Different regeneration mechanisms in the rostra of aulacocerids (Coleoidea) and their phylogenetic implications Helmut Keupp1 * & Dirk Fuchs2 1Freie Universität Berlin, Institut für Geologische Wissenschaften (Fachrichtung Paläontologie), Malteser-Str. 74-100, Haus D, 12249 Berlin, Germany; Email: [email protected] 2Freie Universität Berlin, Institut für Geologische Wissenschaften (Fachrichtung Paläontologie), Malteser-Str. 74-100, Haus D, 12249 Berlin, Germany; Email: [email protected] * corresponding author 77: 13-20, 4 figs., 1 table 2014 Shells with growth anomalies are known to be important tools to reconstruct shell formation mechanisms. Regenerated rostra of aulacocerid coleoids from the Triassic of Timor (Indonesia) are used to highlight their formation and regener- ation mechanisms. Within Aulacocerida, rostra with a strongly ribbed, a finely ribbed, and a smooth surface are distin- guishable. In the ribbed rostrum type, former injuries are continuously discernible on the outer surface. This observa- tion implicates that this type thickens rib-by-rib. In contrast, in the smooth rostrum type, injuries are rapidly covered by subsequent deposition of concentric rostrum layers. This fundamentally different rostrum formation impacts the assumed monophyly of Aulacocerida. According to this phylogenetic scenario, either the ribbed rostrum type (and consequently also the strongly folded shell sac that sec- rets the rostrum) of Hematitida and Aulacocerida or the smooth rostrum type of xiphoteuthidid aulacocerids and bel- emnitids evolved twice. Assumption of a paraphyletic origin of aulacocerid coleoids might resolve this conflict since the ribbed rostrum could have been inherited from Hematitida and gradually smoothened within the lineage Aulacocerati- dae—Dictyoconitidae—Xiphoteuthididae—Proostracophora. Received: 21 December 2012 Subject Areas: Palaeontology, Zoology Accepted: 01 August 2013 Keywords: Coleoidea, Aulacocerida, Triassic, Indonesia, Timor, palaeopathology, rostrum formation, regener- ation, phylogeny Introduction Palaeopathology usually involves individual growth irregu- modification boundaries are often determinable only after larities of mainly mineralised hard parts. Causes for these comparisons with anomalies. Such information is fur- anomalies can be endogene (genetical, pathological, para- thermore very useful for an objective evaluation whether sitic) or exogene (abiotic or biotic interactions like injuries characters possess a taxonomic/systematic significance or or epibionts) factors. Palaeopathological phenomena are not. therefore adequate sources for aut- and synecological in- terpretations. Essential aspects for autecology are forma- tion mechanisms und functionality of characters, whose 14 Helmut Keupp & Dirk Fuchs Table 1: Age, number of studied specimens, number of anomalies, family assignment, and rostrum morphology of examined aulacocerids. The variety of possible interpretations of palaeopathologi- General morphology, stratigraphy cal phenomena in ectocochleate Cephalopoda have been and systematic position of Aulaco- recently summarised by Keupp (2012). As a result of their different bauplans, regeneration mechanisms of shells or cerida within Coleoidea shell parts considerably differ in ecto- and endocochleate cephalopods. In the former, the regenerative mantle epi- The order Aulacocerida Stolley, 1919 is typified by a longi- thelium is located inside the shell, whereas in the latter it is conic phragmocone, which is covered from outside by a located both inside and outside. Within the endocochleate thin primordial rostrum and a solid rostrum proper Coleoidea, whose shell is fully covered by regenerative ep- (Jeletzky 1966; Bandel 1985). In contrast to Belemnitida, ithelium (the so-called shell sac), different healing modifi- both rostral layers have been described as originally arago- cations can additionally indicate different soft part con- nitic (Cuif & Dauphin 1979; Bandel & Spaeth 1988). More structions, which itself might bear a systematic relevance. aulacocerid key characters are a tubular body chamber (or A good example for the systematic significance of shell final chamber) without evidence of a forward projecting anomalies is given for the case of aulacocerid rostra from proostracum typical of belemnitids and obviously widely the Middle and Upper Triassic of Timor (Indonesia). spaced (i.e., remarkably long) chambers. The initial cham- ber (protoconch) is spherical and completely sealed by a closing membrane (Jeletzky 1966; Bandel 1985). The pri- Material mary phragmocone wall (conotheca) is 4-layered (a thin inner prismatic layer, a thick middle layer of tabular nacre, During a field campaign in summer 2007 to the river sys- a thin outer prismatic layer, an outermost organic layer). tem of the Oë Bihati near Baun (SW Timor), the senior The marginal siphuncle is located ventrally. Aulacocerid author, together with W. Weitschat (Hamburg) and R. Ve- soft parts are unknown. it (Velden), collected numerous aulacocerid rostra from Unambiguous aulacocerids shells are recorded from Middle and Upper Triassic (and Lower Jurassic) erratic the Early Triassic to the Late Jurassic. Some shells tenta- boulders (see also Keupp 2009). tively interpreted as belonging to aulacocerids have been Three rostra (= 1.3 %) among in total 236 studied ros- described also from Permian (e.g., Gordon 1966) and tra of Aulacoceras sulcatum showed regenerated shell injuries Carboniferous deposits (Flower & Gordon 1959; Jeletzky (Table 1). Within the few records of dictyoconitid aulaco- 1966; Doyle 1990; Doguzhaeva et al. 2010). Owing to cerids there was no evidence of anomalies. Similarly, 111 slight differences in shell morphologies, some of these atractid rostra from the Ladinian and Carnian/Norian lack presumed Palaeozoic aulacocerids have been later re- any evidence of shell repairs. However, a single specimen moved from Aulacocerida and placed within Hematitida collected during an earlier field trip at Bekal–Nassi (Ti- (Doguzhaeva et al. 2002). Despite of this systematic sepa- mor) was available for the present study. The rostrum of ration, Aulacocerida is very similar to Hematitida in hav- Atractites lanceolatus exhibits a knee-like fracture (Mietchen ing a longiconic phragmocone, a tubular body chamber, et al. 2005). and an aragonitic rostrum proper, which is why a reliable The present material is deposited in the collection H. distinction is very difficult and only possible in well pre- Keupp (SHK) at the Freie Universität Berlin. served material. Different regeneration mechanisms in the rostra of aulacocerids (Coleoidea) 15 In the past, aulacocerids have often been considered as Ribbed rostra of Hematites and Aulacoceras consist of radial- hook-less (Engeser 1990). Indeed, hooks (micro-onychi- ly arranged longitudinal ribs or folds. Adjoining ribs are in tes) and aulacocerid shells have never been observed asso- contact (or with very narrow interspaces) and each rib ciated. However, if the abovementioned Palaeozoic cole- shows a lamellar growth pattern (Figs. 1a, g, j). According- oids indeed represent early aulacocerids, the abundant ly, the ribbed rostrum must have been formed in a radially presence of hooks in Carboniferous, Permian and Triassic folded shell sac. Whereas the latter rostra can be described deposits would indicate that aulacocerid arms were like- as strongly ribbed, some aulacocerid forms such as Dicty- wise armed with hooks – as in belemnitids. oconites exhibit very fine-ribbed rostra (Fig. 1b–d, h). On The aulacocerid/hematitid character combination the other hand, smooth rostra – like in belemnitid rostra – “longiconic phragmocone with a tubular body chamber, a consist of lamellar layers, which must have been succes- small spherical protoconch, a marginal siphuncle, and tab- sively formed in an unfolded shell sac (Fig. 1e–f, i, l). ular nacre in the conotheca” suggest a primitive character Hence, the ribbed rostrum type thickens rib-by-rib, state and therefore a close phylogenetic relationship with whereas the smooth type concentrically grows around a ectocochleate Bactritida (Kröger et al. 2011). nucleus. Smooth rostra characterises the aulacocerid fami- Traditionally, Aulacocerida (and Hematitida) has been ly Xiphoteuthididae (e.g., Atractites). grouped together with Phragmoteuthida (Middle Triassic– Owing to this fundamentally different formation strat- Lower Jurassic), Belemnitida (Upper Triassic–Upper Cre- egy, both rostrum types should have a very high systemat- taceous), and Diplobelida (Upper Jurassic–Upper Creta- ic (and phylogenetic?; see discussion below) value. Both ceous) as Belemnoidea. However, the latter taxon most rostral growth modalities must have additionally a strong probably represents a paraphylum (Fuchs 2006; Fuchs et influence on shell regenerations so that their differences al. 2010; Kröger et al. 2011). should especially become evident in repaired rostra. Phragmoteuthida is characterised by a ventrally open- ed body chamber, which resulted in the development of a weakly mineralised proostracum (Donovan 2006; Dogu- Regeneration patterns after zhaeva & Summesberger 2012). Apart from this, phrag- moteuthids lack a solid rostrum. Belemnitida mainly differ traumatic injuries from Aulacocerida/Hematitida by the possession of a proostracum and a predominantly calcitic rostrum proper. Similar to sepiids, whose cuttlebone is well known to Diplobelida can be easily distinguished from Aulacocerida show shell regenerations caused by unsuccessful predator