INTRODUCTION
Parasite communities of extant hosts are the result of repeated additions and losses of parasite species during evolutionary history and geological time. Developing over time, the ecological and biological characteristics of hosts and parasites determine host colonization, intra-host diversification and rates of parasite extinction, consequently influencing parasite community richness and diversity. The study of the factors influencing host-parasite relationships in vertebrates has been gaining the interest of ecological and evolutionary parasitologists in recent years, as illustrated in the increasing number of papers published on this subject in the last 2 decades (e.g. Poulin and Rohde, Reference Poulin and Rohde1997; Krasnov et al. Reference Krasnov, Shenbrot, Khokhlova and Degen2004; Korallo et al. Reference Korallo, Vinarski, Krasnov, Shenbrot, Mouillot and Poulin2007; Verneau et al. Reference Verneau, Du Preez and Badets2009).
Host age, size, diet, habitat, behaviour, depth distribution and geographical range have all been recognized as variables influencing richness and diversity of parasite communities (e.g. Poulin and Rhode, Reference Poulin and Rohde1997; Lile, Reference Lile1998; Poulin and Morand, Reference Poulin and Morand2000; Luque et al. Reference Luque, Mouillot and Poulin2004; Aguirre-Macedo et al. Reference Aguirre-Macedo, Vidal-Martínez, González-Solís and Caballero2007; Korallo et al. Reference Korallo, Vinarski, Krasnov, Shenbrot, Mouillot and Poulin2007; Luque and Poulin, Reference Luque and Poulin2008). However, there is no consensus about the relative importance of the host's traits in the evolution and diversification of the parasite fauna in general, suggesting that relationships between host features and parasite diversity are specific for each host-parasite association in a particular area (Aguirre-Macedo et al. Reference Aguirre-Macedo, Vidal-Martínez, González-Solís and Caballero2007; Korallo et al. Reference Korallo, Vinarski, Krasnov, Shenbrot, Mouillot and Poulin2007; Luque and Poulin, Reference Luque and Poulin2008). Moreover, universal patterns in parasite community diversity have not yet been found (Poulin, Reference Poulin2007) and, for a given host group (e.g. marine fishes), the same host characteristic has shown a different influence, for instance, whereas Raibaut et al. (Reference Raibaut, Combes and Benoit1998) and González and Poulin (Reference González and Poulin2005) found no relationship between fish body size and parasite species richness, a positive relationship between these two variables was found by Timi and Poulin (Reference Timi and Poulin2003).
These inconsistencies might be due to different sources of data or to data being collected and/or analysed by different people using different methods (Poulin and Rhode, Reference Poulin and Rohde1997; Luque et al. Reference Luque, Mouillot and Poulin2004; Luque and Poulin, Reference Luque and Poulin2008). However, parasite species found within a host's distribution range are variable between geographical areas, limiting the number and type of species that can be acquired, regardless of host's characteristics, which is suggestive of an important role for host and parasite biogeography (Luque et al. Reference Luque, Mouillot and Poulin2004; Luque and Poulin, Reference Luque and Poulin2008). On the other hand, host and parasite phylogeny influence the composition of extant parasite communities (Poulin and Morand, Reference Poulin and Morand2000; Aguirre-Macedo et al. Reference Aguirre-Macedo, Vidal-Martínez, González-Solís and Caballero2007; Poulin, Reference Poulin2007) since hosts might have inherited their parasites from their ancestors by 1 of the 4 main categories of evolutionary events (Verneau et al. Reference Verneau, Du Preez and Badets2009): co-speciation (synchronous divergence of host and parasite lineages over geological time); duplication (within host speciation of parasite lineages without the concomitant speciation of host lineages); host-switching (the parasite lineage colonizes a phylogenetically or ecologically close host species to its ancestral host); or extinction (following host speciation only 1 daughter species is parasitized by the ancestral parasite lineage – ‘missing the boat’ hypothesis – or both daughter species preserve the ancestral parasite lineages but it goes extinct in 1 of them – ‘drowning on arrival’ hypothesis). As such, and for a better understanding of the relative importance that host biology, ecology and phylogeny and environmental characteristics have on parasite richness and diversity patterns, these should be analysed in a group of related species inhabiting the same area (Luque et al. Reference Luque, Mouillot and Poulin2004; Korallo et al. Reference Korallo, Vinarski, Krasnov, Shenbrot, Mouillot and Poulin2007) and integrating both ecological and phylogenetic approaches (Poulin and Morand, Reference Poulin and Morand2000). Despite this recognition, most studies have focused solely on the effects of ecological factors (e.g. Krasnov et al. Reference Krasnov, Shenbrot, Khokhlova and Degen2004; Korallo et al. Reference Korallo, Vinarski, Krasnov, Shenbrot, Mouillot and Poulin2007) and the few that take into account the importance of host phylogeny have usually considered it a confounding variable as closely related hosts are likely to harbour similar parasite communities (e.g. Poulin and Rhode, 1997; Luque and Poulin, Reference Luque and Poulin2008).
In the present study, the macroparasite fauna of 14 species of flatfish (Pleuronectiformes) inhabiting the Portuguese coast was investigated to assess the relative importance of several host features in the diversity of their macroparasite assemblages. Although monophyletic (Azevedo et al. Reference Azevedo, Oliveira, Pardo, Martinez and Foresti2008), this group of fishes comprises many different species, genera and families, displaying a wide variety of ecological and life-history patterns, including biological, environmental and dietetic requirements, and a relatively high species diversity within the Portuguese coast (25 species belonging to 13 genera and 5 different families). Therefore, Pleuronectiformes constitute a valuable group to evaluate the relative importance of host biology, ecology and phylogeny on the structure of host-parasite associations given the identical parasite species pool and environmental characteristics.
The richness (number of species), taxonomic distinctness (Δ+) and the variance of taxonomic distinctness (Λ+) (Clarke and Warwick, Reference Clarke and Warwick1998, Reference Clarke and Warwick1999; Warwick and Clarke, Reference Warwick and Clarke2001) of each ecto- and endoparasite assemblage infecting a host species were calculated and the relative effect of 27 host characteristics on the patterns found were assessed by redundancy analyses in order to evaluate: (1) which host characteristic(s) are more important in the diversity found; (2) whether host phylogeny is, indeed, a good predictor of the structure of parasite communities, i.e. if closer hosts present more similar parasite assemblages; (3) whether the influence of a host characteristic varies according to the parasite taxa and guilds or if it is always identical independently of the type of parasites found; and (4) whether general patterns can be found at this local scale.
MATERIALS AND METHODS
Sampling procedures
Fourteen flatfish species, belonging to 5 different families (Table 1), were collected seasonally between January 2003 and June 2005 from commercial fishing vessels operating with gill nets and bottom trawls along the Portuguese coast. Each fish was measured (total length) and sexed prior to examination of its skin, fins, gills, nostrils and mouth cavity for the detection of ectoparasites. Digestive tract, liver, heart, spleen, gonads and mesenteries were carefully inspected under a stereomicroscope to search for endoparasites. Collected parasite specimens were then identified to the lowest possible taxonomical level, depending on their maturation stage and on the number of individuals available. Parasite collection and identification were always performed by the same researcher, reducing the variance in data associated with different sources of information and different methodologies.
Mean abundance of each taxon infecting each host
Food items, gonads and otoliths were also collected from the same individual hosts and the data on feeding ecology, reproduction and growth were used as the hosts’ ecological and biological characteristics (Table 2).
Host characteristics of the flatfish species included in the analyses
Sources:
1 present study
2 Munroe (Reference Munroe, Gibson and 2005)
3 Froese and Pauly (Reference Froese, Pauly and 2009)
4 Deniel (Reference Deniel1981)
5 Cabral et al. (Reference Cabral, Catarino, Figueiredo, Garcia and Henriques2003)
6 Teixeira et al. (Reference Teixeira, Pinheiro and Cabral2009)
7 Teixeira et al. (Reference Teixeira, Batista and Cabral2010)
8 Teixeira and Cabral (Reference Teixeira and Cabral2010).
* Proportion of the size range of fish examined in the present study considering the size at first maturity as the minimum and the maximum size reported as the maximum.
+ Degrees south are indicated by (–).
Data analyses
Mean abundance of each macroparasite taxa within each host species was calculated according to the method of Bush et al. (Reference Bush, Lafferty, Lotz and Shostak1997), i.e. the mean number of parasites of a given taxon per host. It has been shown that mean abundance values of macroparasite taxa infecting Pleuronectiformes are unaffected by a sample size larger than 40 individuals (Marques and Cabral, Reference Marques and Cabral2007). Thus, no corrections for sample sizes were performed despite the wide range of host individuals surveyed within each species. To evaluate the diversity of macroparasite assemblages, the total number of taxa (richness), taxonomic distinctness (Δ+) and the variance of taxonomic distinctness (Λ+) were calculated considering ectoparasites and endoparasites separately as they have different transmission pathways and life cycles (MacKenzie, Reference MacKenzie2002). Cases where only 1 parasite individual infected a single host individual (Gyrodactylus sp., Anisakis typical (Diesing, 1860), Cucullanus heterochrous Rudolphi, 1802, Lernaeocera sp.) were not included in the calculation of diversity indices, as they might be accidental infections.
For 2 host species presenting identical macroparasite richness some features may promote the acquisition of a broader taxonomic range of parasites whereas others may contribute to a narrower range. Hence, these assemblages may be different and, therefore, the present study considers measures of diversity that take species relatedness into account, i.e. Δ+ and Λ+. Moreover, whereas richness is highly dependent on sample size, ‘taxonomic’ measures are not (Clarke and Warwick, Reference Clarke and Warwick1998). Taxonomic distinctness (Δ+) is the expected path length between any 2 species in a sample and is calculated as the number of steps taken to reach a common taxon (Clarke and Warwick, Reference Clarke and Warwick1998, Reference Clarke and Warwick1999; Warwick and Clarke, Reference Warwick and Clarke2001): 0 for the same species, 1 for different species of the same genus, 2 for different genera within a single family. The variance in taxonomic distinctness (Λ+) reflects the representativeness of each taxon in the assemblage (Warwick and Clarke, Reference Warwick and Clarke1995). Λ+ can only be computed when at least 3 parasite taxa are found in a host sample (it always equals zero with 2 species); however, it is only reliable for assemblages with at least 4 parasite taxa (Luque et al. Reference Luque, Mouillot and Poulin2004). Macroparasite taxa were placed within the Linnaean taxonomic classification according to their species, genus, family, order, class, phylum and kingdom as indicated in the Natural History Museum host-parasite database (Gibson et al. Reference Gibson, Bray and Harris2008) and the European register of marine species database (MarBEF Data System, 2008) and Δ+ and Λ+ were calculated using PRIMER 5.2.8 (Clarke and Gorley, Reference Clarke and Gorley2001). The higher the values of Δ+ the lower the relatedness between the macroparasite species found within a host. The higher the values of Λ+, the lower the evenness within macroparasite assemblages.
A total of 27 phylogenetic, biological and ecological host characteristics expected to have some degree of influence on the composition of macroparasite assemblages and infection levels of each macroparasite taxa within each host species were used (Table 2). Given that mean abundance and diversity values were not expected to be influenced by differences in sample size, as these were all larger than 40 individuals (Marques and Cabral, Reference Marques and Cabral2007), sample size was not considered as a variable in our study. The total number of genera and species known worldwide for each host family, as well as host family itself, were considered as phylogenetic characteristics. Reproductive season, zoogeographical classification and the proportional length of the gut of each host species were transformed from nominal to ordinal variables in the analysed data matrix. The minimum latitude reported for each species was transformed into negative values for southern latitudes in order to allow the calculation of host latitudinal range. The size range of the fishes examined in this study, in relation to the size range reported for each species, was also included as a variable given that body size has been shown to influence macroparasite communities (e.g. Luque et al. Reference Luque, Mouillot and Poulin2004; Muñoz et al. Reference Muñoz, Grutter and Cribb2006) and very large individual hosts were only obtained for some of the host species inhabiting the Portuguese coast. Differences in the amount of epidermal mucus segregated by the 14 species were also detected when examining them for ectoparasites and, therefore, it was considered as a variable in ectoparasites’ analyses. Although all Pleuronectiformes included in the present study are in the same trophic level (values between 3 and 4 according to Froese and Pauly (Reference Froese, Pauly and 2009)), their different depth distribution and maximum size might result in different predation pressures exerted on them by other species (e.g. marine mammals) that, in turn, will influence the probability of endoparasite species completing their life cycle. Therefore, predation pressure can affect macroparasite assemblage composition and infection levels and was also included as a variable. Because endoparasites are differently distributed along the gut, or found in particular regions within it (e.g. pyloric caeca), the length of the gut in relation to host total length and the morphological differentiation of the gut were considered as variables in the analyses of endoparasite assemblages. The number of different prey ingested and the inclusion or absence of certain groups of taxa in the diet were also used in endoparasite analyses.
Redundancy analysis (RDA) was applied to investigate the relationships between host characteristics and the observed patterns of mean abundance, richness, Δ+ and Λ+. RDA is generally used to visualize linear correlations between explanatory variables, response variables and samples and to determine the relative importance of each explanatory variable and its significance in the variation observed in a data set. This multivariate technique was used to assess the relative importance of the phylogenetic, biological and ecological characteristics of hosts in (1) the taxa found infecting them and the mean abundance of each taxon and (2) the richness and diversity of their macroparasite assemblages. Analyses were performed in CANOCO 4.5 (ter Braack and Smilauer, Reference ter Braak and Smilauer2002) and the significance of each host characteristic to the identified patterns evaluated by permutation tests considering α=0·05.
RESULTS
Of the 4482 flatfish examined, 53 macroparasite taxa (23 518 individuals) were found (Table 1). Although Digenea, Cestoda and Nematoda were the most frequent and abundant groups infecting almost every host, Monogenea, Acanthocephala, Copepoda, Isopoda, Pentastomida and Hirudinea were also present. Whereas the majority of the taxa infected several hosts, others were more specific and found in only 1 or 2 flatfish species; nevertheless, and even for the less specific macroparasites, a considerably higher mean abundance was generally observed in one of the host species (Table 1). The largest values of mean abundance found for ectoparasites were registered in Acanthochondria cornuta (Müller, 1776) and Lepeophtheirus pectoralis (Müller, 1776), both infecting Platichthys flesus (Linnaeus, 1758). Otodistomum sp. infecting Solea kleinii (Bonaparte, 1832) and Acanthocephaloides propinquus (Dujardin, 1845) infecting Solea senegalensis Kaup, 1858 presented the highest mean abundance values obtained for endoparasites.
Richness of ectoparasite species ranged between zero, in Microchirus variegatus Desoutter, 1990, and 7, in S. senegalensis and Solea lascaris (Risso, 1810) (Tables 1 and 3). Values of taxonomic distinctness (Δ+) and its variance (Λ+), only computed for hosts presenting at least 4 ectoparasite taxa represented by more than 1 individual, were higher in Dicologlossa cuneata (Moreau, 1881) and Solea solea (Linnaeus, 1758), respectively (Table 3). The ectoparasite species infecting D. cuneata were, therefore, the least related and comprised of Monogenea, Copepoda and Hirudinea (Table 1) whereas the ectoparasite assemblage of S. solea was the less taxonomically even, with most of its ectoparasites being Copepoda (Table 1). This proportion of Copepoda also contributed to S. solea presenting the lowest value of Δ+, i.e. the assemblage comprising the most related ectoparasite taxa (Table 3). The lowest Λ+ was found in S. lascaris (Table 3) indicating that this host harboured the most taxonomically even ectoparasite assemblage.
Diversity of parasite assemblages within each host species
Endoparasite assemblages presented the highest richness in D. cuneata, S. lascaris and S. senegalensis, all infected by 14 taxa, although 3 of the parasite taxa found in S. lascaris were only represented by 1 individual and can, therefore, be considered accidental infections. The lowest endoparasite richness was found in Scophthalmus maximus (Linnaeus, 1758), infected by only 2 species (Tables 1 and 3). Citharus linguatula (Linnaeus, 1758) presented the highest Δ+ and S. lascaris the highest Λ+ (Table 3). Although presenting only 6 taxa, the endoparasite assemblage of C. linguatula comprised at least 1 species of each group, thus it was the most distinct. On the contrary, the endoparasite assemblage of S. lascaris was comprised of many taxa of the same genus or family (Table 1). S. senegalensis presented the lowest value of Δ+, related to the many species within the Digenea and Cestoda infecting this host, and C. linguatula the lowest Λ+, in agreement with its low number of Digenea, Cestoda, Acanthocephala and Nematoda parasites (Table 1).
The redundancy analysis (RDA) performed on ectoparasite taxa (Fig. 1A) evidenced the zoogeographical classification of S. senegalensis, S. solea, S. lascaris and S. kleinii (subtropical species, Table 2) as the main characteristic influencing their higher mean abundance of Bomolochus soleae (Claus, 1864), Caligus brevicaudatus Scott, 1901, Entobdella solea van Beneden and Hesse, 1864 and Hemibdella soleae van Beneden and Hesse, 1863 (all in the left half of the RDA diagram). A larger latitudinal range was directly related to the higher mean abundance of A. cornuta and L. pectoralis found in P. flesus (right half of the diagram). Although the first (horizontal) axis explained 90% of all variation in host characteristics-ectoparasite taxa relationships, the second (vertical) axis accounted for another 10% of all variation, suggesting that the ectoparasite assemblages of these 5 hosts are influenced by adults inhabiting brackish environments (except S. lascaris) and by spawning in colder seasons (autumn-winter or winter-spring) (Table 2). The size and direction of the vectors ‘Life span’, ‘Number of genera’ and ‘Number of species’ was also indicative of the importance of these variables in the type and mean abundance of ectoparasite taxa found in Solea spp. (Fig. 1A). However, the permutations test performed (1000 permutations) revealed that only zoogeography and latitudinal range of hosts had a significant effect (P=0·010 and P=0·017, respectively). When the RDA was performed on the diversity of the ectoparasite assemblages (Fig. 1B) only the 5 host species infected with at least 4 ectoparasite taxa could be considered. Entering brackish waters, reproducing in autumn-winter and attaining a larger size than other analysed species were the most important characteristics determining the richness and variance in taxonomic distinctness of S. solea and S. senegalensis (right half of the diagram) in comparison to the other 3 host species. Because the first RDA axis explained 99% of the variation in host characteristics-ectoparasite taxa, all the other host-characteristics represented in the diagram contributed little to the attained diversity values, despite the size of the vectors displayed in the diagram. The permutations test also revealed that only the maximum size reported was significantly important for the observed pattern (P=0·036).
Redundancy analysis (RDA) between host biological, ecological and phylogenetic characteristics as independent variables and the mean abundance of each ectoparasite taxa (A) or ectoparasite assemblage diversity measures (B) as dependent variables. Vectors represent host characteristics, circles represent host species, open triangles represent ectoparasite taxa (A) and filled triangles represent diversity measures (B). Acronyms for parasite and host taxa are given in Tables 1 and 2, respectively. Acronyms for diversity measures are: TaxDis, taxonomic distinctness; varTaxDis, variance in taxonomic distinctness. Grey symbols correspond to clumped hosts and ectoparasite taxa whose acronyms are not shown due to space constrictions.
The first 2 RDA axes explained only 64% of the variance in host characteristics – endoparasite mean abundance relationships, evidenced by the almost identical size of the vectors representing the more important variables, nevertheless S. kleinii, S. senegalensis, S. lascaris and D. cuneata appeared separated from the remaining species (Fig. 2A). Whereas the higher mean abundance of Otodistomum sp. and Anisakis simplex (Rudolphi, 1809) found in S. kleinii were mainly related to the larger proportion of size range examined and larger latitudinal range of this Soleidae compared to the other 4 (Table 2), inhabiting brackish waters and spawning in autumn-winter were the 2 host characteristics most contributing to the higher mean abundance of Scolex pleuronectis Müller, 1788, Acanthocephaloides propinquus (Dujardin, 1845) and Cucullanus campanae Lebre & Petter, 1985 in S. senegalensis (Fig. 2A, Table 2). Nevertheless, feeding on Crustacea and ingesting a higher number of different prey also contributed to the mean abundance of the endoparasite taxa found in S. senegalensis and to the differentiation between S. kleinii and the other 4 Soleidae along the first RDA axis (Fig. 2A). These 2 characteristics were also the main ones influencing the higher mean abundance of Derogenes varicus (Müller, 1784), Hemipera sp., Macvicaria soleae (Dujardin, 1845), Prosorhynchus crucibulum (Rudolphi, 1819) Odhner, 1905, Didymobothrium rudolphii (Monticelli, 1890), Nybelin, 1922, Nybelinia lingualis Cuvier, 1817, Acanthocephaloides geneticus (Buron, Renaud and Euzet, 1985), Acanthocephalus incrassatus (Molin, 1858), Hysterothylacium sp. and Pentastomida found in D. cuneata and S. lascaris. The differentiation of S. senegalensis, D. cuneata and S. lascaris along the second RDA axis (Fig. 2A) was also related, to some extent, to the higher growth coefficient obtained for S. lascaris (Table 2). Still, ingesting Crustacea was the only variable with a significant effect on the patterns found (P=0·032). When the RDA was performed on endoparasite diversity, each index was mostly influenced by a type of host characteristics (Fig. 2B): whereas a higher number of different prey and ingesting Mollusca, Crustacea and Echinodermata influenced the higher endoparasite richness and taxonomic distinctness found in some hosts (right half of the diagram), inhabiting rocky bottoms and having a higher growth coefficient were important to the higher variance in taxonomic distinctness found in other hosts (left of the diagram). Although the first axis explained 57% of the variance found, the second contributed with 43%, suggesting that preying on Polychaeta and Cephalochordata and belonging to a family with a higher number of genera were also important characteristics influencing the higher values of richness and variance in taxonomic distinctness found in the majority of the Soleidae and in P. flesus (lower half of the diagram); gut differentiation also contributed to the values of taxonomic distinctness found in C. linguatula and Arnoglossus laterna (Walbaum, 1792) (upper half of the diagram). However, none of the host features presented an overall significant effect in the pattern found.
Redundancy analysis (RDA) between host biological, ecological and phylogenetic characteristics as independent variables and the mean abundance of each endoparasite taxa (A) or endoparasite assemblage diversity measures (B) as dependent variables. Vectors represent host characteristics, circles represent host species, open triangles represent endoparasite taxa (A) and filled triangles represent diversity measures (B). Acronyms for parasite and host taxa are given in Tables 1 and 2, respectively. Acronyms for diversity measures are: TaxDis, taxonomic distinctness; varTaxDis, variance in taxonomic distinctness. Grey symbols correspond to clumped hosts and endoparasite taxa whose acronyms are not shown due to space constrictions. The acronyms for Solea lascaris and its endoparasite taxa, whose symbols are contained in the small square in the left side of diagram A, are outlined in the larger square on the bottom left of the diagram.
DISCUSSION
Results showed considerable differences in the macroparasite assemblages of the 14 analysed flatfish species. Since this study was based on data collected in a single area by the same researchers and all data were analysed using standard methods by the same research team, all these factors can be excluded as possible causes for the observed differences in mean abundance, richness, taxonomic distinctness and its variance between hosts (Luque et al. Reference Luque, Mouillot and Poulin2004). Therefore, these ought to be related to host and parasite phylogenetic, biological and ecological characteristics within the study area, which overall determined the likelihood of host colonization, parasite speciation and extinction and the survival of parasite lineages within new hosts.
Since this study was performed in a group of related hosts, a certain similarity in parasite assemblages was expected (e.g. Poulin and Rhode, Reference Poulin and Rohde1997; Luque et al. Reference Luque, Mouillot and Poulin2004; Luque and Poulin, Reference Luque and Poulin2008), especially for those hosts with similar ecology (Muñoz et al. Reference Muñoz, Grutter and Cribb2006). However, assemblages were not more similar between species within the same genus than between species of different genera or families (e.g. endoparasite assemblage of S. solea and its congeners vs endoparasite assemblage of D. cuneata and S. senegalensis or D. cuneata and P. flesus). In addition, species sharing life-history patterns and feeding ecology, such as S. senegalensis and S. solea (Teixeira and Cabral, Reference Teixeira and Cabral2010) also presented differentiated endoparasite and ectoparasite assemblages even if they had 8 species in common. Nevertheless, and with the exception of C. campanae, B. soleae, C. brevicaudatus and H. soleae, these were generalist parasites that also occur in non-flatfish hosts (Marques et al. Reference Marques, Santos and Cabral2009 and references therein). On the other hand, considering other closely related species presenting similar diets but different biological traits and ecological requirements (e.g. S. lascaris and S. senegalensis), ecto- and endoparasite assemblages were more alike than those of less-related hosts with similar biology and ecology (e.g. D. cuneata and S. lascaris). These results seem to agree to some extent with those of Lile (Reference Lile1998), that found a higher similarity in endoparasitic helminth communities of less-related hosts (P. flesus and Hippoglossoides platessoides (Fabricius, 1780)) than between more-related hosts (P. flesus and Glyptocephalus cynoglossus (Linnaeus, 1758)) in 4 Pleuronectidae from Norway, and to those of Muñoz et al. (Reference Muñoz, Grutter and Cribb2006), that found no consistent pattern in endoparasite communities of 5 Labridae from Lizard Island (Australia).
RDA results revealed that the diversity of macroparasite assemblages of the 14 analysed flatfish species can be explained by only a few host features. Of the characteristics considered in each analysis, only a few were identified as important, as depicted by the size of the vectors in the RDA diagrams. Moreover, depending on considering parasite taxa mean abundance or the diversity of ecto- or endoparasites assemblages (a) different host characteristics have different influence – e.g. zoogeography and latitude range of the host were the most important characteristics determining ectoparasite infection levels but the diversity of ectoparasite assemblages were mostly influenced by habitat, reproductive season and maximum size of the host. (b) A given host characteristic has a different importance according to the macroparasites’ guild analysed – e.g. the maximum size attained by the host was significantly important in the diversity of ectoparasite assemblages but not in the diversity of endoparasite assemblages in the same host species. (c) The same host characteristic may have different effects according to host and parasite taxa. For example, a wider latitude range in host distribution was important in the P. flesus – A. cornuta/L. pectoralis relationships, contributing to higher mean abundance of these parasites, but had the contrary effect on the relationships between D. cuneata or S. kleinii and their ectoparasites, as evidenced by the direction of the vector ‘Latitude range’ regarding these host-parasite pairs. Therefore, these results support the hypothesis that different diversity measures are related with different host characteristics, as suggested in previous studies (Poulin and Rhode, Reference Poulin and Rohde1997; Raibaut et al. Reference Raibaut, Combes and Benoit1998; Luque et al. Reference Luque, Mouillot and Poulin2004; Aguirre-Macedo et al. Reference Aguirre-Macedo, Vidal-Martínez, González-Solís and Caballero2007; Korallo et al. Reference Korallo, Vinarski, Krasnov, Shenbrot, Mouillot and Poulin2007; Luque and Poulin, Reference Luque and Poulin2008). However, these studies used data collected from different sources and areas, fewer host characters and only ecto- or endoparasite taxa, unlike the present study. In addition, as RDA was used, it was possible to assess the relative importance of each host characteristic to each diversity measure and determine why some flatfish species were more parasitized than others by the same macroparasite taxa. The latter, to our knowledge, has only been performed previously with the same objective in fish hosts by Aguirre-Macedo et al. (Reference Aguirre-Macedo, Vidal-Martínez, González-Solís and Caballero2007), on unrelated fish.
Hosts with subtropical distribution and latitudinal ranges around 50º presented ectoparasite assemblages clearly differentiated from those of hosts with polar (P. flesus) or temperate (Scophthalmus spp.) distributions and wider latitude ranges, suggesting an important link between host and macroparasite geographical distribution despite most of the analysed ectoparasite taxa being subarctic or northern temperate (Marques et al. Reference Marques, Santos and Cabral2009). On the other hand, feeding on a higher number of different prey and including Crustacea in the diet appeared to be key factors determining the endoparasite assemblages of D. cuneata, S. lascaris and S. senegalensis, even if Soleidae are not very selective and usually prey on the more abundant items (Link et al. Reference Link, Fogarty, Langton, Gibson and 2005). In fact, the influence of diet preferences in macroparasite infection levels of 7 Soleidae along the Portuguese coast has been examined by Marques et al. (Reference Marques, Santos and Cabral2006) and these authors found that much of the variation of prevalence and mean abundance was related to the type and quantity of food ingested as well as by habitat use. The present study corroborates the results found for those 7 Soleidae, and expands the importance of host diet to other Pleuronectiformes, revealing that even when other ecological, biological and phylogenetic factors are considered, diet is the most important and significant host feature determining the composition of endoparasite assemblages.
Taxonomic distinctness and richness of ectoparasite assemblages appeared to be mostly influenced by the maximum size and reproductive season of the host as well as by the host inhabiting brackish waters. Although the significant effect of the hosts’ maximum reported size in ectoparasite diversity is in agreement with the positive correlations between this host characteristic and ectoparasite richness and diversity in some marine fish (e.g., Timi and Poulin, Reference Timi and Poulin2003) and related with the fact that larger fish have a larger surface for ectoparasite attachment (Luque et al. Reference Luque, Mouillot and Poulin2004), hosts presenting small body lengths, such as D. cuneata and M. azevia, harboured a more diverse ectoparasite assemblage than S. maximus and S. rhombus, which were the largest analysed flatfish species. Despite living in brackish and marine environments, which is suggested to contribute to higher ectoparasite richness and taxonomic distinctness as hosts can acquire parasite species in both environments, P. flesus presented low ectoparasite diversity. Thus, the importance of these 2 host characteristics has to be taken with care since the RDA was conducted on a small number of host species (since only 5 were infected with at least 4 different ectoparasites), therefore restricting the conclusions about the relative importance of this characteristic in the Pleuronectiformes inhabiting the Portuguese coast.
Results also showed that phylogeny must play a role, as Muñoz et al. (Reference Muñoz, Grutter and Cribb2006) and Aguirre-Macedo et al. (Reference Aguirre-Macedo, Vidal-Martínez, González-Solís and Caballero2007) pointed out in their studies. Although the hosts’ phylogenetic variables were not significant, RDAs revealed some importance of the number of host species and genera. Moreover, the most diverse assemblages were found in one of the most speciose families within the Pleuronectiformes, i.e. the Soleidae. However, the host belonging to the family comprising the largest number of species, the Bothidae A. laterna was infected by few taxa. Thus, and similarly to that described for ectoparasites in Mediterranean fishes (Raibaut et al. Reference Raibaut, Combes and Benoit1998), small mammals (Krasnov et al. Reference Krasnov, Shenbrot, Khokhlova and Degen2004) and bats (Bordes et al. Reference Bordes, Morand and Ricardo2008), the sympatric occurrence of closely related hosts on the Portuguese coast (13 species occurring in this area) might have contributed to the higher richness and taxonomic distinction of macroparasite assemblages, allowing the infection of closely related hosts sharing ecological and biological features by host-switch, increasing lateral transfer in ectoparasites and favouring parasite speciations (Verneau et al. Reference Verneau, Du Preez and Badets2009). Therefore, our data corroborate the suggestion that host proximate factors are more important than host species-wide characteristics in ectoparasite assemblage diversity (Patterson et al. 2008) suggesting a similar relationship for endoparasite assemblage diversity.
RDA results found for the diversity of endoparasite assemblages revealed that a more diverse diet (type and number of different prey ingested) contributes to a higher diversity of endoparasite assemblages. This has been referred for other marine fishes’ endoparasite assemblages (Lile, Reference Lile1998; Luque et al. Reference Luque, Mouillot and Poulin2004; Muñoz et al. Reference Muñoz, Grutter and Cribb2006; Aguirre-Macedo et al. Reference Aguirre-Macedo, Vidal-Martínez, González-Solís and Caballero2007; Luque and Poulin, Reference Luque and Poulin2008) as feeding on a more diverse array of taxa exposes hosts to a broader pool of Digenea, Cestoda and Nematoda macroparasites which use those taxa as intermediate hosts. Because Crustacea are the first or second intermediate hosts of these three macroparasite groups (Marcogliese, Reference Marcogliese2004) a diet including them must contribute to a higher richness of endoparasite assemblages. However, and since some parasites can only be acquired by ingesting Polychaeta (e.g. Dichelyne minutus (Rudolphi, 1819) and C. heterochrous, (Koie, Reference Koie2001)), including this group of taxa in the diet was also a major factor in Pleuronectiformes’ endoparasite assemblages diversity.
The wider taxonomic range obtained in the macroparasite assemblages of Soleidae when compared to all the other families may, therefore, be explained by some of their phylogenetic, biological and ecological characteristics, each contributing to the acquisition and/or diversification of a broad range of parasite species. The complexity of host-macroparasite relationships, the importance of phylogenetic and biological characteristics and the unpredictability of host-parasite associations due to host characteristics contributing differently to the observed patterns of diversity, highlight the importance of using multivariate analysis to unveil these relationships. Results also suggest that the observed diversity might be more related to host-parasite co-evolution than first thought, emphasizing the need to focus future research towards the historical processes shaping these relationships. One way to achieve this might be via the reconstruction of host and parasite phylogenies, dating the several evolutionary events leading to the extant relationships and integrating this with host ecological data and parasite life-cycle characteristics.
ACKNOWLEDGEMENTS
The authors would like to thank three anonymous referees for their valuable comments and suggestions to improve the manuscript.
FINANCIAL SUPPORT
This study was partially supported by the Fundação para a Ciência e a Tecnologia (J.F.M., Grant SFRH/BD/8983/2002), (C.M.T., Grant SFRH/BD/19319/2004) and co-funded by the European Union through the FEDER – Fisheries Programme (MARE).
