CYLINDROPSYLLIDAE SARS, 1909

Huys, Rony & Conroy-Dalton, Sophie, 2006, Revision of the genus Evansula T. Scott, 1906 (Copepoda, Harpacticoida, Cylindropsyllidae) with a description of three new species, Zoological Journal of the Linnean Society 147 (4), pp. 419-472 : 420-421

FAMILY CYLINDROPSYLLIDAE SARS, 1909

Various authors ( Huys, 1992; Huys & Conroy-Dalton, 1993; Martínez Arbizu & Moura, 1994) have remarked on the polyphyletic status of the family Cylindropsyllidae sensu Lang (1948) and the artificiality of its subfamilial division. These problems were partly resolved by upgrading the Leptastacinae to family level ( Huys, 1992). In a controversial paper on the phylogenetic position of the Cylindropsyllinae and the systematic status of the Leptopontiinae, Martínez Arbizu & Moura (1994) elevated the latter to the family level and subsumed the Cylindropsyllinae within the Canthocamptidae on the basis of the similarity in the male P3 endopod and the female genital field – two characters they regarded as synapomorphies for the Canthocamptidae . If this is to be adopted as a measure leading to a more natural phylogenetic system, their subsequent failure to find any characters supporting a sister-group relationship between both taxa is, in view of the monophyletic status of the Cylindropsyllinae , a real contradiction in terms. The authors arrived at their conclusions by employing the outmoded character sets selected by Lang (1948) to define the Cylindropsyllinae and the Canthocamptidae . No new phylogenetically informative characters were identified, nor was a detailed analysis of the genera in each family group taxon undertaken. Martínez Arbizu & Moura (1994) used the following two characters to support the incorporation of the Cylindropsyllinae into the Canthocamptidae :

1. Homology of apophysis on the three-segmented P3 endopod in the male. Although the derivation of this apophysis in the Cylindropsyllidae and at least the freshwater component of the Canthocamptidae is now fully understood and proven homologous ( Martínez Arbizu & Moura, 1994; R. Huys & S. Conroy- Dalton, unpubl. data), it needs to be stressed that this modification is not a synapomorphy for the Canthocamptidae , as it is found in a wider group of families, including the Louriniidae , Huntemanniidae , Rhizothrichidae , and in all probability also the Cletodidae . Therefore, it cannot be used as evidence for the allocation of the Cylindropsyllinae to the Canthocamptidae .

2. Similarity in structure of the female genital field with a posteriorly displaced copulatory pore. The authors claim that this posterior position is displayed by all genera of the Canthocamptidae , and that they also observed this character in a number of cylindropsyllid genera such as Evansula , Cylindropsyllus Brady and Stenocaris Sars. In reality, the copulatory pore is not posteriorly displaced in the primitive genus Evansula (see below) or the more derived Stenocaris (R. Huys, pers. observ.), and does not represent the ancestral state for the family, but has evolved secondarily only in the Cylindropsyllus – Cylinula Coull lineage. Therefore, the similarity found in both families is a result of convergence.

Moura & Pottek (1998) regarded the pore patterns on the female genital somite (in particular the pore clusters or ‘sieves’) as an additional synapomorphy supporting a relationship between the Canthocamptinae and Cylindropsyllinae , but this evidence is not convincing. Pore triplets flanking the copulatory pore have been recorded in other families of the canthocamptoid complex, such as the Ancorabolidae (e.g. Conroy-Dalton, 2001), Cletodidae (e.g. Gee, 1994, 1999; Gee & Huys, 1996) and Huntemanniidae ( Dahms & Pottek, 1992) . In addition to their evolutionary labile nature within families, the phylogenetic significance of genital pore clusters is further limited by their presence in taxa that are not related to the Canthocamptidae and its allies, such as the Miraciidae ( Huys & Böttger-Schnack, 1994) and Clytemnestridae ( Huys & Conroy-Dalton, 2000) .

There is as yet no evidence questioning the monophyly of the Cylindropsyllidae . The unique sexual dimorphism on the male P2 exopod, the structure of the P4, and the composite seta V of the caudal ramus are distinctive synapomorphies for the family. Conversely, it is at present not clear at all what characters define the Canthocamptidae . Martínez Arbizu & Moura’s (1994) proposal to amalgamate the Cylindropsyllinae within the Canthocamptidae is also an unwise one for pragmatic reasons. Currently, three subfamilies exclusively containing freshwater genera are recognized in the Canthocamptidae : Canthocamptinae, Morariinae and Epactophaninae. Pesta (1932) lumped all the marine and brackish-water canthocamptids in the catch-all taxon ‘Halocanthocamptinae’, a family group name that, for nomenclatural reasons, is unavailable as it was not based on the stem of an available genus (in fact the genus name Halocanthocamptus has never been proposed). Lang (1948) argued strongly against the recognition of this subfamily and finally abandoned it. The marine genera, however, have never been placed satisfactorily and are still floating in the system of the Canthocamptidae . This state of affairs has inspired other people to include even more marine genera in this family and, in one instance, an entire subfamily was transferred ( Por, 1986). The Canthocamptidae , once a morphologically homogeneous, primarily freshwater-inhabiting family, currently contains 55 genera, 18 of which are found in fully marine or estuarine habitats ( Boxshall & Halsey, 2004; Karanovic, 2004; Karaytu ğ & Huys, 2004). Paradoxically, subsuming the Cylindropsyllidae within this family would increase this number to 27, half of the total number of genera.

Transferring a monophyletic group to a larger polyphyletic group without considering the possible implications this may have is generally regarded as bad practice. The systematic arrangement of the Canthocamptidae has not reached any stability and cannot be expected to do so within the foreseeable future. The Cylindropsyllinae , on the other hand, is a sharply delimited group with definite characteristics that deserves separate family rank. It is not even at all established that the Canthocamptidae , or any subgroup currently contained in it, is the sister taxon of the Cylindropsyllidae , as the characters that would support such a relationship are found in other taxa as well. It is the misapplication of the shared presence of very few characters that has caused the Canthocamptidae to become a repository for anything that could not fit the diagnoses of the other families contained in the second volume of Lang’s (1948) monograph. A lack of decisive facts permits every taxonomist to have their own opinion about the rank of groups. However, any objective observer cannot fail to notice the unfortunate consequences of amalgamating various cletodid genera with doubtful affiliations in a loosely defined group, assigning this group the rank of subfamily and accommodating it in the Canthocamptidae ( Por, 1986) . The concept of the ‘Halocanthocamptinae’ introduced by Pesta (1932) deserves to be abolished, as Lang (1948) suggested. Yet from the recent addition of the marine Hemimesochrinae ( Por, 1964), Cylindropsyllinae ( Martínez Arbizu & Moura, 1994) , and a group of genera formally designated ‘ Canthocamptidae incertae sedis ’ ( Por, 1986), it seems that this unsatisfactory taxonomic practice is being kept alive by certain taxonomists. Pending a full-blown phylogenetic analysis of the canthocamptoid complex and the arrival of molecular sequence data, we retain the Cylindropsyllidae as a valid family comprising the following genera: Cylindropsyllus Brady, 1880 ; Evansula T. Scott, 1906b ; Stenocaris Sars, 1909 ; Cylinula Coull, 1971 ; Boreopontia Willems, 1981 ; Stenocaropsis Apostolov, 1982 ; Willemsia Huys & Conroy-Dalton, 1993 ; Navalonia Huys & Conroy-Dalton, 1993 ; and Selenopsyllus Moura & Pottek, 1998 . Of these, the genus Cylindropsyllus is possibly paraphyletic and Stenocaris undoubtedly polyphyletic (R. Huys & S. Conroy- Dalton, unpubl. data).