Chondrichthyes species in deep waters of the Mediterranean Sea *

The fish fauna in the Mediterranean Sea consists of 664 species, of which 86 are represented by chondrichthyes (Quignard and Tomasini, 2000). Most of the cartilaginous species are distributed on the continental shelf and on the slope. Only 4 of them dwell exclusively below 1000 m and only 1 exclusively at depths greater than 2000 m (Fredj and Maurin, 1987). In the past two decades studies on Mediterranean chondrichthyes have increased markedly, although


INTRODUCTION
The fish fauna in the Mediterranean Sea consists of 664 species, of which 86 are represented by chondrichthyes (Quignard and Tomasini, 2000).Most of the cartilaginous species are distributed on the continental shelf and on the slope.Only 4 of them dwell exclusively below 1000 m and only 1 exclusively at depths greater than 2000 m (Fredj and Maurin, 1987).
In the past two decades studies on Mediterranean chondrichthyes have increased markedly, although  SARDÀ, G. D'ONGHIA, C.-Y. POLITOU and A. TSELEPIDES (eds.)Chondrichthyes species in deep waters of the Mediterranean Sea* they mostly regard species distributed from the continental shelf to the upper slope (Fredj and Maurin, 1987).National and international programmes of bottom trawl surveys carried out routinely since the early 1980s have mostly contributed to the knowledge on the distribution and abundance of demersal sharks, rays and skates (Allué et al., 1985;Bertrand et al., 2000, Relini et al., 2000).Although cartilaginous fishes represent by-catch of the commercial fisheries in the Mediterranean, their importance as indicators of the "top down" effect in the food webs and overexploitation by fishery is currently under discussion for this basin (Serena et al., 2002).Their position at the top of the marine food webs in addition to their life-history k-strategies make chondrichthyes more susceptible to overfishing and to man-induced environmental changes (e.g.Stevens et al., 2000;Vacchi and Notarbartolo di Sciara, 2000).In this respect, a specific international plan of action has recently been proposed for the conservation and management of Mediterranean cartilaginous fish (Serena et al., 2002).The importance of such an action plan is linked to the awareness of the fragmentary knowledge on their biology and ecology, even for the more common species.
At present a comprehensive picture of the geographic and vertical distribution of deep-sea cartilaginous fish in the Mediterranean is lacking.Considering the broad area covered by the Mediterranean below 1500 m and the several technical limitations for sampling bathyal and abyssal fish species (Merrett and Haedrich, 1997), specific research programmes are required in order to gather scientific information on the species living in the various basins.
The DESEAS project, financed by the EC, gave the opportunity to investigate three different areas (the Balearic Sea, the western Ionian and the eastern Ionian) of the Mediterranean Sea down to 4000 m.This paper presents the results of deep-sea chondrichthyes species collected during this research.In particular, the abundance of each species is reported as well as the depth-size trends, size structure and biological aspects on the three most abundant species: Galeus melastomus, Centroscymnus coelolepis and Etmopterus spinax.
The sampling was conducted with the otter trawl Maireta System (OTMS) using the R/V García del Cid (38 m long; 1500 HP).The OTMS (vertical opening 1.8-2 m; wing spread 14 m) was trawled by a single warp and operated with a pair of rectangular iron otter boards (1.20 x 2.0 m; 450 kg).A detailed description of the net can be found in Sardà et al. (1998).The initial and final tow time were measured using a SCANMAR sonar system down to 1500 m, and the same parameters were assumed below 1500 m.The vessel speed and geographic position were measured using differential GPS.The trawling was carried out during daylight hours.The number of hauls carried out in each area, with the indication of the depth, is reported in Table 1.Tow duration ranged from 1 to 3 hours, but data were standardised to one hour of trawling.The results obtained in number were also standardised to one km 2 for subsequent numerical processing.
Individuals were identified on board following the nomenclature reported in Whitehead et al. (1984).For each species the relationships between abundance (N km -2 ) and depth were computed.Total length (TL) of each individual was measured to the nearest cm and weight to the nearest g.Sex and the maturity stage of the gonads were determined macroscopically according to Stehmann (1998).
The relationship between size and depth was evaluated for G. melastomus, C. coelolepis and E. spinax by means of regression analysis.The population structure by depth was performed for these species.The length-frequency distribution by maturity stage was computed for G. melastomus.

Species composition, abundance and biomass
During the DESEAS survey a total of 8 chondrichthyes species belonging to 5 families were collected between 600 and 2800 m depth.The depth range and the number of hauls in which the species were found together with the total number of the individuals caught and their total biomass are reported in Table 2. G. melastomus, Dalatias licha and E. spinax were caught in all three areas.In addition, DEEP WATER CHONDRICHTHYES SPECIES IN THE MEDITERRANEAN SEA 155 In all three study areas the Scyliorhinidae and Squalidae families represented about 74 and 24% respectively of the chondrichthyes collected.
The most abundant species were G. melastomus and E. spinax.The former was caught down to depths of 1500 m, with the highest biomass in Balearic waters.E. spinax showed the widest vertical distribution in the eastern Ionian Sea (down to 2200 m) and the greatest biomass in Spanish waters.C. coelolepis showed the deepest distribution (from 1500 to 2800 m) and was only caught in the Balearic Sea.
The relationships of abundance (N km -2 ) with depth of each species are shown in Figure 2. In the Balearic Sea, the abundance of G. melastomus, D. licha and E. spinax decreased with depth while C. coelolepis showed the highest density between 2500 and 2999 m.In the western Ionian Sea both G. melastomus and D. licha were found to be more abundant between 1000 and 1499 m, while E. spinax was mostly captured at depths between 500 and 999 m.In the eastern Ionian Sea G. melastomus exhibited the highest abundance in the 1000-1499 m depth stratum, whereas D. licha and E. spinax were mainly found in the shallower depths.The latter species also showed noteworthy density values between 1500 and 1999 m in both the western and eastern Ionian Sea.

Galeus melastomus Rafinesque, 1810
Both small and large individuals were found across the vertical gradient surveyed.Significant variations in the depth-size trend were only observed in all three areas for females (Fig. 3).The size composition of males and size by maturity stage of gonads in females are presented in Figure 4.The broadest size range was shown in the Balearic waters, with the females up to 62 cm and males up to 55 cm.In this area the most abundant specimens were between 50 and 60 cm TL.A flat distribution with a greater number of small individuals was found in the western Ionian Sea.The largest female and male measured 55 cm and 52 cm respectively.In the eastern Ionian Sea the maximum sizes were 51 cm in females and 48 cm in males.Here the specimens were mostly distributed around the size classes of 40-48 cm.
Immature and maturing females (stages I and II), mature females (stages III and IV) and females with egg capsules (stages V and VI) were caught in all three areas (Fig. 4).Females bearing egg capsules were caught from 800 to 1500 m and were between 44 and 55 cm in the Balearic Sea; they were collected from 800 to 1500 m and between 44 and 55 cm in the western Ionian Sea; 2 females with egg capsules of 45 and 50 cm TL were sampled at a depth of 600 m in the eastern Ionian Sea.

Centroscymnus coelolepis Bocage Capello, 1864
For Centroscymnus coelolepis, the relationship between size and depth is shown in Figure 5.The  two smallest individuals were caught in shallower waters.The size structure of this shark showed a wide range, with females and males that measured 27 to 64 cm and 49 to 54 cm respectively (Fig. 6).
The percentage of immature and maturing females was about 68%.The remaining percentage consisted of 4 females at the "candle" stage (with unsegmented yolk content) and one pregnant female.The former were caught from 2500 to 2800 m depth and were between 59 and 64 cm TL.The gravid female was found at a depth of 1500 m.It was 60 cm TL and contained 8 embryos that measured between 15 and 18 cm and weighed between 47 and 55 g.

Etmopterus spinax (Linnaeus, 1758)
The depth-size trend of E. spinax was analysed for the whole sampled population since a low number of individuals was caught in each area.A significant increase in size with depth was only shown in males (p < 0.05) (Fig. 7).The largest specimens were caught off the Balearic Islands (34-48 cm) and the youngest individuals (10-32 cm TL) in the western Ionian.In the eastern Ionian the size-range was 11-40 cm TL (Fig. 8).The females were found to be larger than the males.In the Balearic Sea eight mature females were collected between 800 and

DISCUSSION
The DESEAS survey represents the first recent investigation on deep-sea chondrichtyes species carried out in three areas of the Mediterranean Sea with the same gear and during the same period.All the species caught in the three study areas are typically demersal species living over sandy and muddy bottoms (Fischer et al., 1987).Although a different number of species was collected in the three study transects, it is not possible to draw definite conclusions on the species diversity differences between the three Mediterranean areas from the present results.All the species collected are known in both the western and eastern Mediterranean (Compagno, 1984;Whitehead et al., 1984;Golani, 1986-87;Fischer et al., 1987;Stefanescu et al., 1992a;Relini et al., 2000;Goren and Galil, 2002;Jones et al., 2003).The differences observed in the presenceabsence between the study areas might be due to the different number of hauls carried out by depth and area and/or to the occasional findings depending on the gear used.In addition, the number of species in each area could have been increased by employing other sampling techniques with different selectivity for the cartilaginous fishes.However, the number of species collected during this study is even higher than that captured in the western Mediterranean using three types of gear with a higher number of stations (OTSB, Agassiz trawl, longline) (Stefanescu et al., 1992a).
Three species (G.melastomus, E. spinax and D. licha) were found in all three areas.Their distribution throughout the Mediterranean basin is well known even in deep waters (Whitehead et al., 1984;Tortonese, 1956;Stefanescu et al., 1992a;Golani, 1986-87;Jones et al., 2003).Although during this research C. monstrosa was not sampled in the western Ionian Sea, its occurrence has frequently been recorded along the southeastern Italian coasts (Matarrese et al., 1996;Relini et al., 2000).C. coelolepis was only caught in the Balearic Sea, confirming its wide vertical distribution in the western Mediterranean.In this area it was caught as far as the maximum depth of 2863 m (Grey, 1956) and recently down to 2853 m (Clò et al., 2002).In the Catalan Sea, it was previously found from about 1400 to 2300 m (Carrassón et al., 1992;Stefanescu et al., 1992b) and off the Balearic Islands down to 1800 m (Moranta et al., 1998).The occurrence of C. coelolepis in the eastern Mediterranean was recorded using lander platforms equipped with baited cameras: between 1500 and 2500 m in the Cretan Sea and between 2300 and 3850 m in the Rhodes Basin (Priede and Bagley, 2000).C. granulosus and H. griseus were only collected in Greek waters, though their occurrence throughout the Mediterranean has long been known (e.g.Fischer et al., 1987).The former was captured by longline from a depth of 1400 m off the coast of Israel (Golani, 1986-87).The presence of both species in the eastern Mediterranean has also been recorded in underwater photographic surveys (Gilat and Gelman, 1984;Priede and Bagley, 2000;Jones et al., 2003).Gilat and Gelman (1984) recorded C. granulosus from 280 to 1490 m, with a density peak at 1080 m, and H. griseus from a depth of 750 m.Jones (Priede and Bagley, 2000;Jones et al., 2003) reported the occurrence of both sharks between 1500 and 2500 m in the Cretan Sea and between 2300 and 3850 m in the Rhodes Basin.
The overall highest abundance of the chondrichthyes considered during the DESEAS cruise was observed at less than 1500 m.The abrupt decrease in abundance below this depth might reflect the reduction of food availability as reported for deep-sea fish assemblages studied in the western Mediterranean Sea (Stefanescu et al., 1993;Moranta et al., 1998).In fact, the trophic resources between the upper and middle slope benthic boundary would be mostly based on the energetic transfer system due to mesopelagic organisms, as reported in the Atlantic ocean (e.g.Gordon, 1979;Merrett and Haedrich, 1997).In this respect, G. melastomus has a diet based on epibenthic (mainly Calocaris macandreae) and benthopelagic species (euphausiids, ommastrephidae, Pasiphaea multidentata, mesopelagic fishes), while C. coelolepis almost exclusively consumes cephalopods and, unlike the Atlantic populations (Mauchline and Gordon, 1983), the Mediterranean ones do not show strong scavenger habits (Carrassón et al., 1992).The diet of E. spinax is mainly based on cephalopods (Wurtz and Vacchi, 1978).
The occurrence of large mobile sharks at depths greater than 1500 m, revealed in the both western (e.g.Stefanescu et al., 1992a; present study) and eastern Mediterranean basin (e.g.Jones in Priede and Bagley, 2000;Jones et al., 2003;and in the present study), confirms the presence on the lower slope of a further trophic system based on food arriving from the upper layers of the water column, mostly represented by animals with migratory habits (Isaacs and Schwartzlose, 1975;Mahaut et al., 1990).
The abundance recorded for G. melastomus, and to a lesser extent for E. spinax, confirms that these two species are the most abundant deep-sea demersal sharks in the Mediterranean Sea (e.g.Bertrand et al., 2000;Relini et al., 2000;Baino et al., 2001).Both species are widespread in this basin, although the former is considered absent from the northern and central Adriatic.Moreover, both species were collected in deep waters of the western (Carrassón et al., 1992;Stefanescu et al., 1992b;Moranta et al., 1998) and eastern Mediterranean basin (Galil and Goren, 1994;Goren and Galil, 1997;Jones et al., 2003).The present finding of E. spinax down to 2200 m represents the greatest depth at which the species has been collected by trawling in the Mediterranean Sea.However, Jones et al. (2003) captured this shark in the Cretan Sea using baited traps as deep as 2230 m.Furthermore, Jones (Priede and Bagley, 2000) recorded the presence of E. spinax, together with G. melastomus, by means of baited cameras between 2300 and 3850 m in the Rhodes Basin.
Catches of G. melastomus in the Balearic and western Ionian Sea consisted of both small and large individuals, as reported by Ungaro et al. (2001) for the southern Adriatic Sea.In the eastern Ionian Sea the samples only consisted of medium-large specimens.Stefanescu et al. (1994) only collected large females at depths of 1200 m in the Catalan Sea.With regard to the size-related depth distribution, no trend was shown in males while a decreasing size with depth was detected in females.According to Carrassón et al. (1992), the occurrence of small specimens (< 20 cm) at depths greater than 1000 m was shown both in the Balearic Sea and the western Ionian Sea.However, considering that the bulk of the population on the edge of the continental shelf and on the upper slope of the western Ionian Sea consists of young individuals (Tursi et al., 1993), the spawning in this shark seems to occur over a wide depth range.
For E. spinax, Macpherson and Duarte (1991) indicated a general trend of increasing size with depth.Jones et al. (2003) did not find any clear depth-size trend in specimens measured photographically.In the present study a smaller-shallower pattern seems to be evident, since large individuals were captured over the whole depth range.Large specimens of E. spinax are also distributed in shallower waters than those investigated in this study (Sion et al., 2000).In addition, since mature females were collected in the uppermost depths, the smallershallower trend might be the result of an inshore migration of adults during the spawning period.
An increase in size with depth was shown in C. coelolepis.However, no definitive conclusion can be drawn due to the small number of specimens collected.Only one pregnant female of this shark was found at 1500 m depth.In the northeast Atlantic, pregnant females were found almost exclusively in waters shallower than 1200 m (Girard and Du Buit, 1999;Clarke et al., 2001).This would explain the fact that pregnant females can move to shallower bottoms to give birth.Embryo sizes are in the range reported in the literature for the Mediterranean Sea (Clò et al., 2002) and much smaller than those observed by Girard and Du Buit (1999) for the Atlantic Ocean (around 30 cm).On the other hand, the maximum lengths usually recorded in the Mediterranean Sea are much smaller than those reported by these authors in the Atlantic Ocean (122 cm).
The distribution of the individuals in deep areas where trawling does not occur in the Mediterranean Sea makes the deep sharks less vulnerable to overexploitation.In fact, the reduction in numbers of several selachian species, mostly on the continental shelf, seems to be related to the development of trawl fishing (Bertrand et al., 2000;Relini et al., 1999Relini et al., , 2000)).According to Stevens et al. (2000), sharks are particularly vulnerable to overexploitation because of their slower growth, delayed maturity, low fecundity and in general k-selected life-history strategies.Indeed, several species of shark and skate that once were widespread and abundant are now uncommon and rare in Italian waters (Vacchi and Notarbartolo di Sciara, 2000).Some fishing activities targeting cartilaginous fishes, such as the deep-water fishery for bluntnose sixgill shark (Hexanchus griseus) in the Italian Ionian Sea, have almost disappeared.The fact that G. melastomus and E. spinax are the most abundant shark species in the Mediterranean and show a stable size structure from trawl fishing catches might be related to their distribution at depths that are not exploited by trawling, thus making them less vulnerable to fishing.
Considering the deep distribution of C. coelolepis, the Mediterranean population of this shark could still be pristine, though its abundance in the whole basin is completely unknown.This could also be true for other deep-water sharks which are occasionally caught as by-catch during bottom trawls, or using longlines and driftnets, indicating that further data collection in deep-water using different sampling tools is required in order to improve our knowledge on the Mediterranean deep-sea ichthyofauna.
FIG. 2.  -Relationship between abundance (N km -2 ) and depth of chondrichtyes species collected in the three study areas during the DESEAS survey.(The number of hauls where the species were collected is indicated).

TABLE 1 .
-Number of hauls by depth strata carried out in the Balearic Sea, in the western and eastern Ionian Sea during the DESEAS survey.

TABLE 2 .
-Chondrichtyes species collected in the Balearic Sea, in the western and eastern Ionian Sea during the DESEAS survey, with indication of depth range, number of hauls in which the species were caught (n.), total number (N h -1 ) and total biomass (kg h -1 ).
one species was collected exclusively in the Balearic Sea (C.coelolepis), one in the western Ionian (Oxynotus centrina) and two in the eastern Ionian (Hexanchus griseus and Centrophorus granulosus).