The fish biomass in the mesopelagic (200–1,000 m depth) and bathypelagic (1,000–4,000 m depth) zones constitutes a large fraction of the estimated total global fish biomass, particularly in the upper mesopelagic zone (200–600 m depth) where most of the abundance is concentrated (Sarmiento-Lezcano et al., 2022). The lower mesopelagic zone (600–1,000 m depth) and the bathypelagic zone on the other hand are usually inhabited by fewer organisms, although they are generally of larger sizes, increasing the proportional biomass (Badcock, 1970; Badcock and Merrett, 1976; Roe and Badcock, 1984; Olivar et al., 2012; Olivar et al., 2017). These species are the main prey of many large pelagic migrators (Schaber et al., 2022) and are also expected to be harvested for human consumption in the near future as a source of marine fat and protein (e.g., Gjøsaeter and Kawaguchi, 1980; FAO, 1997; FAO, 2001; Hidalgo and Browman, 2019; Grimaldo et al., 2020). Moreover, meso- and bathypelagic fishes contribute to the biological carbon pump, exporting organic carbon from the euphotic zone to the meso- and bathypelagic layers in the case of migratory fish, and remineralizing (breakdown or transformation of organic matter molecules into their simplest inorganic forms) in the case of non-migratory fishes that remain at depth (Lampitt et al., 2008; Guidi et al., 2015; Sarmiento-Lezcano et al., 2022). Cyclothone spp., are the most abundant among the non-migratory group (Sarmiento-Lezcano et al., 2022) and the most abundant vertebrates on the planet (Ahlstrom et al., 1984; Nelson et al., (2016) are known to contribute significantly to the non-migrant deep scattering layers detected acoustically (Peña et al., 2014).

Acoustic methodology combined with net trawling is often employed to estimate the abundance of aquatic organisms (Simmonds and MacLennan, 2005). However, an accurate abundance estimate is not yet available for meso- and bathypelagic species due to poor capture efficiency (Koslow et al., 1997; Kloser et al., 2009; Yasuma and Yamamura, 2010; Kaartvedt et al., 2012), vessel and platform avoidance (Peña, 2019; Peña et al., 2020), and acoustic uncertainties (Davison et al., 2015; Peña, 2019; Proud et al., 2019). A global mesopelagic fish biomass estimate using 38 kHz data (11–15 Gt; Irigoien et al., 2014) was an order of magnitude higher than a previous estimate from micronekton nets (1 Gt; Gjøsaeter and Kawaguchi, 1980). However, Irigoien et al. (2014) did not consider the large contribution of C. braueri to the most prominent scattering layer at 38 kHz (400–700 m) in many areas of the world (Peña et al., 2014; Ariza et al., 2016; Peña, 2019). Their resonant target strength at those depths and their low individual biomass would considerably reduce the estimated biomass. The acoustic contribution of non-migrating Cyclothone spp. to mesopelagic acoustic layers has often been underestimated due to the use of nets targeting larger fishes. However, their abundance often masks the echo of migrating species at 38 kHz (Peña et al., 2020).

The use of several frequencies in acoustics allows classifying areas of the echogram as gas-bearing, fluid-like, or elastic-shelled echotraces (Stanton et al., 1996), employing scattering models and clustering techniques (Peña and Calise, 2016; Peña, 2018; Peña et al., 2023a; Peña et al., 2023b). Scattering models reproduce the expected echo of an organism at different frequencies based on body morphology and behavior (Ona, 1990; Ona, 1999). Acoustic backscatter depends firstly on the existence of a swimbladder or gas enclosure. When gas is present, up to 95% of the acoustic backscatter is due to the swimbladder (Foote, 1980). Swimbladder size and depth in the water column determine both the intensity of the echo and the spectra (trend with frequency; Peña et al., 2014). Other important input parameters for the scattering models are orientation, size, shape, and material properties (density and sound contrast with the surrounding water). Therefore, knowledge of the swimbladder characteristics of the target species is of paramount importance for acoustic data elucidation.

Marshall (1951; 1960) described the swimbladder of several meso- and bathypelagic fishes, showing its structure from larval to adult stages. Larval stages of most mesopelagic fish dwell in the epipelagic layers (Olivar et al., 2018) and have swimbladders containing gas (Marshall, 1951; Marshall, 1960). For some species, such as bristlemouths of the genus Cyclothone and some myctophids, the organ regresses with age and is filled with fat in late adults (Marshall, 1960). Some information on the structures of Cyclothone swimbladders (C. braueri, C. signata, C. acclinidens, C. livida, and C. microdon) was described by Marshall (1960). However, there is little information about other bristlemouth species present in the Atlantic Ocean and the Mediterranean Sea, or on the evolution of swimbladder shape and gas/lipid content with age. The use of terms such as inflated, contracted, or regressed swimbladder without a detailed explanation gives rise to confusion in the literature. In fact, the same species (C. braueri) with similar length were described as non-gas-bearing (Kleckner and Gibbs, 1972; Escobar-Flores et al., 2018) or gas-bearing glanding (Marshall, 1979; Peña et al., 2014), highlighting the need for further studies of the swimbladder of these species.

The main swimbladder organs are the gas gland and the rete mirabile. The gas gland allows oxygen exchange between the blood and lumen of the gas bladder. The rete mirabile (a bundle of closely packed venous and arterial capillaries attached to the gas gland and entering the swimbladder anteriorly) supplies blood to the gas gland and maintains a pressure differential between the blood and gas bladder through a countercurrent system (Marshall, 1960; Denton, 1961; Kuhn et al., 1963). A functional swimbladder has well-developed cells in the gas gland and the rete mirabile. On the other hand, non-functional swimbladders present atrophied cells in those areas due to the contraction of the gas gland and rete mirabile (Horn, 1975; Neighbors, 1992). Physoclistous fishes (species without a pneumatic duct connecting the swimbladder with the gut) must have a well-developed gas gland and a rete mirabile with enough resorptive area to fill and empty the gas bladder through the blood cells.

The aim of the present study was to analyze the swimbladder properties of six bristlemouth species of the genus Cyclothone in the meso- and bathypelagic layers of the North Atlantic Ocean and Mediterranean Sea. We analyzed the swimbladder location within the body and its shape, the increase in fat content, and swimbladder degradation with length and body mass density. One of the main objectives was to obtain an indication of swimbladder functionality by species and lengths, that is, the probability of the swimbladder containing gas.

The biological data examined in this study were collected from three research cruises in the Atlantic: Bathypelagic (24 May–23 June 2018), from northwest Africa (20°N 20°W), to the south of Iceland (60°N 20°W), SUMMER (30 September–18 October 2020), from the western Mediterranean to the Atlantic around the Iberian Peninsula, and IDEADOS (12 July–29 July 2010), from the Northwest Mediterranean Sea (40°N 2°E to 40°N 2.45°E) on board the R.V. “Sarmiento de Gamboa” (see Figure 1 ). Specimens used to examine swimbladders were collected during the Bathypelagic and SUMMER surveys; those for body-mass density came from the SUMMER cruise, while distribution data were obtained on the Bathypelagic and IDEADOS cruises. Samples were obtained with the “Mesopelagos” midwater trawl net (Meillat, 2012; Castellón and Olivar, 2023) and MOCNESS-1 m² zooplankton net (Wiebe et al., 1985) during day and night hauls carried out from the surface to 700–800 m during the SUMMER and IDEADOS cruises, and to 1,900 m on the Bathypelagic cruise. The Mesopelagos net had an average mouth opening of 29–35 m² and a total length of 58 m. The gear consists of graded-mesh netting starting at 30 mm and ending at 4 mm. A multi-sampler (Castellón and Olivar, 2023) was installed at the cod-end to collect samples from five different depth layers. Depth samples on the Bathypelagic cruise were obtained from 1,900 to 1,200 m, 1,200 to 800 m, 800 to 500 m, 500 to 200 m, and 200 to 0 m, except for station 7 that only reached 600 m depth (see Sarmiento-Lezcano et al., 2022). For the SUMMER cruise, the depth layers were 700–600 m, 600–500 m, 500–400 m, 400–300 m, 300–200 m, 200–100 m, and 100–0 m. On both cruises, the ship speed during the hauls was maintained at 2–2.5 knots. Specimens collected with the Mesopelagos net were immediately identified, measured (standard length, SL), and used for body mass density (ρ) ( Table 1 ).

The MOCNESS-1 m² net was fitted with a 0.2 mm mesh size, and the ship speed during deployment and retrieval was maintained between 1.5 and 2.5 knots to obtain a net angle between 40° and 50°, and the winch retrieval rate was fixed at 0.3 m s⁻¹ on the IDEADOS, Bathypelagic, and SUMMER cruises (see Olivar et al., 2012 for the IDEADOS cruise, Sarmiento-Lezcano et al., 2022 for the Bathypelagic cruise, and Olivar et al., 2022 for the SUMMER cruise). The sampling layer depths were the same as for the Mesopelagos on the Bathypelagic and SUMMER cruises (see above), and for IDEADOS samples were obtained from 900 to 800 m, 800 to 700 m, 700 to 600 m, 600 to 500 m, 500 to 400 m, 400 to 300 m, 300 to 200 m, and 200 to 0 m depths. Biological samples were preserved in 5% buffered formalin, and specimens were sorted out and measured later in the laboratory. Shrinkage correction for formalin preservation was not applied because all our specimens were transformations, juveniles (>12 mm of SL, see Olivar et al., 2018), or adults (according to fish length at 50% maturity (L₅₀); Miya and Nemoto, 1986; McKelvie, 1989; Miya and Nemoto, 1991), and shrinkage in pelagic fishes is negligible at those developmental stages where the degree of ossification is high (Theilacker, 1980). Samples from this net were used for histology and morphology.

Individuals collected during the Bathypelagic cruise were used to measure the sizes of individuals and the morphological swimbladder measurements. Body-mass density (ρ) and histology examinations were carried out on fish collected during the SUMMER cruise ( Table 1 ). Information on distribution, abundance, and size frequency distributions was gathered from the Bathypelagic and IDEADOS cruises. The abundances within each layer of the water column are given as the number of individuals per 1,000 m⁻³ of water filtered by the net in each sampled layer.

Five species of the Cyclothone genus were collected in the North Atlantic Ocean (C. braueri, C. pseudopallida, C. pallida, C. livida, and C. microdon), while only C. pygmaea and C. braueri were collected in the Mediterranean Sea. The most abundant species were C. braueri, C. livida, and C. microdon, accounting for >80% of total Cyclothone, while in terms of biomass, the major contributions were from C. microdon, C. livida, and C. braueri. Shallower species (C. braueri and C. pseudopallida) were found between 200 and 600 m depths, while the deeper species (C. pallida and C. microdon) were mainly found below 800 m depths. C. microdon increased in abundance northward of 30°N, and Cyclothone livida was mainly sampled in the oceanic upwelling off Northwest Africa from 400 to 1,900 m depths. In terms of biomass, most of C. braueri biomass was observed between 200 and 600 m depths from 20 to 60°N. Cyclothone pseudopallida and C. livida showed a wider biomass vertical distribution (200–800 m depths and 450–1,900 m depths, respectively), but the latter was only found from 18° to 45°N. The biomass of C. microdon showed the widest spatial (18°–54°N) and vertical distributions (450–1,900 m depths).

Standard length ranges ( Figures 3 , 4 ) were narrower for C. braueri, C. pseudopallida, and C. pygmaea (range: 10–35 mm SL) than for the other species (range: 11–55 mm SL). No significant differences in length were found between the day and night samples. Most species showed a bimodal size distribution at every 200 m depth stratum. Shallower species were smaller than deeper species (Shallower, 200–400 m: 10–33 mm SL; 400–600 m: 10–33; 600–800 m: 9–23 mm SL; 800–1,000 m: 10–35 mm SL; Deeper, 200–600 m: 14–42 mm SL; 800–1,200 m: 11–51 mm; 1,200–2,000 m: 15–55 mm SL).

Histological examination of the swimbladder features was carried out for C. braueri (n = 80), C. pseudopallida (n = 14), C. pallida (n = 55), C. pygmaea (n = 18), C. livida (n = 8), and C. microdon (n = 19). The swimbladder of these species is located above the gonads and the gastro-intestinal system (GI) and below the spinal cord (SC) and notochord (NT). It begins at the level of the gastro-intestinal system and ends around the origin of the pelvic fin ( Figure 5 ). C. braueri and C. pseudopallida are shallower, semi-transparent species for which the swimbladder was measured with a microscope, while darker species comprised the second group that were analyzed exclusively through histology (no information on swimbladder size). In general, all species showed gas gland and rete mirabile (as physoclistous fishes; Figure 6 ). However, the deeper species with larger sizes presented reduced gas gland and rete mirabile (0.032 and 0.004 mm², respectively), i.e., reduced total and cell area, a lower number of capillaries, and smaller capillaries than other deep species of smaller size (C. pallida; 0.053 and 0.003 mm²) (see Figures 7 and Table 3 ), indicating a deterioration in swimbladder functionality with age and depth.

The body mass density (ρ) of the four Cyclothone spp. (C. braueri, C. pseudopallida, C. pygmaea, and C. pallida) sampled from the surface to 700 m depth in the Mediterranean Sea ranged from 1.055 to 1.072 g cm⁻³ (1.061 ± 0.004 g cm⁻³; g = 1.026–1.045). Significant differences were observed among the three species (F_2,124 = 14.19, p<0.001; C. braueri: 1.061 ± 0.004; C. pallida: 1.057 ± 0.003; and C. pseudopallida: 1.062 ± 0.004 g cm⁻³; Table 7 and Figure 8A ), where the Tukey test showed significant differences between C. braueri and C. pallida (p<0.001). There was a slight negative correlation between Cyclothone body mass density and standard length (r = −0.39; N = 127; p<0.001) and depth (r = −0.39; N = 127; p<0.001). Although we found a rather low coefficient of determination (R²), the correlation between Cyclothone body-mass density and standard length was significant (r² = 0.15; p<0.001), and significant differences in body-mass density could be established for C. pseudopallida (p<0.05, Tables 5 , 7 and Figure 8 ). The species of Cyclothone with lower mass density (C. pseudopallida and C. pallida) were sampled at 450 and 650 m depth, respectively, while species with high ρ values (C. braueri and C. pygmaea) were found between 350 and 550 m ( Figure 8B ).

This study provides basic biological parameters (length frequency distribution, swimbladder location within the body, fat tissue content, morphology, and morphometry) and acoustic properties (body mass density, density contrast, and gas volume required for neutral buoyancy) of bristlemouth species analyzed in the North Atlantic Ocean and Mediterranean Sea that are needed to improve the accuracy of backscattering models. Species diversity was different between the Atlantic (C. braueri, C. pseudopallida, C. pallida, C. livida, and C. microdon) and the Mediterranean (C. braueri and C. pygmaea) as previously reported in the literature (Goodyear et al., 1972; Badcock and Merrett, 1976; Badcock, 1982; Roe and Badcock, 1984; Olivar et al., 2012; Sarmiento-Lezcano et al., 2022). However, the only coincident species (C. braueri) had similar length and depth distributions in both areas. Our results as well as previous research showed that Cyclothone spp. can be grouped according to their distribution in the water column. Shallower species such as C. braueri and C. pseudopallida were found in the upper mesopelagic layer (Sarmiento-Lezcano et al., 2022), C. pygmaea in the lower mesopelagic layer (Goodyear et al., 1972), and C. pallida, C. livida, and C. microdon were observed in the meso- and bathypelagic layers (Badcock and Merrett, 1976; Badcock and Merrett, 1977; Marshall, 1979; Miya and Nemoto, 1986; Sarmiento-Lezcano et al., 2022). The Cyclothone species occupying the shallower layers tend to be more “transparent,” while those found in deeper layers are “darkly colored” (Badcock, 1984). Cyclothone species are non-migrant organisms (Badcock and Merrett, 1976; Olivar et al., 2017; Sarmiento-Lezcano et al., 2022), widely distributed in the water column (from the mesopelagic to the bathypelagic zone), whose vertical distribution could be conditioned by water temperature, oxygen (Sarmiento-Lezcano et al., 2022), feeding type (Bernal et al., 2015; Thompson and Kenchington, 2017), and size of individuals (Agersted et al., 2021).

The size range of Cyclothone captured in the North Atlantic Ocean and Mediterranean Sea was like previous observations in the Atlantic Ocean (Equatorial and Tropical Atlantic, Olivar et al., 2017; Eastern of the Mid-Atlantic Ocean, Agersted et al., 2021). Badcock and Merrett (1976) showed a positive correlation between size and depth for C. braueri, C. pseudopallida, and C. microdon. In our study, we observed a similar correlation in C. pallida, C. livida, and C. microdon as larger individuals were found in the deeper layer (bathypelagic zone). Shorter species displayed a shallower location in the water column, close to the productive epipelagic layer. Deeper species appeared more adapted to minimizing energy expenses by limiting consumption rates and energy intake (Maynard, 1982). Species such as Cyclothone pallida, C. livida, and C. microdon showed a bimodal size distribution mainly in the bathypelagic zone, while C. pygmaea showed a multimodal size distribution between 600 and 800 m depth. Therefore, although deeper species are prone to be larger, less abundant modes of shorter species often cohabit at the same depths.

The swimbladder of shallower species was measured directly thanks to their semi-transparency. Measurements of swimbladder length, tilt angle, aspect ratio, and volume were lower in Cyclothone spp. than in other larger mesopelagic species previously examined (Kloser et al., 2002; Yasuma et al., 2003; Yasuma et al., 2006; Yasuma et al., 2010; Scoulding et al., 2015; Sobradillo et al., 2019; Escobar-Flores et al., 2020). Although the studied C. pseudopallida were larger than C. braueri (19.06 ± 4.02; 17.68 ± 3.63 mm SL, respectively), the lengths of their swimbladders were similar (2.25 ± 0.70; 2.13 ± 0.69 mm, respectively). Both species showed a negative allometric growth of the swimbladder versus body length. That is, the body grows relatively faster than the swimbladder. C. braueri, on the other hand, showed relatively faster growth of the swimbladder. Yasuma et al. (2010) observed that the swimbladder of myctophids showed positive allometric growth after the onset of gas secretion into the swimbladder. However, regressive growth or atrophy of the swimbladder is also common among myctophids (e.g., Diaphus theta and Symbolophorus californiensis) and a few other mesopelagic fishes during later stages (Butler and Pearcy, 1972; Neighbors and Nafpaktitis, 1982; Yasuma et al., 2010). The equivalent spherical radius (ESR) values calculated in this study from direct swimbladder measurements (C. braueri: 0.27–1.03 and C. pseudopallida: 0.36–0.87 mm) agreed with those modeled for Cyclothone spp. by Agersted et al. (2021) (0.25–1.16 mm) and with previous measurements for C. braueri (0.6 mm) in Peña et al. (2014).

Histology is a technique often employed to study the swimbladder and its different components, such as the gas gland and rete mirabile, in non-transparent species (Marshall, 1960; Denton, 1961; Kuhn et al., 1963; Horn, 1975; Neighbors, 1992). This study shows that shallower Cyclothone spp. (C. braueri and C. pseudopallida) have functional physoclistous swimbladders with well-developed gas glands and rete mirabiles and with numerous capillaries. However, deeper, and larger Cyclothone species (C. pygmaea, C. pallida, C. livida, and C. microdon) present a fat-invested swimbladder and a contraction in the gas gland and rete mirabile. On the other hand, the coexistence of two length modes at the same depth strata seen above implies that a small percentage of the organisms sampled in deeper waters could still have some gas in their swimbladder. Horn (1975) also observed a regression of the gas gland in stromateoid fishes. As these species are larger than Cyclothone spp., their gas gland area (0.6–8.7 mm²) and rete mirabile (0.4–2 mm) are also proportionally longer and start regressing at larger sizes (40–130 mm SL). Macdonald (1975) found a correlation between rete mirabile length and depth in the case of non-regressed swimbladders and a reduction in the number of capillaries in deep species due to degeneration of the swimbladder, as also found for Cyclothone spp. in our study (see Table 3 ).

The accumulation of lipids in deeper species provides stable buoyancy and a long-term energy reserve (Butler and Pearcy, 1972). Horn (1975) found that the swimbladder degenerates with increasing fish size, first becoming non-functional and ultimately existing only as a small remnant or disappearing completely. Similarly, Neighbors (1992) reported the atrophy and degradation of cellular components in the swimbladder with age, as observed in this study. It is difficult to establish the exact length at which the swimbladder becomes non-functional for deeper Cyclothone species, and we have no evidence of the gas content, but the high fat content and the indications of swimbladder degradation found in this study suggest that a high percentage of those individuals have no gas in their swimbladders. Swimbladder regression is a gradual process that makes it difficult to determine the exact length at which functionality is lost and the swimbladder stops to secrete or reabsorb gas. However, the large biomass of these species in the lower mesopelagic and bathypelagic zones, especially at northern latitudes (Sarmiento-Lezcano et al., 2022), and the low-scattering layers recorded acoustically (Peña, 2019), suggest that no gas is present in their swimbladders. Davison (2011) allocated small and large fish species to three groups based on the presence of gas in the swimbladders. The author observed that fish in group I showed increasing gas volume with increasing size, while fish in group II exhibited a decrease in the percentage of gas in the swimbladder with age, filling it with lipids. Species in group III never have inflated swimbladders. Davison (op. cit.) also observed the presence of gas in specimens of shallower species in the upper mesopelagic layer (C. pseudopallida, range: 18–43 mm, and C. signata, range: 15–35 mm), but not for deeper species in the bathypelagic layer (C. acclinidens and C. pallida). Marshall (1960) reported the presence of fat tissue in all Cyclothone spp. Examined, with the swimbladder fully adipose in adults of C. braueri (larger than 35 mm) and C. livida (larger than 37 mm). We did not find this organ regression in C. braueri, probably because our largest specimens were smaller than 31 mm. By contrast, we found swimbladder regression in C. pygmaea, C. pallida, C. livida, and C. microdon starting at 16.5, 17.7, 16, and 21.4 mm SL, respectively. Our predictive model determined that 95% of the C. pygmaea, C. pallida, C. livida, and C. microdon swimbladders were occupied by fat tissue at 29.2, 35.8, 40.7, and 36 mm SL, respectively (see Figure 4 ). We found fat tissue content in the swimbladder was closely related to fish size and depth location in this deeper species (the larger and deeper, the more fat content), in contrast with the high gas content in the upper mesopelagic zone for C. braueri and C. pseudopallida. Moreover, semi-transparent species showed relatively faster growth in body length than swimbladder volume and equivalent spherical radius. On the other hand, body-mass density and gas required for neutral buoyancy were higher in C. braueri and C. pygmaea collected between 350 and 550 m than in C. pallida and C. pseudopallida sampled in the 450–700 m layer.

Meso- and bathypelagic fish swimbladder morphology can differ among species as intra- and interspecific variability is common (Marshall, 1951; Marshall, 1960). A functional physoclistous swimbladder in meso- and bathypelagic species is composed of a well-developed gas gland and a rete mirabile with numerous capillaries (Marshall, 1960; Denton, 1961; Kuhn et al., 1963; Horn, 1975; Neighbors, 1992). However, in some cases the swimbladder regresses and replaces its gas with fat tissue as the organism grows (Marshall, 1951; Marshall, 1960). Thus, the swimbladder volume occupied by gas decreases, and their walls thicken (Kleckner and Gibbs, 1972). Some mesopelagic fishes of the Myctophidae family also accumulate fat tissue in their swimbladders, although, in contrast to Cyclothone spp., adults of myctophid species with fat-invested swimbladders (S. leucopsarus and D. theta) can perform extensive vertical migrations without the need to regulate the swimbladder to the surrounding pressure (Taylor, 1968).

Density and sound speed contrast between the fish body and the surrounding seawater are also important parameters for acoustic scattering models that vary with species, morphology, and vertical distribution (Agersted et al., 2021). The proportion and composition of lipids in mesopelagic species vary with area or season as well as within species (Butler and Pearcy, 1972; Kayama and Nevenzel, 1974). Cyclothone density contrast values obtained in this study (1.02 to 1.04) were within previous reported ranges (0.98 to 1.07) for mesopelagic (Stenobrachius leucopsarus, S. nannochir, Lampanycrus rirteri, L. regalis, D. theta, Tarletonbeania crenularis, Protomyctophum thompsoni, P. crockery; Butler and Pearcy, 1972) and epipelagic fishes (G. morhua; Chu et al., 2003). Although there was some indication of density contrast variation with depth, the limited depth strata evaluated in this study (<700 m depth) precluded us from establishing a clear relationship. However, it has been reported that Gonostomatidae fish species contain 3.1%–30% of lipids in relation to dry wet in the mesopelagic zone and 44% in the bathypelagic zone (Nevenzel and Menon, 1980). Body mass density differences between mesopelagic and bathypelagic Cyclothone species were also reported in previous studies (Childress and Nygaard, 1973; Kayama and Ikeda, 1975; Nevenzel and Menon, 1980; Bailey and Robison, 1986; Neighbors, 1988; Stickney and Torres, 1989; Childress et al., 1990; Davison, 2011), with shallower Cyclothone species (mesopelagic C. pseudopallida and C. signata) individuals showing a range of 1.048–1.078 g/ml) and the Cyclothone with regressed swimbladder (meso- and bathypelagic C. acclinidens and C. pallida individuals showing a range of 1.038–1.078 g/ml) (Davison, 2011; Agersted et al., 2021).

The gas required for neutral buoyancy (VG) in this study (0%–4%) was within the range (0%–5%) reported for mesopelagic species with different standard length and body density in Davison (2011), including some Cyclothone species such as C. pseudopallida VG: 2.1%–3.1%, ρ: 1.0500–1.0600; SL: 18–43 mm), C. signata (VG: 1.9%–4.7%, ρ: 1.0475–1.0775; SL: 15–35 mm), C. atraria (VG: 1.2%–2.9%, ρ: 1.0400–1.0575; SL: 21–47 mm), C. acclinidens (VG: 1.0%–4.7%, ρ: 1.0375–1.0775; SL: 25–61 mm), and C. pallida (VG: 1.2%–2.9%, ρ: 1.0400–1.0575; SL: 30–68 mm). A slight decrease in VG with depth was observed for organisms collected between 350 and 550 m and those sampled between 450 and 700 m.

Cyclothone spp. are known to account for much of the echo of the very dense scattering layer detected at 400–700 m depth at 38 kHz (Peña et al., 2014; Ariza et al., 2016) and could be responsible for deeper layers detected in the bathypelagic zone (Peña et al., 2021). To further pursue the acoustic study of these species, we need to employ accurate scattering models. The swimbladder gas content determines whether a gas-bearing or fluid-like model needs to be employed (Peña et al., 2023a, Peña et al., 2023b). This study contributes to a better understanding of the swimbladder regression process in Cyctlothone spp. and the areas of the northeast Atlantic Ocean and the Western Mediterranean Sea where each acoustic typology could be more expected. However, it also points out the need for the development of intermediate models that consider a partially regressed swimbladder that is filled with both lipids and gas. Further knowledge of the point at which gas is no longer needed (when lipids provide enough buoyancy) is also required.

In summary, swimbladder histological analysis revealed a physoclistous swimbladder structure in the six Cyclothone spp. analyzed. For upper mesopelagic species (C. braueri and C. pseudopallida), the non-regressed swimbladder remained throughout development. However, for the species in the lower mesopelagic and bathypelagic zones (C. pygmaea, C. pallida, C. livida, and C. microdon), the presence of fat tissue varied within species and sizes. A regression model estimated that for this last group of species, the swimbladder of Cyclothone larger than 36–41 mm SL would be completely invested in fat tissue. Thus, while all Cyclothone species possess a swimbladder, only small fish in early stages and semi-transparent species inhabiting the upper mesopelagic zone can be expected to contain gas. These species should be acoustically modeled as gas-bearing organisms. Species in the lower mesopelagic zone have some degree of swimbladder regression but could contain some gas. As the swimbladder area with no fat tissue showed irregular shapes, acoustic models representing prolate spheroids are probably not accurate for these organisms. On the other hand, around 90% of the abundance in the bathypelagic zone (mostly of C. microdon) is expected to have a gas-depleted swimbladder (particularly in northern latitudes), thus behaving acoustically as a fluid-like species. All the information gathered in this study can be used to improve the acoustic backscattering modeling of Cyclothone spp. and thus better infer knowledge from recorded echograms.

Publicly available datasets were analyzed in this study. This data can be found here: Olivar et al. (2022) https://doi.org/10.1594/PANGAEA.947631 and Sarmiento-Lezcano et al. (2023) https://doi.org/10.1594/PANGAEA.95494.

The animal study was reviewed and approved by the Spanish Ministry of Science and Technology proyect (Bathypelagic, CTM2016-78853-R).

AS-L contributed to the species identification, net data collection and processing and carried out the swimbladder measurement and the density experiment. MO contributed to the species identification, data analyses, supervision, and acted as a cruise leader in the SUMMER cruise. MJC performed and examined the histological analysis of the species. MC participated in the density experiment. SH-L designed and promoted the Bathypelagic project and acted as cruise leader in the Bathypelagic cruise. AC designed and controlled the MESOPELAGOS net during both cruises. MP designed and supervised this study and contributed to data curation, data processing, and editing of the manuscript. All authors contributed to article writing and approved the submitted version.