3D-Analysis of a non-planispiral ammonoid from the Hunsrück Slate: natural or pathological variation?

Position of the epicoles and its relationship with non-planispiral coiling

In the model, the first whorl lies on top of the second whorl. The median plane of the first whorl seems to lie between 1 and 2 mm above that of the second whorl in the model, which must have been even greater before compaction (see discussion). The 3D-model (see Figs. 3 and 4) therefore substantiates the previously hypothesized suspicion (De Baets et al., 2013b) that the specimen is not entirely coiled planispirally. Five clusters (A-E) of colonial sclerobionts can be recognized in the 3D-model (see arrows in Figs. 3 and 4).

At least three clusters (C–E) can be recognized on the phragmocone. Additional clusters (A, B) can be found on the final demi-whorl. In the inner whorls, the sclerobionts are located ventrally on both sides of the whorl cross-section (see Figs. 3C–3E, 4C–4E). Their direction of growth and budding follows the spiral axis of the ammonoid shell. The earliest recognizable sclerobionts with respect to the growth direction of the ammonoid (clusters D–E) coincide with the position where non-planispirality can be first recognized (Fig. 5).

More importantly, there is evidence that clusters (C, D and E) are growing on both sides of the ammonoid (Fig. 6). The last demi-whorl of the host ammonoid was not infilled with pyrite, therefore the growth patterns of the sclerobionts in clusters A and B cannot be established with certainty.

Elongated components like the dacryoconarids present in the substrate (marked in yellow in Fig. 7) along with the I. opitzi specimen are often orientated along the direction of the paleo-current (Hladil, Čejchan & Beroušek, 1991; Hladil et al., 2014; Gügel et al., 2017). Neither the dacryoconarids nor the epicoles show a preferential orientation with respect to the substrate. Sclerobiont clusters C, D, and E do show a preferential orientation with respect to the spiral axis of the ammonoid shell. Unnamed, small, and bulky components visible in the matrix are probably pryitic nodules of different sizes.

Morphology of the runner-like sclerobionts

The mode of preservation of the I. opitzi specimen, pyritization and internal mould preservation, makes examination of fine details or microstructure of the sclerobionts impossible. Yet the microCT permits three dimensional examination of the branching patterns of the sclerobiont colonies. The original settlement location of the colonial organism cannot be distinguished, but the branching pattern of colonies grow in a direction largely parallel to the direction of the aperture of the host.

The colonies in all clusters have the same taxonomic affinity. Zooids in the colonies are long and tubular, and curve slightly near the zooecial opening. Zooids are larger than those commonly observed in bryozoans and bud in alternating directions. The diameter of the more 3D-preserved tubes is typically around 1.5 mm, but this might have been artificially augmented by compaction (De Baets et al., 2013b). The tubes widen in the direction of growth expand distally and slightly contracted at the apex giving them a club-like appearance, characters typical of hederelloids (Elias, 1944).

Branching morphology in the established colonies on the host are diverse despite many shared characteristics. Morphologies of zooids exhibited in clusters A through D are generally more elongate and acutely curved away from the uniserial plane to those in cluster E which contain zooids that curve more dramatically along multiple, pluriserial branches of the colony. Some of this variation could have been amplified by differential compaction in shales (Ross, 1978; Briggs & Williams, 1981).

Synvivo vs. post-mortem encrustation

Three possible scenarios can explain encrustation on both sides of the ammonoid by a sclerobiont colony, clearly visible in clusters C, D and E:

(1) In vivo encrustation: encrustation on both sides of the ammonoid by sclerobiont colonies C, D, and E and the direction of the growth of those sclerobionts matching the growth of the ammonoid, speak for an in vivo encrustation—at least for sclerobiont clusters C–E. A lack of a similar directional growth for all sclerobiont colonies on the ammonoid and with dacryoconarids in the surrounding substrate, as well as a lack in widespread encrustation across the ammonoid shell, further support the likelihood that the ammonoid was encrusted before the ammonoid settled on the seafloor (see Fig. 8).

(2) Post-mortem drift: encrustation of such large colonies solely during post-mortem drift seems unlikely due to the small size of our specimen (105 mm diameter). This is below the 200 mm limit listed for long floating cephalopod shells by Wani et al. (2005) and Rakociński (2011). Interestingly, our specimen would have even a much smaller phragmocone volume than coiled ammonoids at equal diameters used in these experiments. More importantly, the extra weight of the sclerobionts should have made it sink even sooner than a non-encrusted shells so that it would not have resulted in encrustation on both sides. Post-mortem drift is deemed rare for ammonoids in general; most are implied to sink rather rapidly (Maeda & Seilacher, 1996). Furthermore, its vertical position in the water column should have been affected by asymmetric encrustation resulting in a non-vertical position of the shell after loss of soft-part parts which should have led to a more asymmetrical distribution of the sclerobionts as they preferentially attach to the lower part of drifting shells (Donovan, 1989).

(3) Resedimentation or reelaboration/reworking on the seafloor: usually encrustation of the sediment-free side of the ammonoids which mostly end up horizontal on the seafloor—is taken to be characteristic unless reelaboration/reworking of shells or internal moulds happened (Macchioni, 2000). In our cases, we have encrustation on both sides by the same colonies and transport by currents and reworking seems unlikely due the completeness of our specimen and similarities in preservation with other specimens of its taphonomic group (De Baets et al., 2013b). All their characteristics speak for a relatively rapid burial and infilling with pyrite during early diagenesis (before shell dissolution). Furthermore, the same sclerobionts (C–E) started growing on the venter to both sides of the lateral sides which speak rather for a simple encrustation history rather than multiple generations of epibionts encrusting each sediment-free side sequentially (at least one before and after resedimentation/reworking and possibly more). No evidence for resedimentation or reworking (abrasional features) is present in any of the 82 studied specimens of Ivoites.

Pathological variation in the morphology of I. opitzi

The non-planispiral coiling of this specimen, a unique occurrence in the species, occurs at the position of the sclerobiont clusters D and E. This pathological variation in host growth is also consistent with encrustation of the cephalopod during its life-time. Even if the deviation in coiling plane is only 1–2 mm now, it might have been substantially greater before burial and compaction. Whorl thickness alone has estimated to reduce up to 248% in some specimens (De Baets et al., 2013b), so originally these deviations could have been up at least 2.5–5 mm. If we artificially increase the thickness to 200%, we obtain a maximum deviation of the median plane of one whorl to the next of about 2 mm (2.5 mm if we augment the thickness to 250%). The same deviations can be observed within a single whorl (Fig. 9). Considering that the specimen has been extremely flattened (one whorl has been pressed on the other). At least another whorl thickness, the differences in whorl thickness between this whorl and the next which would make it a total of about 2 times this amount, ca. 4 mm (5 mm is we artificially augment the whorl thickness to 250%). This is a rather conservative estimate as we cannot know exactly how the specimen was compressed between the whorls.

Sclerobionts in clusters D and E are in a position that would have been hard to encrust if the subsequent whorl had already grown at the time of encrustation (e.g., when it was lying on the seafloor, the space between whorls around the venter would only have been between around 5 mm, complicating sclerobiont settling. Our specimen suggests that the second whorl lies on top of the first whorl (e.g., in the direction of the more heavily encrusted side, see Fig. 3) while the last whorl probably lies again below the second whorl (e.g., the coiling had almost normalized until the next encrustation by sclerobiont cluster C). The encrustation and its slightly different weight distribution across the venter would predict a deviation of the whorl initially in the direction of the encrusters weight and subsequently in the opposite direction (this seems to be still visible in our specimen despite it been heavily flattened). Such a pattern would thus be expected if the specimen was encrusted in vivo for the first time slightly before it completed about one whorl (Checa, Okamoto & Keupp, 2002). The non-planispiral coiling observed in this specimen could not be produced by flattening or deformation—not even if the specimen was embedded obliquely (see discussion in Material and Methods). Furthermore, clusters D–E are positioned closely before the position where deviation of planispiral coiling can first be recognized and indicate that the coiling was induced by the sclerobiont encrustation. If these clusters grew on the ammonoid during life, as is also suggested by growth on both sides of the shell, the deviations from planispiral coiling would have been unavoidable if they are encrusted heavier on one side than on the other (Checa, Okamoto & Keupp, 2002).

The fact that sclerobiont clusters C, D, and E of runner-like epicoles are growing on both sides of the ammonoid and that clusters (D–E) occur slightly before the position where the deviations from planispiral coiling—where a whorl lies directly on top of the subsequent whorl—can be first recognized are also consistent with an encrustation of the ammonoid during its lifetime.

This specimen, therefore, documents the oldest direct evidence for in vivo encrustation of ammonoids. The previous record holders were Paranarcestes, Latanarcestes and Sellanarcestes from the Upper Emsian interpreted to be encrusted with auloporid corals during their lifetime, also evidenced by deviations from normal planispiral coiling in the host and by the subsequent growth of the ammonoid over the coral (Klug & Korn, 2001; De Baets, Keupp & Klug, 2015).

Identity of the encrusters

Hederelloids are a problematic group of runner-like sclerobionts, which occur in the fossil record from the Silurian through Permian and are most diverse in the Devonian (Solle, 1952; Solle, 1968; Taylor & Wilson, 2007). Hederelloids have traditionally been treated as cyclostome bryozoans (Bassler, 1939; Elias, 1944; Solle, 1952; Solle, 1968; Dzik, 1981), but were redefined based upon differences in branching patterns, skeletal microstructure, lack of an astogenetic gradient, and wide range in tube diameters (Bancroft, 1986; Wilson & Taylor, 2001; Taylor & Wilson, 2007). They are currently mostly interpreted to be closely related to phoronids (Taylor & Wilson, 2007; Taylor, Vinn & Wilson, 2010; Frey et al., 2014). Both auloporid corals and hederelloids are uncommon in the middle Kaub Formation as their life habit requires a solid substrate upon which to settle; these were rare within the clay environments of the Hunsrück Slate (Bartels, Briggs & Brassel, 1998).

The nature of the pyritic preservation of the host specimen, I. opitzi, does not make it possible to look at the microstructure or fine details of the sclerobionts, but the general morphology supports that all clusters of sclerobionts share a taxonomic affinity. Runner-like sclerobionts common in the Devonian include auloporid coral, paleotubuliporid bryozoans, and hederelloids. The sclerobionts encrusting I. opitzi are colonial, with elongated zooids, lateral branching/budding patterns, and relatively large tube diameters which speak for their identification as hederelloids rather than auloporids or bryozoans (Elias, 1944; Bancroft, 1986; Wilson & Taylor, 2006; Taylor & Wilson, 2007).

The runner-like encrusters were initially thought to be auloporid corals (De Baets et al., 2013b), which are known to encrust brachiopods (Zapalski, 2005; Mistiaen et al., 2012) and ammonoids (Klug & Korn, 2001) during their lifetime. Some auloporid taxa have been confused with hederelloids in the past (Fenton & Fenton, 1937; Elias, 1944) and can be hard to differentiate when taphonomic conditions have degraded the quality of the specimen due to gross morphological similarities.

Hederelloids have been reported to encrust externally shelled cephalopods before (Thayer, 1974; Brassel, 1977; Bartels, Briggs & Brassel, 1998; Frey et al., 2014), but these are, to our knowledge, the first reported to encrust an ammonoid in vivo. Sclerobionts can provide also important information on paleoecology, sedimentary environments and taphonomy, both when they encrust shells in vivo or after death of their host (Baird, Brett & Frey, 1989; Kacha & Šaric, 2009; Rakociński, 2011; Brett et al., 2012; Wilson & Taylor, 2013; Luci & Cichowolski, 2014; Wyse Jackson, Key & Coakley, 2014; Luci, Cichowolski & Aguirre-Urreta, 2016).

The presence of five distinct clusters of hederelloids suggests that colonization of the host by sclerobionts happened numerous times (see Figs. 4 and 6). Not all colonizations show clear evidence of in vivo encrustation; clusters A and B were likely the last to settle on the specimen—based upon their location on the shell and the ontogeny of the host—but also do not show evidence of post-mortem encrustation. Clusters A and B, however, are the largest, which suggest that they had sufficient nutrients to settle, establish, and grow for an extended period of time. This is unlikely to have occurred in the benthos after the death of the I. opitzi specimen, but is not impossible. However, because the size of the hederelloid colonies in clusters A and B exceed that of the other colonies, clusters C, D and E were likely not living concurrently with clusters A and B.

Implications for paleoenvironment of Ivoites

Some authors have attributed the rarity of encrusters in coiled ammonoids compared with Paleozoic (orthoconic) nautiloids to their ability to (keep) clean or use chemical defenses against encrusters (Donovan, 1989; Davis, Klofak & Landman, 1999; Keupp, 2012). Potentially, early ammonoid were more similar to their orthoconic nautiloid and bactritoid relatives. Furthermore, mechanical removal of epizoans might have proven more difficult in loosely coiled ammonoids like Ivoites as the previous whorl might have been out of reach of the cephalopod-arm complex (see Fig. 8). However, these suggestions are all quite; it is impossible to find direct evidence for chemical or mechanical cleaning or defensive behavior in this fossil group. More importantly, these mechanisms are also unnecessary to be invoked for such differences which could equally be explained by the fact that many nautiloids would have lived in more shallow environments—which show higher encrustation rates in general (Brett et al., 2012; Smrecak & Brett, 2014; Smrecak, 2016). Modern Nautilus shows differences in encrustations between different environments too (Landman et al., 1987).

The sedimentary environment of the Hunsrück Slate have been widely interpreted, ranging from shallow sedimentation on tidal flats, at depths from around storm-wave base to significantly deeper (Solle, 1950; Seilacher & Hemleben, 1966; Erben, 1994; Bartels, Briggs & Brassel, 1998). The sedimentary environment in the central Hunsrück Basin is now thought to be rather complex; depressions between sedimentary fans provided environments below storm wave base and sills which could locally extend into the intertidal zone (Etter, 2002). The beds with the exceptionally preserved fossils and hemipelagic fauna (ammonoids, dacryoconarids) in the Bundenbach-Gemünden area have been interpreted to be mostly deposited just below storm-wave base to depths around 100 m (Sutcliffe, Tibbs & Briggs, 2002; Stets & Schäfer, 2009). Even in these regions, more sandy layers with neritic brachiopods are intercalated, suggesting occasional shallower depths above storm wave base. It is now generally accepted that the maximum depth was within the photic zone due to the presence in some layers of receptaculitid algae, which are interpreted to be closely related to green algae, and good visual capabilities of arthropods (Bartels, Briggs & Brassel, 1998; Etter, 2002; Selden & Nudds, 2012). Based on the latter, maximum depths of 200 m have suggested (Rust et al., 2016).

Hederelloids are typical encrusters found in Devonian photic zone environments (Brett et al., 2012). They are most diverse and abundant in shallower facies, but persist into the deep euphotic zone (Smrecak, 2016). The presence of hederelloids on the specimen, in combination with other evidence described earlier, support an interpretation of in vivo encrustation. The pelagic life habit of I. opitzi would allow colonies of hederelloids to settle and grow on the shell with some success. In contrast, at depths of 100+ meters, hederelloid encrustation, and sclerobiont encrustation in general, is significantly less common (e.g., Brett et al., 2011; Brett et al., 2012; Smrecak & Brett, 2014). Thus, presence of multiple colonies of hederelloids on the specimen support in vivo encrustation at depths within the photic zone, and lends further support for depositional conditions in line with current interpretations that the Hunsrück Slate interpreted was deposited near storm-wave base (Bartels, Briggs & Brassel, 1998; Sutcliffe, Tibbs & Briggs, 2002).

De Baets et al. (2013b) found encrusted ammonoid shells to be rare (only 6 of 342 studied ammonoids: about 2%). Only two specimens of 82 specimens of Ivoites (2%) were found to be encrusted with hederelloids; our specimen and an additional specimen of Ivoites sp. which was interpreted to be encrusted post-mortem (Bartels, Briggs & Brassel, 1998). Taphonomic or collection biases (Wyse Jackson, Key & Coakley, 2014) are unlikely to explain the low percentages of encrustation as the ammonites and epibionts are principally preserved in the way (e.g., pyritic compound moulds in our case). So far, bivalves, brachiopods, bryozoans, crinoids, hederelloids and tabulate corals have been reported to encrust conchs of ammonoids or other externally shelled cephalopods from the Hunsrück Slate s.s. or middle Kaub Formation (Brassel, 1977; Bartels, Briggs & Brassel, 1998; Jahnke & Bartels, 2000; Kühl et al., 2012a; De Baets et al., 2013b), but these have mostly thought to have happened post-mortem due to their heavy encrustation on one side of the fossils or the encrustation of the structures which would normally be covered with soft-parts (Bartels, Briggs & Brassel, 1998; Jahnke & Bartels, 2000; De Baets et al., 2013b). Heavy encrustations seem to be more common in nautiloids (Bartels, Briggs & Brassel, 1998; Jahnke & Bartels, 2000; Kühl et al., 2012a), but these have so far only been qualitatively studied. In environments between 100 and 200 m encrustation is generally low, which also consistent with a rare in vivo encrustation of our specimen which likely swam in shallower depths when the first encrusters settled rather than the commonly reported post-mortem encrustation. Additional studies on epicoles on ammonoid shells and other shells from the Hunsrück Slate would be necessary to further corroborate these hypotheses.

Implication for mode of life

Loosely coiled early ammonoids are mostly treated as poor swimmers based on their poor streamlining with high drag (Westermann, 1996; Klug & Korn, 2004; Klug et al., 2015a); additional limitations imposed by epizoa on streamlining and shell orientation might be (even) less important in these forms than in normally coiled ammonoids. The fact that our specimen survived at least three separate encrustations—as evidenced by different settlement locations of the clusters of hederelloids—and growth deformations associated with earlier encrustation further corroborates this idea, although further investigations on additional specimens would be necessary to confirm this hypothesis. Most hederelloid colonies generally grow along the spiral direction and do not cross from one whorl to the next, which provide additional evidence that they encrusted the ammonoid during its lifetime.

We cannot entirely rule out a post-mortem encrustation of clusters A–B. Hederelloid growth in those clusters is preferentially orientated away from the aperture of the ammonoid conch, as opposed to those of the inner whorls (cluster C–E) which are preferentially orientated towards it. Associated dacryoconarids do not show a preferential orientation with respect to the substrate (as would be expected in the case of current alignment: Hladil, Čejchan & Beroušek, 1991) or the hederelloids. This does not necessarily speak against encrustation during the lifetime of the ammonoid by clusters A and B as the terminal uncoiling of the ammonoid is interpreted to have influenced the life orientation from an upturned aperture in the inner whorls to a downturned aperture (see Fig. 8) during the terminal uncoiling at the end of the ontogeny (Klug & Korn, 2004; De Baets et al., 2013b; Klug et al., 2015a). We know the ammonoid specimen reached adulthood because it terminally uncoils, which is typical for many taxa of Anetoceratinae and interpreted as a sign of adulthood (De Baets, Klug & Korn, 2009; De Baets et al., 2013b; De Baets, Klug & Monnet, 2013; Klug et al., 2015b).

Hederella is known to encrust another ammonoid from the Hunsrück Slate, but this is interpreted to have happened post-mortem (Brassel, 1977; Bartels, Briggs & Brassel, 1998). There is no evidence that this happened in vivo in the second specimen as these are located on an incompletely preserved body chamber, and have been interpreted as encrusting the inside of the shell (Bartels, Briggs & Brassel, 1998)—but this should be further tested with µCT. Additional studies would be necessary to confirm if our specimen is an isolated case of in vivo encrustation or part of a more common phenomenon.

Implications for taxonomy

Defining pathological specimens as species can have important taxonomic implications (Spath, 1945). According to Spath (1945), such ammonoid species should remain valid, but no new type can be chosen while the holotype is still in existence. Others, like Haas (1946), claim assigning a pathological specimen might undermine the status of the species, considering the morphology of the species is described based on a pathological specimen, and that a new neotype should be selected. It is important to note that several authors have erected different ammonoid species based on small differences in coiling. As the only known non-planispirally coiled specimen of I. opitzi is likely pathological, an author like Haas (1946) might have suggested to designate another specimen as type for this species to avoid ambiguity. However, our study highlights that non-planispirality does not belong to normal intraspecific variation of this species, nor to the taxonomic definition of this taxon. Paratype SMF-HF 940, which was collected from the same locality as the holotype (De Baets et al., 2013b), would be the best candidate among the paratypes. Reassigning a neotype has recently been intensively discussed for the holotype of Homo floresiensis (Kaifu et al., 2009; Eckhardt & Henneberg, 2010), but the severity of deviation has to be considered in this specimen. However, both aspects (electing a neotype for pathological specimens and its dependence of the severity of the pathology) are not specifically discussed in the International Code of Zoological Nomenclature (ICZN). According to Article 75.1 of ICZN, “the neotype is …designated under conditions …when no name-bearing type specimen is believed to be extant …”. In this case, if the holotype, even when pathologically deformed, is extant, the proposal of neotype is not granted. A proposal could be submitted to the ICZN to resolve the use of pathologically-induced morphological variation of holotypes, but this falls outside the aim of our study. Furthermore, we are confident that the original type specimen belongs to same species as the paratypes as it completes the same amount of whorl before uncoiling, has a similar rib spacing and only differs from other specimens in its minor coiling deviations (De Baets et al., 2013b). Such minor coiling deviations are not considered sufficient to erect a new species by us and other authors - even if these would not be of pathological nature (Dietl, 1978; De Baets et al., 2013b; De Baets et al., 2015). As non-planispiral coiling was also not part of the original diagnosis, we feel it is unnecessary to submit a proposal to appoint a new type specimen, which might not be allowed anyway.