Mangroves and seagrass beds are considered nurseries for juvenile fish, but little experimental evidence exists to elucidate which factors make them attractive habitats. A multifactorial field experiment on the use of these habitats by juvenile reef fish and their behaviour was performed during daytime with experimental units (EUs: 1 × 1 × 0.8 m), each representing a unique combination of the factors structure, shade, and food, using artificial seagrass leaves (AS) and artificial mangrove roots (AM). Diurnally active herbivores were most abundant in EUs containing food, and grazed on algae growing on the structures, but were not attracted to structures in the absence of food. The most abundant diurnally active zoobenthivores (Eucinostomus spp.) were present in highest numbers in any EU with food, where they fed on zoobenthos or rested on the bottom. The nocturnally active zoobenthivore/zooplanktivore Ocyurus chrysurus and the diurnally active piscivore Sphyraena barracuda were primarily attracted to structure, in which they rested and were not observed to feed. Haemulon flavolineatum was mainly attracted to AS, Lutjanus mahogoni was attracted to AS or shade, whereas L. apodus, O. chrysurus and S. barracuda were found in AM as well as in AS. The data suggest that during daytime, herbivores and diurnally active zoobenthivores are probably attracted to mangroves and seagrass beds primarily by food, and nocturnally active zoobenthivores by structure (in interaction with shade) that offers shelter from predation. S. barracuda is also attracted primarily to structure, but the larger individuals probably use this for ambush predation rather than for protection. In conclusion, our experiment clarifies that presence of structure, food and shade significantly contribute to the attractiveness of mangroves and seagrass beds to juvenile reef fish.

Exploiting several fish species simultaneously reduces variability in daily catches. The reduction depends on the number of species, the catch-frequency distributions of individual species, and the level of co-occurrence of species in the catch. We explore theoretically the reduction of variability (coefficient of variation; CV) in the total catch by combining the distributions of daily catches of individual fish species, including zero catches, into a total catch frequency distribution. Theoretical findings are tested with an example from a stationary lift-net fishery for schooling small pelagic species around Ambon Island in the Central Moluccas, Indonesia. This fishery catches over 30 species, all with high daily variability (CV = 2.2-13.4). The reduction of variability in the total catch (CV = 1.7) is a result of the dominance and independent occurrence of the three main species. We conclude that in this fishery the information value of the total catch as an indicator of the catches of the individual species is low.Résumé : L'exploitation simultanée de plusieurs espèces de poissons peut réduire la variabilité des captures journalières. L'importance de cette réduction dépend du nombre d'espèces, de la distribution de fréquence des espèces dans la récolte et du niveau de la co-occurrence des espèces dans la récolte. Nous examinons en théorie la réduction de la variabilité (coefficient de variation; CV) dans les captures totales en combinant les distributions des captures journalières d'espèces particulières, incluant les captures nulles, dans la distribution de fréquence des captures totales. Un exemple tiré d'une pêche commerciale à carrelet de petits poissons pélagiques vivant en bancs près de l'Îsle Amboine (Moluques, Indonésie) permet de tester les résultats théoriques. Les récoltes incluent plus de 30 espèces de poissons, chacune avec une variabilité journalière élevée (CV = 2,2-13,4). La réduction de la variabilité dans la capture totale (CV = 1,7) est le résultat de la dominance et de l'occurrence indépendante des trois espèces principales. Dans cette pêche commerciale, la valeur prédictive de la capture totale comme indicateur de la capture des espèces particulières est basse.[Traduit par la Rédaction] van Oostenbrugge et al. 1043