The anatomy and development of the tails at the posterior part of the mantle were studied in several groups of Recent and extinct coleoid cephalopods; substantial differences in their formation were revealed. Males of the Recent loliginid squid Alloteuthis spp. form their tail by increased growth of the anterior part of the gladius with simultaneous growth of the posterior mantle. As a result, the gladius rolls longitudinally in the tail forming a pseudoconus. The attenuated tail in males of the squid Lycoteuthis springeri (Lycoteuthidae) is supported from inside by the special rod-like apical vacuolated cartilage. Adults of both sexes of recent Onykia robsoni and O. robusta form a carrot-shaped flexible chitinous rostrum supporting the attenuated tail. Adults of several Jurassic belemnites formed an elongated epirostrum posterior to their orthorostrum; the structures differed in growth and microstructure. Counts of growth microincrements within the orthorostrum and epirostrum were used to date their formation and estimate the age of belemnites. The development of the long rigid tail and the corresponding shift of the fin to the middle part of the mantle streamline the body and possibly facilitate the animal's movement in the water by gliding. The analogous tail formation in several independent groups points to its adaptive nature for the development of a more mobile adult phase in species of coleoid cephalopods.
Apart from incirrate octopods, recent coleoid cephalopods possess a muscular fin on the posterior part of the mantle. Most non-octopod cephalopods have a rhomboidal, oval or slightly attenuated apical tip of the mantle and fin that create a streamlined body to move effectively through the water. However some species develop a special morphological feature at the posterior end of the body. The apical tip of the mantle elongates (sometimes together with the posterior part of the fin) forming an attenuated ‘tail’, the size of which can sometimes exceed that of the mantle (Naef, 1921–23; Nesis, 1987).
The tail may form at different phases of ontogeny. For example, larval and early juvenile squid of the family Chiroteuthidae form a long tail that is supported from inside by the gladius. Narrow membranes on each side of the tail serve as longitudinal fins. The tail is broken near the tip of the fin in adult chiroteuthids (Vecchione, Robison & Roper, 1992). A similar tail with narrow longitudinal membranes is present in adults of the mesopelagic squid Joubiniteuthis portieri (Joubin, 1916) (Young & Roper, 1969). In adults, the tail may be formed in both sexes as in the squids Uroteuthis bartschi Rehder, 1945 (Loliginidae), Batoteuthis skolops Young & Roper, 1968 (Batoteuthidae) and Onykia robsoni (Adam, 1962) (Onychoteuthidae) or just in one sex, commonly in males as in the squids Alloteuthis spp. (Loliginidae), Lycoteuthis springeri (Voss, 1956) (Lycoteuthidae) and in the cuttlefish Sepia confusa Smith, 1916 (Sepiidae) (Nesis, 1987). The ethological and ecological roles of the tail in these recent adult cephalopods are unknown.
Elongation of the posterior body in adults is also known in fossil coleoids. Several Jurassic belemnites (suborder Belemnitina), belonging to the genera Dactyloteuthis, Acrocoelites, Megateuthis, Pleurobelus, Micropassaloteuthis, Salpingoteuthis, Holcobelus and Cuspiteuthis (=Youngibelus) developed a calcitic orthorostrum with dense lamellation during their juvenile and subadult ontogenetic phases (Müller-Stoll, 1936; Doyle, 1985; Bandel & Spaeth, 1988; Schlegelmilch, 1998; Fuchs, 2012; Weis, Mariotti & Riegraf, 2012). Upon reaching the adult phase, however, the growth pattern of the rostrum changed by developing an elongate epirostrum, which differs by its wider lamellation and its different composition. It has been suggested that the epirostrum was composed mainly of aragonite, which during diagenesis was transformed into calcite or dissolved (Bandel & Spaeth, 1988; Dullo & Bandel, 1988). The ontogenetic growth of the rostrum was completed by covering of the epirostral stage with calcitic external layers, similar in structure to the orthorostrum. As a result, the posterior part of their mantle elongated, presumably forming a ‘tail’. The epirostrum-bearing belemnites appeared first in the Pliensbachian (Early Jurassic, ca. 194 Ma), and radiated during Toarcian-Bajocian (Early/Middle Jurassic, ca. 187–166 Ma). Epirostra (albeit shorter than in Jurassic animals) were also found in Early Cretaceous (Albian, ca. 113 Ma) belemnites of the suborder Pachybelemnopseina, for example in Neohibolites minimus (Miller & Lister, 1826) (Spaeth, 1971). The existence of specimens with the same size orthorostrum either with or without epirostrum in Cuspiteuthis tubularis (Young & Bird, 1822) suggested possible sexual dimorphism in the ‘tail’ development similar to modern squid (Doyle, 1985). It has been hypothesized that the epirostrum might have served as additional balance for some specialized organ like an enlarged hectocotylus in males or brood pouch in females (Doyle, 1985), but no evidence to confirm this hypothesis has been provided. The age of adult specimens bearing epirostra was estimated to range between one and two years by counting growth microincrements in the rostra of the Middle Jurassic belemnite Megateuthis gigantea (Schlotheim, 1820) (Dunca et al., 2006). However, the time of formation of various parts of the rostrum remains unknown.
The main aim of the present study is to analyse and compare the structure and development of the ‘tail’ in the posterior tip of the mantle in fossil and recent coleoid cephalopods. Additionally, the duration of the epirostrum development, as well as other parts of the guard in Jurassic epirostrum-bearing belemnites has been estimated by counting growth microincrements. Based on ontogenetic changes in body shape, a new hypothesis on the possible ecological role of the tail in adult coleoid cephalopods is suggested.
MATERIAL AND METHODS
Sixty-two specimens of Alloteuthis subulata (Lamarck, 1798) were sampled from landings in the fishing port of Vigo (Galicia, north-western Spain) in April 2012 and frozen until further analysis in the laboratory. Both subadult and adult males and females ranging from 60 to 110 mm mantle length (ML) with and without the attenuated tip of the mantle were examined. Fifty-five specimens of A. africana Adam, 1950 were collected during a research cruise carried out on the Moroccan shelf (depths 80–110 m) under the Canary Current Large Marine Ecosystem project in July 2012. Sizes of subadult and adult males ranged from 60 to 152 mm ML, females from 55 to 95 mm ML. After capture, all specimens were fixed in 6% formalin.
Two specimens of Lycoteuthis springeri were borrowed from the collections of the Department of Invertebrate Zoology, National Museum of Natural History, Washington D.C. (USNM). The adult male of 83 mm ML (USNM 1205374) and female of 50 mm ML (USNM 575112) were caught in the Gulf of Mexico and preserved in isopropyl alcohol.
Rostra of five species of Jurassic belemnites from the palaeontological collections of the Luxembourg Museum of Natural History (MNHNL) were analysed including 11 specimens (MNHNL BEL440–450) of Megateuthis suevica (Klein, 1773), 3 specimens (MNHNL BEL435–437) of M. elliptica (Miller, 1826) (from Bajocian deposits in Rumelange, Luxembourg), 2 specimens (MNHNL BEL438–439) of M. rhenana (Oppel, 1857) (from Toarcian deposits in Kayl, Luxembourg), 9 specimens of Dactyloteuthis semistriata (Münster, 1830) (MNHNL BEL426–434) and 8 specimens (MNHNL BEL418–425) of Acrocoelites riegrafi Doyle, 1992 (from Toarcian deposits in Dudelange, Luxembourg).
Tail morphology and anatomy
In all squid studied, the dorsal ML was measured to the nearest 1 mm, and sex and maturity stage were identified using the maturity scale of Lipinski (1979). Specimens in good condition were photographed from the dorsal and ventral sides. Dimensions and morphology of the tail, mantle and fin were analysed and sketched from the photographs of specimens of different sizes and sexes. The posterior mantle and tail were dissected to investigate the internal structure and position of the gladius. Then, the gladius was dissected completely free of the mantle, embedded in a gelatin-glycerin mixture in small Petri dishes and covered by a cover glass. Different parts of the gladius (Fig. 1A) were photographed using an Olympus D70 camera and Olympus SZX12 dissecting microscope. Their microstructure was examined under transmitted light with an Olympus BX51 compound microscope at 50–100× magnification.
Belemnite rostrum microstructure
Belemnite rostra were embedded in a fillable casting resin (Martins Plastic Ltd) to form rectangular blocks. The blocks were mounted on the chuck holder for cutting with a Buehler 1000 Isomet precision saw. From each rostrum, three to four successive longitudinal sections of 1–1.5 mm thickness were made near the main longitudinal axis. Then the sections were ground using 600 grit waterproof sandpaper and polished using 1200 grit sandpaper under running water on a Buehler Metaserve 2000 grinder/polisher. Longitudinal sections were then put on glass slides, embedded in clear casting resin and covered with a coverslip.
Usually, one section that had been cut nearest the longitudinal axis of the guard was chosen to examine the microstructure. Growth increments were examined and counted under ×100 magnification of the compound microscope, as described by Wierzbowski (2013). Wherever possible, growth increments were counted separately in juvenile rostrum, orthorostrum and epirostrum (terminology of Müller-Stoll, 1936) (Fig. 1B). The juvenile rostrum (= nepionic stage of Pugaczewska, 1961; Doyle, 1990; = earliest juvenile rostrum of Weis et al., 2012) was usually visible as the more opaque zone near the protoconch and primordial rostrum. Growth increments were clearly visible by transmitted light under the microscope in the juvenile rostrum and major part of the orthorostrum. The microstructure in external parts of the orthorostrum and of the whole epirostrum was usually nontransparent. Growth increments in these parts could be observed and counted under reflected light using a powerful fibre-optic light source (Olympus KL-1500). If the long and sometimes bent epirostrum was offset from the main longitudinal axis, growth increments were examined in two or even three adjacent sections of the same rostrum; some feature (like a distinct crack) in the guard microstructure was noted to enable the observations to be positioned consistently from one section to another.
Growth increments were counted independently by two readers (AA and ZS) to minimize error. If the difference between the two counts was <5%, the mean number of increments was accepted. If the difference was >5%, a third count was made. If the difference between the third and second counts was still >5% the section was rejected from the age estimations, if <5%, it was accepted.
Alloteuthis subulata (Loliginidae)
Juvenile and immature specimens of both sexes had a pointed posterior part of the fin and mantle (Fig. 2F). The tail started to form in maturing animals of 55–58 mm ML. Mature females developed a short tail not exceeding 15–20% ML in length (Fig. 2E). The short tail of small mature males (70–80 mm ML) was similar to that of mature females. All large mature males (110–130 mm ML) had an exceptionally long and attenuated posterior part of the mantle (40–45% ML) with the very narrow posterior parts of the fin reaching about half the tail length. During ontogeny, the fin appeared to change its position on the mantle from posterior terminal (immature animals) and subterminal (mature females and small males) to middle and even slightly anterior in large mature males (Fig. 2D).
The gladius of juvenile and immature squid had a typical ‘loliginid’ shape (Bizikov, 2008) with wide lateral plates and cone flags and a relatively short rachis. At the posterior end of the conus there was a small but well resolved triangular rostrum (Fig. 3B). In mature females and small mature males, the posterior part of the cone flags started to bend around the internal wall of the tail, forming a tube with opened ventral side (pseudocone). In large mature males, the elongate pseudocone occupied and therefore supported the whole tail (Fig. 3C). The gonad entered the pseudocone. The rostrum was located at the most posterior end of the pseudocone, being similar in size and shape to that of small mature males and females. This indicated that the gladius lengthens at the anterior part of the proostracum, as the rostrum was small and did not increase much in length posteriorly.
Alloteuthis africana (Loliginidae)
Patterns of ontogenetic development of the tail and gladius in A. africana were similar to those in A. subulata. The tail started to develop in maturing animals of both sexes at 60–70 mm ML (Fig. 2A). Mature females and small males had a short tail supported inside by the well resolved pseudocone of the gladius (Fig. 2B). Large mature males (130–150 mm ML) had a long, thin tail attaining approximately half the total ML (45–48% ML). The mantle in A. africana was relatively narrower than in A. subulata, giving the body a ‘slender’ appearance (Fig. 2). The small but distinct rostrum was situated at the posterior tip of the conus of the gladius in all animals studied (Fig. 3A).
Lycoteuthis springeri (Lycoteuthidae)
The adult female of L. springeri had a pointed fin at the posterior part of the mantle (8% ML), not attenuated into the tail (Fig. 2H). The posterior end of the gladius was situated near the posterior third of the fin. The tip of the mantle was supported by a short, carrot-shaped, vacuolated, apical cartilage. The adult male had an attenuated posterior mantle, forming a distinct tail (30% ML; Fig. 2G). Similar to the female specimen, the posterior part of the gladius was located near the posterior third of the fin. The attenuated tail was supported by an elongate, vacuolated, apical cartilage. The gladius in both sexes contained a small cap-like rostrum located on the top of the cone (Fig. 3D). Because of the attenuated tail, the fin occupied the middle position along the mantle in male, whereas in female it had a terminal position (Fig. 2G, H).
Dactyloteuthis semistriata (Megateuthididae)
Out of nine specimens studied, two had an exceptional state of preservation of their juvenile rostrum including protoconch, first several septa of the phragmocone and primordial rostrum (Fig. 4A, B). The primordial rostrum did not show any growth increments. The juvenile rostrum had semitranslucent microstructure, consisting of 17–21 well-defined narrow increments (Fig. 4C). It was possible to trace every increment in the juvenile rostrum with correspondent septa in the phragmocone, meaning that in early juveniles they were deposited at the same time and rate. Outside the juvenile rostrum, the increments became wider (Fig. 4D) and more translucent. The width of the increments gradually decreased to the periphery of the orthorostrum. There were some stress marks within the orthorostrum microstructure, but they were irregular and consisted of from 5 to 25 increments. The number of increments in the translucent part of the fully developed orthorostrum ranged from 187 to 343 (mean 277). The opaque outer part of the orthorostrum consisted of an additional 12–112 increments (mean 73.7). The total number of growth increments in the fully developed finger-shaped orthorostrum varied from 278 to 458 (mean 377) increments (Table 1).
|Specimen ID||Rostrum diameter, mm||Number of increments|
|Specimen ID||Rostrum diameter, mm||Number of increments|
Two specimens did not have any developed epirostrum in the posterior orthorostrum (Fig. 4B); the increment number in these specimens varied between 313 and 331 (including the mean 19 increments counted on the juvenile rostrum above). One specimen had just begun to develop the epirostrum (61 increments), which lacked an external covering. The others had either broken or almost intact epirostra with a complete covering, meaning that the rostrum of those belemnites had finished its ontogenetic development. Characteristic of D. semistriata, the epirostrum's chopstick shape differed markedly from the finger-like shape of the orthorostrum (Fig. 4E). The growth increments in the epirostrum were visible only in a region near the external covering (Fig. 4F). The microstructure of the middle part of the epirostrum was quite amorphous and in most specimens did not have well-resolved growth increments. Two specimens had cracked epirostra and the inner part of the epirostra was filled with sediment or cement up to the boundary of the orthorostrum (Fig. 4E). The maximum number of growth increments in the epirostrum was estimated to be up to 288–374, but even these epirostra had broken tips. The external covering had 18–34 (mean 25) opaque growth increments.
The total number of growth increments within the rostra of D. semistriata varied between 312 increments in the specimen with complete orthorostrum and at least 752 increments in the specimen with complete epirostrum.
Acrocoelites riegrafi (Megateuthididae)
The rostrum of this species had a chopstick shape similar to that of D. semistriata, but their microstructure was very different (Fig. 5A). Four of eight specimens had a well-preserved juvenile rostrum around the spherical protoconch, primordial rostrum and first several septa of the phragmocone (Fig. 5B). The total number of growth increments in the juvenile rostrum varied from 14 to 18 (mean 16) and was therefore slightly less than in D. semistriata (Table 1). The orthorostrum microstructure was translucent with well-defined growth increments (Fig. 5C). The number of increments within the complete orthorostrum varied from 190 to 319 (mean 226). The outer part of the orthorostrum was also translucent, unlike that in D. semistriata. The boundary between orthorostrum and epirostrum was not sharply delimited, but the epirostrum had an opaque microstructure (Fig. 5D). The pattern of deposition of growth increments was similar in orthorostrum and epirostrum, however the middle (axial) part of the epirostrum had a more amorphous microstructure than did the orthorostrum (Fig. 5E, F). In three specimens, the middle epirostrum contained sediment, which might have appeared there through the cracks of the external part of the epirostrum (Fig. 5G). The total number of growth increments in the epirostrum varied between 55 and 210 (mean 124), being slightly less than in D. semistriata. The epirostrum did not have a well resolved external covering as in D. semistriata.
The total number of growth increments in completely formed rostra of A. riegrafi varied between 190 (in specimen without epirostrum) to at least 429 increments with fully developed epirostrum.
Megateuthis spp. (Megateuthididae)
The rostrum microstructure in all three species of Megateuthis spp. studied was quite similar (Fig. 6A). The juvenile rostrum was semi-opaque, with 18–29 (mean 21) growth increments (Fig. 6B). Outside the juvenile rostrum, the microstructure of the orthorostrum became more transparent, with clear and well-resolved growth increments medially and peripherally (Fig. 6C, D). However, the area near the longitudinal axis of the orthorostrum was much more opaque than the peripheral area (Fig. 6A). The boundary between the orthorostrum and epirostrum was not sharply delimited. In fact, it questioned whether the rostrum was divided into ortho- and epirostral parts, or is one structure (Fig. 6A). However, in the ‘epirostral’ (posterior) part of the rostrum, the central area had opaque and often deformed growth layers (Fig. 6G). Close to the rostrum tip, the central part of the ‘epirostrum’ was often completely destroyed and filled with sediment from either the cracked tip or some transverse cracks in the area (Fig. 6E, F). It is notable that the phragmocone cavity was filled with the same sediments. If there were some cracks near the longitudinal axis of the juvenile rostrum and orthorostrum, the growth layers in these areas would also be deformed as in the epirostrum (Fig. 6B). The outer part of the ‘epirostrum’ always had clearly defined and nondeformed growth increments, making it possible to count all of them up to the rostrum tip. There was no distinct external covering of the ‘epirostrum’ as observed in D. semistriata.
As it was not possible to distinguish the microstructure of orthorostrum and epirostrum, a total number of growth increments was estimated. In small specimens (M. rhenana, 14 mm in diameter) only 101 increments were observed. In large specimens of M. suevica (26–32 mm in diameter), up to 528–585 increments were observed. The largest belemnites of M. suevica (39 mm in diameter) had 948 and 1084 increments in their rostra (Table 1).
Types of tail development
Our study combined with some literature data show that the attenuated posterior tip of the mantle and fin (‘tail’) is not homologous in the various extinct and Recent coleoid cephalopods. At least four different types of tail development can be distinguished.
The first type of tail development characterizes adult squid of the genus Alloteuthis—A. subulata and A. africana (Fig. 7A). The elongated tip of the mantle forms in mature specimens of both sexes, a short tail in females and a long tail in large males. From the inside, the tail is supported by the chitinous gladius with lateral plates and cone flags curved longitudinally to form a tube. The tube is not closed on the ventral side and forms a so-called ‘pseudocone’ (Bizikov, 2008). The pseudocone is therefore almost the same length as the tail. As the gladius cone bears a small rostrum on its tip, the gladius does not grow posteriorly. Both subadult squid without the pseudocone and adult squid with a long pseudocone have a similar-sized rostrum at the tip of their gladius. Naef (1921_23: 222) suggested that the muscular part of the tail grows posteriorly, shifting the fin to the middle of the mantle, whereas the gladius continues to grow anteriorly, with its posterior end being “gradually pushed posteriorly”. However, this would require that the gladius be moved within the shell sac; this does not happen in other species of squid (Bizikov, 2008). Obviously, further observation of the tail elongation is required to elucidate how the tail develops in Alloteuthis species. As a result of tail elongation, the body proportions and especially the position of the fin (in the middle of the mantle) in adult males differ markedly from those of subadult males and adult small males and females where the tail is not elongate.
The second type of tail development is observed in Lycoteuthis springeri (Fig. 7B). In adults of this squid, the posterior mantle and fin are not supported by the gladius as in Loliginidae. Instead, the tip of the gladius cone is located proximal to the anterior third of the fin. The apical tip of the mantle and fin are supported from the inside by a flexible chitinous/cartilaginous rod that forms outside the shell sac and therefore cannot be considered as a part of the gladius (Bizikov, 2008). This so-called ‘vacuolated apical cartilage’ is a derivative of the fin cartilage (Bizikov, 2008). The vacuolated apical cartilage is short in females, but is elongate and supports the attenuated tail and fin in adult males. Therefore, both the external muscle layer and internal vacuolated apical cartilage grow in a posterior direction in males, forming the distinct tail. Tail formation also ‘displaces’ the fin to the middle of the male mantle, as in Alloteuthis species. Similar development of the vacuolated apical cartilage has been recorded in other oegopsid squid, such as Ommastrephidae and Gonatidae; in the latter it also forms a noticeable tail (Bizikov, 2008; Arkhipkin, Bizikov & Fuchs, 2012).
The third type of tail development characterizes coleoids with a developed rostrum of the gladius (Fig. 7C), such as the squids Onykia and Kondakovia (Onychoteuthidae) (Bizikov, 2008). Notably, juveniles of these squid have a relatively short rostrum and no tail, whereas immature and mature adults, especially in the species O. robsoni and O. robusta (Verrill, 1876), have a well-developed tail (Jereb & Roper, 2010). In these species, the tip of the gladius cone is located approximately at the posterior third of the fin, quite similar to its position in L. springeri. However, the tail (which includes the attenuated tip of the mantle and fin) is supported from the inside by the chitinous/cartilaginous rostrum. The microstructure and chemical composition are the same throughout the rostrum (Bizikov & Arkhipkin, 1997).
The same type of tail development is found in all species of belemnites studied (Fig. 7C). In large belemnites belonging to the genus Megateuthis, the development of the rostrum (and presumably tail) was most similar to the modern O. robusta. However, the microstructure and composition of the rostrum in Megateuthis species differed. Most probably, the middle part along the apical line of the rostrum contained a high proportion of organic components (chitinous/cartilaginous?) that were not preserved after the animal died, contrary to the calcified periphery of the rostrum. Quite specific patterns of growth layer deformation indicate the ‘amorphous’ nature of the middle axial layers that had been present since the early juvenile phase in these belemnites. In adults, deformation of the axial rostrum layers indicates that the apical part of the rostrum (especially the epirostrum) may have been flexible, as in the rostrum-bearing modern squid. A similar effect is observed in modern Sepia officinalis Linnaeus, 1758, where the live animal has an aragonitic rostrum covered by chitinous layers. In dead animals, only the aragonitic part of the rostrum is fossilized.
The belemnite A. riegrafi had a solid calcified orthorostrum, because no deformation of growth increments/layers has been observed. However, the epirostrum may have had a similar structure to that of Megateuthis species, with the middle axial part containing either deformed or completely destroyed growth layers, with the inner space of epirostrum filled with sediments or cement. The general shape and pattern of growth increment deposition in the orthorostrum and epirostrum were not different; therefore there was no clear boundary between these two rostral parts. The most ‘extreme’ development of the tail was observed in the belemnite D. semistriata. These belemnites had a finger-like fully calcified orthorostrum until the subadult phase, without any deformations of growth increments anywhere. Then, they started to develop an epirostrum that was very different in microstructure and chemical composition (aragonitic according to Bandel & Spaeth, 1988) from the orthorostrum. Most probably, it had a higher organic content, being slightly calcified only in the outer part. At the end of the adult phase, the flexible organic part was sealed by a special calcified covering layer. When the epirostrum was cracked or chipped after death and subsequent burial of the animal, sediments filled the epirostral cavity that would have appeared after decomposition of the organic components.
The fourth type of the tail development is observed in the cuttlefish S. confusa (Carleton & Robson, 1924). Adult males of this species develop a muscular tail at the posterior apex of their mantle. The tail is formed by the fused posterior parts of the fins, which are elongated until they are as long as the total ML. No elastic or hard structure supports this tail inside the body.
Time and period of tail development
Apart from some deepwater squid where the tail develops in paralarval forms and then either disappears in adults (Chiroteuthidae, Vecchione et al., 1992) or is retained throughout life (Joubiniteuthidae, Young & Roper, 1969), all other squid studied here form the tail in the adult (and possibly mature) phase of their ontogeny.
Unfortunately, neither the gladius lateral plates in Alloteuthis species nor the vacuolated apical cartilage in L. springeri revealed distinct growth increments, making it impossible to estimate the timing of tail development in these squid. However, these tails appear only in mature males in both species (Jereb & Roper, 2010). So, the tail forms in these squid over a relatively short period.
The rostrum microstructure of all the studied belemnites revealed growth increments both in the orthorostrum and epirostrum. In species without a clear boundary between these parts, it was possible to estimate the total number of increments. The largest numbers of increments (1000–1080) were observed in the large rostra (39–40 mm in diameter) of M. suevica (= M. gigantea). This was twice as many as observed in M. gigantea by Dunca et al. (2006). This difference in total growth increment counts might be explained by the fact that those authors counted them in transverse sections through the widest part, therefore missing growth increments formed in the more apical parts of the epirostrum. Dunca et al. (2006) suggested that growth increments formed daily, based on the observation that 15 and 30 increment groupings (‘bundles’) possibly reflected fortnightly periods of shell growth controlled by lunar cycles. Assuming the approximate starting point of epirostrum formation by the presence of either deformed or missing growth increments in the axial part of the rostrum, it could take 300–400 d for the epirostrum to form completely in large adult Megateuthis and up to 210 d in smaller A. riegrafi. In D. semistriata with a clear boundary between orthorostrum and epirostrum, it could take up to one year to form the long rostrum. So, the examination of microstructure and growth increment counts in our study showed that the epirostrum grew for a much longer time than had been supposed by examination of wider growth increments in the epirostrum without counting them (Müller-Stoll, 1936; Bandel & Spaeth, 1988).
Reasons for tail formation: from dart to glider?
Morphological and anatomical differences in tail development between mature females and small mature males vs large males in A. subulata were described by Naef (1921–23). It is known that formation of the long tail in large males of A. subulata increases the length of the animal with little corresponding increase in body weight (Rodhouse, Swinfen & Murray, 1988). However, neither of these studies, nor others describing tail formation (Laptikhovsky, Salman & Moustahfid, 2005; Anderson et al., 2008) and sexual dimorphism in growth rates in Alloteuthis species (Lipinski, 1986; Arkhipkin & Nekludova, 1993) have revealed or suggested any reasons for tail formation in adult squid. Bizikov & Arkhipkin (1997) suggested that the long rostrum supporting the tail in O. robusta might serve as a shock absorber during swimming. The rostrum elongation in extinct Jurassic belemnites has been suggested to counterbalance the development of specialized reproductive organs (Doyle, 1985).
The tail develops in adults of either planktonic (Taoninae in Cranchiidae, Nesis, 1987) or nektonic squid that do not have powerful mantle muscles for long migrations. Among them, Alloteuthis species are among the smallest (in weight) loliginids without strong mantles. The same may also apply to micronektonic L. springeri and one of the smallest ommastrephids Ornithoteuthis antillarum Adam, 1957, whose adults also develop a tail (Nesis, 1987). Large onychoteuthids O. robsoni and O. robusta become ammoniacal squid that have significant amounts of ammonium chloride in their flesh contributing to their neutral buoyancy, but at the same time compromising their swimming abilities (Voight, Pörtner & O'Dor, 1995). Notably, the tail does not develop in any powerful swimming squid such as the majority of Ommastrephidae that all have a ‘dart’-shaped body with the terminal fin mainly serving to balance/stabilize the body position during jet propulsion (Hoar et al., 1994).
Adult coleoids with a fully developed tail acquire a streamlined ‘arrow’-like body shape with relatively thin mantle and a more centrally positioned fin compared to subadult animals without a tail. Such a streamlined body shape with a shifted fin should enhance the gliding abilities of these squid; it should therefore facilitate long migrations without substantial contribution of the mantle muscles. Large males of A. subulata (because of their longer life span) are known to take part in gene exchange between various seasonal cohorts (Rodhouse et al., 1988). It could also be that the large males migrate between different groups/subpopulations, providing not only a temporal but a spatial connection between them. Large males of other loliginids such as Doryteuthis gahi (d'Orbigny, 1835) that have a more streamlined body shape than their small counterparts could play similar roles in connectivity of subpopulations (Fig. 8). Acquisition of a streamlined body in large adult squid O. robusta might aid them in travelling long distances from their main feeding grounds in subpolar northern Pacific waters to their spawning grounds located in tropical waters around Hawaii (Wakabayashi et al., 2007).
Little is known about the swimming abilities of belemnites, with the exception of a few case studies on specific taxa (Monks, Hardwick & Gale, 1996; Hewitt, Westermann & Judd, 1999; Lewy, 2009). It is however assumed that species with ‘bulky’ rostra (such as finger-shaped rostra of Jurassic D. semistriata) and cone-shaped rostra (as in Megateuthis species) were less able swimmers and migrated shorter distances than species with spindle-shaped rostra from the Middle-Late Jurassic and Cretaceous (Doyle & Howlett, 1989: 179). We hypothesize that development of the epirostrum in adults of Jurassic species might help them acquire a streamlined body and therefore facilitated their spatial migrations, by analogy with modern squid. This strategy, however, would not have been as effective as the slender spindle-like rostrum shape, which became dominant during the Middle-Upper Jurassic while the Belemnitina declined in the Middle Jurassic (e.g. Stevens, 1965; Mutterlose, 1986; Doyle, 1987, 1994; Doyle & Howlett, 1989; Doyle & Pirrie, 1999; Weis et al., 2012; Weis, Mariotti & Wendt, 2014).
The posterior mantle and fins in adults of some species of Recent and extinct coleoid cephalopods form an attenuated tail. Anatomically, the tail is supported from the inside by a pseudocone of the gladius (as in the myopsid Alloteuthis), a vacuolated apical cartilage (as in the oegopsid Lycoteuthidae and Gonatidae) or a long rostrum of the internal shell (as in the oegopsid Onychoteuthidae and Jurassic epirostrum-bearing belemnites). The development of the long, rigid tail with corresponding shift of the fin to the middle part of the mantle makes the body more streamlined and possibly facilitates the animal's gliding movements. The analogous tail found in several independent groups suggests that it is adaptive. A more mobile adult ontogenetic phase would be able to provide both spatial and temporal connection between various subpopulations of these previously less active coleoid cephalopods.
We thank Julio Portela (Vigo, Spain) for providing the samples of Alloteuthis africana and A. subulata, Jean-Paul Fayard (Metz, France) and Kurt Meiers (Losheim, Germany) for their donations of additional belemnite rostra, and Wolfgang Riegraf (Münster, Germany) for discussions on epirostral development in belemnites. This work benefits from the support of the project ‘Coleoid cephalopods in the European Mesozoic: systematics, evolution and paleobiogeography of fossil coleoids’ at the National Museum of Natural History, Luxembourg.