Latimeria chalumnae : University of Tübingen, Lehrstuhl für Spezielle Zoologie: CCC 162.11, 351mm pup, 40μm sections, Heidenhahn’s Azan stain. Slides are numbered rostral to caudal and referred to with the prefix T.

American Museum of Natural History: 32949. CCC 29.1, 303 mm pup, 50 μm sections, Weigert’s haematoxylin and van Gieson’s picro-fuchsin stain. An account of the history and organisation of this section series is provided by Northcutt and Bemis (1993). Slides are numbered caudal to rostral with the prefix RC.

Sphenodon punctatus : specimen and staining details as in Ung and Molteno (2005): sequences from a serially sectioned hatchling head *.

Neoceratodus forsteri : photographs of multiple section series as listed in Ziermann et al. (2018)*.

Neoceratodus forsteri : formalin fixed specimens, details in Diogo et al. (2016).

Necturus maculosus : alcohol specimen, commercial supplier; author’s collection, SVL 90mm.

Latimeria chalumnae : Scripps institute of Oceanography, CCC 88, MRI scan accessed from digitalfishlibrary.org, scan details at * http://www.digitalfishlibrary.org/library/ViewSpecies.php?id=124

Muséum Nationale d’Histoire Naturelle (MNHN), Paris. CCC94. Scan details in Mansuit et al. (2021), scan accessed at https://www.morphosource.org/concern/media/000113919?locale=en.

Neoceratodus forsteri : California Academy of Sciences SU 18139, specimen and scan details * http://www.digitalfishlibrary.org/library/ViewSpecies.php?id=271.

Protopterus aethiopicus : California Academy of Sciences CAS 46377, specimen and scan details * http://www.digitalfishlibrary.org/library/ViewSpecies.php?id=237.

Latimeria chalumnae : MNHN CCC27, phosphomolybdic acid staining, staining and scan details in Dutel et al. (2013) *.

Polypterus delhezi : phosphotungstic acid (PTA) staining, specimen, staining and scan details in Molnar et al. (2017).

Amia calva : stage 29 larva, PTA staining; specimen, staining and scanning details in (Funk et al. 2021)*.

Salamandra salamandra : https://www.morphosource.org/media/000345888 *

Necturus maculatus : https://www.morphosource.org/concern/media/000346019?locale=en *

Plethodon cinereus : https://www.morphosource.org/concern/media/000345972?locale=en *

Chiloscyllium punctatum : PTA staining, specimen and scan details in Coates et al. (2019). *

Polyodon spatula : PTA staining, specimen and scan details in Metscher (2009). *

Latimeria chalumnae : fetus, South African Institute for Aquatic Biodiversity, CCC 202.

pup, MNHN, CCC 29.5.

pup, Zoologische Staatssammlung München (ZSM) CCC 162.21.

Scan details for the above 3 Latimeria series are given in Dutel et al. (2019), and scans were downloaded from http://paleo.esrf.eu.

Latimeria chalumnae : ZSM, CCC 162.21, pelvic girdle and fin, scan details in Mansuit et al. (2021). *

Scyliorhinus canicula : specimen and scan details in Dearden et al. (2021). *

Callorhinchus milii : specimen and scan details in Dearden et al. (2021). *

The PTA stained microCT of Polypterus delhezi is available on request to the author.

The spiracular organ is found on the medial side of the spiracular chamber, close to the orbital artery and enclosed in a recess in in the ‘afacial’ or ‘affacial’ eminence of Jarvik (1980) on the otic shelf of the oto-occiptial moiety of the braincase (Fig. 2), in which a surface impression is evident in skeletal preparations (Fig. 3). The afacial process bears a toothplate in Latimeria, and Jarvik (1965) refers to it as dermal bone applied to the endochondrally ossified otic shelf; the afacial process does have a more delicate trabecular structure than the ossification in the otic shelf. Histologically the spiracular organ conforms with the appearances in elasmobranchs (Barry and Boord 1984), in which it has elongated cells with basal nuclei and pale-staining cytoplasm (on haematoxylin and eosin staining), with an amorphous material occupying part of the lumen, formed by the cilia. Cytological detail is lacking in the relatively thick sections and aging staining of both the Latimeria section series, together with some loss of structure from delayed fixation after death, but the residua of these features remain visible (Fig. 3). The spiracular organ forms a narrow tube that is measured at 0.75mm length in CCC 29.1. With knowledge of the spiracular organ’s location, its nerve, a branch of the otic lateral line nerve, can be traced further than was evident to Northcutt and Bemis (1993). The nerve passes out from the braincase alongside the combined facial and anteroventral lateral line trunk, passes ventral to the jugular vein and then traverses the bony interface between the otic shelf and the afacial process, as does the nearby orbital artery. The last, intraosseus, section of the course of the nerve cannot be confidently seen on the histological sections or synchrotron CT series, but its line of passage can be inferred (Fig. 4). On one side of one of the specimens used here (CCC 29.5), the nerve follows an alternative or aberrant course, which actually makes it easier to follow: it leaves the cranial cavity through a separate foramen lying dorsal to the exit of the facial nerve, passes dorsal to the jugular vein and then ventrally on the lateral side of the vein the enter the dorsal aspect of the otic shelf-afacial process interface and travels directly ventrally through this to the spiracular organ. Fig. 4 includes both paths of the SO nerve, and the whole distribution of the otic lateral line nerve.

The tissue-stained CT series of CCC27 was made after phosphomolybdic acid staining; this is very similar to the effects produced by phosphotungstic acid (PTA), which allows neuromast and related sensory structures such as the inner ear to become selectively enhanced on microCT scans (Schulz-Mirbach et al. 2013). In the CCC27 series the neuromast tissue of the SO is evident as an intense plaque of enhancement at the identical position to that noted histologically (Fig. 5 A, B). On the synchrotron series the indentation (or fossa) for the SO on the rostral surface of the afacial process can be identified; the slice image of the fetus CCC 202 is shown in Fig. 5 C, D: the afacial process is not fully developed at this stage and the opening of the SO faces laterally rather than rostrally.

Figure 2. 

Latimeria chalumnae CCC29.1, spiracular organ in histological sections,. Slide RC 657 in progressive magnifications A–C Abbreviations: jv, jugular vein; VIIp, palatine ramus of facial nerve; oa, orbital artery; ap, afacial process; so, spiracular organ. Scale bar: A=10mm

Figure 3. 

Latimeria chalumnae , drawings reproduced with permission from Millot and Anthony (1958). A, B (planche XXV) (rostral view of oto-occipital braincase) showing afacial process on otic shelf, and fossa for the spiracular organ. C, D (planche XX) lateral view showing afacial process.

Figure 4. 

Latimeria chalumnae CCC 29.5, 3D reconstructions from synchrotron tomography, showing the otic lateral line nerve and its distribution. A anterolateral view; B dorsal view. Relationships among spiracular organ, orbital artery and palatine nerve are observed in A. Nerve branches too small to appear in the reconstruction are indicated with dashed lines. Abbreviations: OLL, otic lateral line; OLLn, otic lateral line nerve; OLLn,a, anterior ramus otic lateral line nerve; OLLn, p, posterior ramus; Olln, s, spiracular ramus of otic lateral line nerve; jv, jugular vein; so, spiracular organ; VIIp, palatine ramus of facial nerve; ADLLn, anterodorsal lateral line nerve; Vp, profundus ramus of Vth cranial nerve; Vs, sensory ramus of V nerve; Vm, motor ramus of V nerve; VII+ AVLLn, facial + anteroventral lateral line nerves. Terminology after Northcutt and Bemis (1993).

Figure 5. 

A, B Latimeria chalumnae adult, CCC 27, phosphomolybdic acid stained CT scan, demonstrating uptake of the stain in the spiracular organ. C, D Latimeria chalumnae fetus, CCC 202, showing spiracular organ. The afacial process is not yet developed. Abbreviations: ap, afacial process; fso, fossa for spiracular organ; s, spiracular cavity; so, spiracular organ. Scale bars: A=50mm, C=1mm.

Scyliorhinus cannicula and Chiloscyllium punctatum: in these image series, the SO is seen at the base of the spiracular chamber on its medial wall (Chiloscyllium, Fig. 6 A, B). The orbital artery and its lateral branch are in close apposition to the SO.

Polypterus : In the PTA microCT series of the head of Polypterus delhezi, no neuromast tissue could be found within the spiracular chamber, but the expected enhancement of superficial and canal neuromasts, and the sensory structures on the snout known as the ampullae of Lorenzini, is confirmed, indicating absence of the SO. However, in the developing Polypterus senegalus a placode is seen within the spiracle in sections of stage 26 and at the location of the spiracle in whole-mount preparations of stage 29 (Diedhiou and Bartsch 2009), suggesting that the SO originally forms and is lost in later development (Robert Cerny, personal communication). In microCT series of Amia the SO is present on the medial wall of the spiracular chamber dorsally (Fig. 6C), consistent with previous reports (Goodrich 1930).

Figure 6. 

A Chiloscyllium punctatum (bamboo shark), phosphotungstic acid (PTA) stained microCT section with spiracular organ within a diverticulum close to the oral opening of the spiracular canal. B Chiloscyllium punctatum, adjacent section to A, showing orbital artery and its branch in apposition to spiracular organ. C Amia calva (bowfin), PTA stained microCT section, with pseudobranch ventrally and spiracular organ dorsally within the spiracular canal. Abbreviations: so, spiracular organ; oa, orbital artery; oa, l, lateral branch of orbital artery; ps, pseudobranch; s, spiracular canal. Scale bars: A=5mm, C=1mm.

In Polyodon the spiracular organ is evident within a lateral diverticulum of the main spiracular chamber.

Callorhinchus and Sphenodon are mentioned by O’Neill et al. (2012) as unconfirmed instances of the SO, based on single reports. In the Callorhinchus synchrotron CT series, the residual, enclosed spiracular cavity is seen in a position identical to that reported by de Beer and Moy-Thomas (1935), immediately ventral to the lateral edge of the palatoquadrate at the level of the dorsal elements of the first branchial arch. In Sphenodon, a cavity lined by cuboidal epithelium is seen in a position identical to that described by Werner (1963), dorsal to the columella auris and adjacent to the stapedial artery. The section series available is incomplete in this area such that the whole cavity could not be followed, and the neuromast tissue demonstrated by Werner was not seen; all other features of Werner’s description are, however, confirmed. In the Neoceratodus section series the SO is as described by Bartsch (1994) as a neuromast structure within a very small, enclosed remnant of the spiracular canal; an artery wraps closely around the SO here, and is identified by Bartsch as the efferent mandibular artery.

The UB in Latimeria is seen as a shallow pouch or diverticulum from the ventral mucosa of the posterior pharynx, on the left side only. Findings are the same in both section series: a narrow lumen leads in from the mucosal surface to glands in small acini, mostly with their own lumen and not in obvious continuity with the main lumen of the pouch (Fig. 7). The total antero-posterior length of the UB in CCC 29.1 is 1.0cm, and in CCC 162.11is 0.84cm. In comparative material, the UB is seen as with very similar morphology in the posterior pharynx on the left side in the synchrotron series of Scyliorhinus.

Figure 7. 

Latimeria chalumnae , CCC 29.1, histological sections. A slide RC893; B magnified location of ultimobranchial gland in A; C RC885; D RC 879 (more rostral sections), showing gland tissue as terminal branches of the central duct of the gland. Abbreviation: ub, ultimobranchial gland. Scale bar A=10mm.

M. cucullaris

In histological series (Fig. 8 A–E), m. cucullaris is found as a thin and discontinuous sheet in the expected position of this muscle, dorsal and exterior to the levatores arcuum branchialium; in its cranial end the muscle is represented by a thin fibrous sheet with a few muscle fibres radiating caudally from its origin on the ‘processus post-oticus’ of Millot and Anthony (1958) and Jarvik (1954) of the braincase, and the body of the muscle becomes more apparent toward its insertion on the pectoral girdle at the dorsal end of the cleithrum. The synchrotron CT images at the same ‘pup’ stage have findings identical to the histological series (Fig. 8 F).

The MRI imaging used here is the same as that used by Sefton et al. (2016), but different conclusions are reached. Again, the m. cucullaris is seen (most clearly in horizontal plane image reconstruction, rather than axial or sagittal) as a narrow sheet in the typical position of this muscle, as just described in the histological slides. Muscle tissue on this imaging stops short of the bony pectoral girdle, implying a flat tendon connecting these, or insertion of the muscle into the epaxial fascia. 3D reconstructions from the fetus synchrotron imaging and the adult MRI are seen in Fig. 9.

In the synchrotron series the m. cucullaris is clearly seen again more prominent caudally, and is relatively larger in the ‘fetus’ (CCC202) (Fig. 9 A,B) than in the ‘pup’ (Fig. 8F) or adult (Fig. 9 C,D) specimens.

The nerve supply of m.cucullaris could be traced in the synchrotron scan of CCC29.5: a slender branch arises from the vagus just distal to the second branchial branch, and passes dorsally in the space between the epaxial muscles and the levatores arcuum, in parallel with slender branches from the posterior lateral line nerve to the posterior lateral line. Innervation of m. cucullaris in fishes is not well described, but the situation in Latimeria as described here is very similar to the innervation of m.cucullaris in ganoid fishes (Norris 1925).

The levatores arcuum are displayed in both fetus and adult reconstructions in Fig. 9: levator arcus branchialis 5, as described by Millot and Anthony (1958), is a small muscle which arises from the pectoral girdle and meets and is partly enveloped by levator arcus branchialis 4 as both converge on the dorsal aspect of the epibranchial plate at the dorsal ends of ceratobranchialia 3+4. There is, however, a discrepancy between the text and figures of Millot and Anthony (1958): the text is as just described, but the figure (planche VII) shows levator arcus branchialis 5 as a more elongated structure inserting on the dorsal tip of the 5^th ceratobranchial. The latter situation was not seen in any of the specimens used here.

Figure 8. 

Latimeria chalumnae (A–E) CCC 162.11, histological sections. A (T971), overview; B magnified section of dorsal branchial region showing m. cucullaris. C (T1139); D (T1143); E (T1216): progressively more caudal representative sections, E being at the level of the pectoral girdle. F Latimeria chalumnae CCC 29.5, synchrotron CT section at a level similar to C. Abbreviations: cuc, m. cucullaris; ao, distal fibres of adductor operuli; lab 3+4, common belly of levator arcus branchialis 3 and 4; lab 3, 4, 5, levator arcus branchialis 2–5; o, opercular chamber; omo, m. omohyoideus; X, vagus nerve; PLLn, posterior lateral line nerve; PLL, posterior lateral line; pg, pectoral girdle. Scale bars: A=10mm; F=1mm.

Figure 9. 

Latimeria chalumnae , 3D reconstructions: A and B: CCC 202 (fetus), synchrotron CT, and C and D, CCC 88 (adult), MRI scan, showing relationships of m. cucullaris to branchial levators and adjacent structures. Colours are similar in A, B and C, D; in C, D the epibranchial cartilages are segmented separately from the ceratobranchials, all are together in A, B. Nomenclature for the muscles of the hyomandibula and operculum in C, D as in Millot and Anthony (1958); in the scan images these three muscles form a continuous broad sheet. Abbreviations: pr-oto, location of the processus post-oticus of the braincase; br, branchial cartilages, cb, ceratobranchials, ch, ceratohyal; eb, epibranchials; pb1, suprapharyngobranchial 1; h, hyomandibula; ih, interhyale; s, sympletic; pg. pectoral girdle; cuc, m. cucullaris; omo, m. omohyoideus; lab 1–5, levator arcus branchialis 1–5 (lab 3 and 4 have a common dorsal belly which gives rise to separate muscles close to the dorsal arches); ah, m. adductor hyomandibulae; lo, m. levator operculi, ao, m. adductor operculi. Scale bars: A, B=1mm; C, D=50mm

Lateral elevator of the pelvic fin

Musculus levator lateralis of the pelvic fin is a thin, short sheet of muscle passing between the hypaxial body muscle and the antero-lateral border of the proximal fin (Fig. 11A), and was identified in MRI sections of Latimeria as described by Diogo et al. (2016). This is almost identical to a muscle in this position in Neoceratodus, the abductor dorsolateralis of Diogo et al. (2016), designated the superficial ventrolateral abductor of Young et al. (1989) and Boisvert et al. (2013). Latimeria dissection material available to us for Diogo et al. (2016) was damaged in this area, but Huby et al. (2021) could not confirm such a muscle in dissected material and doubted its existence. In the synchrotron series of the pelvic fin, a small sheet of muscle is clearly seen in this position (Fig. 11A) and it can also be identified on a different MRI series made available by Mansuit et al. (2021) (CCC162.21) from their study of skeletal elements of the fin. A specimen of Neoceratodus was dissected here and confirmed the presence of this small muscle (Fig. 11B).

Figure 10. 

Neoceratodus forsteri , MRI scan: A–C reconstruction of muscles connected to the cranial rib in A caudomedial; B medial; and C lateral views. D slice image to demonstrate some of the structures segmented in A–C. E Latimeria chalumnae CCC 202, synchrotron tomography, slice image of segmentation to show muscle proposed as m. omohyoideus. Scale bars A–D=10mm, E=1mm. Abbreviations: ac, anocleithrum; cbc, cardiobranchial cartilage; cl, clavicle; cm, cleithrum; cor, m.coracomandibularis; cr, cranial rib; omo, m. omohyoideus; ps, parasphenoid; rs, m. rectus abdominis, superficial lamina; rd, m. rectus abdominis, deep lamina; sh, m. sternohyoideus.

Figure 11. 

A Latimeria chalumnae CCC162.21, synchrotron CT section through base of pelvic fin with muscles labelled, confirming the presence of m. levator lateralis (Diogo et al. 2016), comparable to the m. abductor dorsolateralis of Neoceratodus (Diogo et al. 2016) (‘superficial ventrolateral abductor’ of Young et al. 1989, Boisvert et al. 2013). B Neoceratodus forsteri, superficial dissection of body wall at origin of pelvic fin, lateral view. Scale bars: A=5mm, B=1cm.

The identification of the spiracular organ of Latimeria has a number of interesting implications. The distribution of the SO across vertebrates is reviewed by O’Neill et al. (2012); a simple vertebrate phylogeny with distribution of the SO updated from O’Neill et al. (2012) with the data presented here is given in Fig. 1. Among fishes this structure is present in elasmobranchs, non-teleost actinopterygians with the exception of bichirs, and absent in teleosts. The SO is also present in holocephalans and lungfish, is spite of the early closure of the spiracular cleft. The function of the SO (or paratympanic organ of tetrapods) is unknown, but is presumed to be mechanosensory in elasmobranchs, where its relation to the hyomandibula has suggested a function in positional information of that cartilage (Barry and Boord 1984). In avians a role in barometric or altimetric sensation has been proposed but not confirmed experimentally (Neeser and von Bartheld 2002). From the distribution of the SO across vertebrate phylogeny, it is reasonable and certainly more parsimonious to assume that the SO is a plesiomorphic feature among vertebrates (Gardiner 1984). The significance of the multiple losses of the SO is difficult to establish in the absence of functional information; investigative tools such as identification of placodes are now available, but have not yet been applied across a wide spectrum of vertebrates with focus on the SO placode, so the preliminary findings cited here in Polypterus are of interest.

A close spatial association between the SO and the orbital artery or its homologue, the stapedial artery of tetrapods, has not been noted previously. The orbital artery is a branch of the paired dorsal aorta, and passes lateral and inferior to supply the orbit and facial structures. Here this association is observed in Latimeria, Chiloscyllium, and Sphenodon; in Neoceratodus, the named orbital artery is a different structure to that of other fish, and the original territory of the typical orbital artery is taken over by branches from the ventral aorta. In neognathous birds, a close apposition of the paratympanic organ and stapedial artery is seen (Starck 1995). The situation in crocodylians is not as clear; combining the findings of Neeser and von Bartheld (2002) and Kundrát et al. (2009), the stapedial artery passes adjacent to the dorsal recess of the tympanic cavity, where the paratympanic organ lies. The significance of this relationship across these disparate taxa is not clear, and could be examined more closely with further data. Whether there is a functional association is an interesting question: a role in cardiovascular regulation could be proposed.

Recognition of the SO in Latimeria could throw light on fossae and foramina in the spiracular area in extinct taxa. The afacial process of Jarvik (1980) is clearly present in Diplocercoides (Nesides), a Devonian coelacanth that lies remote from Latimeria on all estimations of coelacanth phyogeny (Toriño et al. 2021). In Diplocercoides, Jarvik noted a groove or fossa in the anterior face of the afacial process and proposed this housed the orbital artery, which it may indeed have done, but perhaps more likely contained the spiracular organ. In Eusthenopteron, Jarvik (1980) proposed a spiracular organ in a much more dorsal position (Jarvik 1980; Fig. 206) than that identified here in Latimeria; Jarvik was using Amia as a baseline for a plesiomorphic osteichthyan fish, and in Amia and in other non-teleost actinopterygians the spiracular organ is found in a considerably more dorsal position than the Latimeria SO (Goodrich 1930; Fig. 6C). Latimeria would have been a better model, at least in this respect, for Jarvik’s hypotheses about Eusthenopteron. Latimeria is closer phylogenetically to Eusthenopteron (in Fig. 1 it would appear on a branch from the stem leading to tetrapods), and the skull structure is much more similar, with the two-part cranium and intracranial persistence of the notochord. Jarvik (1980) was not clear about which phylogenetic hypothesis he was entertaining, but was clearly supporting Amia as a descendent of palaeoniscid fish and thus having arisen close to the base of the actinopterygian tree. Several small fossae are seen in Jarvik’s figures of Eusthenopteron on the medial and posterior aspect of the spiracular chamber at a similar level to the SO of Latimeria, and it would be interesting to investigate these with modern imaging.

The identification of the nerve to the SO in Latimeria, a discrete branch of the otic lateral line nerve, also has interest in the study of other taxa; I will offer some examples. As pointed out by Forey (1998) and Dutel et al. (2012), there are a number of foramina or fossae on the lateral face of the otic shelf in fossil coelacanths, and the identity of these can be difficult to work out. Among tetrapodomorphs, a foramen on the otic shelf in Megalichthys was designated by Romer (1937) as ‘hypotic branch of the facial nerve’, and this was followed for a similar foramen in the osteolepidid Gogonasus andrewsae by Long et al. (1997). However, the term ‘hypotic nerve’ is difficult to place among other taxa: none of the accounts of the cranial nerves in fish either by the classical anatomists or in more recent investigations have a facial nerve branch of this name or location. Allis (1934) considered ‘hypotic’ as an archaic term for ‘otic’ in the context of lateral line nerves; an otic lateral line nerve branch in this territory would presumably only be supplying a spiracular organ, as the otic lateral line in these taxa is in the usual location, dorsolateral on the cranial roof. The ‘hypotic’ nerve foramen in these taxa could in fact be for the nerve to the SO, or even the SO itself. A search for possible osteological correlates of the SO in other tetrapodomorphs and in osteichthyans may also be rewarding; just to give one further example, blind pits within the spiracular cavity of the early osteichthyan Ligualepis (Basden and Young 2001; Clement et al. 2018) could be considered with the SO in mind.

The location and morphology of the ultimobranchial body (or organ) as demonstrated here in Latimeria is very similar to that described for chondrichthyans, and the structure and location of the gland closely resembles that of elasmobranchs in which the basibranchial copula has not occupied most of the space between the last hypobranchials, such as Squalus (Camp 1917) and Torpedo (van Bemmelen 1885), and as seen here in Scyliorhinus. Persistence of the UB on the left side only is characteristic of all groups of chondrichthyans and lungfish (Shinohara-Ohtani and Sasayama 1998). The location of the UB in Latimeria is as expected from its phylogenetic position: it originates as a small outpouching of the pharynx in relation to the afferent artery of the last branchial arch, and remains in that location after development, as is the case in chondrichthyans, non-teleost actinopterygians and lungfish. The location of the UB in a range of fish, including Latimeria, can be correlated with hox gene retentions and deletions that occur after the multiple whole-genome duplications that occur over the various fish lineages (author’s observations, manuscript in preparation). No particular functional correlate of UB location can be deduced, and in fact the function of calcitonin itself is far from clear across the vertebrate spectrum (Hirsh and Baruch 2003).