femelles étaient le Lepidopus caudatus, le Scomber japonicus et le Trachurus picturatus, que l’on a détecté dans plus de 50 % des contenus stomacaux de chaque sexe. Toutefois, il semble y avoir une différence notable entre la principale proie préférée par chaque sexe, le L. caudatus chez les femelles et le Capros aper chez les mâles. Pour ce qui est des céphalopodes, l’Ommastrephes bartramii prédominait chez les femelles (22 % des contenus stomacaux), et l’Argonauta argo et l’Histioteutis dofleini (14 %) chez les mâles. Un G-test a servi à comparer la fréquence d’apparition des groupes de proies chez les deux sexes. Aucune différence significative n’a été observée dans l’alimentation entre les mâles et les femelles. Toutefois, la présence plus fréquente de céphalopodes à déplacement rapide, comme l’Ommastrephes bartramii et le Todarodes sagittatus dans l’alimentation des femelles, comme leur absence dans celle des mâles, pourrait indiquer quelque limitation physiologique et quelque différence de comportement entre les sexes. En fait, s’il existe des différences quant au choix de l’habitat des mâles et des femelles, il pourrait également y avoir des différences quant à leur disponibilité à la pêche de palangre. RESUMEN Este documento presenta algunos resultados sobre la dieta del pez espada, Xiphias gladius, en las aguas en torno a Azores. De los 82 estómagos examinados (58 hembras y 24 machos), el 11% estaban vacíos (9 estómagos). En cuanto a los otros 73, se compara entre los sexos la composición por presas y algunas de las peculiaridades biológicas de las especies de presas. En 51 hembras, se identificó un total de 11 especies de peces que contribuían al 65,5% del peso total del estómago. También se identificó un total de 15 especies de cefalópodos (34,4%). En 22 machos se identificó un total de 9 especies de peces (91,4% del peso total del estómago), así como 4 especies de cefalópodos (8,5%). Los teleósteos componían el grupo más importante de presas en el contenido estomacal, tanto de las hembras como de los machos. En términos de frecuencia de presencia de las especies de peces más importantes en la dieta de hembras y machos Lepidopus caudatus, Scomber japonicus y Trachurus picturatus, representaban un porcentaje superior al 50% del contenido estomacal de cada sexo. No obstante , parece existir una diferencia notable sobre la especie de presa preferente en cada sexo. L. caudatus, en hembras y Capros aper en los machos. Entre los cefalópodos, la especie más importante en las hembras era Ommastrephes bartramii, que estaba presente en el 22% del contenido estomacal, y en los machos Argonauta argo y Histioteuthis dofleini (14%). Se empleó un G-test para establecer una comparación de la ocurrencia de grupos presa entre los sexos. No se encontraron diferencias significativas en la dieta entre machos y hembras. Sin embargo, la mayor ocurrencia de especies de cefalópodos de natación rápida, como Ommastrephes bartramii y Todarodes sagittatus en la dieta de las hembras, y su ausencia en la dieta de los machos, podría indicar la existencia de algunas limitaciones fisiológicas y diferencias de comportamiento entre ambos sexos. De hecho, si hay diferencias entre machos y hembras en cuanto a la selección del hábitat, su disponibilidad a la pesquería de palangre de superficie también podría ser distinta. Keywords: Swordfish, Sex, Diet, Behaviour, Fishery Introduction The swordfish is a very aggressive and voracious mesopelagic teleost (Palko et al., 1981; Chalabi & Ifrene, 1993). It is an opportunistic predator feeding mainly on pelagic schooling fishes (Tibbo et al., 1961) as well in schooling cephalopods (Guerra et al., 1993). This species also shows cannibalism during the larval and post-larval stages (Arata, 1954). The swordfish presents some external specialised features according to its feeding habits, namely the shape of the sword (Tibbo et al., 1961; Scott & Tibbo, 1968; Toll & Hess, 1981), the reduced size of the lower mandible (McGowan, 1988) and the presence of large eyes (Carey & Robinson, 1981). Swordfish’s diet throughout their geographical range has been described by numerous authors, e.g. Tibbo et al. (1961), Scott and Tibbo (1968), Scott and Tibbo (1974), Toll and Hess (1981), Hess and Toll (1981), Stillwell and Kohler (1985), Azevedo (1990), Moreira (1990), Chalabi and Ifrene (1993), Guerra et al. (1993), Clarke et al. (1995), Hernández-Garcia (1995), Relini et al. (1995), Barreto et al. (1996), among others. No differences were found between the diets of female and male (Toll & Hess, 1981, Stillwell & Kohler, 1985, Guerra et al., 1993; Barreto et al., 1996). However, there are few studies that compare the ecology of prey and predator in each sex. In this sense, the main objective of this paper is to analyse the prey species composition in the diet of each sex, from a sample taken in Azores area. In addition, the existence of a relationship between prey ecology and swordfish individual behaviour was also verified. Material and methods The stomach contents of 82 swordfish were analysed. From these 58 were females (86- 233 cm Ljfl) and 24 were males (99-174 cm Ljfl) (Figure 1). The sample was taken during the commercial landings of the Azores longline fleet. For each individual sampled, size, weight and sex were registered along with the fishery data (gear, month and area if possible). Stomachs were removed immediately after sampling. After being deep frozen, the contents of each stomach were thawed out and laboratory proceedings were performed according to Toll and Hess (1981) and Herrán (1988). For each stomach, the basic contents (fish, crustaceans, cephalopod and unidentified items) were sorted, fixed in 10% formalin and transferred to 70% ethyl alcohol for storage. Prey items were identified to the lowest possible taxonomic category. Fish were identified from otoliths and bones using keys provided in Smale et al. (1995) and author’s collection. Scomber sp. that showed hook marks and/or occurred quasi intact in the stomach contents were assumed to be bait and rejected for analysis. Lower beaks were used as the primary means for classification of cephalopods and beak identity was established by available keys provided in Clarke (1962;1986), Pérez-Gandaras (1986), Voss (1969;1985), Wolff (1982) and author’s collection. In some cases, upper beaks, the known geographical distribution and the existence of monotypism were used (Clarke, 1986). The otolith length (OL) and the lower rostral length (LRL) were measured by stereoscopic microscope or digital calliper. The weights (g) of the teleosts from which otoliths came were then estimated from OL’s by using relationships provided in Isidro (1990), Moreira (1990), Clarke et al. (1995) and Smale et al. (1995). The weights (g) of the cephalopods from which lower beaks came where then estimated from LRL’s by using relationships published in Clarke (1962; 1986). Nematodes were found in the majority of the stomachs and in some of them, in large quantities, but these were assumed to be parasites and were not considered as prey. The following indices were used to quantify the diet in each sex: -Frequency of occurrence as the number of stomachs where a certain prey item occurred, expressed as percentage to the total number of stomachs (f%) (Hyslop, 1980). -Weight percentage as the wet weight of a particular prey item relative to the total weight of all prey, in the whole sample (w%) (Herrán, 1988) -Relative frequency (in number) as the number of a certain prey item expressed as percentage to the total number of prey, in the whole sample (n%) (Herrán, 1988). A comparison of the diet of each sex was made applying the G-test to the frequency of occurrence of the prey groups. The statistic G was corrected by Williams’s correction factor (Gadq) (Zar, 1984). To analyse the prey importance in the diet of each sex, the prey alimentary factor, Q (*) (Hureau, 1970) was used according to Geistedoerfer prey classification (Geistedoerfer, 1975). This author presented a new proposal of the Hureau prey grouping, introducing the prey frequency of occurrence in prey classification. (*) Q=(%n * %w) (Hureau, 1970) Results From the 82 stomachs examined, 11% were empty (9 stomachs). In the remaining 73, a total of 11 fish species, belonging to 11 families and 11 genera, was identified, which corresponded to 467 specimens. A total of 15 cephalopod species belonging to 13 families and 17 genera was identified, corresponding to 83 specimens. One Penaeidae family was also identified in female’s stomach contents, which corresponded to 1 specimen. Due to the low crustacean representativity in the stomach contents, this prey group was not considered in the present analysis (Table I and II). In the 51 females observed, a total of 11 fish species was identified (256 specimens) and contributed to 65.5% of the total stomach weight. A total of 15 cephalopod species was also identified (65) and contributed to 34.4% of the total stomach weight. In 22 males, a total of 9 fish species was identified (211 specimens) and contributed to 91.4% of the total stomach weight. Four cephalopod species (18) were also identified, contributing to 8.5% of the total stomach weight. Figure 2 and 3 show the frequency of occurrence of teleosts prey and the respective length interval found in female and male’s diet, respectively. The majority of identified species were common on both sexes, occurring in a wide length interval. L. caudatus (Silver scabbard-fish), S. japonicus (Spanish mackerel) and T. picturatus (Horse mackerel) were the only species that occurred in more than 50% of stomach contents, in each sex. C. aper (Boar-fish), D. argenteus and Myctophidae (lanterfishes), small species, showed a higher occurrence in male’s stomach contents than on females. Figure 4 and 5 show the frequency of occurrence of the common cephalopod preys in the diet of both sexes and the respective length interval found in female and male’s diet, respectively.