The Arrangement of the Peripheral Olfactory System of Pleuragramma antarcticum: A Well-Exploited Small Sensor, an Aided Water Flow, and a Prominent Effort in Primary Signal Elaboration

Simple Summary How animals perceive their surrounding environment is crucial to their reactions and behavior. Olfaction, among others, is one of the more important senses for wide-range communication and in low-light environments. This study aims to give a morphological description of the peripheral olfactory system of the Antarctic silverfish, which is a key species in the coastal Antarctic ecosystem. The head of the Antarctic silverfish is specialized to assure that the olfactory organ keeps in contact with a large volume of water, even when the fish is not actively swimming. The sensory surface area and the number of neurons in the primary olfactory brain region show that this fish invests energy in the detection and elaboration of olfactory signals. In the cold waters of the Southern Ocean, the Antarctic silverfish is therefore likely to rely considerably on olfaction. Abstract The olfactory system is constituted in a consistent way across vertebrates. Nasal structures allow water/air to enter an olfactory cavity, conveying the odorants to a sensory surface. There, the olfactory neurons form, with their axons, a sensory nerve projecting to the telencephalic zone—named the olfactory bulb. This organization comes with many different arrangements, whose meaning is still a matter of debate. A morphological description of the olfactory system of many teleost species is present in the literature; nevertheless, morphological investigations rarely provide a quantitative approach that would help to provide a deeper understanding of the structures where sensory and elaborating events happen. In this study, the peripheral olfactory system of the Antarctic silverfish, which is a keystone species in coastal Antarctica ecosystems, has also been described, employing some quantitative methods. The olfactory chamber of this species is connected to accessory nasal sacs, which probably aid water movements in the chamber; thus, the head of the Antarctic silverfish is specialized to assure that the olfactory organ keeps in contact with a large volume of water—even when the fish is not actively swimming. Each olfactory organ, shaped like an asymmetric rosette, has, in adult fish, a sensory surface area of about 25 mm2, while each olfactory bulb contains about 100,000 neurons. The sensory surface area and the number of neurons in the primary olfactory brain region show that this fish invests energy in the detection and elaboration of olfactory signals and allow comparisons among different species. The mouse, for example—which is considered a macrosmatic vertebrate—has a sensory surface area of the same order of magnitude as that of the Antarctic silverfish, but ten times more neurons in the olfactory bulb. Catsharks, on the other hand, have a sensory surface area that is two orders of magnitude higher than that of the Antarctic silverfish, while the number of neurons has the same order of magnitude. The Antarctic silverfish is therefore likely to rely considerably on olfaction.

Notothenioidei is the only fish sub-order that has adapted to the coldest ocean on the Earth: the Antarctic Ocean. The morphology of at least some parts of the olfactory system of notothenioids has been investigated over the years [16,[25][26][27][28][29][30][31][32][33]. The Antarctic silverfish Pleuragramma antarcticum is the dominant pelagic fish in coastal Antarctica and a keystone species in local ecosystems. Unlike most notothenioids, the Antarctic silverfish is associated with water columns, with eggs developing in a unique environment-platelet ice-and adults living in the midwaters over the continental shelf. Its pelagic lifestyle, unusual among the swim bladder-less Notothenioidei, is made possible by relevant morphological modifications, selected on an evolutionary timescale in a process of secondary pelagization. [34][35][36]. On the other hand, the importance of migrations around the Antarctic continent for this species seems clear. Although its life cycle has been only partially disclosed, a homing behavior was hypothesized for the Antarctic silverfish [37], with reproductive migrations driven by environmental cues from open waters to selected coastal areas. Even if it is not yet proven to date, it may be possible that homing is assisted by olfaction. The use of the specific name P. antarcticum follows the currently recognized change of nomenclature [38]. The OB of P. antarcticum has been the object of a study focused on the sensory areas of the brain [39]. In that article, the size of the OB was evaluated and compared to other brain regions and the OB of other species.
This work aims to illustrate the anatomy of the peripheral olfactory system of adult P. antarcticum, describing the olfactory chamber, the olfactory organ, the ON, and the OB.

Sampling and Tissues Preparation
The heads from six adult specimens of the Antarctic silverfish Pleuragramma antarcticum were obtained from two different operations. Two specimens were collected at Cape Hallett (Ross Sea, Antarctica) during the Western Ross Sea Voyage 2004 [40] onboard the RV Tangaroa. The other four were collected at Iselin Bank (Ross Sea, Antarctica) during the Tangaroa Ross Sea Voyage 2019 [41]. The size of the specimens was not recorded after capture and only the heads were fixed in paraformaldehyde 4%, rinsed in PBS, and stored in 70% ethanol for anatomical and histological investigations. To determine the size of the specimens a linear function between standard length (SL) and brain length (BL, from the olfactory bulb to the obex) was based on the dataset from [39]: SL = 20.811 BL − 97.568 (mm) (1) To make comparisons with data from the literature where the total length (TL) of the fish is indicated, this linear function for the length was obtained for P. antarcticum from [42]: TL = 1.092 SL + 0.284 (cm) (2) The calculated sizes of the P. antarcticum specimens are reported in Tables 1 and 2. At the time of the analysis, the heads were dissected to isolate the brain and the olfactory organ. Alternatively, the olfactory organ was not removed, and the histological analysis was performed on the anterior part of the snout after five days of decalcification in Osteodec (Bio-Optica, Milano, Italy).

Gross Morphology
One olfactory organ for each specimen was analyzed through a stereomicroscope Leica DMRB (Leica Microsystems, Wetzlar, Germany) equipped with a Moticam 10+ camera (Motic Europe, Barcelona, Spain) to count the olfactory lamellae, to evaluate the OSA, and to measure the olfactory nerve thickness using ImageJ [43]. As specified, all the measurements were collected after fixation and storage in 70% ethanol.

Histology
The isolated olfactory organ and nerves, or the whole anterior part of the head, were dehydrated in ethanol, paraffin embedded, and 5 µm cut using a microtome, according to standard protocols. The slides were stained using Hematoxylin-Eosin, Masson's Trichrome, and Azan Trichrome (Bio-Optica, Milano, Italy). The stained sections were observed using a transmitted light microscope Leica DMRB equipped with a Moticam 3+ (Motic Europe, Barcelona, Spain) or through a transmitted light microscopy Olympus BX60 equipped with a Microvisioneer (Esslingen am Neckar, Germany) camera and acquisition software.

Isotropic Fractionator
Each OB (one hemisphere) was weighed and analyzed according to the isotropic fractionator technique for cell and neuron counting [44]. Briefly, each OB was homogenized in 1% Triton in 40 mM trisodium citrate. After centrifugation at 10,000 rpm for 45 minutes, the pellet, which contained all the extracted nuclei, was resuspended in a known volume of phosphate buffer saline (pH 7.4) and 4 ,6-diamidino-2-phenylindole (DAPI) for nuclei detection. The solution with stained nuclei was continuously stirred to become isotropic, and samples from the solution were used for nuclei counting in a hemocytometer under an epifluorescence microscope (Leica DMRB) equipped with a Moticam 3+ (Motic Europe, Barcelona, Spain). The total number of nuclei in the solution, and thus in the OB, was calculated. An aliquot of extracted nuclei solution was treated for immunocytochemistry using a mouse anti-Neuronal Nuclei (anti-NeuN) antibody (Millipore, Darmstadt, Germany, Mab377, 1:100) overnight at room temperature. NeuN is a neuronal marker and can be used as a marker for neuronal nuclei in vertebrates [45][46][47]. After rinsing in PBS and centrifuging at 10,000 rpm for 3 min, the pellet was incubated with a secondary antibody goat antimouse conjugated with the fluorochrome Dylight ® 488 (Immunoreagent, Raleigh, NC, USA, GtxMu-003-D488NHS). At least 100 nuclei were evaluated both in UV light for DAPI and in blue light for the Dylight ® 488, and the percentage of neurons and non-neuronal cells was evaluated. The mitral cells, large interneurons present in the OB of vertebrates, are one of the few neuronal populations with a NeuN-negative cell nucleus [48]. Nevertheless, the number of mitral cells has been considered negligible for the evaluation of neuron number in the OBs-at least in mammals [49].

Building of the Dataset
The quantitative measurements and counts obtained for the adults were then compared with data from other teleosts acquired from the literature. Teichmann [17] reports the OSA for 11 species of teleost fish. One of those species is the European eel, Anguilla anguilla, the OSA of which was also evaluated in another article [15]. As the average size of the fish considered (Teichmann [17]: average SL 51 cm; Atta [15]: average SL 54 cm) and the average OSA evaluated (Teichmann [17]: average OSA for two olfactory organs 575 mm 2 ; Atta [15]: average OSA for two olfactory organs 424 mm 2 ) are not identical in the two articles, but similar enough, we here decided to consider numbers from [17], for homogeneity with the other 10 species. In [17], several individuals for each species were analyzed and the average values for standard length and whole OSA are also given. Ferrando et al. [16] report the total length and the OSA of one olfactory organ of one specimen of the nototheniid species: Dissostichus mawsoni. Data from the present paper, from [16], and from [17] are gathered in Table 1. It is noteworthy that all the measures are average values from more than one specimen of the same species, except for D. mawsoni, where only one specimen was considered. Moreover, the body size reported for the specimen of D. mawsoni is the total length while, for all other fishes, the standard length is indicated.
The teleost species used in Table 1 for the comparison with P. antarcticum were chosen because of the availability of data from the literature. Many are mainly freshwater species (Phoxinus phoxinus, Gobio gobio, Squalius cephalus, Tinca tinca, Nemachilus barbatulus, Exos Lucius, Lota lota, Perca fluviatilis), two species are fully or partially anadromous (respectively Salmo irideus and Gasterosteus aculeatus), one species (Anguilla anguilla) is catadromous, and one species is marine (D. mawsoni). Beside the anadromous and catadromous species, some others are characterized as P. antarcticum, by their spawning migrations during their lifetime (S. cephalus, P. fluviatilis, D. mawsoni). For only some species is occurrence in groups documented (S. cephalus, G. aculeatus) [50].
In the literature, the isotropic fractionator technique has been applied to the OB of only one specimen of a teleost species: D. mawsoni. Unfortunately, the anti-NeuN antibody failed to work on that species, so the number of cells in the OB of that specimen is known, but not the percentage (and number) of neurons [16]. Considering non-teleost fish, the isotropic fractionator technique has been applied, to date, to the OB of two catshark species: Scyliorhinus canicula and Galeus melastomus [51]. Here, the number of cells and the cell density in the OB of P. antarcticum are compared to the same parameters in D. mawsoni and the two catshark species. Moreover, the number of neurons, the neuron density, and the ratio of other cells/neurons in the OB of P. antarcticum are compared to the same parameters in the two catsharks. In [51], another interesting parameter was also calculated: the number of neurons in the OB normalized to square millimeters of OSA. This parameter, which puts together two variables from the sensor and the first relay, promises to be quite interesting when available for several vertebrate species. We here calculated it for adult P. antarcticum. The measured and calculated numbers are presented in Table 2. Some scatterplots were obtained using the ggplot2 R package [52], to present graphically the data in Tables 1 and 2.  Table 2. Measurements and counts for 4 species of fish from the present study and the literature. TL = the total length was measured for fresh fish for all the specimens but P. antarcticum specimens, where it was calculated on the basis of paraformaldehyde-fixed and ethanol 70%-preserved brains (See Materials and Methods). RoL = Rosette Length; RoW = Rosette Width; LN = lamellar number for one rosette; ES = epithelial surface area for one rosette; OB = mass of the olfactory bulb (one hemisphere); OB cells = number of cells in one OB; OB neu = number of neurons in one OB; Neu% = percentage of neurons in the OB cells; Cell/mg = number of cells per milligram of OB tissue; Neu/mg = number of neurons per milligram of OB tissue.

The Olfactory Organ of Adult P. antarcticum Is an Asymmetrical Rosette
The olfactory organ of the adult specimens of P. antarcticum (average TL 18.2 ± 2.6 cm) was an olfactory rosette characterized on average by 24 ± 2 lamellae, which were unevenly arranged in two rows along a central raphe. The length of the raphe was about 3.1 ± 0.3 mm in the specimens analyzed here, and it ran cranio-caudally. As the rosette was obliquely oriented within the olfactory chamber, one the parasagittal plane, the medial row of lamellae was dorsomedial, while the lateral one could be said to be ventrolateral. The ventrolateral row had two more lamellae than the dorsomedial, giving the olfactory rosette an asymmetrical shape that is not commonly seen in fish (Figures 1a and 2a,b). Only one of the observed rosettes had a difference of three lamellae instead of two between the ventrolateral and the dorsomedial row.

The Olfactory Organ of Adult P. antarcticum Is an Asymmetrical Rosette
The olfactory organ of the adult specimens of P. antarcticum (average TL 18.2 ± 2.6 cm) was an olfactory rosette characterized on average by 24 ± 2 lamellae, which were unevenly arranged in two rows along a central raphe. The length of the raphe was about 3.1 ± 0.3 mm in the specimens analyzed here, and it ran cranio-caudally. As the rosette was obliquely oriented within the olfactory chamber, one the parasagittal plane, the medial row of lamellae was dorsomedial, while the lateral one could be said to be ventrolateral. The ventrolateral row had two more lamellae than the dorsomedial, giving the olfactory rosette an asymmetrical shape that is not commonly seen in fish (Figures 1a and 2a,b). Only one of the observed rosettes had a difference of three lamellae instead of two between the ventrolateral and the dorsomedial row.

The Nasal Chamber Has a Reinforced Roof and Is Connected to Accessory Nasal Sacs
The cranial part of the olfactory rosette, corresponding to the dorsomedial row of the lamellae, was connected to the roof of the olfactory chamber, anterior to the opening of

The Nasal Chamber Has a Reinforced Roof and Is Connected to Accessory Nasal Sacs
The cranial part of the olfactory rosette, corresponding to the dorsomedial row of the lamellae, was connected to the roof of the olfactory chamber, anterior to the opening of the nostril (Figure 2a,c). Posteriorly to this point of connection, the rosette was anchored only basally, and it stood free in a quite large olfactory chamber, opened dorsally in a single nostril (Figure 2a,d). Posteriorly to the nostril opening, the roof of the olfactory chamber-which overhung the rosette-was characterized by particularly compact connective tissue (Figure 3a,b and File S1). The main olfactory chamber was in continuity with three accessory nasal sacs. One accessory nasal sac was dorso-medially developed, and was quite small with respect to the other two; it could also have had the role of a branch of the main olfactory chamber (Figure 3a). The second sac was ventro-medially developed, presented with some diverticula, and reached the roof of the oral cavity-being separated from it only by a thin layer of tissue (Figure 3a,c). The third accessory nasal sac was ventro-lateral ( Figure 3a). chamber-which overhung the rosette-was characterized by particularly compact connective tissue (Figure 3a,b and File S1). The main olfactory chamber was in continuity with three accessory nasal sacs. One accessory nasal sac was dorso-medially developed, and was quite small with respect to the other two; it could also have had the role of a branch of the main olfactory chamber (Figure 3a). The second sac was ventro-medially developed, presented with some diverticula, and reached the roof of the oral cavity-being separated from it only by a thin layer of tissue (Figure 3a,c). The third accessory nasal sac was ventrolateral ( Figure 3a).

The Olfactory Organ Surface Area of an Adult P. antarcticum Is Almost 50 mm 2
The sensory olfactory epithelium covered almost the whole surface of the olfactory organ alongside the flat faces of the lamellae and the interlamellar curves. It had a thickness of 38.4 ± 0.6 µm. The non-sensory epithelium was localized along the free edge of the lamellae (Figure 4). Measuring the surface area of the flat faces of each lamella of one olfactory organ in order to obtain the OSA also provided a good evaluation of the sensory surface area (SSA). The measured OSA in one olfactory rosette from the adult specimens analyzed here ranged from 15 to 32 mm 2 . Each fish had, overall-on average-47.4 ± 11.9 mm 2 of OSA, and probably of SSA too. All the measures are reported in Table 2. In Figure 5, the average value for P. antarcticum is plotted with measurements from the literature on other teleost species (see Materials and Methods).

In P. antarcticum the Olfactory Nerve Is Long, while the Olfactory Tract Is Short
The lamellae had a thin lamina propria where the bundles of the axons of the olfactory sensory neurons, i.e., the fila olfactoria, were visible. The bundles from each lamella gathered in a ribbon of fibers at the base of each lamellar row. More deeply, under the raphe, the two ribbons of nerve fibers joined and formed the olfactory nerve that ran toward the CNS (Figure 6). The olfactory nerve had a diameter of about 400 µm and it appeared to be divided into bundles (Figure 6c). Each olfactory nerve reached one of the olfactory bulbs. Olfactory bulbs can be considered sessile in P. antarcticum as they had an

In P. antarcticum the Olfactory Nerve Is Long, while the Olfactory Tract Is Short
The lamellae had a thin lamina propria where the bundles of the axons of the olfactory sensory neurons, i.e., the fila olfactoria, were visible. The bundles from each lamella gathered in a ribbon of fibers at the base of each lamellar row. More deeply, under the raphe, the two ribbons of nerve fibers joined and formed the olfactory nerve that ran toward the CNS (Figure 6). The olfactory nerve had a diameter of about 400 µm and it appeared to be divided into bundles (Figure 6c). Each olfactory nerve reached one of the olfactory bulbs. Olfactory bulbs can be considered sessile in P. antarcticum as they had an extremely short olfactory tract connecting them to the rest of the telencephalon. In the specimens analyzed, the average mass of one OB (one hemisphere) was 1 ± 0.5 mg. The numbers of cells and neurons in each OB of P. antarcticum are reported in Table

In the Olfactory Bulb of P. antarcticum, Many Cells Elaborate the Signal from a Small Olfactory Organ
In the specimens analyzed, the average mass of one OB (one hemisphere) was 1 ± 0.5 mg. The numbers of cells and neurons in each OB of P. antarcticum are reported in Table 2, together with data from [16] regarding D. mawsoni, and from [51] regarding two catsharks. The mass of the OB and the number of cells and neurons (Figure 7a) had the same order of magnitude in P. antarcticum and catsharks of comparable body size. The number of neurons in the OB normalized for the OSA, instead, ranges in catsharks from 102 to 103, while it reached 104 in P. antarcticum.  Table 2. Quantitative data from olfactory rosette and bulb of P. antarcticum (present work), and G. melastomus and S. canicula [51].

Discussion
The peripheral olfactory organ of P. antarcticum is a rosette, which is a quite common shape in fish in general. The shape of the raphe, which is elongated, and the distribution of the lamellae, in two rows connected in the posterior part, correspond to the G arrangement, as indicated by [5]. Nevertheless, the olfactory rosette of P. antarcticum could be considered a modified G-type, because of the asymmetry of the two rows of lamellae.
The lamellar number reported here (which was, on average, 24 for each rosette and 48 for the whole adult specimen) is likely to increase ontogenetically, and thus we can expect to find a lower number in younger specimens. The ontogenetic increase is expected because lamellae have a smooth surface, being without secondary folds on their surface, and a positive correlation between the number of olfactory lamellae and body length is found in fish without secondary folds [13]. Eastman [25] reported a lamellar number of 22-26 for this species, but without specifying the range of body sizes of the analyzed specimens.
The presence of accessory nasal sacs, which are also close to the roof of the oral cavity, suggests a possible pumping mechanism of the sacs to create the water current in the olfactory chambers, exploiting mouth and head movements. In fact, the accessory nasal sacs are often described in monotrematous species, such as P. antarcticum, and are present in all notothenioids [32,33].
The OSA, and the SSA, were on average 24 mm 2 for each rosette, and 48 mm 2 for the organism. To evaluate if this is a large or small area, a comparison with other vertebrates was needed. A comparison with other teleost species from the literature [16,17] is shown in Table 1 and Figure 5. Data from [17] and from [16] regard the OSA, because the distribution of the sensory and non-sensory epithelium was examined in those articles, and the surface areas regard the lamellae. Considering the OSA, P. antarcticum was out of the confidence interval calculated for the regression curve, and its surface area was lower than expected for a teleost of that body size ( Figure 5). The sensory and non-sensory epithelium coverage on the olfactory lamellae of teleosts was described by [5] and divided into four types. Type I is similar to that seen in P. antarcticum and shows the sensory epithelium continuously covering the sides of the lamellae, leaving the non-sensory epithelium on the edge of the lamellae. Types II, III, and IV instead show a more fragmentary distribution of the sensory epithelium. Yamamoto [5] described these different types in several teleost species, none of which are present in our dataset. The four species of Anguilliformes analyzed by [5] showed a distribution of type I; the same was seen for the two species of Gadiformes and the five species of Cypriniformes. This could suggest that, for at least some of the species in Figure 5, the OSA is a good proxy for the SSA-which is more informative about sensory function. In Figure 5, the species are indicated with different colors according to their order. As the number of species is relatively low, and some orders are represented by only one species, it would be difficult to draw possible correlations between ecological tracts-such as the migratory behavior that characterize some of the species-and the OSA.
The SSA was evaluated for other vertebrates in the literature. In Chondrichthyes, species with a body size ranging from 210 to 2300 mm had an SSA ranging from about 2000 to 120.000 mm 2 [53]. The comparison with non-fish vertebrates could be performed only on a body-mass base. The body mass of the P. antarcticum specimens analyzed here was not measured, but they could be grossly calculated using the regression function from [42]: Thus, the P. antarcticum specimens analyzed here had a calculated body mass of 20-32 g. In mammals, the SSA has been measured in many species [54][55][56][57][58][59]. Although a review of the SSA of mammals is out of the scope of this work, from the partial literature cited, several mammalian species have a total SSA of the same order of magnitude as that of P. antarcticum: 2 species from the order Afrosorcida (range of SSA about 51-82 mm 2 , range of average body mass about 80-100 g), 2 species from Chiroptera (range of SSA about 32-68 mm 2 , range of average body mass about 16-28 g), 12 species from Eulipotyphla (range of SSA about 52-92 mm 2 , range of average body mass about 6-35 g), 1 species from Primates (SSA about 55 mm 2 , average body mass about 246 g), and 29 species from Rodentia (range of SSA about 29-104 mm 2 , average body mass about 7-195 g). All these numbers, regarding the SSA and the body mass, were obtained from the literature [54,57,58,60,61]. The SSA of P. antarcticum, which is two orders of magnitude lower than that of Chondrichtyes of the same body size, is one order of magnitude lower than expected among teleosts, but it is quite similar to mammalian species with a comparable body mass.
The comparison of SSAs brought up other parameters that should be considered: the ORN density (i.e., ORNs per mm 2 )-which can be reflected by the olfactory epithelium thickness and also by the cell size-the number of sensory cilia or microvilli that the ORNs bear, and the level of expression of olfactory receptors-which represent the actual part of the "sensory surface" that the odorant molecules link to.
The thickness of the ON should reflect the number of ORNs and could allow other comparisons with other species; this requires some assumptions that are not certain-for example, that all the axons have the same diameter, and that the proportion of connective tissue among the bundles is similar. The ON diameter in a D. mawsoni specimen of 143 cm TL was evaluated to be 2-2.5 mm, and so 5-6 times larger than that of the adult P. antarcticum presented here. A quantitative analysis of the ON would require a transmitted electron microscope (TEM) observation as, for example, was carried out for Esox lucius [23].
At least for mammals, the absolute and relative size of the OB as a proxy of the olfactory capability of a species has been questioned [62]. The number of neurons has been proposed as a more reliable parameter to consider [48,60]. In adult P. antarcticum, the number of neurons in both the OBs was 2.65 × 10 5 ± 1.24 × 10 5 . The mouse Mus musculus, whose SSA of about 47 mm 2 [57] is roughly the same as P. antarcticum, has 3.89 × 10 6 ± 1.25 × 10 6 neurons in the two OBs [48]. In M. musculus, both OB mass (about 2 mg in P. antarcticum and about 14 mg in M. musculus) and the number of neurons in the OB, are one order of magnitude higher than in P. antarcticum. It is noteworthy that M. musculus is considered a macrosmatic species-at least among mammals. The comparison of P. antarcticum to the two catsharks [51], which are the only fish species for which the number of neurons in the OB is available, shows that this number is comparable in fish with a similar SL and comparable OB mass (Figure 7 and Table 2), although those fishes have a notably larger SSA (Table 2). This difference in SSA is shown also in Figure 7, where the number of neurons in the OB of P. antarcticum, normalized to the mm 2 of surface area of the organ, was remarkably higher than those of the catsharks. This suggests that a large number of neurons elaborate the information coming from a relatively small (compared to catsharks) sensor in P. antarcticum.

Conclusions
Overall, the quantitative anatomy of the olfactory system of P. antarcticum presented in this study could indicate that this species relies on olfaction for important life tasks. Particularly, the presence of a number of neurons in the OB of this species, equal to that of catsharks and only tenfold less than that of macrosmatic mammals, suggests the necessity of high efficiency in odor first elaboration. Considering that the "expensive tissue hypothesis" [63] has been verified in teleosts too [64], the advantage of an efficient odor elaboration is likely to counterbalance the high energy cost of neural cells.