear swollen in several places, posterior margins with two pairs of lateral lobes, ventral pair slightly anterior to dorsal pair (Fig. 1B, E), ventral lobes simple subspherical swellings, dorsal lobes each comprised of two confluent swellings that appear to protect basal portions of egg sacs (Fig. 1E).Posterior processes allantoid (Fig. 1C), 1.68 ± 0.28 mm wide (n = 10) at midpoint, 2.23 ± 0.09 mm long (n = 10), much shorter than main portion of trunk (Fig. 1B, C), attached just ventral to perianal swelling (Fig. 1E).Abdomen a small bilobed perianal swelling located ventral to openings of oviducts (Fig. 1E).Egg sacs allantoid, multiseriate (Fig. 1A-C), 1.47 ± 0.05 mm wide (n = 11) at midpoint, 4.43 ± 0.25 mm long (n = 11).No males observed or collected.
Twelve most parsimonious trees (tree length = 16 steps, consistency index = 0.875) resulted from the cladis-tic analysis of Sphyriidae.One of the 8 morphological characters (Character 1; see Tables 1, 2) used in the analysis was parsimony uninformative, i.e., a basal synapomorphy for Sphyriidae.One strict consensus tree (Fig. 2A) and one 50% majority rule consensus tree (Fig. 2B) resulted from a consensus of the aforementioned 12 shortest trees.None of the aforementioned trees was fully resolved.Nevertheless, Periplexis and Paeonocanthus were depicted as sister taxa on all trees (e.g., see Fig. 2A, B).Lophoura and Sphyrion were depicted as sister taxa on 50% of the most parsimonious trees, while Lophoura and Driocephalus were depicted as sister taxa on the other 50% of all most parsimonious trees.Paeon group A and Paeon group B were depicted as independent taxa on all  trees (e.g., see Fig. 2A, B).Most non-terminal branches on all trees were supported by 1 synapomorphy and no branch was supported by more than 2 apomorphies.

DISCUSSION
We assigned our specimens to Driocephalus cerebrinoxius based on three lines of evidence.First, amongst sphyriids, D. cerebrinoxius has a uniquely shaped trunk and cephalothorax (Diebakate et al. 1997).Second, D. cerebrinoxius is the only sphyriid that infects the olfactory organ of elasmobranchs and only copepod known to infect the olfactory lobe (Diebakate et al. 1997).Third, it is not unusual for sphyriid species to infect multiple species or for them to possess wide geographic distributions, e.g., see Ho (1992) regarding the distribution of Sphyrion lumpi (Krryer, 1845).Based on this species assignment, this report is a new host record (M. manazo) and new ocean record (western Pacific) for D. cerebrinoxius.However, it should be noted that Yamaguchi et al. (2003) reported these specimens, as well as others collected from the same host species captured off Maizuru, Japan, in the Sea of Japan, as an unidentified genus and species of Sphyriidae.Yamaguchi et al. (2003) reported the prevalence of this sphyriid infecting M. manazo as 3.3 and 1.0 in Tokyo Bay (213 hosts examined) and off Maizuru (93 hosts examined) respectively.In that report (loc.cit.) the infection site of the copepods was erroneously switched.In fact, D. cerebrinoxius (reported as Sphyriidae gen.sp.) was collected from the nostril (i.e., olfactory sac) while Perissopus oblongatus was collected from the fin.The authority and date for P. oblongatus was not provided by Yamaguchi et al. (2003) but it was likely (Wilson, 1908).Perissopus oblongatus (Wilson, 1908) is considered a junior synonym of Achtheinus oblongus Wilson, 1908 as proposed by Ho (1975) and recently endorsed by Benz et al. (2003).Diebakate et al. (1997)  Ovigerous female character states for eight characters used in the phylogenetic analysis of Sphyriidae Wilson, 1919; codes 0, 1, 2, and 3 within data matrix identify particular character states for each character as defined in Table 1; code 0 indicates a plesiomorphy rather than an absence, code M indicates missing (unknown) or ambiguous data; Ommatokoita Leigh-Sharpe, 1926 served as the outgroup; number of species accumulated from information in Ho et al. (2003), Turner et al. (2003), andBoxshall (2004) * The description of this appendage by Wilson (1919: p. 460) as, "one-jointed papilla on either side of mouth tube, a mere stump, apparently immovable," stymies character state designation.† Wilson (1908; p. 460)   Hewitt (1965), Kabata (1965), Hogans (1986b), and Ho et al. (2003) expose intrageneric and intraspecific variation that prevents certain allocation as code 0 or 1.
al. 2005) suggest that D. cerebrinoxius may infect these species elsewhere in temperate or tropical coastal waters.Furthermore, given that D. cerebrinoxius has been reported infecting representatives of three families of carcharhiniforms (Carcharhiniformes), it is likely that future host records may include additional species of these "ground sharks."Diebakate et al. (1997) did not propose an explicit hypothesis regarding the phylogenetic relationship of Drio-cephalus to other sphyriid genera.Therefore, it was our intension to use our new observations and the literature to extend the phylogenetic analysis of Sphyriidae published by Dojiri and Deets (1988) to include Driocephalus.However, our review of the characters and character states used by Dojiri and Deets (1988) sometimes placed us at odds and ultimately prompted us to produce a new analysis.Given the relatively widespread adoption of the phylogenetic hypothesis for Sphyriidae presented by Dojiri and Deets (1988) and especially the interest shown in its associated hypotheses regarding host-parasite coevolution and parasite historical ecology (e.g., see Ho 1992, Paterson and Poulin 1999, Roberts and Janovy 2000, Boxshall 2000, 2004), we consider it important to herein list our concerns regarding some of the characters used in that analysis.
Character A (body condition): Dojiri and Deets (1988) considered sphyriids to exhibit a rotated body condition, i.e., torsion.While torsion has been mentioned as a characteristic of a few sphyriids (see Kabata 1979), Wilson (1919), Kabata (1979), Benz (1993), andBoxshall (2004) did not mention torsion as a synapomorphy for Sphyriidae.Wilson (1919) provided the best description of torsion within Sphyriidae, mentioning that this body twisting stems from the parasite growing while changing its path through the host to affect contact with a particular host structure.Thus the existence of torsion in sphyriids is dictated by environmental factors and may vary (Wilson 1919).
Character B (egg sac length): Dojiri and Deets (1988) considered the egg sacs of sphyriids to be elongate relative to those of O. elongata.However, we consider the egg sac length of O. elongata relative to either total body length or trunk length to be within the length range or longer than the eggs sacs of sphyriids.For example, relative to trunk length, the egg sacs of P. vaissierei (see figure 15 in Delamare-Deboutteville and NuZes-Ruivo 1954) are similar to those of O. elongata (see figure 34.17 in Roberts and Janovy 2000), while relative to total body length, the egg sacs of O. elongata are longer than those of P. vaissierei (loc.cits.).
Characters K and L (genital complex): we did not agree with allocation of the genital complex character states ovoid, gradually expanding, and discoid in the analysis of Dojiri and Deets (1988).The trunk (i.e., genital complex in Dojiri and Deets 1988), of O. elongata appears (see Kabata 1979, Benz et al. 1998) more like the gradually expanding condition noted by Dojiri and Deets (1988) for taxa such as Opimia, Paeon, and Tripaphylus than it does the ovoid condition coded for Lophoura, Paeonocanthus, Periplexis, and Sphyrion.Dojiri and Deets (1988) also considered the trunk of Paeon spp. as gradually expanding, but the trunk of three of eight Paeon spp.(P.australis, P. triakis, and P. veriscolor) are not (Wilson 1919, Kabata 1993, Castro Romero 2001).And, Dojiri and Deets (1988) considered the trunk of Paeonocanthus antarcticensis (Hewitt, 1965) to be ovoid, but illustrations in Hewitt (1965), Kabata (1965), Hogans (1986b), andHo et al. (2003) together denote trunk variation that challenges confidence in allocating a state of either ovoid or gradually expanding for Paeonocanthus.
Characters R and S (cephalothorax): we felt unsure operationally regarding the definitions used by Dojiri and Deets (1988) regarding these characters, especially in light of how little is known regarding the homology of various components of the sphyriid cephalothorax.For example, the cephalothorax of some Lophoura spp.(e.g., see Ho andKim 1989, Boxshall 2000) appears bulbous to us, while other congeners seem to possess cephalothoracic protuberances (e.g., see Hogans and Dadswell 1985).
Character V (gut diverticulae): Dojiri and Deets (1988) considered Sphyrion and Lophoura spp. to possess a synapomorphy in the form of anastomosing gut diverticulae (see Najarian 1952).Kabata (1979) mentioned this feature in one sentence of general discussion as being unique to Sphyrion and Lophoura and Wilson (1919) mentioned the aforementioned taxa as sharing this characteristic while Paeon did not.However, the literature contains no information regarding the form of gut diverticulae amongst representatives of Driocephalus, Norkus, Opimia, Paeonocanthus, Periplexis, or Tripaphylus or the outgroup Ommatokoita.
Character W (first antenna): Dojiri and Deets (1988) considered the form of the adult female first antenna to be known for members of Opimia and Paeon.However, the only mention of this structure in Opimia (see Wilson 1908) is confusing and does not allow character state designation.Regarding Paeon, we are unaware of literature detailing the structure of the adult female first antenna of its species.
Character X (second antenna): Dojiri and Deets (1988) considered the form of the adult female second antenna to be biramous for Paeon spp.However, the literature contains no details regarding the condition of this appendage for 50% of Paeon spp., i.e., P. asymboli, P. australis, P. triakis, and P. versicolor.
No cladogram resulting from our analysis was unambiguously congruent with the cladogram for sphyriid genera presented by Dojiri and Deets (1988) even in light of taxonomic progress since 1988 (cf.Fig. 2A, B with C).Our consideration of Paeon as being comprised of 2 groups (see Tables 1, 2) was supported by the fact that all of our trees depicted Paeon group A and Paeon group B as being independent from one another (Fig. 2A).In our analysis, monotypic Driocephalus (whose representative is only known from elasmobranchs) was placed as the sister taxon to Lophoura (all of whose representatives infect teleosts) on 50% of all shortest trees and as a sister taxon to a clade comprised of Lophoura and Sphyrion (both only known to infect teleosts) on the remaining 50% of shortest trees.None of our cladograms depicted Sphyriidae as being comprised of 2 clades (as did the results of Dojiri and Deets 1988), one consisting of taxa that infect only elasmobranchs (i.e., representatives of Norkus, Opimia, Tripaphylus, Paeon, and Driocephalus) and the other consisting of taxa that infect only teleosts (i.e., representatives of Periplexis, Paeonocanthus, Lophoura, and Sphyrion).
In all, our results do not convincingly support any of the host summary or ecological summary cladograms presented by Dojiri and Deets (1988), albeit one could surmise how additional parsimony informative characters might alter our results to support some or even all of the results of Dojiri and Deets (1988) (cf.Fig. 2A, B with C).Nevertheless, with our cladograms being supported by so few synapomorphies, we do not consider any of them robust enough to in turn support a consideration of host associations or ecological associations amongst sphyriids.And, based on the highly modified and morphologically simple habitus of sphyriid adult females and the difficulty in recognizing homology regarding their character states, as well as the lack of information regarding the morphology of sphyriid adult males and the relatively modified and simple morphology of known sphyriid adult males as compared to the adult males of many other siphonostome taxa (e.g., Kroyeriidae, Eudactylinidae, Caligidae, and others; see Kabata 1979, Benz 1993, Deets 1994), we believe that a robust phylogeny for members of Sphyriidae is best sought using molecular tools.Until such a phylogeny is realized, it seems premature to conduct analyses regarding the historical ecology (e.g., host associations, habitat associations, environment associations, biogeographic associations, etc.) of sphyriids.