Vol. 531: 207–219, 2015 MARINE ECOLOGY PROGRESS SERIES Published July 2 doi: 10.3354/meps11347 Mar Ecol Prog Ser Feeding ecology of two squid species from the western Mediterranean M. Valls1,*, M. Cabanellas-Reboredo2, I. Uranga1, A. Quetglas1 1Instituto Español de Oceanografía, Centre Oceanogràfic de les Balears, Moll de Ponent s/n, Apdo. 291, 07015 Palma, Spain 2Instituto Mediterráneo de Estudios Avanzados, IMEDEA (CSIC-UIB), C/Miquel Marqués 21, 07190 Esporles, Illes Balears, Spain ABSTRACT: The squid Loligo vulgaris (LV) and L. forbesii (LF) are 2 cephalopod species occur- ring in the Atlantic Ocean and the Mediterranean Sea. Bathymetric segregation allows the co- existence of both species, with LV preferentially inhabiting the shallow shelf and LF living on the shelf-break and upper slope grounds. In this paper, the feeding habits of LV and LF were studied for the first time in the Mediterranean, by means of stomach content analysis (1452 and 900 indi- viduals of LV and LF, respectively). The main objective was to determine the diet of both species, analysing temporal and ontogenetic diet changes and inferring predator−prey interactions. Fish were by far the most important prey in both squid, followed by crustaceans and cephalopods. Prey composition revealed the bathymetric segregation of both species in the Mediterranean. Whereas LV preferentially consumed typical coastal species of sparids and gobiids, LF preyed on slope inhabitants such as myctophids and euphausiids. Ontogenetic shifts of diet occurred in both squid, but took place at contrasting sizes, suggesting that the factors triggering them might be species- specific. The diet of small-sized LV individuals was more dependent on bottom-living organisms than in large individuals, which preyed mainly on benthopelagic fish. During the main reproduc- tive period in spring, LV increased the consumption of highly nutritive prey such as polychaetes (nereidids). Size-related differences in LF diet during the second half of the year indicated a deeper distribution of large individuals, preferentially preying on mesopelagic species and being thus involved in benthic−pelagic coupling. KEY WORDS: Diet analysis · Stomach contents · Loligo spp. · Ontogenesis · Seasonal variation · Predator–prey interactions · Bathymetric segregation Resale or republication not permitted without written consent of the publisher INTRODUCTION Guerra & Rocha 1994, Pierce et al. 1994, Coelho et al. 1997, Wangvoralak et al. 2011). It is accepted that The squid Loligo vulgaris Lamarck (1798) and L. squid have a large trophic impact on other species in forbesii Steenstrup (1857) are 2 cephalopod species the food web and top-down control from squid to their with few external morphological differences and prey can be high (Coll et al. 2013, Young et al. 2013). which inhabit the eastern Atlantic Ocean and Medi- They are also important fishery resources. Both spe- terranean Sea (Jereb & Roper 2010). They are im- cies are taken as by-catch of the Mediterranean bot- portant prey for a number of marine predators such as tom trawl fishery (Guerra & Rocha 1994, Relini et al. pelagic and demersal fish and marine mammals (e.g. 1999, Sifner & Vrgoc 2004, Massutí & Reñones 2005), Morte et al. 1997, Peristeraki et al. 2005, Bearzi et al. while L. vulgaris also supports important artisanal 2011). In turn, they prey on a broad spectrum of spe- and recreational fisheries (Guerra & Rocha 1994, cies, especially fish, but also crustaceans, cephalopods Morales-Nin et al. 2005, Cabanellas-Reboredo et al. and polychaetes (Roper et al. 1984, Collins et al. 1994, 2014a). Thus, the role of these squid, as predator and *Corresponding author: [email protected] © Inter-Research 2015 · www.int-res.com 208 Mar Ecol Prog Ser 531: 207–219, 2015 prey, and the interactions they have with other spe- whether these species display differences in diet re - cies, are key factors in the trophodynamics of marine lated to sex, season or ontogenetic growth. Diet com- ecosystems (Rocha et al. 1994, Navarro et al. 2013), position studies constitute a crucial first step in order and adequate knowledge about their role and interac- to better understand trophic interactions, which in tions is important to allow appropriate resource man- return allow building robust, meaningful marine food agement (Moreno et al. 2013). web models (Christensen & Walters 2004, Coll et al. Diet composition and feeding ecology of these lolig- 2006, 2008, Moreno et al. 2013). inids are well documented in Atlantic waters. They mainly feed on fish, with little frequency variation but different species composition depending on the region MATERIALS AND METHODS (Roper et al. 1984, Collins et al. 1994, Guerra & Rocha 1994, Pierce et al. 1994, Coelho et al. 1997, Wangvo- Species sampling ralak et al. 2011). Seasonal and daily spatial migrations, related to reproduction and feeding, are known to oc- Individuals from both species were sampled cur in both species (Rocha & Guerra 1999, Cabanellas- monthly from commercial fishing boats off the Reboredo et al. 2012, 2014b). However, to date nothing Balearic Islands (western Mediterranean) between is known about the diet of L. vulgaris and L. forbesii January 2009 and January 2010 (N = 984 Loligo vul- in the Mediterranean, although different aspects of garis and 693 L. forbesii). Squid were obtained from their life cycle (e.g. growth and reproduction) are re - bottom trawlers operating between 50 and 750 m and latively well studied both in the western (Mangold- small-scale boats generally fishing down to 100 m. Wirz 1963, Worms 1979, Wurtz & Giuffra 1989, Sán - Species were identified on the basis of macroscopic chez & Demestre 2010) and central (Ragonese & Jereb external characters (e.g. tentacular club) after Roper 1986, Sifner & Vrgoc 2004) regions. In the Mediterran- et al. (1984). Whereas L. vulgaris were collected from ean Sea, both species show a clear bathymetric segre- trawl (76%) and artisanal fishery (24%), L. forbesii gation, as L. vulgaris preferentially inhabits waters were all obtained from the trawl fishery. Additional shallower than 200 m and L. forbesii is found at depths samples were taken from the fishery-independent between 100 and 600 m, slightly overlapping on the MEDITS bottom-trawl surveys (Bertrand et al. 2002) deep continental shelf (Quetglas et al. 2000). conducted in the study area during early summer All available information on squid diet is based on from 2007 to 2010 (N = 468 L. vulgaris and 207 L. stomach content analysis. In general, gut content forbesii). For each individual, the following measure- analysis fails to provide information on long-term ments were taken: dorsal mantle length (ML, to the feeding habits (Jackson et al. 2007) and neglects some nearest mm), total weight (TW, to the nearest 0.1 g) dietary materials. These analyses are further biased and sex (male, female, undetermined). in cephalopods, which reduce the food to hardly rec- ognizable pieces (Hyslop 1980, Boyle & Rodhouse 2005). Despite these shortcomings, stomach content Stomach sampling and diet indices analysis remains the main source of data for prey items and provides useful information on predator Prey items from gut contents were analysed under feeding habits and ecology (Clarke & Kristensen 1980, a binocular microscope and identified to the lowest Laptikhovsky et al. 2010, Miller et al. 2013). Although possible taxon considering key morphological fea- isotopic analysis performs better than dietary analysis tures (i.e. otoliths or claws) using published guides in revealing assimilated food, it does not provide in- (Zariquiey 1968, Lombarte et al. 2006) and our own formation on predator−prey interactions at the species reference collections. The number of examples of level (Winemiller et al. 2007, Young et al. 2015). prey within a prey category was recorded and the We investigated the feeding habits of L. vulgaris following indices calculated to analyse the diet and and L. forbesii for the first time in the Mediterranean feeding intensity (Hyslop 1980): (1) relative abun- Sea by means of stomach content analysis. The diets dance (%N), calculated as the percentage of each were expected to reflect the bathymetric segregation prey item compared to the total number of prey of the 2 species in the area, yet some competition items; (2) frequency of occurrence (%O), calculated should exist at depths where the squid overlap. The as the percentage of stomachs containing each prey main objective was to analyse the feeding habits of item compared to the total number of stomachs these 2 squid, to determine differences and similari- containing food; and (3) vacuity index (%v), or the ties in their feeding ecology. We also investigated percentage of empty stomachs. Valls et al.: Feeding ecology of Mediterranean squid 209 To determine the feeding strategy, niche breadth arithmetic average clustering (Clarke & Gorley 2006) was calculated using Levins’ standardized index Bi: by calculating Bray-Curtis similarity resemblance matrices. Analyses were done using abundance data. 1 ⎛ 1 ⎞ In order to reduce the weight of numerically domi- (1) Bi = ⎜ −1⎟ n −1 p2 nant species, a prior square root transformation of ⎝ ∑ j ij ⎠ the data was performed. The significant groups were where pij is the proportion of diet of predator i that is determined using the SIMPROF test (p < 0.01) made up of prey j, and n is the number of prey cate- (Clarke & Gorley 2006). gories (Krebs 1999). The values of pij are determined from a matrix (e.g. Table S1 in