MARINE ECOLOGY PROGRESS SERIES Vol. 334: 263–272, 2007 Published March 26 Mar Ecol Prog Ser Schooling behaviour of small pelagic fish: phenotypic expression of independent stimuli Patrice Brehmer1,*, François Gerlotto2, Charline Laurent1, Pascal Cotel3, Alina Achury4, Birane Samb5 1Centre de Recherche Halieutique Méditerranéenne et Tropicale (IRD; IFREMER and University of Montpellier II), 1 Avenue Jean Monnet, BP 171, 34203 Sète, France 2Institut de Recherche pour le Développement, c/o Instituto del Mar del Peru, Esq. Gamarra y Gral Valle s/n, Chucuito, La Punta, Callao, Peru 3Institut de Recherche pour le Développement, Sainte Clotilde, 97490, Île de La Réunion, France 4Fundación La Salle, Estación de Investigaciones Marinas Isla Margarita, Isla de Margarita, Punta de Piedras 6318, Venezuela 5Institut Sénégalais de Recherche Agronomique, Centre de Recherche Océanographique de Dakar-Thiaroye, BP 2241, Senegal ABSTRACT: We compared echo sounder measurements of fish schools on Venezuelan and Sene- galese coastal shelves. In both areas Sardinella aurita was the predominant aggregative species; the stocks are known to be genetically independent. Persistent school typology is described by analysing 7 acoustic surveys in Venezuela and comparing with pioneer observations in Senegal. The echo types encountered in both countries were comparable, with S. aurita and their associated spe- cies having comparable schooling behaviour. School morphology appears to be more related to spe- cies than to global characteristics of their ecosystems, as fish school descriptors were similar for both stocks. Nevertheless, S. aurita may adopt different school morphologies. Differentiation between day and night fish schools was mainly due to morphological variables, as determined by fractal dimen- sion. Echo types were different between day and night in both Venezuela and Senegal. This analysis shows that several independent parameters can influence school morphology. However, although schooling behaviour is temporally highly variable, it shows several constant, typical 2D morphologi- cal structures, i.e. echo types. The observation of small pelagic fish schools occurring during the night is not unusual according to the literature, although a strict diurnal schooling behaviour, with scatter- ing during the night, is often reported. Schooling behaviour appears to be a phenotypic expression depending on biotic and abiotic stimuli, intrinsic characteristics and motivation. KEY WORDS: Behaviour · Diel cycle · Echo type · In situ observations · Phenotypic expression · Schooling · Small pelagic fish Resale or republication not permitted without written consent of the publisher INTRODUCTION (Marchal & Petitgas 1993, Hjellvik et al. 2004). Never- theless, in situ measurements of basic behaviour, such The aggregation patterns of small pelagic fish pro- as schooling and diel aggregative patterns, are com- duce clusters of fish schools where most of the bio- plex to obtain and to interpret. The most common diel mass is concentrated. Describing such aggregative schooling pattern of small pelagic species consists of patterns is essential to modelling of the in situ school- fish scattering during the night and gathering in ing behaviour of fish (Azzali et al. 1985, Viscido et al. schools during the day (Shaw 1961, Blaxter & Holliday 2004). Schooling behaviour must be considered in the 1969, Muiño et al. 2003, Milne et al. 2005) although management of fisheries of pelagic species because of there are numerous exceptions (Bertrand et al. 2004, its impact on the estimation of catchability coefficients 2006). *Email: [email protected] © Inter-Research 2007 · www.int-res.com 264 Mar Ecol Prog Ser 334: 263–272, 2007 Acoustic survey methods commonly employed in MATERIALS AND METHODS pelagic fisheries can be used to explore fish schooling behaviour, providing standardized in situ samplings of Comparative approach using 2 independent fish schools (Fréon & Misund 1999, Simmonds & MacLennan stocks. The study areas were situated on both edges of 2005). Using acoustic surveys, Petitgas & Levenez (1996) the northern tropical Atlantic Ocean, in shelf waters of demonstrated the existence of different schooling pat- Senegal and Venezuela (Fig. 1). Sardinella aurita is a terns within a fish community, establishing an ‘echo type dominant schooling planktivore in the Eastern Vene- classification’ of schools in the Senegalese small pelagic zuela and South Senegalese continental shelves (Fréon fish community. In this study, to determine whether & Mendoza 2003). Among the ca. 20 most common these results are relevant to other communities, we per- pelagic species in Senegal (engraulids, clupeids, formed a comparative analysis of schooling behaviours carangids, scombrids) the main commercial landings of similar fish assemblages from different geographical are of S. aurita, followed by Engraulis encrasicolus, areas. Following definitions by Gerlotto et al. (2004), we Ethmalosa fimbriata, and Sardinella maderensis. In studied school ‘morphology’, represented by external Venezuela, there are at least 67 different small pelagic and global school characteristics (e.g. overall dimen- species (16 engraulids, 14 clupeids, 29 carangids, sions) and school ‘structure’, defined by the internal 8 scombrids) but the bulk of the biomass consists of arrangement of individual fish inside schools, e.g. vac- 4 clupeids (S. aurita, Harengula clupeoa, Etrumeus uoles and nuclei. Continuous in situ acoustic measure- teres, Opisthonema oglinum), 1 anchovy (Cetengraulis ments of diel schooling behaviour of small pelagic fish edentulus), 3 horse mackerels (Trachurus lathami, (Fréon et al. 1996, Bertrand et al. 2004) were obtained Decapterus punctatus and Selar crumenophtalmus) from the same vessel in Senegalese and Venezuelan wa- and 1 scombrid (Scomber japonicus) species. In both ters (Brehmer 2004). A comparative approach employing countries, S. aurita represents ca. 80% of the pelagic 2 different acoustic methodologies allowed comparison biomass. Three sets of acoustic data were used, includ- of school morphologies in 2 genetically independent ing the original echo traces of Petitgas & Levenez tropical fish populations (Chicki 1995). The questions ad- (1996). Echo type classifications obtained from the dressed in this study were: (1) Is small pelagic fish school Venezuelan database were compared with the Sene- morphology the result of a given strategy to occupy a galese results (Petitgas & Levenez 1996). Also, fish given space depending on the environmental conditions school descriptors observed on both fish stocks or a simple phenotypic expression? (2) Do independent (S: Sénégal; V: Venezuela) were compared between stocks of fish share similar diel schooling dynamics? night and day hours. Classification of Venezuelan small pelagic fish echo types. An acoustic database from the eastern Venezuelan coastal shelf was compiled by combin- ing data from 7 surveys (Table 1). A Simrad EKS or EK 400 echo sounder was used. Both echo sounders emitted at 38 kHz with a pulse length of 0.6 ms and a Time Varied Gain function in 20 Log R. An Agenor echointegrator was used. The entire Venezuelan data- base included a total of 7736 single Elementary Sampling Distance Units (ESDU = 1 nautical mile) with a predom- inant echo type structure, over a total of 14 500 n mile analysed. Visual analysis of fish school targets on echograms, as defined by Petitgas & Levenez (1996), was employed to clas- sify echo types. Petitgas & Levenez (1996) applied this visual echo classifi- cation to a set of 4 Senegalese acoustic Fig. 1. The study areas, situated in the northern tropical Atlantic Ocean. surveys representing 3794 ESDUs. Nine Sardinella aurita constitute the main small pelagic schooling species, with 2 in- school types were defined: (1) small dependent stocks: Eastern Venezuela and Southern coast of Sénégal schools ‘SSM’, (2) tower shaped schools Brehmer et al.: Schooling behaviour of small pelagic fish 265 Table 1. List of acoustic surveys used to constitute the Venezuelan data base for echo type classification, with 14 500 Elementary Sampling Distance Units of 1 n mile. The Varget surveys were used for echo integration by shoal processes (6236 fish schools) Survey Year Country RV Sounder Echointegrator n Data process Echoven2 1986 Venezuela Capricorn EKS Agenor 4032 Ecoven3 1985 Venezuela La Salle EKS Agenor 2062 Ecoven4 1986 Venezuela La Salle EKS Agenor 2401 Echo type Ecoven5 1986 Venezuela La Salle EKS Agenor 965 discrimination Ecoven6 1987 Venezuela La Salle EKS Agenor 2109 by ESDU Fepelor1 1988 Venezuela La Salle EKS Agenor 1406 Fepelor2 1988 Venezuela La Salle EK 400 Agenor 1525 Varget1 1999 Sénégal Antéa EK 500 Movies+ 4606 Echointegration Varget2 1999 Venezuela Antéa EK 500 Movies+ 1630 by shoal ‘STW’, (3) compact schools ‘SCO’, (4) pole shaped 4606). Short term variability and circadian changes schools ‘SPO’, (5) fluffy schools ‘SFL’, (6) zigzag shaped were taken into account, and night and day periods schools ‘SZZ’, (7) thermocline shaped layers ‘LT’, (8) were differentiated (Fréon et al. 1996). macro structures ‘MS’ (not used by Petitgas & Levenez The ‘echo integration by shoal’ process was 1996 because they were too infrequent) and (9) scat- employed (Scalabrin & Massé 1993, Petitgas et al. tered fish ‘FS’. In an ESDU more than one target type 2001, Guillard et al. 2004), which did not take into con- can be recorded. In such cases, the corresponding sideration the individual fish but rather the structure