The historical background regarding the establishment of the taxa “Ancorabolidae”, “Ancorabolinae” and “Laophontodinae” was summarised by George (2006a) and recently updated by Lee and Huys (2019). Since the family’s erection by Sars (1909), its monophyletic status has been generally accepted, although several authors have raised concerns, particularly at the beginning of the 21^st century (e.g. Conroy-Dalton 2003a, 2004; George 2006a; Gheerardyn and George 2010; Gheerardyn and Lee 2012). More recently, George and Müller (2013) presented a detailed discussion on some characteristics indicating the paraphyletic status of the “Ancorabolidae”. As shown by these authors (and already indicated by Conroy-Dalton (2004) and George (2006a)), the monophyletic status of the “Ancorabolidae” was questioned by the ambiguity of its presumed autapomorphies. George (2006a) listed three characters as potential “ancorabolid” autapomorphies [plesiomorphic states in square brackets]:

1. Female A1 at most 5-segmented [female A1 at least 6-segmented];
2. Bases of P2–P4 transversely elongated [bases not elongated transversely];
3. P5 exp longitudinally elongated [P5 exp not elongated longitudinally].

A fourth derived character, originally identified by Lang (1948), i.e. the loss of the antennary exopod, was rejected by George (2006a), since several “ancorabolid” species are now known to bear a small, 1-segmented A2 exp.

When discussing the systematic relation of Echinocletodes Lang, 1936 with the “Ancorabolidae”, George and Müller (2013) qualified characters A–C, confirming their ambiguity – these characters are not exclusively found in the “Ancorabolidae” but widely distributed amongst several harpacticoid families and characters B and C are, in addition, not found in all “Ancorabolidae”. Moreover, these authors discussed the complexity of character B, highlighting three distinct features, namely (i) reduction and shortening of P2–P4 endopods; (ii) lateral elongation of the bases themselves, leading to a lateral shift of the exopod; (iii) shortening of the coxae, resulting in a coxa-basis borderline that does not enclose the whole basis. Consequently, George and Müller (2013) rejected the monophyletic status of the “Ancorabolidae”.

The extensive phylogenetic analyses of harpacticoid major taxa, provided by Willen (2000) and Seifried (2003), support Lang’s (1948) assignment of the “Ancorabolidae” to the Podogennonta Lang, 1944 (Fig. 2, branch *). Although being highly derived if compared with the diagnostic characters of podogennontan groundpattern (cf. Willen 2000; Seifried 2003), the “Ancorabolidae” share apomorphies of the Podogennonta, such as the reduction of the mxp enp-2, with the mxp endopod being 1-segmented (Willen 2000; Seifried 2003). Similarly, in some “Ancorabolidae” (“Laophontodinae”; “Ancorabolinae”: Ancorabolus-lineage sensu Conroy-Dalton and Huys 2000; Ancorabolina), the shape of the P1 corresponds to that of the podogennontan groundpattern. This important affiliation is discussed in detail below (Chapter II).

Within the Podogennonta, however, the position of the “Ancorabolidae” remains uncertain. They yet may be assigned to Willen’s (2000) “taxon II”, which encloses all Podogennonta except the (?)Harpacticidae Dana, 1846, (?)Latiremidae Bozic, 1969 and Pseudotachidiidae Lang, 1936 (cf. Willen 2000: 198, fig. 82) (altogether pooled in Fig. 2, branch **).

Lang (1948) united the “Ancorabolidae”, Laophontidae T. Scott, 1904 and “Cletodidae” in the supra-familiar taxon Cletodidimorpha Lang, 1948, whose name was changed to Cletodoidea Bowman & Abele, 1982 (Bowman and Abele 1982). However, when splitting the “Cletodidae” into several different families, Por (1986) placed the “Ancorabolidae” and Laophontidae into the Laophontoidea T. Scott, 1904, with the “Cletodidae” in its own supra-family Cletodoidea. Subsequently, Huys (1990a) established the Laophontoidea uniting the Adenopleurellidae Huys, 1990, Cristacoxidae Huys, 1990, Laophontidae, Laophontopsidae Huys & Willems, 1989 and Orthopsyllidae Huys, 1990. Huys (1990a) stated that the “Ancorabolidae” does not share any of the laophontoid apomorphies and, therefore, rejected the hypothesis of a closer “Ancorabolidae”–Laophontoidea-relationship. Later, Huys and Lee (1998/99) added the Normanellidae Lang, 1944 to the Laophontoidea (Fig. 2, branch ***), leading to further systematic complications. Although the Laophontoidea is not registered in the World Register of Marine Species (Walter and Boxshall 2020), its validity as a monophylum is generally accepted (e.g. George 2006a; Kihara and Huys 2009; Huys and Kihara 2010). This was also the case for the assumption that the “Ancorabolidae” is more closely related to the “Cletodidae” than to the Laophontoidea (Conroy-Dalton 2003a, 2004). Nevertheless, George (2006a) followed Lang’s (1948) and Por’s (1986) hypothesis of a close relationship between the “Ancorabolidae” and Laophontidae, extending it to the Laophontoidea. This was based on the shape of the P1 and a re-evaluation of laophontoid apomorphies, concluding that a close relationship between the Laophontoidea and “Ancorabolidae” still holds (George 2006a).

Against the background discussed above, the Laophontoidea and “Cletodidae” were selected as outgroups of the “Ancorabolidae” for this current phylogenetic analysis. Further comparison aimed to uncover the phylogenetic relationships amongst all three groups. A matrix of 150 phylogenetically-relevant morphological characters was created (Table 1), including the presumed “ancorabolid” autapomorphies A–C. This enabled the morphological comparison of all “ancorabolid” species with each other, as well as with the Laophontoidea, “Cletodidae” and other podogennontan taxa (e.g. the Ameiridae, Argestidae and Tetragonicipitidae).

This analysis suggests, as discussed below, that the monophyly of the “Ancorabolidae” cannot be maintained. Thus, Lang’s (1936a) suggestion of starting anew was adopted. Confirming George and Müller’s (2013) doubts, it is shown that the “Ancorabolidae” forms a polyphylum and neither the “Ancorabolinae” nor “Laophontodinae” are monophyletic. Moreover, the “Ancorabolidae” and “Cletodidae” must be re-ordered: the “Laophontodinae” (plus Ancorabolina = Laophontodinae ●) and the Ancorabolus-lineage constitute a monophyletic taxon Ancorabolidae ●, whilst the Ceratonotus-group sensu Conroy-Dalton (2001) forms part of the Cletodidae ●. Additionally, ancorabolid ●, cletodid ● and laophontoid species together form a well-supported clade within the Podogennonta, as proposed by Lang (1948) and George (2006a). This hypothesis is supported by and based upon Willen (2000) and Seifried’s (2003) phylogenetic characterisation of the Podogennonta, particularly the shape and development of the P1.

The shape, size and setation of the first swimming leg is an important indicator of systematic relationships within the Harpacticoida and particularly in the Podogennonta (Willen 2000) and has been the object of significant controversial discussion with regards to ancorabolid taxonomy (e.g. Lang 1948; Huys 1990a; Conroy-Dalton and Huys 2000; Conroy-Dalton 2001; George 2006a). As summarised by George (2006a), there have been two opposite hypotheses, namely (i) the interpretation of a prehensile P1 in the Ancorabolus-lineage as a derived stage (Conroy-Dalton and Huys 2000) and (ii) the assumption of the respective P1 as an ancestral state (Lang 1948; George 2006a). The inclusion of the Laophontoidea and “Cletodidae” as outgroups in this current analysis has proved informative to this discussion.

According to Willen (2000) and Seifried (2003), the hypothetical podogennontan ancestor (Fig. 2, branch *) bore a P1 (Fig. 3A) characterised by eleven derived features (Table 1: characters 1–11). The endopod consists of an elongated enp-1 (Table 1, character 1), a small enp-2 (Table 1, character 2) and enp-3 (Table 1, character 3), resulting in a P1 endopod longer than the P1 exopod. Additionally, the outer spine of enp-3 (Fig. 3A, element I-en) was shifted apically (Table 1, character 4) and transformed into a claw (Table 1, character 5); the middle apical seta (Fig. 3A, element 5-en) became geniculated (Table 1, character 6) and the inner apical seta (Fig. 3A, element 4-en) was diminished in size (Table 1, character 7). Thus, the endopod became prehensile. In the exopod, both apical setae of exp-3 (Fig. 3A, elements VI, VII) became geniculated (Table 1, characters 8 and 9), whilst the 2 inner setae (Fig. 3A, elements 3-ex and 4-ex) reduced completely (Table 1, characters 10 and 11). Such a P1 is still retained in several podogennontan taxa, such as the Ameiridae (part.) and Tetragonicipitidae (part.). However, within Willen’s (2000) “taxon II”, a further development of the P1 can be observed. In this study, the data suggest the following scenario:

1. In the hypothetical ancestor of the Laophontoidea–Cletodoidea-clade (L–C-clade, cf. III/A), the endopod became 2-segmented and enp-1 underwent elongation (Fig. 3B), becoming longer than the whole exopod. Additionally, the proximal seta of enp-1 (Fig. 3B, element 1-en) and the short apical seta of enp-2 (Fig. 3B, element 4-en) were lost (Table 1, characters 12–15).
2. The Laophontoidea (Fig. 2, branch ***) and Cletodoidea (cf. III/B) were separated by further evolution in the Cletodoidea, with the P1 exopod losing the inner element on exp-2 (Fig. 3C, element 2-ex) and the proximal outer spine of exp-3 (Fig. 3C, element III) (Table 1, characters 20 and 21). These elements are still present in the Laophontoidea (Fig. 3B).

Within the Cletodoidea, the ancestor of the Ancorabolidae ● (cf. III/C) retained the P1 condition shown in Fig. 3C, as did the subordinated Laophontodinae ● (cf. III/D), as still detectable in the laophontodin genus Tapholaophontodes. The sister-group of the Ancorabolidae ●, the Cletodidae ● (cf. III/V), may be characterised by the following further development of the P1:

1. The formerly elongated enp-1 strongly reduced in size, becoming, at most, as long as exp-1, whilst enp-2 increased in length (Fig. 3D; Table 1, characters 127–130). Overall, the P1 endopod in the Cletodidae ● is shortened considerably in comparison with the podogennontan ancestor. If Willen’s (2000) and Seifried’s (2003) hypothesis regarding the podogennontan P1 groundpattern is adopted, the P1 of the Cletodidae ● exhibits several supplementary deviations. However, these are ambiguous and await more detailed evaluation, specifically: the enp-2 apical claw (Fig. 3D, element I-en) reverted to an outer pinnate spine in the subapical position; the enp-2 geniculate apical seta (Fig. 3D, element 5-en) reverted to a bipinnate or biplumose seta. Thus, the P1 in the Cletodidae ● lost its prehensile state, being the endopod at most as long as the exopod.

The secondary transformation of highly specialised elements (I-en, element 5) into a more primitive, pre-podogennontan state may sound somewhat implausible. However, as discussed by several authors (e.g. Mayr 1975; Ferrari 1988; Fiers 1990; Boxshall and Huys 1998; Wägele 2001; Seifried 2003), such reverse transformation may apparently occur, it is in fact:

…not simply a regain of the older state, but instead a new state that merely resembles the older plesiomorphic one. (Seifried 2003: 209/210).

This hypothesis serves as the basis for the secondary transformation of elements I-en and 5-en in the Cletodidae ● presented here. The alternative was that the specialisation of elements I-en and 5-en into a claw (combined with an apical shift) and a geniculated seta, respectively, do not form part of the podogennontan groundpattern as postulated and well-supported by Willen (2000) and Seifried (2003). The rejection of a secondary transformation of a prehensile P1 back to a non-prehensile P1 in the Cletodidae ● would, however, mean that it could no longer be assigned to the Podogennonta, but should be placed between the Podogennonta and the Neobradyidae Olofsson, 1917, if following Seifried’s (2003) systematic concept. As shown below, such displacement of the Cletodidae ● would cause a conflict with respect to several other derived characters that are shared with the Ancorabolidae ● (cf. III/A, III/B). It is therefore postulated that the podogennontan groundpattern of the P1 constitutes the evolutionary basis for the Ancorabolidae ● and Cletodidae ●. Consequently, the drastic transformation of the P1 in both the “Cletodidae” and the Ceratonotus-group is regarded as an evolutionary novelty that has been inherited from a common ancestor. That premise forms the basis of the here presented new phylogenetic concept.

1. The Ancorabolidae ● splits into two sister-groups, the Laophontodinae ● (cf. III/D) and the Ancorabolinae ● (cf. III/E). The latter is characterised by a further deviation of the prehensile P1, with loss of the distal inner seta (Fig. 3E, element 2-en) of enp-1 (Table 1, character 61). This is evidenced by the retention of element 2-en in Tapholaophontodes (La ophontodinae ●) and in Limnocletodes (Cletodidae ●). However, within two sub-clades–in the A–P-clade (cf. III/F) and the Cletodinae subfam. nov. (cf. III/Y) – element 2-en is subsequently reduced and is considered here as a convergent development.

To summarise, the following conclusions, with respect to P1 development, can be made: (i) the Laophontoidea, the “Ancorabolidae” and the “Cletodidae” evolved from a common ancestor with a P1 characterised by four apomorphies (Fig. 3B; Table 1, characters 12–15); (ii) the “Ancorabolidae” and the “Cletodidae” descend from a common ancestor characterised by two further apomorphies missing from the supposed sister-group Laophontoidea (Fig. 3C; Table 1, characters 20–21); (iii) the “Cletodidae” and the Ceratonotus-group share a P1 characterised by four further apomorphies (Fig. 3D; Table 1, characters 137–140), whilst the “Laophontodinae” and the Ancorabolus-lineage derive from a common ancestor that retained the P1 as shown in Fig. 3C; (iv) with the loss of element 2-en on P1 enp-1, the ancestor of the Ancorabolus-lineage (= Ancorabolinae ●) evolved a further apomorphy (Table 1, character 61), which also occurred convergent in the A–P-clade and in Cletodinae subfam. nov. (Fig. 3D, E).

This scenario provides the foundation for the phylogenetic analysis below, in which a large number of additional derived characters are discussed.

Figure 3. 

Schematic representation of hypothesised groundpatterns for the shapes of the first swimming leg P1 in each proposed taxon. A. Podogennontan (after Willen 2000; Seifried 2003), B. Laophontoidea–Cletodoidea-clade, C. Cletodoidea, D. Cletodidae ●, E. Ancorabolidae ●. Black crosses indicate apomorphies; Roman numerals = outer spines, Arabic numerals = inner setae; ex = exopod, en = endopod. Explanations given in the text.

Characters 1–11 of Table 1 focus on the derived state of the podogennontan P1 as compared with the remaining Harpacticoida. It is not the aim of this study to evaluate the phylogenetic status of the Podogennonta, which has been done extensively by Willen (2000) and was confirmed later on by Seifried (2003). Thus, characters 1–11 merely illustrate that the here treated taxa–even those that present further deviations–can generally be assigned to Podogennonta.

The phylogenetic evaluation starts with “taxon II” of Willen’s (2000) system. Compared with other podogennontan “taxon II” taxa (Ameiridae, Argestidae, Tetragonicipitidae), there is a group formed by the Laophontoidea, “Ancorabolidae” and “Cletodidae”.

Remarks: The analysis explicitly excludes two “ancorabolid” taxa, namely Echinocletodes Lang, 1936 and Patagoniaella Pallares, 1968. Echinocletodes has been excluded from “Ancorabolidae” by George and Müller (2013) and is currently assigned to the “Cletodidae” (Walter and Boxshall 2020); it is therefore not further discussed here. The assignment of Patagoniaella to the “Ancorabolidae” has been questioned already by George (2006a). Recently, Lee and Huys (2019) suggested its affiliation to the “Cletodidae”. Its phylogenetic status is discussed in detail below (cf. III/CC).

The monophyly of the Laophontoidea–Cletodoidea-clade (L–C-clade) clade is supported by eight unambiguous autapomorphies (Table 1, characters 12–19; Fig. 2, branch A). Characters 12–15 refer to the development of the P1 and were discussed above (cf. Chapter II).

Character 16, female A1 at most 8-segmented: according to Seifried (2003), the podogennontan female A1 consists of nine segments. This agrees with the groundpattern of Oligoarthra (Willen 2000), which is now synonymised with Harpacticoida (Khodami et al. 2017). Within Podogennonta, oligomerisation of the female A1 takes place in several taxa. Families such as the Ameiridae Boeck, 1865, Argestidae Por, 1986, Canthocamptidae Brady, 1880, Tetragonicipitidae Lang, 1944 and Thalestridimorpha sensu Willen (2000) include several species whose female antennules have a reduced number of segments, down to a minimum of six segments (cf. Boxshall and Halsey 2004 for overview).

The L–C-clade also reflects such oligomerisation in the A1. In the Cletodoidea, the female A1 consists of at most five segments (e.g. Conroy-Dalton 2004; George 2006a, 2017, 2018; George and Gheerardyn 2015) and the laophontoid Adenopleurellidae, Cristacoxidae, Laophontopsidae and Orthopsyllidae are characterised by a 4-segmented female A1 (Huys and Willems 1989; Huys 1990a, b, c). Nonetheless, as several species of Normanellidae and Laophontidae present female antennules with up to six and eight segments, respectively (e.g. Willen 1992; Huys 1990a; Lee and Huys 1999a; Kihara and Huys 2009), it is concluded that the ancestor of the L–C-clade carried an 8-segmented female A1. Compared with the podogennontan groundpattern, that number of segments is regarded here as deviation.

Character 17, A2 with allobasis: an antennar allobasis is formed by the fusion of the basis and the first endopodal segment (Lang 1948). This derived condition is found in all representatives of the L–C-clade but not in the groundpattern of other “taxon II”-Podogennonta (e.g. Ameiridae, Tetragonicipitidae): exceptions occur in two species of Normanella Brady, 1880 (Laophontoidea: Normanellidae), namely N. bolini Lang, 1965 and N. pallaresae Lee & Huys, 1999, in which the basis and enp-1 are discernible by a transverse suture (Lang 1965; Packmor and Riedl 2016).

Another conspicuous feature of the A2 is the 1-segmented antennar exopod, which is observable in the whole L–C-clade. This might be interpreted as apomorphic character when compared with Ameiridae (cf. Boxshall and Halsey 2004) and Thalestridimorpha (Willen 2000). Nevertheless, a 1-segmented A2 exopod is heterogeneously distributed over the Podogennonta (e.g. Cletopsyllidae Huys & Willems, 1989, Rhizotrichidae Por, 1986, Tetragonicipitidae). Thus, whilst the phylogenetic relationship of the L–C-clade with the remaining “taxon II”-Podogennonta remains unclear, a 1-segmented A2 exopod cannot be used for phylogenetic comparison, as it might indicate a close relationship within the Podogennonta at a higher taxonomic level.

Character 18, A2 endopod with only one slender seta accompanying the two spines at the distal edge: Within the relatively derived Podogennonta (Thalestridimorpha, Tetragonicipitidae; cf. Willen 2000), the A2 endopod still bears, apart from two strong spines, two very slender and fine juxtaposed setae. In the L–C-clade, however, one of these setae is reduced; this is seen as an autapomorphy of that clade compared with the remaining Podogennonta.

Character 19, endopods of P2, P4 and female P3, slender and 2-segmented: In the groundpattern of the Podogennonta, swimming legs have 3-segmented endopods (cf. Willen 2000). However, the Laophontoidea, “Ancorabolidae” and “Cletodidae” show 2-segmented P2, P4 and female P3 endopods. Moreover, the endopods are much more slender than the respective exopods. Additionally, enp-1 becomes shorter than enp-2. In addition to the named taxa, this endopodal shape is also observed in the Argestidae (part.), Cletopsyllidae, Rhizotrichidae (part.) and Tetragonicipitidae. Nevertheless, these latter taxa are considered as distinct monophyla (Kunz 1984; Por 1986; Huys and Lee 1998/99; Corgosinho and Martínez Arbizu 2010; Packmor 2013; Gheerardyn and George 2019), owing to the absence of autapomorphies of the L–C-clade. Thus, the development of slender, 2-segmented P2–P4 endopods in the Argestidae (part.), Cletopsyllidae, Rhizotrichidae (part.), Tetragonicipitidae and in the L–C-clade is regarded as convergence. For the L–C-clade, the development of such endopods forms part of a set of derived characters. It is assumed here that their development took place in the common ancestor of the Laophontoidea, “Ancorabolidae” and “Cletodidae”, being synapomorphic for these three taxa.

Remarks: Character 19, along with characters 33, 36, 62, 63, 64, 78, 84, 98, 112, 113, 123, 124 and 143 constitutes a character complex, as it pools the respective character changes (seta reduction, seta/segment elongation or shortening) of three different pairs of swimming legs. However, to date, they have always been observed together; if, in future, new species present morphological changes in only single pairs of swimming legs, they can be split into single characters.

As stated by Lang (1948), the “Ancorabolidae” and “Cletodidae” form a monophyletic major taxon, shown here to be unambiguously supported by 19 autapomorphies (Table 1, characters 20–38; Fig. 2, branch B). These evolved in a hypothetical common ancestor being exclusively present in “Ancorabolidae” and “Cletodidae” species, but missing in the sister-group Laophontoidea.

Characters 20 and 21 refer to the P1 and have been discussed above (Chapter II).

Character 22, Rostrum: According to Huys and Boxshall (1991), the harpacticoid groundpattern has a distinct rostrum, being defined at its base and this, therefore, represents the ancestral state within the Harpacticoida. While several Laophontoidea still retain this plesiomorphic state (e.g. Huys and Willems 1989; Huys 1990a; Willen 1992, 1995), it is fused to the cephalothorax in all species of the Cletodoidea and interpreted here as an autapomorphy for that taxon.

Characters 23–26, development of the female antennule: The most plesiomorphic female A1 within the “Ancorabolidae” and “Cletodidae” consists of five segments (character 23), being a clear apomorphy compared with the Laophontoidea, which retains an up to 8-segmented female A1 (cf. III/A, character 16). The female A1 in the Cletodoidea is quite characteristic, with segments 1–3 of almost equal length, segment 4 very small (character 24) and (at least partly) overlapped by the acrothek of the previous segment (character 25) and segment 5 of nearly the same length as segments 1, 2 and 3, respectively, but more slender (Fig. 4A–D). Aesthetascs are present on the third (character 26) and fifth segments. The presence of an aesthetasc on the third segment is also found in several laophontoid species (e.g. Normanellidae, Cristacoxidae, Adenopleurellidae; cf. Huys 1990a, b; Huys and Lee 1998/99); however, in several Laophontidae (e.g. Afrolaophonte chilensis Mielke, 1985, Bathylaophonte Lee & Huys, 1999, Heterolaophonte minuta (Boeck, 1872)), the aesthetasc arises from the fourth segment, even if the A1 consists of only five segments, for example, in Amerolaophontina reducta (Coull & Zo, 1982) (cf. Mielke 1985; Fiers 1991; Willen 1992; Lee and Huys 1999a). Thus, in the laophontoid groundpattern, the position of the first aesthetasc is on the fourth antennular segment, whilst its location on the third segment occurred later as a convergence.

Figure 4. 

Antennulae (A1) of A. Enhydrosoma parapropinquum Gómez, 2003, B. Cletodes meyerorum George & Müller, 2013, C. Bicorniphontodes bicornis (A. Scott, 1896), D. Ancorabolina divasecunda Gheerardyn & George, 2010. Arrows indicate gradual elongation. Modified after Gómez 2003 (A), George and Müller 2013 (B), George and Gheerardyn 2015 (C), Gheerardyn and George 2010 (D). No scales.

In the “Cletodidae”, the above-described female A1 is found in all genera with the exception of Intercletodes (four segments), Limnocletodes (four segments), Scintis (four segments) and Sphingothrix (three segments). In most species, the 5-segmented A1 is rather short and sturdy, as in Enhydrosoma (Fig. 4A), although it may be elongated and narrower, as in Cletodes (Fig. 4B).

Remarks: An exception is the “cletodid”(?) genus Barbaracletodes Huys, 2009. The females of B. barbara Becker, 1979 and B. carola Becker, 1979 bear a 7-segmented A1 (Becker 1979), being a huge exception within the “Cletodidae”. Nevertheless, the systematic status of Barbaracletodes is absolutely uncertain. Huys et al. (1996) removed it to the Canthocamptidae, which was adopted by Boxshall and Halsey (2004). Additionally, Gee (1998) stated that Barbaracletodes cannot be maintained within the “Cletodidae”; however, like Huys et al. (1996), he provided no further justification. Therefore, Wells (2007; cf. also Huys 2009) retained the genus in the “Cletodidae” as genus incertae sedis. In the present contribution, Barbaracletodes is not included in the phylogenetic analysis.

In the “Ancorabolidae”, a 5-segmented A1 is found in the subfamily “Laophontodinae”, specifically in Algensiella, Ancorabolina (Fig. 4D), Bicorniphontodes, Laophontodes (Fig. 4C), Rostrophontodes, and Tapholaophontodes. Instead, all remaining “Ancorabolidae” bear only four or even three segments in the female A1. Therefore, within the “Ancorabolidae”, a 5-segmented female A1 has always been regarded as the ancestral state (e.g. Lang 1936, 1948). As noted for the “Cletodidae”, a gradual elongation and cutting-back of the female A1 is also observed in the “Ancorabolidae”.

Due to the very specific shape of the female A1 in the “Cletodidae” and “Ancorabolidae”, it is considered here to have developed in a common ancestor of both taxa. Comparison with the sister-group Laophontoidea (and even with other Podogennonta) clearly reveals that the reduction towards a 5-segmented A1, in combination with the above-described additional morphological particularities, forms a derived state that is shared only by the “Cletodidae” and “Ancorabolidae”, thus constituting a complex synapomorphy for both taxa. The reduction of antennular segments within Laophontoidea (see above) is–analogue to the position of the aesthetasc on the third segment–consequently interpreted here as convergence.

Character 27, A2 exopod with at most three setae: while the hypothetical ancestor of the Laophontoidea is characterised by an A2 exopod bearing four setae (Laophontidae, Normanellidae (part.), Orthopsyllidae; cf. Huys 1990a), the groundpattern of the here-postulated Cletodoidea consists of at most three exopodal setae as documented for the cletodid genus Limnocletodes by Gee (1998). This is seen as autapomorphic for the Cletodoidea.

Characters 28 and 29, md palp 1-segmented, with at most six setae: the hypothetical ancestor of the Laophontoidea bore a md palp with both the exopod and endopod well-developed, a condition that is retained in, for example, Archilaophonte maxima Willen, 1995 (Laophontidae) and in the Normanellidae, whilst in the Cristacoxidae, the md endopod is retained (cf. Huys 1990a; Willen 1995; Lee and Huys 1998/99). In the Cletodoidea, however, both the endopod and exopod are incorporated into the basis, resulting in a unilobate md palp (character 28). Moreover, at least seven setae are found on the md palp of the ancestral laophontoid (two basal setae, one exopodal and four endopodal setae), as still observed in Normanella (cf. Lee and Huys 1998/99; Packmor and Riedl 2016). In comparison, all representatives of the Cletodoidea bear at most six setae on the md palp, which is regarded as autapomorphic for that group.

Character 30, mxl endopod and exopod fused with basis: while in the laophontoid groundpattern, the mxl bears both an endopod and exopod (e.g. Archilaophonte Willen, 1995, Heterolaophonte minuta, Bathylaophonte), these are lost in the Cletodoidea. This is considered as an autapomorphy of that taxon.

Character 31, mx endopod with at most two setae: in the laophontoid groundpattern, the mx bears three endopodal setae (e.g. Bathylaophonte, Normanella; cf. Huys and Lee 1998/99; Lee and Huys 1999a, b; Packmor and Riedl 2016), whilst all members of the Cletodoidea carry only two setae on the mx endopod. This latter state is regarded as a shared deviation and thus as an autapomorphy for the group.

Character 32, mxp syncoxa with at most one seta: the Laophontoidea present 2–3 apical setae on the mxp syncoxa (e.g. Laophontidae (part.), Normanellidae (part.); cf. Willen 1992, 1995; Huys and Lee 1998/99; Packmor and Riedl 2016), whilst in the Cletodoidea only one seta is retained. This is interpreted as autapomorphy of the taxon.

Character 33, extreme reduction of P2–P4 endopods: the above-discussed development of the P2–P4 endopods (cf. III/A, character 19) suffered a further deviation in the Cletodoidea: the enp-2 is transformed into a relatively slender and cylindrical segment (cf. Figs 5A–D, 6B, C). Although noted in other podogennontan taxa (e.g. Argestidae (part.)), in combination with the additional derived characters of Cletodoidea, this occurs in all known cletodoid species and is, therefore, considered as inherited from their last common ancestor and thus interpreted as an autapomorphy of that suprafamily.

Figure 5. 

A. P3 of Tapholaophontodes rollandi Soyer, 1975, B. P3 of Calypsophontodes macropodia (Gee & Fleeger, 1986); dotted circle indicates approximate position and size of coxa; C. P3 of Laophontodes sarsi George, 2018, D. P4 of Probosciphontodes stellata Fiers, 1988. Modified after Mielke 1985 (A), Gheerardyn and Lee 2012 (B), George 2018 (C), Fiers 1988 (D). No scales.