Polycladida, Lang, 1884 [viz. Faubel, 1983/, 1984

Bahia, Juliana, Padula, Vinicius & Schrödl, Michael, 2017, Polycladida phylogeny and evolution: integrating evidence from 28 S rDNA and morphology, Organisms Diversity & Evolution (New York, N. Y.) 17 (3), pp. 653-678 : 674-675

Polycladida reclassification

The position of traditionally acotylean Theama and Cestoplana in our trees, grouping with or within Cotylea , is an important finding of our work and calls for a revision of the classification of polyclad suborders. Since Lang (1884) proposed both suborders, many authors disputed the character B exceptions^ in Cestoplana ( Laidlaw 1903c; Marcus & Marcus 1966; Faubel 1984, Curini-Galletti et al. 2008), but all conservatively let the genus in Acotylea. Some authors emphasized that suborders were non-monophyletic because of the lack of synapomorphies ( Faubel 1984). However, the synapomorphies are there ( Fig. 3 ). The main problem was that the whole superfamily systematics ( Faubel 1983, 1984; Prudhoe 1985) was based on one or few characters. Lang (1884) already criticized such a procedure as inadequate to get to a natural system of classification; in his opinion, systematists should look to the whole set of characters present in a polyclad. For nomenclatural stability, we do not create new names here but rather emend the traditional concept and maintain both suborders as monophyletic groups. We thus transfer both families, Cestoplanidae and Theamatidae , to Cotylea based on the morphological and anatomical characters ( Table 2) shared with this suborder and based on new molecular data presented here. We also use as evidence that both genera grouped in our trees with morphologically similar taxa, such as Boninia ( Fig. 8 ), which do not show clear clusters of eyespots either ( Marcus and Marcus 1968). Additionally, our results are corroborated both by Rawlinson and Stella (2012) who showed a Cestoplana clustered with cotyleans (28S rDNA) instead of in the Acotylea branch and by Laumer and Giribet (2014), who showed Theama grouped with cotyleans instead of acotyleans in a Bayesian tree (28S and 18S rDNA and cytochrome B data). The hypothesis tests ( Table 3) are compatible with our hypothesis of a monophyletic Cotylea with Cestoplana and Theama as part of it.

As Acotylea diagnosis, we consider the presence of Lang’ s vesicle, male reproductive system directed backwards, tentacles, when present, nuchal. Eyespot groups can be nuchal, cerebral, and marginal ( Table 2). Cotylea are diagnosed by the presence of an adhesive structure behind the female gonopore, cement gland pouch, male reproductive structures directed forward or perpendicular, and cerebral, marginal/frontal, or pseudotentacular eyespot groups ( Table 2). Cases of species that presented adhesive structures inside Acotylea showed that structure beside or before the female gonopore. Thus, we did not consider them to be homologous to more posterior adhesive structures.

Our results indicate that some characters traditionally used for polyclad classification have more potential, in the sense of phylogenetic signal, than others. For example, the type of prostatic vesicle and eyespot distribution vary in closely relat- ed groups and the same states may appear independently in non-related groups, having a patchy distribution (Telford et al. 2015). Rawlinson and Litvaitis (2008) had pointed out that the two competing systems were not totally supported by their cladistic analyses, but no further details were given. However, some characters (e.g., tentacular or nuchal eyespot groups) considered by Prudhoe (1985) and Lang (1884) had more weight in systematics than the ones used by Faubel (1983, 1984). This can be possible because they are related to the nervous system, which has been proved a more robust character in evolutionary studies (Quiroga et al. 2015). Gross morphology, followed by sensory organs, was regarded by Rawlinson and Litvaitis (2008) as the most informative character, in comparison with features of the digestive and reproductive systems. This was corroborated by our results ( Figs. 7 and 8 ) and contradicts Laidlaw (1903c), who thought that the prostatic vesicle was the most informative character, followed by nuchal tentacles and eyes. Marcus and Marcus (1966) and Bock (1913) proposed eyespots as most informative in Acotylea, followed by orientation of male structures and male vesicular structures, and in Cotylea , pharynx would be most informative and then eyespots. Both ideas were also contradicted by our results. As suggested by Laidlaw (1903c) and Rawlinson and Litvaitis (2008), and other authors as well, reproductive characters are useful in the genus level, and color pattern can also be informative at species level ( Litvaitis et al. 2010). Since cladistic and molecular results pointed to the gross morphology (diagnosis or general morphology) as an important combination of characters, we suggest that it should be also considered in defining superfamilies, an idea already proposed by Lang (1884) on the earlier phylogenetic hypothesis. Therefore, some more parsimonious arrangements (focused on monophyletic groups) than previous systems, according to combined molecular and morphological data, are indicated here, as a new system.

In general, Acotylea can be divided in three groups that correspond to a new superfamily-level classification ( Fig. 7 ). One has rounded body, nuchal tentacles, and cerebral, nuchal, and sometimes marginal eyespots, which includes Hoploplana , Imogine , Stylochus , Idioplana , Planocera , and Paraplanocera ( Fig. 7 ) and their corresponding families Hoploplanidae , Stylochidae , Pseudostylochidae , and Planoceridae . This group is to be called superfamily Stylochoidea Poche, 1926 , as this is the oldest family name included in the group. Another group has oval to elongated body, without tentacles, and with cerebral, nuchal, and marginal eyespots. It includes Adenoplana and Phaenocelis ( Fig. 7 ) and their families, Discocelidae and Cryptocelidae . This is to be called Cryptoceloidea superfam. nov., by Cryptocelidae and Discocelidae being equally old families, but the latter gets priority by alphabetical order. And finally, one group has elongated body, without tentacles, with only cerebral and nuchal eyespots, and with true interpolated prostatic vesicle and Lang’ s vesicle, which includes Armatoplana , Notoplana , Leptoplana , Echinoplana , and Melloplana ( Fig. 7 ). The last group is to be called superfamily Leptoplanoidea Faubel, 1984 , as also used in other systems, due to nomenclatural priority; it includes families Stylochoplanidae , Notoplanidae , Leptoplanidae , Gnesioceridae, and Pleioplanidae .

On the other hand, Cotylea can be divided into five superfamilies ( Fig. 8 ). One group with extremely elongated body, frontal eyespots, and true interpolated prostatic vesicle contains Cestoplana ; this is to be called superfamily Cestoplanoidea Poche, 1926 ( Fig. 8 ). Another group, with rounded body, frontal cerebral, and marginal eyespots, includes Pericelis ( Fig. 8 ), to be called Periceloidea superfam. nov.. A clade with extremely to mildly elongated body and frontal and cerebral eyespots includes Theama , Chromyella , and Boninia ( Fig. 8 ) and the corresponding families Theamatidae , Amyellidae , and Chromoplanidae . This group is to be called Chromoplanoidea superfam. nov., as the oldest family included is Chromoplanidae . Another group with elongated body, tubular pharynx, and cerebral and marginal eyespots includes Prosthiostomum , Enchiridium , and Amakusaplana ( Fig. 8 ). This includes the family Prosthiostomidae and is to be called Prosthiostomoidea superfam. nov.. One last group with oval body, and cerebral and tentacular eyespots, and developed marginal tentacles includes all Pseudocerotidae and Euryleptidae ( Fig. 8 ). This group is to be called superfamily Pseudocerotoidea Faubel, 1984 . None of the proposed superfamilies were rejected in the hypothesis tests and are thus considered likely ( Table 3).

Among the families that were not sampled in this or other studies are the acotyleans Enantidae, Stylochocestidae , Limnostylochidae , and Apidioplanidae , which present mixed sets of morphological characters ( Faubel 1983, 1984; Prudhoe 1985). The families Euplanidae , Latocestidae , and Didangiidae may contain basal cotylean genera. Ilyplanidae , Mucroplanidae , Palauidae , Notocirridae , and Camdimboididae possibly belong to Leptoplanoidea . Discoprosthidae and Callioplanidae may once be included in Stylochoidea . Polyposthiidae and Plehniidae possibly belong to Cryptoceloidea . The family Gnesiocerotidae , despite being sampled in this study, also needs more sampling as it presents mixed morphological characters among its genera. The families Leptoplanidae , Pleioplanidae , Notoplanidae , and Stylochoplanidae deserve a separate study and should be more extensively sampled in order to be better understood and possibly divided into monophyletic groups. For the moment, we stay with the separate family approach by Faubel (1983) but cannot conclude much about the relationships between these historically confused groups. Among the cotyleans, the families Opisthogeniidae , Ditremageniidae , Dicteroidae , Diposthidae , and Stylochoididae still need to be tested, with molecular and more morphological information. Anonymidae , Laidlawiidae , and Euryleptididae possibly belong to Pseudocerotoidea .

In general, most of the early morpho-anatomical studies ( Lang 1884; Laidlaw 1903c; Bock 1913; Hyman 1955; Marcus and Marcus 1966) were successful in classifying systematic units on genus level. There are some exceptions (e.g., Pseudoceros , Leptoplana ) that were later solved by improved working methods that allowed better observation of living and fixed specimens (Newman and Cannon 2003). As indicated by Prudhoe (1985, p. 101), our results suggest that the genus Hoploplana should have its own family, Hoploplanidae , since it was recovered far away in the tree from the Leptoplanidae (Figs. 2 and 3). Prudhoe also pointed out that this family might be related to Planocera , which was corroborated by molecular data (Figs. 2 and 3). The genus Idioplana should also be included in a separate family from the Stylochidae , as proposed by Faubel (1983; Pseudostylochidae ), and not be treat- ed as a subfamily (Prudhoe 1985), since in our trees, it appears as a reciprocally monophyletic group. The position of the Pseudoceros clade as a sister group of the Pseudobiceros and Thysanozoon clade is in accordance with results of Litvaitis and Newman (2001). The relation between those two genera, however, is different, according to our molecular results; Thysanozoon appears as a sister group to the Pseudobiceros, Yungia , and Maiazoon clade, instead of originating from a Pseudobiceros clade ( Litvaitis and Newman 2001). Morphology-based cladistic results by Rawlinson and Litvaitis (2008) also pointed to a non-monophyletic Pseudobiceros .

Figures

Fig. 3 Polycladida species-level 28S rDNA ML phylogenetic tree with morphological characters plotted as potential apomorphies

Fig. 8 Cotylea 28S rDNA phylogeny showing genus gross morphology and the new systematic division of monophyletic superfamilies. Figures from Marcus (1947, 1949, 1950, 1952), Marcus and Marcus (1968), and Newman and Cannon (1994, 1996, 2000) (un- der permission of Taylor & Francis Ltd.; http://www. tandfonline.com)

Fig. 7 Acotylea 28S rDNA phylogeny showing genus gross morphology and the new systematic division of the monophyletic superfamilies. Figures from Hyman (1939b, 1939c), Hyman (1953), Marcus (1949, 1950, 1952), Marcus and Marcus (1968), and Prudhoe (- 1978) (under permission of University of Chigago Press, Australian Museum Scientific Publications and Bulletin of the American Museum of Natural History)