INTRODUCTION
The NW Mediterranean continental margin is the historical focus of a strong fishing pressure, which started during the middle of the past century (Sardà et al., Reference Sardà, Cartes and Norbis1994). Large proportions of species caught by commercial bottom trawls have no market value and are discarded (Sánchez et al., Reference Sánchez, Demestre and Martín2004). Several of these by-catch items are of ecological importance because of their large abundance and overall biomass within the catch itself (Alverson et al., Reference Alverson, Freeberg, Murawski and Pope1994). Therefore, any human impact assessment in the continental margins requires detailed data on the life history of both target and non-target species, in order to foster an ecosystem management approach to fishery, as required by modern marine EU policies (Marine Strategy Framework Directive; 25-0-2008). However, the available biological data on by-catch species are scarce worldwide (Hall & Mainprize, Reference Hall and Mainprize2005). The Mediterranean tonguefish Symphurus nigrescens and Symphurus ligulatus are examples of this situation. Both species are of ecological importance in local demersal communities given their abundance (Stefanescu et al., Reference Stefanescu, Morales-Nin and Massuti1994, Fernandez-Arcaya et al., Reference Fernandez-Arcaya, Ramirez-Llodra, Rotllant, Recasens, Murua, Quaggio-Grassiotto and Company2013a). Therefore, their populations often appear in large numbers within commercial hauls (Carbonell et al., Reference Carbonell, Massutí, Reñones and Álvarez1995).
The available information on S. nigrescens and S. ligulatus reproductive biology is scarce. These two species are the only pleuronectiform of the family Cynoglossidae inhabiting the Mediterranean slope (Lloris et al., Reference Lloris, Rucabado, Cerro, Portas, Demestre and Roig1984), between 100–700 m and 400–1300 m, respectively (Massutí et al., Reference Massutí, Stefanescu and Morales-Nin1995; Fernandez-Arcaya et al., Reference Fernandez-Arcaya, Ramirez-Llodra, Rotllant, Recasens, Murua, Quaggio-Grassiotto and Company2013a). Symphurus nigrescens is distributed throughout the Mediterranean (Blauchot, Reference Blauchot1987), while S. ligulatus is restricted to the western (Matallanas, Reference Matallanas1984) and eastern basins (Koukouras, Reference Koukouras2010). The two species are mesobenthonic (Massutí et al., Reference Massutí, Stefanescu and Morales-Nin1995) and feed on benthic invertebrates (Macpherson, Reference Macpherson1978; Cau & Deiana, Reference Cau and Deiana1979). Symphurus nigrescens is normally associated with the asteroid Brisingella coronata and the coral Funicula quadrangularis (Maurin, Reference Maurin1962), while S. ligulatus is commonly found on bamboo coral (Isidella elongata) communities (Maurin, Reference Maurin1962; Torchio, Reference Torchio1971; Tortonese, Reference Tortonese1975).
Mediterranean demersal resources are exposed to a continuous seasonal and spatial clinal variation acting on individual species biology (Aguzzi et al., Reference Aguzzi, Company, Bahamon, Flexas, Tecchio, Fernandez-Arcaya and Canals2013a, Reference Aguzzi, Costa, Ketimaier, Angelini, Antonucci, Menesatti and Companyb; Fernandez-Arcaya et al., Reference Fernandez-Arcaya, Ramirez-Llodra, Rotllant, Recasens, Murua, Quaggio-Grassiotto and Company2013a). In the NW Mediterranean slope, particle fluxes from the photic zone to the deep sea increase in autumn–winter in response to major coastal storms and during spring–summer water stratification prevents such a flux, decreasing the overall energy availability in deep-sea realms (Lopez-Fernandez et al., Reference Lopez-Fernandez, Bianchelli, Pusceddu, Calafat, Sanchez-Vidal and Danovaro2013). In this context, S. nigrescens apparently spawns between June and October (Macpherson, Reference Macpherson1978), whereas S. ligulatus spawns between May and November (Cau & Deiana, Reference Cau and Deiana1979). However, a more in-depth analysis of their reproductive strategy has not yet been conducted. Accordingly, our aim is to present a comparative analysis along a bathymetric gradient of the size frequency distribution and reproductive patterns of S. nigrescens and S. ligulatus in the NW Mediterranean over an annual cycle.
MATERIALS AND METHODS
Sample area and data collection
Megafaunal sampling was conducted between April 2003 and May 2004 in the Blanes canyon and the adjacent open slope (Figure 1). A total of 37 hauls were conducted between 300 and 1500 m depth with commercial and scientific vessels, even though the stratum between 701 and 800 m depth was not sampled. The individuals of Symphurus nigrescens and S. ligulatus were caught in 28 of those hauls each (Table 1). Fish were caught using a commercial Otter Trawl Bottom System (OTBS) and its scientific equivalent, the Otter Trawl Maireta System (OTMS), both with a cod-end net of 12 mm stretch mesh size (Sardà et al., Reference Sardà, Cartes, Company and Albiol1998). The catchability of these two different gears does not differ significantly in either the catch composition or the size frequencies of the species sampled (Sardà et al., Reference Sardà, Recasens, Abelló, Rotllant and Molí2002).
Fig. 1. Study area and bottom trawl fishing stations (lines).
Table 1. Depth range (m) and season for the number of individuals caught, measured, sexed and macro and microscopically analysed of Symphurus nigrescens (Sn) and Symphurus ligulatus (Sl). SPR, spring; SUM, summer; AUT, autumn; WIN, winter.
On board, the length of individuals was measured from the mouth tip to the edge of the tail to the nearest 0.5 cm and they were weighed to the nearest 0.1 g. They were immediately dissected and their sex was determined by macroscopic examination of the gonads (see below).
Population distribution
In order to obtain population density estimation, catch data for both species were standardized to the unit of swept area (ind. km−2), as estimated from Scanmar sensors. These sensors provide information on horizontal average opening of gear doors and the distance travelled between the initial and final position of the gear. To assess the occurrence of a differential bathymetric distribution between the two species, standardized density data were pooled by every 100 m depth interval and represented by box-plotting.
Kolmogorov–Smirnov non-parametric tests were used to evaluate the occurrence of significant depth-related differences in species' size-frequency distribution. An Analysis of Variance (ANOVA) was used to test overall differences in size between the two species. Chi-squared tests were used to analyse sex ratios in relation to depth. Additionally, Mann–Whitney non-parametric tests were used to assess the occurrence of any significant sex dimorphism in size.
Reproductive biology
Sex was determined by macroscopic examination of the gonads. Females of both species were classified into a five-stage maturity scale according to their external appearance, based on currently accepted standardized procedures (Brown-Peterson et al., Reference Brown-Peterson, Wyanski, Saborido-Rey, Macewicz and Lowerre-Barbieri2011 modified by Fernandez-Arcaya et al., Reference Fernandez-Arcaya, Recasens, Murua, Ramirez-Llodra, Rotllant and Company2012): I: immature (with undifferentiated gonads); II: regenerating-developing; III: spawning capable; IV: actively spawning; and V: regressing.
The reproductive period in both species was established as the variation in the percentage frequency of the females’ maturity stages. The Gonadosomatic Index (GSI) formula was then computed over the year (Anderson & Gutreuter, Reference Anderson, Gutreuter, Nielsen and Johnson1983):
$${\rm GSI}\, = \,({\rm Wg})/({\rm Wt} - {\rm Wg}) \times 100$$where, Wg is the gonad weight and Wt is the total gonad-free weight.
In order to validate the visually classified maturity stages, a subsample of the females was dissected for gonad extraction (76 for S. nigrescens and 64 for S. ligulatus) (Table 1). The gonads were preserved in 10% buffered formaldehyde for histological analysis. The males were not considered in this analysis due to the difficulty in extracting the entire gonad. Samples were dehydrated in graded alcoholic baths, embedded in paraffin wax and sectioned by microtome. The sections of 7 μm were stained with Harris’ haematoxylin and eosin. The oocyte developmental stages were described according to Brown-Peterson et al. (Reference Brown-Peterson, Wyanski, Saborido-Rey, Macewicz and Lowerre-Barbieri2011) (modified by Fernandez-Arcaya et al., Reference Fernandez-Arcaya, Recasens, Murua, Ramirez-Llodra, Rotllant and Company2012). The average size range of oocytes was calculated measuring the diameters of 100 oocytes of each oocyte developmental stage from 10 different females using Sigma Scan Pro4. Three mature individuals of each species were also selected to analyse the oocyte size frequency distribution. A transverse histological section from one gonad of each individual was used to measure the diameters of all the types of oocytes present with a visible nucleus using image software Sigma Scan Pro4. Finally the histogram of only one of the selected individuals of each species was included in the study (Figure 7) (169 oocytes for S. nigrescens and 335 for S. ligulatus).
RESULTS
Population structure characteristics
A total of 606 Symphurus nigrescens and 405 S. ligulatus were caught in 37 hauls (see Table 1). The analysis of species distribution with depth revealed a differential and only partial overlap between their populations (Figure 2). Symphurus nigrescens occurred in high number at the shallowest sampling depth (i.e. between 330 and 420 m), while S. ligulatus showed a deeper range, with a maximum density between 500 and 600 m depth. For the latter species few (N = 8) individuals were caught down to 800 m depth.
Fig. 2. Box-plot showing the bathymetric distribution of Symphurus nigrescens (S.n) and S. ligulatus (S.l).
In both species, the overall sex ratio was significantly different (Chi-squared test: S. nigrescens χ 2 = 5.76; S. ligulatus χ 2 = 4.84; P < 0.05) being females:males ratio of 1:0.6 in both species. Also, in both species, the females were significantly larger than males (Mann–Whitney U-test: S. nigrescens U = 36330.5, N females = 265, N males = 163; S. ligulatus U = 18464, N females = 184, N males = 121). When the sex ratio was analysed in relation to depth (Table 2), the results showed that S. nigrescens presented a neutral sex ratio at its shallower distribution (i.e. 301–400 m, and 401–500 m) (Chi-squared test: 301–400 m, χ 2 = 2.56; 401–500 m, χ 2 = 0.04; P > 0.05), while females dominated in deeper ranges (i.e. 501–600 m) (Chi-squared test: 501–600 m, χ 2 = 21.16; P < 0.05). Symphurus ligulatus also showed a greater proportion of females at deeper depths (i.e. below 500 m) (Chi-squared test: 501–600 m, χ 2 = 4; 601–700 m, χ 2 = 17.64; 801–1200 m, χ 2 = 4; P < 0.05), while males were predominant at shallower depths (i.e. 301–400 m) (Chi-squared test: 301–400 m, χ 2 = 4; P < 0.05) (see Table 2).
Table 2. Percentage of females and males over total mature individuals by each sampling depth for Symphurus nigrescens (Sn) and S. ligulatus (Sl)
*Significant differences in the sex ratio between females and males (Chi squared, P < 0.05).
a The deepest ranges have been grouped because of the low number of individuals.
Table 3. Description of macroscopic and microscopic features in relation to the ovary developmental stages in Symphurus nigrescens and S. ligulatus. AVtg, advance vitellogenic oocyte; EVtg, early vitellogenic oocyte; GVM, germinal vesicle migration; PG, primary growth oocyte; POF, postovulatory follicle; Sn, Symphurus nigrescens; Sl, Symphurus ligulatus (based on Brown-Peterson et al., Reference Brown-Peterson, Wyanski, Saborido-Rey, Macewicz and Lowerre-Barbieri2011 modified by Fernandez-Arcaya et al., Reference Fernandez-Arcaya, Recasens, Murua, Ramirez-Llodra, Rotllant and Company2012).
Both species showed a size increase according to depth with a higher proportion of females (i.e. as the largest individuals) on their lower realms (Figure 3). However, the observed pattern was not statistically significant (Kolmogorov–Smirnov test, P > 0.05). The size frequency distributions of S. ligulatus between 301–400 m, 401–500 m and below 701 m depth were discarded because of the small sample (N < 20).
Fig. 3. Total size-frequency distribution histograms at different depth intervals for Symphurus nigrescens and S. ligulatus. White bars are immature (undifferentiated), grey bars are females, black bars are males and grey bars with lines are sexually undetermined individuals.
The analysis of size frequency distribution through seasons for S. nigrescens (Figure 4) showed significant differences in autumn (Kolmogorov–Smirnov; D = 0.15; P < 0.05), where the smallest individuals were also captured (i.e. 4.5 cm). In spring and winter, in contrast, individuals of intermediate size were observed (i.e. 6.5 cm and 10 cm, respectively). Symphurus ligulatus also presented differences in its seasonal size distribution. The higher number of small individuals was captured in winter, while in spring the largest proportion of large individuals was found (see Figure 4). No significant differences in the size-frequency distribution between all seasons except in the individuals captured in summer and winter (Kolmogorov–Smirnov; D = 0.689; P < 0.05) were found. The size comparative analysis showed an overall larger size for S. nigrescens than S. ligulatus (ANOVA; F = 63.9; P < 0.005).
Fig. 4. Total size-frequency distribution by season for Symphurus nigrescens and S. ligulatus. White bars are immature (undifferentiated), grey bars are females, black bars are males and grey bars with lines are sexually undetermined individuals.
Reproductive biology
Germinal cells showed similar morphology but differed in size for both species. The vitellogenic oocytes of S. nigrescens were significantly larger than those of S. ligulatus (Mann–Whitney U-test: U = 4432, N 1 = 60, N 2 = 103; P < 0.001). The characteristics of the different gonad developmental stages and the range of oocyte sizes for each maturity stage of the two species are shown in Table 3 and Figure 5.
Fig. 5. Histological sections of ovaries of Symphurus nigrescens (A–B) and S. ligulatus (C–D) at different stages. (A): stage II, developing-regenerating; (B): stage III, spawning capable; (C): stage IV, actively spawning; (D): stage V, regressing. ATR, atretic follicle; AVtg, advanced vitellogenic oocyte; EVtg, early vitellogenic oocyte; GVM, germinal vesicle migration; PG, primary growth oocyte (based on Brown-Peterson et al., Reference Brown-Peterson, Wyanski, Saborido-Rey, Macewicz and Lowerre-Barbieri2011 modified by Fernandez-Arcaya et al., Reference Fernandez-Arcaya, Recasens, Murua, Ramirez-Llodra, Rotllant and Company2012). Scale bars = 100 μm.
Oocytes development in S. nigrescens presented a higher seasonality in ovary development than S. ligulatus (Figure 6). The former species presented a peak of spawning in summer, when 65.3% of females were found in stage IV (i.e. spawning capable), lasting until the autumn months. In spring and winter, all analysed females were in stage II (i.e. developing gonads). The highest GSI values (4.8%) were found in the summer months, coinciding with the most active spawning period (see Figure 6). Symphurus ligulatus showed a semi-continuous reproductive cycle with the presence of females in stages between III and IV (i.e. spawning capable and actively spawning, respectively), starting in spring and lasting until autumn (see Figure 6). However, in winter the large majority of females (94%) were in stage II (developing phase). The GSI data were confirmed by the histological results, which showed a similar trend throughout the year (see Figure 6).
Fig. 6. Ovary maturity stages and gonadosomatic index (GSI) by season for Symphurus nigrescens (Sn) and S. ligulatus (Sl). II–VI: developing-regenerating stage; III: spawning capable stage, IV: actively spawning stage, V: regressing stage.
The oocyte size-frequency distribution of S. nigrescens and S. ligulatus (Figure 7) was continuous in both species according to the criteria of Murua & Saborido-Rey (Reference Murua and Saborido-Rey2003). There was no hiatus between previtellogenic and vitellogenic oocytes, and a random mixture of oocytes at every stage was present in all specimens analysed.
Fig. 7. Oocyte-size frequency histograms for Symphurus nigrescens and S. ligulatus on one spawning capable individual (stage IV).
DISCUSSION
This is the first in-depth study of the reproductive biology of Symphurus nigrescens and S. ligulatus. Both species showed similar reproductive strategies, with asynchronous ovarian organization and similar gonads morphology. However, their depth–size frequency distribution, spawning temporality and oocyte size varied, suggesting that interspecific competition and the different environmental conditions (over their respective depth axes) influence the overall population dynamism.
Both species showed different distribution maxima (i.e. S. nigrescens on the upper slope above 400 m and S. ligulatus on the middle slope between 500–600 m). Variability in environmental factors related to depth, sediment type and diverse feeding habits may account for such a difference (Massutí et al., Reference Massutí, Stefanescu and Morales-Nin1995). Symphurus nigrescens and S. ligulatus apparently bury themselves partially within the sediment, both moving onto it in search of a common and major item for their diet, the decapod Calocaris macandreae (Macpherson, Reference Macpherson1978; Cau & Deiana, Reference Cau and Deiana1979). The inter-specific competition over space use and associated trophic resources may account for a process of bathymetric segregation in these two ecologically related species, as has been already reported in other local strictly demersal Monkfish Lophius budegassa and L. piscatorius (Colmenero et al., Reference Colmenero, Aguzzi, Lombarte and Bozzano2010), other Macrourids (Carrasón & Matallanas, Reference Carrassón and Matallanas2002; Fernandez-Arcaya et al., Reference Fernandez-Arcaya, Rotllant, Ramirez-Llodra, Recasens, Aguzzi, Flexas, Sanchez-Vidal, Lopez-Fernandez, Garcúa and Company2013b), as well as in crustacean decapods (Aguzzi et al., Reference Aguzzi, Bahamon and Marotta2009).
In this study we observed an increase in size for both species according to the sampling depth, confirming the general ‘bigger–deeper’ trend phenomenon of continental margin megafauna (Polloni et al., Reference Polloni, Haedrich, Rowe and Hovey Clifford1979). The observed size–depth related pattern could be attributable to a depth partitioning between sexes. Sex ratio showed a higher proportion of females at greater depths, which have also a larger size than males. Similar results have been found in other species distributed on the NW Mediterranean continental margin (Trachyrincus scabrus, Fernandez-Arcaya et al., Reference Fernandez-Arcaya, Recasens, Murua, Ramirez-Llodra, Rotllant and Company2012; Nezumia aequalis, Fernandez-Arcaya et al., Reference Fernandez-Arcaya, Rotllant, Ramirez-Llodra, Recasens, Aguzzi, Flexas, Sanchez-Vidal, Lopez-Fernandez, Garcúa and Company2013b). The presence of larger females during their spawning period at deeper depths (i.e. summer-autumn in S. nigrescens and from spring to summer in S. ligulatus) suggest that both Cynoglossids might undertake migrations toward deeper areas when spawning. Seasonal modifications in fish species distributions seem to commonly occur in this area (Aguzzi et al., Reference Aguzzi, Company, Bahamon, Flexas, Tecchio, Fernandez-Arcaya and Canals2013a) and have been associated to the seasonal modulation of behaviour over reproductive cycles (Aguzzi & Company, Reference Aguzzi and Company2010; Aguzzi et al., Reference Aguzzi, Company, Costa, Menesatti, Garcúa, Bahamon, Puig and Sardá2011).
The shallower-dwelling S. nigrescens showed a stronger seasonal reproductive pattern than S. ligulatus. The spawning females of S. nigrescens were concentrated mainly in the summer months, coinciding with other observations of this species in other Mediterranean areas (Sabatés & Maso, Reference Sabatés and Maso1992; Sabatés & Olivar, Reference Sabatés and Olivar1996; Somarakis et al., Reference Somarakis, Ramfos, Palialexis and Valavanis2011) and with the reported overall spawning seasonal pattern of the Cynoglossidae family (Gibson, Reference Gibson2005). The reproductive activity of the shallowest distributed species, S. nigrescens, could be strongly affected by the spring-summer phytoplankton blooms (Aguzzi et al., Reference Aguzzi, Company, Bahamon, Flexas, Tecchio, Fernandez-Arcaya and Canals2013a, Reference Aguzzi, Costa, Ketimaier, Angelini, Antonucci, Menesatti and Companyb). In contrast, S. ligulatus showed a prolonged spawning period, ranging from spring to autumn, and coinciding with the general trend of continuous reproductive cycle found for several species inhabiting between 400–800 m depth in the same study area (Fernandez-Arcaya et al., Reference Fernandez-Arcaya, Ramirez-Llodra, Rotllant, Recasens, Murua, Quaggio-Grassiotto and Company2013a). Reproductive characteristics have evolved in response to environmental conditions. However, the factors and processes that ultimately control and shape the reproductive timing of flatfish are still being discussed (Gibson, Reference Gibson2005).
Related to the ovarian organization, it appears to be asynchronous for both species as shown by the oocyte-size distribution, which is characterized by a random mixture of oocytes at every stage of development. A clear separated stock of oocytes between advanced yolked oocytes and hydrated oocytes occurs only at hydration (Murua & Saborido-Rey, Reference Murua and Saborido-Rey2003). In addition, even though they are not very apparent, we observed two slightly discrete modes of oocyte sizes in both species between early vitellogenic and advanced vitellogenic oocytes. However a deeper analysis of the oocyte size frequency distribution should be conducted in order to confirm these observations. If the presence of these two modes was confirmed, together with the observed regressing phase from autumn to spring in S. nigrescens and in spring in S. ligulatus, it would suggest that both species have some kind of seasonal modulation in the production of their oocytes. Such a result suggests that during the females' maturation process a continuous recruitment of previtellogenic oocytes into a ‘vitellogenic oocyte clutch’ might occur. However, that recruitment process seems to cease after the spawning season. Asynchronous ovarian organization has been related to income-breeding species, the energy for reproduction of which is financed by the available energy (Stearns, Reference Stearns1992). The peak of reproductive activity of S. nigrescens and S. ligulatus is linked to the greater availability of C. macandreae in summer (Cartes & Maynou, Reference Cartes and Maynou1994). Our results suggest that both species might use the demographic increase of Calocaris as a timing cue to set their reproduction synchronously, in a way to successfully satisfy the energy requirements for their reproduction.
The spawning of small hydrated oocytes and the presence of an oil drop in hydrated oocytes (not quantified here) suggests that S. ligulatus and S. nigrescens produce pelagic eggs, which are the most common type of eggs in flatfishes. Benthic eggs would otherwise have been larger in these deep-water species due to the usually low temperature and the poor food availability of their environment (Gage & Tyler, Reference Gage and Tyler1992). Flatfish are not an exception and in a comparative analyses conducted among species of this group, an increase in egg size with depth has been observed (Gibson, Reference Gibson2005). However, the species of the present study did not follow a similar depth-driven pattern in their size. Although we could not measure hydrated eggs, the deeper dwelling species, S. ligulatus, showed smaller oocyte size than the shallower species, S. nigrescens. The larger size of the shallower species S. nigrescens and the higher food availability at shallower depths might encourage the development of larger oocytes.
ACKNOWLEDGEMENTS
The authors would like to thank the Officers and crews of the RV ‘García del Cid’, FV ‘Montse III’ and FV ‘Virgen del Vilar’ for helping to provide samples and the data necessary for our study. We also thank J.A. Garcia for producing the study area map. This study was conducted within the RECS projects funded by the Spanish Science and Innovation Ministry CICYT (REN02/04556/C02/MAR to JBC) and the EU-FP7 project HERMIONE (Grant Agreement 226354).
