In the Java sea, pelagic fish are exploited by purse seiners. The resource is dispersed and has to be aggregated before being fished. Light is used to reach this objective. The purse seiners are mobile and fish in ‘packs’. When concentrations of fish are found the fishing vessels gather in clusters. The spatial characterization of these clusters provides better understanding of the fishing strategy followed by the purse seiners. A point process approach is used to understand the clustering process and determine the life time of the clusters. The calculation of a density function defines the surface influenced by the cluster and estimates the intensity of the exploitation inside it. Clustering is the mark of the fishing phase that is defined by the significant divergence of the L-function from the Poisson process. In the fishery, such a result is observed when at least 15 purse seiners are gathered. Usually, the fishing phase lasts 4–6 days. The surface covered by the clusters represents a small part (15–25 %) of the surface that is accessible to the fleet. In these areas, the fishing intensity can reach high values (160 vessels × nautical mile–2).

Lagoon fish in New Caledonia are mainly caught by artisanal fisheries and subsistence fishing. Reef fish are the major component of this catch. The present study aimed at estimating these reef fish standing stocks and at finding the main factors influencing the distribution of these fish. Sampling of 904 stations was stratified according to three zones (north, east and west) and three reef types (barrier, intermediate and fringing). Fish communities exhibited bold heterogeneity in their distribution, showing higher biomass (maximum of 447 g·m–2) and total standing stock (43000 tonnes) in the north zone than in the east and west zones. Similarly, observed patterns were dependent on reef types: higher biomass and total standing stock being observed on barrier reefs than on intermediate or fringing reefs. The total standing stocks, which were about 65000 t, were mainly composed of herbivorous fish families such as the Acanthuridae and Scaridae. The differences in the patterns of distribution of species, individuals and standing stocks between reef types may be explained by variations in terrestrial influences and reef morphology, whereas differences among zones were most likely due to accessibility of fishing areas and fishing pressure. The latter is almost non-existent in the north zone, which can thus be considered to be almost unexploited commercially. This most likely explains the high proportion, 77 %, of long-lived species in the biomass of this zone. The results might have implications in management of reefs elsewhere in the South Pacific, for which similar data are only scarcely available.

Fish stocks found on lagoon bottoms and near reefs are characterized by a high diversity and heterogeneous habitat which make stock estimation difficult. In particular, it is necessary to combine several methods in order to evaluate the major components of these stocks. The present study aimed at estimating reef fish stocks in the Northern Province of New Caledonia, a region where they represent a major target for the local fishermen. These estimates were based on experimental fishing with handlines and bottom longlines. Handlines were used to assess the stocks near reefs, and longlines for those in areas away from reefs and on lagoon soft bottoms. Handline stations (363) were sampled in three different regions (west, north and east lagoons) and three biotopes (nearshore, middle lagoon and barrier reef). A total of 104 species were caught, the major families being Lethrinidae, Lutjanidae and Serranidae. Species composition and catch per unit of effort (CPUE) varied among regions and biotopes. In particular, the north zone showed characteristics of an unexploited area with very high yields. An increase in CPUE and fish size from the coast towards the barrier reef and with increasing depth was observed in all regions. Correlations between CPUE from handlines and the numerical density and biomass estimates from underwater visual censuses (UVC) performed on nearby reefs were significant. A total of 206 longline stations were sampled in the west and east regions. A total of 80 species were caught, the major families being the same as those caught in stations with handlines with the addition of Carangidae. The CPUE of longlines was significantly correlated with numerical density and biomass estimates made by UVC along the longlines. Stock estimates based on the correlations between CPUE from handlines or longlines and UVC estimates indicate that 90 % of the lagoon stock were found on soft bottoms and near reefs. Reefs, despite biomasses that were five times larger than soft-bottom and near-reef areas, made only a small contribution to total stock. The ratio between density and CPUE was highest for the lowest densities, thus indicating that fishing efficiency increased with density. Comparison of handline catch performed around reefs and UVC data from nearby reefs suggests that there were bold relationships between the fish assemblages of these two biotopes.

The potential yields of demersal reefal and lagoonal finfish, which are valuable for trade or consumption in the Northern Province of New Caledonia, are unknown. Fishing pressure was estimated from tally-sheets of professional fishermen to assess trade fishing and a household consumption survey to assess subsistence fishing. Total yield was estimated to be 1326 tonnes in 1995, 94 % of which correspond to subsistence fishing. Most of the catches were taken from the east and west coasts; in the north zone (Bélep Archipelago) catches were very low. From this fishing pressure and the total stock values, the maximum sustainable yield (MSY) was estimated to be 12600 t, which was about 10 % of the total stocks assessed in the Northern Province (138300 t). This MSY was dominated by five families, namely Lethrinidae, Acanthuridae, Scaridae, Serranidae and Lutjanidae. The results suggest that fishing effort in the Northern Province of New Caledonia could increase without endangering the stock. However, on a smaller geographical scale, some locations (Koné and Népoui on the west coast), reef habitats (near-reef areas) or species groups (fish caught by line) were already being intensively fished. Fishing techniques should be diversified so that part of the fishing effort be redirected toward the least heavily exploited groups of species and biotopes.

From April 1995 to April 1996, the annual reproductive cycle of male Eurasian perch Perca fluviatilis was studied at the Lindre Center (Moselle, France). At monthly intervals, five males (mean body weight of 133 ± 43 g and total length of 175 ± 9 mm) were caught. From sampled organs, the gonadosomatic (GSI), hepatosomatic (HSI) and viscerosomatic indexes (VSI) were calculated and plasma testosterone (T) and 11-ketotestosterone (11KT) levels were measured. After the spawning period in mid-April, GSI and HSI dropped and VSI increased to 3.8 ± 0.1 %. In September, GSI reached its maximum (8.5 ± 1.8 %). At this time, spermatocytes, spermatids and spermatozoa were abundant, whereas in June only spermatogonia were observed. During winter, GSI was stable at approximately 5 %, HSI reached its maximum (1.9 ± 0.3 %) and VSI was low (2.6 ± 0.2 %). From April to November 1995, plasma T and 11KT concentrations were low (< 0.5 ng·mL–1). Plasma T levels increased significantly in December and reached peak levels (12.3 ± 2.1 ng·mL–1) in January, then decreased in February and increased again until spawning in April (6.8 ± 2.1 ng·mL–1). This second elevation could be related to the beginning of a new spermatogenic cycle. Plasma levels of 11KT increased significantly from October to February 1996 (4.9 ± 1.1 ng·mL–1). From February to the spawning period (April 1996), plasma 11KT decreased significantly, but values were significantly higher than those measured in fall. Males were spermiating from January to spawning in April.

The effect of feeding time was studied in rainbow trout fed different dietary levels of fat. Fish were fed either 1 h after light on in the morning or 1 h after light off in the evening with a low energy diet (LE, 6 % lipid) or a high energy diet (HE, 23 % lipid). Regardless of the diet, apparent digestibility and post-prandial protein synthesis were higher in fish fed in the morning than in those fed at the beginning of the night. In fish fed the LE diet in the morning, growth performance and nutrient retention efficiency tended to be higher than in those fed at the beginning of the night. In contrast, fish fed the HE diet in the morning had lower protein growth rate, protein content and protein retention efficiency than those fed in the evening. These results suggest that protein metabolism might be involved in the effect of feeding time on growth and that there is an interaction between the time of feeding and dietary level of fat on growth.

Pirapatinga given submaximal rations grow faster when fed at night than when fed during daylight hours. It was hypothesised that these fish were day–night conformers, with a nocturnal acrophase, and expended more energy during daytime feeding. In order to test this hypothesis, telemetry was applied to study the behaviour of cultured pirapatinga (stocking biomass of 40 kg·m–3; 27 ± 1 °C; 6 mg O2·L–1; 13.5 L/10.5 D) under three feeding schedules (diurnal versus nocturnal feeding [12 g·d–1·fish–1 over 10 h] versus fasting) that were evaluated in succession over 1 week each. This implied a feasibility study to test the adequacy of tagging fish with surgically implanted transmitters (ventrolateral incision, posterior to the pelvic girdle). There was no mortality or tag loss over 12 months following surgery, and abdominal incisions healed within 4 weeks. Under all three feeding schedules, activity increased at dawn, peaked during daytime, decreased at twilight and reached a minimum during the late night. Diurnal and twilight activity levels were similar under all three schedules, whereas nocturnal activity was significantly lower with daytime feeding. These observations indicate that pirapatinga have a diurnal activity acrophase, which is little influenced by food availability, and imply that the higher growth of fish fed at night does not originate from lower energetic expenditures.

In order to elucidate the population structure of the deep-sea shrimp (Pandalus borealis) in the NE Atlantic, 32 subsamples and 3865 individuals were analysed for allozymic variation. They were caught at various locations in the Barents Sea, in waters off Svalbard, Jan Mayen and Iceland, and in fjords along the Norwegian coast. Only three enzymes (malate dehydrogenase, phosphoglucomutase and glucosephosphate isomerase) of the 22 initially tested showed a combination of gel images that could be interpreted with confidence and allozymic variation. The locus coding for malate dehydrogenase was by far the most polymorphic. Samples caught within the Barents Sea and in the Svalbard area showed no significant heterogeneity in allele frequencies, supporting earlier suggestions of only one population of P. borealis in the Barents Sea. Genetic differentiation was found, however, between Norwegian fjords and the Barents Sea, and among fjords.