Typology and behaviour of tuna aggregations around fish aggregating devices from acoustic surveys in French Polynesia

Aquat. Living Resour. 13 (2000) 183 192 2000 Ifremer/CNRS/INRA/IRD/Cemagref/Éditions scientifiques et médicales Elsevier SAS. All rights reserved S0990744000000516/FLA Typology and behaviour of tuna aggregations around fish aggregating devices from acoustic surveys in French Polynesia Erwan Josse a *, Laurent Dagorn b, Arnaud Bertrand b a Centre IRD de Bretagne, BP 70, 29280 Plouzané, France b Centre IRD de Montpellier, BP 5045, 34032 Montpellier cedex 1, France Received 8 November 1999; accepted 20 March 2000 Abstract Eighty-seven two-hour acoustic surveys (radius 0.8 nautical mile, vertical range 0 500 m) around 17 fish aggregating devices (FADs) were conducted in French Polynesia between December 1995 and February 1997. Associated tuna densities were calculated using two different techniques: echo counting when the fish had sufficient distances from each other and echo integration when the fish swam close together (in schools). No acoustic detection of tuna was observed during 27 of the 87 surveys, representing 81 % of all the nocturnal surveys and 15 % of the diurnal ones. The 60 other surveys showed three different classes of aggregations: (1) deep scattered fish, observed 45 times, (2) intermediate scattered fish, observed 16 times, and (3) shallow schooling fish, observed 16 times. Sometimes aggregations of different classes were observed beneath the same FAD. The size of the fish inside the aggregations (determined from target strength values), the distance between the individuals, and the depth of the fish all decreased from deep scattered fish to shallow schooling fish (100 300 m for deep scattered fish, 50 150 m for intermediate scattered fish, and above the depth of 50 m for shallow schooling fish ). Fish densities also varied according to the class of aggregations: 7.3, 26, and 801 fish per km 3 on average for deep scattered fish, intermediate scattered fish, and shallow schooling fish, respectively. The highest densities were observed during daytime, while night-time observations indicated a variety of situations, from the absence of individuals to large amounts of fish. 2000 Ifremer/CNRS/INRA/IRD/Cemagref/Éditions scientifiques et médicales Elsevier SAS tuna fish / fish aggregating devices / acoustics / aggregation / behaviour / French Polynesia Résumé Typologie et comportement des agrégations thonières autour de dispositifs de concentration de poissons à partir de prospections acoustiques en Polynésie française. Quatre-vingt-sept prospections acoustiques, d une durée de deux heures, ont été effectuées entre décembre 1995 et février 1997 en Polynésie française autour de 17 dispositifs de concentration de poissons (DCP). Les densités en thons, entre la surface et 500 m de profondeur et dans un rayon de 0.8 mille nautique autour des DCP, ont été estimées par écho-comptage en présence de poissons dispersés, ou par échointégration en présence de poissons agrégés en bancs. Lors de 27 prospections, (représentant respectivement 81 % et 15 % des prospections nocturnes et diurnes effectuées), aucune détection de thon n a été observée. Les 60 autres prospections ont montré 3 différents types d agrégation : (1) le type «poissons dispersés profonds» observé 45 fois, (2) le type «poissons dispersés intermédiaires» observé 16 fois et (3) le type «poissons en bancs superficiels» observé 16 fois. Différents types d agrégation ont parfois été observés sous un même DCP. La taille des poissons (déterminée à partir des valeurs d indice de réflexion individuel ou TS), la distance entre les individus et la profondeur des détections diminuent, alors que les densités augmentent, entre les types «poissons dispersés profonds» (7.3 poissons par km 3 entre 100 et 300 m de profondeur), «poissons dispersés intermédiaires» (26 poissons par km 3 entre 50 et 150 m) et «poissons en bancs superficiels» (801 poissons par km 3 entre la surface et 50 m). Les densités les plus fortes ont été observées de jour, alors que de nuit, différentes situations, depuis l absence totale de détection jusqu à la présence de densités élevées ont été rencontrées. 2000 Ifremer/CNRS/INRA/IRD/Cemagref/Éditions scientifiques et médicales Elsevier SAS thon / dispositif de concentration de poissons / acoustique / agrégation / comportement / Polynésie française *Corresponding author. E-mail address: Erwan.Josse@ird.fr (Erwan Josse).

184 E. Josse et al. / Aquat. Living Resour. 13 (2000) 183 192 Figure 1. Geographical localization of FADs where echo surveys were conducted. FAD position. 1. INTRODUCTION Many types of tropical tuna fisheries take advantage of the fact that tuna aggregate around floating objects. For the last two decades, anchored man-made floating objects, known as fish aggregating devices (FADs), have greatly helped to develop and maintain artisanal tuna fisheries, especially in tropical islands of the Pacific and Indian Oceans. At the same time, purse seine fisheries extend the use of drifting FADs to increase catches. Knowledge of fish behaviour and spatial structuring of tuna aggregation under floating objets has become a main stake in the management of tuna fisheries. Therefore, several approaches have already been used to study tuna aggregations. Ultrasonic telemetry techniques were used in Tahiti, Hawaii, and in the Indian Ocean to observe the fine-scale vertical and horizontal movements of FAD-associated skipjack Katsuwonus pelamis (Linnaeus, 1758), yellowfin Thunnus albacares (Bonnaterre, 1788) and bigeye Thunnus obesus (Lowe, 1839) tuna (Cayré and Chabanne, 1986; Holland et al., 1990; Cayré, 1991; Marsac et al., 1996; Bach et al., 1998; Josse et al., 1998; Marsac and Cayré, 1998; Brill et al., 1999). This method provides useful information about the behaviour of individuals. This behaviour however, may not always reflect the behaviour of entire aggregations and it appears necessary to observe aggregations and not only individuals. In this study we present acoustic observations of tuna aggregations with a discrimination between the echoes of tuna or tuna-like species and other ones. The objective of this work is to define the vertical and horizontal structures of tuna aggregations around FADs as well as their time dynamics, in order to identify aggregation patterns. 2. MATERIALS AND METHODS The experiments were conducted in French Polynesia between December 1995 and February 1997 within the framework of the Ecotap programme (Studies of tuna behaviour using acoustic and fishing experiments), a joint project between two French research institutes (Ifremer: Institut français de recherche pour l exploitation de la mer, and IRD: Institut de recherche pour le développement), and a French Polynesian institute (SRM: Services des ressources marines). Acoustic surveys were carried out around FADs anchored some nautical miles off the main inhabited islands of the Society and Tuamotu Archipelagos (figure 1). Similar surveys were carried out off the Marquesas Islands, around an instrumented oceanographic buoy anchored approximately 200 nautical miles from the nearest land. Data were collected onboard the 28 m IRD Research Vessel Alis equipped with a SIMRAD EK500 echo sounder (version 4.01). The sounder was connected to a SIMRAD ES38B hull-mounted, split-beam transducer producing pulse duration of 1.0 ms at 38 khz. The beam angle was 6.9. Calibration of acoustic equipment was performed with a 60 mm copper sphere using the standard procedure recommended by the manufacturer (Simrad, 1993). Measurements of the acoustic noise level according to the vessel speed were used to define the optimal speed survey (i.e. 7 knots, see Josse et al., 1999). Three survey patterns were defined, based upon a maximum survey time a priori fixed at 2 h (figure 2): transect 1, a star survey pattern with eight branches, each 0.8 nautical mile long and repeated twice (figure 2a),

E. Josse et al. / Aquat. Living Resour. 13 (2000) 183 192 185 Figure 2. Survey patterns used during acoustic surveys around FADs in French Polynesia. FAD position; start of the survey; end of the survey. transect 2, a star survey pattern with eight branches, each 1.2 nautical miles long, without duplicate (figure 2b), transect 3, a star survey pattern with twelve branches, each 1.0 nautical mile long, without duplicate (figure 2c). The transect 1 pattern, already used during previous acoustic surveys around FADs in French Polynesia (Depoutot, 1987; Josse, 1992; Bach et al., 1998), allows an increase in the number of runs close to the FAD and exploration of an area a priori wide enough to take into account the whole aggregation. The two other survey patterns make the navigation close to the FAD easier, and allow extension of the area prospected. On the other hand, the number of runs close to the FAD is reduced. All the acoustic measurements were carried out between the surface and a depth of 500 m. SIMRAD EP500 software (Simrad, 1994) was used to record, via ETHERNET on a personal computer (PC), acoustic and navigation data from the EK500 echo sounder. In order to estimate the tuna densities associated with FADs, surveyed areas were partitioned into 30 or 45 angular sectors based upon the survey pattern used. Each angular sector was then subdivided into volumes, using the sector s distance to the FAD (0.1 nautical mile increments) and an arbitrary depth category. Depth categories included one 40-m layer from depths between 10 and 50 m, and nine 50-m layers for depths between 50 and 500 m (see Josse et al., 1999). In each elementary sampling unit thus defined, the densities, expressed as a number of fish per unit volume, were then determined by echo counting in the presence of scattered fish, or by echo integration in the presence of schools. All analyses were limited to a radius of 0.8 nautical mile around the FAD, irrespective of the type of survey pattern carried out. The applicability and validation of the two techniques (echo counting and echo integration) used to estimate tuna densities around FADs were presented in Josse et al. (1999). Echo counting is a technique which allows to obtain direct quantitative estimations of fish density (Kieser and Ehrenberg, 1990), provided that the fish are sufficiently distant from each other so that their echoes can be discriminated. A split-beam system allows direct application of this technique (Misund, 1997), following three steps. The first step consists of identifying and counting all the fish. The EP500 trace tracking procedure allows the automated recognition of a single fish detected over one or more successive pings. In a second step, each identified fish must be allocated to the elementary sampling unit corresponding to its space location. EP500 specifies the depth of each identified target, but its geographical position must be researched in the raw data files. The third step consists of converting the number of fish obtained in a basic sampling unit into a density value. This step requires the knowledge of the water volume sampled by the acoustic beam which is characterized by its beam angle. A split-beam system allows easy determination of the beam angle from the angular co-ordinates associated with individual echoes. Thus, the beam angle was determined using all the angular data associated with tuna and non-tuna echoes observed during the different surveys around FADs (see Josse et al., 1999). In echo integration, contrary to echo counting, no restriction of use related to dispersion of fish occurs. This method can thus apply when fish are aggregated in schools. The acoustic densities were extracted from each sampling unit by using the EP500. The acoustic densities were then transformed into absolute densities, which requires the knowledge of the average target strength index (TS) of the detected fish. The TS values were extracted from each survey using the EP500 trace tracking procedure. An average TS was then calculated and used to transform the acoustic

186 E. Josse et al. / Aquat. Living Resour. 13 (2000) 183 192 Table I. Target strength values (TS) for yellowfin (Thunnus albacares) and bigeye tuna (Thunnus obesus) from the literature. Species Fork length (cm) Estimated weight (kg) Average TS(dB) References Thunnus albacares 60 4 34.8 Bertrand et al., 1999a, b 90 14 33.0 108 25 30.4 120 30 26.1 Thunnus obesus 49.9 (*) 3 32.8 Josse and Bertrand, in press 50.1 (*) 3 31.9 110 30 24.4 Bertrand et al., 1999a, b 130 50 21.4 (*) Mean value densities into absolute densities (expressed as a number of fish per volume unit). Knowledge of fish TS is of prime importance for acoustic estimation both with echo counting and with echo integration. When echo counting is used for tuna biomass estimation, it is necessary to count only echoes from tuna and to exclude other echoes. When echo integration is used, it is necessary to know the mean TS value of the detected targets (i.e. the mean TS values of the detected tuna) to convert acoustic densities into tuna densities. Experiments were carried out within the framework of the Ecotap programme to determine TS values for yellowfin and bigeye tuna (see Bertrand et al., 1999a, b; Josse and Bertrand, in press; and table I). During the Ecotap programme, 87 acoustic surveys were carried out around 17 FADs. Some surveys were done in the same 24-hours period on the same FAD to collect information on the temporal evolution of the aggregations and therefore, construct series of surveys. Tuna were acoustically observed during 60 surveys. A visual analysis of the echograms of each survey, coupled to a search of the individual targets with the EP500 software, showed differences between 3-D organisation of tuna around FADs. According to these differences, the detection of tuna was classified into three categories: (1) presence of deep scattered fish (45 times); (2) presence of intermediate scattered fish (16 times); and (3) presence of shallow schooling fish (16 times). Sometimes, two or more categories were observed together in the same survey. When one or both of the two first categories were observed with the last one, it was not possible to quantitatively characterize each category. For the first two classes, the echo counting technique was systematically employed. For the last category, echo integration was systematically applied as it was not always possible to identify all the individuals. 3. RESULTS 3.1. No acoustic detection of tuna In 27 surveys (31 % of the surveys) no tuna were detected. Seventeen of them were nocturnal surveys and ten were diurnal surveys, which corresponds to 81 % and 15 % of the surveys, respectively. In particular, around two FADs in the Society Archipelago, observations were made over 24-hour periods, which indicates that no fish were around or beneath those two FADs during a period of 24 hours. In the other series of observations where in some surveys no fish were detected, no sequential pattern of aggregation types found before or after the absence of fish can be drawn. All the three types of aggregations were found in those series of surveys. 3.2. Deep scattered fish This type of aggregation was observed forty five times: thirty two times alone, ten times in the presence of intermediate scattered fish and three times in the presence of shallow schooling fish. We will only show results when shallow schooling fish were not simultaneously observed, because in those cases it was not possible to quantitatively differentiate the aggregation types. The average density per survey was 7.3 fish per km 3, which corresponded to 25 individuals per survey (table II). Highest densities were observed close to the FAD. They then decreased and the minimum values were found between 0.5 and 0.6 nautical mile from the FAD. High densities were also found at the edge of the sampling area (0.8 nautical mile). Fish were observed from the depth of 10 m down to 500 m, but 92 % of the fish was distributed between 100 m and 300 m. The average distance between two fish along a same branch was 240 m (number of observations n = 29, SD = 340 m). TS values varied between 34.4 and 19.0 db, giving a mean value of 23.0 db. Two peaks of TS values were observed, the first one between 26 and 28 db, and the second one between 20 and 22 db (figure 3a). The comparison with TS values found by Bertrand et al. (1999a, b) indicates that these fish are likely to be large tuna of more than 100 cm fork length (FL). However, the acoustic techniques do not make it possible to identify the species directly, without access to complementary data such as, for example, simultaneous fishing data.

E. Josse et al. / Aquat. Living Resour. 13 (2000) 183 192 187 Table II. Deep scattered fish : densities (d: number of fish km 3 ) and numbers of fish (n) per depth and distance to the FAD strata*. Depth strata (m) Distance to the FAD strata (nautical mile) 0.0 0.1 0.1 0.2 0.2 0.3 0.3 0.4 0.4 0.5 0.5 0.6 0.6 0.7 0.7 0.8 0.0 0.8 10 50 d n 50 100 d 12.9 0.2 n 0.1 0.1 100 150 d 12.5 7.7 20.1 7.6 7.7 26.5 37.6 18.2 n 0.1 0.1 0.5 0.3 0.4 1.9 3.0 6.3 150 200 d 56.7 32.1 21.3 37.0 21.3 11.0 13.4 16.1 n 0.3 0.5 0.6 1.4 1.0 0.7 1.1 5.6 200 250 d 43.4 35.4 19.8 25.2 15.5 8.6 6.1 32.1 19.1 n 0.2 0.6 0.5 0.9 0.8 0.5 0.4 2.6 6.6 250 300 d 58.2 17.4 13.5 8.5 3.5 5.0 6.9 24.0 12.1 n 0.3 0.3 0.4 0.3 0.2 0.3 0.5 1.9 4.2 300 350 d 10.2 3.0 3.0 8.9 11.5 5.0 n 0.1 0.1 0.1 0.5 0.9 1.7 350 400 d 2.6 0.4 n 0.1 0.1 400 450 d n 450 500 d n 10 500 d 19.8 9.5 7.9 8.0 5.5 3.4 4.0 12.1 7.3 n 1.0 1.5 2.1 3.0 2.6 2.0 2.8 9.6 24.5 * These values are means calculated from 42 echo surveys. 3.3. Intermediate scattered fish This type of aggregation was observed sixteen times: two times alone, ten times in the presence of deep scattered fish, and four times with the presence of both deep scattered fish and shallow schooling fish. These four surveys will not be considered in the present results. An average density per survey of 26 fish per km 3 characterizes the 12 other surveys, i.e. 87 fish per survey (table III). The highest densities were found close to the FAD (the values are 40 times higher than those encountered for deep scattered fish ). These densities rapidly decreased with an increasing distance from the FAD, to drop to zero at 0.4 nautical mile. The densities were observed from the surface down to 200 m, but 94 % of the fish were found between 50 and 150 m. The average distance between two fish along a same branch was 39 m (n = 100, SD = 72 m). The TS values varied between 40.3 db and 18.7 db giving a mean value of 30.6 db and a single mode between 36 and 34 db (figure 3b). These values likely correspond to tuna of less than 100 cm FL, according to Bertrand et al. (1999a, b) and Josse and Bertrand (in press). 3.4. Shallow schooling fish This type was observed sixteen times around only two FADs. These aggregations are characterized by an average density of 801 fish per km 3, i.e. 2 708 fish per survey (table IV). The higher densities were located close to the FAD and between the depths of 10 and 50 m. Despite some small schools observed in the same vertical range but further from the FAD, densities were very low outside this area. The fish detected in the first 0.1 nautical mile around the FAD and between the depths of 10 and 50 m represented 72 % of the tuna detected. The average distance between two fish could not be calculated for this class as it was often lower than the acoustic beam diameter (2RTan(α/2)), where α is the beam angle and R the depth). For instance, at a depth of 50 m, two fish 6 m away from one another could not be distinguished. The TS values varied between 45.9 db and 18.8 db for a mean value of 32.6 db. Two modes were observed. The first between 42 and 40 db and the second between 32 and 28 db (figure 3c). Troll-fishing around one of the two FADs indicated that the aggregations were mainly composed of juvenile bigeye tuna with an average FL of 50 cm (Josse and Bertrand, in press). 3.5. Temporal dynamics The biomass observed for the shallow schooling fish type was at a maximum after sunrise. It then decreased during daytime. The night-time biomass varied between an absence of fish to a great quantity of aggregated fish, but even in this last case, the biomass was lower than the one observed during daytime (figure 4). This cycle was also observed for the two

188 E. Josse et al. / Aquat. Living Resour. 13 (2000) 183 192 Figure 3. Distribution of TS values for the three types of aggregation: (a) deep scattered fish, (b) intermediate scattered fish and (c) shallow schooling fish. other aggregation types, but was more obvious for shallow schooling fish because of the high observed densities. 4. DISCUSSION Despite a significant scientific interest in tuna associated with floating objects, the structure and the dynamics of the aggregations are badly known. Several studies have focused on the individual behaviour of tuna around FADs using acoustic telemetry (Cayré and Chabanne, 1986; Holland et al., 1990; Cayré, 1991; Marsac et al., 1996; Bach et al., 1998; Josse et al., 1998; Marsac and Cayré, 1998; Brill et al., 1999), but very few have been dedicated to the direct observation of tuna aggregations beneath FADs. Cillauren (1994) studied tuna aggregations around FADs anchored off Efate Island in Vanuatu from artisanal troll-fishing data. Mixed schools of skipjack and yellowfin tuna were mostly present within 200 m of the FAD, with exceptional captures beyond 500 m. Yellowfin tuna were caught closer to the FADs than skipjack tuna. The highest catches of yellowfin tuna were taken towards midday, whereas an increase in skipjack tuna catches occurred one hour before sunset. These results however, do not necessarily reflect the behaviour of the entire aggregation. Results obtained from troll-fishing operations depend on the feeding motivation of fish and on the fish-gear interactions. Acoustic data, on the contrary, provide information about the biomass without using fishing gear. Depoutot (1987) and Josse (1992) made acoustic observations around FADs in French Polynesia using a SIMRAD EYM single beam echo sounder with a frequency of 70 khz. The echo sounder used did not allow the collection of data below a depth of 120 m. Moreover, in the absence of any data on tuna TS, the authors could not distinguish between tuna or tuna-like species echoes and other species echoes. From one series of surveys, Depoutot (1987) concluded that in shallow waters (above 40 m), the biomass is mostly concentrated around the FAD during the night, while during daytime, fish are observed below 60 m. Josse (1992) observed acoustic detection between 10 and 60 m during daytime, but no detection during the night. This daytime biomass appeared in the morning and disappeared in the afternoon. Considering tuna catches by local fishermen using the drop-stone fishing technique (Moarii and Leproux, 1996), the author also concluded there was a secondary concentration encountered during daytime, below 150 m. Other acoustic observations were done around FADs in French Polynesia using a BIOSONICS model 102 dual beam echo sounder with a frequency of 120 khz (Bach et al., 1998). For the first time, the presence of isolated fish down to the depth of 250 m, the maximum vertical range of the echo sounder used, was observed. However, it is difficult to compare those results with those of the present study as the authors did not observe the same vertical range and, above all, could not distinguish tuna or tuna-like species echoes from other ones. Other small fish of various families (Carangidae, Balistidae, etc.) are known to aggregate around FADs in this region, a few metres below the surface, and could therefore be observed by the echo sounder. Moreover, during night-time, the detection may also correspond to organisms of the Sound Scattering Layer (Myctophidae, etc.) that ascend to the surface layer after sunset. The scientific echo sounder used in the present study allows determination of the TS of the organisms. Low values characterize small fish while higher TS values represent larger fish (see Bertrand et al. 1999a; b and Josse and Bertrand (in press) for tuna TS values). However, it is not possible to acoustically identify the species without complementary data. The size of the

E. Josse et al. / Aquat. Living Resour. 13 (2000) 183 192 189 Table III. Intermediate scattered fish : densities (d: number of fish.km 3 ) and numbers of fish (n) per depth and distance to the FAD strata*. Depth strata (m) Distance to the FAD strata (nautical mile) 0.0 0.1 0.1 0.2 0.2 0.3 0.3 0.4 0.4 0.5 0.5 0.6 0.6 0.7 0.7 0.8 Total 10 50 d 902 14 n 4 4 50 100 d 5 017 818 683 219 194 n 27 13 18 8 66 100 150 d 2 401 158 45 n 13 3 16 150 200 d 207 3 n 1 1 200 250 d n 250 300 d n 300 350 d n 350 400 d n 400 450 d n 450 500 d n 10 500 d 852 100 70 22 26 n 45 16 18 8 87 * Means calculated from 12 echo surveys. Table IV. Shallow schooling fish : densities (d: number of fish km 3 ) and numbers of fish (n) per depth and distance to the FAD strata*. Depth strata (m) Distance to the FAD strata (nautical mile) 0.0 0.1 0.1 0.2 0.2 0.3 0.3 0.4 0.4 0.5 0.5 0.6 0.6 0.7 0.7 0.8 0.0 0.8 10 50 d 451 425 2 309 339 8 161 405 1 338 7 612 n 1 946 30 7 0.2 8 23 86 2 100 50 100 d 55 219 2 020 361 201 1 020 n 298 33 10 12 352 100 150 d 18 233 5 597 1 070 10 5 633 n 98 90 29 0.4 0.2 218 150 200 d 2 687 329 57 n 14 5 20 200 250 d 42 232 113 24 n 0.2 4 4 8 250 300 d 618 60 12 n 3 1 4 300 350 d 600 51 12 n 3 0.8 4 350 400 d 387 6 n 2 2 400 450 d n 450 500 d n 10 500 d 44 788 1 034 174 13 0.5 34 33 109 801 n 2 365 164 46 5 0.2 20 23 86 2 708 * Means calculated from 16 echo surveys.

190 E. Josse et al. / Aquat. Living Resour. 13 (2000) 183 192 Figure 5. Typology of aggregations around a fish aggregating device: (a) deep scattered fish, (b) intermediate scattered fish and (c) shallow schooling fish. For each type of aggregation, a schematic histogram of TS values distribution observed is represented. Figure 4. Temporal evolution of the aggregated biomass on a 24-hours basis from acoustic data. In solid line: average situation observed during the first part of the day. In dotted lines: different situations observed in the afternoon and during the night. tuna comprising the aggregations, the distance between two individuals, and the depth of fish decrease from the type deep scattered fish to the type intermediate scattered fish, then to the type shallow schooling fish (figure 5), while the density of fish increases. This typology argues for a size-dependent behaviour of fish around FADs. Larger fish occupy the deepest waters and do not really seem to form a structured aggregation. Smaller fish are found in shallower waters and form schools that stay close to the FAD. This size-dependent vertical distribution is commonly accepted for fish schools or fish aggregations, but has rarely been confirmed by in situ observations (Parrish, 1989). Anchored FADs are known to have some effects on tuna movements (Cayré and Chabanne, 1986; Holland et al., 1990; Cayré, 1991; Kleiber and Hampton, 1994; Bach et al., 1998; Josse et al., 1998; Marsac and Cayré, 1998), implying consequences for the spatial distribution and residence time of fish in an area. If it is assumed that all tuna that are present in an area can be attracted by FADs located within that area, all the different individuals (species and sizes) should be found around or beneath a FAD. The behaviour of fish however, might be different depending on the species and the size of fish, and the proportion of each category might therefore change. Data collected on some small artisanal fishing boats in Tahiti showed interesting species composition (table V). Local fishermen use two different techniques to catch fish around FADs. The more frequently used fishing technique is the drop-stone technique (Moarii and Leproux, 1996) where hooks are set between the depths of 150 and 350 m, depending on the strategy and current conditions. Such depths correspond to the depth strata where deep scattered fish were observed by acoustic methods. The other technique is trolling to catch fish in the first few metres below the surface. All the captures are made during daytime. Around 85 % of the overall catch in number of fish are taken with the first fishing technique, i.e. deep fishing. These results are in agreement with our results, as aggregated schools of small individuals were very rarely observed around FADs in the Tuamotu or the Society Archipelagos. When fish were detected by the echo sounder, the deep scattered fish type was always observed, which shows that large fish in deep waters are more frequently encountered than smaller schooling fish. Fish caught by the drop-stone technique are likely to belong to this class. Composition of FAD-associated captures indicates that the deep scattered fish are likely to be mainly large albacore Thunnus alalunga (Bonnaterre, 1788) and yellowfin tuna (> 100 cm fork length). Fish of this category showed feeding motivation beneath FADs as they reacted to the baited hooks. Echo counting was used by Bertrand and Josse (in press) to estimate the longline tuna abundance in the French Polynesia exclusive economic zone (EEZ) between the surface and a depth of 500 m. A density of 1.3 fish per km 2, 2.7 fish per km 3, was found. This value was obtained inside an area where all the FADs of the present study were located. If only the deep scattered fish class is considered, the average density around FADs anchored in nearshore area (7.3 fish per km 3 ) was nearly 2.7 higher than the density observed in the open ocean without FADs. It is not possible, however, to determine if these differences were due to an attractive effect of FADs or to a nearshore versus offshore effect of the biological environment. The intermediate scattered fish class was very rarely observed alone. It was most of the time observed in association with deep scattered fish. As fishermen do not frequently set their hooks in depths

E. Josse et al. / Aquat. Living Resour. 13 (2000) 183 192 191 Table V. Catches made around FADs by small artisanal fishing boats in Tahiti in 1997*. Species Fishing gear Number of fish Total weight Mean weight n % kg % (kg) Thunnus alalunga Drop-stone 615 69.8 11 675 62.6 19.0 Thunnus albacares Drop-stone 125 14.2 5 382 28.8 43.1 Thunnus obesus Drop stone 3 0.3 177 0.9 59.0 Katsuwonus pelamis Trolling 56 6.4 109 0.6 1.9 Coryphaena hippurus Trolling 44 5.0 427 2.3 9.7 Acanthocybium solandri Trolling 2 0.2 22 0.1 10.7 Small tuna Trolling 25 2.8 75 0.4 3.0 Billfish Trolling 11 1.2 796 4.3 72.4 * The total fishing effort during the year for all these boats was 361 fishing days. The category small tuna represents either yellowfin (Thunnus albacares) or bigeye (T. obesus) tuna of less than 5 kg. The category billfish gathers the sworfish Xiphias gladius (Linnaeus, 1758), the blue marlin Makaira mazara (Jordan and Snyder, 1901) and the striped marlin Tetrapterus audax (Philippi, 1887). corresponding to the acoustic detection of these fish, artisanal fishing data cannot be used to determine the species and size composition of this class. TS values show that fish are smaller than those composing the deep scattered fish group, and the few experimental catches made during the scientific surveys tend to indicate that these fish are mainly yellowfin tuna with a fork length between 60 and 100 cm, with some albacore tuna with less than 100 cm fork length. Experimental troll-fishing showed that fish composing the shallow schooling fish were mainly small bigeye and yellowfin tuna (mean size of 50 cm fork length, see Josse and Bertrand, in press). The two series of surveys made around the FAD anchored north of the Marquesas Islands always showed the presence of shallow schooling fish while only one series of surveys around one FAD anchored in the Society Archipelago exhibited the presence of this type of aggregation. The observation of shallow schooling fish of small bigeye and yellowfin tuna might be due to the northern location of this buoy, in an area where small tuna are likely to be more abundant than in the Tuamotu or Society Archipelagos. Series of several two-hour surveys over the same FADs and over a short period (24 hours for instance) allowed the study of the temporal evolution of the aggregated tuna biomass. The maximum biomass seems to occur during daytime and a total of 81 % of the nocturnal surveys did not show the presence of fish around FADs. This result may confirm a behaviour pattern reported by different authors: some fish tagged with an ultrasonic device spent the daytime at the FAD site and left it at night (Holland et al., 1990; Cayré, 1991; Bach et al., 1998; Marsac and Cayré, 1998). Some nocturnal surveys however, indicated the presence of fish aggregated around FADs, which also confirms some other sonic tagging experiments where fish spent the entire duration of the track associated with the FAD, day and night (Cayré and Chabanne, 1986; Bach et al., 1998). Even if the present acoustic data tend to confirm that fish are mostly aggregated beneath FADs during daytime, the presence of associated fish during night-time is in agreement with the variety in diel horizontal patterns observed for individual fish. Individuals composing an aggregation might behave differently, as shown by the different patterns observed using ultrasonic telemetry. The cooccurrence of different individual behaviours should lead to different dynamics for the aggregations, as observed in the current study. The number of nocturnal surveys in this study however, was inferior to the number of diurnal surveys (21 versus 66), and more comparative observations between daytime and nighttime aggregations should be developed. 5. CONCLUSION This data represents the first acoustic characterization of tuna aggregations around FADs. The three different types of aggregations i.e.: deep scattered fish, intermediate scattered fish and shallow schooling fish, seem to be related to different sizes (and maybe species) of fish, as shown by the TS values. The structure and the dynamics of some aggregations observed with this acoustic tool complement results on fine-scale movements of individuals observed by acoustic telemetry. If we add the fact that acoustics also provide information about the biological environment of fish (Josse et al., 1998), we thus have two useful and complementary tools to improve our knowledge on tuna behaviour around floating objects. These tools should be used together for studying tuna aggregations, especially on drifting floating objects. Acknowledgements. The Government of French Polynesia and the co-operating research organisations funded this research. The authors sincerely thank the officers and crew of the R/V Alis for providing valuable help during all the cruises achieved in the Ecotap programme. They would like to thank all their colleagues from Ifremer, IRD and SRM who worked with them during this program in French Polynesia. Special thanks to Ah-Soy Asine for collecting the

192 E. Josse et al. / Aquat. Living Resour. 13 (2000) 183 192 catch data from local fishermen. The authors are grateful to Tony Lewis from the Secretariat of the Pacific community for revising the English of this paper. The authors are also grateful for the helpful insights and comments of the two referees. An oral presentation of this work was made during the international symposium Tuna Fishing and Fish Aggregating Devices which was held in Martinique (French West Indies) from 15 to 19 October 1999. References Bach, P., Dagorn, L., Josse, E., Bard, F.X., Bertrand, A., Misselis, C., 1998. Experimental research and fish aggregating devices (FADs) in French Polynesia. SPC Fish Aggregating Device Inform. Bull., 3 19. Bertrand, A., Josse, E., in press. Acoustic estimation of longline tuna abundance. ICES J. Mar. Sci.. Bertrand, A., Josse, E., Massé, J., 1999. Preliminary results of acoustic target strength measurements of bigeye (Thunnus obesus) and yellowfin tuna (Thunnus albacares). In: Séret, B., Sire, J.-Y. (Eds.), Proc. 5th Indo-Pac. Fish Conf., Soc. Fr. Ichtyol., Paris, pp. 443 450. Bertrand, A., Josse, E., Massé, J., 1999. In situ acoustic target-strength measurements of bigeye (Thunnus obesus) and yellowfin tuna (Thunnus albacares) by coupling split-beam echosounder observations and sonic tracking. ICES J. Mar. Sci. 56, 54 60. Brill, R.W., Block, B.A., Boggs, C.H., Bigelow, K.A., Freund, E.V., Marcinek, D.J., 1999. Horizontal movements and depth distribution of large adult yellowfin tuna (Thunnus albacares) near the Hawaii Islands recorded using ultrasonic telemetry: implications for the physiological ecology of pelagic fishes. Mar. Biol. 133, 395 408. Cayré, P., 1991. Behaviour of yellowfin tuna (Thunnus albacares) and skipjack tuna (Katsuwonus pelamis) around fish aggregating devices (FADs) in Comoros Islands as determined by ultrasonic tagging. Aquat. Living Resour. 4, 1 12. Cayré, P., Chabanne, J., 1986. Marquage acoustique et comportement des thons tropicaux (albacore: Thunnus albacares et listao: Katsuwonus pelamis) au voisinage d un dispositif concentrateur de poissons. Océanogr. Trop. 21, 167 183. Cillauren, E., 1994. Daily fluctuations in the presence of Thunnus albacares and Katsuwonus pelamis around fish aggregating devices anchored in Vanuatu. Oceania Bull. Mar. Sci. 55, 581 591. Depoutot, C., 1987. Contribution à l étude des dispositifs de concentration de poissons à partir de l expérience polynésienne, Orstom Tahiti Notes Doc. Océanogr. 33. Holland, K.N., Brill, R.W., Chang, R.K.C., 1990. Horizontal and vertical movements of yellowfin and bigeye tuna associated with fish aggregating devices. Fish. Bull. 88, 493 507. Josse, E., 1992. Different ways of exploiting tuna associated with fish aggregating devices in French Polynesia. In: International workshop on the ecology and fisheries for tuna associated with floating objects and on the assessment issues arising from the association of tuna with floating objets. Inter-Am. Trop. Tuna Comm., La Jolla, California. Josse, E., Bertrand, A., in press. in situ acoustic target strength measurements of tuna associated with a fish aggregating device. ICES J. Mar. Sci.. Josse, E., Bach, P., Dagorn, L., 1998. Simultaneous observations of tuna movements and their prey by sonic tracking and acoustic surveys. In: Lagardère, J.P., Bégout-Anras, M.L., Claireaux, G. (Eds.), Advances in Invertebrates and Fish Telemetry. Hydrobiologia 371/372, pp. 61 69. Josse, E., Bertrand, A., Dagorn, L., 1999. An acoustic approach to study tuna aggregated around fish aggregating devices in French Polynesia: methods and validation. Aquat. Living Resour. 12, 303 313. Kieser, R., Ehrenberg, J.E., 1990. An unbiased, stochastic echo-counting model. Rapp. P-v Réun. Cons. Int. Explor. Mer 189, 65 72. Kleiber, P., Hampton, J., 1994. Modeling effects of FADs and islands on movement of skipjack tuna (Katsuwonus pelamis): estimating parameters from tagging data. Can. J. Fish. Aquat. Sci. 51, 2642 2653. Marsac, F., Cayré, P., 1998. Telemetry applied to behaviour analysis of yellowfin tuna (Thunnus albacares Bonnaterre, 1788) movements in a network of fish aggregating devices. In: Lagardère, J.P., Bégout-Anras, M.L., Claireaux, G. (Eds.), Advances in Invertebrates and Fish Telemetry. Hydrobiologia 371/372, pp. 155 171. Marsac, F., Cayré, P., Conand, F., 1996. Analysis of small scale movements of yellowfin tuna around fish aggregating devices (FADs). In: Anganuzzi, A.A., Stobberup, K.A., Webb, N.J. (Eds.), Proc. Expert Consultation on Indian Ocean tunas, 6th session, Colombo, Sri Lanka, 25 29 Sept.1995, Indo-Pacific Tuna Development and Management Programme (IPTP), Collective, Vol. 9, pp. 151 159. Misund, O.A., 1997. Underwater acoustics in marine fisheries and fisheries research. Rev. Fish Biol. Fish. 7, 1 34. Moarii, G., Leproux, F., 1996. The drop-stone technique used by artisanal fishermen in French Polynesia. SPC Fish Aggregating Device Inform. Bull. 1, 16 18. Parrish, J.K., 1989. Layering with depth in a heterospecific fish aggregation. Environ. Biol. Fish. 26, 79 85. Simrad, SIMRAD EK500 1993. Scientific echo sounder operator manual, Simrad Subsea A/S Horten, Norway. Simrad, SIMRAD EP500 1994. Echo processing system, Simrad Subsea A/S Horten, Norway.

Aquat. Living Resour. 13 (2000) 181 2000 Ifremer/CNRS/INRA/IRD/Cemagref/Éditions scientifiques et médicales Elsevier SAS. All rights reserved S0990744000010676/EDI Editorial Mechanisms and effects of the aggregation of tuna by Fish Aggregating Devices (FADs) The associative behaviour of large pelagic fish, tuna and related species, is strongly influenced by the presence of fixed or drifting, natural or artificial floating objects such as FADs. The number of FAD for the development or maintenance of tuna fisheries, both oceanic industrial and small coastal fisheries, is increasing in such a way that at present one quarter of the worldwide production of tuna (one out of four million tons produced annually) is dependent upon the use of FAD. This evolution is, in some respects, worrying and has justified holding an international meeting on the theme Tuna fishing and FAD. This symposium took place in October 1999 at La Martinique (French West Indies). It follows an international working group organized by the IATTC in Jolla (California) in 1996 (Scott et al., 1999). The papers presented during the symposium are published in two different publications: Proceedings Les Actes de Colloques (Le Gall et al., 2000) bring together regional summaries (Pacific, Atlantic and Indian oceans and Mediterranean sea), questions relating to the technology of FAD and associated fishing techniques, the behaviour of tuna towards fishing detection equipment, the impact of FAD on the behaviour of fishermen; This special issue of the journal Aquatic Living Resources: Mechanisms and effects of the aggregation of tuna by Fish Aggregating Devices (FADs), attempts to provide some understanding of the determinism of the association between attracting objects and large pelagic fish, to propose a model of tuna movement, to describe the preferential distribution of species and size of aggregated fish, to compare the specific duration of residence in the different types of aggregation (FAD, sea mounts, etc.). The need to conduct research programs on an international level in order to progress in the understanding of this phenomenon has been emphasized. What is the role of FAD in the oceanic environment: refuge from predators, concentration of food, signs of rich environment, points of spatial reference and orientation support, meeting point (Fréon and Dagorn, 2000)? Amongst the hypotheses developed to explain the aggregation of pelagic fish on FAD, some have been the subject of presentations and discussions concerning the experimental designs already implemented or which will serve as the basis for future scientific investigations. The main research presented in this special issue deals with: the changes in associative behaviour of tuna in relation to FAD through the joint use of underwater acoustic prospecting (echo sounders, sonar) and detecting by ultrasonic devices; the analysis and model building of movement and migration surrounding FAD networks and of exchanges between offshore areas and coastal fisheries; the effects of FADs on the ecology of tuna, and in particular the alteration of their place in the food chain system caused by FADs; the impact of changes in the means of operation on production in specific tuna fisheries in terms of size levels and species diversity; the organisation and emergence of industrial and local networks, and their effect on the marketing of fish products. Jean-Yves Le Gall (Énsar, Rennes, France) Patrice Cayré (IRD, Paris, France) Marc Taquet (Ifremer, La Réunion, France) References Fréon, P., Dagorn, L., 2000. Review of fish associative behaviour: toward a generalisation of the meeting point hypothesis. Rev. Fish Biol. Fish. (sous presse). Le Gall, J.-Y., Cayré, P., Taquet, M. (Eds.), 2000. Pêche thonière et dispositifs de concentration de poissons. Actes Colloq. Ifremer, 28. Scott, M.D., Bayliff, W.H., Lennert-Cody, C., Schaefer, K., 1999. Proceedings of the international workshop on the ecology and fisheries for tunas associated with floating objects. Inter-Am. Trop. Tuna Comm., Spec. Rep. 11.

Aquat. Living Resour. 13 (2000) 182 2000 Ifremer/CNRS/INRA/IRD/Cemagref/Éditions scientifiques et médicales Elsevier SAS. All rights reserved S0990744000010846/EDI Editorial Méchanismes et incidences de l agrégation des thons par les dispositifs de concentration de poissons Le comportement associatif des grands poissons pélagiques, thons et espèces proches, est fortement influencé par la présence d objets flottants fixes ou dérivants, naturels ou artificiels tels les dispositifs de concentration de poissons (DCP). L accroissement du nombre des DCP pour le développement ou le maintien des pêcheries thonières industrielles océaniques et artisanales côtières est tel qu actuellement, un quart de la production mondiale de thons (soit un million sur les quatre millions de tonnes produites annuellement) est dépendant de l utilisation des DCP. Cette évolution est, à certains égards, inquiétante et a justifié la tenue d une réunion internationale sur le thème «Pêche thonière et DCP». Ce colloque s est déroulé en octobre 1999 à La Martinique (Antilles françaises). Il a fait suite à un groupe de travail international organisé par la Commission inter-américaine des thons tropicaux (IATTC) à La Jolla (Californie) en 1996 (Scott et al., 1999). Les communications présentées lors de ce colloque font l objet de deux publications distinctes : les Actes de Colloques (Le Gall et al., 2000) regroupent des synthèses régionales (océans Pacifique, Atlantique, Indien et mer Méditerranée), les questions de technologie des DCP et de techniques de pêche associées, le comportement des thons vis-à-vis des engins de pêche, l impact des DCP sur les ressources exploitées, ainsi que la modification induite par le développement des DCP sur le comportement des pêcheurs ; ce numéro thématique de la revue Aquatic Living Resources tente d apporter des éléments de compréhension du déterminisme de l association entre objets attracteurs et grands pélagiques, de proposer une modélisation des déplacements des thons, de décrire la répartition préférentielle des espèces et des tailles des poissons agrégés, de comparer les durées de résidence spécifiques au sein de différents types d agrégations (DCP, monts sous-marins), etc. La nécessité de conduire des programmes de recherche à l échelle internationale afin de progresser dans la compréhension de ce phénomène a été soulignée. Quel est le rôle des DCP dans le milieu océanique : refuge contre les prédateurs, concentration de nourriture, indicateurs de richesse du milieu, points de référence spatiale et supports d orientation, points de rencontre (Fréon et Dagorn, 2000)? Parmi les hypothèses élaborées pour expliquer l agrégation des poissons pélagiques sur les DCP, certaines ont fait l objet de présentations et de discussions concernant les protocoles expérimentaux déjà mis en œuvre ou qui serviront de base aux prochaines investigations scientifiques. Les principaux travaux présentés dans ce numéro thématique portent sur la modification du comportement associatif des thons en relation avec les DCP par : l utilisation conjointe de la prospection acoustique sous-marine (sondeurs, sonars) et du suivi par marquage ultrasonique ; l analyse et la modélisation des déplacements et des migrations autour des réseaux de DCP ainsi que des échanges entre les zones du large et les pêcheries côtières ; l incidence sur l écologie des thons, et en particulier, leur place dans le réseau trophique modifié par les DCP ; l impact de la modification des modes d exploitation sur la production des pêcheries thonières spécifiques en termes d échelles de tailles et de diversité d espèce ; l organisation et l émergence de filières industrielles et artisanales, jusqu à la commercialisation des produits de la pêche. Jean-Yves Le Gall (Énsar, Rennes, France) Patrice Cayré (IRD, Paris, France) Marc Taquet (Ifremer, La Réunion, France) Références Fréon, P., Dagorn, L., 2000. Review of fish associative behaviour: toward a generalisation of the meeting point hypothesis. Rev. Fish Biol. Fish. (sous presse). Le Gall, J.-Y., Cayré, P., Taquet, M. (Eds.), 2000. Pêche thonière et dispositifs de concentration de poissons. Actes Colloq. Ifremer, 28. Scott, M.D., Bayliff, W.H., Lennert-Cody, C., Schaefer, K., 1999. Proceedings of the international workshop on the ecology and fisheries for tunas associated with floating objects. Inter-Am. Trop. Tuna Comm., Spec. Rep. 11.

Aquat. Living Resour. 13 (2000) 233 240 2000 Ifremer/CNRS/INRA/IRD/Cemagref/Éditions scientifiques et médicales Elsevier SAS. All rights reserved S0990744000010664/FLA Food consumption of tuna in the Equatorial Atlantic ocean: FAD-associated versus unassociated schools Frédéric Ménard ad *, Bernard Stéquert a, Alex Rubin a, Miguel Herrera c, Émile Marchal b a IRD, Centre de recherches océanologiques, BP V 18 Abidjan, Ivory Coast b IRD, Institut océanographique, 195 rue Saint-Jacques, 75005 Paris, France c IOTC, PO Box 1011 Victoria, Mahe, Seychelles d present address: IRD, Dep. HEA, BP 5045, 34032 Montpellier cedex 1, France Accepted 3 July 2000 Abstract Since 1991, fishing operations on tuna schools associated with drifting Fish Aggregating Devices (FADs) have become widespread in the purse seine fishery in the Gulf of Guinea. In the offshore South Sherbro area (0 5 N, 10 20 W), FAD-associated catches represent about 75 % of the total catch. This FAD fishery exploits concentrations of skipjack mixed with a smaller amount of bigeye and yellowfin tuna of similar size (46 cm), and some large yellowfin. Catches on unassociated tuna schools are mainly composed of large yellowfin in breeding phase and skipjack. Here we studied tuna diet in relation with the aggregation mode (FAD-associated or unassociated tuna schools), species, and size. The stomach contents of around 800 fish were analysed. Numerous empty stomachs were found, especially in fish caught under FADs. Diets were similar for all small-size tuna sharing the same aggregation type. Small tuna mainly feed on Vinciguerria nimbaria (Photichthyidae), a mesopelagic fish of the micronekton, whereas large tuna mainly feed on Scombridae, mixed with Cubiceps pauciradiatus (Nomeidae) when they were caught in unassociated schools. The feeding habits of tuna are discussed with emphasis on the behavior of V. nimbaria. Estimations of the daily ration of similarly sized tuna with the same aggregation mode were very close. The low estimated rations for small, FAD-associated tuna show that logs do not have a trophic function, but rather are a refuge. In contrast, FADs seem to influence the diet of large tuna because of the Scombridae prey that probably is associated to the FAD. 2000 Ifremer/CNRS/INRA/IRD/Cemagref/Éditions scientifiques et médicales Elsevier SAS tuna feeding / daily ration / Fish Aggregating Devices / Vinciguerria nimbaria / purse seine / Atlantic Ocean Résumé Alimentation des thons dans l Atlantique équatorial : comparaison sous DCP et en bancs libres. Depuis 1991, la pêche sous des Dispositifs de Concentration de Poissons (DCP) dérivants mis à l eau par les senneurs s est très fortement développée dans le golfe de Guinée. Dans la zone hauturière Sud Sherbro (0 5 N, 10 20 W), elle représente 75 % des captures totales. Cette pêche sous DCP exploite des concentrations plurispécifiques constituées surtout de listaos, et d albacores et de patudos de taille similaire (46 cm), associés à quelques grands albacores. Les captures sur bancs libres sont constituées de gros albacores en reproduction et aussi de listaos. On compare ici l alimentation des thons dans cette zone selon le type d agrégation (bancs libres et DCP), l espèce et la taille. Près de 800 contenus stomacaux ont été analysés. Les estomacs vides sont très nombreux, surtout sous les DCP. Le régime alimentaire varie suivant le type d agrégation. Les petits thons se nourrissent essentiellement de Vinciguerria nimbaria (Photichthyidae), un poisson mésopélagique du micronekton dont le comportement particulier dans la zone est discuté. Les Scombridae sont les proies principales des grands thons sous DCP dérivants, alors qu en bancs libres, ils se nourrissent aussi de Cubiceps pauciradiatus (Nomeidae). Les rations journalières des thons de même taille et capturés de la même façon ne montrent pas de différences importantes entre espèces. Les faibles rations estimées pour les petits thons sous DCP montrent que les épaves ne jouent pas un rôle alimentaire mais représentent plutôt un refuge. En revanche, les DCP semblent avoir un rôle dans l alimentation des grands thons qui se nourrissent de petits Scombridae associés eux-mêmes aux DCP. 2000 Ifremer/CNRS/INRA/IRD/Cemagref/Éditions scientifiques et médicales Elsevier SAS alimentation des thons / ration journalière / Dispositifs de Concentration de Poissons / Vinciguerria nimbaria / pêche à la senne / océan Atlantique *Corresponding author. E-mail address: menard@ird.fr (Frédéric Ménard).

234 F. Ménard et al. / Aquat. Living Resour. 13 (2000) 233 240 1. INTRODUCTION In the Eastern Tropical Atlantic, tropical tuna, mainly yellowfin (Thunnus albacares), skipjack (Katsuwonus pelamis), and bigeye (Thunnus obesus), are exploited by surface purse seine fishery covering the whole area. Two main fishing modes are used: the search and the catch of free swimming tuna schools (mostly big-size tuna in mono-specific unassociated schools), and sets under artificial floating objects (drifting Fish Aggregating Devices, FADs), set afloat by the purse seiners. FADs concentrate tuna (mostly mixed concentrations of small-size tuna dominated by skipjack), but also other pelagic species associated with FADs: Acanthocybium solandri, billfish, Balistidae, Sphyranea barracuda, Coryphaena sp., Elagatis bipinnulata, Kyphosidae and sharks (Ménard et al., 2000). In the early 1990s, fishing operations on tuna schools associated with drifting FADs became widespread in the Eastern Tropical Atlantic (Ariz et al., 1993). At the moment catches using this fishing mode reach more than 50 % of the total surface catch (ICCAT data base). Very few studies have undertaken the analysis of the impact of FADs on the feeding of tuna. Hunter and Mitchell (1967) studied the stomach contents of skipjack and yellowfin tuna associated with drifting logs (natural FADs) in the offshore waters of Costa Rica, and they showed that tuna do not feed on fish associated with the log. Other studies focus on the diet of yellowfin tuna associated with anchored FADs in the Hawaiian waters (Brock, 1985) and in the American Samoa (Buckley and Miller, 1994), or associated with anchored payaos in the Philippines (Yesaki, 1983; Barut, 1988). No link has been demonstrated between feeding behavior of yellowfin tuna and anchored FADs, except in the work of Yesaki (1983), who found cannibalism of large yellowfin tuna on small-size conspecifics associated with the anchored FADs. The purpose of this paper is to study the tuna diet in relation with the aggregation type (unassociated tuna schools versus tuna concentrations around drifting FADs) in the South Sherbro area (0 5 N and 10 20 W) in the Equatorial Atlantic. The South Sherbro area developed into a major seasonal FAD fishing zone in the Eastern Tropical Atlantic (Ménard et al., 1999, Ménard et al., 2000) with high catches from November to January. During these three months, sets under FADs represented about 75 % of the total catch, and exploited mixed concentrations of small-size tuna (46 cm fork length), and some large yellowfin and bigeye tuna (less than 10 % of the catches in weight). Sets of unassociated tuna schools mainly occurred from February to April and targeted large-size yellowfin tuna, at this period in breeding condition, in the Gulf of Guinea (Albaret, 1977; Bard and Capisano, 1991). The influence of drifting FADs on the diet of tuna is investigated by comparing the stomach contents of tuna caught in the South Sherbro area, taken into account the type of association, the Figure 1. Sampling positions from which stomachs were collected in 1995, 1996, 1997 and 1998 in or near the South Sherbro area. Area of each circle is proportional to the number of stomachs. species, and the size. Daily rations are estimated following the methodology of Olson and Boggs (1986). Feeding habits of tuna are discussed with emphasis on the behavior of their main prey, Vinciguerria nimbaria, a mesopelagic fish of the micronekton. 2. MATERIALS AND METHODS Stomachs were collected in 1995, 1996, 1997, and 1998 from tuna caught during daylight hours in or near the South Sherbro area (figure 1). The samples were collected by scientific observers on board vessels, by scientific technicians at the port of Abidjan during landings, or at tuna canneries in Abidjan. In order to study the influence of the size of the fish, the data was grouped in two size classes (smaller than 90 cm and bigger than 90 cm) that represent two distinct groups in the size catch distribution. Most of the fish were smaller than 90 cm, so only a few yellowfin tuna with a size greater than 90 cm have had their stomach collected and analysed. The length of each fish was measured and the stomach was either preserved in formalin or deep frozen. Stomachs were weighed, food was removed, and then contents were weighed and sorted according to the digested prey and its degree of digestion. The prey was identified to the lowest possible taxon. Four degrees of digestion were used according to the state of the prey. The degree of digestion four, the most advanced state of digestion, occurred in 60 % of our analyses. Recognizable prey items were divided into six major categories for the fish that constituted the dominant phylum: Vinciguerria nimbaria (Photichthyidae), Scombridae (represented by Auxis thazard and Katsuwonus pelamis), Cubiceps pauciradiatus (Nomeidae), epipelagic fish (Balistidae, Clupeidae, Diretmidae and Exocetidae), mesopelagic fish (Astronesthidae, Gonostomatidae, Myctophidae, Nemichthyidae, Paralepididae and Gempylidae), and undetermined fish when the digestion state was too advanced. Planktonic crustaceans (mainly Euphausiaceae and

F. Ménard et al. / Aquat. Living Resour. 13 (2000) 233 240 235 natantia Decapods) constituted a separate category, as was the case for the Cephalopods (exclusively squids of the Teuthoidae family). The latter category was grouped with the remainder preys (undetermined pulp, tunicates such as Salpidae). The stomachs that contained only mucus or parasites were considered as empty. To take into account the multivariate nature of the stomach contents data, we performed a correspondence analysis on the data matrix of the individual stomach contents of the six small tuna predators (skipjack, bigeye and yellowfin smaller than 90 cm, unassociated and FAD-associated), and on eight of the prey items. Scombridae was eliminated as an item because it was observed only once in the diet of a FAD-associated bigeye. Empty stomachs were also eliminated. Then a simple binary designation of presence or absence of each prey item was used in order to extract patterns from this data set. In order to estimate daily rations, the process of gastric retention needed to be taken into account, i.e. the relationship between the amount of food remaining in the stomach and the time elapsed since the feeding. Such a process depends on the type of prey and on the prey sizes, and several models have been proposed; the most widely used assumes that gastric evacuation is exponential (Elliott and Persson, 1978). Olson and Mullen (1986) and Olson and Boggs (1986), in their very comprehensive study on the feeding of the yellowfin tuna of the Eastern Tropical Pacific, have proposed a food consumption model that does not require an a priori assumption of exponential gastric evacuation. Data from gastric evacuation experiments were fitted best by linear regression models for the different prey used to feed experimental tuna (mackerel Scomber japonicus, squid Loligo opalescens, smelt Hypomesus pretiosus, and nehu Stolephorus purpureus). The authors proposed to calculate the daily ration from the estimation of a feeding rate r (g h -1 ), computed as: W i rˆ = i A i with A i (h), the integral of the function that fits the gastric evacuation rate (i.e. the average amount of time required to evacuate the average proportion of ingested prey of type i present in the stomach at any time); W i (g), the average wet weight of prey of type i found in the stomach of the predator. In the W i calculus, empty stomachs are taken into account because, as a rule, they reflect the natural feeding conditions of the population. According to Maldeniya (1996) and due to the lack of experimental data for the gastric evacuation rates for the identified prey of the tuna in the South Sherbro area, we used the A i values as they were estimated by Olson and Boggs (1986) with respect to the ecology of the species (table I). Assuming that tuna mainly feed during the daylight hours, the daily food intake is calculated by multiplying Wrˆ by 12 h. Daily rations correspond to the daily food intake expressed as a percentage of the average weight of the corresponding tuna predator. The weight of the stomach contents can be adjusted in order to take into account the stress due to the fishing. During the fishing operation, digestion continues until the death of the fish or until the preservation of the stomach (stress may also lead to a partial or total regurgitation of the stomach). Following Olson and Boggs (1986), we used an average duration of gear retention (2 h) in conjunction with gastric evacuation rates to calculate such an adjustment of the weight of the stomach contents at the time of gut removal (table I). Furthermore, to take into account the wet mass loss of the food in the stomachs observed within a few minutes after eating (Magnuson, 1969), rˆ values were adjusted. Olson and Boggs (1986) interpreted the differences between the 100 %-value and the estimated intercept values of the gastric evacuation in linear models (table I) as a quick water loss. Thus, for each food item rˆ values were increased by these differences. 3. RESULTS The number of empty stomachs (85 %) was very high among tuna caught under FADs (table II). This was not the case for unassociated tuna, where the proportion of empty stomachs reached 25 %. Figure 2 displays the status of the stomachs versus the time of the day. Under FADs, most of the tuna were caught before 8h00 in the morning, i.e. at dawn (7h00 at 15 W in the South Sherbro area), but the number of empty stomachs remained high, independent of the time. Non-empty stomachs were sampled before 08h00, 33 for a total of 305 sampled from 26 fishing sets. The composition of the prey and its degree of digestion varied greatly. The prey in the fullest stomach (around 1 kg) was a Scombridae recently eaten (degree of digestion: one) by a large-size yellowfin tuna. For unassociated tuna schools, non-empty stomachs of three bigeye and nine yellowfin tuna coming from the same fishing set occurring at 7h30, were analysed. The fish had fed almost only on Vinciguerria nimbaria and the degree of digestion ranged from one to four. The average weights of the stomach contents were expressed as a percentage of the body mass of the predator (table III). The mass adjustments for gear retention led to an increase in the mean mass of the observed stomach contents between 14.8 and 23.5 %. Under the FADs, our estimations were very close for skipjack, small-size yellowfin (smaller than 90 cm) and bigeye tuna. Values of the stomach content mass were higher, but showed a great variability, for smallsize unassociated tuna. Furthermore, the stomach content (% of body mass) decreased for increasing fish

236 F. Ménard et al. / Aquat. Living Resour. 13 (2000) 233 240 Table I. Parameter values used in the estimations of the daily rations and for the adjustments of the stomach contents*. Prey category A i Intercept Slope Experimental food species Vinciguerria nimbaria 2.24 0.727 0.0693 mackerel Cubiceps pauciradiatus Mesopelagic fish Epipelagic fish Undetermined fish Scombridae 5.29 0.856 0.1182 nehu Cephalopods 4.48 0.847 0.0800 squid Crustacea 3.77 0.805 0.0859 mean of mackerel, squid, smelt, nehu Other prey * Intercepts and slopes are the linear regression parameters of the proportion of food remaining as a function of time (h) passed since feeding; from Olson and Boggs (1986). sizes (from 2.42 % to 0.47 %). Such a tendency was not found for yellowfin tuna caught under FADs. Figure 3 displays the results of the correspondence analysis performed on the individual stomachs of the small-size tuna and their prey items. The first two axes explain 40 % of the inertia (first axis 22 %, second axis 18 %). The observed pattern shows that the second axis distinguishes tuna caught under FADs from unassociated tuna. Thus, the diet in the study area appears to be similar for all small-size tuna from the same aggregation type. Vinciguerria nimbaria and crustaceans are characteristic prey of unassociated tuna, whereas the intake of epipelagic fish and undetermined fish characterizes FAD-associated tuna. Afterwards all individual stomachs were grouped by tuna predator. Figure 4 displays the mass proportions of each prey item. Fish, including undetermined fish species, were the main prey. Small-size tuna (smaller than 90 cm) mainly fed on Vinciguerria nimbaria, supplemented with cephalopods for tuna caught under FADs. Their diet showed a greater variability than for small tuna caught in unassociated schools. Cubiceps pauciradiatus and Scombridae formed the most important prey for large-size yellowfin tuna of unassociated schools (though with 65 % undetermined fish species), whereas large yellowfin under FADs almost exclusively fed on Scombridae (frigate tuna and skipjack). When samples coming from the same set of unassociated tuna schools were analysed together, we observed a high homogeneity in the prey and in the indices of digestion (degree of digestion). Under FADs, stomach contents rarely contained the same category of prey with the same degree of digestion, and for a given school, empty stomachs and more or less full stomachs were observed simultaneously. Daily rations were estimated to be approximately 1.17 % of the body mass for the small-size tuna caught under FADs, and the ration was about 6.12 % for the small tuna caught in unassociated schools (table IV). Table II. Number of tuna stomachs (N) sampled in the South Sherbro area and percentages of empty stomachs. FADs Unassociated schools Yellowfin N 69 36 % empty 65.2 16.7 Skipjack N 333 115 % empty 91 27 Bigeye N 191 32 % empty 82.7 25 Total N 593 183 % empty 85.3 24.6 Figure 2. Tuna stomach samples versus time of the day with the proportion of empty stomachs.

F. Ménard et al. / Aquat. Living Resour. 13 (2000) 233 240 237 Table III. Observed and adjusted stomach contents (% of body mass) and body mass (g) of tuna from which stomach samples were taken in the South Sherbro area. Body mass (g) Observed Stomach contents Adjusted stomach contents Min. Max. Mean Min. Max. Mean Min. Max. Mean FADs skj 790 8 224 2 046 0 7.82 0.19 0 9.68 0.23 bet 796 11 530 3 142 0 6.86 0.22 0 8.30 0.27 yft < 90 cm 812 12 735 3 137 0 2.97 0.13 0 3.68 0.16 yft > 90 cm 15 040 81 003 46 167 0 4.66 0.85 0 5.32 0.98 Unassociated schools skj 850 15 220 2 852 0 6.04 1.12 0 7.51 1.36 bet 1 101 8 119 2 772 0 3.20 0.57 0 3.98 0.70 yft < 90 cm 1 564 2 753 2 266 0 4.04 1.96 0 5.00 2.42 yft > 90 cm 30 365 90 262 59 186 0 1.67 0.39 0 2.05 0.47 skj: skipjack; bet: Bigeye; yft: yellowfin tuna. The latter estimation included the high value of 16.03 % computed for yellowfin tuna with a size smaller than 90 cm, that was probably overestimated because of the lack of samples for this tuna category (table II). Nevertheless, the daily ration was about five times higher than the one computed for small tuna under FADs. Figure 5 displays the mass proportions of adjusted stomach contents and of daily food intake, for the small-size tuna grouped by aggregation type. For large yellowfin tuna, the ration was slightly greater under FADs (3.04 %) than in unassociated schools (2.59 %). 4. DISCUSSION Vinciguerria nimbaria is a major chain in the local food web of the South Sherbro area, as was previously suggested by Roger and Marchal (1994). This mesopelagic fish of the micronekton has already been reported to be the most important forage item of skipjack and yellowfin tuna (Alverson, 1963; Legand et al., 1972; Dragovich and Potthof, 1972; Kornilova, 1981; Borodulina, 1982), but also of Bryde s whales (Kawamura and Hamaoka, 1981), and of spinner porpoise Stenella longirostris (Perrin et al., 1973). In the South Sherbro area, V. nimbaria represented 63 % of the daily food intake of the small tuna caught in unassociated schools, and 49 % for the small tuna under FADs (figure 5). Because tuna are generally considered to be opportunistic feeders, V. nimbaria must be abundant and accessible to tuna. This fish is often seen as a typical mesopelagic species, diving to depths of 500 m or more during the day, and migrating to the 0 90 m layer at night. Marchal and Lebourges Figure 3. Correspondence analysis: projections on the factorial plane (two first axis with 22 and 18 % of the inertia, respectively) of the eight prey items and of the gravity centers of the six categories of small-size tuna predators; vinci: Vinciguerria nimbaria; cubi: Cubiceps pauciradiatus; meso: Mesopelagic fish; epi: Epipelagic fish; und-fish: Undetermined fish; crust: Crustacean; ceph: Cephalopod; other: other prey.

238 F. Ménard et al. / Aquat. Living Resour. 13 (2000) 233 240 Figure 4. Mass proportions of prey found in skipjack (SKJ), bigeye (BET), yellowfin (YFT) stomachs, sampled in the South Sherbro area. (1996) reported that the adult population of V. nimbaria have a peculiar diel behavior in the South Sherbro area, concentrating in mono-specific dense schools in the upper layers during the day where they become available to tuna predation. At night, they concentrate at or below the thermocline (80 m) mixed with other fish, squids and crustaceans. Lebourges- Dhaussy et al. (2000) showed that schools of V. nimbaria remained at the surface by day when a stable hydrological situation occurred with a well-mixed surface layer and a strong Deep Chlorophyll Maximum. V. nimbaria has a short life span (six to seven months) and a maximum standard length of 55 mm; instantaneous biomass was estimated by acoustic data to be around the megaton (unpublished data). This biomass, consisting of nearly-adult fish, renewed during the fishing period, must sustain the high concentrations of small tuna. The explanation for the peculiar behavior of the adults may be found in their feeding activity (Lebourges-Dhaussy et al., 2000). The tuna purse seine fishery operates only during daylight hours. We therefore did not dispose of all the samples needed for covering a full diel cycle. However the feeding of surface tuna that are visual predators, takes place mainly during the daylight hours (Schaefer et al., 1963; Legand et al., 1972; Ortega-Garcia et al., 1992; Roger, 1994; Maldeniya, 1996), and very little (or not at all) at night for small-size surface tuna. Large bigeye and yellowfin tuna have been observed to feed during the night (Josse et al., 1998). Still, they are probably more active feeders during the daytime than at night (Kobayashi and Yamaguchi, 1971). In the Indian Ocean, Roger (1994) found that all the stomachs sampled from sets under FADs before sunrise were empty. Our results suggest that almost all the tuna caught by the purse seine fishers early in the morning have not fed yet, or have fed recently around dawn. In a great number of stomachs V. nimbaria very often seemed the only prey found, clearly showing that small tuna fed on this fish while schooling. Such Table IV. Daily food intake (g) and daily rations (%) of tuna with respect to aggregation mode, species and size; skj: skipjack; bet: Bigeye, yft: yellowfin tuna. Daily food intake (g) Daily ration (%) FADs skj 23.6 1.16 bet 39.8 1.27 yft < 90 cm 27.8 0.89 yft > 90 cm 1 404.3 3.04 Unassociated schools skj 157.1 5.51 bet 133.6 4.82 yft < 90 cm 363.2 16.03 yft > 90 cm 1 530.9 2.59 FADs < 90 cm 29.3 1.17 Unassociated schools < 90 cm 170.5 6.12

F. Ménard et al. / Aquat. Living Resour. 13 (2000) 233 240 239 Figure 5. Mass proportions of adjusted stomach contents and of daily food intake, for the small-size tuna (skipjack, yellowfin, bigeye) of both aggregation types sampled in the South Sherbro area; m: mean weight (g) of the adjusted stomach contents; n: number of samples. schools of V. nimbaria were only observed at the surface during daylight, whereas V. nimbaria was mixed in the upper layers at night together with other fish, squids and crustaceans coming from deeper layers. However, acoustic observations showed that they also concentrated in wide loose aggregations at night, most of the time inside or just below the layer. It is thus difficult to know whether they are mixed or not with other species. Another point is the role of the photophores, that fish such as V. nimbaria have, in the predation process: do they attract predators at night, or do they fight them off, especially when these fish concentrate? Until now, no direct observations have been able to answer this question. The degree of digestion may give some insights but this requires a precise measuring of the time elapsed between the beginning of the fishing operation and the preservation of the stomach. In our analysis, we thus chose a feeding period of 12 h, but we do not want to conclude that none of the tuna feed at night, because this may have been an artefact of the sampling. Tuna are considered to have high food requirements due to their very high metabolic rates (Olson and Boggs, 1986; Dickson, 1995). However, few studies have undertaken analyses to estimate the food consumption of tuna from data about stomach contents sampled in the field. The order of magnitude of the tuna s food intake in our results is similar to the one estimated by Olson and Boggs (1986) for yellowfin tuna of the Eastern Tropical Pacific Ocean (from 3 to 5 % of the fish s body mass, depending of the size of the fish, but assuming a daily feeding period of 24 hours), and to those computed by Maldeniya (1996) for yellowfin in Sri Lanka waters (from 2.1 to 5.5 %). Such estimates strongly depend on the model chosen for food consumption and gastric evacuation that are similar in these latter approaches. Young et al. (1997), using an exponential rate of gastric evacuation (after Elliott and Persson, 1978 and Boisclair and Marchand, 1993), found an overall estimate of 1 % per day for the southern bluefin tuna (Thunnus maccoyii), although these authors indicated that daily ration estimates of farmed fish could reach 7 %. Seven percent of the weight of the fish was also the maximum capacity of the stomach of captive skipjack tuna, (Magnuson, 1969), but during an entire day they ate around 15 % of their body weight. Assuming that the feeding activity is associated with the amount of food found in the stomach, the proportion of empty stomachs (85 % versus 25 % for unassociated schools) and the low daily ration estimates for the small-size tuna under FADs indicate that these fish do not feed under drifting FADs. Indeed, the biomass of the potential prey under the FADs, available for small-size tuna, is low. The tuna have to leave their FAD during the day and may form free swimming loose schools in order to feed actively on Vinciguerria nimbaria (not associated with the FAD). They may come back under floating objects they use as a refuge or meeting point (for review see Fréon and Misund, 1999). The large-size yellowfin tuna caught under FADs mainly feed on Scombridae themselves associated with the FAD and constituting a high biomass available for predation (Yesaki, 1983). It seems that FADs have a refuge function for small tuna, and a trophic function for large tuna, and probably for other associated pelagic species such as billfish and sharks. References Albaret, J.J., 1977. La reproduction de l albacore (Thunnus albacares) dans le golfe de Guinée. Cah. Orstom sér. Océanogr. 15, 389 419. Alverson, F.G., 1963. The food of yellowfin and skipjack tuna in the Eastem Tropical Pacific Ocean. Bull. Inter- Amer. Trop. Tuna Comm. 7, 293 396. Ariz, J., Delgado, A., Fonteneau, A., Gonzales Costas, F., Pallares, P., 1993. Logs and tuna in the Eastern Tropical Atlantic. A review of present knowledge and uncertainties. ICCAT Coll. Sci.Vol. Pap. 40, 421 446. Bard, F.X., Capisano, C., 1991. Actualisation des connaissances sur la reproduction de l albacore (Thunnus albacares) en océan Atlantique. ICCAT Coll. Vol. Sci. Pap. 36, 158 181.

240 F. Ménard et al. / Aquat. Living Resour. 13 (2000) 233 240 Barut, N.C., 1988. Food and feeding habits of yellowfin tuna Thunnus albacares (Bonnaterre, 1788), caught by handline around payao in the Moro Gulf. FAO, Indo-Pacific Tuna Development and Management Programme, IPTP/ 88/WP/18. Boisclair, D., Marchand, F., 1993. The guts to estimate daily ration. Can. J. Fish. Aquat. Sci. 50 1969-1975. Borodulina, O.D., 1982. Food composition of yellowfin tuna Thunnus albacares. J. Ichthyol. 21, 38 46. Brock, R.E., 1985. Preliminary study of the feeding habits of pelagic fish around Hawaiian fish aggregating devices, or can fish aggregation devices enhance local fish productivity? Bull. Mar. Sci. 37, 40 49. Buckley, T.W., Miller, B.S., 1994. Feeding habits of yellowfin tuna associated with fish aggregation devices in American Samoa. Bull. Mar. Sci. 55, 445 459. Dickson, K.A., 1995. Unique adaptations of the metabolic biochemistry of tuna and billfishes for life in the pelagic environment. Env. Biol. Fish. 42, 65 97. Dragovich, A., Potthoff, T., 1972. Comparative study of food of skipjack and yellowfin tuna off the coast of west Africa. Fish. Bull. 70, 1087 1110. Elliott, J.M., Persson, L., 1978. The estimation of daily rates of food consumption for fish. J. Anim. Ecol. 47, 977 991. Fréon, P., Misund, O., 1999. Dynamics of pelagic fish distribution and behaviour: effects on fisheries and stock assessment. Fishing News Books, Blackwell. Hunter, J.R., Mitchell, C.T., 1967. Association of fishes with flotsam in the offshore waters of Central America. Fish. Bull. 66, 13 29. Josse, E., Bach, P., Dagorn, L., 1998. Simultaneous observations of tuna movements and their prey by sonic tracking and acoustic surveys. Hydrobiologia 371/372, 61 69. Kawamura, A., Hamaoka, S., 1981. Feeding habits of the Gonostomatid fish, Vinciguerria nimbaria collected from the stomach of Bryde s whales in the Southern North Pacific. Bull. Plankton Soc. Jap. 28, 141 151. Kobayashi, H., Yamaguchi, Y., 1971. Feeding ecology and hooking tendency of tuna and marlins in the Eastern Equatorial Pacific. Bull. Jpn. Soc. Sci. Fish. 37, 83 89. Kornilova, G.N., 1981. Feeding of yellowfin tuna, Thunnus albacares, and bigeye tuna, Thunnus obesus, in the equatorial zone of the Indian Ocean. J. Ichthyol. 20, 111 119. Lebourges-Dhaussy, A., Marchal, E., Menkès, C., Champalbert, G., Biessy, B., 2000. Vinciguerria nimbaria (micronekton), environment and tuna: their relationships in the Eastern Tropical Atlantic (ETA area). Oceanol. Acta 23, 515 528. Legand, M., Bourret, P., Fourmanoir, P., Grandperrin, R., Guérédrat, J.A., Michel, A., Rancurel, P., Repelin, R., Roger, C., 1972. Relations trophiques et distributions verticales en milieu pélagique dans l océan Pacifique intertropical. Cah. Orstom sér. Océanogr. 10, 303 393. Magnuson, J.J., 1969. Digestion and food consumption by skipjack tuna (Katsuwonus pelamis). Trans. Am. Fish. Soc. 98, 379 392. Maldeniya, R., 1996. Food consumption of yellowfin tuna, Thunnus albacares, in Sri Lankan waters. Env. Biol. Fish. 47, 101 107. Marchal, E., Lebourges, A., 1996. Acoustic evidence for unusual diel behaviour of mesopelagic fish (Vinciguerria nimbaria) exploited by tuna. ICES J. Mar. Sci. 53, 443 447. Ménard, F., Hervé, A., Fonteneau, A., 1999. An area of high seasonal concentrations of tuna: 2-4 N 10-20 W. The site of the PICOLO programme. In: Beckett J. (Ed.) Proc. ICCAT Tuna Symp. Part 1, 405-420. Ménard, F., Fonteneau, A., Gaertner, D., Nordstrom, V., Stéquert, B., Marchal, E., 2000. Exploitation of small tuna by a purse seine fishery with fish aggregating devices and their feeding ecology in an Eastern Tropical Atlantic ecosystem. ICES J. Mar. Sci. 57, 525 530. Olson, R.J., Boggs, C.H., 1986. Apex predation by yellowfin tuna (Thunnus albacares): independent estimates from gastric evacuation and stomach contents, bioenergetics, and Cesium concentrations. Can. J. Fish. Aquat. Sci. 43, 1760 1775. Olson, R.J., Mullen, A.J., 1986. Recent developments for making gastric evacuation and daily ration determinations. Env. Biol. Fish. 16, 183 191. Ortega-Garcia, S., Galvan-Mangana, F., Arvizu-Martinez, J., 1992. Activity of the Mexican purse seine fleet and the feeding habits of yellowfin tuna. Cienc. Mar. 18, 139 149. Perrin, W.F., Warner, R.R., Fiscus, C.H., Holts, D.B., 1973. Stomach contents of porpoise, Stenella spp., and yellowfin tuna, Thunnus albacares, in mixed-species aggregations. Fish. Bull. 71, 1077 1092. Roger, C., 1994. Relationships among yellowfin and skipjack tuna, their prey-fish and plankton in the tropical western Indian Ocean. Fish. Oceanogr. 3, 133 141. Roger, C., Marchal, E., 1994. Mise en évidence de conditions favorisant l abondance des albacores, Thunnus albacares, et des listaos, Katsuwonus pelamis, dans l Atlantique Equatorial Est. ICCAT Coll. Vol. Sci. Pap. 32, 237 248. Schaefer, M.B., Broadhead, G.C., Orange, C.J., 1963. Synopsis on the biology of yellowfîn tuna Thunnus (Neothunnus) albacares (Bonnaterre) 1788 (Pacific Ocean). FAO Fish. Rep. 6, 538 561. Yesaki, M., 1983. Observations on the biology of yellowfin (Thunnus albacares) and skipjack (Katsuwonus pelamis) tuna in the Philippine waters. FAO, Indo-Pacific Tuna Development and Management Programme, IPTP/83/ WP/7. Young, J.W., Lamb, T.D., Le, D., Bradford, R.W., Whitelaw, A.W., 1997. Feeding ecology and interannual variations in the diet of southern bluefin tuna, Thunnus maccoyii, in relation to coastal and oceanic waters off eastern Tasmania, Australia. Env. Biol. Fish. 50, 275 291.