Chelidonura, A. Adams, 1850

Zamora-Silva, Andrea & Malaquias, Manuel António E., 2018, Molecular phylogeny of the Aglajidae head-shield sea slugs (Heterobranchia: Cephalaspidea): new evolutionary lineages revealed and proposal of a new classification, Zoological Journal of the Linnean Society 183, pp. 1-51 : 42-44

CHELIDONURA S. S. A. ADAMS , 1850 AND THE NEW GENERA BIUVE , CAMACHOAGLAJA AND MARIAGLAJA

Chelidonura s.l. has been assembled together based often on the shared presence of a narrow and elongated body, a tri- or quadrilobed anterior shield, calcified shell, a small non-eversible buccal bulb, a single lobed mucous gland and a simple penis ( Rudman, 1971, 1973, 1974; Gosliner & Williams, 1972; Gosliner, 1981; Ortea & Martínez, 1997). Rudman (1974) and Gosliner (1980) discussed the anatomy of several chelidonurids and disagreed regarding the disposition of the labial glands and the presence/absence of a genital ganglion, while agreeing on the uniformity of the reproductive system of Chelidonura , giving C. hirundinina as a typification for the whole genus ( Gosliner, 1987).

Anthes & Michiels (2007) were the first to hint at the possible non-monophyly of the genus, a result similarly obtained by Malaquias et al. (2009) and Oskars et al. (2015). Nevertheless, none of these works aimed to discuss the systematics of Aglajidae and it was only with the work by Camacho-García et al. (2014) that the paraphyletic status of the genus was thoroughly emphasized and discussed. The latter authors recovered a ‘superclade’ with Aglaja + Navanax + three sub-clades of Chelidonura and considered the possible synonymization of one of the clades of Chelidonura (the one containing the type species C. hirundinina ) and Navanax with Aglaja , but low statistical support and the recognizable synapomorphies of Navanax ( Gosliner, 1980) prevented synonymization of the three names. They considered a possible introduction of new names for the sub-clades of Chelidonura , but the authors found it premature because, with the exception of the Atlantic Chelidonura , the sub-clades were poorly supported.

In this study, we included sequences of 54% of the recognized valid species of Chelidonura and retrieved four clades: (1) Chelidonura s.s. with type species C. hirundinina from Mauritius (PP/BS = 0.97/71). This clade is dominated by IWP taxa, but includes two representatives from the WA ( C. cubana and C. ‘hirundinina’; Table 4). (2) The new clade Mariaglaja (type species M. alexisi from the Philippines), which includes representatives only from the IWP (PP/BS = 1/97) and corresponds in part to the sub-clade A. 2.2 in Camacho-García et al. (2014) (see Systematic results for diagnosis). (3) The clade Camachoaglaja (type species C. africana from the Canary Islands; Martínez et al., 2002) (PP/BS = 1/100) includes the majority, but not all the Atlantic species of Chelidonura s.l. (two species are part of Chelidonura s.s.; Fig. 2A, B; Table 4). The separation between Chelidonura s.s., Camachoaglaja and Mariaglaja based on external features is difficult ( Table 5). The main differences are molecular, anatomical and in part conchological. The shell of the three genera is spoon-shaped, but in Camachoaglaja , it has a highly modified apex with outward-pointed projections on the protoconch ( Fig. 5J; Ortea et al., 1996). The shells of Chelidonura s.s. and Mariaglaja are of similar size and shape with an acuminated or rounded whorl ( Fig. 5G, I; Table 5). (4) The last clade comprises the IWP ‘ C. fulvipunctata ’ and probably a cryptic species from the Hawaiian Is (PP/BS = 1/100; Figs 1, 2), for which we here introduce the genus name Biuve [type species B. fulvipunctata ( Baba, 1938) from Kii, Japan; Table 4]. This genus is also morphologically similar to Chelidonura s.s. ( Rudman, 1974; Gosliner, 1980); the main differences are molecular and the presence of a W-shaped mark in the anterior part of the cephalic shield (see Systematic results section). Species of this clade were not included in the molecular phylogeny of Camacho-García et al. (2014) and were here rendered sister to the radula-bearing new genus Mannesia restricted to the eastern Atlantic (PP/BS = 0.95/75; discussed below; Fig. 2A, B).

THE RADULATE GENERA: ODONTOGLAJA RUDMAN, 1978 AND THE NEW GENUS MANNESIA

Three species of aglajids with radula have been described so far, namely Odontoglaja guamensis ( Rudman, 1978) , O. sabadiega Ortea et al., 2002 and O. mosaica Gosliner, 2011 . Gosliner et al. (2008: 48) referred to a white species of Chelidonura from South Africa ( Chelidonura sp. 3 ) with what the author has labelled as a ‘vestigial radula’.

The evolutionary implications of radula loss in the Aglajidae have been extensively discussed by Rudman (1978) and Camacho-García et al. (2014). The first work considered Odontoglaja to be the basal or sister lineage to all other aglajids and the presence of a radula to be an ancestral feature that was lost once in the evolutionary history of the family, being therefore absent in all other lineages. Camacho-García et al. (2014), including only IWP radulate species, corroborated the view that Odontoglaja is basal within the family, but because of the vestigial presence of a radula in the aforementioned specimen of white Chelidonura ( Gosliner et al., 2008) , they speculated about possible alternative scenarios to explain the evolution of this character, suggesting (1) an eventual regain of the radula in some lineages, (2) a widespread presence of this feature in juvenile stages not detected so far, with consequent lost in adults, or (3) a possible misidentification of specimen ( Chelidonura sp. 3 ) of Gosliner et al. (2008) with a philinid.

Nevertheless, both Anthes & Michiels (2007) and Malaquias et al. (2009) have retrieved different phylogenetic scenarios with Odontoglaja branching off in more derived positions usually sister to lineages of Chelidonura s.s. The phylogeny presented here is the first to include both IWP and Atlantic lineages of ‘ Odontoglaja ’ and establishes a new paradigm in the evolution of radula-less lineages in Aglajidae . This does not necessarily conflict with the idea that radula-less lineages are basal, but also does not support this view. What seems clear is that the radula was either lost or regained at least twice in the evolution of aglajids. Stochastic reappearance of lost structures due to reactivation of ‘dormant’ genes has been documented ( Scheltema, Kerth & Kuzirian, 2003; Golubtsov, Dzerjinskii & Prokofiev, 2005) and could be influenced by effects of dietary preferences ( Wägele & Klussmann-Kolb, 2005) upon particular food items ( Zamora-Silva & Malaquias, 2016).

The Atlantic radula-bearing species of ‘ Odontoglaja ’ is sister to the new IWP radula-less genus Biuve , whereas both IWP species of Odontoglaja clustered together elsewhere in the trees ( Fig. 2A, B). Odontoglaja proper with its type species O. guamensis includes two formally described species (the latter and O. mosaica ), but our COI phylogeny revealed a third molecular lineage (Supporting Information, Fig. S2) highlighting a potential higher diversity in the IWP.

The Atlantic radula-bearing species is here ascribed to the new genus Mannesia (with type species M. sabadiega from the Canary Islands; Table 4). Morphologically, Mannesia differs from Odontoglaja in the shape of the body (both have elongated bodies, but Odontoglaja is more tubular and slimmer than Mannesia ), in the surface of the body (numerous tubercles are present only in Odontoglaja ; Rudman, 1978; Gosliner, 2011; Fig. 3E, K) and in the cephalic shield [rounded anteriorly and not overlapping posteriorly in Odontoglaja ( Fig. 3K), whereas in Mannesia , it is bilobed with quadrangular lobes and overlaps posteriorly ( Fig. 3E)] ( Table 5).

Anatomically, both genera differ in the number of radular rows (<10 in Mannesia ;> 20 in Odontoglaja ) and on the shell, which is spoon-shaped with a conspicuous, broad and extended whorl in Odontoglaja ( Fig. 5H), whereas Mannesia has a spoon-shaped shell with prolonged and semi-wide open whorl ( Figs 5G, 6D). Therefore, in order to reflect the present phylogeny, we introduce the new genus name Mannesia (see Systematic results section for diagnosis).