Among Mesozoic coleoids (cephalopods with internal hard parts that include octopuses, squid, and cuttlefish), belemnoids represent the most abundant and best documented clade. In the Jurassic, it is mainly the Belemntitida that can regionally be found in rock-forming numbers, but remains of the Diplobelida also occur occasionally. Despite their abundance, direct evidence for the swimming and hunting behaviour of these extinct cephalopods is extremely rare1,2,3,4.

Here, we document four cases of Early Jurassic coleoids, which all hold a small bony fish in their arm crowns (Fig. 1, Suppl. Figs 2, 4). Such preservation requires exceptional taphonomic conditions, for which the Early Jurassic Posidonienschiefer is world-renowned5,6,7,8,9,10. Repeated hypoxic to anoxic bottom water conditions11,12 decelerated decay processes and allowed pyritization or phosphatization of soft tissues that are only rarely fossilized otherwise13,14,15,16. These conditions persisted over a long time and in a rather vast area from southern France via northern Switzerland throughout much of Germany and into Great Britain; therefore, these widely distributed conditions allowed the formation of one of the most famous Konservatlagerstätten8,17, the Posidonienschiefer ( = Posidonia Shale or ‘Schistes cartons’).

Figure 1
figure 1

Left: Photo Clarkeiteuthis conocauda with Leptolepis bronni in its arm crown, as displayed at the Urweltmuseum in Holzmaden, Germany, Toarcian, Kirschmann Quarry. Right: Drawing of the same specimen, combining drawings of slab and counterslab.

From this point of view, it is not surprising that the first rostrum-bearing belemnite preserved with soft-parts was found in the Posidonienschiefer6,18,19. Some other cephalopods also show anatomical details of soft tissues that are usually not preserved (e.g., digestive tract in ammonites6; musculature and gills in non-belemoid coleoids20,21,22). Despite the extreme scarcity of cephalopods preserved with their prey, the specimen presented here was never described in detail and only figured once without detailed discussion23. Recently, a second specimen documenting the same behaviour was published10.

In the Posidonienschiefer, molluscs represent the most common macrofossils; among those, bivalves are very abundant followed by ammonites and belemnites5,6. Other molluscs such as non-belemnoid coleoids and nautilids are rarer and complete specimens preserved with intact arm crowns are very rare. The diplobelids described here have strong arm-hooks and a proostracum and thus belongs to the proostracum-bearing belemnoids (in contrast to, e.g., aulacoceratid belemnoids24). Belemnoid coleoids share a conch with mineralized shell layers including the calcitic, aragonitic or bimineralic rostrum25,26,27,28,29,30, which surrounds posterior parts of the chambered, cone-shaped phragmocone (aragonitic). It has been hypothesized that diplobelids could achieve neutral buoyancy using their gas-filled phragmocone31.

In many respects, the mode of life of belemnoid coleoids is still poorly understood but the differences in anatomy, proportions and spatial distribution of masses of the soft body and internal structures suggest diverse modes of life and habits32. Accordingly, we use the opportunity of the availability of these exceptionally preserved specimens of Clarkeiteuthis conocauda to (1) describe these specimens with their prey, (2) to discuss their taphonomic history and (3) implications for its feeding behaviour. Stimulated by this finding, we further assess (4) differences in the mode of life of C. conocauda and the Jurassic belemnitids Passaloteuthis bisulcata and Hibolithes semisulcatus (Belemnitida; Fig. 2, Suppl. Figs 3, 4) by reconstructing the centres of mass and buoyancy in order to make inferences on their syn vivo orientation in the water column.

Figure 2
figure 2

Passaloteuthis bisulcata, as displayed at the Urweltmuseum in Holzmaden, Germany, Toarcian. The specimen is complete preserving its rostrum, phragmocone (partially ontop of the rostrum), proostracum, phosphatized soft-tissue remains, arms with hooklets and the ink sac at the side of the rostrum. Its somewhat upended appearance can be explained by its taphonomical history, where the carcass sank into the unconsolidated sediment with the rostrum first and the other parts obliquely above it. This corroborates that the rostrum density was high during life already.


Four specimens of the coleoid Clarkeiteuthis conocauda with the teleost fish Leptolepis bronni in their arm crown were available for study, but only one of those is kept in a public museum (Museum Hauff, Holzmaden, Germany) and thus is chosen here for the detailed description. The C. conocauda has a total length of 210 mm. Together with the teleost fish L. bronni, the two animals from the Posidonienschiefer of Holzmaden have a total length of 240 mm including arms (Fig. 1, Suppl. Fig. 7). The total length of coleoids varied during life, because they could stretch and contract the tentacles to some extent; we provide this value because we cannot directly measure the actual mantle length.

Clarkeiteuthis conocauda


The aragonitic phragmocone is around 64 mm long, but only 44 mm of it is preserved since the apex of the phragmocone is missing. Twelve chambers are visible, each between 4 and 5 mm wide. The last chamber measures around 40 mm in diameter. The apical angle of the crushed and flattened phragmocone is 50°, which suggests an originally lower angle prior to compaction of 30–35°21. The middle part of the phragmocone is covered by phosphatized remains of the skinny mantle as well as unidentified black remains.


The originally chitinous proostracum is only partially fossilized, 90 mm long and 5–9 mm wide. Taking the proportions of the animal into account, the proostracum should be 80 up to 110 mm long21. Possibly, only the thicker posterior parts are preserved either due to taphonomic alteration or loss during preparation. Its remains are straight and spatulate, have a mother of pearl-like shine, are slightly translucent and show straight lines along its length. It emerges from the last phragmocone chamber and reaches the top of the ink sac (posterior part of soft body). There is no evidence of a 3-lobed proostracum typical for phragmoteuthid belemnoids, thus confirming the diplobelid affinity21.

Mantle remains

Clarkeiteuthis conocauda shows two different types of preservation of mantle remains. Just anterior to the proostracum, two patches of the muscular mantle are visible. They have an oval shape and show the transverse striation, which is characteristic for Mesozoic coleoids and thus also for those of the Posidonienschiefer6,19,21. Both patches are 26 mm long and 14 mm wide and end just posterior of the jaws. Three patches of thin mantle remains lie on top of the phragmocone. Their striation is less distinct than that of the two patches mentioned first. Two of the posterior patches lie on the sides of the phragmocone and the last one covers a part of the proostracum. Additionally, a black patch lies in the middle of the phragmocone, which likely also represents mantle material33, perhaps pyritized in this case.

Ink sac and duct

The ink sac lies in the middle of the specimen and is partly covered by the proostracum. It has an oval form and is c. 23 mm long and 10 mm wide. From its right border, the 3 mm wide ink duct emerges anterolaterally to the right, then bends to the left and crosses the middle of the specimen for 14 mm. The ink duct is partially covered by mantle remains over 13 mm (dotted line in Fig. 1). A black patch that measures 23 × 20 mm lies at the anterior end of the ink duct; it is partly covered by muscular mantle and might represent leaked ink remains.

Arm hooks

Around 70 arm hooks are visible on the slab and 60 on the counterslab (many are probably identical, i.e. represent fragments of the same hook). Some of the hooks are arranged in rows of two, roughly documenting the dimensions of the arms and the number of arm hooks per arm (≈30, i.e. ca. 300 in total)34,35. The size of the arm hooks is diverse with smaller hooks at the distal ends and proximal bases of the arms and larger hooks in the middle. Their lengths vary from 1 to 5 mm with a base 0.1 to 1 mm wide. At least two different types of micro-onychites are present: Larger hooks (up to 4 mm long), where the uncinus is strongly bent as well as 1 to 5 mm long onychites, where the uncinus is just slightly curved and the shaft is straight21. Their black colour and structure suggest that the hooks are carbonized similar to other coleoids from the Posidonienschiefer35.


The carbonized jaw remains are crescent-shaped and located directly posterior of the arms where the head used to be. They are c. 10 mm wide and 8 mm high. Their preservation is poor and prohibits a detailed interpretation of jaw morphology; anteriorly, however, there is a V-shaped darker structure, which likely is a part of the more strongly sclerotized part of the outer and inner lamella, possibly of the lower jaw34,36. The fact that this structure shows two tips points at the possibility that the jaws are still in situ, and one tip belongs to the lower jaw and the other to the upper jaw.

Leptolepis bronni

The teleost fish is about 120 mm long (when straightened out), flattened and partially fragmented. Its head measures around 34 mm in length and 20 mm in height in its compacted state. The head thus makes up almost 30% of the body length. In the head region, the orbit, jaw and operculum are discernible as well as seven rays from the pectoral fin. Just posterior to the head, eight vertebrae can be seen accompanied by six ribs. The vertebral column shows several fractures and kinks, most of which probably have post mortem-causes (taphonomic) such as compaction. The vertebral column continues after a gap of around 5 mm. The course of the deformed vertebral column is well visible caudally. Remarkably, the parts on both sides of this gap each follow roughly the directions of arms adjacent to it. The most distinct kink in the vertebral column occurs exactly in this gap, and this kink is the closest to the jaw of the diplobelid.

Although the vertebral column is visible, the total number of vertebrae cannot be distinguished due to the strong compactional deformation of the specimen. Both dorsal and anal fin rays are visible near the vertebral column, with fin ray length between 1 and 5 mm. In case of the dorsal fin, the length of the distal rays is smaller than that of the dorsal ones.

Fossilized behaviour

As mentioned above, four specimens of Clarkeiteuthis conocauda from Holzmaden were examined, which are all completely preserved including soft-tissue remains. All four hold a small teleost fish in their arm crowns. To our knowledge, these represent – apart from an octobrachian (Glyphiteuthis libanotica) from the Upper Cretaceous Lebanon Limestones37 – the first and oldest findings of a cephalopod that has been preserved together with its vertebrate prey within its arm crown (stomach contents with vertebrate remains have occasionally been reported from fossil coleoids1,37,38,39,40 and once prey within the buccal mass41). The association of small teleosts within arm crowns of C. conocauda can be considered as evidence for fossilized predatory behaviour, because both species are rather rarely preserved in articulation otherwise, and thus, an accidental association of such fossils is unlikely; more importantly, the same joint occurrence in always the same way in four cases corroborates the interpretation as fossilized behaviour. If currents had transported the fish to the coleoid carcass, the arms would have been bent in current direction and the fish would probably have come to a rest somewhere else next to the diplobelid with the long body axes of both animals aligned in parallel.

The interpretation of the fish as the last prey is further corroborated by behavioural biology studies of recent squids. For example, Illex illecebrosus captures its prey with both head and tail protruding from the squid’s arms42. To immobilize the prey, squids cut the vertebral column of their fish prey by a jaw bite near the head43. In three of the four cases, the fossil fish displays a distinct kink in vertebral column in close proximity to the beak of the diplobelid, thereby documenting this behaviour for Clarkeiteuthis.

A possible explanation as to how these specimens became preserved is the phenomenon of post capture sinking or rising, which has been reported from, e.g., the giant squid Architeuthis dux44,45 and from Illex illecebrosus43. Some pelagic animals are slightly negatively buoyant and start sinking when inactive. Architeuthis has ammonium chloride-solution-filled tissues that neutralize buoyancy46, but if its prey is negatively buoyant or tries to escape downward, the squid gets pulled downward as well. Consequently, if Clarkeiteuthis focused on its prey instead of staying in the same water level, they likely began to sink, or, alternatively, the cephalopod descended on purpose in order to get out of sight of other predators. This downward movement was potentially accelerated by the reduction of the gas volume in the teleost’s swim bladder by the increasing hydrostatic pressure. Thereby, the living coleoids with the probably already dead fish reached the poorly oxygenated bottom waters and suffocated, which is supported by the ongoing embrace of the fish by the diplobelid in all four cases. Post capture sinking is not only corroborated by the completeness of the cephalopod and its prey but also by the fact that sinking can still be seen in recent cephalopod species as well as in other fossilized specimens like mating lacustrine turtles47. Recently, Mapes et al.48 coined the term ‘distraction sinking’ for aquatic organisms that were occupied by activities such as mating, feeding, or fighting; this activity distracted their attention from the onset of a sinking process, whereby they potentially reached deeper water layers poor in oxygen (as in the case of the Clarkeiteuthis described here) or poisonous (as in the case described by Joyce et al.47).

Cephalopods usually pull prey towards them by contracting their arms or tentacles49,50,51. Did Clarkeiteuthis also contract its arms after it had caught the fish? We measured the lengths of the visible arms of 14 specimens and calculated their mean length (Fig. 3). As a proxy for body size we used the distance between the most proximal arm hook and the last formed septum. With measurements of only 14 specimens with only four with a fish in their arm crown, the statistical power is poor. Additionally, arms can be contracted for several reasons besides holding prey; it might occur due to distress, linked with swimming movements or else52,53. In Fig. 3, mean arm length is plotted versus body size. The four specimens holding a fish are among the specimens with the most contracted arms. There is only one specimen with similarly contracted arms, but the reason for the contraction is not evident. It is possible, that the prey is not preserved, not visible or escaped before the squid settled on the sediment but there might be other reasons to explain the contracted arms of this specimen (Nr. 2 in Table 1). In any case, these findings suggest that Clarkeiteuthis indeed contracted its arms to pull its teleost prey towards its mouth to cut the spine and then feed on it.

Figure 3
figure 3

Relative arm lengths in the four specimens of Clarkeiteuthis conocauda with Leptolepis bronni in its arm crown and ten specimens of Jurassic diplobelids without prey. Note that the specimens with fish tend to have rather contracted arms.

Table 1 Lengths of arms (arm numbering is arbitrary), phragmocone and body of published and museum specimens of Clarkeiteuthis.

Remarkably, only one type of prey has been found in the arm crown of Clarkeiteuthis to our knowledge. The specimen presented here, the one published by Klaschka10 as well as the two additional specimens (Suppl. Fig. 2) all hold the small teleost Leptolepis in their arms. This raises the question for prey specificity. With four specimens available, the statistical power is low, but it is remarkable that, for example, no specimen with invertebrate prey is known and that it is always the same fish that has been caught. Articulated fish skeletons are rare in the Posidonienschiefer, but among all fishes, Leptolepis is one of the more common genera. Being one of the smallest, it was one of the very few species Clarkeiteuthis could probably handle without the risk of becoming injured or killed. In contrast to other small contemporary fish such as Tetragonolepis, Leptolepis lacked thick scales. With this reduced armour, it was probably an easier prey to catch and hold for the diplobelids by inserting their hooks in the skin and flesh rather than into those species with thick scales. Late Jurassic relatives (Leptolepides) of the Toarcian Leptolepis occur in large fossilized groups suggesting that they often lived in larger schools54. Thus, Leptolepis was possibly quite abundant and a likely prey but their delicate skeletons became quickly disarticulated. In any case, these joint occurrences suggest an overlapping habitat in the water column of both cephalopod and teleost.

Mode of Life

In order to reconstruct the syn vivo body orientation and the mode of life of Clarkeiteuthis conocauda, different approaches were employed. The first approach included the determination of the centres of buoyancy and mass of Clarkeiteuthis conocauda, Passaloteuthis bisulcata and Hibolithes semisulcatus (these taxa were chosen because of the well known phragmocone and rostrum proportions). In Clarkeiteuthis, the centre of buoyancy lies more or less in the centre of its body, around 32 mm anterior to the edge of the last septum of the phragmocone (Fig. 4, Suppl. Fig. 1). The specimen had its centre of mass in the middle of its body, about 40–41 mm anterior to the phragmocone, which is very close to its centre of buoyancy with around 8–9 mm separating them. The centre of buoyancy of Hibolithes lied around 66 mm anterior to the phragmocone (Fig. 4). In Hibolithes, the centre of mass was situated in the middle of the body, about 68 mm anterior to the phragmocone. This implies that its centres of mass and buoyancy were only about 2 mm apart.

Figure 4
figure 4

Localizations of the centres of mass (blue cross) and buoyancy (black cross) in Clarkeiteuthis conocauda, Passaloteuthis bisulcata and Hibolithes semisulcatus along the body axis in relation to their body parts.

In Passaloteuthis bisulcata, the centre of buoyancy lied 33 mm anterior to the phragmocone, whereas the centre of mass lied 35 mm anterior to it. As in Hibolithes, the centres of mass and buoyancy were situated around 2 mm apart from each other.

These values vary depending on the values used for the mass of the phragmocone and rostrum density. If the phragmocone was entirely filled with gas (which implies estimated phragmocone weights of 2.32 g for Clarkeiteuthis, 1.68 g for Passaloteuthis and 1.49 g for Hibolithes), the centre of mass lay closer to the head. If the phragmocone was filled with 30% liquid and 70% gas (which implies estimated weights of 13.49 g for Clarkeiteuthis, 7.99 g for Passaloteuthis and 10.99 g for Hibolithes), the centre of mass shifted backwards55. With a centre of mass in the middle of the body and close to the centre of buoyancy, the animal would have been able to turn its body into a horizontal position but it could rather easily turn in every direction. Accordingly, the animal presumably could maintain a stable horizontal position using its fins.

Mean body densities of Clarkeiteuthis, Passaloteuthis and Hibolithes were calculated to verify the calculations of their modes of life of using the values listed in Table 2. At a total length (including arms) of approximately 213 mm, Clarkeiteuthis weighed around 145.8 to 146.6 g and its body had a volume of 161.6 to 162.2 cm3 which implies a mean density of 0.90 up to 0.904 g/cm3. At a total length of approximately 300 mm (including arms), Passaloteuthis weighed around 170.9 to 179.3 g, had a body volume of 175.2 to 175.7 cm3 and thus a mean density between 1.004 and 1.051 g/cm3. At a total length (including arms) of about 456 mm, Hibolithes weighed around 547.8 to 560.5 g, had a body volume of 539.7 to 541.4 cm3 and a mean density of 1.011 to 1.035 g/cm3. Compared to the density of seawater of 1.025 g/cm3 at 3.5% salinity, Hibolithes would have been nearly neutrally buoyant with a phragmocone filling of about 12 to 53% of its volume; similarly, Passaloteuthis would have been neutrally buoyant with a phragomocone filling of around 33 to 73% and the diplobelid would have been positively buoyant up to a phragmocone filling of nearly 82 to 86%.

Table 2 The table shows reconstructions of volumes and densities of Passaloteuthis bisulcata, Hibolithes semisulcatus, and Clarkeiteuthis conocauda, which were used for the calculations of the centre of mass and buoyancy of this species29,30.

The second approach to reconstruct the mode of life for diplobelids and belemnitids are actualistic comparisons with recent cephalopods such as cranchiids and other teuthids. Due to their comparatively thin muscular mantle, the modern cranchiids are buoyant but poor swimmers, which often position themselves with the head down in the water column56,57. By contrast, modern non-buoyant teuthids (e.g. loliginids, ommastrephids or onychoteuthids) are mostly very muscular and can migrate over long distances. Belemnitids might have had a similar mode of life due to their anatomical features.

The Cranchiidae vary in mantle length between 100 and 2000 mm57,58. The sometimes translucent animals have an elongated, conical or cylindrical shape, a thin but muscular mantle and a short head with large eyes59. Their fins vary in size and shape depending on species and habit. The thin fins are either widely separate, small and paddle-shaped or medium to large and round60. Cranchiids are neutrally buoyant due to large coelomic cavities, which are filled with an ammonium chloride-solution (NH4Cl59,61). They are able to change their orientation in the water column by using their fins rather than by contracting the mantle cavity, i.e. hyponome action61. Due to these features, cranchiids are not very active swimmers but position themselves with the head down in the water column. In that respect, the diplobelids might have had a similar mode of life due to their supposed syn vivo position.


Reconstructing the mode of life of Jurassic coleoids is of interest because they are quite common in many localities and because the rostrum of belemnitids is widely used for examinations of stable isotope ratios to measure palaeotemperatures62,63. Such isotope data recently turned out to be far from trivial to interpret26,63. Other evidence for coleoid habitats and behaviour is rare; partially, this roots in the scarcity of complete specimens including soft parts, which are largely limited to Konservatlagerstätten32,46.

Most researchers suggest that belemnitids used their large and supposedly heavy calcitic rostra to obtain a horizontal swimming position with further support from the fins31. Information on the mode of life of diplobelids without a massive rostrum was missing.

According to our buoyancy calculations of Clarkeiteuthis, the centre of gravity lies in the middle of the animal and moderately close to the centre of buoyancy using the values listed in Table 2. This suggests that diplobelids could easily change their orientation from oblique to vertical. This also depended on the amount of chamber water in the phragmocone. If the phragmocone was partially filled with liquid, the orientation would have been rather horizontal to oblique. If the diplobelid maintained a vertical body orientation, this would imply that they were less active swimmers and thus rather planktonic with the capability of occasional short bursts of rapid swimming movements (to catch prey or to escape predators).

Belemnoteuthids such as Acanthoteuthis were suggested to have been fast swimmers32, which is partially based on a misinterpretation of the term ‘buoyant squid’. In contrast to Klug et al.32, Acanthoteuthis was probably rather nektoplanktonic, as suggested by the shape and dimensions of its statoliths, the thin aragonitic rostrum and the probably rather thin mantle (poorly preserved compared to the thicker mantle in, e.g., Plesioteuthis or Leptotheuthis).

In belemnitids with their characteristic partially calcitic rostrum64,65, the centre of mass and buoyancy were very close to each other according to our reconstructions. This result suggests that belemnitids with a comparatively long rostrum such as Passaloteuthis and Hibolithes could easily turn their body in each direction and thus also horizontally. The well-developed fins32,64 enabled the animal to steer, implying an at least temporally nektonic mode of life. Also, accepting that the rostrum supported fins31,65, this suggests that larger fins were present in belemnitids than in diplobelids, thus providing more dynamic lift31.

The calculations of the mean density for all three cephalopods show that their phragmocones were partially filled with liquid, because only then, they attained neutral buoyancy57; it is unclear, however, if these cephalopods had other organs controlling their average body density, thus making them neutrally (like the more passive planktonic squids) or slightly negatively buoyant (like nektonic ommastrephids and loliginids) with less chamber liquid. For Passaloteuthis and Hibolithes, the density of 1.004 to 1.051 g/cm3, is close to that of seawater, which corroborates the hypothesis of neutral or near-neutral buoyancy. With a low mean density of possibly 0.9 g/cm3, Clarkeiteuthis was rather strongly positively buoyant (which is unlikely) and thus, a higher amount of chamber liquid in the phragmocone is quite likely. These differences in density resulted from the small phragmocone and large rostrum of Passaloteuthis and Hibolithes in comparison to Clarkeiteuthis with a large phragmocone but a small sheath-like rostrum. This implies that both taxa regulated buoyancy by the quantity of chamber liquid31,46. From its morphology (rostrum, mantle) and the spatial distribution of mass, we conclude that Clarkeiteuthis might have had a mode of life similar to modern neutrally buoyant squids such as the cranchiids, oblique to vertical orientation of the body axis, a life as ambush predator (arm hooks, teleost prey) that stayed inactive most of the time but was capable of short pulses of fast swimming to catch prey.

If diplobelids were neutrally buoyant, it would have been possible that they changed their orientation in the water column similar to modern cranchiids by the use of their fins and funnel in addition to contracting their thin muscular mantle66. Assuming the presence of rather well-developed fins (concluded from rostrum size32) and a thick muscular mantle, belemnites were probably capable of using the fins and mantle contractions for more enduring and rapid swimming for, e.g., active pursuit hunting. Taking the even thicker muscular mantle in other Jurassic coleoids such as Plesioteuthis, Trachyteuthis or Leptotheuthis into account, these gladius-bearing octobrachians were likely the best swimmers among the Jurassic coleoids.


We have described adult individuals of Clarkeiteuthis conocauda, which caught small teleost fish of the species Leptolepis bronni. This diplobelid cephalopod was able to catch and hold on to the fish with its numerous arm hooks (Fig. 5). Having caught the fish, it likely first contracted its arms to bring the fish towards its mouth. Like some modern cephalopods, these diplobelids probably quickly killed the fish by cutting its spine, which is documented in distinct kinks in the vertebral column in three of the four specimens. The kink in the spine lies the closest to the cephalopod’s mouth. Once the teleost prey was immobilized, both animals started to sink, perhaps due to the deflated swim bladder of the fish, because the cephalopod tried to move out of sight of other predators or it stopped active movements. Both animals sunk into the oxygen depleted bottom waters where the cephalopod suffocated (the low oxygen conditions explain the exceptional preservation). These specimens provide therefore evidence for their predatory behaviour and indirect evidence for post capture sinking, which is a special case of distraction sinking47. Thus, diplobelids played an ecological role as neutrally buoyant medium-sized ambush predators in Jurassic food webs. The fact that all four specimens of Clarkeiteuthis were fossilized with a Leptolepis in their arm crown suggests that there possibly was some sort of prey specificity. Further indirect evidence for the hunting behaviour comes from their body orientation in the water during life; reconstructions of the centres of mass and buoyancy suggest a possible vertical orientation for diplobelids in contrast to a preferentially horizontal orientation of belemnitids.

Figure 5
figure 5

Reconstruction of Clarkeiteuthis conocauda after catching the small teleost Leptolepis bronni. Note the horizontally swimming individuals of Passaloteuthis bisulcata near the water surface. Artwork by CK.