- Top of page
- Material and Methods
Abstract: The palaeohistological study of the calcified internal organ of Axelrodichthys araripensis Maisey, 1986, a coelacanthiform from the Lower Cretaceous of Brazil (Crato (Aptian) and Santana (Albian) formations of the Araripe Basin), shows that the walls of this organ consist of osseous blades of variable thickness separated from each other by the matrix. This indicates that, in the living individuals, the walls were reinforced by ossified plates, probably separated by conjunctive tissue. This calcified sheath present in Axelrodichthys, as well as in other fossil coelacanths, lies in ventral position relative to the gut and its single anterior opening is located under the opercle, suggesting a direct connection with the pharynx or the oesophagus. The calcified organ of Axelrodichthys, like that of other fossil coelacanths, is here regarded as an ‘ossified lung’ and compared with the ‘fatty lung’ of the extant coelacanth Latimeria. The reinforcement of the pulmonary walls by the overlying osseous blades could be interpreted as a means of adapting volumetric changes in the manner of bellows, a necessary function for ventilation in pulmonary respiration. Other functional hypotheses such as hydrostatic and/or acoustic functions are also discussed.
F ossil coelacanths are known since 1822, when the palaeontologist G. Mantell described the genus Macropoma from the Upper Cretaceous of Sussex, England (Agassiz 1839). Subsequently, many species have been referred to this group of sarcopterygians, now known to have a temporal range from the Devonian to the Recent (Forey and Cloutier 1991; Forey 1998).
Among the differences between the extant coelacanth Latimeria and most fossil, coelacanths is the presence, in the fossils, of a visceral calcified structure currently named in literature as ‘bladder’ (e.g. Woodward 1891). This structure is known in the Palaeozoic genera Coelacanthus, Caridosuctor, Rhabdoderma, Hadronector and in the Mesozoic genera Axelrodichthys, Mawsonia, Macropoma, Undina, Coccoderma, Lybis, Laugia, Swenzia, and Piveteauia. The Carboniferous genera Allenypterus and Polyosteorhynchus also present a calcified bladder (Lund and Lund 1985) (contrary to a previous assumption of one of us (Clément 2005)). The condition (either absence or presence of a calcified organ) is unknown in all other coelacanth taxa. Nevertheless, it seems to be absent in the two well-preserved Mesozoic genera Whiteia and Diplurus. Although Whiteia is very common in the Triassic nodules of Madagascar, none of the studied specimens show any trace of calcified internal organ, whereas the rare genera Piveteauia and Coelacanthus, from the same localities, do possess such an organ. When entirely preserved, this calcified organ occupies the length of the abdominal cavity, reaching back as far as the pelvic fins (Text-fig. 1A). Latimeria possesses a tubular, fat-filled organ (about 4 cm in diameter and 45 cm in length in adults, usually filling the entire length of the abdominal cavity) that is mostly situated in a dorsal position relative to the gut but with a direct link to the ventral side of the oesophagal. Some anatomists have called this organ ‘lung’ or ‘fatty lung’ (Millot et al. 1978). The physiological function of the calcified organ in fossil coelacanths and its homology with the fatty lung of Latimeria remained to be proven.
Figure TEXT-FIG. 1.. Axelrodichthys araripensis. A, Adult specimen, ‘Josa collection’ deposited at the Laboratoire de Paléontologie, MNHN, left lateral view showing the calcified bladder (arrow head), scale bar represents 50 mm. B, juvenile specimen, UERJ-PMB33, showing the calcified bladder somewhat distorted (arrow head), scale bar represents 40 mm. C, a very young specimen, MPSC-287, from Brito and Martill 1999, showing the bladder (arrow head), scale bar represents 10 mm.
Download figure to PowerPoint
To provide factual data for this long unresolved issue (i.e. is the calcified organ of fossil coelacanths homologous to the ‘lung’ of the recent Latimeria?), we present a histological study of the calcified organ of the fossil Mesozoic coelacanth, Axelrodichthys araripensis.
The following interpretations, and those of previous authors (e.g. Williamson 1849; Woodward 1891; Maisey 1986), of the role of the calcified bladders of fossil coelacanths should be considered cautiously as there is little or no fossil evidence for any particular function.
Material and Methods
- Top of page
- Material and Methods
We have studied the histological structure of fossilized ‘bladder’ preserved in different specimens of Axelrodichthys araripensis. This species is relatively common in the Lower Cretaceous Santana Formation of the Araripe Basin, northeastern Brazil (Campos and Wenz 1982; Maisey 1986), being also known from the slightly older Crato Formation of the same Basin (Brito and Martill 1999).
To observe the histological structure of the fossils, about 1 cm thick slices were cut in the middle of the bladder. Then, each slice was embedded in Stratyl (Chronolite 2060). Cross and horizontal sections (according to the antero-posterior axis of the calcified organ) were cut with a saw (‘Isomet’ or ‘Brot’ for the larger samples), glued on a glass slide and then ground to the appropriate thickness. The ground sections were observed in transmitted natural and polarized light.
The external surface of the bladder of young Axelrodichthys specimens reaching less than 200 mm TL (total length: the distance between the most anterior point of the snout to the most posterior point of the caudal fin) was observed using a scanning electron microscope (SEM).
A high-resolution computerized axial tomography scanning (CAT scan) of the specimen MNHN C.20 (male, 130 cm TL; for further information see Bruton and Coutouvidis 1991) was made at the Centre Hospitalier Intercommunal of Villeneuve-Saint-Georges (France). Scan parameters are as following: 120 kV, 158 mA, slice thickness = 0.8 mm, 1807 views.
High-resolution X-ray computed tomography is a nondestructive and noninvasive technique that has the unique ability to image a combination of bone, cartilage and soft tissue (liver, muscles, fat, blood vessels, etc.). These CAT images are suited to three-dimensional reconstruction. The 3D image processing software MIMICS (Materialise’s Interactive Medical Image Control System) has been used to create 3D virtual reconstructions.
To make ourselves perfectly clear, the ossified organ of the fossil coelacanths (previously called ‘internal osseous viscus’, ‘ossified stomach’, ‘air bladder’, ‘gas bladder’, ‘swim bladder’) will be here simply defined as a bladder, whose definition is an inflated and hollow sac.
Institutional abbreviations. MNHN, Muséum national d’Histoire naturelle, Paris; MPSC, Museu de Paleontologia de Santana do Cariri; UERJ, Universidade do Estado do Rio de Janeiro.
Referred material. †Axelrodichthys araripensis. MNHN ‘Josa collection’ (adult specimen from the Santana Formation); MPSC-287 (juvenile specimen from the Crato Formation); UERJ-PMB33 (juvenile specimen from the Santana Formation); UERJ-PMB 143 (adult specimen from the Santana Formation). Latimeria chalumnae MNHN C. 20 (adult male caught the 19th of June 1960 at Itsoundzou, Comoro Islands, Indian Ocean, and preserved ever since in a 7 per cent diluted formalin solution).
Anatomical abbreviations used in the text-figures. Bo, primary bone; G, gut; Ga, limestone matrix; GO, fatty organ; OD, oesophagal diverticulum; Oe, oesophagus; St, stomach.
The bladder of Axelrodichthys is a well-calcified structure, easily observed in specimens preserved in lateral view (Text-fig. 1A–C). This structure is situated in the ventral part of the body and its anterior part turns up where it is covered by the opercle. At this level, its anterior extremity opens by a median orifice as the neck of a bottle and may have opened into the pharynx or communicated with the oesophagus as proposed by Woodward (1891). This anterior opening seems to be the only aperture of the calcified organ; its posterior extremity is generally more or less pointed, but always closed (Text-fig. 1A). This bladder is generally divided into an anterior and a posterior chamber, separated by a constriction (Maisey 1986; Forey 1998). More than one constriction can also occur (see Text-fig. 1A, where two constrictions are present). In adult individuals, the walls of this bladder comprise a series of superimposed bony plates (Text-fig. 2A–D), as in other fossil coelacanths such as the Triassic Piveteauia madagascarensis and the Jurassic Swenzia latimerae (Clément 1999, 2005, 2006). Each plate is gently concave on its internal side with the longest plates located on the peripherical area of the organ, the smallest and thinnest ones being preserved along the inner surface of the walls of the bladder. The thickness of the laminae decreases regularly towards their margin, and both surfaces of the laminae are smooth (Text-fig. 2A, B). These features show that the wall of the calcified bladder comprises several sheets that are spatially organized like the layers of an onion.
Figure TEXT-FIG. 2.. Axelrodichthys araripensis. A, ‘Josa collection’ deposited at the Laboratoire de Paléontologie, MNHN, photograph of right lateral view showing the posterior chamber of an ossified bladder; scale bar represents 10 mm. B, same specimen as in A, photograph of left lateral view; scale bar represents 10 mm. C, Section of an uncrushed ossified bladder (from Clément, 1999, fig. 6); scale bar represents 5 mm. D, Section of a more or less crushed bladder. UERJ-PMB 143; scale bar represents 5 mm.
Download figure to PowerPoint
In young Axelrodichthys individuals, generally <200 mm TL, the bladder is unossified, being sometimes preserved as a compact mass of diagenetic calcium phosphate (Text-fig. 1B, C). Soft tissues composing these walls were preserved as eodiagenetic replacement by calcium phosphate (Text-fig. 3A, B) during fossilization. This process is linked to early bacterial growth and the decay of soft tissues (Martill 1988; Briggs et al. 1993; Briggs 2003). It occurs rapidly and often predates significant tissue decay. As a consequence, high fidelity replication of tissues may occur, especially when the replacing crystallite size is very small (<1 μm), as it is the case here. Observed in SEM, this phosphatic mass reveals the spatial organization of the walls of the bladder. They consist of several strata of fibres, with a diameter of 10–20 μm; these fibres were probably collagenic during life. In each sheet, the fibres are parallel to each others and their direction changes from an angle of about 90 degrees between two successive strata (Text-fig. 3A). The Jurassic genus Swenzia also presents parallel striations on the superimposed bony plates whose directions seem to have a radiating arrangement on the plate (Clément 2005, fig. 6C; Clément 2006).
Figure TEXT-FIG. 3.. Axelrodichthys araripensis, UERJ-PMB 33. A, general view showing several strata of collagen fibres whose orientation change from one layer to the next one. Scale bar represents 50 μm. B, detail of the insert showing magnified collagen fibres. The fossilization processes have preserved the morphology of the fibres. Scale bar represents 10 μm.
Download figure to PowerPoint
In adult specimens of Axelrodichthys, the smooth ossified laminae of the fossilized ‘bladders’ are relatively well preserved (Text-fig. 2A–D), except when some epigenetic mineralization has occurred, but epigenized laminae are relatively scarce in the whole organ. In the preserved laminae, the histological details are clearly visible. The walls of the bladder are made up by layers of primary cellular bone, separated by layers of limestone matrix (Ga, Text-fig. 4A, C, E). The bony tissue is either pseudo-lamellar or lamellar cellular bone (Text-fig. 4C, E, G). The osteocytes are typically star shaped, with numerous cytoplasmic processes, the main direction of which is orthogonal to the bone lamellae (Text-fig. 4G–I). They are located between two successive lamellae giving them a flat shape. Their average thickness is 15–30 μm, and their main depth is about 3–6 μm. In some thicker areas, the bony layers can be vascularized (Text-fig. 4D–F) and the walls of the cavities can show remodelling bone (= secondary bone) (Text-fig. 4F).
Figure TEXT-FIG. 4.. Axelrodichthys araripensis. Ground cross section (transmitted natural light) in the calcified wall of the bladder. A, section through 11 bony laminae of various thickness separated by the limestone matrix (Ga). The upper and lower surfaces of each lamina look regularly smooth. The sixth and ninth laminae are very thin because of their proximity of their border. Scale bar represents 200 μm. B, section through an epigenized bony lamina. The various stages of the epigenesis process are very clearly because of the concentric ridges (arrows). The osseous organization has wholly disappeared. Scale bar represents 50 μm. C, section shows three bony laminae (1–3) separated by the limestone matrix (Ga). The upper and lower surfaces of each laminae look regularly smooth. The middle lamina is faintly epigenized. Scale bar represents 100 μm. D, section through a bony lamina crossed by several more or less regular artefact cracks (arrow heads). Several medial vascular canals are obvious (arrows). Scale bar represents 50 μm. E, section through two bony laminae: a very thick (below) and a very thin (above) (arrow head), the latter having been cut near its margin. The thickest lamina shows pseudo-lamellar bone and several vascular canals and/or cavities (asterisks) (Ga = limestone matrix). Scale bar represents 100 μm. F, enlargement of a section showing two bony laminae, separated by a very thin layer of matrix (white arrow head). The upper lamina is constituted of primary bone (Bo) only. The second one shows three large vascular cavities (white asterisks) the wall of which is constituted of secondary bone (white and black arrows). The black arrowhead points towards osteocytes lacunae. Scale bar represents 30 μm. G, enlargement of a bony lamina constituted of primary bone showing numerous osteocytes. Scale bar represents 30 μm. H, detail of an osteocyte from the region localized by the black asterisk in Text-figure 3E. Scale bar represents 10 μm. I, Detail of osteocytes from the region localized in Text-figure 3F, showing the canalicles that start from the osteocytes lacunae and are more numerous on the lower surface of the cells (arrows). Scale bar represents 10 μm.
Download figure to PowerPoint
- Top of page
- Material and Methods
In 1849, Williamson (pl. 43, figs 29–30) figured accurate drawings of horizontal and vertical sections of the ossified bladder (×350 magnifications of the ‘internal osseous viscus’) of the Cretaceous coelacanth Macropoma mantelli. Williamson (1849, pp. 463–464) claimed that the walls of this bladder ‘...consisted of true laminated bony tissue’ and that ‘Except in cases of diseased ossification, the existence of an internal thoracic or abdominal viscus, having hard parietes of true bone, is an anomaly, which, as far as I am aware, has hitherto presented no parallel in nature’. One hundred and sixty years after Williamson, our histological observations confirm that the walls of the fossil coelacanth bladder are made of true cellular and vascularized bone. One of the first questions to be placed is how, in adult specimen bladders, bone could be developed from tissues belonging to the walls of the bladder? It is generally considered that the whole swim bladder of teleosts is either an evagination of the gut or that the evagination of the endodermal tissue may invade a group of unconnected mesoderm cells (coming from the splanchnopleura) then forming the outer layers of the bladder (Hoar 1937 in Pelster 2004); therefore, the walls of the bladder have a double origin.
We have no available data about the embryological origin of the bladder in fossil coelacanths neither of the ‘fatty organ’ in Latimeria. However, one may consider an analogy with the second hypothesis (e.g. double embryological origin: endodermal for the breathing epithelium and mesodermal for the walls in part). The presence of an ossification process of the wall of the ‘air bladder’, in physostomid fish (Marshall 1962; Parmentier et al. 2008) is thus understandable, although surprising, but not exceptional. Besides, the mineralized structure of the so-called ‘rocker bone’ situated at the anterior part of the swim bladder of the carapid teleosts (Perciformes) is made by a specialized mineralized conjunctive tissue, possibly of chondroid bone (Parmentier et al. 2008) if not of true bone.
How can we interpret this structure in fossils? What function performed the calcified organ in fossil coelacanths? We can suppose that the ossified organ was filled either with fat for buoyancy control (like the ‘fatty organ’ of Latimeria) or with gas for breathing function. Could this large empty cavity protected by numerous independent bony plates also have had other functions as a specialized bladder for an auditory function or a sound production?
The calcified bladder as an air bladder or a lung?
The position and shape of the bony organ in the fossil coelacanthids suggest a homology with the anatomical complex: vestigial lung (= oesophagal diverticulum) + fatty organ of Latimeria. The anterior part of the ossified organ appears to be situated somewhat more ventrally in the body cavity than its posterior part (Maisey 1986). It is congruent with a ventral position of the anteriormost part of the lung (originating from the ventral side of the oesophagus), posteriorly followed by a dorsal turn up, leading to a dorsal position of the posterior part of the lung in relation to the gut. The fatty organ of Latimeria, enclosing anteriorly the vestigial lung, presents a such dorsal turn up (Text-fig. 5A–D).
Figure TEXT-FIG. 5.. Latimeria chalumnae, specimen MNHN C.20. External morphology of lung and diverticulum. A, drawing of the anatomy of the lung, modified from Robineau (1987). (G, gut; GO, fatty organ; OD, oesophagal diverticulum; Oe, oesophagus; St, stomach). B, three-dimensional reconstruction of the whole anatomy of the fish by means of axial computed tomography, showing the entire gut and the oesophagal diverticulum in the visceral cavity (arrow) in left lateral view. Scale bar represents 20 cm. C, Detail of the oesophagal diverticulum (arrow) and fatty organ (asterisk). Reconstruction of the anterior anatomy of the fish in left lateral view. Scale bar represents 10 cm. D, Detail of the oesophagal diverticulum (left lateral view). Scale bar represents 4 cm. B–D, (Materialise MIMICS v.12.1 software).
Download figure to PowerPoint
The distribution of lungs in vertebrate phylogeny suggests that the latter are primitive osteichthyan structures (Farmer and Jackson 1998), and even possibly for gnathostomes because putative respiratory organs (‘lungs’) have been reported in placoderms (Janvier et al. 2007), that act in synergy with gills. This allows a bimodal respiration (Brainerd 1994b), i.e. functional lungs + functional gills, that results from either an increasing metabolism activity or an increase in the body mass. Such a bimodal respiration was described especially in lungfishes (Grigg 1965a, b; Liem 1986), and polypterids (Brainerd et al. 1989). It should be noticed that lepisosteids and amiids (Farmer and Jackson 1998) and some rare osteoglossomorphs, elopomorphs, ostariophyses and esocoïdes possess respiratory gas bladders rather than true lungs (Graham 1997) and that these organs are dorsal (and not ventral) evaginations of the oesophagus. Currently, it is considered that the presence of a ventral oesophagal diverticulum is plesiomorphic for Osteichthyes. In a general way, this organ does not fossilize like most soft tissue internal organs, except in fossil coelacanthids when the walls of this oesophagus pouch are biomineralized.
The calcified organ as a specialized lung
In Latimeria, the ventral oesophagal diverticulum (Text-fig. 5A–D; Millot et al. 1978) is considered as a vestigial lung, the external conjunctive layer of which being secondarily filled with oil to form the so-called ‘fatty organ’ (Millot et al. 1978); it appears to be an adaptation, a loss of its breathing function and an improvement of the buoyancy of the fish for deep-water habitat, because this species lives at a depth of several hundred metres (Forster 1974; Fricke and Plante 1988; Forey 1998).
When preserved in an uncrushed condition, the calcified lung in Axelrodichthys is usually hollow with a geode-like lining of calcite (Text-fig. 2C). Maisey (1986) suggested that the internal cavity was filled with fatty tissue during life. However, in some of the specimens examined, the lung has not resisted compaction during fossilization, although a lining of calcite is still present (Text-fig. 2D). The presence of fatty tissue within the bladder of Axelrodichthys, if this assumption is correct, lets suppose that fossil coelacanths had an ossified fatty organ (homologous to the soft fatty organ of Latimeria) and probably a vestigial lung (supposedly in the same position and maybe already at same degree of reduction than in Latimeria, although no trace of this soft vestigial lung has ever been recognized in fossils). An ossified organ filled with fatty tissue should have had the same function as the fatty organ of Latimeria that is for the hydrostatic balance.
The bladder in adult Axelrodichthys is a very well-ossified structure, formed by numerous superimposed, multi-layered, bony plates, separated from one another by an unossified connective tissue. This pattern suggests that this structure had a somewhat variable volume, whereby the large plates moved over each other to accommodate certain volumetric changes and acting as bellows. The role of these movable bony layers in a bladder filled with fat would be difficult to interpret, and it makes more sense that the bladder was filled with gas rather than oil or fat. If so, the volumetric variation was probably linked with elastic and muscular properties of the lung’s walls, facilitating the ventilation of breathing gas on the respiratory epithelium.
Accepting that the anatomical complex (vestigial lung + fatty organ) is homologous to the bony organ of Axelrodichthys, one may suppose that these structures in the fossil taxon should be active as an interface for gas exchange, like the lung of extant lungfish Neoceratodus (Günther 1871; Grigg 1965a, b; Thomson 1968; Burggren and Johansen 1986) and extant polypterids (Horn and Riggs 1973; Brainerd et al. 1989; Farmer and Jackson 1998). It is also known that the dependence upon aerial respiration, in the extant Protopterus and Lepidosiren, increases dramatically as their body mass increases (Liem 1986). This does not imply that Axelrodichthys, as all other fossil coelacanths with ossified lungs, was necessarily a strictly air breather but that in some circumstances, it could effectively use the lung to support more intense activity. Contrary to extant coelacanth species, Axelrodichthys lived in lagoonal or epicontinental shallow marine environments and was probably subjected to adverse conditions of shallowing or hypoxia.
The lung was closely surrounded by ossified plates. Connective tissue connexions between the external surface of the lung and the internal side of the laminae permitted the increase in the pulmonary volume to swallow the air when the muscles attached to the external surface of the laminae were contracted. The protecting bony laminae should not have been a problem for the gas exchanges if we assume that the folded mucous membrane was inside the ossified body. Arteries and veins could have most probably passed between the laminae (Text-fig. 4A, C) or through the bony lamina itself because large vascular cavities and canals have been observed within the laminae (Text-fig. 4F).
Air respiration necessitates the differentiation of an accurate ventilation system. In the extant Neoceratodus, exhalation of air is effected by contraction of the smooth muscle components of the lung, assisted by its natural elasticity provided by elastin fibres present in both connective tissue and smooth muscle (Grigg 1965a). It is interesting to note that the Mesozoic coelacanth Diplurus, one of the rare well-preserved coelacanths lacking a calcified organ, also shows greatly elongated pleural ribs (Schaeffer 1948). According to Schaeffer (1952, p. 48), ‘the functional significance of this elongation is obscure’. If we postulate that fossil coelacanths had a functional lung, we can assume that these elongate pleural ribs played the same role as the ribs of tetrapods, i.e. the contraction of the inter-pleural rib muscles enlarged the pulmonary cavity and the reduced air pressure in the cavity causes air to enter the lung. Such a ventilation system, homoplasic with that of the tetrapods, could explain the absence of calcified lung in Diplurus. The primitive actinopterygian Polypterus uses recoil respiration (Brainerd et al. 1989). The walls of the lungs are partly composed of a muscular layer especially in the posterior part where it is thicker and it could occur in ventilation function (Poll and Dewattines 1967). This extant biological model also uses the deformation of the stiff integument, constituted of interlocking rhomboid scales, for sucking air into the lung (Brainerd 1994a). Conversely, when the bladder of Axelrodichthys inflated during air inhalation, the osseous plates could store the tension strength and released it as an elastic energy facilitating air expulsion. So, the bony walls of the lung in Axelrodichthys (and in other coelacanthids with an osseous bladder) can play the same function as does the integument of polypterids for air ventilation.
The calcified organ as an auditory organ
A large empty cavity protected by numerous independent bony plates could have had an auditory function. In fossil coelacanths, the anterior extension of the ossified organ extends under the opercle and seems to reach the back of the head in some specimens. Such anterior position may suggest some possible relation with the inner ear, as in many teleosts (Tavolga 1971). The anteriormost part of the bladder could have joined the skull as in clupeomorphs. Some bones (equivalent to the Weberian apparatus) or modified vertebrae could have connected the bladder to the inner ear as in otophysans and gonorynchiforms (Lecointre and Nelson 1996). Furthermore, both gonorynchiforms and otophysans present a swim bladder divided into anterior and posterior chambers, a morphology also seen in the ossified bladder of Axelrodichthys. More importantly, the coelacanths present a communication, via the canalis communicans, between the two inner ears, as seen in the ostariophysans (Millot and Anthony 1965; Bernstein 2003). In Latimeria, the canalis communicans emerges from the vestibule, at the transition point of the saccule and lagenar recess, passing backward and medially to complete a semiloop between the two inner ears. The commissure of these two canals is obvious on the posterior side of the skull, between the foramen magnum and the notochordal canal. According to Millot and Anthony (1965), a short median diverticule issued from the commissure is posteriorly directed. These authors proposed a regression of the auditory apparatus on the basis of palaeontological data and a possible link with the calcified bladder of the fossil coelacanths Undina and Laugia. The posterior wall of the skull of fossil coelacanths presents foramina for the canalis communicans and a transverse groove for the communicans commissure. This communicans commissure is even much more developed in fossil coelacanths (especially Palaeozoic genera such as Diplocercides) than in Latimeria. This character most probably developed early in the evolutionary history of coelacanths (Bernstein 2003).
In Latimeria, a membrane (‘innervated end organ’) covering the foramen at the sacculo-lagenar orifice suggests that the canalis communicans is a perilymphatic duct rather than an endolymphatic duct as in some teleosts (Bernstein 2003). This membrane in Latimeria might be responsive to very small pressure changes between the endolymphatic cavity and the perilymphatic duct. An auditory function of the ossified bladder is thus possible if we assume that the canalis communicans was linked to the ossified organ by any kind of soft tissue: ligament, canal, anterior expansion of the epithelium of the ossified organ or posterior expansion of the communicans commissure (in this latter hypothesis, the median diverticule of the communicans commissure in Latimeria could thus be considered as a vestigial remain of this expansion).
Such a connection might have transmitted vibrations from the bladder to the labyrinth of the ear. In fossils, the wall of the bladder, composed of closely set independent rounded plates, might have amplified the sensitivity to the difference of external pressure as well as increased the amplitude of the vibrations. A sound-transmitting apparatus from the anterior part of the ossified bladder and the canalis communicans was not necessarily very long if we assume that the ossified bladder was anteriorly extended, as seen in fossils, but it had to bypass the notochord. However, this auditory function is highly hypothetical, because no trace of Weberian-like apparatus or of modified vertebrae have been recorded so far in fossil coelacanths; however, a nonfossilized link cannot be totally excluded.
The calcified organ as a sound production organ
Such a cavity, supposedly filled with gas, could also have played a role as a resonance chamber. In this case, the independent plates might have been shaken the ones on the others, creating a rattle noise, the cavity having the same function as a resonator. Some teleosts produce sounds (Marshall 1962), generally for agonistic and/or courtship behaviour (Tavolga 1971; Fine 1997). ‘Drums’ and ‘Croakers’ (Sciaenidae) have special muscles attached to their swim bladder for sound production (see Schneider 1962; Tavolga 1971); some catfishes also possess a modified swim bladder for sound production (Fine et al. 1997; Fine and Ladich 2003), and they use specialized muscles on the upper surface of an elastic bony spring to create vibrations from the swim bladder. Moreover, certain species among the carapids are able to generate sounds thanks to the so-called ‘rocker bone’ a specific mineralized formation situated at the anterior part of the swim bladder (Rose 1961; Courtenay and McKittrick 1970; Parmentier et al. 2003). Extant coelacanths seem to have a remarkable social behaviour (Fricke et al. 1991). Sound communications could have played a role in the early coelacanth communities.
- Top of page
- Material and Methods
The abdominal ossified organ of fossil coelacanths and the complex (oesophagal diverticulum + fatty organ) of Latimeria are most probably both of pulmonary origin. Although this homology between the ossified organ of †Axelrodichthys and the pulmonary apparatus of Latimeria seems to be acceptable, it is impossible to determine whether the ossified organ was filled (1) with oil (as previously assumed) + an anterior vestigial lung, (2) with gas + an anterior vestigial lung and (3) with breathing epithelium + ventilated air + a well-developed and functional lung. Assuming that the ossified organ was filled with gas, coelacanths would have had to inhale air at the surface as do the extant Dipnoi. If its function is most probably hydrostatic in the living coelacanth Latimeria, it is quite improbable in fossil coelacanths. Actually the Bernstein’s (2003) hypothesis of an adaptation to ground-dwelling or deep-sea fishes to better withstand the high water pressure is not reliable with the usually low-depth palaeoenvironments of fossil coelacanths. Here, we favour the air-breathing hypothesis, and then the calcified organ is a remnant of a specialized functional lung; although other hypotheses such as auditory, sound production and mineral elements storage could be possible, certain of them being able to act in synergy. Such multifunctionality is known in some neotropical silurids: buoyancy, audition, sound production. As a matter of fact, a breathing function of the ventral oesophagal pouch is a plesiomorphic character and complements the gill breathing. In the coelacanthids with a bony bladder, the ossified walls represent a specialization to improve breathing ventilation.
Acknowledgements. We thank Marc Herbin, Marie-Madeleine Loth, Lúcio P. Machado, and Christiane Chancogne for all their help during this project. We especially thank John G. Maisey, Jésus Alvarado-Ortega, and an anonymous referee for their carefully reviewing of this manuscript as well as D. Martill for improving the English style. P.M.B.’s research has been partially supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), a fellowship from the Department des Milieux et Peuplements Aquatiques – Muséum national d’Histoire naturelle, and a PROCIENCIA research grant.