Feeding ecology of the Stomiiformes (Pisces) of the northern Mid-Atlantic Ridge. 1. The Sternoptychidae and Phosichthyidae

Progress in Oceanography 130 (2015) 172–187

Contents lists available at ScienceDirect

Progress in Oceanography

journal homepage: www.elsevier .com/locate /pocean

Feeding ecology of the Stomiiformes (Pisces) of the northernMid-Atlantic Ridge. 1. The Sternoptychidae and Phosichthyidae

http://dx.doi.org/10.1016/j.pocean.2014.11.0030079-6611/� 2014 Elsevier Ltd. All rights reserved.

Abbreviations: AM, amphipods; AZ, Azorean Zone; CAP, Canonical Analysis ofPrincipal components; CE, cephalopods; CG, Charlie Gibbs Fracture Zone; CH,chaetognaths; CL, cladocerans; CO, copepods; DW, dry weight; EGoM, eastern Gulfof Mexico; EU, euphausiids; F, frequency of occurrence; FI, fish; FS, FaradaySeamount Zone; IRI, index of relative importance; L, ‘‘Large’’ size class; N, number;N MAR, northern Mid-Atlantic Ridge; OS, ostracods; OT, other taxa; PERMANOVA,PERmutational Multivariate ANAlysis Of VAriance; PT, pteropods; Rinst, instanta-neous ration; RR, Reykjanes Ridge; S, ‘‘Small’’ size class; SL, standard length; W,weight; WW, wet weight.⇑ Corresponding author. Tel.: +351 292 207 800; fax: +351 292 207 811.

E-mail addresses: [email protected], [email protected] (V. Carmo).

Vanda Carmo a,⇑, Tracey Sutton b, Gui Menezes a, Tone Falkenhaug c, Odd Aksel Bergstad c

a Centre of IMAR of the University of the Azores, Department of Oceanography and Fisheries, LARSyS Associated Laboratory and MARE – Marine and Environmental SciencesCentre, Rua Prof. Doutor Frederico Machado 4, 9901-862 Horta, Portugalb Nova Southeastern University Oceanographic Center, Dania Beach, FL 33004, USAc Institute of Marine Research, Flødevigen, N-4817 His, Norway

a r t i c l e i n f o a b s t r a c t

Article history:Received 5 August 2014Received in revised form 6 November 2014Accepted 6 November 2014Available online 13 November 2014

Comprehensive trophic studies in the vast mid-oceanic regions are rare compared to coastal and fisher-ies-oriented investigations. Field sampling conducted by the multidisciplinary, international Census ofMarine Life project MAR-ECO, namely the 2004 G.O. Sars cruise, has generated one of the largest openocean deep-pelagic sample collections ever obtained. With the overall goal of understanding carbon flowprocesses within and through the deep-pelagic nekton associated with the northern Mid-Atlantic Ridgesystem (N MAR), quantitative trophic analyses were conducted in order to identify the major intraspecificpatterns in diet of characteristic members of the midwater fish community. Diets of five abundant speciesof zooplanktivorous fishes were examined in detail in terms of prey taxonomy and variability in space,ontogeny and diel cycle. Two major patterns of feeding were identified. Pattern 1 included three speciespreying primarily on copepods, Argyropelecus hemigymnus, Maurolicus muelleri and Vinciguerria attenuata,the former two of which revealed spatial differences in diet with latitude, mostly likely related tolatitudinal prey distributions and densities. Maurolicus demonstrated ecological differences in diet thatmirrored phenotypic variation North and South of the Subpolar Front, an ‘‘oceanic species concept’’question that warrants further research. Pattern 2 included two species feeding primarily on amphipods,Argyropelecus aculeatus and Sternoptyx diaphana, both of which showed ontogenetic variability in feedingprimarily related to specific amphipod taxon sizes, rather than prey switching to other major prey taxa.This is the first study that highlights the importance of amphipods in the diets of these species. All fishspecies showed selectivity in prey choice, possibly related to competition with the other major nektoncomponents along the N MAR, namely the Myctophidae and other zooplanktivorous Stomiiformes. Dailyration fell within the expected values for midwater fishes (1–4% of body weight) with the exception ofS. diaphana, suggesting that this species is unique among the zooplanktivores – either its ration is threetimes higher than the other species, or it eats one-third as often (i.e., every 3 days). Given the high percentstomach fullness observed throughout the diel cycle, we believe the former to be the case, which is thefirst estimate of its kind for a midwater fish. In order to facilitate further quantitative research onmesopelagic carbon cycling, detailed prey length/weight regressions are presented here.

� 2014 Elsevier Ltd. All rights reserved.

Introduction

Since organic matter is almost exclusively produced byphotosynthesis in the surface layers, studying the pathways bywhich living or dead energy is transferred to great depths is ofgreat importance to better understand deep-sea ecosystems(Vinogradov and Tseitlin, 1983; Herring, 2002). The transfer ofenergy and biomass to the deep depends mostly on the passivesinking of organic matter, but also on the vertical migrations oforganisms (Angel, 1997a). Mesopelagic fishes play significant rolesin the foodweb as predators of mesozooplankton and micronekton

V. Carmo et al. / Progress in Oceanography 130 (2015) 172–187 173

and as prey of other fishes (Balanov, 1994), especially in the openocean. There, carbon sources are quite limited (e.g., Hopkins andBaird, 1977; Clarke, 1982; Balanov, 1994; Balanov et al., 1994),particularly, in the oligotrophic central ocean gyres where thereis not much energy available for sustaining life in the deep(Angel, 1997b).

In 1977, Hopkins and Baird reviewed the feeding ecology ofmidwater fishes in the world ocean and concluded that most linesof investigation of the subject were still embryonic andemphasized the need for more diet analysis research. Over nearlyfour decades later, our knowledge of the feeding of the mostabundant species has improved progressively, but thorough andintegrated foodweb studies considering many predator and preycomponents are still scarce. Some comprehensive diet studies onmidwater fishes were published from most oceanic regions(e.g., Appelbaum, 1982; Hopkins et al., 1996 – Atlantic; Clarke,1978, 1980, 1982; Gordon et al., 1985; Gaskett et al., 2001; Yang,2011 – Pacific; Butler et al., 2001 – Indian; Pakhomov et al., 1996– Southern Ocean). However, the majority of the studies providedscattered data on certain species or groups mainly near or off thecontinental shelf and slope (e.g. DeWitt and Cailliet, 1972;Merrett and Roe, 1974; Hopkins and Baird, 1985; Dalpadado andGjøsæter, 1987), while the central oceanic regions were lessinvestigated (Hudson et al., 2014). Integrated studies of wholecommunities like that of Hopkins et al. (1996) are rare.

Mid-ocean ridges are vast, global features distributed in alloceans, but unexpectedly few studies were dedicated to the faunainhabiting such areas (Bergstad and Godø, 2003). Under theMAR-ECO project (Patterns and Processes of the Ecosystems ofthe northern Mid-Atlantic, 2003–2010; www.mar-eco.no), thepresent study was initiated to investigate the diets of some ofthe most numerically and biomass-dominant mesopelagicfishes of the northern section of the Mid-Atlantic Ridge, theStomiiformes. In particular, we aim to characterize the relativecontributions of different food sources to the diets of members ofthe families Sternoptychidae and Phosichthyidae and identifyintraspecific patterns. Later, we will investigate major interspecificlinkages (predator–prey relationships) and elaborate a conceptualfood-web model with the overall goal of better understandingthe carbon flow processes within and through the deep-pelagicnekton associated with the northern Mid-Atlantic Ridge.

Material and methods

Field sampling

The study material was obtained on the 2004 MAR-ECO cruiseon the RV G.O. Sars, the first focused deep-pelagic sampling effortsin the area of the Mid-Atlantic Ridge. A detailed description of themethods and cruise strategy was provided by Wenneck et al.(2008). The vessel zigzagged along the northern Mid-Atlantic Ridge(N MAR) from Iceland to the Azores (Leg 1) and back (Leg 2),sampling pre-determined locations (‘‘super-stations’’) in four mainzones: Reykjanes Ridge (RR), Charlie Gibbs Fracture Zone (CG), Far-aday Seamount Zone (FS), and Azorean Zone (AZ) (Fig. 1). Leg 1 (5June–3 July) was dedicated to pelagic sampling, while Leg 2 (4July–5 August) focused on the demersal fauna. Net samplingincluded several types of gear in an effort to sample a wide sizerange and taxa and allow discrete-depth sampling down to>3000 m. The depth strata considered in this study followedSutton et al. (2008): 1: 0–200 m, 2: 200–750 m, 3: 750–1500 m,4: 1500–2300 m and 5: >2300 m. Time of capture of studyspecimens are indicated herein as hours/minutes elapsed sincemidnight, local time (0000–2359). For point-sampling at thesuper-stations during Leg 1, two different multi-codend mid-water

trawls were used: a large pelagic fish trawl (Aakratrawl, 22-mmcod-end mesh) and a 36-m2 mouth area macrozooplankton trawl(‘‘krill trawl’’, 3–6 mm mesh). Additionally, a very large pelagictrawl (Egersund trawl, 50-mm cod-end mesh) was used foropportunistic sampling of particular targets located acoustically.On Leg 2, the bottom trawl used was a Campelen 1800 (‘‘shrimptrawl’’, 22/40-mm mesh). Specimens used in this study wereobtained by all these different gears. Schematics and details onthe sampling gears were given by Wenneck et al. (2008).

Sample processing and preservation

Catches from each net deployment and depth stratum werehandled separately. Total weight was recorded at sea using amotion-compensating scale (±0.1 g). Fishes were then sorted tothe lowest possible taxonomic level, counted and frozen, or in caseof rarity or taxonomic uncertainty, preserved in 10% formalin:sea-water. In the Bergen Museum, samples were fixed in formalin andlater transferred to 70% ethanol and incorporated into the curatedcollection, from which specimens for this study were retrieved.This selection included specimens of four species of the familySternoptychidae, Argyropelecus aculeatus Valenciennes 1850,Argyropelecus hemigymnus Cocco 1829, Maurolicus muelleri(Gmelin, 1789), Sternoptyx diaphana Hermann 1781, and onespecies of family Phosichthyidae, Vinciguerria attenuata (Cocco,1839). These were the most abundant members of each fish familyin the study area (Sutton et al., 2008), and thus putatively mostimportant in an ecosystem-based context.

Laboratory methods

Diet analysis methods followed Hopkins et al. (1996) andSutton and Hopkins (1996). Each fish was measured to the nearestmillimeter standard length (SL), weighed to the nearest milligram(wet weight, WW) and then dissected under a binocularstereoscope. The digestive tract was removed. The stomach wasdefined as the thick, muscular, distensible and pigmented anteriorsection of the digestive tract extending from the esophagus to theorigin of the intestine (Hopkins and Baird, 1981; Kinzer and Schulz,1988). Food in the mouth or esophagus was not included in theanalyses to avoid possible bias due to net feeding (Hopkins andBaird, 1975; Lancraft and Robison, 1980; Clarke, 1982; Roe andBadcock, 1984; Sutton, 2005). Data from the intestines were notconsidered in this study. Before removing the contents from eachstomach, an assessment of its fullness was made using thefollowing arbitrary scale (0–5) adapted from Sutton and Hopkins(1996): empty (0), <25% full (1), 25% to <50% full (2), 50% to<75% full (3), 75% to <100% full (4), totally full and extended stom-ach (5). Subsequently, the stomach was dissected under the stereo-scope and the contents were removed and sorted into major taxa.

Prey identification

All prey items were identified to the lowest possible taxonomiclevel, carefully counted to avoid duplication and then classifiedaccording to their state of digestion as: fresh/undigested (1),partially digested but still recognizable (2) and very digested/unrecognizable (3) (Tyler and Pearcy, 1975; Balanov et al., 1994).Intact and partially digested larger prey items were kept in vials,while smaller items were fixed in a labeled glass slide for athorough search for hard parts that enabled identification. Theslides were prepared with a glycerin:fuchsin-acid stain andobserved under higher magnification using a microscope with anattached camera. A photo-catalog of over 3500 images of the differ-ent parts and whole prey was built.

Fig. 1. Map of the study area along the northern mid-Atlantic ridge, including stations sampled (numbers) and main areas used for spatial analysis.

174 V. Carmo et al. / Progress in Oceanography 130 (2015) 172–187

Prey measurements

Measurements to the nearest 0.1 mm of prey in vials were madewith a binocular stereoscope using an ocular micrometric scale.Measurements of prey and hard parts (lm) on glass slides weremade using a microscope:camera system and the PC Software ZeissAxio Vision �, after calibration. Regarding different prey taxa, thefollowing measurements were taken: standard length of fishes;dorsal mantle length or beak length of cephalopods; back of eyesocket to tip of telson (excluding terminal spines) of decapodsand mysids; tip of rostrum to tip of telson (excluding terminalspines) of euphausiids; anterior end of eyes to tip of uropods ortelson (depending which was longer) of amphipods; valve lengthof ostracods; prosome length of copepods; maximum shell lengthof pteropods; total length of cladocerans, chaetognaths andpolychaetes; total length of individual bracts of siphonophores(fragmented from calycophoran or physonect colonies); and diam-eter of eggs.

Prey biomass estimation

Hard part lengths were converted into body size of whole preyusing regressions determined from own data or the literature(Table A.1 – Supplementary Material). Body lengths were con-verted to dry weight (DW) of fresh, undigested prey, dependingon the regressions available. This conversion was either direct, orvia estimation of the WW first and then converting to DW, either

by another equation or by adjusting for % water content. In theabsence of a known regression for a certain prey, availabledata for the closest taxon or body morphology were used. Whenhard parts were identifiable, but measuring was not possible ornot directly related to body size (e.g., euphausiids’ ommatidia),part counts were converted to biomass using the average DW ofthat taxon, calculated from all stomachs where it occurred withina species of predator (e.g., the average biomass of all euphausiidsidentified). The biomass of unidentifiable material was estimatedfor each predator species by averaging the total DW of stomachswith equal fullness status containing identified prey.

Major prey taxa categories

Prey were grouped into 15 major categories for calculatingthe indices of prey importance: amphipods (AM), cephalopods(CE), chaetognaths (CH), cladocerans (CL), copepods (CO),decapods (DE), euphausiids (EU), fishes (FI), gelatinous plankton(GP), mysids (MY), polychaetes (PO), pteropods (PT), ostracods(OS), unidentified crustaceans (UC) and unidentified material(UM).

Indices of prey importance

For every prey category of each species of predator fish, threeindices were calculated, percentage in number %N, percentage inweight %W and frequency of occurrence %F, as follows:

V. Carmo et al. / Progress in Oceanography 130 (2015) 172–187 175

%Ni ¼total number of prey item i

total number of all prey items� 100

%Wi ¼total dry weight of prey item i

total dry weight of all prey items� 100

%Fi ¼number of stomachs including prey item i

total number of positive stomachs� 100

These indices were combined into an index of relative impor-tance (IRI, Pinkas et al., 1971), calculated for each prey category:

IRIi ¼ ð%Ni þ%WiÞ �%Fi

The latter is represented by the area of a rectangle obtained byplotting the three previous indices on a 3-way graph (Moku et al.,2000). This index is dimensionless so the IRI values per se aredifficult to interpret and used only for illustration of major pat-terns. To allow comparisons between food categories and species,the IRI was standardized to %IRI:

%IRIi ¼IRIi � 100Xn

i ¼ 1IRIi

where n is the total number of food categories to be considered at acertain taxonomic level, and i a specific prey category (Cortés,1997). Prey with a %IRI < 1 were considered rare or occasional preyand were represented within the category ‘‘Others’’.

Statistical analysis

Most of the subsequent analysis was performed on PRIMER 6software for Windows, with the PERMANOVA add-on (Clarke andGorley, 2006). A predator–prey matrix representing biomass(DW) of the lowest possible identified prey taxa (variables)ingested by each fish containing food in the stomach (samples)was built. Samples were standardized by their total. Due to thehigh number of zeros, data transformation was attempted andtested, but since the outcome was similar with or without trans-forming, raw data were used instead. A resemblance matrix forthe samples was calculated using the Bray–Curtis similarity index.The experimental design was unbalanced, so several Type III PER-mutational Multivariate ANalyses Of VAriance (PERMANOVA)(Anderson et al., 2008), were performed, testing the null hypothe-ses that there were no differences in diet within each species as afunction of sample location (MAR-ECO box) and size class. The twofactors were considered as fixed and interactions were excludedfrom the model. Depth stratum and phase of the diel cycle werenot included in this analysis, as there were not enough combina-tions of each factor, and because individual fishes routinelymigrate through multiple depth strata in one diel cycle. Whenthe resulting model for one of the factors revealed a p > 0.25 andthe estimates of components variation were negative, that factorwas excluded and the model run again. A SIMPER analysis wasperformed when PERMANOVA results revealed significance inorder to understand which prey taxa contributed more to the dif-ferences between two groups. Furthermore, a Canonical Analysisof Principal coordinates (CAP) (Anderson et al., 2008) was usedwhen more than two groups were significantly different in orderto evaluate the distinctiveness of these a priori groups in multivar-iate space and the discriminant power of each prey taxa. The role ofthe variables (prey taxa) in separating the groups was visualized bysuperimposing vectors of multiple correlations with the canonicalordination axes.

Ontogenetic trends

In order to detect dietary shifts with fish growth, the lengthdata for the predator fishes were converted into size classes andconsidered as a factor in PRIMER. Due to low sample numbers insome species, only two species-specific size classes (‘‘S’’ smalland ‘‘L’’ large) were defined (Table 1). Size classes were determinedby roughly dividing the size distribution in half, splitting evenlythe number of individuals into those two classes.

Estimation of the instantaneous ration

The instantaneous ration (expressed as the % of body weight afish eats per feeding interval) was determined for each species ofpredator by the ratio between the sum of the DW of all prey itemsand the sum of the DW of all fishes with positive stomachs (i.e.,fishes with empty stomachs were excluded) multiplied by 100.

Rinst ¼P

DW ðprey itemsÞPDW ðpredatorÞ � 100

Daily ration (defined as the % of body weight a fish eats per day)was determined based on the estimated instantaneous rationmultiplied by the number of times a species eats per day. Sincethere were not enough data in this study to cover the whole dielcycle, feeding periodicity was estimated based on available litera-ture. In the absence of previous feeding chronology studies for acertain species, a reference value of 1–4% daily ration for midwaterfishes was applied instead (Feagans-Bartow and Sutton, 2014).

Results

The five fish species examined in this study were collectedprimarily within depth strata 1 and 2 (<750 m), with only 10% ofthe total individuals captured deeper. Sample size (positive, non-empty stomachs) ranged between 50 and 100 individuals, whereasstomach emptiness varied from 0% to 18%, in Argyropelecus aculea-tus and Maurolicus muelleri, respectively (Table 1). The number ofprey items in a stomach ranged between 1 and 291 (the maximumin M. muelleri) and the number of major taxon categories in astomach ranged between 1 and 6 (the maximum in Sternoptyxdiaphana). The average values are given for each predator speciesin Table 1. PERMANOVA test results for diet comparisons betweensize classes and areas of sampling within each species aresummarized in Table 2. The lowest possible prey taxa identifiedin the guts of each species and their indices of prey importanceare listed in the appendix (Table A.2 – Supplementary Material).

Argyropelecus aculeatus

The diet of Argyropelecus aculeatus was dominated byamphipods (66.6%IRI), that together with pteropods and ostracodsconstituted 94.5%IRI of the diet of this species (Fig. 2). Amphipodswere, by far, the most frequent taxon (occurring in 80.0% of thestomachs), pteropods and ostracods were important in terms ofnumbers (together making up for 58.4%N) and frequency of occur-rence (over 1/3 of the stomachs, each), while much rarer prey suchas fish and cephalopods were more relevant in terms of weight. A.aculeatus had the lowest average number of prey per stomach con-taining food of the species studied (less than five food items). Thisnumber did not fluctuate much between samples, with the excep-tion of a few individuals that appeared to have consumed dozens ofpteropods in the same feeding event, the maximum being 29pteropods in one stomach (10 Clio spp. and 19 Limacina inflata).Specimens available were exclusively from the Azores, so dietcomparison was only possible between sizes (Table 2), revealing

Table 1Species data, including the number of fishes analyzed and the size classes attributed to each species in order to test for ontogenetic changes in diet. S – small, L – large.

Predator taxon Speciescode

Sizerange(cm SL)

Size classes Depth range (m) Total no.fish

No.positivestomachs(withfood)

% EmptyStomachs

Avg.stomachcondition(all) ± SD

Avg. positivestomachcondition± SD

Totalno.prey

Max. no.prey perstomach

Avg. no. prey(per positivestomach) ± SD

Avg. no.prey taxacategories(per positivestomach) ± SD

S L

SternoptychidaeArgyropelecus aculeatus Aa 2.2–6.0 <4.00 P4.00 0 to >2300

(bottom)50 50 0.0 3.20 ± 1.05 3.20 ± 1.05 233 29 4.7 ± 5.1 2.0 ± 1.0

Argyropelecushemigymnus

Ah 1.9–3.6 <2.75 P2.75 0 to >2300 87 72 17.2 1.32 ± 0.96 1.60 ± 0.82 377 17 5.2 ± 3.8 2.3 ± 1.0

Maurolicus muelleri Mm 2.1–5.4 <4.30 P4.30 0–2300 122 100 18.0 2.21 ± 1.69 2.70 ± 1.47 1838 291 18.4 ± 37.1 2.0 ± 0.9Sternoptyx diaphana Sd 1.5–4.0 <2.75 P2.75 0–2300 90 77 14.4 2.58 ± 1.72 3.01 ± 1.46 848 126 11.0 ± 17.6 2.7 ± 1.3

PhosichtyidaeVinciguerria attenuata Va 3.3–4.5 <3.80 P3.80 0–750 75 69 8.0 1.85 ± 1.31 2.01 ± 1.25 476 23 6.9 ± 5.4 2.3 ± 1.0

Table 2PERMANOVA test results for possible comparisons of diet with size and location for five stomiiform fish species. Significant results (p < 0.05) presented in bold.

Species Available samples(location)

Possible comparisons considering thenumber of samples per category

N df SS MS Pseudo-F (main tests) ort (pairwise tests)

p (perm) Uniquepermutations

Argyropelecus aculeatus AZ Size (Aa_S, Aa_L) 50 1 9923.3 9923.3 2.787 0.0166 9923

Argyropelecus hemigymnus RR, CG, AZ Size (Ah_S, Ah_L) 72 1 2599.0 2599.0 0.653 0.8191 9935Location (RR + CG, AZ) 72 1 7018.5 7018.5 1.773 0.0421 9931

Maurolicus muelleri RR, CG, FS, AZ Size (Mm_S, Mm_L) 100 1 2401.1 2401.1 0.8045 0.5545 9944Location (RR, CG, FS, AZ) 100 3 60690.0 20230.0 6.778 0.0001 9899Pairwise:

AZ, FS 50 1.601 0.0225 9924AZ, CG 47 1.532 0.0363 9943AZ, RR 43 2.553 0.0002 9930FS, CG 57 2.301 0.0011 9932FS, RR 53 3.006 0.0001 9941CG, RR 50 1.407 0.0827 9943

Sternoptyx diaphana CG, FS, AZ Size (Sd_S, Sd_L) 77 1 12180.0 12180.0 3.274 0.0028 9928Location (FS, AZ) 74 1 4950.5 4950.5 1.331 0.1988 9925

Vinciguerria attenuata FS, AZ Size (Va_S, Va_L) in AZ only 65 1 5065.3 5065.3 1.232 0.2204 9927

176V

.Carmo

etal./Progress

inO

ceanography130

(2015)172–

187

Fig. 2. Plot of the percentage in number %N, percentage in weight %W andfrequency of occurrence %F of the major prey categories in the diet of Argyropelecusaculeatus. The combination of these three indices represents the Index of RelativeImportance (IRI) of each prey, here standardized to %IRI. Prey taxa: AM –amphipods, PT – pteropods, OS – ostracods, FI – fish, CE – cephalopods, OT – others.

V. Carmo et al. / Progress in Oceanography 130 (2015) 172–187 177

significant differences with ontogeny (p = 0.016). These differenceswere mainly related with the proportion of amphipods found inthe stomachs (Table 3), their contribution summing 55.7% of thetotal dissimilarity between the groups. Themisto compressa andPhronima sedentaria occurred mostly in the stomachs of largerindividuals, while non-identified and smaller-sized Hyperiideaand Platyscelus spp. were more important for smaller fish. Ptero-pods, including Clio spp. and Limacina inflata, were predominantin the diet of small-sized specimens. Myodocopid ostracods weremore important as food for smaller size classes, as opposed toHalocypridids that occurred chiefly in stomachs of largerindividuals. Fish prey occurred in the diet of larger-sizedA. aculeatus, and included not only the consumption of otherhatchetfishes of the same genus (A. hemigymnus), but alsodemonstrated cannibalistic behavior. Remarkably all A. aculeatusstomachs contained prey, of which 78.0% were half-full to fulland distended (stomach conditions 3–5). Additionally, conditions

Table 3SIMPER (Similarity Percentages) routine results, showing the prey taxa that contributed moaculeatus (Aa_S and Aa_L) and Sternoptyx diaphana (Sd_S and Sd_L). Average dissimilaritcontributing variables at a cut-off percentage of 90%.

Prey taxon Aa_S Aa_L Contribution % Cumulative % PAverage% biomass

Average% biomass

Themisto compressa 35.07 42.60 27.94 27.94 TPhronima sedentaria 5.42 27.13 17.87 45.81 HHyperiidea 10.25 1.31 6.79 52.60 CClio spp. 8.36 0.41 5.27 57.87 PLimacina inflata 7.32 0.49 4.67 62.54 EOegopsida 3.92 3.71 4.51 67.05 THalocyprididae 1.39 5.16 3.82 70.87 NMyodocopa 5.97 0.22 3.75 74.62 OArgyropelecus aculeatus 0.00 4.00 2.46 77.08 CVinciguerria spp. 4.00 0.00 2.46 79.54 SGelatinous prey 3.64 0.29 2.41 81.95 PArgyropelecus hemigymnus 0.00 3.73 2.30 84.25 HPlatyscelus spp. 3.36 0.00 2.07 86.32 HPteropoda 2.86 0.00 1.76 88.08 PEuphausiidae round-eyed 0.22 2.59 1.71 89.79 EPhysonectae 0.00 2.72 1.68 91.46 E

4 and/or 5 were recorded at all times of capture (Fig. 3). Most preywere well digested, but 42.5% were still in good condition(digestion state 2). Fresh prey were noted only in daytime, bothat <750 m depth (1000 and 1800 h) and near bottom at >2300 m(1700 h). Near bottom individuals (>1500 m) comprised 12 large-sized specimens that were captured throughout the diel cycle(except for 1400 h).

Argyropelecus hemigymnus

Argyropelecus hemigymnus fed largely on copepods (80.8%IRI,93.1%F), complementing its diet with ostracods (8.7%IRI) and occa-sionally fish, chaetognaths and euphausiids (Fig. 4). Copepods andostracods dominated numbers (together 78.0%N) and occurrence,and fish was an important food item in terms of weight(41.5%W). The number of prey items and taxonomic groups foundin positive stomachs did not oscillate much in this species. Nodifferences in diet with size were detected in our dataset (Table 2).Given the low number of fish caught in the northern samples(seven in each), data from RR and CG were pooled together (con-sidering there were no differences between these two areas testeda priori, p = 0.193) and compared with the Azores, which revealed asignificant difference (p = 0.042). Differences in diet with latitudein A. hemigymnus were mainly explained by the proportion of cal-anoids found in the stomachs (Table 4). The summed contributionof the calanoid copepod taxa to dissimilarity between groups wasover 50%. Euchaetidae, namely Paraeuchaeta norvegica and Aetidi-dae were more important in the diet of the northern fish, whilethe genus Pleuromamma, namely P. gracilis, were predominant inthe diet of the southern individuals. Moreover, fish prey prevailedin the north, while halocyprid ostracods, euphausiids (includingNematoscelis megalops) and pteropods were exclusive of the south-ern diet. Regarding stomach fullness, 65.5% were either empty orcontained less than 25% of food (condition 1) and none was filledup completely. Additionally, 69 A. hemigymnus individuals werenot dissected (42% of the available specimens lent by the BergenMuseum) because their stomachs were everted. Fullness washighest at 1800 h, although samples were not available fornighttime hours (Fig. 5). Approximately 71% of the prey were ina very advanced state of digestion. Fresh prey were detected onlybetween 0 and 750 m depth, both at 1800 h and around noon. Indi-viduals from the 1900 h samples were caught at depths of>1500 m, but all others belonged to strata 1 and 2.

st to the dissimilarity between the small (S) and large (L) size classes of Argyropelecusy between the groups = 81.2% and 86.3%, respectively. Results shown for the higher

rey taxon Sd_S Sd_L Contribution % Cumulative %Average %biomass

Average %biomass

hemisto compressa 11.42 32.99 21.11 21.11yperiidea 20.43 11.44 13.83 34.93haetognatha 11.44 8.31 9.12 44.05latyscelus ovoides 8.71 8.45 8.46 52.51uphausiidae 11.25 5.09 8.05 60.56eleostei 2.75 8.59 6.16 66.72ematoscelis megalops 3.60 5.18 4.63 71.36egopsida 2.60 4.69 4.08 75.44alanoida 5.10 0.51 3.18 78.62tylocheiron maximum 2.06 2.64 2.62 81.23latyscelus spp. 3.96 0.00 2.29 83.52alocyprididae 1.01 2.67 2.08 85.61yperia macrocephala 2.37 0.00 1.37 86.98hronima sedentaria 0.50 1.81 1.31 88.29uphausiidae round-eyed 0.42 1.68 1.19 89.48uphausia khroni 0.00 1.74 1.01 90.48

0

2

4

6

8

10

0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5

N

Stomach fullness

0000h 1000h 1400h 1700h 1800h

*

**

*

*

*

Fig. 3. Stomach fullness chronology for Argyropelecus aculeatus. Symbols: daylight, nighttime, ⁄ presence of fresh prey (digestion status 1).

Fig. 4. Plot of the percentage in number %N, percentage in weight %W and frequency of occurrence %F of the major prey categories in the diet of Argyropelecus hemigymnus.The combination of these three indices represents the Index of Relative Importance (IRI) of each prey, here standardized to %IRI. Prey taxa: CO – copepods, OS – ostracods, FI –fish, CH – chaetognaths, EU – euphausiids, OT – others.

178 V. Carmo et al. / Progress in Oceanography 130 (2015) 172–187

Maurolicus muelleri

Maurolicus muelleri fed abundantly on copepods, which com-prised as much as 91.0%IRI of the diet, and dominated in numbers,weight and frequency (85.1%W and 90.0%F). Secondary prey werecladocerans (29.2%N), euphausiids (8.2%W) and ostracods, eachtaxon occurring roughly in 1/4 of the stomachs analyzed (Fig. 6).On average, over 18 prey items were found in each positivestomach of this species (Table 1), the maxima consisting of twoindividuals caught at Faraday Seamount, one with 99 cladoceransand 112 calanoid copepods in its stomach and the other containing150 cladocerans plus 141 calanoids. Cladocerans were exclusive

prey of FS fish (except for one case in the AZ). We did not find evi-dence of changes in feeding with size in M. muelleri, but latitudinaldifferences in diet between all locations studied were highlysignificant, apart from the northern sector (Reykjanes Ridge andCharlie-Gibbs Fracture Zone). In the canonical analysis, CAP1 axisseparated the northern samples (RR and CG) from the southernFS and AZ, while the second axis seemed to isolate FS from theother areas (Fig. 7), which mirrors the PERMANOVA pairwise testresults. In fact, the CAP permutation test was highly significant(tr(Q_m0HQ_m) = 1.129, p = 0.0001; d2

1 ¼ 0:640, p = 0.0001) indicat-ing a strong distinctiveness in diets between geographic areas. Inthe northern areas (RR and CG), large predatory copepods of the

Table 4SIMPER (Similarity Percentages) routine results, showing prey taxa that contributedmost to the dissimilarity in diet between the northern (RR + CG) and southernsamples (AZ) of Argyropelecus hemigymnus. Average dissimilarity between thegroups = 88.2%. Results shown for the higher contributing variables at a cut-offpercentage of 90%.

Prey taxon RR + CG AZ Contribution % Cumulative %Average %biomass

Average %biomass

Calanoida 32.55 15.30 19.49 19.49Teleostei 12.40 7.65 10.15 29.64Pleuromamma spp. 5.60 13.91 9.96 39.60Myodocopa 6.78 8.66 7.68 47.28Euchaetidae 11.58 0.94 6.95 54.23Chaetognatha 8.17 6.80 6.86 61.09Heterorhabdus spp. 4.27 5.58 5.14 66.22Aetideidae 8.09 0.20 4.67 70.89Halocyprididae 0.00 5.77 3.27 74.16Paraeuchaeta norvegica 5.30 0.00 3.00 77.17Euphausiidae 0.00 4.20 2.38 79.55Gelatinous prey 2.49 1.34 2.08 81.62Pteropoda 0.00 3.36 1.90 83.53Pleuromamma borealis 3.03 0.00 1.72 85.24Pleuromamma gracilis 0.00 2.99 1.70 86.94Herbivorous copepod 2.76 0.00 1.56 88.50Sternoptychidae 0.00 1.70 0.96 89.47Nematoscelis megalops 0.00 1.66 0.94 90.41

Fig. 6. Plot of the percentage in number %N, percentage in weight %W andfrequency of occurrence %F of the major prey categories in the diet of Maurolicusmuelleri. The combination of these three indices represents the Index of RelativeImportance (IRI) of each prey, here standardized to %IRI. Prey taxa: CO – copepods,CL – cladocerans, EU – euphausiids, OS – ostracods, OT – others.

V. Carmo et al. / Progress in Oceanography 130 (2015) 172–187 179

family Euchaetidae, namely Paraeuchaeta norvegica, andeuphausiids prevailed in the diet (Fig. 7a). Pteropods were morecorrelated with feeding at Faraday Seamount while the copepodsCandacia spp. and Pleuromamma spp. had the greatest correlationswith the southern box (AZ). Significant differences in diet withontogeny were not detected. However, all specimens from theAzores belonged to the smaller size class and this might bereflected in the correlation of small-sized copepods with this areawhen compared with the Euchaetidae, including Paraeuchaetanorvegica, from the northern boxes (RR and CG) (Fig. 7b). Stomachfullness was distributed evenly among the five fullness categories.Approximately 14% of the stomachs were completely filled(condition 5), but �74% of all the prey were highly digested.Looking at the chronology of stomach fullness (Fig. 8), at 0600 hand 1600 h, stomachs were either empty or practically void, withhighly digested items (condition 3), as opposed to a series ofincreasingly filled stomachs containing fresh prey or nearly so(condition 1 or 2). At 1000, 1200 and 1800 h, stomach fullnesswas generally low and all prey were half-digested or unrecogniz-able. The 11 individuals captured at 1300 h belonged to deep sam-ples (>1500 m, at FS), while the remaining specimens were caughtbetween 0 and 750 m.

0

2

4

6

8

10

0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5

*

*

0 1 2 3Stomach f

N 0400h 1000h 1400h0600h 1200h

*

Fig. 5. Stomach fullness chronology for Argyropelecus hemigymnus. Symbo

Sternoptyx diaphana

Sternoptyx diaphana fed predominantly upon amphipods andeuphausiids (89.8%IRI), frequently supplemented with ostracods,copepods and chaetognaths (occurring in nearly 1/3 of the stom-achs each). Cephalopods occurred quite rarely, but were importantin terms of weight (Fig. 9). The stomachs of this species weregenerally characterized by a high number of prey (11, on average,but variable), often belonging to several different taxa (Table 1).The extreme cases were two individuals from the AZ, one contain-ing 126 amphipods in its stomach and another holding 75 preybelonging to five different major prey categories (mostly amphi-pods, but also a few euphausiids, copepods, ostracods and fish).S. diaphana had the highest number of prey categories per positivestomach of all species studied (average and maximum). Of the S.diaphana specimens analyzed there were differences in feedingwith ontogeny. Given the very few positive stomachs from theCG area, spatial comparison was only possible between FS andAZ, and the test was not significant. Ontogenetic variation in thediet of S. diaphana was based mainly in the different proportionof amphipods in the diet, between size classes (52.2% of the totaldissimilarity between groups) (Table 3). Themisto compressa and

4 5 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5ullness

1700h 1800h1500h 1900h

*

ls: daylight, f sunset, ⁄ presence of fresh prey (digestion status 1).

Fig. 7. Canonical Analysis of Principal coordinates (CAP) of the spatial variability inthe diet of Maurolicus muelleri, showing prey taxa with multiple correlations > 0.3.Samples labeled according to MAR-ECO box sites (a) and fish size classes (b).

180 V. Carmo et al. / Progress in Oceanography 130 (2015) 172–187

Phronima sedentaria were more important for larger individualsthan the other hyperiids (including Platyscelus and Hyperia macro-cephala). Non-identified euphausiids including larval forms (suchas furcillia stages) dominated in the diet of smaller fish, while moredeveloped forms (both bilobed and round eyed species) occurred

0

2

4

6

8

10

12

14

0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5

N 0600h 1000h 1200h

*

Stomach f

Fig. 8. Stomach fullness chronology for Maurolicus muelleri. Symb

mainly in larger fish stomachs. Also, large S. diaphana individualspreyed more on ostracods (Halocyprididae), cephalopods(Oegopsida) and fish (Teleostei), when compared with smaller sizeclasses that preyed more on copepods and chaetognaths. Roughlyhalf (47.8%) of S. diaphana stomachs were more than 50% full(conditions 3–5), and �18% were fully extended. This species hadthe highest percentage of fresh prey of all species studied(14.0%). There were full and/or distended stomachs (conditions 4and 5) at all times and depths, as well as fresh prey throughoutthe diel cycle at most times and all depths (Fig. 10). Of the 77individuals examined in this study, most were captured between0 and 750 m, while 18 were caught at depths between 750 and2300 m (three near the bottom).

Vinciguerria attenuata

The stomachs of Vinciguerria attenuata primarily containedcopepods and ostracods (86.3%IRI), and to a lesser extentamphipods and euphausiids (Fig. 11). Copepods dominated allparameters (%N, %W, %F) occurring in 91.3% of the stomachs ana-lyzed. Ostracods followed in occurrence (65.2%F) and numbers(25.2%N) and amphipods in weight (33.6%W). The average numberof prey items in the stomach of this species was approximatelyseven, usually belonging to two or three major taxa. There wasonly one positive specimen of V. attenuata from FS, and therefore,differences between sizes could only be tested in the AZ area,and these were not significant. Most (74.7%) of the stomachs wereless than half full, but of these only 8% were empty. Most preyitems were highly digested. The individuals caught at 0300 h hadlow stomach fullness and all prey were highly digested (Fig. 12).The highest stomach fullness (4 and 5) and least digested prey (sta-tus 1 or 2) belonged to fish captured at daylight, between 1000 and1800 h. All fish were caught at <750 m.

Instantaneous ration and daily ration

The instantaneous ration calculated for each species, andexpressed as percentage of body weight (dry weight biomass,DW) was compared with the reference 1–4% daily ration formidwater fishes (Feagans-Bartow and Sutton, 2014) and is pre-sented in Table 5. Estimated instantaneous rations fell within thereference daily ration interval for all species, with the remarkableexception of S. diaphana, which was an order of magnitude higherthan all others.

0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5

1300h 1800h1600h

**

*

**

ullness

ols: daylight, ⁄ presence of fresh prey (digestion status 1).

Fig. 10. Stomach fullness chronology for Sternoptyx diaphana. Symbols: daylight, nighttime, ⁄ presence of fresh prey (digestion status 1).

Fig. 9. Plot of the percentage in number %N, percentage in weight %W and frequency of occurrence %F of the major prey categories in the diet of Sternoptyx diaphana. Thecombination of these three indices represents the Index of Relative Importance (IRI) of each prey, here standardized to %IRI. Prey taxa: AM – amphipods, EU – euphausiids, OS– ostracods, CO – copepods, CH – chaetognaths, CE – cephalopods, OT – others.

Fig. 11. Plot of the percentage in number %N, percentage in weight %W and frequency of occurrence %F of the major prey categories in the diet of Vinciguerria attenuata. Thecombination of these three indices represents the Index of Relative Importance (IRI) of each prey, here standardized to %IRI. Prey taxa: CO – copepods, OS – ostracods, AM –amphipods, EU – euphausiids, OT – others.

V. Carmo et al. / Progress in Oceanography 130 (2015) 172–187 181

0 1 2 3 4 5 0

2

4

6

8

10

12

14

0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5

N 0300h 1000h 1200h 1400h 1800h1700h

Stomach fullness

Fig. 12. Stomach fullness chronology for Vinciguerria attenuata. Symbols: daylight, nighttime, ⁄ presence of fresh prey (digestion status 1).

Table 5Estimates of instantaneous rations and potential periodicity in feeding when compared with reference daily rations (average mass of all meals eaten per day in relation to fishmass) for midwater fishes (1% and 4%). DW = dry weight biomass.

Predator species Total DWof predatorswith positivestomach (g)

Total DWof prey (g)

Instantaneousration(% of bodyweight)

Periodicity in feeding based on a reference daily ration

1% 4%

Number oftimes a fisheats per day

Intervalbetweenmeals(approx. hours)

Periodicityin feeding(approx. days)

Number oftimes a fisheats per day

Intervalbetweenmeals(approx. hours)

Periodicityin feeding(approx. days)

Argyropelecus aculeatus 27.97 1.02 3.64 0.3 87 1�/3.5 days 1.1 22 1�/dayArgyropelecus hemigymnus 8.44 0.14 1.68 0.6 40 1�/1.5 days 2.4 10 2–3�/dayMaurolicus muelleri 29.90 0.82 2.74 0.4 66 1�/2.5 days 1.5 16 1.5�/daySternoptyx diaphana 17.04 2.01 11.82 0.1 284 1�/12 days 0.3 71 1�/3 daysVinciguerria attenuata 6.81 0.17 2.54 0.4 61 1�/2.5 days 1.6 15 1.5�/day

182 V. Carmo et al. / Progress in Oceanography 130 (2015) 172–187

Discussion

Argyropelecus aculeatus

Earlier studies in the Atlantic and the Pacific (Table 6) foundthat smaller-sized Argyropelecus aculeatus fed mainly on ostracodsand copepods and larger individuals on pteropods, euphausiids andfish, while amphipods were relatively rare food items. In contrast,in our analysis (AZ area), hyperiid amphipods were the major foodcomponent, copepods were consumed in low amounts by smallerfish (<40 mm), and were absent in the stomachs of larger individ-uals (>40 mm). The diet was complemented with pteropods(namely Limacina) and conchoecid ostracods by the smaller sizeclasses, and by fish in larger specimens, taxa common with earlierstudies. Similarly, cannibalism was previously reported in this spe-cies by Hopkins and Baird (1985). In other areas, A. aculeatus was anighttime or dusktime feeder (eastern Gulf of Mexico – EGoM:Hopkins and Baird, 1985; NE Atlantic: Merrett and Roe, 1974).Our data suggest A. aculeatus may be a daytime feeder in theAzores area, but given the major dusk and nighttime gaps in oursampling, we may have missed another feeding period in the dark.Cyclic feeding once a day is consistent with a daily ration ofapproximately 3.5–4% (Table 5).

Argyropelecus hemigymnus

The Argyropelecus hemigymnus diet consisted primarily ofcopepods (mainly calanoids, including the genus Pleuromamma),ostracods and fish, which is consistent with prior reports for the

species (Table 6). The diet varied mainly in the proportion of cala-noid copepod taxa ingested among the northern and southernstudy areas. Feeding chronology studies have reported that A.hemigymnus feeds predominantly during the day (afternoon or lateafternoon) and evening (EGoM: Hopkins and Baird, 1985; NEAtlantic: Merrett and Roe, 1974) with a decline in stomach fullnessthroughout the night (Clarke, 1978), which seems consistent withour data. Cyclic feeding once a day, according to our estimates ofinstantaneous ration, would correspond to 1.5–2% daily ration forthis species (Table 5), but if some regurgitation occurred (see Sec-tion ‘Sources of error’), a more reasonable estimate should be >2%.

Maurolicus muelleri

Maurolicus muelleri fed predominantly on calanoid copepodsand euphausiids, as recurrently reported for this species (Table 6),but cladocerans were also important (%N and %F) for larger sizedindividuals (only at FS). Cladocerans are fast-growing opportuniststhat may display strong population increases under favorableenvironmental conditions, usually in summer, after the springbloom, namely in the Azores area (Carmo et al., 2013). Given theexceptionally high numbers of cladocerans found in somestomachs, it is possible there was a swarm in FS during thesampling period. These crustaceans were typical of the epipelagicassemblages (0–200 m) of the N MAR, between stations 10 and36 (see Fig. 1; MAR-ECO unpublished data). Cladocerans werefound in the stomachs of fish caught at 0–200 m in early morning(0600 h), and then much deeper (1500–2300 m) at midday(1300 h), which suggests strong vertical migration by M. muelleri.

Table 6Results of the present study compared with previous work on the same species for different locations.

SpeciesSize classes(this study)

SL (cm)Dominant prey (secondary prey in parentheses) Location

Major prey categories

Present study (%IRI) Previous studies

Argyropelecusaculeatus

Aa_S <4.00Amphipods, pteropods,(ostracods)

Ostracodsa,b,c,d; copepodsa,b,c; (euphausiidsb;pteropodsa,b; decapod larvaea)

aWN Atlantic; aE PacificEquatorial; a,b,dE Gulf of Mexico;a,cNE AtlanticAa_L P4.00 Fishes, amphipods Fishesa,b,d; pteropodsa,b,d; euphausiidsa,b; (cephalopodsd)

Argyropelecushemigymnus

Ah_S <2.75Copepods, ostracods, fish(euphausiids, chaetognaths)

Copepodsa,b,c,e,f,g,h; ostracods a,b,c,d,f,g; fish e; chaetognathsa,h; (euphausiids e; salps f)

aWN Atlantic; aN Pacific; aPacificSubantarctic; aEN Pacific Central;a,b,dE Gulf of Mexico;eMediterranean; a,c,f,g,hNE AtlanticAh_L P2.75

Fish, copepods (chaetognaths,euphausiids, ostracods)

Copepods a,b,e,f,g,h; Ostracods a,b,d,f,g; fishe; chaetognathsa,g,h; (euphausiids e; salps f; pteropods e,g)

Maurolicusmuelleri

Mm_S <4.30 Copepods, euphausiids Copepodsd,g,i,j,k,m,n; euphausiidsd,j,k,m,n; decapodlarvaem; cladoceransk,l; veliger bivalvial; (ostracodsi;amphipodsg, i,k; pteropodsk; eggsg)

g,j,l,mNE Atlantic; iN Red Seaand Gulf of Aden; kJapan Sea; nETasmania; dE Gulf of MexicoMm_L P4.30 Copepods, euphausiids

(cladocerans)Euphausiidsj,n; copepodsj,n; cladoceransl; veligerbivalvial

Sternoptyxdiaphana

Sd_S <2.75

Amphipods, euphausiids,(ostracods, copepods,chaetognaths)

Copepodsa,b,f,o, amphipodsa,b,d,f,o; ostracodsb,f,o;euphausiidsa,f,o; (chaetognathsa, b,o; decapodab,o;polychaetab)

a,oWN Atlantic; a,oSE Atlantic;a,oVenezuelan-Caribbean; 1,15NWAtlantic Pocket; b,dE Gulf ofMexico; f NE Atlantic and Centralequatorial Atlantic; o PacificSubantarcticSd_L P2.75

Amphipods, euphausiids,(ostracods, copepods,chaetognaths)

Amphipodsa,b,d,o; euphausiidsa,b,f,o; fisha,n; copepodsa,f,o; (chaetognathsb,o; decapodab,o; polychaetab)

Vinciguerriaattenuata

Va_S<3.80 Copepods, amphipods,

euphausiids Copepodsd dE Gulf of MexicoVa_L P3.80 Amphipods, copepods, ostracods

References:a Hopkins and Baird (1977).b Hopkins and Baird (1985).c Merrett and Roe (1974).d Hopkins et al. (1996).e Jespersen (1915).f Kinzer and Schulz (1988).g Mauchline and Gordon (1983).h Roe and Badcock (1984).i Dalpadado and Gjøsæter (1987).j Gjøsæter (1981).k Ikeda et al. (1994).l Rasmussen and Giske (1994).

m Samyshev and Schetinkin (1973).n Young and Blaber (1986).o Hopkins and Baird (1973).

V.Carm

oet

al./Progressin

Oceanography

130(2015)

172–187

183

184 V. Carmo et al. / Progress in Oceanography 130 (2015) 172–187

Cladocerans were formerly described as food items of this speciesby Ikeda et al. (1994) in the southern Japan Sea and Rasmussen andGiske (1994) in the NE Atlantic (Masfjorden, W Norway), both dur-ing the summer period.

There were significant differences in diet between sites, withfeeding in the two northern areas (RR and CG) being similar, butquite different from the FS and the AZ areas. This was mainlyrelated to the proportion of calanoid copepods in the diets, as seenin A. hemigymnus. Fish from the AZ region exhibited some morpho-logical differences compared to specimens caught farther north,and this was reflected in the size frequency distribution. Parinand Kobyliansky (1996) suggested the presence of two differentspecies of Maurolicus in our study area: Maurolicus amethystino-punctatus, occurring from the Mediterranean to the adjacent east-ern N Atlantic (Portugal to Senegal) including the Azores, while the‘‘true’’ M. muelleri form occurred in the eastern N Atlantic, above40�N. All specimens from the Azores area used in this study fitthe morphological description (namely gill rakers count and pig-mented intestine) of the deep-bodied M. amethystinopunctatus,while the remaining individuals belonged to the slender-bodiedM. muelleri form. Despite these morphological differences, recentgenetic analyses (Rees et al., unpublished data) have concludedthat these are morphotypes of a single species, M. muelleri, whichis the taxonomic identity we have adopted in this paper. However,as a result of these morphological differences, all of the short-bod-ied specimens from the AZ were placed in the small size class,while individuals from the RR, CG and FS were placed in both largeand small size classes. It is possible that the detected spatial differ-ences in diet between the AZ and the northern areas could berelated to ecophenotypic variation, an ‘‘oceanic species concept’’question (sensu Gibbs, 1986) that warrants further investigation.

Earlier reports of the diel feeding rhythm of M. muelleri areinconsistent; both nighttime feeding (Young and Blaber, 1986;Rasmussen and Giske, 1994) and no diel feeding rhythm(Samyshev and Schetinkin, 1973; Gjøsæter, 1981; Mauchline andGordon, 1983) have been reported. Our data suggest that therecould be at least two main periods of feeding for this species alongthe N MAR, one during the night/early morning and another inmid-afternoon, with asynchronous feeding by individuals withinthe population (possibly segregating by depth and/or size). If feed-ing occurs twice a day, daily ration would be closer to 5.5% (twicethe instantaneous ration). A daily ration of 4% would be consistentwith one primary meal per day, with snacking in between the mainfeeding periods (Table 5).

Sternoptyx diaphana

Sternoptyx diaphana relied mainly on amphipods (Hyperiidea)and euphausiids (generally bilobed-eye species) as prey, but itsdiet was the most diversified of the species studied in terms ofmajor prey taxa, and included secondary prey such as ostracods,copepods and chaetognaths. Copepods, nevertheless, were the pri-mary prey items in the diet of this species reported in most of theprior studies, especially for smaller individuals, while fish preywere important for larger size classes (Table 6). Hyperiideanamphipods (namely Platyscelus) and bilobe-eyed euphausiids(e.g., Nematoscelis and Stylocheiron) are reported as prey in thisand prior studies. Diet changed with ontogeny, namely the propor-tion of amphipods consumed among size classes. Fresh prey andhigh stomach fullnesses were detected at all times of sampling(despite the absence of data between 1900 and 2300 h) probablyindicating that S. diaphana foraged throughout the diel cycle. Thispattern is consistent with that reported by Hopkins and Baird(1985), who found no evidence of a diel feeding rhythm for thisspecies, supported by the observation that S. diaphana has a largeswim bladder, but poorly developed gas gland, typical of species

that do not migrate. Assuming 4% as an upper reference dailyration of midwater fishes, this species would feed only once every3 days (see Table 5), which would be aberrant for a zooplanktivo-rous fish. Given the high percentage of fresh prey and distendedstomachs throughout the day, a daily ration based on thecalculated instantaneous ration could be at least 12% (one mealper day). The pattern observed may reflect an ‘‘opportunisticsnack’’ feeding habit where large numbers of prey are ingested sev-eral times during the day, whenever there is a prey encounter.

Vinciguerria attenuata

Vinciguerria attenuata’s main prey items were copepods(especially calanoids, including Pleuromamma), conchoecidostracods and amphipods. A diet essentially consisting of copepods(61–80%) was reported by Hopkins et al. (1996). We did not findany previous feeding chronology studies for this species, butVinciguerria nimbaria and Vinciguerria poweriae have been shownto be strong vertical migrators and daytime feeders off Hawaii(Clarke, 1978, 1982) and the former also in the tropical Atlantic(Champalbert et al., 2008). Our results suggest a similar pattern,and assuming a cyclic diel rhythm for the AZ area, a realistic dailyration for V. attenuata would be close to 2.5% (Table 5). Legand andRivaton (1969) estimated a daily ration of approximately 5% forV. nimbaria in the E Indian Ocean.

Sources of error

Several considerations should be made regarding the validity ofour results.

For biomass estimations, prey dry weight was calculated basedon length data instead of the direct measure of the prey WW. Thisprocedure is intended to reduce the error derived from the level ofdigestion, although the estimated biomass is not necessarily pre-cisely equivalent to dry weight of living organisms, as bodyshrinkage and losses of organic matter and salts may have resultedfrom storage in preservatives (Ahlstrom and Thrailkill, 1963;Giguère and St-Pierre, 1989; Kuhlmann et al., 1982; Põllupüü,2007). Prey sizes determined in case of fragmentation were conser-vative. Also, when paired or multiple structures of approximatelythe same size were found they counted as the minor number ofprey possible. The estimations shown here should, therefore, beinterpreted as a minimum estimation of the prey size, numberand biomass ingested by these fishes.

A high percentage of A. hemigymnus had everted stomachs andwere removed from the analysis, which may to some extent reflectregurgitation of a portion of their meal during hauling of thesampler. Also some zeros might be false negatives, influencingthe results. A. hemigymnus had the lowest stomach condition,and no full and distended stomachs (condition 5) were found,hence the instantaneous ration was quite low when comparedwith the other sternoptychids. Stomach eversion was previouslyreported in A. hemigymnus in the central equatorial Atlantic byKinzer and Schulz (1988) and in the NE Atlantic by Mauchlineand Gordon (1983), but these authors did not consider it verysignificant, although Jespersen (1915) found it quite common inthe Mediterranean. This phenomenon is likely a gear issue.Rectangular trawls are pulled slowly, and thus eversion issomewhat rare (Sutton, unpublished data). Dual warp trawls, onthe other hand, are much more rapidly hauled in, and fishes areprobably killed on contact, eliminating the possibility of gasreabsorption on the way up, which was probably the case for thisspecies. Thus, results presented here for A. hemigymnus must beinterpreted with some caution. If regurgitation did occur andaffected all prey types in the same way, the described patterns offeeding should be representative, even though all quantities might

V. Carmo et al. / Progress in Oceanography 130 (2015) 172–187 185

be underestimated. This is not likely to have occurred in the otherspecies studied, because everted stomachs were virtually absent.Whenever a stomach was exposed or ruptured that specimenwas not considered in the analysis.

Net feeding is another possible source of bias, although mostmidwater fish diet studies consider it negligible (e.g., Hopkinsand Baird, 1977; Roe and Badcock, 1984; Sutton, 2005). Hopkinsand Baird (1985) and Hopkins et al. (1996) dismissed this kind ofcontamination for all sternoptychids investigated here. Further,reports of intensive net feeding on our target fish species are notpresent in the literature. Given the frequently poor external condi-tion of the specimens examined, which suggests the animals wereunder severe stress or even dead shortly after capture, it isreasonable to assume they were not actively feeding, as suggestedby others (e.g., Collard, 1970; Hopkins and Baird, 1973, 1975;Sutton, 2005). Extensive net feeding is even more unlikely judgingby the advanced state of digestion of most prey. Discrete-depthsampling also decreases exposure of fish to potential prey andthe risk of contamination (Lancraft and Robison, 1980). Reflexivegulping obviously occurred in a few cases, but it was easily detect-able by the presence of fresh prey in the mouth or esophagus, andthese were removed from the analysis. Individuals with rupturedbodies and guts, especially the case of many M. muelleri, were alsoexcluded.

As in all diet studies, differences in digestibility may causecertain taxa to stand out more than others, because their hardparts resist digestion. The luminous metasomal organ of Pleuro-mamma copepods is an excellent example, this black spot beingeasily identified in the guts even when all other traces of theanimal were gone. Gelatinous prey, on the other hand, are oftenunidentifiable in the stomachs, especially after chemicalpreservation, and are therefore potentially underestimated.Nevertheless, no cnidocytes were detected in the microscopicexamination of the glass slides, so these fish species are notexpected to feed extensively at least on cnidarians. Stable isotopeanalysis and genetic studies can complement stomach contentanalysis in this regard.

This multidisciplinary study of the N MAR was not designed totest for feeding chronology, since it was not possible to sample theentire diel cycle in all locations and for all species. Nevertheless,the available data permitted some consideration of this aspect,resulting in a preliminary attempt to describe the probable feedingcycles of our study subjects along the N MAR.

General patterns

Dominant prey groups were in general agreement with previ-ous diet studies of the species, except for A. aculeatus and S. diaph-ana. The ontogenetic differences detected in A. aculeatus and S.diaphana were more related with the quantity of a certain taxawithin a prey type (most notably the amphipods), rather than a fullswitch of diet, although with increasing size the former replacedpteropods with fish, while the latter complemented its diet withlarger amounts of fish and cephalopods. The spatial differencesnoted in A. hemigymnus and M. muelleri were associated with thedifferent densities of calanoid taxa, their chief prey, in the differentsampling areas.

Feeding selectivity was apparent in all species, given the differ-ent proportion of zooplankters in the stomachs, when comparedwith potential prey of comparable size present in abundance inthe water column. Amphipods, the main prey taxa of A. aculeatusand S. diaphana, were far less abundant than copepods, which wererarely preyed upon. Along the N MAR, the average amphipod/cope-pod ratio was 1:8 in the epipelagic and 1:15 at 200–1500 m depth(MAR-ECO unpublished data). Other key elements of the zooplank-ton that were merely taken as secondary prey by these two species,

such as euphausiids, ostracods and chaetognaths, were also morenumerous than amphipods in the zooplankton assemblages ofthe N MAR. While it is possible that the gear used for MARmesozooplankton studies (a Multinet, 0.25 m2, 180 lm) mayunderestimate the amphipod prey present in the water column,the abundance of this taxon would not likely approach that ofcopepods, the major net zooplankton components in the WorldOcean (Mauchline, 1998). Regarding the copepod-eaters, selectiv-ity was also noticeable, calanoids being preferred in relation topoecilostomatoids and cyclopoids, the latter only rarely consumed,but all co-occurring in abundances of the same order of magnitudealong the N MAR, from the surface down to 2500 m (MAR-ECOunpublished data). Concerning the calanoids, the differences indiet with latitude detected in M. muelleri and A. hemigymnusreflected the availability of the prey species along the N MAR, asdetermined by Gaard et al. (2008), where the copepod assemblagesin RR and CG were grouped together and were significantly differ-ent from the FS and AZ assemblages. In the Azores, the prevalenceof Pleuromamma spp. in the diet of A. hemigymnus and M. muelleriand also Candacia spp. in the latter, is explained by the higherabundance of these copepods in that area, while the Euchaetidae,namely Paraeuchaeta norvegica, an artic-boreal species, predomi-nated in the northern areas (RR and CG according to Gaard et al.,2008) (see Table 4 and Fig. 4). Selective feeding might be drivenby competition with other species that co-occur in the first severalhundred meters of the water column, like the myctophids and evenother Stomiiformes, the major biomass and numerical componentsof the mesopelagic layer along the N MAR (Sutton et al., 2008).

In general, two main feeding patterns were detected in the spe-cies studied: the copepod-eaters (A. hemigymnus, M. muelleri, andV. attenuata), which exhibited spatial differences in diet along theN MAR, following the calanoid copepod distribution patterns(except for V. attenuata, which was only caught in the AZ); andthe amphipod-eaters (A. aculeatus and S. diaphana), caught onlyin the southern areas (FS and AZ), where diversity is higher notonly in prey species but also in competitors, perhaps leading tospecialization in feeding on a less abundant zooplanktonic group.

The estimated instantaneous rations fall within the values cal-culated by Sutton and Hopkins (1996) for stomiids that feed onmicronekton/copepods (Rinst = 2.1%) or micronekton/euphausiids(Rinst = 4.4) in the eastern Gulf of Mexico and are comparable tothe daily rations estimated by Hopkins and Baird (1981) for thesternoptychid Valenciennellus tripunctulatus (3.7–5.6%) in the Gulfof Mexico and Caribbean. For the species studied, the upper inter-val of 4% reference daily ration for midwater fishes seems reason-able in most cases, given the low vacuity index, the generally highstomach fullness, and the number of prey consumed per feedingevent. The large exception was S. diaphana, which exhibited a Rinst

(=11.85%), an estimate larger than the myctophid-eating stomiidsin Sutton and Hopkins (1996) study (Rinst = 7.6%). These resultssuggest that more physiological and life history studies are neededfor this and related species in order to explain this exceptionalfinding.

Acknowledgments

The authors would like to thank the crew and scientific mem-bers of the R/V G.O. Sars for handling and processing the samples,Ingvar Byrkjedal and Gunnar Langhelle for their precious assis-tance at the Bergen Museum, Thomas Wenneck and Filipe Porteirofor their thorough explanations of the MAR-ECO database andmethodologies used in the cruise, the later also for the taxonomicvalidations of some species and Ricardo Medeiros for the maps.This work was funded by the MAR-ECO project and by‘‘CONDOR—Observatory for Long-Term Study and Monitoring ofAzorean Seamount Ecosystems’’ project, co-financed by the EEA

186 V. Carmo et al. / Progress in Oceanography 130 (2015) 172–187

Grants Financial Mechanism—Iceland, Liechtenstein and Norway(PT0040/2008). The first author was supported by a PhD Grant(SFRH/BD/31693/2006) of the Portuguese Foundation for Scienceand Technology (FCT) and by a Grant of the Regional Governmentof the Azores, the Estagiar-L Program (FSE/PROEMPREGO) co-funded by DOP & Center of IMAR of the University of the Azores.We also acknowledge FCT – PEst/OE/EEI/LA0009/2011–2014 – Pro-ject LA 9 (LARSyS). IMAR-DOP/UAc is Research and DevelopmentUnit # 531 and LARSyS – Associated Laboratory #9 partially fundedby the Portuguese Foundation for Science and Technology (FCT)through pluriannual and programmatic funding schemes (OE,COMPETE, FEDER, POCI2001, FSE) and by the Azores Directoratefor Science and Technology (DRCT). This project was also sup-ported by a Grant to T. Sutton by the NSF Ocean Sciences Division– Biological Oceanography Program (OCE 0623551).

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.pocean.2014.11.003.

References

Ahlstrom, E.H., Thrailkill, J.R., 1963. Plankton volume loss with time of preservation.California Co-operative Oceanic Fisheries Investigations Reports 9, 57–73.

Anderson, M.J., Gorley, R.N., Clarke, K.R., 2008. PERMANOVA+ for PRIMER: Guide toSoftware and Statistical Methods. PRIMER-E, Plymouth, UK.

Angel, M.V., 1997a. Pelagic biodiversity. In: Ormond, R.F.G., Gage, J.D., Angel, M.V.(Eds.), Marine Biodiversity: Patterns and Processes. Cambridge University Press,U.K., pp. 35–68.

Angel, M.V., 1997b. What is the deep sea? In: Randall, D.J., Farrell, A.P. (Eds.), Deep-Sea Fishes. Academic Press, San Diego, pp. 2–42.

Appelbaum, S., 1982. Studies on food organisms of pelagic fishes as revealed by the1979 North Atlantic Eel Expedition. Helgoländer Meeresuntersuchungen 35,357–367.

Balanov, A.A., 1994. Diet of common mesopelagic fishes of the Bering Sea. Journal ofIchthyology 34, 73–82.

Balanov, A.A., Gorbatenko, K.M., Gorelova, T.A., 1994. Daily feeding dynamics ofmesopelagic fishes in the Bering Sea during summer. Journal of Ichthyology 34,85–99.

Bergstad, O., Godø, O.R., 2003. The pilot project ‘‘Patterns and processes of theecosystems of the northern Mid-Atlantic’’: aims, strategy and status.Oceanologica Acta 25, 219–226. http://dx.doi.org/10.1016/S0399-1784(02)01203-3.

Butler, M., Bollens, S.M., Burkhalter, B., Madin, L.P., Horgan, E., 2001. Mesopelagicfishes of the Arabian Sea: distribution, abundance and diet of Chaulioduspammelas, Chauliodus sloani, Stomias affinis, and Stomias nebulosus. Deep SeaResearch Part II: Topical Studies in Oceanography 48, 1369–1383.

Carmo, V., Santos, M., Menezes, G.M., Loureiro, C.M., Lambardi, P., Martins, A., 2013.Variability of zooplankton communities at Condor seamount and surroundingareas, Azores (NE Atlantic). Deep Sea Research Part II: Topical Studies inOceanography 98, 63–74. http://dx.doi.org/10.1016/j.dsr2.2013.08.007.

Champalbert, G.A., Kouamé, B., Pagano, M., Marchal, E., 2008. Feeding behavior ofadult Vinciguerria nimbaria (Phosichthyidae), in the tropical Atlantic (0�–4�N,15�W). Marine Biology 156, 79–95. http://dx.doi.org/10.1007/s00227-008-1067-z.

Clarke, T.A., 1978. Diel feeding patterns of 16 species of mesopelagic fishes fromHawaiian waters. Fishery Bulletin 76, 495–513.

Clarke, T.A., 1980. Diets of fourteen species of vertically migrating mesopelagicfishes in Hawaiian waters. Fishery Bulletin 78, 619–640.

Clarke, T.A., 1982. Feeding habits of stomiatoid fishes from Hawaiian waters.Fishery Bulletin 80, 287–304.

Clarke, K.R., Gorley, R.N., 2006. PRIMER v6: User Manual/Tutorial. PRIMER-E,Plymouth, UK.

Collard, S.B., 1970. Forage of some eastern Pacific midwater fishes. Copeia 1970,348–354.

Cortés, E., 1997. A critical review of methods of studying fish feeding based onanalysis of stomach contents: application to elasmobranch fishes. CanadianJournal of Fisheries and Aquatic Sciences 54, 726–738. http://dx.doi.org/10.1139/cjfas-54-3-726.

Dalpadado, P., Gjøsæter, J., 1987. Feeding ecology of the lanternfish Benthosemapterotum from the Indian Ocean. Marine Biology 99, 555–567.

DeWitt Jr, F.A., Cailliet, G.M., 1972. Feeding habits of two bristlemouth fishes,Cyclothone acclinidens and C. signata (Gonostomatidae). Copeia 1972, 868–871.

Feagans-Bartow, J.N., Sutton, T.T., 2014. Ecology of the oceanic rim: pelagic eels askey ecosystem components. Marine Ecology Progress Series 502, 257–266.http://dx.doi.org/10.3354/meps10707.

Gaard, E., Gislason, A., Falkenhaug, T., Søiland, H., Musayeva, E.I., Vereshchaka, A.,Vinogradov, G.M., 2008. Horizontal and vertical copepod distribution andabundance on the Mid-Atlantic Ridge in June 2004. Deep Sea Research Part II:Topical Studies in Oceanography 55, 59–71. http://dx.doi.org/10.1016/j.dsr2.2007.09.012.

Gaskett, A.C., Bulman, C.M., He, X., Goldsworthy, S.D., 2001. Diet composition andguild structure of mesopelagic and bathypelagic fishes near Macquarie Island,Australia. New Zealand Journal of Marine and Freshwater Research 35, 469–476. http://dx.doi.org/10.1080/00288330.2001.9517016.

Gibbs Jr., R.H., 1986. The stomiod fish genus Eustomias and the oceanic speciesconcept. In: Pierrot-Bults, A.C., van der Spoel, S., Zahuranec, B.J., Johnson, R.K.(Eds.), Pelagic Biogeography, vol. 49. UNESCO Technical Papers in MarineScience, pp. 98–103.

Giguère, L., St-Pierre, J., 1989. Can we estimate the true weight of zooplanktonsamples after chemical preservation? Canadian Journal of Fisheries and AquaticSciences 46, 522–527.

Gjøsæter, J., 1981. Life history and ecology of Maurolicus muelleri (Gonostomatidae)in Norwegian waters. Fiskeridirektoratets Skrifter, Serie Havunders¢kelser 17,109–131.

Gordon, J.D.M., Nishida, S., Nemoto, T., 1985. The diet of mesopelagic fish from thePacific coast of Hokkaido, Japan. Journal of Oceanographical Society of Japan 41,89–97. http://dx.doi.org/10.1007/BF02109178.

Herring, P., 2002. The Biology of the Deep Ocean. Oxford University Press, Oxford,pp. 314.

Hopkins, T.L., Baird, R.C., 1973. Diet of the hatchetfish Sternoptyx diaphana. MarineBiology 21, 34–46.

Hopkins, T.L., Baird, R.C., 1975. Net feeding in mesopelagic fishes. Fishery Bulletin73, 908–914.

Hopkins, T.L., Baird, R.C., 1977. Aspects of the feeding ecology of oceanic midwaterfishes. In: Andersen, N.R., Zahuranec, B.J. (Eds.), Oceanic Sound ScatteringPrediction. Plenum Publishing Corporation, pp. 325–360.

Hopkins, T.L., Baird, R.C., 1981. Trophodynamics of the fish Valenciennellustripunctulatus. I. Vertical distribution, diet and feeding chronology. MarineEcology Progress Series 5, 1–10.

Hopkins, T.L., Baird, R.C., 1985. Feeding ecology of four hatchetfishes(Sternoptychidae) in the eastern Gulf of Mexico. Bulletin of Marine Science36, 260–277.

Hopkins, T.L., Sutton, T.T., Lancraft, T.M., 1996. The trophic structure and predationimpact of a low latitude midwater fish assemblage. Progress in Oceanography38, 205–239. http://dx.doi.org/10.1016/S0079-6611(97)00003-7.

Hudson, J.M., Steinberg, D.K., Sutton, T.T., Graves, J.E., Latour, R.J., 2014. Myctophidfeeding ecology and carbon transport along the northern Mid-Atlantic Ridge.Deep-Sea Research Part I 93, 104–116. http://dx.doi.org/10.1016/j.dsr.2014.07.002.

Ikeda, T., Hirakawa, K., Kajihara, N., 1994. Diet composition and prey size of themesopelagic fish Maurolicus muelleri (Sternoptychidae) in the Japan Sea.Bulletin of Plankton Society of Japan 41, 105–116.

Jespersen, P., 1915. Sternoptychidae (Argyropelecus and Sternoptyx). Report on theDanish Oceanographical Expeditions 1908-10 to the Mediterranean andadjacent seas 2, 1–41.

Kinzer, J., Schulz, K., 1988. Vertical distribution and feeding patterns of midwaterfish in the central equatorial Atlantic II. Sternoptychidae. Marine Biology 99,261–269. http://dx.doi.org/10.1007/BF00393252.

Kuhlmann, D., Fukuhara, O., Rosenthal, H., 1982. Shrinkage and weight loss ofmarine fish food organisms preserved in formalin. Bulletin of the NanseiRegional Fisheries Research Laboratory 14, 13–18.

Lancraft, T.M., Robison, B.H., 1980. Evidence of postcapture ingestion by midwaterfishes in trawl nets. Fishery Bulletin 77, 713–715.

Legand, M., Rivaton, J., 1969. Cycles biologiques des poissons mésopélagiques del’est de l’océan Indien. Troisième note: Action Prédatrice des PoissonsMicronectoniques. Cahiers de l’Office de la Recherche Scientifique etTechnique Outre-Mer – Serie oceanographie VII, 29–45.

Mauchline, J., 1998. The biology of calanoid copepods. Advances in Marine Biology33, 1–710.

Mauchline, J., Gordon, J.D.M., 1983. Diets of clupeoid, stomiatoid and salmonoid fishof the Rockall Trough, northeastern Atlantic Ocean. Marine Biology 77, 67–78.

Merrett, N.R., Roe, H.S.J., 1974. Patterns and selectivity in the feeding of certainmesopelagic fishes. Marine Biology 28, 115–126. http://dx.doi.org/10.1007/BF00396302.

Moku, M., Kawaguchi, K., Watanabe, H., Ohno, A., 2000. Feeding habits of threedominant myctophid fishes, Diaphus theta, Stenobrachius leucopsarus and S.nannochir, in the subarctic and transitional waters of the western North Pacific.Marine Ecology Progress Series 207, 129–140. http://dx.doi.org/10.3354/meps207129.

Pakhomov, E.A., Perissinotto, R., McQuaid, C.D., 1996. Prey composition and dailyrations of myctophid fishes in the Southern Ocean. Marine Ecology ProgressSeries 134, 1–14.

Parin, N., Kobyliansky, S., 1996. Diagnoses and distribution of fifteen speciesrecognized in genus Maurolicus Cocco (Sternoptychidae, Stomiiformes) with akey to their identification. Cybium 20 (2), 185–195.

Pinkas, L., Oliphant, M.S., Iverson, L.K., 1971. Food habits of albacore, bluefin tuna,and bonito in California waters. Fishery Bulletin 152, 1–105.

Põllupüü, M., 2007. Effect of formalin preservation on the body length of copepods.Proceedings of the Estonian Academy of Sciences: Biology, Ecology 56 (4), 326–331.

V. Carmo et al. / Progress in Oceanography 130 (2015) 172–187 187

Rasmussen, O.I., Giske, J., 1994. Life-history parameters and vertical distribution ofMaurolicus muelleri in Masfjorden in summer. Marine Biology 120, 649–664.http://dx.doi.org/10.1007/BF00350086.

Roe, H.S.J., Badcock, J., 1984. The diel migrations and distributions within amesopelagic community in the north east Atlantic. 5. Vertical migrations andfeeding of fish. Progress in Oceanography 13, 389–424.

Samyshev, E.Z., Schetinkin, S.V., 1973. Myctophids: feeding patterns of some speciesof Myctophidae and Maurolicus muelleri caught in the sound dispersing layers inthe northwestern Africa area. Annales Biologiques. International Council for theExploration of the Sea (ICES): Copenhagen 28, 212–215.

Sutton, T.T., 2005. Trophic ecology of the deep-sea fish Malacosteus niger (Pisces:Stomiidae): an enigmatic feeding ecology to facilitate a unique visual system?Deep Sea Research Part II: Topical Studies in Oceanography 52, 2065–2076.http://dx.doi.org/10.1016/j.dsr.2005.06.011.

Sutton, T.T., Hopkins, T.L., 1996. Trophic ecology of the stomiid (Pisces: Stomiidae)fish assemblage of the eastern Gulf of Mexico: strategies, selectivity and impactof a top mesopelagic predator group. Marine Biology 127, 179–192.

Sutton, T.T., Porteiro, F.M., Heino, M., Byrkjedal, I., Langhelle, G., Anderson, C.I.H.,Horne, J., Søiland, H., Falkenhaug, T., Godø, O.R., Bergstad, O.A., 2008. Vertical

structure, biomass and topographic association of deep-pelagic fishes inrelation to a mid-ocean ridge system. Deep Sea Research Part II: TopicalStudies in Oceanography 55, 161–184.

Tyler, H.R., Pearcy, W.G., 1975. The feeding habits of three species of lanternfishes(family Myctophidae) off Oregon, USA. Marine Biology 32, 7–11.

Vinogradov, M., Tseitlin, V., 1983. Deep-sea pelagic domain (aspects ofbioenergetics). In: Rowe, G.T. (Ed.), Deep-Sea Biology. Wily-Interscience, NewYork, pp. 123–165.

Wenneck, T.de L., Falkenhaug, T., Bergstad, O., 2008. Strategies, methods, andtechnologies adopted on the RV GO Sars MAR-ECO expedition to the Mid-Atlantic Ridge in 2004. Deep Sea Research Part II: Topical Studies inOceanography 55, 6–28. http://dx.doi.org/10.1016/j.dsr2.2007.09.017.

Yang, M-S., 2011. Diet of Nineteen Mesopelagic Fishes in the Gulf of Alaska. U.S.Dep. Commer., NOAA Tech. Memo. NMFS-AFSC-229, 67 p.

Young, J.W., Blaber, S.J.M., 1986. Feeding ecology of three species of midwater fishesassociated with the continental slope of eastern Tasmania, Australia. MarineBiology 93, 147–156.