PSEUDOCYCLOPOIDEA, Giesbrecht, 1893

Bradford-Grieve, Janet M., Boxshall, Geoffrey A. & Blanco-Bercial, Leocadio, 2014, Revision of basal calanoid copepod families, with a description of a new species and genus of Pseudocyclopidae, Zoological Journal of the Linnean Society (Zool. J. Linn. Soc.) 171 (3), pp. 507-533 : 529-531

POSITION OF PSEUDOCYCLOPOIDEA IN CALANOIDA

Genetic data and a revised phylogeny confirm the basal position of a newly defined superfamily Pseudocyclopoidea (which now includes families previously assigned to the synonymized Epacteriscoidea) within the Calanoida , as well as improving the organization and resolution of the relationships among superfamilies.

The revised molecular phylogeny ( Fig. 12 ) reinforces the results obtained in the previous genebased study ( Blanco-Bercial et al., 2011), although the lack of members of the basal families, the epibenthic Ridgewayiidae , Boholinidae , Pseudocyclopidae , and Epacteriscidae in the previous analysis was problematic. With the inclusion of sequences from the superfamily Pseudocyclopoidea , the reconstructed phylogeny agrees partially with the topology of early morphologybased phylogenies ( Andronov, 1974; Park, 1986). Here superfamilies Pseudocyclopoidea , Augaptiloidea , and Centropagoidea sequentially split off from a main stem. The remaining superfamilies form a single clade, the topology of which is similar to the topology described in Bradford-Grieve et al. (2010) based on morphological data.

The lower support than that found by Blanco-Bercial et al. (2011), evident in many parts of the reconstruct- ed phylogeny, could result from incomplete gene coverage of the new taxa added (Appendix S1). It is likely that the addition of COI and cyt b to the clades where they are missing would add stronger support to the analyses. For example, the inclusion of mitochondrial genes can improve resolution at deeper nodes ( Fisher-Reid & Wiens, 2011; Cornils & Blanco-Bercial, 2013). We found that without the addition of mitochondrial gene sequences from P. ambiguus sp. nov., not only was the phylogeny not recovered, as it is here, but the resulting superfamilies were also polyphyletic or paraphyletic in some cases (data not shown).

All superfamilies were recovered as monophyletic, even in cases where very divergent groups were included, supporting the conclusions drawn from morphological characters ( Ho, 1990; Huys & Boxshall, 1991). This fact is very significant in Centropagoidea, where the two divergent families Diaptomidae (the only entirely freshwater family) and Pseudodiaptomidae clustered with the other Centropagoidea in a single clade.

The resolution of monophyletic clades, representing the superfamilies Eucalanoidea, Megacalanoidea, and Bathypontioidea, sister to the Clausocalanoidea and Spinocalanoidea, does not agree with morphological studies ( Andronov, 1974; Park, 1986; Bradford-Grieve et al., 2010). Although interesting, this result should be considered with caution, because the family identified as intermediate between these two clades ( Ryocalanidae ; superfamily Ryocalanoidea ) is missing from the molecular analysis, and its addition might result in changes to the present topology of the phylogeny. Thus, the presented revised molecular phylogeny provides testable hypotheses for future work.

PHYLOGENY OF PSEUDOCYCLOPOIDEA

Among the pseudocylopoidean genera, the morphologybased phylogenetic signal was possibly obscured by the high degree of homoplasy and may have interfered with accurate tree inference. Some of the homoplasy in our data may be the result of the possibility that we are not always dealing with homologous characters or that the character states are not accurately recorded in the literature. It is also possible that some characters reflect lifestyle rather than preserving a phylogenetic signal. In this analysis, it was not possible to determine homologies in some characters relating to the mandible upon which feeding niche strongly impacts ( Itoh, 1970). We note, however, that certain types of modification of the mandible, maxilla, and maxilliped are strongly linked to the two major clades ( Pseudocyclopidae and Epacteriscidae ), and so may contain a robust phylogenetic signal.

In the Pseudocyclopidae , the mandible generally has small teeth (although Exumellina and Stargatia have two elongate ventral teeth) and the endopod is well developed, often with four setae on segment 1 and more than nine setae on segment 2. The maxilla has a normally developed basis with an elongate endite and normal endopod setae. The maxilliped usually has an elongate endopod that is furnished with normal setae. We note that this family contains genera that live in open water habitats and the only freshwater groundwater genus, as well as marine cave-dwellers. It is deduced from the form of the mouthparts that many of these genera are fine-particle feeders. Those that deviate in having reduced mandibular endopod setation appear to have other modes of feeding. For example, Exumella seems to be a benthic scavenger ( Jaume & Boxshall, 1995), and Exumellina and Stargatia , which also have paddle-like modification to the mandible and maxillule endopods, may be raptorial feeders feeding on delicate prey ( Fosshagen & Iliffe, 1998). A large subset of these genera also have female leg 5 modified so that the mode of articulation of exopod segment 3 directs this segment into the midline, or ensures that it has an even greater arc of movement ( Badijella , Pseudocyclops , Ridgewayia , Robpalmeria , Normancavia , Brattstromia , Exumella, Exuminella , Pinkertonius gen. nov., and possibly Hondurella , Placocalanus , and Stargatia ). It is tempting to hypothesize that these genera are adapted to digging in sediment.

In the Epacteriscidae , the mandible generally has the ventral tooth enlarged and separated by a much larger gap than exists between the remaining teeth; an even more specialized gnathobase is found in Epacteriscus , where the cutting blade has a prominent extension bearing sharp teeth that extends well out from the body ( Fosshagen, 1973). Most species have a reduced endopod that is either one-segmented with only one seta or is absent (except for Balinella , Bofuriella , Erebonectes , and Miheptneria , which have two-segmented endopods). The endopods of the maxillae and maxillipeds are very condensed and nearly always have spine-like setae, with a few exceptions ( Miheptneria , Bomburiella , and Edaxiella ), and the maxilla basis is enlarged and has a very low-profile endite. Taken together, these character states indicate that most genera are carnivores.

The basal taxa in each family ( Pinkertonius gen. nov. and Miheptneria ) appear to be adapted to fineparticle feeding. Their mandibles have short teeth, and the endopods of the maxilla and maxilliped bear normal setae. The remaining genera in each clade have rather uniform modifications to the mandible, maxilla, and maxilliped. A similar situation exists in the family Heterorhabdidae , within which distinct transformations occur from small-particle feeders such as Disseta , to highly specialized carnivores, such as Heterorhabdus and Neorhabdus . This evolutionary shift in feeding type is accompanied by a strong reduction or loss of dorsal teeth on the mandible and an increase in size of the ventralmost tooth ( Nishida & Ohtsuka, 1996). The phylogenetic analysis of the Heterorhabdidae by Ohtsuka, Soh & Nishida (1997) recovered the small particle-feeding genera as basal branches and the specialized carnivorous genera as terminal branches. We infer that carnivory has evolved independently within all of these families.

Our phylogenetic analysis and conclusions concerning the taxonomic hierarchy at the base of the calanoid phylogeny are similar to the suggestions of Andronov (2007). Nevertheless, we have raised the rank of his family and subfamily names. Our analysis places Miheptneria in the same taxon as that of Andronov (2007), but separates out Azygonectes , the affinities of which he was not sure about, along with Erebonectoides and Caiconectes . We consider these genera currently to be incertae sedis (of uncertain placement). There is reasonably strong evidence for a differential diagnosis of a new family based on Caiconectes . It has a uniquely primitive setation pattern on the endopod of leg 1. It is the only calanoid with seven setae on the third endopodal segment. It also is the only calanoid we are aware of with five setae on the basis of the maxilla (most others have a maximum of four setae). In addition to these unique plesiomorphies, there is a cluster of other shared maximum plesiomorphic states relating to the distribution of aesthetascs on the antennules in both sexes, especially on segments XIX and XX (Appendix S5a, b). In addition, there are some apomorphies (e.g. reduced setation on the mandibular endopod and maxillule, and the coarse outgrowths of the long feeding setae on the maxilliped). At least on the basis of morphology, we would expect Caiconectes to be robustly recovered as the basal offshoot of the Calanoida in future analyses. On the other hand, Azygonectes and Erebonectoides is a potentially unstable group. These two genera share many key characters with other Epacteriscidae : presence of an aesthetasc on segment IV in female; the form of the mandible palp, with a reduced endopod; the asymmetry of caudal seta VI in the female, and of the caudal rami in the male; and endopod segment 1, of at least leg 3, with its distal outer corner bifid or trifid. But these character states alone are obviously not enough (on balance) to cluster them with the Epacteriscidae in this tree.

The Pseudocyclopoidea are characterized by numerous homoplasious character states that ensure that the phylogenetic signal in the data is weak; therefore, none of the higher level taxa definitions is based on unique synapomorphies. The wide range of combinations of character states among the taxa in this superfamily hints at these taxa being a sparse sampling of a once much more diverse, ancient taxon, from which the ancestors of the Augaptiloidea , Centropagoidea, and a clade containing the remaining superfamilies evolved ( Fig. 12 ). Part of their evolutionary capacity may have includ- ed the possession of character states that were easily reversed or independently reacquired, hence the currently observed diversity of character state combinations. Although the addition of new taxa or the revision of some character states may change relationships at the base of the tree, in our judgement, the pseudocyclopid and epacteriscid clades are likely to remain intact.

The old concepts of the monogeneric Pseudocyclopidae and Boholinidae were based, at least in part, on the obvious external separation of the paired gonopores of the adult female (e.g. Huys & Boxshall, 1991: fig. 2.2.16). A similar arrangement was noted in at least some members of the Arietellidae , but the variation in structure of the female genital system within the family prompted Ohtsuka, Boxshall & Roe (1994) to recognize five major trends involving fusion of copulatory pores to form a single common pore, various migra- tions of gonopores and copulatory pores, and the asymmetrical enlargement of copulatory pores. Despite this variability in structure, Bradford-Grieve et al. (2010) were only able to use three characters based on female genital systems in their phylogenetic analysis of the Calanoida , and two of those were based on seminal receptacles and ducts. Variability in female genital structures can be found even in the more derived calanoid taxa, such as the clausocalanoidean family Stephidae . Unlike all of its congeners, the adult female of Stephos vivesi Jaume, Boxshall, Gràcia, 2008 possesses a pair of separate gonopores. Jaume et al. (2008) interpret- ed this condition as secondary, possibly derived by the loss of the genital operculum concealing the paired gonopores and their subsequent migration and separation. This serves to highlight the scale of intrafamilial variability in certain calanoid families, and the variability within the revised and enlarged concept of the family Pseudocyclopidae should be interpreted from this perspective.

It is interesting to note that the two taxa that are basal to the Epacteriscidae and Pseudocyclopidae live in the open ocean at depth, and that the terminal taxa are cave-dwelling in the Epacteriscidae and cavedwelling, shallow-water, or even groundwater-dwelling in the Pseudocyclopidae . Thus, it appears that any hypothesis about the sequence of events surrounding the colonization of anchialine cave environments may be exactly the opposite from that proposed by Boxshall & Jaume (2000) for the Misophrioida . The misophrioid family Speleophriidae currently comprises eight genera and 19 species, almost all of which occur only in coastal anchialine habitats. The exception is Archimisophria Boxshall, 1983 , which contains two species, both found in the deep hyperbenthic community of the tropical Atlantic. This genus is not basal within the family, and is recovered as the sister taxon of the cave-dwelling genus Expansophria Boxshall & Iliffe, 1987 ( Boxshall & Jaume, 2000). In the case of the speleophriids, Boxshall & Jaume (2000) inferred that the presence of species in the deep sea was secondary, and that they were probably descended from shallow-water ancestors.

Figures

Figure 12. Phylogram of taxa of the copepod order Calanoida and out-groups. The topology and branch length correspond with the RAxML maximum-likelihood tree. Numbers at nodes indicate percentage of bootstrap recovery (when> 50%)/Bayesian posterior probability (when> 0.95, except for the Augaptiloidea/Centropagoidea node). For the superfamily groupings, see Appendix S1.