INTRODUCTION
Cestodes belonging to the order Tetraphyllidea have been considered to be oioxenous (exhibiting strict host specificity) (e.g. Williams, Reference Williams1960, Reference Williams1961, Reference Williams1964, Reference Williams1966, Reference Williams1968, Reference Williams1969). Williams (Reference Williams1966) noted that no mature Echeneibothrium spp. have been reported from more than 1 rajid host species, and only on rare occasions have immature specimens been recovered from 2 rajid host species. For the sister genus Pseudanthobothrium Baer, Reference Williams1956 (Caira et al. Reference Caira, Jensen and Healy1999, Reference Caira, Jensen, Healy, Littlewood and Bray2001), we were unable to identify published accounts regarding host specificity. However, in an earlier study, we assessed the host specificity of mature specimens of 2 Pseudanthobothrium spp., using anatomical observations and partial sequence data of a variable region (D2) from the large subunit of nuclear ribosomal DNA (LSU), and established that Pseudanthobothrium n.sp. and P. hanseni infect different ecological pairs of host species (Randhawa et al. manuscript submitted). Randhawa et al. (manuscript submitted) also reported, on the basis of morphological features, that E. vernetae also occurs in 2 different host species (Leucoraja erinacea and L. ocellata). These findings question the strictness of the host-parasite relationship for adult Pseudanthobothrium and Echeneibothrium species.
Assessments have been made from mature specimens possessing the necessary morphological characters for species-level identification and subsequently confirmed by the molecular data (Randhawa et al. manuscript submitted). On the other hand, the identity of immature specimens can only be determined using molecular tools, since species-diagnostics are based solely on adult features (e.g. Euzet, Reference Euzet, Khalil, Jones and Bray1994). For studies investigating the specificity of parasites to be comprehensive, however, accurate species-level identifications of immature specimens are necessary to determine at what stage of the host-parasite interaction does specificity occur. Molecular markers are invaluable tools for measuring and assessing the specificity patterns of host-parasite relationships (Anderson et al. Reference Anderson, Blouin and Beech1998). In cestodes, the D1–D3 region of the LSU is useful for discriminating among species (Mariaux and Olson, Reference Mariaux, Olson, Littlewood and Bray2001; Olson et al. Reference Brickle, Olson, Littlewood, Bishop and Arkhipkin2001; Reyda and Olson, Reference Reyda and Olson2003), and the D2 domain, a divergent and rapidly evolving region of the LSU (Harper and Saunders, Reference Harper and Saunders2001), has been used previously to differentiate between tetraphyllidean species (Brickle et al. Reference Brickle, Olson, Littlewood, Bishop and Arkhipkin2001; Agusti et al. Reference Agusti, Aznar, Olson, Littlewood, Kostadinova and Raga2005). This molecular marker allows us to identify accurately immature specimens to species, therefore determining when host specificity is established in these host-parasite relationships.
There are 2 main views of host specificity: host range and quantitative measure. The first, host range, is the most commonly used concept (Poulin, Reference Poulin1998) and relates to the number of different host species infected by a single parasite species at a given stage of its life-cycle (Euzet and Combes, Reference Euzet and Combes1980; Holmes, Reference Holmes1987; Lymbery, Reference Lymbery1989; Combes, Reference Combes1995, Reference Combes2001; Poulin, Reference Poulin1998). The concept of filters was introduced by Euzet and Combes (Reference Euzet and Combes1980) to illustrate the 4 parameters responsible for delimiting the host range of parasites, thus defining the degree of specificity of these parasites. The ‘encounter filter’ is defined as the probability of contact between a given parasite species and potential hosts and includes a biodiversity parameter (geographical component) and a behaviour parameter (spatial component). Host species absent from the ecosystem of a parasite (biodiversity parameter) are excluded from the host range of the parasite. Similarly, host species whose behaviour (behaviour parameter) renders contact with infective stages of the parasite impossible are excluded from the host range of the parasite. The ‘compatibility filter’ is defined as the probability of a parasite establishing in the host following encounter and includes a resource parameter and a defence parameter. Host species not providing the adequate spatial resources (e.g., attachment surface or interspecific competition) or metabolic resources (e.g., glucose) to meet the needs of the parasite are excluded from the host range of the parasite (resource parameter). This type of exclusion is also referred to as host unsuitability. Host species, whose immune factors or other mechanisms prevent the establishment of the parasite, are excluded from the host range of the parasite (defence parameter). This type of exclusion is also referred to as host resistance or host restriction (e.g. Rohde and Rohde, Reference Rohde and Rohde2005). Therefore, of all potential host species, only a subset is encountered by the parasite, and of that subset, only the parasite species compatible with the host can establish.
The second view measures specificity by quantifying prevalence, abundance and mean intensity of infection by parasites in different host species (Rohde, Reference Rohde1980, Reference Rohde1994, Reference Rohde and Rohde2005; Lymbery, Reference Lymbery1989; Rohde and Rohde, Reference Rohde and Rohde2005) and relating these parameters to the phylogenetic relatedness between infected host species (e.g. Poulin and Mouillot, Reference Poulin and Mouillot2003, Reference Poulin and Mouillot2004; Krasnov et al. Reference Krasnov, Shenbrot, Khokhlova and Poulin2004; Rohde and Rohde, Reference Rohde and Rohde2005). These measurements would provide information on host preference of parasites and accidental infections. Although the traditional view of host specificity (i.e. host range) is adhered to here, prevalence and intensity of infection data are presented in recognition of their utility in discussing the ecological implications of our findings on the host distribution of mature and immature specimens of the four parasite species studied herein.
From June 2002 to September 2004, 84 L. erinacea (Mitchill, 1825), 25 Malacoraja senta (Garman, 1885), 11 Amblyraja radiata (Donovan, 1808), and 7 L. ocellata (Mitchill, 1815) were collected from Passamaquoddy Bay and waters surrounding the West Isles of the Bay of Fundy, NB, Canada. As a result of our parasitological survey: mature specimens of Pseudanthobothrium n.sp. and Echeneibothrium vernetaeEuzet, Reference Euzet1956 were recovered from L. erinacea and L. ocellata; P. hanseniBaer, Reference Williams1956 was recovered from A. radiata and M. senta; E. dubium abyssorum Campbell, 1977 and E. canadensisKeeling and Burt, Reference Keeling and Burt1996 were recovered from A. radiata; Zyxibothrium kamienae Hayden and Campbell, 1981 was recovered from M. senta, and Grillotia sp. was recovered from all 4 rajid skate species. Approximately 250 immature specimens of Pseudanthobothrium spp. and Echeneibothrium spp. were also recovered from the 4 rajid skate hosts.
In this study, the host distribution of both mature and immature specimens of 2 Pseudanthobothrium spp. and two Echeneibothrium spp. was assessed, using the partial sequence of the D2 domain of the LSU as a molecular marker, to gain insights into the stage of the host-parasite relationship where specificity becomes apparent (onset). The results indicate that, for the species studied here, specificity in Pseudanthobothrium specimens occurs prior to attachment, whereas specificity in Echeneibothrium specimens occurs post-attachment.
MATERIALS AND METHODS
Collections and examination of material
Skates were collected during Otter trawls on board the W. B. Scott R/V and CCGS Pandalus III, and identified using keys and descriptions from Scott and Messieh (Reference Scott and Messieh1976) and Scott and Scott (Reference Scott and Scott1988). Skates were maintained in a holding tank or ‘live well’ on the vessels and subsequently kept live at the research facilities of the Huntsman Marine Science Centre (HMSC) in St Andrews, NB, until examination. This generally occurred within 24 h of their capture. Skates were pithed and access to internal organs was achieved by cutting out the ventral body wall. Spiral valves from 24 L. erinacea, 11 A. radiata, ten M. senta, and seven L. ocellata were retrieved and examined immediately by making a mid-ventral incision through the whorls of the mucosal sheet, from the rectum straight up to the pyloric stomach along the ventral blood vessel, thus exposing 2 surfaces with distinct chambers separated by a mucosal flap. These spiral valves were then placed in saline in a large Petri, or culture, dish and examined using a binocular dissecting microscope. Both mature and immature parasites were removed from the spiral valves, and cleaned in fresh saline prior to being processed. For the purpose of this study, immature worms are defined as those encompassing a morphological gradient between that resembling a plerocercoid to that of specimens with evident strobilation, but lack of sexual features. Only scoleces, both attached and detached, were counted to determine the number of parasites present in each individual spiral valve and the attachment-site of those attached was noted. Both mature and immature worms were fixed in hot, almost boiling 70% ethanol and stored in fresh 95% or absolute ethanol. Scoleces of both mature and immature specimens were retained (stored in 70% ethanol) as vouchers for each of the specimens. Other spiral valves, preserved for later examination, were injected with 10% formalin and treated as described by Randhawa et al. (manuscript submitted).
This study included sequence data from 64 mature specimens, including the 51 mature Pseudanthobothrium specimens from the Northwest Atlantic sequenced by Randhawa et al. (manuscript submitted), and 92 immature specimens (Table 1). Immature Echeneibothrium specimens were not recovered from L. ocellata. Additionally, since the 3 mature E. d. abyssorum specimens recovered from A. radiata were used for physiological experiments (results to be published elsewhere), this parasite species was not included in the analyses. Voucher material for immature worms is deposited with the NB Museum (Table 1).
Inventory of specimens used for assessment of host distribution with voucher and GenBank Accession numbers
a Sequences obtained during this study.
b Includes 1 specimen from 1997 collections.
c Sequences obtained from Randhawa et al. (manuscript submitted).
d Includes 6 specimens from 1997 collection.
Molecular characterization and analysis
Genomic DNA was extracted using standard techniques (Devlin et al. Reference Devlin, Diamond and Saunders2004). The 5′ end of the large subunit ribosomal DNA (LSU) was amplified as reported by Harper and Saunders (Reference Harper and Saunders2001), using the Ex-Taq polymerase PCR kit (Takara Bio Inc., Otsu, Shiga, Japan). The amplicons were purified from 0·8% electrophoresis grade agarose (MP Biomedicals, Aurora, OH, USA) gels as described by Saunders (Reference Saunders1993) and sequenced using the T16 forward primer (Harper and Saunders, Reference Harper and Saunders2001), following the method of Randhawa et al. (manuscript submitted), the ‘ABI PRISM® Big Dye™ Terminator Cycle Sequencing Ready Reaction Kit v.3.1’ in a 16 capillary 3100 Genetic Analyzer (Applied Biosystems). In order to test the accuracy of the data obtained from a single primer, the other strand (T30 reverse primer; Harper and Saunders (Reference Harper and Saunders2001) ) was sequenced for 2 mature E. vernetae from L. erinacea and 1 from L. ocellata; 1 mature E. canadensis from A. radiata; and 1 immature specimen from each of the 4 host species. This region included the D2 domain of the LSU.
Sequence data were edited using Sequencher 4.5 (Gene Codes Corporation, ©1991–2005) and subsequently aligned using MacClade 4.07 (Maddison and Maddison, Reference Maddison and Maddison2005). The transversional model (TVM) was determined to provide the best fit to the data based on Modeltest 3.7 (Posada and Crandall, Reference Posada and Crandall1998; Posada and Buckley, Reference Posada and Buckley2004). The neighbour-joining algorithm, implemented in PAUP v.4.0b10 (Swofford, Reference Swofford2002), was used for the visual display of the within-species variation versus the between-species differences. The purpose of this analysis was not to construct phylogenies; rather it served to assign specimens to a particular cluster, each of which represents a different species.
RESULTS
The size of the amplicons was ∼1850 bp. The region sequenced was ∼800 bp, of which 541 bp (central) were used for sequence alignment and analysis. Sequences resolved as 4 distinct clusters assignable to: Pseudanthobothrium n.sp., P. hanseni, E. vernetae and E. canadensis (Fig. 1). The dissimilarity between species, expressed as percentage of nucleotide difference, was between 2·59 and 8·69%, whereas within-species variation was 0–0·74%. All 9 sequences assignable to those from mature E. vernetae specimens were identical, regardless of whether specimens were recovered from L. erinacea or L. ocellata (Figs 1 and 2). Additionally, all 3 sequences assignable to those from mature E. canadensis were identical (Figs 1 and 2). The genetic distance between mature specimens of both Echeneibothrium spp. was 2·96%, whereas that between mature specimens of both Pseudanthobothrium spp. was 2·59–3·33%. The within-species variation among the 30 mature specimens of Pseudanthobothrium n.sp. was <0·55%, whereas no variation was observed among the 22 mature specimens of P. hanseni.
Phylogram (neighbour-joining) displaying four clusters: one for the included specimens for Pseudanthobothrium n.sp., one for P. hanseni, one for Echeneibothrium canadensis, and one for E. vernetae. Each cluster is accompanied by maturity level (mature or immature), host species, voucher numbers and GenBank Accession numbers for individual isolates.
Of the 38 sequences from immature Pseudanthobothrium recovered from L. erinacea, all were assignable to Pseudanthobothrium n.sp., as were the 3 sequences for immature Pseudanthobothrium isolates recovered from L. ocellata (Figs 1 and 2). Of the 24 sequences from immature Pseudanthobothrium recovered from M. senta, all were assignable to P. hanseni, as were the 3 sequences of immature Pseudanthobothrium recovered from A. radiata (Figs 1 and 2). The sequences of all 16 immature Echeneibothrium specimens recovered from L. erinacea were assignable to those from E. vernetae; of the sequences from 8 Echeneibothrium specimens recovered from A. radiata, 5 were assignable to those from E. vernetae and 3 were assignable to those from E. canadensis (Figs 1 and 2).
The genetic distance between immature and mature specimens for all 4 species was <0·74% for Pseudanthobothrium n.sp., <0·18% for P. hanseni, 0% for E. canadensis, and <0·18% for E. vernetae. These values are all within the range expected for the variable D2 domain of the LSU within tetraphyllidean species as recorded here and reported previously (Brickle et al. Reference Brickle, Olson, Littlewood, Bishop and Arkhipkin2001; Reyda and Olson, Reference Reyda and Olson2003; Agusti et al. Reference Agusti, Aznar, Olson, Littlewood, Kostadinova and Raga2005; Randhawa et al. manuscript submitted). A summary of genetic differences is presented in Fig. 2.
Matrix summarizing the number of actual nucleotide differences (out of 541 bp) among sequences of Pseudanthobothrium n.sp., P. hanseni, Echeneibothrium canadensis and E. vernetae for both mature and immature specimens analysed from the four different host species. IPpLe, Immature Pseudanthobothrium n.sp. ex Leucoraja erinacea; PpLe, Mature Pseudanthobothrium n.sp. ex L. erinacea; IPpLo, Immature Pseudanthobothrium n.sp. ex L. ocellata; PpLo, Mature Pseudanthobothrium n.sp. ex L. ocellata; IPhAr, Immature P. hanseni ex Amblyraja radiata; PhAr, Mature P. hanseni ex A. radiata; IPhMs, Immature P. hanseni ex Malacoraja senta; PhMs, Mature P. hanseni ex M. senta; IEcAr, Immature Echeneibothrium canadensis ex A. radiata; EcAr, Mature E. canadensis ex A. radiata; IEvLe, Immature E. vernetae ex L. erinacea; EvLe, Mature E. vernetae ex L. erinacea; EvLo, Mature E. vernetae ex L. ocellata; IEvAr, Immature E. vernetae ex A. radiata.
Prevalence, defined as the proportion of hosts examined infected with one or more individuals of a given parasite species (Margolis et al. Reference Margolis, Esch, Holmes, Kuris and Schad1982; Bush et al. Reference Bush, Lafferty, Lotz and Shostak1997), of Pseudanthobothrium spp. is high among all 4 rajid host species (60·0–85·7%), whereas that of Echeneibothrium spp. is lower and more variable (18·2–52·4%). The prevalences of E. canadensis in A. radiata and E. vernetae in L. ocellata are <30% (Table 2). The intensity of infection, defined as the mean number of parasites of a given species per infected host (Margolis et al. Reference Margolis, Esch, Holmes, Kuris and Schad1982; Bush et al. Reference Bush, Lafferty, Lotz and Shostak1997), of Pseudanthobothrium spp. is almost double that of Echeneibothrium vernetae (9·9–14·2 versus 6·6–8·5 per infected host, respectively). The intensity of infection of E. canadensis is 1·5, with a range of 1 or 2 specimens per infected A. radiata. All prevalence, intensity of infection and range data are presented and summarized in Table 2.
Summary of the prevalence and intensity of infection (range) for Pseudanthobothrium n.sp., P. hanseni, Echeneibothrium canadensis and E. vernetae
a Immature E. vernetae and E. canadensis specimens from A. radiata were not included as only a fraction were identified using the molecular marker. Others could not unequivocally be identified due to absence of species-diagnostic features.
DISCUSSION
Using a variable region of the LSU, we were able to assign unequivocally immature tetraphyllideans to known species. The present results indicate that mature E. vernetae specimens are found in both Leucoraja spp., whereas mature E. canadensis specimens are restricted to A. radiata. The lack of sequence variation among mature specimens of Echeneibothrium spp. is consistent with the within-species variation reported by Randhawa et al. (manuscript submitted) among mature specimens of both Pseudanthobothrium n.sp. (<0·31%) and P. hanseni (<0·47%) over 643 bp. These findings indicate that, similarly to mature isolates of Pseudanthobothrium n.sp. and P. hanseni (see Randhawa et al. manuscript submitted), mature isolates of E. vernetae are shared by 2 species of rajid skates. The recovery of mature Pseudanthobothrium n.sp., P. hanseni and E. vernetae, each from 2 host species, challenges the dogma surrounding the strict host specificity of these parasites (generally accepted as 1 tetraphyllidean species being restricted to 1 host species). Recording E. vernetae from both L. erinacea and L. ocellata contradicts Williams (Reference Williams1966), who stated that: “… no mature specimens of any one species of Echeneibothrium have been found in more than one host species …” (p. 268). Conversely, E. canadensis seems to exhibit strict (or oioxenous) host specificity, however, increased sampling effort is required in order to confirm this observation as only 3 mature specimens were recovered during this study. The present results highlight the need to extend studies investigating the host specificity of tetraphyllidean cestodes to other genera (also see Randhawa et al. manuscript submitted).
For both Pseudanthobothrium n.sp. and P. hanseni, levels of within-species variation (0·18–0·55%) and between-species differences (2·59–3·33%) in sequence for immature specimens were consistent with those reported for adult specimens of the same species (0·31–0·47% and 2·64–3·42%, respectively, for 643 bp.) (Randhawa et al. manuscript submitted). Also, levels of within-species variation and between-species differences (0·18% and 2·96–3·14%, respectively) for immature specimens of E. vernetae and E. canadensis were consistent with those reported from mature specimens in this study (0% and 2·96%, respectively). Thus, all immature specimens were assigned unequivocally to 1 of the 2 Pseudanthobothrium spp. or 1 of the 2 Echeneibothrium spp. Molecular results indicated that all immature specimens of Pseudanthobothrium n.sp. were restricted to L. erinacea and L. ocellata, whereas those of P. hanseni were restricted to A. radiata and M. senta. These results are consistent with observations of mature parasite-host relationships for these species (Randhawa et al. manuscript submitted), which suggests that host specificity for species of the genus Pseudanthobothrium is expressed early in this host-parasite relationship. Either plerocercoids of Pseudanthobothrium n.sp. are not encountered by A. radiata and M. senta (and similarly plerocercoids of P. hanseni are not encountered by L. erinacea and L. ocellata) or some mechanism(s) prevents their attachment, and therefore establishment of either plerocercoid in the ‘wrong’ host. Differing host behaviours, such as substrate preferences and different feeding habits of hosts, close the encounter filter of the host spectrum, whereas the absence of parasite adaptations necessary to overcome host defences or the incompatibility of parasite adaptations to host resources (spatial or metabolic) close the compatibility filter of the host spectrum. All 4 skate species are known to be sympatric over their geographical range (McEachran and Musick, Reference McEachran and Musick1975; McEachran et al. Reference McEachran, Boesh and Musick1976, and references therein). However, L. erinacea and L. ocellata prefer sandy and gravely bottoms (Packer et al. Reference Packer, Zetlin and Vitaliano2003a, Reference Packer, Zetlin and Vitalianob), whereas M. senta prefers soft, muddy substrate (Packer et al. Reference Packer, Zetlin and Vitaliano2003c). Amblyraja radiata shows little substrate preference (Packer et al. Reference Packer, Zetlin and Vitaliano2003d) but is positively associated to M. senta (McEachran and Musick, Reference McEachran and Musick1975; McEachran et al. Reference McEachran, Boesh and Musick1976), whereas this ecological species-pair is negatively associated with the L. erinacea and L. ocellata ecological species-pair (McEachran and Musick, Reference McEachran and Musick1975). These substrate preferences (behaviour parameter of the encounter filter) and corresponding prey biota explain, at least in part, the presence of Pseudanthobothrium n.sp. in 1 ecological species-pair and its absence from the other, and vice versa for P. hanseni.
Furthermore, molecular results indicated that all immature specimens of E. canadensis were specific to A. radiata, whereas those of E. vernetae were recovered from A. radiata and L. erinacea. The specificity of immature E. canadensis is not surprising, since it reflects that of mature specimens, but should be considered as preliminary for reasons stated earlier. The low prevalence and intensity of infection of immature Echeneibothrium specimens in A. radiata, and non-recovery from L. ocellata, lead to an overestimation of the specificity of E. canadensis. High specificity is possibly an artefact of inadequate sampling (Poulin, Reference Poulin1998). Williams (Reference Williams1966) also stated that: “… only on very rare occasions were immature specimens of a [Echeneibothrium] species found in two species of Raja …” (p. 268). Although the recording of immature E. vernetae specimens from A. radiata and L. erinacea supports the potential for recovering immature Echeneibothrium spp. from 2 rajid hosts (Williams, Reference Williams1966), it questions the rarity of this event. Of the 8 immature Echeneibothrium specimens recovered from A. radiata, 5 were assignable to E. vernetae, a parasite specific to L. erinacea and L. ocellata once mature. Additionally, the recovery of mature E. vernetae from L. ocellata implies that immature E. vernetae occur in this host and should be recovered with intensified sampling effort. These results suggest that host specificity of the genus Echeneibothrium is not apparent from our sampling of immature specimens and that worms can establish (or attach), but are not able to mature (or develop) in the ‘wrong’ host species.
Host resistance (restriction) occurs when the host's defence mechanisms (e.g. immune system) prohibit the development of the parasite, i.e., death or detachment of the parasite occurring pre-establishment (or post-attachment) or prohibit the maturation of the parasite. Host resistance, or restriction, has been shown experimentally by exposing Acanthobothrium quadripartitum, a tetraphyllidean cestode, to serum from the ‘wrong’ host, which led to the death of 80% of the worms within 2 h (McVicar and Fletcher, Reference McVicar and Fletcher1970) (an example of death occurring post-attachment); and by transferring the host-specific monogenean Entobdella soleae onto the ‘wrong’ host, which led to it detaching within 30 h (see Rohde and Rohde, Reference Rohde and Rohde2005) (an example of detachment occurring post-attachment). The presence of immature, and the absence of mature, E. vernetae specimens from A. radiata caused by host resistance is a plausible hypothesis, however, an experiment exposing immature and mature E. vernetae specimens to various rajid skate sera is necessary for its confirmation.
Neither host incompatibility nor host resistance hypotheses have been tested here, therefore, neither can be confirmed nor ruled out as an explanation for the presence of immature E. vernetae in A. radiata and the specificity of E. canadensis. Parasitological surveys of wild rajid skates are inadequate in addressing hypotheses of host resistance or unsuitability. Therefore, experimental infections are the only valid alternatives for testing either of these two hypotheses. Support for the host resistance hypothesis would assume that immature E. vernetae in A. radiata were recently acquired infections (hours or days) and would not be ‘able’ or ‘allowed’ to establish over the long term. Host unsuitability would assume that E. canadensis larvae are encountered by M. senta, but are unable to attach. It is also possible that host resistance or host unsuitability is/are involved in the specificity of Pseudanthobothrium spp., but was not observed. Dissections only provide a glimpse into an otherwise dynamic relationship between hosts and parasite populations/communities and do not offer the means necessary to make strong inferences on past assemblages (hours or days), therefore reiterating the importance of experimental infections in order to test these hypotheses.
Furthermore, knowledge of the ecology of skate hosts allows certain inferences to be made. Keeling and Burt (Reference Keeling and Burt1996) reported that the prevalence and intensity of infection for E. canadensis were 13·7% and 1·4 E. canadensis per infected A. radiata (range of 1 to 2), respectively. Prevalence, intensity of infection and range for E. canadensis reported herein are consistent with those published (Keeling and Burt, Reference Keeling and Burt1996). Although A. radiata and M. senta are sympatric and have a high coefficient of association (McEachran and Musick, Reference McEachran and Musick1975), M. senta has specialized its feeding habits in response to possible competition with A. radiata for resources (McEachran et al. Reference McEachran, Boesh and Musick1976). This has led to lower diversity of prey species in the diet of M. senta (McEachran et al. Reference McEachran, Boesh and Musick1976) even though they share some of the more abundant prey species (Packer et al. Reference Packer, Zetlin and Vitaliano2003c, Reference Packer, Zetlin and Vitalianod). This suggests that P. hanseni may be transmitted via one of these common prey items (possibly an amphipod), as it is a cestode common to both rajid skates (Randhawa et al. manuscript submitted), whereas E. canadensis may be transmitted via a larger prey item specific to A. radiata (possibly infauna, e.g. polychaete worm). It is assumed here that the larger prey items are less abundant (number of individuals) in the host diet and that both Pseudanthobothrium spp. and Echeneibothrium spp. are transmitted in similar numbers during each infection event, thus explaining the greater intensity of infection of Pseudanthobothrium spp. compared to that of Echeneibothrium spp.
Euzet (Reference Euzet1956) did not publish prevalence and intensity data for E. vernetae when he described this species based on material collected by Linton (Reference Linton1889) from L. erinacea, nor did Linton (Reference Linton1924) from material collected between 1905 and 1913 from L. erinacea, which he described as E. variabile (later recognized as E. vernetae by Euzet in Reference Euzet1956), therefore contributing little to the understanding of the ecology of the parasite. Prevalence of infection for E. vernetae reported herein (52·4% in L. erinacea vs 28·6% in L. ocellata) suggests that L. erinacea is the preferred host for this cestode in the area sampled and that the relatively high abundance (3·5 in L. erinacea and 2·4 in L. ocellata) of the parasite indicates that infections are not acquired accidentally in either host species. Experimental infections tracking the development of E. vernetae in both host species and investigating host suitability (e.g. stunted growth in 1 host species, lower biotic potential in 1 host species, etc.) would provide useful information for host preference. If differential fitness is observed, then it could be assumed that one host species is the preferred host, whereas the other is less suitable and may provide evidence for the trade-off hypothesis (trade-off between adapting to a new host species versus the ability to reach high abundance in that host species) (see Poulin and Mouillot, Reference Poulin and Mouillot2004).
It is generally accepted that ecological factors often drive host specificity (e.g. Holmes, Reference Holmes, Esch, Bush and Aho1990; Rohde and Rohde, Reference Rohde and Rohde2005) and that the availability of ‘suitable’ hosts is necessary for the successful colonization of a new host species (Poulin, Reference Poulin1992). Leucoraja erinacea is one of the commonest demersal fishes in the Northwest Atlantic (Packer et al. Reference Packer, Zetlin and Vitaliano2003a, and references therein) whereas L. ocellata is sympatric to L. erinacea over most of its range (McEachran and Musick, Reference McEachran and Musick1975; Packer et al. Reference Packer, Zetlin and Vitaliano2003b) its abundance is much lower than that of L. erinacea in the Bay of Fundy and Passamaquoddy Bay, as shown by number of fish examined for this study (84 vs. 7, respectively) and reported in Packer et al. (Reference Packer, Zetlin and Vitaliano2003a, Reference Packer, Zetlin and Vitalianob). This relative abundance is consistent with abundance data over the shared range of both sympatric Leucoraja spp. as reported in Packer et al. (Reference Packer, Zetlin and Vitaliano2003a, Reference Packer, Zetlin and Vitalianob). Higher prevalence of E. vernetae in L. erinacea may be an artefact of relative abundance of both skate species. Host relative abundance (or host availability) can skew host-use by parasites and render host specificity indices unreliable (Poulin, Reference Poulin1998). It is therefore suggested herein that investigations of host specificity should provide the measures of prevalence and intensity (when available) to gain a better understanding of parasite ecology and so that host specificity can be measured and compared across studies.
In summary, the onset of host specificity differs between Pseudanthobothrium spp. and Echeneibothirum spp. Specificity in Pseudanthobothirum spp. occurs prior to attachment, whereas specificity in Echeneibothrium occurs post-attachment. Estimating the onset of host specificity was predicted by determining the specificity of mature specimens of each parasite species and comparing the host distribution of immature specimens of those same species. The host specificity of immature P. hanseni, Pseudanthobothrium n.sp. and E. canadensis mirrors that of mature specimens of their respective host species. The recovery of immature E. vernetae from L. erinacea and A. radiata was unexpected and somewhat surprising. Since A. radiata is not host to mature specimens of E. vernetae, the presence of immature cestodes of that species in A. radiata may indicate that host restriction or resistance is involved, allowing the attachment of the parasite, but not its establishment.
We are grateful to HMSC staff, in particular: F. Purton for his technical assistance, availability and professionalism; D. Parker for his technical assistance; Peter Rose and Jeff Markey for assistance with collection of specimens; and E. Carter who went beyond the call of duty as Captain of the W. B. Scott R/V. The help of D. Loveless and W. Minor, Mate/Engineer and Captain of the CCGS Pandalus III, respectively, and Department of Fisheries and Oceans Canada (DFO) personnel in collecting is also gratefully acknowledged. The senior author thanks L. LeGall for her advice and help with GenBank submissions, as well as T. Moore and M. Surette for their technical assistance in the Saunders' Lab. The HMSC and Department of Biology of UNB provided lab space and research facilities. Funding and support to H. S. R. was provided through UNB Graduate Teaching and Research Assistantships, one UNB R. C. Frazee Research Scholarship at HMSC, a W. B. Scott Graduate Research Scholarship in Ichthyology, two Marguerite and Murray Vaughan Graduate Fellowships, and one UNB Alumni Student Merit Award. The Natural Sciences and Engineering Research Council of Canada (NSERC) provided financial support though operating/discovery grants to G. W. S. and M. D. B. B.; a Steacie Fellowship to G. W. S.; and a Major Facilities Access Grant to HMSC. The Canada Research Chairs Program (CRC) also provided financial assistance to G. W. S.
