Paralaophontodes was assigned to Laophontodinae by Lang (1965), which has never been questioned in the past decades. Nevertheless, the systematic status of Laophontodinae remains unclear, as that taxon is mainly characterized by plesiomorphic character states when compared with its putative sister group Ancorabolinae Sars, 1911 (e.g. Conroy-Dalton 2004, George 2006, Gheerardyn and George 2010, Gheerardyn and Lee 2012). Yet, Gheerardyn and George (2010) detected three derived features, namely (i) a spinulose outer “bump” on the second antennular segment, (ii) the lengthways elongation of the P1 coxa, and (iii) the transformation of the P1 exp-2 outer element from a bipinnate spine into a bare, geniculate seta (cf. George and Müller 2013). These features may constitute autapomorphies of Laophontodinae, condition that still has to be verified in detail for all laophontodin representatives. However, the taxon Paralaophontodes fits all named derived features. Thus, its allocation to Laophontodinae persists undoubted; future studies may elucidate the systematic status of and within Laophontodinae.

When erecting Paralaophontodes, Lang (1965) already noted a strong similarity of the then assigned species P. echinatus and P. robustus with Laophontodes armatus and L. hedgpethi, based mainly on the “armature of the body” (Lang 1965, p. 538). He chose not to place all four species in one genus because L. armatus and L. hedgpethi retain a 3-segmented P1 exp, a long apical seta in addition to the apical claw on the P1 enp-2 and endopods on the swimming legs P2–P4 (but see remarks on the re-description of L. armatus above), whilst P. echinatus and P. robustus present a 2-segmented P1 exp, no apical long seta on P1 enp-2, and at least the swimming legs P2 and P4 lack an endopod. Mielke (1981) first recognized that all these characters were synapomorphic for the then known Paralaophontodes species (P. echinatus, P. exopoditus, and P. robustus). Five years later, the status of Paralaophontodes was again addressed, independently by Baldari and Cottarelli (1986) and Fiers (1986). Baldari and Cottarelli (1986) described P. elegans from a Philippine island and provided a then updated generic diagnosis (but excluding Fiers’ (1986) work). Fiers (1986) provided an excellent re-description of P. echinatus (as P. echinata) and made a brief comparison of Paralaophontodes with other laophontodin species, stating that within the genus Laophontodes an “armatus-group” enclosing Laophontodes armatus, L. hedgpethi and L. psammophilus might be the sister-group of Paralaophontodes. Although this pointed to paraphyletic states for both genera, Fiers (1986) did not present any further phylogenetic argument, so a sound hypothesis could not be made. Later, Fiers (1988) re-stated his assumption of a sister-group relationship between Laophontodes [part.] and Paralaophontodes, but he again did not provide enough detail. It was George (1993) who first suggested a monophylum Paralaophontodes to include not only the then valid species P. echinatus, P. elegans, P. exopoditus, and P. robustus [George (1993) overlooked that Wells and Rao (1987) had synonymized P. robustus with P. echinatus] but also Laophontodes armatus, L. hedgpethi and L. psammophilus Soyer, 1974. The species described here is the first new species since the work by Baldari and Cottarelli (1986) and Fiers (1986).

Based on the summary given above, it is concluded that a monophylum Paralaophontodes is, according to Mielke (1981), phylogenetically justified by the following autapomorphies [plesiomorphic states in square brackets]:

A. P1 exp 2-segmented [P1 exp 3-segmented];

B. P2 lacking endopod [P2 with at least 1-segmented endopd];

C. P4 lacking endopod [P4 with at least 1-segmented endopod].

All known Paralaophontodes species share these apomorphies, whilst they are missing from the Laophontodes species treated here (Fiers’ [1986] “armatus-group”; but see the remarks in the description of L. armatus). Nevertheless, apomorphies A–C are to some extent weak, being widely scattered not only in Laophontodinae but also in most Ancorabolinae, its supposed sister-group, and other harpacticoid taxa. More precisely, a 2-segmented P1 exopod (apomorphy A) resembling that of Paralaophontodes in shape and ornamentation is also present in Ancorabolina George, 2006, and in all genera of the Ancorabolus-lineage sensu Conroy-Dalton and Huys, 2000. In fact, most other members of Ancorabolinae also show this kind of P1 exopod: Arthuricornua Conroy-Dalton, 2001, Ceratonotus Sars, 1909, Dendropsyllus Conroy-Dalton, 2003, Dorsiceratus Drzycimski, 1967 [part.], Polyascophorus George, 1998, Pseudechinopsyllus George, 2006, and Touphapleura Conroy-Dalton, 2001. Yet, each of these taxa present numerous autapomorphies (cf. George 1998, Conroy-Dalton and Huys 2000, Conroy-Dalton 2001, 2003a, George 2006a, b, c, Gheerardyn and George 2010, Gheerardyn and Lee 2012, George and Müller 2013, George and Gheerardyn 2015) absent from Paralaophontodes, so it is likely they are not closely related. It must be supposed that the reduction (in a phylogenetic, not an ontogenetic, sense) of a 3-segmented to a 2-segmented P1 exopod has convergently occurred more than once. This is also noted for apomorphy B: apart from many laophontodin taxa (Algensiella Cottarelli and Baldari, 1987, Laophontodes multispinatus Kornev and Chertoprud, 2008, Lobopleura ambiducti Conroy-Dalton, 2004, Probosciphontodes Fiers, 1988, Tapholaophontodes Soyer, 1974) (Pallares 1968b, Soyer 1974, Cottarelli and Baldari 1987, Fiers 1988, Conroy-Dalton 2004, Kornev and Chertoprud 2008), the P2 endopod is also absent in several ancorabolin taxa (Arthuricornua anendopodia Conroy-Dalton, 2001, Ceratonotus steiningeri George, 2006, Dendropsyllus, Echinopsyllus Sars, 1909, Polyascophorus martinezi (George, 1998), P. monoceratus George, Wandeness and Santos, 2013, Pseudechinopsyllus) (Conroy-Dalton 2001, 2003a, b, George 1998, 2006a, b, George et al. 2013). Even apomorphy C is also present in other laophontodin species (e.g. Patagoniaella Pallares, 1968, Probosciphontodes, Tapholaophontodes). Thus, none of the autapomorphies so far assigned to Paralaophontodes are unambiguous, particularly because a reduction of segments or elements is quite common in Copepoda (Huys and Boxshall 1991), explaining its heterogeneous distribution across the copepod taxa.

To evaluate phylogenetic relationships, synapomorphies, i.e. unique derived characters of the treated taxa must be recognized (Hennig 1982, Ax 1984, Sudhaus and Rehfeld 1992, Wägele 2001). In the case of Paralaophontodes, it seems somewhat peculiar that for many years the rather crude apomorphies A–C above have been used to justify a monophylum Paralaophontodes whilst complex and unique features such as the two transverse triangular elongations on cphth, the tuft of “hairy” setules dorsally on cphth, and the characteristic dorsal ornamentation of the body somites have been neglected. Even authors who noted the strong similarity between Paralaophontodes and corresponding Laophontodes species (e.g. Lang 1965, Fiers 1986, 1988) did not recognize or consider them to be of phylogenetic relevance. Here a justification for a monophylum Paralaophontodes is presented based on a combination of unambiguous autapomorphies, demonstrating that Laophontodes armatus, L. hedgpethi and L. psammophilus must be displaced into Paralaophontodes.

All species treated in the present contribution (Fig. 13) share 16 distinct and exclusive apomorphies (Table 3), which are therefore supposed to have evolved in a common ancestor. For outgroup comparison, remaining Laophontodinae was chosen (Table 3, “outgr.”) and, where being appropriate, even taxa of the further phylogenetic surroundings (Ancorabolinae, Cletodidae [part.]) were included.

Character 1 – Rostral tip distinct, knob-like: Compared with other Laophontodinae, those species considered here exhibit a protruded rostrum as do Lobopleura and Probosciphontodes, and also several Ancorabolinae (e.g. Ancorabolina, Ancorabolus, Dorsiceratus, Echinopsyllus, Pseudechinopsyllus). Moreover, many other harpacticoid taxa show a (more or less) strongly protruded rostrum (c.f. Lang 1948, Boxshall and Halsey 2004). This suggests a convergent rostral elongation/diminution across the Harpacticoida even where close phylogenetic relationships are lacking. However, Paralaophontodes and Laophontodes armatus, L. hedgpethi and L. psammophilus carry a rostrum of a particular shape: it is triangular with a broad base (Fig. 13, but note Fig. 13F) and a distinct tip that is small and knob-like. This rostrum type is unique not only within Laophontodinae but also within Harpacticoida. It is assumed to have originated in a common ancestor and is therefore regarded as a shared apomorphy.

Character 2 – Cphth with dorso-median ridge extended into 2 posteriorly directed blunt conical elevations: The formation of dorsal sclerotized structures on cphth and body somites has been considered to be one of the characteristic features of the (paraphyletic, cf. George and Müller 2013) family Ancorabolidae Sars, 1909, but similar cuticular processes are also recorded in other harpacticoid taxa (e.g. Argestidae Por, 1986 [part.], Cerviniinae Sars, 1903 [part.], Idyanthidae Lang, 1944 [part.], Laophontidae T. Scott, 1905 [part.]). However, Paralaophontodes and Laophontodes armatus, L. hedgpethi and L. psammophilus share a unique structure dorsally on the cphth. It consists of a strongly sclerotized ridge running medially along the cephalic longitudinal axis. Its posterior half splits into 2 backwardly directed branches that each terminate in a blunt, more or less conical process, carrying a sensillum apically. Although these ridges/processes vary in shape between species, their general appearance is identical, leading to the conclusion that they evolved in a common ancestor, and supporting the hypothesis of a common evolution. This character is considered as autapomorphy of Paralaophontodes.

Character 3 – Cephalic dorso-median ridge covered with tuft of hair-like setules in anterior half: A tuft of hair-like setules covers the anterior part of the cephalic ridge. Lang (1965) noted a similar tuft of “hairy” setules in the genus Echinolaophonte Nicholls, 1941, but this without doubt belongs to Laophontidae. Thus it has evolved independently in Echinolaophonte, while being considered here as synapomorphic for Paralaophontodes, Laophontodes armatus, L. hedgpethi and L. psammophilus. Although Mielke (1981) did not mention this “hairy” tuft in Paralaophontodes exopoditus, examination of individuals from Ednago Island (Papua New Guinea) and from Dahab (Egypt, Red Sea) similarly reveals the presence of small and fine hair-like setules, resembling those described by Fiers (1986) for P. echinatus. Thus, this character can be interpreted as autapomorphic for Paralaophontodes.

Characters 4–8 – P2–P6-bearing thoracic somites with dorsal pair of processes: The presence of dorsal (and often also dorso-lateral and/or lateral) cuticular processes is considered characteristic for Ancorabolinae. In Laophontodinae, however, only Paralaophontodes has dorsal sensilla-bearing processes at the thoracic somites (Fig. 13), all other laophontodin taxa instead may or may not carry small socles (Fig. 14). The dorsal processes in Paralaophontodes are weakest in Laophontodes hedgpethi and in P. robustus, being most apparent in P. elegans and P. exopoditus. Nevertheless, the general development of dorsal processes on the thoracic somites is considered autapomorphic for Paralaophontodes.

Characters 9–11 – First to penultimate abdominal somites with pair of A- or H-shaped processes dorsally: Similar to the circumstances regarding the thoracic somites, Paralaophontodes and Laophontodes armatus, L. hedgpethi and L. psammophilus are exceptions within Laophontodinae in carrying very characteristic processes on the three abdominal somites, but not the telson. The processes are paired, strongly sclerotized, and close together at the dorsal posterior margin of each abdominal somite. These processes are (not always) connected by a transverse cuticular ridge (Fig. 13), forming an “H”-shape (“H-förmige Chitinvorsprünge”, Mielke 1981, p. 92); whilst in P. elegans they are additionally fused at their bases, the “H” therefore becoming a squarish “A” (Fig. 13F). Although the shape and ornamentation of these processes varies between species, and even between individuals of the same species (cf. Figs 2A, A’, 6, 7), the similarity of their general appearance suggests that they evolved in a common ancestor and therefore constitute autapomorphies of Paralaophontodes.

Character 12 – Male antennule 5-segmented: Whereas in Ancorabolinae the subchirocer male antennule retains eight segments (Fig. 15A), it shows only seven segments in Laophontodinae (Fig. 15B) due to the loss of the original penultimate segment. This is therefore being considered apomorphic. Within Laophontodinae, however, this state has been retained in Laophontodes only (Fig. 15B) with further successive segment fusions in other genera. In a first step, segments 6 and 7 are fused, leading to a chirocer male A1 (Algensiella, Calypsophontodes Gheerardyn and Lee, 2012, Lobopleura, Paralaophontodes, Patagoniaella, Probosciphontodes, Tapholaophontodes) (Fig. 15C). In a second step, segments 4 and 3 are fused (Fig. 15D), as seen in Paralaophontodes (Mielke 1981, Fiers, 1986, and e.g. in Laophontodes armatus, see Fig. 11A present contribution). Furthermore, male descriptions of Algensiella boitanii Cottarelli and Baldari, 1987, A. laurenceae (Bodiou and Colomines, 1988), Patagoniaella vervoorti Pallares, 1968 and Tapholaophontodes rollandi Soyer, 1974 (Cottarelli and Baldari 1987, Bodiou and Colomines 1988, Pallares 1968b and Soyer 1974, respectively, and Mielke 1985) include the same type of A1. Thus, the fusion of segments 3 and 4 may constitute a common deviation of all these taxa. Nevertheless, certain caution appears to be advisable, segment 4 is quite minute and may have been overlooked in the original descriptions of the above listed males (e.g. Fiers (1988) in his otherwise excellent description of Probosciphontodes stellata Fiers, 1988 noted a 5-segmented male A1, but this was later revealed to be 6-segmented (Conroy-Dalton 2004) with a minute segment 4). Thus, before interpreting character 12 as synapomorphic for a group of laophontodin taxa, detailed revision of Algensiella, Patagoniaella and Tapholaophontodes is urgently needed. In the meantime, it is regarded as an autapomorphy of Paralaophontodes.

Character 13 – Male antennular swollen segment with strongly developed, tooth-like spine: All known males of Paralaophontodes present a characteristic strong and tooth-like spine situated proximally at the anterior margin of the swollen antennular segment 4 (cf. Fig. 11A). This feature is unique in Laophontodinae and its’ presumed closest relatives (Ancorabolinae, Ancorabolina, Cletodidae), which bear a normal-shaped seta, if at all. Therefore this strong spine is regarded as a derived character and interpreted as an autapomorphy for Paralaophontodes.

Character 14 – Mxp without syncoxal seta: Character 14 poses a certain degree of incertitude. Firstly, in older descriptions of Laophontodinae (mainly Laophontodes species), the maxilliped has been illustrated and described rather roughly, and therefore potentially not observing existing rows of spinules, the typical minute seta that accompanies the maxillipedal claw, and the syncoxal apical seta. Such an assumption seems to be justified when considering the differences between the original descriptions of e.g. Laophontodes bicornis A. Scott, 1896 or L. whitsoni T. Scott, 1902 with recent re-descriptions: In their original descriptions both species lack all of the above mentioned features, but all were observed on re-examination (George and Gheerardyn 2015).

Within the supposed sister-group of Laophontodinae, Ancorabolinae, the loss of the syncoxal seta has occurred in the whole Ancorabolus-lineage, as well as in some genera of the Ceratonotus-group (Echinopsyllus and Pseudechinopsyllus). This points towards a convergent loss of the syncoxal seta in both supposed monophyla. In Ancorabolina and in most Laophontodinae (Calypsophontodes, Laophontodes, Lobopleura, Probosciphontodes) the syncoxal seta is present (but keep in mind the comparison “old vs. recent descriptions”). Apart from the species treated here, the LaophontodinaeAlgensiella, Patagoniaella and Tapholaophontodes also lack the syncoxal seta on mxp, but it seems unlikely that they are closely related to Paralaophontodes and Laophontodes armatus, L. hedgpethi and L. psammophilus since they each lack the other apomorphies listed above. It is therefore hypothesized that in both Ancorabolinae and Laophontodinae the loss of the syncoxal seta of the mxp occurred independently (at least) twice. Considering the other apomorphies shared by the species treated here, however, it appears probable and plausible that this reduction took place in a common ancestor of Paralaophontodes and must therefore be considered as an autapomorphy for that genus.

Character 15 – P1 enp considerably strengthened, transformed into powerful appendage: Laophontodinae and part of Ancorabolinae share a “laophontoidean-like” P1 that is characterized by a 2- to 3-segmented exp of rather small size and slender shape and a 2-segmented, elongated, prehensile enp (cf. George 2006c for detailed discussion). However, Paralaophontodes and the Laophontodes “armatus-group” differ from the remaining Laophontodinae (and also from Ancorabolinae) in a remarkable strengthening of the P1 enp. Such strengthening is characterized by a broadening of the endopodal segments and the apical claw in enp-2, transforming the P1 enp from a slender, rather delicate appendage into a quite robust prehensile medium. This is considered to be autapomorphic for Paralaophontodes.

Character 16 – P1 Enp-2 distinctly elongate, reaching at least half the length of enp-1: As assumed by George (2006c), the basic morphology of ancorabolin and laophontodin P1 includes a P1 enp being about twice as long as the exp, with enp-1 being 4–5 times longer than enp-2. P1 enp-2 carries 3 elements: 1 inner subapical tiny seta, 1 apical claw of increasing strength within Laophontodinae, and 1 apical seta, often being geniculate (George 2006c). Within Laophontodinae, the Paralaophontodes and the three Laophontodes species, L. armatus, L. hedgpethi and L. psammophilus, exhibit a secondary elongation of the P1 enp-2, so that it is at least half the length of enp-1. It is regarded as a shared apomorphy of these species.

Compared to apomorphy A discussed above (see section, ‘The taxon Paralaophontodes reconsidered’: P1 exp 2-segmented), characters 15 and 16 are of greater phylogenetic relevance, as they are not simple reductions of a segment (which occurs quite often within Harpacticoida, i.e. an incongruent character). Characters 15 and 16 are instead diagnostic transformations of particular parts of a swimming leg, and are therefore congruent, detectable exclusively in the Paralaophontodes and the Laophontodes “armatus-group” species.

Additional remarks: As described for Paralaophontodes anjae sp. n., the intercoxal sclerites P2–P4 are strongly reduced. They are of a triangular, thorn-like shape and do not connect the legs. Apart from P. anjae sp. n., this condition mentioned by Mielke (1981, p. 97) only: P. exopoditus shows a hooked process (“hakenartiger Fortsatz”) on the inner coxal margin. Other Paralaophontodes descriptions neither refer to such a process nor mention the loss of the intercoxal sclerites. However, at least Lang’s (1965) illustrations of P2–P4 on Laophontodes hedgpethi do show sclerotized parts in the respective coxal area, suggesting that they correspond to the relicts of former intercoxal sclerites. Also, the re-description of L. armatus reveals a clearly reduced and very narrow, but still bow-like intercoxal sclerite (Fig. 12F, G) connecting both legs. Thus, it is supposed that the intercoxal sclerites in Paralaophontodes are reduced, a condition that would constitute a valuable autapomorphy of that taxon. However, the legs are not separated in all the species treated in the present contribution and, since most descriptions do not specifically refer to the intercoxal sclerites, this feature requires further investigation before an autapomorphy can be definitively established.

The description of Paralaophontodes robustus provided by Bŏzić (1964; as Laophontodes robustus) may perhaps not fulfil completely the current high standard of harpacticoid species descriptions; nevertheless, it is of a sufficient quality to characterize that species unambiguously. It is therefore somewhat surprising that Bŏzić’s (1964) description and the validity of L. robustus (subsequently displaced to Paralaophontodes by Lang [1965]) have been consistently doubted (Lang 1965), resulting in its synonymisation with Paralaophontodes echinatus (Wells and Rao 1987) and the transfer of all associated material (La Réunion, Bŏzić 1964; Mediterranean, Bodin 1964, 1968, Dinet 1971, 1972).

Bŏzić (1964) names (among others) three particular features characterizing the single specimen he described: (i) the fusion of P2 and P3 exps 1–3, (ii) the complete absence of a P3 endopod in the female, and (iii) pronounced dorsal cuticular processes are restricted to the first three abdominal somites. These features are not found in P. echinatus, where (i) the P2 and P3 exps 1–3 are clearly separated, (ii) the female P3 bears a small, knob-like endopod carrying 1 seta, and (iii) pronounced cuticular processes also on all pedigerous somites (Fiers 1986). Based on these strong differences between Bŏzić’s (1964) specimen and P. echinatus, a synonymy of P. robustus and P. echinatus must be refuted categorically. Such rejection is supported by comparing the female genital field: the P6 is developed as a distinct small segment bearing 2 setae in P. echinatus (Fiers 1986) whilst in P. robustus it is fused to the body and lacks any setation (Bŏzić 1964). Thus the species Paralaophontodes robustus (Bŏzić, 1964) is re-established.

Uncertainty persists with respect to the reports of P. robustus in the Mediterranean. Bodin (1964) re-described “Laophontodes armatus(?)” from Lagune du Brusc (south of Sanary-sur-Mer, France) and from Plateau des Chèvres (south of Marseille, France) but transferred it later (Bodin 1968) to Paralaophontodes robustus. Also Dinet (1971, 1972) reported P. robustus from the same region (Ile de Riou, Bay of Marseille, France). However, both authors remark that despite a general strong similarity of their specimens with that from Bŏzić (1964) (e.g., the absence of an endopod even at P3), the Mediterranean individuals present separated exopods on P2 and P3. Assuming that Bŏzić (1964) was not wrong, the fused P2 and P3 exopods constitute a highly valuable apomorphic character for Paralaophontodes robustus. Even the fact that Bŏzić’s (1964) description is based on only one female does not justify rejecting the validity of the species, since this is not uncommon when describing rare Harpacticoida. Bŏzić (1964) undoubtedly described an adult female, so the fusion of the exopods cannot be interpreted as an ontogenetic stage (i.e., not yet separated). Thus, trusting in Branko Bŏzić’s power of observation, two possibilities must hitherto be considered:

The specimen described by Bŏzić (1964) presented malformations in both the P2 and P3. This is possible, harpacticoid specimens do present malformations relatively frequent (George, pers. obs.). The fact that both Bodin (1964, 1968) and Dinet (1971, 1974) explicitly stress the strong resemblance of their Mediterranean material with Bŏzić’s (1964) specimen from La Réunion, particularly with respect to the lost P3 enp, makes the assumption of a malformation in P2 and P3 exopods in Bŏzić’s specimen somewhat plausible. It might therefore be concluded that the specimens of La Réunion and the Mediterranean belong to the same species, i.e. Paralaophontodes robustus.

P. robustus does present fused P2 and P3 exopods as autapomorphic specific character. This would mean that both Bodin’s (1964, 1968) and Dinet’s (1971, 1974) identifications were not correct and the Mediterranean specimens do not represent P. robustus.

Unfortunately, the original material is not available, and therefore additional material from both La Réunion and the French locations is needed for further comparison to determine the correct status of these records.

Paralaophontodes anjae sp. n. shares all the mentioned generic autapomorphies 1–16 of Paralaophontodes. It differs from already known species, primarily by the derived condition of the complete loss of a female P3 endopod. That endopodal loss separates P. anjae sp. n. from all other Paralaophontodes and Laophontodes “armatus-group” species, L. armatus, L. hedgpethi and L. psammophilus, with one exception: P. robustus also lacks a P3 enp in female (see previous section). Nevertheless, P. anjae sp. n. and P. robustus differ regarding the following features: (i) cephalic lateral extensions strongly triangular in P. robustus, but only moderately triangular in P. anjae sp. n.; (ii) pedigerous somites bearing P2–P5 dorsally with tiny socles in P. robustus, while bearing strong sclerotized processes in P. anjae sp. n.; (iii) P1 exp-2 with 4 bare geniculate setae in P. robustus but with 5 bare geniculate setae in P. anjae sp. n.; (iv) P2 and P3 exopodal segments fused in P. robustus, but separated in P. anjae sp. n. Therefore, the erection of a new species to assign the Chilean specimen is well-justified.

Based on the comprehensive phylogenetic discussion given above, it is established that the species Laophontodes armatus, L. hedgpethi and L. psammophilus share apomorphic characters 1–16 (characters 12 and 13 unknown for L. psammophilus) with the currently accepted Paralaophontodes species and are therefore considered to be their close relatives. Fiers (1986) pooled these species into a so-called “armatus-group” and considered it a sister-group of Paralaophontodes, but distinguished them by apomorphic characters A–C. However, retaining the “armatus-group” within Laophontodes overlooks the necessity for characters 1–16 to have evolved twice, convergently, within Laophontodinae: once in the taxon Paralaophontodes and once in the taxon Laophontodes. Given the high phylogenetic relevance of all 16 apomorphies (see Character discussion above) their interpretation instead as shared apomorphies is much more plausible. The assumption of convergent evolution should be therefore applied to the apomorphies A–C previously assigned to Paralaophontodes. Consequently, the reassignment of L. armatus, L. hedgpethi and L. psammophilus from Laophontodes into Paralaophontodes is undertaken, re-naming the corresponding species Paralaophontodes armatus (Lang, 1936) comb. n., P. hedgpethi (Lang, 1965) comb. n., and P. psammophilus (Soyer, 1974) comb. n.

The above presented discussion confirms that the taxon Laophontodes is actually quite a heterogeneous collection of species (George and Gheerardyn 2015) that urgently needs to be re-examined. In addition to previous revisionary work (Lang 1965, Conroy-Dalton 2004, Gheerardyn and Lee 2012), the reassignment of the “armatus-group” into Paralaophontodes is another important step towards an elucidation of the systematics of Laophontodes and even Laophontodinae.

Concerning the differences that led to the “armatus-group” previously being retained in Laophontodes (i.e. the 3-segmented P1 exp, the retention of an apical seta in P1 enp-2, and the presence of endopods in P2 and P4), these are rejected here according to the detailed character discussion presented above. With the transfer of the “armatus-group” into Paralaophontodes, these characters are assumed to have evolved within Paralaophontodes, a condition that is commonplace in Harpacticoida.

While the monophylum Paralaophontodes is in the author’s opinion well-founded, unambiguously supported by autapomorphies 1–16 (see above), the relationships within Paralaophontodes are difficult to discern. Thirty-four additional morphological characters (17–50, Tab. 3) were included in the phylogenetic analysis. Possible phylogenetic relationships inside Paralaophontodes are discussed in detail below and summarized in Figure 16.

A supposed basal position within the genus may correspond to P. hedgpethi comb. n. It is the only species to show the assumed plesiomorphic state in characters 17–19 (the retention of an inner seta in the second exopodal segment of P2–P4) and in character 20 (the retention of an inner seta also in P2 exp-3). It can be justified as a distinct species due to character 21 (the exclusive possession of long and flexible setules on the abdominal dorsal processes: Lang 1965). Outgroup comparison (Ancorabolinae, Cletodidae) confirms the uniqueness of such setules, so they are interpreted as a derived state for P. hedgpethi comb. n.

P. hedgpethi comb. n. is followed by P. psammophilus comb. n., as both species retain a P2 endopod (character 22) as well as an inner seta on P3 and P4 exp-3 (characters 23, 24). In contrast, all remaining Paralaophontodes species share the derived condition that is the complete loss of a P2 endopod, and the loss of inner setae on P3 and P4 exp-3. Nonetheless, P. psammophilus comb. n. can be characterized by an autapomorphy, namely the transverse extension of the P1 basis (character 25) to which the P1 exopod is connected. Such transverse extension is virtually absent within Laophontodinae (exception: Laophontodes gracilipes Lang, 1936) but expressed in Ancorabolina and showing its strongest development in Ancorabolinae.

Next in the systematic hierarchy might be P. armatus comb. n. which, together with P. hedgpethi comb. n. and P. psammophilus comb. n., holds the ancestral states of a series of characters: possession of a geniculated seta on P1 enp-2 (Character 26), P1 exopod still 3-segmented (character 27), P2–P4 exp-3 inner apical seta long and flexible (characters 28–30), P3 female endopod 2-segmented (character 31) and carrying 2 setae (character 32), and P4 still bearing an endopod (character 33). Otherwise, Paralaophontodes armatus comb. n. may be characterized by the presence of 1 (female) or 2 (male) fishbone-like setae at the P5 (character 34). The corresponding setae in the remaining Paralaophontodes species and in the Laophontodinae in general are usually bipinnate. In P. armatus comb. n. these setae show strongly developed pinnae which are fused to the seta, giving a fishbone-like appearance. These fishbone-like setae are considered as an autapomorphy of P. armatus comb. n.

The remaining clade [P. anjae sp. n.—P. echinatus—P. elegans—P. exopoditus—P. robustus] is characterized by the derived states of characters 26–33 (see Table 3, Fig. 16). The relationships within that clade are, however, somewhat ambiguous. While some species may be characterized by unambiguous autapomorphies (P. exopoditus: character 43; P. robustus comb. n.: characters 44, 45; P. elegans: characters 46–48; cf. Table 3), P. anjae sp. n. shares its specific deviation (character 49) with P. robustus, whilst P. echinatus, as yet cannot be characterized by an autapomorphy. Similar incongruence is observed if ascending further in the cladogram (Fig. 16). Paralaophontodes echinatus, P. elegans, and P. robustus appear more closely related, sharing 3 deviations: loss of 1 geniculated seta in P1 exp-2 (Character 35), loss of the abexopodal seta in A2 (character 36), and loss of the antennary exopod (character 37). Furthermore, Paralaophontodes elegans and P. echinatus also appear closely related, showing greatest similarity in the derived shape of their thoracic dorsal processes, which are H-like only in these two species (characters 38–42). In contrast, the allocation of P. exopoditus is difficult. It shares any synapomorphy neither with P. anjae sp. n. nor with any of the other three species (Table 3, Fig. 16).

With respect to characters 36 and 37, it has to be conceded that both characters are somewhat weak. The loss of the abexopodal seta in A2 (36) is also described for the presumed most primitive P. hedgpethi comb. n., and loss of the A2 exopod (37) is also recorded in P. anjae sp. n., P. armatus comb. n. and P. hedgpethi comb. n. (cf. Table 3, Fig. 16). In particular the here supposed convergent loss of the A2exp (37) in four clades – (1) [P. hedgpethi], (2) [P. armatus], (3) [P. anjae sp. n.], and (4) [P. robustus—P. echinatus—P. elegans] – contradicts the principle of parsimony. It would be more parsimonious assuming that the loss of the A2exp occurred once in the Paralaophontodes groundpattern, while reversing subsequently in P. robustus and P. exopoditus. Nonetheless, also the latter implies certain problems. The reduction of appendages is a comparatively common occasion within Copepoda, described as principle of oligomerization (e.g. Huys and Boxhall 1991), whereas the formation of new armature elements happens only quite sporadically in Harpacticoida (Huys 1996), and the secondary development of previously lost segments has not been documented for Copepoda so far. Therefore, manifold convergent loss of the A2exp within Paralaophontodes sounds more plausible and probable than its secondary and independent re-formation.

The resulting partly unsatisfactory systematic resolution within Paralaophontodes may be an effect from limitations in information available on morphological characters, and therefore genetic comparison should be considered in future studies (but see below). Nevertheless, this lack of resolution undoubtedly also results from insufficient and incomplete species descriptions that can and should be resolved in the future. As an example Table 3 includes characters 48 and 50 that, despite their presumed high phylogenetic value, have not been described for P. psammophilus comb. n. and P. robustus. Similarly, other mouthparts that might provide important additional phylogenetic information are not described for all species. This clearly demonstrates the importance of species re-descriptions in the context of phylogenetic analyses.