Ancorabolina, AND ITS SYSTEMATIC STATUS INSIDE

Gheerardyn, Hendrik & George, Kai Horst, 2010, New representatives of the genus Ancorabolina George, 2006 (Copepoda, Harpacticoida, Ancorabolidae) including remarks on ancorabolid phylogeny, Zoological Journal of the Linnean Society 158 (1), pp. 16-55 : 42-53

ANCORABOLINA AND ITS SYSTEMATIC STATUS INSIDE ANCORABOLIDAE

The division of Ancorabolidae into the two subfamilies Ancorabolinae and Laophontodinae by Lang (1944, 1948) has been questioned repeatedly ( Gee & Fleeger, 1986; Conroy-Dalton, 2004; George, 2006c). Obviously, no autapomorphies have been detected for Laophontodinae so far, being defined by plesiomorphies only ( George, 2006c) and therefore constituting a classical paraphylum (cf. Ax, 1984). When carrying out the assignment of A. chimaera to Ancorabolinae , George (2006c) based his argumentation on ten derived characters, which are restricted to (at least certain) Ancorabolinae plus A. chimaera (plesiomorphies in square brackets):

1. Antenna without exopod (all Ancorabolinae ) [antenna exopod present];

2. Basis of P1 transversely elongate (all Ancorabolinae ) [basis transversely not elongate];

3. Cuticular processes on cephalothorax and/or body somites (all Ancorabolinae ) [cephalothorax/body somites without cuticular processes];

4. Cephalothorax with one pair of lateral processes posteriorly (in Echinopsyllus Sars, 1909 , Polyascophorus George, 1998a , and Pseudechinopsyllus George, 2006b ) [cephalothorax without lateral processes];

5. Rostrum small and constricted, resembling that of Arthuricornua Conroy-Dalton, 2001 , Dorsiceratus Drzycimski, 1967 , Polyascophorus , and Touphapleura Conroy-Dalton, 2001 [rostrum triangular, of normal shape and size];

6. Frontal part of cephalothorax forming a ‘peak’ (all Ancorabolinae , cf. George, 2006b) [cephalothorax without ‘peak’];

7. Thoracic somites with a single, long tube pore dorsally (all Ancorabolinae ) [dorsal tube pore absent];

8. Antennule first segment elongate, with long spinules on inner margin (as in Ceratonotus -group sensu Conroy-Dalton, 2001) [first segment not elongate];

9. Telson shorter than broad, trapezoid in shape (as in Ceratonotus -group) [telson rectangular];

10. Furcal rami long and divergent, inserted at outer corners of telson and directed upwards (as in Ceratonotus -group) [furcal rami not divergent, not directed upwards].

Based on new information from the four new species described in this paper and by thorough comparison with Laophontodinae , the following comments can be made:

Character 2

Within the genus Ancorabolina , there is a gradient from a slightly to a strongly pronounced transverse elongation of the basis in P1. In A. belgicae and A. anaximenesi , the basis is only slightly elongated transversely and merely forms a pedestal for the exopod, whereas the basis in A. galeata is strongly elongated transversely. The intermediate condition is found in A. chimaera and A. divasecunda . Because this morphocline is not easily dissolved into discrete character states and does not reach the condition as found in the remaining Ancorabolinae (i.e. with a morphology similar to the basis of P2–P4), we prefer to omit that character in our phylogenetic analysis for the moment.

Character 3

Several members of Laophontodinae , i.e. Paralaophontodes Lang, 1965 , Laophontodes armatus Lang, 1936 , Laophontodes hedgpethi Lang, 1965 , and Laophontodes psammophilus Soyer, 1975 , are characterized by the presence of paired processes along the posterior margins of the cephalothorax, all prosomites, and all urosomites (except for the telson). However, although this character is not restricted to Ancorabolinae , it is clear that its presence in Laophontodinae rather points to the paraphyletic nature of the latter, and illustrates the need for a phylogenetic re-evaluation and the removal of certain taxa from Laophontodinae . Therefore, these cuticular processes should be analysed regarding their homologous development.

Character 4

In addition to Ancorabolina , lateral processes posteriorly on the cephalothorax are present in several other ancorabolin taxa, namely Ancorabolus Norman, 1903 , Arthropsyllus Sars, 1909 , Juxtaramia Conroy- Dalton & Huys, 2000, Breviconia Conroy-Dalton & Huys, 2000 , Echinopsyllus , Polyascophorus , Pseudechinopsyllus , and Uptionyx Conroy-Dalton & Huys, 2000 . However, this character should be considered in greater detail, as there are some arguments contradicting the assumption of the processes being homologous:

1. Some processes deviate considerably regarding their shape and ornamentation. Although those of Ancorabolus , Arthropsyllus , Juxtaramia , Breviconia , and Uptionyx may be homologized without problems ( Conroy-Dalton & Huys, 2000), the processes of the remaining taxa may not (yet?), showing completely different shape and position of elements (sensilla).

2. Each taxon may be characterized by meaningful apomorphies that are not shared by the other taxa. For instance, Ancorabolus , Arthropsyllus , Juxtaramia , Breviconia , and Uptionyx form the so-called ‘ Ancorabolus -lineage’, considered as a monophylum as evidenced by several apomorphies (Conroy- Dalton & Huys, 2000). Also, the remaining genera are well established by unique derived characters (cf. George, 1998a, b, 2006b; Conroy-Dalton, 2003b; Wandeness, George & Santos, 2009).

This suggests a rather convergent development of the lateral processes at least within the so-called ‘ Ceratonotus -group’ ( Conroy-Dalton, 2001; extended by George, 2006b). The pair of posterior, strongly cuticularized processes of the cephalothorax in Ancorabolina are located ventrolaterally and turn backwards (character 16). We consider this character as autapomorphic for Ancorabolina . Another species showing such lateral processes is the laophontodin Laophontodes bicornis A. Scott, 1896 . A more complete comparison between Ancorabolina and La. bicornis is provided below.

Character 5

The development of a small constricted rostrum is restricted to some members of the Ceratonotus -group [ Arthuricornua , Ceratonotus Sars, 1909 (complete loss), Dorsiceratus , Polyascophorus , Touphapleura ], whereas the remaining Ancorabolinae ( Ancorabolus - lineage, Echinopsyllus , Pseudechinopsyllus ), including some species of Ancorabolina , deviate from that rostral type, showing an elongation of the constricted rostrum (cf. Conroy-Dalton & Huys, 2000; George, 2001, 2006b, c; Conroy-Dalton, 2003b; Wandeness et al., 2009; present species descriptions).

Character 7

The presence of single, dorsomedian tube pores on thoracic somites is not restricted to Ancorabolinae but also detectable in some laophontodin species (e.g. Laophontodes maccklintocki Schizas & Shirley, 1994 , Laophontodes spongiosus Schizas & Shirley, 1994 , Lobopleura ambiducti Conroy-Dalton, 2004 , Probosciphontodes Fiers, 1988 ) and therefore worthless for characterization of Ancorabolinae .

Character 8

The elongation of the first antennular segment is least expressed in A. galeata . However, when compared with ‘typical’ Laophontodes T. Scott, 1894 species , such as e.g. Laophontodes whitsoni T. Scott, 1912 , Laophontodes typicus T. Scott, 1894 , Laophontodes gracilipes Lang, 1936 , and La. bicornis , the segment in A. galeata is clearly derived, already showing the shape of all corresponding Ancorabolina species.

Sharpening George’s (2006c) argument, we state that at present from the listed characters, only two (1, 6) may be recognized as true autapomorphies for a monophyletic Ancorabolinae (including Ancorabolina ), being these characters present in all corresponding species. Characters 2 and 3 may further support this assumption, but are presently rather weak in view of the above-discussed difficulties. Furthermore, characters 8, 9, and 10 may even support a closer relationship of Ancorabolina to the Ceratonotus -group. Nevertheless, the two above-listed apomorphies (1, 6) should sufficiently justify the allocation of Ancorabolina into Ancorabolinae , in particular, if the morphological similarity with laophontodin taxa is based on plesiomorphies only, as stated by George (2006c). He based his argument mainly on two additional characters that are usually considered as important apomorphies of Ancorabolinae ( Lang, 1948; George, 2006c):

11. Female antennule at most four-segmented [fivesegmented];

12. Exp-3 of P2–P4 with two outer spines only [with three outer spines].

All Ancorabolinae present these characters, except Ancorabolina , which shares the plesiomorphic conditions with Laophontodinae ( George, 2006c) . Although no autapomorphies for Laophontodinae have been recognized so far, we detected some characters that are widespread in Laophontodinae and also present in Ancorabolina , but absent in the remaining Ancorabolinae :

13. Antennule second segment with outer bump bearing some long spinules [without bump];

14. P1 coxa lengthways elongate [P1 coxa small, square];

15. P1 seta of (former) exp-2 geniculate [element formed as a bipinnate spine];

Character 13

The second segment of laophontodin antennules, as well as those of Ancorabolina , show a rounded, bumplike expansion at the outer margin and bear several long spinules. For example, this structure is quite clear in the original descriptions of Algensiella boitanii Cottarelli & Baldari, 1987 , La. typicus , La. bicornis , La. whitsoni , La. hedgpethi , La. psammophilus , La. spongiosus , Laophontodes mourois Arroyo, George, Benito & Maldonado, 2003 , Paralaophontodes exopoditus Mielke, 1981 , Probosciphontodes , Lobopleura Conroy-Dalton, 2004 , and Tapholaophontodes Soyer, 1975 . Its shape, position, and restriction to certain species point to a homologous structure. Thus, although not necessarily constituting an autapomorphy for a monophyletic Laophontodinae , it may point to a closer relationship between some laophontodin species and Ancorabolina if proved to be homologous. This would weaken the assumption of Ancorabolina as a member of Ancorabolinae , where the outer margin of the (former) second antennular segment is straight without a tuft of spinules.

Character 14

The remarkable lengthways elongation of P1 coxa is present in all species of Ancorabolina . Compared with the remaining Ancorabolinae showing a small and nearly square coxa, this elongation can be considered as apomorphic. However, in view of the remarkable transformation of ancorabolid natatorial legs, both types of coxa – the elongate as well as the shortened one – may constitute derived stages that evolved from a primitive ancestor, which possibly showed cletodidlike coxae and bases on its swimming legs. Although it is a unique and derived character of Ancorabolina compared with the remaining Ancorabolinae , a lengthways elongate P1 coxa is rather common within Laophontodinae . Several laophontodin taxa [e.g. Probosciphontodes ptenopostica Fiers, 1988 , Paralaophontodes echinata (Willey, 1930) , Pa. exopoditus , La. armatus , La. hedgpethi , La. macclintocki ] show such an elongate coxa. Although such derivations may occur independently in different taxa, their presence in laophontodin taxa and Ancorabolina may be an indication of a closer relationship of Ancorabolina with certain members of Laophontodinae . However, as the phylogeny of that supposed paraphylum is not resolved, any assumption remains speculative. Furthermore, especially in older descriptions of laophontodin species, the P1 coxa has not been drawn completely and this hampers comparison.

Character 15

For this character the same argument applies as for character 14. Compared with Ancorabolinae , the transformation of the corresponding spine into a geniculate seta may constitute a true apomorphy for Ancorabolina . All remaining Ancorabolinae retain a normal-shaped element, i.e. a nongeniculate, bipinnate spine – this being the ‘typical’ and plesiomorphic condition widely distributed in Harpacticoida . However, in most species of Laophontodinae , the outer armature element of exp-2 (in the threesegmented exopod) or the most proximal element of exp-2 (in the two-segmented exopod, i.e. in Pa. exopoditus and Tapholaophontodes remotus Cottarelli & Baldari, 1987 ) is a geniculate seta, as in Ancorabolina . The descriptions of certain species in Laophontodinae are quite poor and probably missed the geniculation of this seta (i.e. in Laophontodes propinquus Brady, 1910 , Laophontodes latissimus Brady, 1918 , La. gracilipes , Laophontodes ornatus Krishnaswamy, 1957 , and Paralaophontodes elegans Baldari & Cottarelli, 1986 ). Furthermore, Pa. elegans and Pa. echinata only bear four setae on the second segment of their two-segmented exopod, which implies that one seta has been lost. Also, Patagoniaella vervoorti Pallares, 1968 forms an exception with a nongeniculate outer spine (but cf. George, 2006c).

The difficulties in allocating Ancorabolina within Ancorabolidae clearly demonstrate the urgent need for a phylogenetic re-evaluation of the subfamily Laophontodinae . Furthermore, phylogenetically important characters (such as the geniculation of setae, the presence of minute setae, etc.) probably have been missed in most of the older descriptions, implying the need for redescription of known species. Presently, we retain the genus Ancorabolina in the subfamily Ancorabolinae based on two true synapomorphies (1, 6) for the corresponding species.