411 



Feeding habits of European hake 



(Merluccius merluccius) 



in the central Mediterranean Sea 



Paolo Carpentieri 



Francesco Colloca 



Department of Animal and Human Biology 



University "La Sapienza" 



Viale dell'Umversita 32 



00185 Rome, Italy 



E-mail address (for P. Carpentieri) paolo.carpentieri@uniromal it 



Massimiliano Cardinale 



Institute of Marine Research 

 National Board of Fisheries 

 P.O. Box 4 

 45 332, Lysekil, Sweden 



Andrea Belluscio 



Giandomenico D. Ardizzone 



Department of Animal and Human Biology 

 University "La Sapienza" 

 Viale dell'Universita 32 

 00185 Rome, Italy 



European hake (Merluccius merluc- 

 cius) is an important predator of 

 deeper shelf-upper slope Mediterra- 

 nean communities. It is a nectoben- 

 thic species distributed over a wide 

 depth range (20-1000 m) throughout 

 the Mediterranean Sea and the north 

 east Atlantic region (Fisher et al., 

 1987). Notwithstanding the ecologi- 

 cal and economic importance (Oliver 

 and Massuti, 1995) of hake in the 

 Mediterranean, many aspects of its 

 biology (e.g., recruitment and repro- 

 duction), due to multiple spawning 

 (Sarano, 1986) and the current state 

 of exploitation, are poorly understood 

 (Arneri and Morales-Nin, 2000). 



Recent studies on hake feeding 

 habits in the Mediterranean (Papa- 

 costantinou and Caragitsou, 1987; 

 Bouaziz et al., 1990; Oliver and Mas- 

 suti, 1995) have focused on 0-3 age 

 groups using data from trawl catch- 

 es (Recasens et al., 1998; Colloca et 

 al., 2000). For this reason, trophic 

 habits of older individuals (Bozzano 

 et al., 1997) and possible ontogen- 

 esis-related diet changes are almost 



unknown. Therefore, in this study 

 we combined samples from trawl and 

 gillnet fisheries collected in the same 

 fishing ground (Colloca et al., 2000) 

 to address these issues. 



Materials and methods 



The study area is located off the cen- 

 tral western coasts of Italy, cover- 

 ing 13,404 km 2 between 20 and 700 

 meters depth (outer boundaries: lati- 

 tude 40°52'64, longitude 13°23T3; lat- 

 itude 42°20'30, longitude 11°16'32). 



Monthly size-stratified samples 

 were obtained from spring 1997 to 

 winter 1998 both from bottom-trawls, 

 gillnet commercial-vessels, and from 

 commercial landings. Trawlers catch 

 mainly 0-2 year-old juveniles; they 

 rarely capture adults (Aldebert et al., 

 1993; Abella et al., 1997; Ardizzone 

 and Corsi, 1997). The gillnet fishery 

 exploits mainly adults of the species 

 (>25 cm TL). 



Caught fish were kept on ice, 

 subsequently frozen to prevent di- 



gestion of their stomach contents, 

 taken to the laboratory, measured 

 (total length: TL) to the nearest 

 1 mm, and weighed to the nearest 

 0.01 g. Sex and maturity stage were 

 also recorded. Maturity state was 

 determined by macroscopic analysis 

 of the gonads by using the maturity 

 scale for partial spawners (Holden 

 and Raitt, 1974). 



Stomachs were removed and their 

 contents weighed to the nearest 

 0.001 g. Prey items were identified 

 and sorted into taxonomic groups to 

 the species level whenever possible. 

 When the state of digestion was 

 more advanced, prey were checked 

 and grouped into unidentified fish, 

 cephalopods, or crustaceans. The de- 

 gree of digestion of the prey was not 

 considered in the analysis. Empty 

 stomachs and those with partially 

 everted or unidentified contents were 

 excluded from the total sample. 



With the exception of the largest 

 individuals (grouped into two het- 

 erogeneous length classes), all re- 

 maining hakes in the sample were 

 grouped into 5-cm length classes. 

 The study of size-related diet varia- 

 tions was based on these groups. The 

 contribution of each food item to the 

 diet of these fish length groups was 

 evaluated by using the index of rela- 

 tive importance (IRI, Pinkas et al., 

 1971) as modified by Hacunda (1981): 

 IRI= F{N + W). 



This index, expressed as 



IRI%=IRI 



- IIRI . 



100, 



incorporates the percentage by num- 

 ber (N%), wet weight (W7<), and fre- 

 quency of occurrence (F% ) (Hyslop, 

 1980). Hierarchical cluster analysis 

 and nonmetric multidimensional scal- 

 ing (NMDS), based on Bray-Curtis 

 similarity and on the IRI%, were 

 used for classification and ordination 

 of hake size classes (Clarke and War- 

 wick, 1994). 



Manuscript submitted 27 April 2003 

 to the Scientific Editor's Office. 



Manuscript approved for publication 

 13 December 2004. 



Fish. Bull. 103:411-416(2005). 



