New evidence of functional suckers in belemnoid coleoids (Cephalopoda) weakens support for the ‘Neocoleoidea’ concept

The idea of a sister-group relationship between the cephalopod superorders Decabrachia (squids and cuttleﬁshes) and Vampyropoda (octopods, cirroctopods and vampire squid) has been widely accepted since Jeletzky (1966: fig. 2). In their phy- logenetic analyses, Berthold & Engeser (1987) and Engeser & Bandel (1988) regarded the presence of suckers as a character state that supports the monophyly of what they called the ‘sucker-bearing Coleoidea’. In 1997, Haas introduced the terms ‘Neocoleoidea’ for coleoid groups with living representa- tives and ‘Palaeocoleoidea’ for a group of hook-bearing coleoids also known as ‘Belemnoidea’, which was thought to have disappeared at the end of the Cretaceous. Like Berthold & Engeser (1987) and Engeser & Bandel (1988), Haas (1989: text-ﬁg. 114, ﬁg. 1) detected only one character state uniting the ‘Neocoleoidea’, the ‘presence of suckers’. Although von Boletzky (1992) found strong arguments against the monophy- letic origin of the group later called ‘Neocoleoidea’, Young, Vecchione & Donovan (1998: ﬁg. 1) followed Haas (1997) and adopted the ‘Neocoleoidea’ concept. Phylogenetic analyses including morphological and/or molecular

The idea of a sister-group relationship between the cephalopod superorders Decabrachia (squids and cuttlefishes) and Vampyropoda (octopods, cirroctopods and vampire squid) has been widely accepted since Jeletzky (1966: fig. 2). In their phylogenetic analyses, Berthold & Engeser (1987) and Engeser & Bandel (1988) regarded the presence of suckers as a character state that supports the monophyly of what they called the 'sucker-bearing Coleoidea'. In 1997, Haas introduced the terms 'Neocoleoidea' for coleoid groups with living representatives and 'Palaeocoleoidea' for a group of hook-bearing coleoids also known as 'Belemnoidea', which was thought to have disappeared at the end of the Cretaceous.
Thus, the 'Neocoleoidea' concept is chiefly based on the presumed absence of suckers in belemnoids, although equivocal evidence for their presence has existed for some time. Mantell (1852: 19, fig. 4) probably observed suckers in Belemnotheutis antiquus Pearce, 1847, a belemnoid from the middle Jurassic (Callovian; 164 Ma) Oxford Clay of Christian Malford (UK). Engeser & Clarke (1988: 139) suggested that these structures instead represented muscles that moved the hooks and that they were unlikely to be functional suckers. In their redescription of the type material of B. antiquus, Donovan & Crane (1992: 286), however, recorded sucker-like structures in a second specimen from the same area and described them as follows: "Where suckers are preserved there are the same number of pairs of hooks as suckers, but it is not clear whether the hooks are rooted in the suckers, as they sometimes appear to be (Pl. 2 fig. 2). There could have been only one row of suckers per arm. Distal suckers are smaller than proximal ones, the largest being about 2 mm in diameter. Several suckers show what could have been a chitinous ring around the outside (Pl. 2 fig. 4)". Their observations led them to conclude that ". . .the presence of suckers in Belemnotheutis shows that suckers are not autapomorphic for Dibranchiata sensu Berthold and Engeser. . ." and that ". . .it is likely that suckers were already present in the common ancestors of the Belemnitida and the living forms". Both Young et al. (1998: 395) and Fuchs (2006: 71, 77, fig. 3.7 -3) later expressed doubts about the existence of functional suckers in belemnoids. We report here a new record of a belemnoid arm crown with exceptionally preserved soft parts including apparent suckers, which again question the monophyly of the 'Neocoleoidea'.
The isolated arm crown comes from the late Jurassic (Tithonian; 150 Ma) limestones of Eichsta¨tt (southern Germany). The specimen (Coll. H. Tischlinger, catalogue no. 98/3) is flattened and consists of slab and counter slab. Most probably, the arm crown belongs to the genus Acanthoteuthis Wagner in Mu¨nster, 1839, a common belemnoid genus in the Eichsta¨tt limestones and close relative of Belemnotheutis Pearce, 1842. It consists of 10 equal arms (maximum arm length 12 cm), each of which shows well-preserved imprints of hooks and suckers (Fig. 1A-F). Observations under UV light indicate that the proximal arm musculature is preserved as calcium phosphate and that five arms cover the remaining arms (Fig. 1B). The plane in which the arm crown has been embedded is unclear; identification of ventral and dorsal arms is therefore impossible. Although distal arm musculature is not preserved, the position of each arm can be reconstructed through the arrangement of hook and sucker imprints (Fig. 1C).
Hook imprints, which are generally filled with calcite, suggest a biserial arrangement as is typical for belemnoids ( Fig. 1D-F). Only few hooks are visible as brownish staining. The shape of the paired hooks is variable along each arm as described by Engeser & Reitner (1992) for Acanthoteuthis from the Eichsta¨tt region.
Apart from the hooks, each arm is equipped with one series (¼ longitudinal row) of circular suckers (because the arm crown consists of part and counterpart, upper as well as lower arms left sucker imprints). The suckers appear very close to the arm bases. The sucker diameter increases gradually from proximal to one half the arm length and then decreases distally. The largest sucker imprints are 5 mm in diameter, the smallest 2 mm. We counted a maximum number of 19 suckers per arm. The best preserved sucker imprints exhibit another circular imprint in their centre (Fig. 1F). This inner circle can be interpreted as the opening of a suction chamber of the acetabulum; this would strongly suggest that this is, indeed, a sucker. If so, the periphery would correspond to the imprints of the infundibulum. It is important to note that there is no evidence of horny rings as are typical for decabrachian suckers. The present suckers seem to be identical to those illustrated and described in Donovan & Crane (1992: pl. 2) but, in contrast to their specimen, our new record unambiguously shows that the hooks are not rooted within the suckers (Fig. 1D-F). Instead, the paired hooks clearly alternate with the unpaired suckers. In particular, the latter observation demonstrates the The hooks (h) on the right are seen in lateral view, while corresponding hooks (h 0 ) on the left just exhibit their bases. Scale bar ¼ 3 mm. F. Detail of a circular sucker imprint from the proximal part of arm 9. The inner circle might be imprinted by the acetabulum (a); the periphery by the infundibulum (i). The associated hook imprint (h) is filled with calcite. Scale bar ¼ 1 mm. morphological and functional independence of suckers and hooks. The present specimen therefore confirms earlier reports of Mantell (1852) and Donovan & Crane (1992) about presumed suckers in belemnoids and further indicates the weakness of the 'Neocoleoidea' concept.
Moreover, our observation contributes to the still unsolved discussion about the homology of suckers, hooks, cirri and trabeculae. Naef (1921Naef ( -1923 regarded belemnoid hooks as homologous with decabrachian hooks. However, as several authors later explained, belemnoid hooks cannot result from secondary modification of a sucker ring as observed in decabrachian hooks (Engeser & Bandel, 1988: 113;Engeser & Clarke, 1988: 135;Young et al., 1998: 395). The apparent absence of horny rings in the Acanthoteuthis suckers supports this assumption. Haas (1989: 182) tried to correlate the arrangement of belemnoid hooks with the arrangement of suckers in vampyropods and decabrachians. However, the arm crown described above clearly shows that the arrangement of hooks and suckers cannot be correlated, because these characters represent two independent morphological and functional units. Haas (1989: 183) concluded further that belemnoid hooks and cirri cannot be homologues. In contrast, Engeser & Bandel (1988: 113) derived belemnoid hooks from cirri and also Young et al. (1998: 395) remarked that ". . .belemnoid hooks could be homologous with cirri. . .". Engeser & Clarke (1988: 135, fig. 1) suggested that most of the hook shaft was covered with soft tissue and that only the terminal claw was naked. Thus, the reduction of hooks without the reduction in the associated soft tissues would have produced flap-like innervated structures very similar to cirri or trabeculae.
Our specimen clearly supports the ideas of Engeser & Bandel (1988), Engeser & Clarke (1988) and Young et al. (1998). Because vampyropod cirri are thought to be homologous with decabrachian trabeculae (Young & Vecchione, 1996: 98;Young et al., 1998: 395), belemnoid hooks may be homologous with both cirri and trabeculae. Each of the three coleoid superorders ('Belemnoidea', Vampyropoda and Decabrachia) exhibits pairs of supposedly tactile organs in close association with suckers. This conclusion is clearly in accordance with Naef's hypothetic 'Protodibranchus'. Naef (1921Naef ( -1923 fig. 23) postulated one series of suckers with a series of 'small, movable processes or palps' on each side of the suckers to be the ancestral character state of the Coleoidea.