Feeding habits of Pagellus acarne ( Sparidae ) in the Gulf of Tunis , central Mediterranean

the feeding habits of the axillary seabream, Pagellus acarne (Risso, 1810), from the Gulf of tunis were investigated in relation to season, sex and fish size (juveniles tl<14.5 cm and adults tl≥14.5 cm). a total of 536 specimens (males, females, unsexed and hermaphroditic), ranging between 11.0 and 25.2 cm tl, were collected with trammel nets from June 2005 to July 2006. their stomach contents were analysed. of the total number of examined stomachs, 279 stomachs were empty (Vacuity index, VI = 52%). the VI did not reveal significant seasonal differences when all of the specimens were analysed together; however, significant seasonal variations were observed for females. the diet of the axillary seabream was composed of 36 different prey species. the most important prey were arthropoda, mollusca and echinodermata. Both the dietary indices and the spearman’s coefficient of correlation indicated seasonal variations in the diet. the diet was more diversified in adults than in juveniles. the axillary seabream fed on a wide range of prey items, endofauna and nekton, and can be considered a carnivorous and euryphagous predator.

P. bellottii has never been recorded in tunisia; P. bogaraveo is confined to the northern coasts of tunisia whereas P. erythrinus and P. acarne are found everywhere (Bradaï, 2000).the main target of local fisheries is the common pandora (2441 tons/year, from coastal and trawl fleets).Conversely, landings of the axillary seabream are rather low along the northern coasts and only mentioned occasionally (lubet and azouz, 1969;Bouhlel, 1978;Bradaï, 2000).only one study is available on the reproduction of this species in the Gulf of tunis (mokrani et al., 2007).however, the economic interest of the axillary seabream has stimulated some biological studies in other mediterranean regions, such as in the ionian sea (andaloro, 1983a), Greek waters (mytilineou, 2000), the algerian sea (Zerouali-khodja and amalou, 2005), around Canary isles (Pajuelo and lorenzo, 2000) and south Portugal (Coelho et al., 2005).since 1990, following the adoption of the strategic necessity to diversify the artificial production of finfish species in the mediterranean (abellán and Basurco, 1999), the culture of P. acarne started in italy (Greco et al., 1995;arculeo et al., 2000) and in spain (dominguez, 2000).the feeding habits of P. acarne have been studied in atlantic (domanevskaya and Patokina, 1984;morato et al., 2001) and mediterranean waters (andaloro, 1983b;Rizkalla et al., 1999).
to improve the stock management of the axillary seabream and to ensure its sustainability, the key parameters relating to its dynamics and exploitation have been defined.natural and fishing mortalities have been estimated in atlantic waters (Pajuelo and lorenzo, 2000).selectivity of the fishing gears has been studied for this species in Greek (Petrakis and stergiou, 1996;tokaç et al., 1998) and Portuguese waters (santos et al. 1995;Campos and fonseca, 2003).
the purpose of this work was to study the feeding habits of the axillary seabream in the Gulf of tunis through stomach content analysis.

Fish sampling
samples of axillary seabream were collected twice per month from the artisanal fleet operating in the Gulf of tunis (fig. 1) with trammel nets of 48 mm mesh size (stretched).the nets were set at night and hauled in the morning.
from June 2005 to July 2006, all the axillary seabream specimens caught were collected and considered for this study.each specimen was measured to the nearest mm (tl, total length), dissected and sexed.the individuals were classified into two size groups, juveniles (tl<14.5 cm) and adults (tl≥14.5 cm), on the basis of the smallest size at maturity (mokrani et al., 2007).the stomach was removed from each specimen and preserved in a 70° alcohol solution.

Stomach content analysis
the stomach contents were washed in a Petri dish and examined under a microscope.the food items were sorted into large taxonomic groups and, when possible, identified to the species level according to Riedel (1983) andfischer et al. (1987).When the state of digestion was not advanced, the number of prey was counted.otherwise, the number of prey was determined from the hard structures.the vacuity index (VI, %), the frequency of occurrence (FO, %) and the numerical percentage of a prey item i (N, %) were calculated according to hureau (1970) and Berg (1979) as follows: where Nv = number of empty stomachs, Ne = total number of examined stomachs, Ndi = number of stomachs containing the prey i, Nnv = total number of stomachs containing some food, Nti = total number of prey i, Ni = total number of ingested prey.the point method proposed by hynes (1950), Berg (1979), Pasquaud et al. (2004), andBouchereau et al. (2006) was used to estimate the ingested biomass of each item.each prey is assigned a number of points (P, %) proportional to its estimated contribution to the whole stomach volume (table1), as follows: where Npi = total number of points of a prey item, Ntp = total number of points.the food preference of axillary seabream was determined by applying the main food index (MFI) (Zander, 1982): Prey were classified based on MFI values as follows: MFI > 75 = preferential prey, 50 <MFI< 75 = main prey, 25 <MFI< 50 = secondary prey and MFI < 25 = accessory prey.
statistical differences of the different indices (VI; FO; N) in relation to season or sex were assessed by the chi-square goodness of fit test (χ²) at p = 0.05.
Variations in the diet of the axillary seabream were estimated using the spearman coefficient of correlation (ρ) (falissard, 2005): where n = rank number and d = difference between ranks.the items were classified according to a decreasing order of the coefficient value.for the same sample, the rank attributed to the equally placed items corresponds to the mean ranks that the items would have had if they had not been identical values of the dietary index.statistical significance of ρ was tested using the t-student test at p = 0.01: Results a total of 536 specimens of P. acarne ranging from 11 cm to 25.2 cm were collected over the entire year (table 2).

Feeding intensity
out of 536 stomachs examined, 273 were empty (vacuity index, VI = 52 %). the vacuity index did not reveal any significant differences over the year (χ² = 3.09; p>0.05), and accounted for 58 % in summer, 48 % in autumn, 49 % in winter and 53 % in spring.however, the vacuity index computed by sex showed significant seasonal variations for females (VI = 51%; χ² = 10.4,p<0.05), but did not present significant seasonal variations for males (VI = 52 %; χ² = 1.10, p>0.05).VI values only differed significantly between sexes in autumn (table 3). the mean annual VI recorded in juveniles (lt<14.5 cm) was 87.8% with a maximum in autumn and spring (100%) and was higher than that of the adults (lt≥14.5cm)

Diet composition
the stomach contents of the axillary seabream were composed of 36 different taxa belonging to nine

Seasonal variations in the diet
dietary indices calculated for each prey indicated that the axillary seabream fed mainly on arthropoda and mollusca over the year (table 5).arthropoda, more frequent and abundant in winter and autumn, showed significant seasonal variations in frequency of occurrence (χ² = 8.64, p<0.05) but not in abundance (χ² = 6.29, p>0.05).the point method showed that arthropoda were more than 50 % of the ingested biomass in spring (P = 66.7%), while in summer and autumn they represented only 40.1% and 35% respectively, which shows significant seasonal variations (χ² = 17.21, p<0.05).mollusca attained the highest frequency of occurrence in summer (FO = 83.3%)and the lowest in winter (FO = 21%), thus revealing significant seasonal variations (χ² = 46, p<0.05).arthropoda constituted the main prey of the axillary seabream in winter (MFI = 70) and secondary prey during the rest of the year, while mollusca were accessory prey all year round (MFI<25).echinodermata were secondary prey (13<MFI<51), and showed seasonal variations in frequency of occurrence (χ² = 50.77,p<0.05), number (χ² = 13.19,p<0.05) and biomass (χ² = 69.58,p<0.05).annelida were absent in the diet in spring, poorly represented in autumn and increased in winter and summer.the other prey contributed very little to the ingested biomass (table 5). the ingestion of fish, Bryozoa and nemathelmintha was seasonal and scarce.the spearman coefficient of correlation confirmed the changes in the diet of the axillary seabream, which indicates seasonal differences in diet.homogeneity of diet was noted between spring and summer and between summer and autumn (table 6).

Food in relation to sex
the diet composition of P. acarne varied seasonally in relation to sex, attaining 19 to 22 taxa of prey for females and 14 to 22 taxa for males (tables 7 and 8).arthropoda showed significant seasonal variations in the ingested biomass in males (χ² = 91.13,p>0.05) but not in females (χ² = 4.38, p<0.05). in summer, mollusca predominated both in number and in frequency of occurrence in both males and females (tables 7 and 8).
the mollusca contribution to the ingested biomass, varied between 0.6% and 4.8% for males and between 1.2% and 7.2% for females and did not show any significant seasonal difference for either sex (χ² (P?) = 6.96, p>0.05; χ² (P/) = 3.28, p>0.05).according to the main food index value, arthropoda constituted the main prey for both sexes in winter, but became the secondary prey in spring and summer for males and accessory for females (tables 7 and 8).however, mollusca were accessory prey for both sexes all year round.echinodermata were always present in the diet of the axillary seabream, and showed the highest indices in autumn for both sexes.these indices varied significantly during the year (χ² (fo?) =15.21, p<0.05; χ² (n?) = 9.94, p<0.05; χ² (P?) = 29.76,p<0.05; χ² (fo/) = 31.61,p<0.05; χ² (n/) = 12.59, p<0.05; χ² (P/) = 41.33,p<0.05).echinodermata were the main prey in males and secondary prey in females in autumn but accessory prey in winter for both sexes.the seasonal variations of the ingested prey and differences between male and female diet were confirmed by the spearman test (table 9).no significant seasonal variations were observed in females (0.383<ρ<0.754).however, in males, variations were observed between autumn and winter and between autumn and spring.

Seasonal variations in relation to fish size
the stomach content analysis of juveniles showed that their diet was poorly diversified in winter, as it consisted of 6 prey taxa, which increased to 10 prey taxa in summer; taxa belonged to arthropoda, mollusca and annelida (table 10).however, the adult diet was more diversified with 10 prey groups in winter and 8 groups in summer.the most representative prey groups were arthropoda, mollusca, annelida and echinodermata (table 11).ingestion of arthropoda showed   3.78, p<0.05).their contribution to the ingested biomass, constituted mainly by amphipoda and mysida (Mysis sp.) in juveniles and decapoda in adults, increased according to size in winter and decreased in summer (tables 10 and 11); there were significant variations between these two seasons only for juveniles (χ² (P, juveniles) =23.45, p<0.05; χ² (P, adults) = 0.165, p>0.05).according to the MFI, arthropoda constituted secondary prey for juveniles both in summer and in winter and main prey for adults in summer and secondary prey in winter.the contribution of mollusca to the ingested biomass in juveniles varied significantly between summer and winter both in frequency of occurrence and in number (χ² (fo) =54.65, p<0.05; χ² (n) = 90.03,p<0.05).this prey group, including mainly bivalve taxa, constituted a secondary prey in summer and an accessory prey in winter.its presence in the diet of the axillary seabream decreased according to size and it became the accessory prey in adults, both in summer and winter.in juveniles, the ingested biomass of annelida showed significant variations between summer and winter (χ² (P) =27.80, p<0.05).this biomass decreased according to size chiefly in winter and therefore annelida were secondary prey in juveniles and accessory prey in adults.echinodermata, absent in the juvenile diet both in winter and summer, seemed to be appreciated indiscriminately by the adults during these two seasons (χ² (f) =0.619, p>0.05; (χ² (n) =1.368, p>0.05; χ² (P) =1.74, p>0.05).they were secondary prey in summer and accessory prey in winter (table 11), whereas all the other taxa were accessory prey.the spearman test indicated differences between the summer and winter diets for both juveniles (ρ = 0.945; t obs =8.17 p<0.01) and adults (ρ = 0.9; t obs =5.85 p<0.01).disCussion the vacuity index of all the specimens together did not show any significant seasonal variation; however, the VI of females was lower than that of males in autumn.this period corresponds to the gonad maturation of the axillary seabream (mokrani et al., 2007).females had a higher feeding intensity than males, as they need to consume more energy for the maturation of their gonads.the high vacuity index values revealed  low feeding intensity in both sexes during spawning and post-spawning periods (i.e.winter-spring).this may be related to the lower availability of prey mainly in winter, which become less active and are closer to the bottom due to the low water temperature in the Gulf of tunis (8-13°C) (daly Yahia, 1998), and are therefore less exposed to predation.in general, the low feeding intensity observed in this study may be due to the fishing gear and/or to the feeding behaviour of fish at the moment of capture, as observed for the Pleuronectiforms caught by static gears (Verheijen and de Groot, 1979). in the Gulf of tunis, the axillary seabream was captured at night, which is when the fish are most active generally because they are feeding.the individuals caught were hauled onboard the following morning.hence, some of them may have remained several hours in the net, and their capture may have occurred before the ingestion of prey or after digestion.as a result, many specimens would have had an empty stomach at the moment they were collected.according to the data obtained, arthropoda, mollusca and echinodermata were the most common prey in the diet of the axillary seabream living in the Gulf of tunis.annelida were secondary prey.other prey, such as, foraminifera, Bacillariophyta, Bryozoa, nemathelmintha and teleostei were of minor importance.this carnivorous type of diet conforms to those of sparidae species belonging to the Diplodus genus reported by several authors (Rosecchi, 1987;Gonçalves and erzini, 1998;Pallaoro et al., 2006;derbal et al., 2007) and of other species of the Pagellus genus in particular (larrañeta, 1964;Rijavec and Županovic´, 1965;ardizzone and messina, 1983;Rosecchi, 1983;Ghorbel, 1996).
the axillary seabream, is considered to be an euryphagous species, as it searches for food in the endofauna and crawling fauna (mollusca, annelida, echinodermata).our results agree with those of Collingnon and aloncle (1960), who reported the presence of a large number of prey from the endofauna (Bivalvia and annelida), fish and crustaceans in the stomachs of the axillary seabream from the atlantic coasts of morocco.the presence of echinodermata in the diet of this species was also reported in the eastern mediterranean along the egyptian coasts (Rizkalla, 1999) and in the atlantic, along the northwestern coasts of africa (Phân and kompowski, 1972) and around azores (morato et al., 2001).the low contribution of fish to the diet of P. acarne from the Gulf of tunis contrasts with the information reported for the same species in the atlantic, where the frequency of occurrence reached 76.3% in spring (morato et al., 2001).the scarcity of fish in the diet of P. acarne observed in this study may be due to the overexploitation of local fish, which might have led to changes in the existing trophic chain at different levels.the axillary seabream probably uses its thick lips to move the sediment around in order to search for prey moving close to or inside the substratum .unlike sparids, P. acarne does not have sharp anterior teeth.Prey is probably sucked up and then ground with the molars or swallowed whole.indeed, linde et al. (2000) showed a strong correlation between the diet and the shape of the premaxillary in sparids, but not between the diet and teeth.
the feeding behaviour of P. acarne was different between sexes. it showed seasonal variations, mainly in females, which could be related to the reproductive cycle. in fact, prey ingestion is adapted to the energetic needs of the fish throughout the year. in contrast, seasonal variations in males were observed only during maturation of their gonads and the spawning period (autumn-winter).Considering the liver as the storage organ of the energetic reserves, our findings are in agreement with the monthly changes in the hepatosomatic index of P. acarne, which are larger in females than in males (mokrani et al., 2006).
the diet of juveniles was less diversified than that of adults.in summer and in winter, Crustacea and echinodermata (i.e.prey with hard carapaces and tests) were absent from the diet of juveniles.their diet is dominated by small organisms, such as amphipoda (gammarus gammarus), mysida (Mysis sp) and bivalve juveniles, which indicates an increase in prey size in relation to fish size.this difference in feeding behaviour between juveniles and adults is also known in other sparids, such as Diplodus sargus (Rosecchi, 1987), D. vulgaris (Pallaoro et al., 2006) andD. annularis (derbal et al., 2007). in adults, the larger opening of the mouth and the bony structure of the maxillaries, premaxillaries and mandibles, which makes them stronger, could allow the capture and ingestion of bigger and harder prey.
in conclusion, in the Gulf of tunis P. acarne is a carnivorous and euryphagous predator.this species feeds on benthic organisms such as arthropoda, mollusca and echinodermata.however, sex-related seasonal variations and diet differences between juveniles and adults were observed.

Table 5 .
-Prey identified in the stomach contents of P. acarne in relation to season.FO, frequency of occurrence; N, numerical percentage; P, point percentage; MFI, main food index; R, classification of prey.

Table 5 (
cont.).-Prey identified in the stomach contents of P. acarne in relation to season.FO, frequency of occurrence; N, numerical percentage; P, point percentage; MFI, main food index; R, classification of prey.

Table 7 .
-Prey identified in the stomach contents of P. acarne females according to season.FO, frequency of occurrence; N, numerical percentage; P, point percentage; MFI, main food index; R, classification of prey.

Table 8
. -Prey identified in the stomach contents of Pagellus acarne males in relation to season.FO, frequency of occurrence; N, numerical percentage; P, point percentage; MFI, main food index; R, classification of prey.

Table 11 .
-Prey identified in the stomach contents of P. acarne adults in relation to season.FO, frequency of occurrence; N, numerical percentage; P, point percentage; MFI, main food index; R, classification of prey.