Sibling species of copepods in association with Mediterranean gorgonians *

Copepods show diversity in habitat and life styles, and this is particularly seen in their diversification as symbionts or parasites of other animals (Gotto, 1979; Raibaut, 1985; Svavarsson, 1990). Those associated with invertebrates show every step, from a free-living, non-transformed creature, via moderately adapted forms, to an unrecognisably transformed, truly parasitic animal. As a result, these copepods display a great morphological variation which can be attributed to the grade of parasitism and to its host specificity as well as to the influence of a geographical isolation. The phenomenon of host specificity is the extent to which a parasite is restricted in the range of hosts that it utilises. Levels of host specificity can range through a continuum from high, with the parasitic species SCI. MAR., 68 (1): 85-96 SCIENTIA MARINA 2004


INTRODUCTION
Copepods show diversity in habitat and life styles, and this is particularly seen in their diversification as symbionts or parasites of other animals (Gotto, 1979;Raibaut, 1985;Svavarsson, 1990).Those associated with invertebrates show every step, from a free-living, non-transformed creature, via moderately adapted forms, to an unrecognisably transformed, truly parasitic animal.As a result, these copepods display a great morphological variation which can be attributed to the grade of parasitism and to its host specificity as well as to the influence of a geographical isolation.
The phenomenon of host specificity is the extent to which a parasite is restricted in the range of hosts that it utilises.Levels of host specificity can range through a continuum from high, with the parasitic species SCI.MAR., 68 (1): 85-96

Sibling species of copepods in association with
Mediterranean gorgonians* occurring on only a single host species, to low, with the parasite occurring on a wide range of phylogenetically unrelated host species (Boxshall, 1998).
The lichomolgoid complex (Copepoda: Poecilostomatoida) includes free living copepods and fairly unspecialised symbiotic copepods.Its recent revision divided this complex into ten families, the majority of them established on the basis of morphological similarity, using exclusively or predominantly a single host category (Humes and Boxshall, 1996).Thus, the large family Rhynchomolgidae Humes and Stock, which comprises 41 genera and more than 230 species, is mainly associated with cnidarians.However, Doridicola Leydig and Critomolgus are the only genera within the family that use a wide range of hosts.The remaining genera live in association with one particular host group within the Cnidaria.
The hypothetical historical (evolutionary) relationships between the species of Doridicola and its three host phyla: cnidarians (Hydrozoa and Anthozoa), molluscs (Bivalvia, Gastropoda and Cephalopoda) and echinoderms (Asteroida and Ophiuroidea) have recently been studied (Ho and Kim, 2001).Although the cladistic analysis was based strictly on morphological information, the cladogram obtained revealed certain degrees of relationship between the symbionts and their hosts.Thus, the study indicates that the ancestral stock of the modern species of Doridicola diverged into two lineages early in its evolution, with one exploring life on octocorallians (alcyonaceans and gorgonaceans) and the other mainly on alcyonaceans and gastropod mollusks.Following Ho and Kim (2001), the only European Doridicola species associated with a gorgonian, Doridicola botulosus (Stock and Kleeton, 1963), would be a descendant of the second lineage.Doridicola botulosus was described living on the Mediterranean gorgonian, Eunicella singularis (Esper, 1791), and has not been reported since then.In the present study, in addition to reporting new records of D. botulosus from the Mediterranean and describing a new European species of Doridicola, a hypothesis dealing with the origin of these two species associated with two different Mediterranean gorgonians is proposed and discussed.

MATERIAL AND METHODS
Fragments from large colonies of the gorgonian Paramuricea clavata were collected from three different infralittoral zones on the Iberian Mediterranean coast: Gibraltar Harbour (Algeciras Bay, Southern Iberian Peninsula); Hormigas Islands (Cabo de Palos, southeastern Iberian Peninsula) and Medes Islands (L'Estartit, northeastern Iberian Peninsula).Fragments from Eunicella singularis were also collected in two of these areas (Fig. 1).
The samples were immediately fixed by adding progressively Formalin (4% in sea water) to make a concentration of approximately 2%.Later, the fixative-sea water was passed through a 100 µm net.The copepods were then recovered from the sediment retained and later preserved in 70% ethanol.For microscopical study, whole specimens were stained with Chlorazole black E (Sigma® C-1144) and dissected under a stereomicroscope.Permanent mounts were made in Lactophenol and sealed with Entellan (Merck® 1.07961.0100).All figures were drawn 86 M. CONRADI et al. with the aid of a camera lucida on a Leica DMLB differential interference microscope.
For scanning electron microscopy (SEM) studies, selected specimens were post-fixed in 2.5% glutaraldehyde in 0.2M cacodylate buffer at pH 7.3 and in 1% OsO 4 in the same buffer and subsequently critical point dried, mounted on stubs, coated with goldpalladium and observed with a PHILIPS XL30.
Remarks: In the labels of this type material the indication of the host [Eunicella verrucosa (Pall.)]differs from that given by the authors in the publication [Eunicella stricta (Bert.)].Currently, E. stricta is a well known junior synonym of Eunicella singularis (Esper, 1791) (see Carpine andGrasshoff, 1975, Weinberg, 1975 for additional nomenclatural details).The host identity was already commented in the original description (Stock and Kleeton, 1963: 247, footnote), although this mistake was not resolved as a new label to be included with the type specimens.
Remarks: According to Carpine and Grasshoff (1975), all Mediterranean references to P. chamaeleon should be considered as Paramuricea clavata (Risso, 1826).The revision of this material revealed that it must be placed in the new species of Doridicola described in this paper as D. comai sp.nov.(see below).
Remarks: This species was described in detail as Lichomolgus botulosus by Stock and Kleeton (1963).Ten years later, it was assigned by Humes and Stock (1973) to the new genus described by these authors (1972), Metaxymolgus, which was later synonymised with the genus Doridicola (Humes and Stock, 1983).
The slender antenna, with two subequal terminal claws nearly as long as the fourth segment, the long spines on lash of the second maxilla, forming a crest (Fig. 2c) and the large (outer) seta on the middle segment of the female maxilliped (nearly 3 times longer than the segment and 7 times longer than the medial seta) (Fig. 2d) are some diagnostic characteristics of D. botulosus.
The specimens of D. botulosus studied in this work differ slightly from those drawn by Stock and Kleeton in the armature of the antenna and the maxillule as well as in the structure of the mandible.Those specimens found in Banyuls have a maxillule armed with 3 setae whilst the newly collected specimens have 4 unequal setae.However, as this seta is small and difficult to see, it is likely that Stock and Kleeton overlooked it.With respect to the second antenna, our specimens have a slender claw-like seta on the third segment (Fig. 2f) which is not clearly shown in the original drawings.There is also some variation in the mandible since the newly collected specimens show a more complex structure on the inner margin of this appendage (Fig. 2a).As D. botulosus is a well described species, we have only drawn those parts which are clearly different from the new Doridicola and those parts which have some kind of variation with respect to the original description.Doridicola comai, sp.nov.

Etymology:
The specific name comai is after our friend and colleague Rafel Coma, who enthusiastically encouraged us to sample the gorgonian Paramuricea clavata in the Strait of Gibraltar.
Among these species, only D. aculeatus, D. botulosus, D. hetaericus, D. sensilis and D. singularipes have the two terminal claws of the antenna subequal in size like D. comai.Nevertheless, four of these five species can be easily separated from D. comai by the combination of the following three features: (1) maxilla with a tiny seta III, a barbuled seta I nearly as long as the setulose seta II; (2) maxilliped with an outer seta barbuled on its outer surface and setulose on both sides of its distal part and nearly or more than three times longer than the medial seta; (3) leg 5 with basal swelling, bearing spinules along its outer surface and carrying two unequal terminal setae.These distinct characteristics set the new species apart from all known congeners but morphologically close to other European species: D. botulosus, which lives on a Mediterranean gorgonian species, Eunicella singularis, that usually inhabits similar habitats to Paramuricea clavata.However, both species can be clearly distinguished by the  5c), the length of the two unequal setae on the middle segment of the maxilliped (Figs.2d, 5d) and the distomedial seta of the third segment of the antenna which is transformed into a claw in D. comai (Fig. 3e) and into a claw-like seta in D. botu-losus (Fig. 2f).As mentioned above, the specimen reported by Stock and Kleeton (1963) associated with Paramurica clavata in Cap Creus (ZMA Co. 100.747) must be assigned to Doridicola comai.

DISCUSSION
It is widely accepted that the many complex interactions in host-parasite co-evolution do not allow taxa to be linked in monophyletic clades solely on the basis of their shared host affiliation (Huys, 2001).However, certain aspects of copepod biology, such as host switching, especially in copepod symbionts where it is a common phenomenon, might throw some light on the group's past history (Gotto, 1998).Some examples are the cyclopoid family Notodelphyidae, commonly associated with ascidians but with some species inhabiting octocorals (Stock and Humes, 1970) resembling their ascidiandwelling relatives, and the family Mantridae, whose species are known to occur in bivalves, and present strong evidence of having arisen from ascidicolous ancestors (Huys, 1990;Ho, 1994).
Although host-switching can sometimes be explained in simple opportunistic terms, the selective advantages of switching hosts are obvious since new hosts mean new horizons of evolutionary opportunity and also provide more chance to survive (Gotto, 1998).Nevertheless, the mechanics of the process has not yet been determined.Ho (1994) believes that competition can mould host specificity in copepods, so that the ancestral cyclopoid in sponges was lost in competition for food and shelter with the siphonostomes, and in bivalves with the poecilostomatoids.Another example of competition, in this case at species level, occurs between the caligids Lepeophtheirus thompsoni and L. europaensis, two parasitic copepods naturally isolated on their hosts, turbot (Psetta maxima L.) and brill (Scophthalmus rhombus L.) respectively.They are able to meet, mate and hybridise on turbot experimentally but they are prevented from doing so by a strong host preference when given a choice (Dawson et al., 2000).This pattern is a result of a greater sensitivity to competition for the generalist species L. europaensis than for the specialist L. thompsoni.
Later on, Gotto (1998) postulated an altered receptor as the mechanism which triggers a new association.According to this hypothesis, the biochemical recognition between copepod and host might be of prime importance, so if the copepod's sensory receptor undergoes some sort of change and does not respond to the cue of the usual host, a switch to a new host would probably occur.However, this receptor would presumably be capable of picking up a different signal on which the searching larva might hopefully home.On arrival at the source, and should conditions prove suitable for invasion and residence, a new alliance may then be born.There is no evidence to sustain this hypothesis, but there are some genera of symbiotic copepods, which live on different hosts, or even different parts of the same hosts, that are rich in sympatric species (Lonning and Vader, 1984).If a population several different habitats, an adaptive polymorphism can be maintained (Levene, 1953).A reduced migration between these different habitats can greatly enhance the stability of this polymorphism (Moran, 1959, Eyland, 1971) and may even lead the population into a process of genetic isolation (Balkau and Feldman, 1973).Lepeophtheirus europaensis can also be taken as an example of adaptive polymorphism since this ectoparasite can live on two different hosts in the Gulf of Lions (Mediterranean, France): the brill (Scophthalmus rhombus), a marine scophthalmid, and the flounder (Platichthys flesus), a pleuronectid inhabiting lagoons (brackish water) (Zeddam et al., 1988).L. europaensis from brills may colonise flounders during their reproductive migrations, but these parasites will encounter unfavourable conditions when the flounders return to the lagoons since they are unable to live at salinities encountered in lagoons (De Meeüs et al., 1990).These differences in adaptation lead to a very low rate of exchange between the two populations, a phenomenon encouraging conditions for protected adaptive polymorphism.Thus, the taxon L. europaensis can represent the beginning of a sympatric speciation currently occurring in the Gulf of Lions (De Meeüs et al., 1992).Another example of speciation are the twin species Astericola clausi and A. asterinae, lichomolgid symbionts of the European asteroids Marthasterias glacialis (O.F.Müller) and Asterina gibbosa (Pennant), which are likely to be derived from a common ancestor (Bocquet et al., 1970;Conradi et al., 1993).
Allopatric speciation through postzygotic genetic incompatibilities has been the dominant view of speciation for the past six decades (Turelli et al., 2001).In this scenario, sympatric speciation has been extremely controversial.Sympatric speciation was defended by a stalwart group of empiricists but thought by many to be implausible and of limited applicability.Some authors understand sympatric speciation as a short-scale allopatric speciation (e.g.switching host, different microhabitats inside the same host, displacement of the reproductive period).Recently, Via (2001) postulates that sympatric speciation can occur under a variety of conditions (see also Turelli et al., 2001).However, the underlying idea is that it might be time for a re-evaluation of the geographical classification of speciation modes in favour of one based primarily on evolutionary mechanisms.
The two Doridicola species associated with Mediterranean gorgonians, D. botulosus and D. comai, are so morphologically similar that they may be regarded as sister species and would represent a good example of speciation producing sibling species by the colonisation of two related hosts.The geographical distribution of E. singularis resembles that of Paramuricea clavata, and they can sometimes be found living together.However, from the ecological point of view, E. singularis is a photophilic gorgonian with a depth range up to 50 meters (Weinberg, 1975).On the other hand, P. clavata is a sciaphilic gorgonian with a deeper distribution, usually being abundant at 50-60 metres in the Mediterranean (Carpine and Grasshoff, 1975).There are some places of the Mediterranean coast where the deeper range of the Eunicella overlaps with the superficial range of Paramuricea, and therefore both species cohabit.
According to cladogenesis (see De Queiroz, 1998), there are two possibilities to explain the origin of these two Doridicola species: 1) the budding model: one of the Doridicola species is the origin of the other, and both species (original and new) coexist in time, naturally isolated on their respective hosts; and 2) the bifurcating model: an ancestral Doridicola species living in the overlapping deep range of these two Mediterranean gorgonians gave rise to two subpopulations living on these related hosts.The original species did not persist in time, but the two subpopulations evolved as different species including their own apomorphies.After Carton's studies, the first possibility (budding model) could explain the case of Astericola clausi and Astericola asterinae (see Bocquet et al., 1970;Carton, 1964).However, both Doridicola species, D. botulosus and D. comai, show apomorphic stages in different appendages (D. botulosus has a more derived set of oral appendages, while D. comai has a more transformed second antenna including a claw on the third segment instead of the claw-like seta present in D. botulosus).Thus, the second possibility (the bifurcating model) could be the most plausible in this case.
part, only the material associated with Paramuricea chamaeleon).