Head anatomy of a lantern shark wet‐collection specimen (Chondrichthyes: Etmopteridae)

In this study, we apply a two‐step (untreated and soft tissue stained) diffusible iodine‐based contrast‐enhanced micro‐computed tomography array to a wet‐collection Lantern Shark specimen of Etmopterus lucifer. The focus of our scanning approach is the head anatomy. The unstained CT data allow the imaging of mineralized (skeletal) tissue, while results for soft tissue were achieved after staining for 120 h in a 1% ethanolic iodine solution. Three‐dimensional visualization after the segmentation of hard as well as soft tissue reveals new details of tissue organization and allows us to draw conclusions on the significance of organs in their function. Outstanding are the ampullae of Lorenzini for electroreception, which appear as the dominant sense along with the olfactory system. Corresponding brain areas of these sensory organs are significantly enlarged as well and likely reflect adaptations to the lantern sharks' deep‐sea habitat. While electroreception supports the capture of living prey, the enlarged olfactory system can guide the scavenging of these opportunistic feeders. Compared to other approaches based on the manual dissection of similar species, CT scanning is superior in some but not all aspects. For example, fenestrae of the cranial nerves within the chondrocranium cannot be identified reflecting the limitations of the method, however, CT scanning is less invasive, and the staining is mostly reversible and can be rinsed out.

Generally, non-invasive methods are prefered for museum collection specimens over dissection. One option for non destructive anatomical examination of soft tissue of museum specimens is computed tomography (CT). Since hard tissue, for example calcified cartilage elements or teeth, are easily distinguishable using CT scanning, soft tissue poses a larger challenge, especially the differentiation of adjacent less calcified cartilage and soft tissue, as both tissue types have similar absorption spectra (Jeffery et al., 2011). Here, we overcome these issues by using diffusible iodine-based contrast-enhanced computed tomography (diceCT). This approach allows for distinguishing different tissue structures and types including challenging transitional areas using the absorption rate of iodine (Metscher, 2009). In this study, we present a non-destructive approach for analysing the cranial anatomy of hard-and soft tissues in a lantern shark species for the first time and compare results to dissection-based anatomical studies of Etmopteridae. New data on the anatomy of the cranial region of Etmopterus lucifer (Jordan & Snyder, 1902) are provided showing the structure and organization of skeletal and muscular elements as well as organs with cerebral innervation such as the ampullae of Lorenzini.

| Materials
The Etmopterus lucifer (ZSM-30813) specimen used in this study is housed at the Bavarian State Collection of Zoology and was collected in 1979 SE of New Zealand (depth 700-720 m) by F. Pfeil. The specimen was fixed in 4% formaldehyde and is preserved in 70% EtOH solution.

| Staining
Only the head of the specimen was used in this study, therefore the specimen was decapitated with a scalpel. The head was immersed in a 1% (w/v) ethanolic iodine solution following the protocol by Metscher (2009). For the staining solution, 1 g I 2 was dissolved in 100 ml denatured (MEK) 100% ethanol.

| CT acquisition and imaging procedures
Three CT acquisitions of the head were performed. One scan was done before immersing in the staining agent as a reference to retain the contrast of the calcareous skeletal elements. A second scan was performed after treatment with the iodine solution for 72 h. The final scan was performed after a total staining duration of 120 h. The sample-in ethanol or in staining solution-soaked condition but surrounded by air-was mounted in a plastic container tightly closed with a lid to prevent drying out and stabilized by pieces of Styrofoam to prevent movements during the scanning process. CT acquisition was performed with a Phoenix Nanotom m (GE Sensing & Inspection). For scanning parameters, see Table 1. The final scan was carried out as 'multiscan' with three overlapping z-direction positions.
The CT reconstruction was conducted using phoenix datos|x2 software. The resulting 16-bit volumes were converted to 8-bit with VGStudio Max® (Volume Graphics) software at adjustment of the histogram. Further 3D graphic procedures (merging of the multiscan, orientation, segmentation, visualization, etc.) were performed with Amira® software (v. 6.4.0, Mercury Computer Systems, Inc.). The three portions of the multiscan were merged after manual registration and subsequent automatic refinement of the alignment. The resulting data set (2394 × 1856 × 4670 voxels) was resampled to half the original resolution. The unstained (Etmop-lucif-ZSM-30813) and the 120 h stained (Etmop-lucif-ZSM-30813_ai2mu) data sets were elastically co-registered to clearly depict the position of the calcified elements in relation to the stained soft tissues (warped together).
The 120 h stained data set was used for segmentation and subsequent surface rendering of organ components. Most segmentation steps were carried out 'manually' by subjective interpretation of structures on 1800 (cross) slices.

| Comparison of CT data sets
The skin of the unstained head (Etmop-lucif-ZSM-30813, of staining, the skin also appears light grey but compared to the previous scan, no significant contrast enhancement on the inner soft tissues was observable (Etmop-lucif-ZSM-30813_ai2mu, Table 1 and Figure 1).
The 120-h data set allows a detailed description of the head anatomy. Contrast and resolution of the data set allowed for distinguishing many structures and tissue types, and, therefore, segmentation with color-coding of different components.

| Chondrocranium
The chondrocranium is a large single element forming a very distinctive structure that contains the brain and associated compo-  (Shirai & Nakaya, 1990a, Figure 2d) is not discernable in our data ( Figure 3).
The orbital region contains a large eye socket occupying about a third of the total cranium ( Figure 3a). The superficial ophthalmic nerve (cranial nerve V) perforates the cranium first at the posterior end of the eye socket, passes along the wall of the orbit and perforates again the chondrocranium at the posterior end of the nasal capsule and opens at the anterior part of the preorbital wall (Figure 3a,c,d). The eye stalk is very short and does not reach the eyeball. The dorsally located suborbital keel process (Figure 3a) appears broad and compact.

| Mandibular arch and dentition
The mandibular arch is composed of the lower Meckel's cartilage and the palatoquadratum (Figures 4 and 5).
The palatoquadratum consists of two antimeres. Both cartilages are connected mesially by soft tissue along the symphysis ( Figure 5c).
In the middle, the palatoquadratum shows an orbital process that ex- The teeth of the lower jaw differ fundamentally in morphology from those of the upper jaw; the base appears flat and rectangular. The cusp is bent laterally and forms a cutting edge (Figure 6a,d).
The cutting edge has no serration. The teeth are overlappingly arranged and are connected by connective tissue (Figure 6d), which supports the replacement of complete tooth rows in the lower jaws of Etmopterus (Figure 6e).
A pair of labial cartilages support the upper and lower lips. The lower labial cartilage is lying lateral at the 10th tooth counting from the symphysis. It protrudes significantly beyond the dorsal mandibular rim (Figure 4c). The upper labial cartilages are situated between the 6th and 7th upper jaw teeth. Compared to the lower labial cartilage, they lie more anteriorly and are curved pointing laterally ( Figure 5a,b,d).

| Hyoid arch
The hyoid arch consists of five individual elements: One central unpaired basihyale, paired ceratohyale and paired hyomandibulae (see

| Brain and nervous system
Our CT data allow identification of individual brain components, that are olfactory bulbs, telencephalon, diencephalon, mesencephalon, to Yopak (2022), the function of the diencephalon in cartilaginous fishes is still not entirely clarified. The author suggests that it is a 'multimodal relay centre' (Yopak, 2022). Further, it seems to regulate a variety of homeostatic functions (e.g. feeding and reproduc-

| Nose
The olfactory rosettes lie in the nasal capsules that connect anteriorly to the orbits (Figure 13e). The individual olfactory rosettes appear kidney-shaped with the inner cavity connected to the nostrils ( Figure 13a,b,e,f). The epithelium as such is folded for surface enlargement and forms the so-called olfactory lamellae (Figure 13d).

| Eyes
The lenses appear as a bright, well-defined structure in the CT slices ( Figure 14a). Their surface appears to be even and not affected by shrinkage. There is a strongly twisted and folded structure surrounding the eyeballs. This presumably is composed of the sclera (incl. scleral cartilage), the choroid (incl. tapetum lucidum) and the retina (Figure 14a,c,d). This tissue complex behind/below the left eye seems to be misplaced in comparison to the one of the right eye. The

| Ampullae of Lorenzini
Generally, the ampullae of Lorenzini open by a dermal porus to the exterior, which receives the electrical stimuli. From these pores, a duct leads into the terminal diverticulum/ampulla (Wueringer, 2012;Wueringer et al., 2021). This in turn consists of several separate circularly arranged chambers in which the signal conversion takes place (Figure 15a-c). A few ampullae and their ducts were segmented and visualized representatively for each region (Figure 15).
The majority of ampullae are located along the dorsal area of the rostrum, above the chondrocranium (Figure 15a). They extend from the rostral process until about the middle of the otic region.
Their ducts leading to the ampullae are relatively variable in length but tend to be somewhat shorter than those of the other regions

| Comparison of CT examinations to previous anatomical studies
Comparative anatomical studies of Etmopteridae and sharks, in general, are low in number, however, our data allow us to observe characteristic differences of Etmopterus described, for example, in Shirai (1992b) to other squaliform sharks in general and within etmopterids in particular.

Another characteristic of Etmopterus along with Euprotomicrus,
Squaliolus and Trigonognathus identifiable on our CT data is a separate fossa for the external rectus muscle (Shirai & Okamura, 1992).
The branchial arches of E. lucifer consist of five pairs of arches, typical for most extant sharks species (Shirai, 1992a). In addition to Shirai's examination of cartilage elements and muscular parts, it was also possible to show and describe essential elements of various sensory organs and the associated nervous system. Especially the brain, the olfactory rosette (containing the olfactory epithelia) and the ampullae of Lorenzini were examined in detail. These structures are often difficult to identify in dissection, as they easily collapse if not supported by surrounding tissue when not prepared in water.
Compared to the subadult E. spinax analysed in Holmgren (1940), the skull of the E. lucifer investigated herein is straighter and more elongated. The skull of the subadult specimen in Holmgren (1940) describes an almost semicircular arc from anterior to posterior (see Holmgren, 1940 fig. 95). Furthermore, the rostrum of E. lucifer appears significantly wider compared to the S. acanthias specimen studied in Marinelli and Strenger (1959) and Shirai (1992b), whereby the specimen examined by Marinelli and Strenger (1959) was subadult as well and significantly smaller compared to the specimen analysed herein.
The skull cartilage of E. lucifer shows a smooth transition from the nasal capsules to the tip of the rostrum and has only a small, separated area towards the rostral process and tapers at this point ( Figure 3c,d). The rostral process is more prominent and larger in S. acanthias. The nasal capsules of E. lucifer occupy a substantial part of the anterior part of the chondrocranium and are significantly larger than those of S. acanthias. They are generally more similar to those of Cirrhigaleus barbifer (Shirai, 1992b). The rostral or subnasal fenestrae of E. lucifer appear significantly larger than those of S. acanthias shown in Marinelli and Strenger (1959) and Shirai (1992b). In contrast to the juvenile E. spinax (Holmgren, 1940), the rostral fenes- In general, the eye sockets of our specimen are larger than those of S. acanthias but are relatively similar dimensioned in the juveniles of E. spinax in Holmgren (1940).
The foramen nervi II in E. lucifer is more prominent than in S. acanthias (Marinelli & Strenger, 1959). The fossa parietalis extends posteriorly within the skull cartilage in two relatively short tubular cavities and two lobe-like branches within the otic region. The depression of the foramen arteriae carotis internae is also visible in the model and corresponds in position and size to that of S. acanthias (Marinelli & Strenger, 1959;Shirai, 1992b), as well as that of E. spinax (Holmgren, 1940). The basicranial keel process is a prominent outgrowth shortly in front of the basal angle. In Etmopteridae it persists until the adult stage, but in our specimen, it appears much wider and more massive than the ones of Aculeola and Centroscyllium (Shirai, 1992b;Shirai & Nakaya, 1990b). In the E. lucifer examined here this keel process serves at the base, in combination with the interorbital wall, as an attachment site for the Musculus suborbitalis, which is relatively prominent. According to Shirai (1992b), the muscle fibres inserted into the palatoquadratum in the family Etmopteridae should not be distinguishable from the M. adductor mandibulae. In the present case, however, the suborbital muscle is very distinct and clearly separated from the mandibular adductor muscle. This muscle seems to be generally used to protrude the jaws. In this way, it is possible to extend the mouth opening. This may be advantageous for feeding or scavenging on larger prey items to saw out pieces, which is a common feeding strategy taking into account the heterodont dignathy and stomach content analyses of E. spinax (Neiva et al., 2006).

| Anatomical adaptations to the deep sea
As indicated by previous studies, inter-specific variation in the morphology, quantity and distribution of ampullae of Lorenzini and their associated structures cannot be strictly correlated with morphological similarities or a close phylogenetic relation, but rather with feeding ecology and habitat Kempster et al., 2012Kempster et al., , 2013. In line with these results, further cranial characteristics may also reflect adaptations to the (deep-sea) habitat of Etmopterus. In

F I G U R E 11
Brain of Etmopterus lucifer, showing the different areas of this part of the central nervous system. (a) Single image of the CT scan shows the structure of the brain in a cross-sectional view, coronal, warped unstained (Etmop-lucif-ZSM-30813) and 120 h stained (Etmop-lucif-ZSM-30813_ai2mu) data sets; the red line in the scheme indicates the plane from which the slice originates in the CT scan, (b) dorsal view of the brain, showing the individual parts of the brain, (c) dorsal view of the brain which illustrates the position within the chondrocranium (latter displayed translucently), (d) lateral view of the brain which illustrates the position within the chondrocranium (latter displayed translucently), (e) lateral view of the brain showing the single lobes and areas, (f) dorsal view of the brain showing the single lobes and areas. Abbreviations: ace, auricle of cerebellum; br, brain; ce, cerebellum; dic, diencephalon; ey, eye; lih, lobus inferior hypothalami; mob, medulla oblongata; msc, mesencephalon; mtc, metencephalon; myc, myelencephalon; och, optical chiasma; ob, olfactory bulbs; otr, olfactory trunk, sc, spinal cord; tec, telencephalon; to, tectum opticum fact, the head anatomy suggests that the olfactory system in combination with the ampullae of Lorenzini are the dominant senses likely used for prey detection. Both are notably enlarged in comparison to S. acanthias, a squaliform not permanently inhabiting the deep sea (Ebert et al., 2021). Kajiura et al. (2010) describe the dorsal-to-ventral pore distribution ratio for multiple shark species. Etmopterus lucifer and S. acanthias show a rather similar distribution. Both species exhibit more pores on the ventral side (approx. ratio dorsal-ventral pore distributions: 0.6 vs. 0.5). This may suggest that both species feed rather benthos orientated or generally tend to approach their prey from above. Considering the ratios of species that can be confidently classified as benthic feeders (e.g. Ginglymostoma cirratum, Stegostoma fasciatum, Carcharias taurus), the ratios are remarkably similar. Therefore, it can be assumed that E. lucifer has a similar pore distribution because of its feeding habits. Also, Kajiura et al. (2010) state that deep sea-inhabiting species with large pore numbers (E. lucifer possesses one of the highest pore numbers of all species in the study) are likely forageing off the sea floor. Etmopterus lucifer has almost four times more ampullae of Lorenzini compared to other species studied by Kajiura et al. (2010). The high number of electroreceptors could facilitate the detection of fast and agile prey, especially in the open water column  as prey items of this opportunistic feeder also encompasses squids and myctophid fishes (Martin & Mallefet, 2022). Further, pores and ducts associated with transporting electric signals to synapses with a larger diameter, which are often found in deep-sea sharks, reduce the electrical impedance along the length of the duct. This probably results in increased sensitivity. This may be an advantage of comparatively small-bodied deep-sea shark species, as a similar electrosensory sensitivity is achieved as in larger-bodied species .
For highly migratory species, specific senses for orientation and navigation are essential, especially in the pelagic realm lacking location-specific landmarks (Meredith et al., 2022). The use of the geomagnetic field as a navigation aid in chondrichthyans was presumed previously (Kalmijn, 1974;Klimley, 1993;Paulin, 1995). strengthened the assumption that there is an active use of the geomagnetic field of the earth in these species (Keller et al., 2021;Newton & Kajiura, 2020). It is not yet fully clarified whether the detection of variation in magnetic fields is conducted via the ampullae of Lorenzini or/and through another (unknown) structure (Anderson et al., 2017;Kalmijn, 1984;Walker et al., 2003). E. lucifer has very long ampullar channels, especially on the lateral sides of the head. According to Sisneros and Tricas (2002)  Etmopterus lucifer (158-1357) occurs in a greater median depth than S. acanthias (0-600 m) (Ebert et al., 2021). At these depths, short wavelengths predominate, i.e. blue light and only a small number of photons is still present . Enlarged eyes frequently evolved in various species inhabiting such depths. In addition to the enlargement for increased photon sensitivity, sensitivity is also shifted to specific wavelengths (Meredith et al., 2022).

Experiments in
That increased sensitivity helps on one hand to efficiently use the low amount of penetrating light and on the other hand to detect blue light emitted via the bioluminescence of prey and the light emission of conspecifics' photophores (Claes et al., 2014;Ebert et al., 2021;Reif, 1985). Kajiura et al. (2010) show that the eye diameter of sharks increases with the depth of occurrence. Deep-sea species seem to have larger eye diameters than shark species found in shallower habitats. In fact, E. lucifer shows the largest relative eye diameter of all Abbreviations: II, fossa of the cranial nerve II; ace, auricle of cerebellum; aos, antorbital shelf; ba, branchial arches; cc, chondrocranium; ce, cerebellum; ins, internasal septum; kp, keel process; na, naris; ob, olfactory bulb; or, olfactory rosette; rp, rostral process; sf, subnasal fenestrae; sp, spine; to, tectum opticum sharks studied in Kajiura et al. (2010). This all together may indicate that the sense of sight for E. lucifer is higher developed than in other species of the Etmopteridae investigated in Kajiura et al. (2010) and certainly more emphasized than in S. acanthias.
The dimensions of the olfactory bulbs were often used to determine the olfactory capacity of a species and would indicate the significance of the olfactory sense in E. lucifer Schluessel et al., 2008Schluessel et al., , 2009Theisen et al., 1986;Zeiske et al., 1987). The sheer size of the nasal capsules does not necessarily allow conclusions on the efficiency of the olfactory system (Gardiner et al., 2012;Meredith et al., 2012;Meredith & Kajiura, 2010), however, Camilieri-Asch,  claim that the dimensions of the olfactory bulbs are correlated with the number of olfactory stimuli through the epithelium. They conclude that the dimensions can be considered a valid proxy for olfactory ability in elasmobranchs. In contrast, other studies state that the size of the olfactory bulbs likely reflects rather a functional specialization taking the habitat and ecology of species into account (Kotrschal et al., 2017;Schluessel et al., 2008;Theiss et al., 2009;Yopak et al., 2015). The functional adaptation matches with results from Yopak et al. (2015), where the standardized residual olfactory bulb size of E. lucifer is consistent with the general trend that deep-sea shark species have large olfactory bulbs, showing a presumably greater reliance on the olfactory sense in environments with low visual cues. The brain areas responsible for processing signals from the olfactory sense and the receptor cells of the ampullae of Lorenzini are notably enlarged compared, for example, to S. acanthias ( Figure 11) further supporting the increased functionality of these senses in a species permanently inhabiting deep-sea habitats.
Several studies showed that galeomorph sharks and myliobatiform rays tend to have larger brains than squalomorph sharks and holocephali (Myagkov, 1991;Northcutt, 1978;Yopak & Lisney, 2012). In general, especially deep-sea inhabiting species show average to below-average brain dimensions in relation to their body size compared to species that tend to live in shallower habitats (Yopak & Montgomery, 2008). According to Kajiura et al. (2010), the F I G U R E 1 4 Involved organ systems of the sense of sight, including the nerves and ciliary muscles of Etmopterus lucifer. (a) CT image shows the colour labelled the Nervus opticus (yellow), the sensory epithel (green), the eyeball with lens (red) and the ciliary muscles (orange); the red line in the scheme indicates the plane from which the single image originates in the CT scan, coronal cut, warped unstained (Etmoplucif-ZSM-30813) and 120 h stained (Etmop-lucif-ZSM-30813_ai2mu) data sets; (b) ciliary muscles, dorsal view (c) lens, sclera, cornea, choroid and suprachoroidea with Nervus opticus, anterior view, the arrow points out the direction of rotation around which the entire right structure is twisted, (d) lens, sclera, cornea, choroid and suprachoroidea, Nervus opticus and brain dorsal, (e) Nervus opticus with the transition to the brain at the optical chiasma, ventral view.

| CON CLUS IONS
This study describes the cranial morphology of an etmopterid shark species in detail using (dice)micro-CT data for the first time.

ACK N OWLED G M ENTS
We would like to thank Ulrich K. Schliewen (Ichthyology, ZSM) for the sampling allowance and Roland Melzer (ZSM) for the fruitful discussion of the project. Eva Lodde Bench (ZSM) is thanked for her help with staining. Two anonymous reviewers are acknowledged for their very constructive feedback.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
All figures are included and visible in the submission.