Population biology of the little gulper shark Centrophorus uyato in Lebanese waters
Abstract
A total of 38 individuals of the heavily exploited little gulper shark Centrophorus uyato were collected from Lebanese coastal waters using bottom longlines and trammel nets of different meshes at depths ranging from 115 to 600 m between May 2013 and February 2014. Their total lengths were between 45 and 94 cm and their total mass was from 870 to 6700 g. The sex ratio was not significantly different from 1:1, with 20 males and 18 females, but bathymetric sexual segregation occurred. Catch per net setting (CNS) was used as a relative abundance index to detect spatial distribution; trammel nets showed largest CNS ranging from 4·9 to 5·45 kg per unit effort in the north and south, at depths from 120 to 200 m, during spring and summer. The mass–length relationships demonstrated negative allometric growth (b < 3) (males: MT = 0.3585LT2·071, r2 = 0·94; females: MT = 0.0239LT2·735, r2 = 0·64). The condition factor as well as the gonado-somatic and hepato-somatic indices of C. uyato in the study area were also calculated. The results are discussed in relation to the distribution, growth and reproduction as well as the management of C. uyato.
Introduction
The Mediterranean Sea is considered one of the seven Chondrichthyan biodiversity hotspots and one of the three main hotspots where the biodiversity of sharks and rays is seriously threatened (Dulvy et al., 2014). Among the 80 species of cartilaginous fish known in the Mediterranean Sea (45 sharks, 34 batoids and one chimaera), 71 were assessed in the frame of the International Union for Conservation of Nature (IUCN) red list and >40% (34 species) are threatened and face an elevated risk of extinction (Cavanagh & Gibson, 2007). The main causes of these threats are linked to anthropogenic activities that degrade marine ecosystems, such as habitat degradation, pollution, littoral construction, tourism and overexploitation (Cuttelod et al., 2008). This last factor particularly affects elasmobranch populations and the fishing activity exerted on these populations is responsible for the reduction in their abundance (Dulvy et al., 2000). The decline of elasmobranch populations in the Mediterranean is a notable consequence of increased fishing effort in terms of engine power and increased capacity of fishing equipment for both artisanal and industrial fisheries. Elasmobranchs can be caught by various fishing methods, although bottom trawling is considered responsible for a large portion of their by-catch and discards throughout the world (Bonfil, 1997).
The taxonomic status of the little gulper shark, Centrophorus uyato (Rafinesque 1810) has long been debated, but it is a valid species according to White et al. (2013), occurring in the Mediterranean Sea along with its larger congener Centrophorus granulosus (Bloch & Schneider 1801). It is a small squaloid shark reaching 110 cm total length (LT) v. 170 cm for C. granulosus, occurring on the continental slopes of the Atlantic Ocean and Mediterranean Sea. Records from the Indo-West Pacific Ocean need confirmation. It is classified as data deficient in the red list of the International Union for the Conservation of Nature, mainly because of the taxonomic problem (Pogonoski & Pollard, 2003).
In fishery catch data, C. uyato is confused with other gulper sharks. Gulper sharks are exploited through by-catch with trammel nets, longlines, gillnets and trawls (Costantini et al., 2000; Pogonoski & Pollard, 2003; Morey et al., 2006) and the fact that they are deep-water sharks inhabiting depths greater than 200 m with late maturity, slow growth and low fecundity makes them susceptible to overexploitation (Compagno, 1984; Stevens et al., 2000). When caught, these sharks are discarded or used as by-product. Their meat is marketed fresh, smoked, dried or salted for human consumption or processed into fish meal, the oil of their liver is a source of squalene (Compagno, 1984), a chemopreventive substance used in cosmetics and pharmacology. The Centrophoridae family has long been subject to overexploitation worldwide; a drastic gulper shark population decline was observed in the north-eastern Atlantic Ocean off the coast of Portugal due to targeting by longliners between 1990 and 2004 (Gibson et al., 2008). In addition, in the north-east Atlantic Ocean, south-west Pacific and Indian Oceans, six Centrophorus Müller & Henle 1837 species have been depleted (Adam et al., 1998; Graham et al., 2001; ICES, 2010). For instance, Centrophorus harrissoni McCulloch 1915 and Centrophorus zeehaani White, Ebert & Compagno 2008 are still subject to by-catch by multispecies fisheries in Australia, despite the measures introduced to end their targeted fishing (ICES, 2005; Forrest & Walters, 2009; White & Kyne, 2010; Graham & Daley, 2011).
No biological studies have been carried out on C. uyato in the Mediterranean Sea because of the taxonomic problem. A few studies exploring only the biology of C. granulosus, were performed in the eastern Mediterranean Sea (Golani & Pisanty, 2000; Megalofonou & Chatzispyrou, 2006), but no studies have been carried out in the coastal waters of Lebanon, the present study area. Some studies in the Mediterranean dealt with the analysis of trace metals (Hornung et al., 1993) and polychlorinated biphenyls (Storelli & Marcotrigiano, 2001) in some organs of C. granulosus. Other Mediterranean studies have evaluated the reproduction of this species, demonstrating its lecithotrophic nature (Guallart Furio & Vicent, 2001). The taxonomy of C. uyato, however, has not been adequately resolved and taxonomic problems have arisen between this species and C. granulosus. Most studies in the Mediterranean (Capapé, 1985; Guallart Furio, 1998; Megalofonou & Chatzispyrou, 2006) refer to the species as C. granulosus.
Lebanese fishing is artisanal and traditional and the most common gears used are trammel nets, longlines, purse-seines and beach seines. Trawling is legally forbidden (Fowler et al., 2005). The majority of the fish caught are bony fish (Majdalani, 2004) and cartilaginous fish landings are particularly low (Fowler et al., 2005). According to Martin et al. (2006), the Lebanese production of elasmobranchs was 60 t in 2003 accounting for 2% of all groups caught. Centrophorus uyato, however, are often seen in Lebanese fish markets as by-catch species that are not discarded and sold as fish fillets. They have a low commercial value. The lack of management measures (and in particular, any specific to elasmobranchs) and the increase of commercial fishing have created the need to manage the stocks of such elasmobranch species in this region.
Biological information necessary for stock assessment is lacking for many of the eastern Mediterranean cartilaginous species, including minimum size, maximum size and average size, as well as length–mass relationships. Nonetheless, these data are necessary for understanding growth rate, age structure and other aspects of population dynamics. Also, conversion factors have a practical value in fisheries management. One measure currently needed is the establishment of minimum size limits and a minimum mass for C. uyato catches. Because sizes must be estimated at sea, means for converting length to mass are essential to fishers. The aim of the present study was to determine the main biological and ecological aspects of the C. uyato population in Lebanese marine waters and contribute to better knowledge of this population.
Materials and methods
Sampling and study area
A total of 38 C. uyato was collected along the coast of Lebanon between May 2013 and February 2014 at depths ranging from 115 to 600 m. The individuals collected came from two sources. The first was catches of the major Lebanese longline fisheries (16 collected individuals) and the second was catches of the scientific survey of the PESCA-Libano programme performed within the framework of the CANA programme, establishing monitoring and sustainable development of the Lebanese Sea, of the Lebanese National Council for Scientific Research (CNRS-L) during the same time period (22 collected individuals). In the latter survey, gillnets and trammel nets (T) with different meshes (26 and 30 mm: T4 and T6, respectively), along with longlines (L), were lowered at different sites with predetermined depths to evaluate the potential of the marine coastal resources to support the Lebanese government in strengthening management of marine resources. The study area showing the major Lebanese ports and Centrophorus catches in the PESCA-Libano scientific survey are shown in Fig. 1. The exact positions of the fishery catches were not available because the majority of Lebanese fishermen have no global positioning system (GPS) instruments, but depths were provided. The specimens were identified according to the two field identification guides of Bariche (2012) and Golani et al. (2006). They were preserved in ice and later frozen at −20° C for examination in the laboratories of the CNRS-L National Centre for Marine Sciences, Batroun.
, size of the circles is proportional to the number of individuals caught). 
, PESCA-Libano hauls where no C. uyato were caught.Morphometric and biological measurements
The specimens were examined for external damage and sexed macroscopically using the male claspers. The external length of the clasper (LC) in males was measured to the nearest mm from the tip of the clasper to the base of the pelvic fin using callipers. All lengths and masses were measured to the nearest cm and g, respectively, using an electronic balance and a measuring tape.
Morphometric measurements were taken from each C. uyato specimen. The total mass (MT), total length (LT) and fork length (LF) were measured prior to dissection. The liver and gonads were removed, along with the expanded uteri in the case of a gravid females. The digestive tract was also removed and the eviscerated mass (ME) noted.
The hepato-somatic index (IH) gives a useful indication of energy reserves because the liver is important for energy storage in fish (Halton et al., 2001). Therefore, the mass of the liver (ML) was measured and used for the computation of IH (Bulow et al., 1978; Adams & McLean, 1985): IH = MLMT−1100
The gonado-somatic index (IG) was used to investigate changes in testes and ovary masses in relation to the mass of fish. Thus, the mass of the gonads (MG), including the masses of the testes, Needham's sac, spermatophoric complex in males and the oviducal glands, accessory glands, ovary and oviducts in females (Gristina et al., 2006), was also measured for the computation of IG (Bougis, 1952): IG = MGMT−1100
Sexual maturity was determined using the instruction manual for the MEDITS survey 6 (MEDITS, 2012). The aplacental viviparous elasmobranch females were classified using 8 maturity stages: 1, immature; 2, developing; 3a, capable of reproducing; 3b, early gravidity; 3c, mid gravidity; 3d, late gravidity; 4a, regressing; and 4b, regenerating. Males were classified into five maturity stages: 1, immature; 2, developing; 3a, spawning capable; 3b, actively spawning; and 4, regressing.
Finally, the condition factor (K) (Fulton, 1904) was also computed using total mass (MT) and total length (LT): K = MTLT−3100, The index is commonly used to compare the condition, fatness, or well-being of fish (Tesch, 1968), based on the assumption that heavier fish of a given length are in better condition (Froese, 2006).
Catch per net setting
During the PESCA-LIBANO survey, longlines with size 5 hooks and trammel nets with meshes of 22, 24, 26, 28 and 30 mm were lowered at specific sites at predetermined coordinates and depths. Catch-per-net setting (CNS, kg) for the C. uyato samples caught during the survey (22 individuals) was calculated from the sum of the total masses of C. uyato caught in a specific gear in relation to the total number of C. uyato caught in the same gear, per unit of time. The CNS was also used to provide relative abundance indices for each area at specific times of the year (seasons) or for each specific gear.
Regressions and statistical analyses
All regressions and statistical analyses were carried out in R (www.r-project.org) and Quantum Geographic Information System (qGIS) software (QGIS Development Team, 2012) was used for the maps.
Total length and MT were used to produce a mass–length relationship in a non-linear power regression (MT = aLTb) (Le Cren, 1951), where a and b were the parameters to estimate using the nls() function in R. The coefficient of determination (r2) was used to evaluate the quality of the model. Non-linear regressions were used to represent any regression of mass on length because mass scales to the cubic power of length (Froese, 2006). In addition, a statistical comparison of mass–length relationships between sexes was performed by applying an ANOVA test using the anova() function in R. A Shapiro-Wilk normality test was used to test for the normality of the sample using the shapiro.test() function. The Mann-Whitney U-test was used to test for differences in length and mass between males and females using the wilcox.test() function in R. A Mann-Whitney U-test was also used to examine the mean differences between the LT and MT according to the type of gear used. Moreover, a Kruskal-Wallis test was used to examine the differences between LT, MT and IG according to sexual maturity and between the biological indices (IG, IH and K) according to season using the kruskal.test() function in R. A post hoc Nemenyi test was used for pair-wise multiple comparisons of the ranked data after significant results were obtained from the Kruskal-Wallis tests using the posthoc.kruskal.nemenyi.test() function of the PMCMR package. The sex ratio was calculated for the whole sample and compared to the 1:1 proportion with a χ2-test (Zar, 2010) using the chisq.test() function in R. A χ2-test was also used to examine the relationships between sex and depth and between maturity stage and depth. The anova.2way.unbalanced() function (Anderson & Legendre, 1999; Legendre & Anderson, 1999) was used to perform a two-way unbalanced ANOVA on the total masses, according to several factors. Two factor modalities were used for region (north:south), season (cold, spring and winter; hot, summer and autumn), depth (0–400, 400–600 m) and sexual maturity (mature:immature) in the latter analysis. Finally, a principal component analysis (PCA) was carried out using the FactoMineR package and the PCAmix() function of the PCAmixdata package. A PCA provides insight into the underlying structure of a large amount of data by synthesizing the relationships between variables.
A PCA was implemented with the following numerical variables: LC, MT, LT, LF, ME, IH, ML, IG, MG, K. The four categorical variables (season, haul, sexual maturity and depth) were included in the analysis as additional variables with the following modalities: the 8 and 5 modalities of sexual maturity for females and males, respectively; 4 modalities of seasons (SPR, spring; SUM, summer; WIN, winter; AUT, autumn); 2 modalities of gears (L, long lines; T, trammel nets) and 3 modalities of depth (0–20], 200–400 and 400–600 m). The results are given for the first two principal components, representing 48·55% of the total variance.
Results
Morphometric measurements and depth distribution
A total of 38 C. uyato (18 females and 20 males), ranging from 45 to 94 cm LT (mean LT ± S.D. = 80·64 ± 9·97 cm, n = 38) and from 870 to 6700 g MT (mean MT ± S.D. = 3791 ± 1313 g, n = 38), were studied (Table I). Significant differences were observed in the LT and MT distributions between male and female individuals (LT: Mann-Whitney U = 1, NMale1 = 20, NFemale2 = 18, P < 0·05; MT: Mann-Whitney U = 0·5, NMale1 = 20, NFemale2 = 18, P < 0·05).
| Number of individuals | LT (cm) | MT (g) | |||||
|---|---|---|---|---|---|---|---|
| Mean ± S.D. | Min | Max | Mean ± S.D. | Min | Max | ||
| Females | 18 | 88·08 ± 3·11 | 81 | 94 | 4997 ± 631·17 | 3400 | 6200 |
| Males | 20 | 73·95 ± 9·24 | 45 | 83 | 2706 ± 616·84 | 870 | 3400 |
The sampled individuals were caught at depths ranging from 115 to 600 m. All male individuals were caught at depths >400 m, whereas females were caught throughout the depth range sampled. Moreover, 61% of female individuals were at a depth of 200–300 m and 47% of male individuals at 500–600 m. Females were also caught at depths of 500 to 600 m, but no individuals were caught at 300–400 m (Fig. 2). Significant differences were observed between the male and female distributions according to depth (χ2 = 16·87, P < 0·01, d.f. = 5), but significant differences were observed between mature and immature individuals according to depth (χ2 = 10·53, P > 0·05, d.f. = 5). In addition, significant differences in mass distributions were observed according to depth (Kruskal-Wallis H = 12·99, P < 0·01, d.f. = 2), with significant differences between the depth classes of 200–400 and 400–600 m (Nemenyi test, P < 0·05).
, male; 
, female; L, longline; T, trammel net.Spatial distribution and fishing gear selectivity
Twenty-two C. uyato were caught using longlines and trammel nets with 26 and 30 mm meshes in the PESCA-Libano survey. The CNS for the survey catches indicated the highest catches were made using trammel nets of both meshes in the north and south with CNS ranging from 4·9 to 5·45 kg per unit effort, at depths from 120 to 200 m during spring and summer. Relatively lower CNS were found for longline catches in the south in spring, at depths from 450 to 600 m (Fig. 3).
) of Centrophorus uyato caught in the PESCA-Libano surveys using trammel nets of (a) 26 and 30 mm meshes (n = 14 C. uyato) and (b) longlines (n = 8 C. uyato). 
, Hauls where no C. uyato were caught.In addition, larger individuals of C. uyato [Fig. 4(a)–(c)] were caught by trammel nets rather than longlines and significant differences were observed between LT and MT according to the type of gear used (LT, Mann-Whitney U = 27·5, NLongline1 = 26, NTrammelnet2 = 12, P < 0·05; MT, Mann-Whitney U = 30·5, NLongline1 = 26, NTrammelnet2 = 12, P < 0·05). A wider size range, however, was caught when using longlines [Fig. 4(a), (b)]. Moreover, a significant difference in terms of the number of individuals caught by sex was also observed between males and females according to the gear used (χ2 = 16·52, P < 0·05, d.f. = 1). All males were caught using longlines and the majority of females using trammel nets (Fig. 2).
, Median; 
, mean; box, 25th and 75th percentiles; whiskers, range; 
, outliers; n, sample size.Sex ratio, length-frequency distribution and mass–length relationship
The sex ratio lacked significant bias and was not significantly different from 1:1 (χ2 = 0·11, P > 0·05, d.f. = 1). The length–frequency distribution indicated the absence of females in size classes <80 cm and their domination of the larger size classes (80 and 90 cm). On the contrary, males occupied the smaller size classes (40–80 cm), there were few in the 80 cm size class and they were absent in the largest size class (90 cm). The 75% of males occupied the 70 cm size class and 72% of females were in the 80 cm size class (Fig. 5).
) and female (
), Centrophorus uyato by 10 cm length classes.The ANCOVA test indicated significant differences in the mass–length relationships of male and female individuals (ANCOVA, F1,1 = 52·89, P < 0·05); therefore the sexes were not combined. The estimated MT–LT parameters of both male and female specimens of the whole sample demonstrated negative allometric growth (b < 3): males, MT = 0.3585LT2·071, r2 = 0·94, n = 20; females, MT = 0.0239LT2·735, r2 = 0·64, n = 18 [Fig. 6(a), (b)].
, All fish sampled; 
, mature fish only.The mass–length relationships for both males and females were also estimated with the immature individuals of stage one and two removed from the sample (few individuals were found in these small size classes, which could bias the relationships). The estimated MT–LT parameters of both male and female specimens without the immature individuals also demonstrated negative allometric growth (b < 3): males: MT = 0.0785LT2·419, r2 = 0·53, n = 16; females: MT = 0·243 LT 2·219, r2 = 0·45, n = 17 [Fig. 6(a), (b)]. These new relationships were valid for sizes between 70 and 85 cm for males and 85 and 95 cm for females.
Sexual maturity and biological indices
All maturity stages were observed during the study period, except stage 2 (developing), stage 3a (capable of reproducing) and stage 4 (regenerating stages for females and regressing stage for males). Significant differences were observed between the mass and length distributions of C. uyato according to each maturity stage for male individuals and for the whole sample (Whole sample LT, Kruskal-Wallis H = 26·25, P < 0·001, d.f. = 6; MT, Kruskal-Wallis H = 26·29, P < 0·001, d.f. = 6; male sample LT, Kruskal-Wallis H = 8·16, P < 0·05, d.f. = 3; MT, Kruskal-Wallis H = 8·81, P < 0·05, d.f. = 3). Nevertheless, no significant differences were observed for the same distributions according to each maturity stage for female individuals (LT, Kruskal-Wallis H = 7·12, P > 0·05, d.f. = 4; MT, Kruskal-Wallis H = 7·12, P > 0·05, d.f. = 4). The post hoc Nemenyi test indicated significant differences in mean LT and MT according to sexual maturity between the male maturity stages 2 and 3b (P < 0·05) and whole-sample stages 2 and 3c (P < 0·05), 3d and 3c (P < 0·05).
The majority of male C. uyato in the study period were actively spawning (stage 3b), with sperm being released from the cloaca upon pressure. The highest mean values of IG, however, were observed in stage 3a when the testes were greatly enlarged and full of sperm [Fig. 7(a)]. In addition, the majority of observed females were in mid gravidity (stage 3c), showing the highest values of IG. Relatively lower IG values were observed in early (stage 3b) and late (stage 3d) gravidities [Fig. 7(b)]. No significant relationship was observed between the IG according to the sexual maturity stages in either male or female samples: male IG, Kruskal-Wallis H = 3·07, P > 0·05, d.f. = 3; female sample IG, Kruskal-Wallis H = 6·77, P > 0·05, d.f. = 4). A significant relationship was observed, however, between IG according to the sexual maturity stages in the whole sample, with significant differences between stages 2 and 3c and between 3c and 3d: whole sample IG, Kruskal-Wallis H = 22·79, P < 0·001, d.f. = 6; Nemenyi test, P < 0·05. Moreover, no significant relationships were observed between LC and LT (Pearson rank correlation, r2 = 0·14, n = 18, P > 0·01). The smallest mature male (LT = 73·5 cm) had LC = 75 mm [Fig. 8(a)]. The male gonad mass ranged from 2 to 100 g and showed no significant increase with LT (Pearson rank correlation, r2 = 0·35, n = 18, P > 0·01) [Fig. 8(b)].
, Median; 
, mean; box, 25th and 75th percentiles; whiskers, range; n, sample size.
Early, mid and late gravid females were caught in spring, summer and autumn 2013, but only early gravid females were caught in winter 2014 in the study area at depths of 100 to 200 m (spring and summer) and 500 to 600 m (autumn and winter). Seven embryos were found in the late gravid females of the sample. All these embryos were easily sexed and had total masses from 52·2 to 153 g and LT from 21 to 30 cm. The smallest mature female had a LT of 85·5 cm. Each of the gravid females bore one embryo except one which was observed in July 2013 bearing two embryos, one male (MT = 121 g; LT = 13 cm) and one female (MT = 123 g; LT = 27·5 cm).
The presence of gravid females occurred in parallel with the increase in IG observed in autumn and spring, with high values of the gonado-somatic index (mean IG ± S.D. = 6·10 ± 1·73, n = 11) during the latter season. The IG of the males showed high values (mean IG ± S.D. = 2·73 ± 0·86, n = 7) in spring compared with other seasons [Fig. 9(b), (e)]. In addition, IH values of the females were lower than those of males throughout the year. The lowest values were in spring, along with the elevated values of IG. The males demonstrated the highest values of IH in spring (mean IH ± S.D. = 57·95 ± 4·56, n = 7) and relatively lower values than the general mean during the other seasons [Fig. 9(a), (d)]. The general mean for the male condition factor throughout the four seasons was less than that of females. Male K values were almost constant, with slight fluctuations throughout the study period, whereas those of females showed wider fluctuations, with the highest values in summer (mean K ± S.D. = 0·77 ± 0·03, n = 2) [Fig. 9(c), (f)]. Finally, there were no significant differences between combined sexes sample IH and season (Kruskal-Wallis H = 0·57, P > 0·05, d.f. = 3) nor between combined sex or male sample K according to season (combined sexes Kruskal-Wallis H = 0·95, P > 0·05, d.f. = 3; male sample Kruskal-Wallis H = 1·16, P > 0·05, d.f. = 3), whereas significant differences were observed in IG (combined sexes sample and male only sample) and IH (male sample) according to seasons during the same period (combined sex IG Kruskal-Wallis H = 9·04, P < 0·05, d.f. = 3; male only IH Kruskal-Wallis H = 12·68, P < 0·01, d.f. = 3; IG Kruskal-Wallis H = 13·60, P < 0·01, d.f. = 3). Females demonstrated no significant difference for all biological indices according to season (IH, Kruskal-Wallis H = 6·72, P > 0·05, d.f. = 3; IG, Kruskal-Wallis H = 2·98, P > 0·05, d.f. = 3; K, Kruskal-Wallis H = 1·83, P > 0·05, d.f. = 3). The post hoc Nemenyi test gave significant differences for IG and IH according to season between the spring and autumn seasons for both indices (P < 0·05) of the male sample, but no post hoc significant differences were observed for IG according to season for the whole sample (P > 0·05).
, Median; 
, mean; box, 25th and 75th percentiles; whiskers, range; o, outliers; n, sample size.Two-way unbalanced anova and PCA
The two-way unbalanced ANOVA showed no significant double interactions between all 2 × 2 factors (two-way ANOVA, P > 0·05). The PCA indicated that the main numerical variables contributing to axis 1 were MT, LF, MG, ME and IG, whereas those contributing to axis 2 were ML and IH. The main categorical variables contributing to axis 1 were depth, gear, sex and sexual maturity, whereas those contributing to axis 2 were season and region. In addition, mature females caught by trammel nets at depths between 0 and 400 m had the largest sizes and gonad masses, in contrast to the immature males on the negative side of axis 1. Regarding axis 2, the largest livers were in individuals caught during spring in the southern region of Lebanon, in contrast to those caught during winter and summer in the northern region on the negative side of the axis [Fig. 10(a), (b)].
), season, gear, sexual maturity and depth: SPR, spring; SUM, summer; WIN, winter; AUT, Autumn; L, longline; T, trammel net; Mat, sexual maturity (for stages, see Fig. 7); [0–200] and [400–600], depth (m).Discussion
Although many factors influence the population dynamics and maintenance of fish populations, the present work only discusses those that could be important for the future management of C. uyato.
Spatial and seasonal distribution, habitat and life cycle
Muñoz-Chapuli (1984) hypothesized that demersal sharks are segregated by size and sex in the Atlantic Ocean. Based on bathymetric distributions of a number of species, including C. granulosus and Galeus melastomus Rafinesque 1810, he hypothesized that such sharks had nursery grounds on the continental shelf to which gravid females migrated, whereas adult males remained at greater depths. Golani & Pisanty (2000) confirmed this assumption for C. granulosus, by observing that the males of this species dominated intermediate depths (550–800 m) and females shallow depths (200–400 m) along the coast of Israel. The present observations agree with these results and suggest the probable bathymetric sexual segregation of C. uyato. The results also show that mature females constitute a substantial proportion of sample size, which indicates that the Lebanese coast is a probable breeding region. The results also suggest that C. uyato is present both north and south of Lebanon. The higher catches observed in the trammel net catches compared with those of the longline catches implies that trammel nets are more relevant to demersal shark by-catch than longlines, but Stergiou et al. (2002) also showed that bottom longlines incidentally catch demersal species, such as Mustelus spp., Squalus spp., Torpedo spp. and some Rajidae. Centrophorus uyato is demersal, therefore, it is likely that bottom longlines also contribute significantly to the by-catch of this species. The effect of gear selectivity was also shown in the present study, where trammel nets caught larger individuals, but with a narrower length range than longlines that caught smaller individuals with a greater length range. A similar pattern for gear selectivity was observed for several elasmobranch species in southern Portugal (Coelho et al., 2005). In addition, the majority of C. uyato females caught by trammel nets in this study differ from the situation in south-eastern Australian waters where C. zeehani and Centrophorus moluccensis Bleeker 1860 mature females are mainly caught by longlines (Graham & Daley, 2011). This can be attributed to the type of hooks, baits and meshes used, as well as the depths at which fishing occurred. Since both gears contribute to catching Centrophorus spp., mitigation measures for both could be used to reduce by-catch. For instance, mitigation measures for sharks and rays caught in nets include shorter soak times to enable by-catch to be released alive and avoidance of critical habitats. In addition, common longline mitigation measures include a variation in the size, shape and composition of hooks (to reduce deep hooking and ensure the corrosion of hooks if not removed from the shark upon release). For example, longline J-hooks baited with squid take a lower catch of sharks, but more turtles (Fowler, 2016).
Size and growth
The maximum length recorded for C. uyato in this study (94 cm) is shorter than that given by White et al. (2013). This variation might be due to sampling or the occurrence of a sub-population in the Levantine basin, where the waters are oligotrophic (Abboud-Abi Saab & Kassab, 1997; Ignatiades, 2005) and which may affect the physiological status and growth of this species. Negative allometric growth was found for both males and females, which is probably due to the domination of large mature individuals in the male and female samples and the absence of a wide length range in the female sample. Variation in allometric coefficients b could also be attributed to the environment and physiological status of C. uyato in the regions studied. In all cases, a sample with a wide length range for both sexes should be analysed to obtain more precise estimation of allometric parameters.
The smallest mature C. uyato male in this study was smaller than the smallest mature female. Although Megalofonou & Chatzispyrou (2006) stated that gonad masses increase rapidly with LT for centrophorids, no significant relationship between male MG and LT of C. uyato was observed in this study. This can be attributed to the domination of mature male individuals and the absence of small to medium-sized individuals from the sample.
Reproduction and biological indices
The presence of gravid females at various stages could indicate that the Levantine area along the coast of Lebanon is a reproductive area for this species. One of the C. uyato females in this study that carried two embryos was caught in July. The sizes of these embryos (LT = 27·5 cm; LT = 13 cm) coincided with those noted by McLaughlin & Morrissey (2005), which indicated late upcoming parturition.
As for the variation in biological indices, it was obvious that IG reached its highest values in early and mid gravidity for females, due to the formation of well-filled rounded uteri with high yolk content, and prior to spawning in males, due to greatly enlarged testes in males, all of which can increase gonad mass and positively affect IG.
The lower mean values of IH in the female sample throughout the study period could be due to the domination of gravid females, which explained the high IG values in spring and autumn. Moreover, it has been shown that IH decreases during maturation and gravidity for some elasmobranchs (Yano, 1995; Demirhan et al., 2010). Therefore, this decline in IH for females, especially in spring, along with the highest IG values, could imply that C. uyato females metabolize liver reserves during gestation. Conversely, the presence of high IH and IG values for male C. uyato in spring were similar to those of Capapé (1980) for Raja asterias Delaroche 1809, where an increase in gonad mass was accompanied by storage of substances in the liver. This could indicate a sexual dimorphism in liver mass and reserve metabolism between male and female C. uyato. In addition, due to the variation of IH in both male and female C. uyato and the fact that this species has a relatively long gestation period of 3 years (as reported by McLaughlin & Morrissey, 2005), it is apparent that this involves metabolic energy expenses for a longer period and the size of the liver changes according to the reproductive stage; an observation also described for a squalid by Yano (1995).
The condition factor can provide information on the general condition of fishes in the habitat in which they live, as well as indicate phenomena such as spawning period, alterations in population density and feeding conditions (Braga, 1986). The condition factor of male C. uyato, showed almost constant values throughout the study period. This is indicative of an wellness among the male population and could indicate a lack of any negative physiological or environmental factors affecting the stock. Nonetheless, the wide fluctuations in K of the females could be due to the physiological stress caused by gestation or mating as some male or female elasmobranchs exhibit low condition factors, indicating probable stress during the mating season (Loefer & Sedberry, 2003).
This deep-water shark in the Levantine basin is taken as by-catch by longlines and trammel nets, at times discarded, otherwise sold at low cost. Precautionary measures to prevent the overexploitation of such deep-water elasmobranchs in Lebanese waters should follow the actions of the European Union, which banned catches of deep-sea sharks in E.U. waters and the General Fisheries Commission for the Mediterranean (GFCM) that banned any fishing below 1000 m depth in the Mediterranean Sea to protect poorly understood deep-sea ecosystems. In addition, noting that catch prevention does not provide a sufficient protection, by-catch and discarding should be controlled through mitigation measures and shark awareness campaigns, with the involvement of all concerned stakeholders.
We are particularly grateful to the Lebanese fishermen whose catches represent a significant portion of the collected data. M. Lteif was supported by the Lebanese government through a doctoral grant. Some data and financial support for this study were also partially provided by the scientific survey under the PESCA-Libano programme performed in the framework of the CANA project of the Lebanese National Council for Scientific Research (CNRS-L).
