Trophic relationships between the jumbo squid ( Dosidicus gigas ) and the lightfish ( Vinciguerria lucetia ) in the Humboldt Current System off Peru

1 Instituto de Ciencias del Mar (ICM-CSIC), Passeig Marítim de la Barceloneta, 37-49, 08003 Barcelona, Spain. E-mail: rigoberto@icm.csic.es 2 Instituto del Mar del Perú (IMARPE), Esquina Gamarra con General Valle s/n, Chucuito, Callao, Peru. 3 Programa de Doctorado en Oceanografía, Departamento de Oceanografía, Universidad de Concepción, Casilla 160C, Concepción, Chile. 4 Centro de Investigaciones Biológicas del Noroeste S.C. Mar Bermejo 195, Colonia Playa, Palo de Santa Rita, La Paz B.C.S. México.


INTRODUCTION
The jumbo squid Dosidicus gigas (d`Orbigny, 1835) is a voracious predator that attacks a great va-riety of prey, including fish, crustaceans and other invertebrates (Ehrhardt et al., 1986, Markaida and Sosa-Nishizaki, 2003, Markaida et al., 2008).It is con-is considered an important organism in the ecosystem due to its voracity and its tendency not to feed selectively (Markaida and Sosa-Nishizaki, 2003).It has become an important commercial resource in the eastern Pacific Ocean.In Peru the jumbo squid fishery has been one of the most important fisheries since 1990 (Yamashiro et al., 1998, Markaida and Sosa-Nishizaki, 2003, Rosas-Luis et al., 2008, Keyl et al., 2008).This species feeds mainly at night, but there are reports of jumbo squid feeding during the day in the California Current System of Mexico (Markaida andSosa-Nishizaki, 2003, Rosas-Luis et al., 2008), and in the Humboldt Current System of Peru (Alegre-Norza, personal observations, 2010), which shows that the species carries out both vertical (Gilly et al., 2006) and horizontal migrations (Markaida et al., 2005).
In the Humboldt Current System of Peru, upwelling events on the coast result in high concentrations of primary and secondary production that are exploited by D. gigas and other species such as mesopelagic fish.A remarkable feature of mesopelagic fish, such as the lanternfish (Myctophidae) and lightfish (Phosichthyidae), is that they perform extensive vertical migrations and form dense patches.Vertical migration is one of the most widespread patterns of animal behaviour in mesopelagic zones (Frank and Widder, 2002), and influences the life histories of non-migrating and migrating fish and cephalopods (mainly the jumbo squid) and the feeding behaviour and spatial distribution patterns of predators (Konchina, 1983, Benoit-Bird and Au, 2002, Bertrand et al., 2002).Aggregations, densities and the geographic distribution of species can be quantified and mapped by acoustic methods, such as sound scattering layers (Lapko and Ivanov, 1994, Luo et al., 2000, Cornejo and Koppelmann, 2006).This is very useful for estimating the abundance of marine organisms.Data from acoustic surveys can also provide biological information, such as spatial distribution patterns and migratory movements.These methods have therefore been used to study the predator-prey relationship when species are well discriminated (Miyashita et al., 2004).For D. gigas, the acoustic method has been standardized and acoustic detection has frequently been used to estimate biomass in Mexico and Peru (Benoit-Bird et al., 2008).However, acoustic detection of the lightfish (Vinciguerria lucetia,Garman, 1899) is always compared with trawls that are monitored and controlled by an acoustic net-recorder (Marchal andLebourges, 1996, Cornejo andKoppelmann, 2006).
Dosidicus gigas is an important commercial resource and like V. lucetia, it is important to marine ecology.V. lucetia is one of the most abundant species of mesopelagic fish in the eastern Pacific Ocean both in tropical and warm waters (Ahlstrom, 1968), and it has been reported in studies on the feeding habits of squid (Schetinnikov, 1986, Schetinnikov, 1989, Markaida and Sosa-Nishizaqui, 2003, Markaida, 2006, Rosas-Luis, 2007).Due to the abundance of D gigas, its feeding behavior and its importance as a fishery resource, and also the dynamism and abundance of V. lucetia, the objective of this study was to determine the trophic relationship between D. gigas and V. lucetia in the pelagic ecosystem in the southeast Pacific Ocean off Peru (Fig. 1), based on the data analysis of acoustic measurements, midwater trawl fishing and stomach content analyses in 2007 and 2008.

Biological data
Three bio-acoustic surveys were carried out by the Instituto del Mar del Peru (IMARPE) aboard the BIC Humboldt and José Olaya Balandra in the Humbold Current System off Peru in March (summer) of 2007 and June-July (autumn-winter) and November-December (spring) of 2008.
Jumbo squid were caught, measured and the stomach contents were sampled.Hard structures and tissues of fish, cephalopods and crustaceans were used to identify prey in 829 stomach, while 108 stomachs were empty.The stomach contents were passed through a 500 m mesh sieve in the Trophic Ecology Laboratory of IMARPE.Observations were made under a binocular microscope (60-120x) over a black and white background to aid identification.
Frequency of occurrence and numeric and gravimetric methods were used to quantify the diet.The frequency of occurrence (%FO) was calculated as the percentage of jumbo squid that fed on a certain prey, the number (%N) was the number of individuals of a certain prey in relation to the total number of individual prey, and the weight (%W) was defined as the weight of a certain prey in relation to the total weight of all the prey (Cailliet, 1977).
Graphs of the index of relative importance (IRI) were plotted to illustrate the diet compositions obtained from scientific cruises (Pinkas et al., 1971) (Fig. 2).The most important prey according to the IRI were included in the plots.IRI= (%N+%W)*%FO.

Acoustic data
Acoustic data were collected with a Simrad EK60 dual frequency quantitative scientific echo sounder that consisted of split-beam transducers of 38 and 120 kHz mounted on the ship's hull, which were calibrated prior to the survey using standard procedures (Foote et al., 1987).The data were processed with Echoview (Simmonds and MacLennan, 2005).The water column was investigated down to depths of 500 m. Figure 3 shows the acoustic survey area and the 11 parallel transect lines.Each transect line crossed the continental shelf to the oceanic zones (about 300 nautical miles from the shore), where sea depths range from 5 to 500 m.A daytime survey (from 1 h after sunrise to 1 h before sunset) and a night-time survey (from 1 h after sunset to 1 h before sunrise) were conducted for each transect line within 24 h.
For mesopelagic fish, acoustic detection with a -70 dB threshold was applied to minimize bias due to noise or non-mesopelagic fish.With this threshold, the nautical-area-backscattering coefficients were recorded along survey tracks at georeferenced elementary distance sampling units of 1 nautical mile each.The result can be considered to represent the biomass of mesopelagic fish (Bertrand et al., 1999, MacLennan et al., 2002).Sometimes several species were found in mixed concentrations so that the marks on the echogram from each species could not be distinguished.The echogram shows that the echo-integrals can provide data about a group of mixed species as one category, but not about the individual species.However, it is possible to make further divisions to species level by referencing the composition of the trawl catches (Nakken and Dommasnes, 1975).To support this process, organism samples were collected at the same time by non-closing pelagic trawls (Length: 55 m: mesh codend: 13 mm).Commercial midwater trawls were used to determine the taxonomic composition of the mesopelagic fish in The nautical-area-backscattering coefficients (S A ) were calculated for each trawl for cells in which fish were present (S A +), which is an index of fish density.We used the following expression for calculating the biomass: s kg = 4p 10 TSkg/10 TS kg = TS -10 log (<w>/1000) V. lucetia: TS= 20 Log L-79.06 (dB), L: 3.5 to 6.5 cm (Gutierrez and Herrera, 1998) D. gigas: TS= 20 Log L-92.82 (dB), L: 65.5 to 93.5 cm (Castillo and Gonzales, 2000) D. gigas: TS= 20 Log L-86.17 (dB), L: 22 to 38 cm (Castillo and Gonzales, 2000) where: TS: target strength, specific to each species.
A: isoparalitoral area (nm 2 ) s kg : retrodispersed acoustic section (kg).C i : instrumental constant of echosounder.<w>: weight average of species (g).L: body length of a scatterer (normally this is the total length for fish, and dorsal mantle length for squid) C: acoustic constant.

Consumption of V. lucetia
The percentage of V. lucetia in the total stomach contents of each jumbo squid was calculated and plotted to show the trophic relationship between these two species.To estimate the consumption rate (Qi ) of V. lucetia by D. gigas, we determined three parameters: (1) the biomass of the predator (Bj); (2) the consumption-biomass relationship of the predator (Q/B)j [taken from Alarcon-Muñoz et al. (2008)]; and (3) the diet composition (DCij) of the prey (i ) in the stomach contents of the predator ( j).The following expression was used to calculate the consumption rate:

RESULTS
A total of 937 jumbo squid was sampled (472 females, 456 males and 9 unknown).The dorsal mantle length (ML) was measured, and ranged in size between 3.2 and 103.5 cm (Table 1).The most usual size of males and females was 10 to 30 cm ML, and large jumbo squid (>70 cm ML) were observed on all research cruises.

Diet description
The stomach contents of D. gigas revealed two main groups: fish and molluscs (mainly cephalopods).Fish were found in practically every stomach (81.62% FO) in 2008 and (18% FO) 2007, and were the most important group in both years.The index of relative importance (IRI) was used to determine the importance of each group in the diet of D. gigas.In 2007 fish and cephalopods (Teuthida) were the most important groups and we observed that D. gigas fed primarily on these two groups (Fish IRI=1841, Teuthida=1238).The situation changed drastically in 2008, when jumbo squid were found to feed mainly on fish (Fish IRI=12795.82and Teuthida=852.5).V. lucetia was the main item in the jumbo squid diet.In 2007 it accounted for 44.1%N,IRI=445.4,and in 2008 it was present in over half of the samples (54% and 69.12% FO) and accounted for 53.3% and 42.6%N (IRI=3734 and 7029).The second most important fish was the myctophid Myctophum nitidulum, which was present in both years, and showed particularly high values in November-December 2008 (IRI=651).Another myctophid in the squid's diet was Diogenichthys laternatus, which was only present in 2008 (IRI 268 and 199) (Table 2).
Cephalopods were next in importance, and were found in 19.4% of stomachs in 2007 and 10.8%, and 5.53% in 2008, and accounted for 19.4% of the prey in 2007 and 4.62% and 0.77% in 2008.D. gigas was the main item in the group as it was found in 6.45% of samples in 2007 and accounted for 8.24% of prey (IRI= 364.3).In 2008 its importance in the diet declined (IRI=479.7 and 69.7)(Table 2).
Other groups were present in the diet such as the crustaceans: Pleuroncodes monodon, pteropoda and protista but they were not significant (IRI less than 35) (Table 2).
The stomach samples collected represent the climatic seasons of Peru: March (summer), June-July (autumn-winter) and November-December (spring).D. gigas mainly fed on V. lucetia in the three seasons (Fig. 2), and the importance of this fish increased throughout the year (IRI summer= 445.4,autumn-winter= 3734 and spring= 7029.2) (Fig. 2, spelling bars).The diet of D. gigas did not vary over the year, and we always found the same prey groups (fish, cephalopods and crustaceans); however, the frequency of occurrence of these groups varied in the two years (Table 2).

Acoustic observation of the distribution patterns of D. gigas and V. lucetia
Typical echograms at 38 and 120 kHz (Fig. 3) allowed us to explain the distribution patterns and the aggregative behaviour of D. gigas and V. lucetia in the Humboldt Current System of Peru.We observed the daily vertical migration of mesopelagic fish distributed in the sound scattering layers.The midwater trawls indicated that these acoustic structures were formed mainly by micro-nektonic organisms such as V. lucetia.In the summer 2007, D. gigas was detected at depths of between 2 and 215 m, and it was observed interacting with V. lucetia (Fig. 3).The distribution of the two species was similar at between 26 and 290 nautical miles (nm) from the seashore.Both species came close to the surface water at night, and migrated to deeper waters during the day.Biomass detection of these species with acoustic methods showed V. lucetia to be a principal component of the mesopelagic system, and its movements were related to those of D. gigas.
Transect F and H in Figure 3 show important interaction between D. gigas and V. lucetia in 2007 and 2008.The echograms were plotted during day and night.The vertical migration of the two species is the main component of the echogram; during the day, they occupied deeper waters from 120 to 300 m depth and at sunset they returned to surface water.Generally, D. gigas and V. lucetia shared the same distribution range at the same times.In the northern area at night they were detected in two areas (257-300 nm and 60-160 nm) near the surface water and at 50 m depth, whereas in the day the distribution was between 160 and 250 nm in deeper water (155-250 m depth) (Transect F).The same day-night pattern was found in the southern area, but the distribution of the two species was between 20 and 160 nm (Transect H).

Consumption of V. lucetia by D. gigas
Table 3 shows the acoustic biomass estimates of D. gigas and V. lucetia obtained in 2007 and 2008.Ac- cording to the Q/B and the diet of D. gigas, predation on V. lucetia by D. gigas was of the order of 5% to 24% of the biomass estimate.The greatest predation impact by D. gigas was in 2008 with 23.3%, which means that the jumbo squid consumed 570 084 tons of V. lucetia.
We averaged the diet consumption values (DCij=0.30),and using this value an estimate of jumbo squid consumption was made for 1998 to 2009 (Fig. 4).We should also highlight that there was a positive correlation between the biomass estimates of D. gigas and V. Lucetia.Accordingly, there was a moderate correlation throughout the entire series from 1999 to 2009 with an r 2 of 0.34 (90% confidence limits, P= 0.05).The biomass estimates of D. gigas and V. lucetia for 2007 and 2008 were used to contrast the trophic relationships.The regression value was less than the complete series (r 2 = 0.28) but in both cases the regression values were positive.This shows that there is a moderate relation between D. gigas and the biomass of V. lucetia, which is expressed in the diet of the squid.

DISCUSSION
Jumbo squid sampled in 2007 and 2008 showed typical trophic behaviour: they fed on fish, cephalopods and crustaceans, and discarded the cannibalism.In general, studies on the feeding behaviour of jumbo squid are influenced by fishing activity because jumbo squid nibble fishery products, which has led to misconceptions about its feeding activity (Nigmatullin et al., 2001, Markaida and Sosa-Nishizaki, 2003, Markaida, 2006, Ruiz-Cooley et al., 2006, Field et al., 2007, Rosas-Luis, 2007, Markaida et al., 2008).However, when the data come from scientific sources, the results are more reliable: cannibalism is reduced or absent (Rosas-Luis, 2007), which is also reflected in this study.Samples of jumbo squid were taken on fishing cruises and the stomach contents were immediately frozen to preserve the tegument, scales and otoliths, which are the main structures used for identifying prey.The main group in the diet of the jumbo squid was found to be mesopelagic fish and the principal prey was the lightfish V. lucetia.
The acoustic biomass estimates made for D. gigas and V. lucetia are considered realistic because the target strengths were contrasted in situ with jig sampling of squid and fish trawls for V. lucetia.It is possible for acoustic signals to be misinterpreted; however different ways of obtaining the best signal for squid have been contrasted and it has been shown that the squid length is the best factor for obtaining the best target strength estimate (Castillo andGonzales, 2000, Benoit-Bird et al., 2008).In fact, other considerations can modify the final biomass values (sex proportion, maturity stage and size) but the error is not significant (Soule et al., 2010).In mesopelagic fish is more difficult to identify the acoustic signal because different fish aggregate at the same time.However, in order to correctly estimate biomass the values obtained with the acoustic method can be contrasted with net trawls, in which the collected organisms are identified and analyzed.This provides a tool for verifying the fish composition in the water column (Marchal andLebourges, 1996, Cornejo andKoppelmann, 2006).Both D. gigas and V. lucetia form aggregations.This characteristic and acoustic detection supported the hypothesis of a strong relationship between these two species, which is reflected in the jumbo squid's feeding activity (Fig. 2) and the vertical migrations shown in the echograms (Fig. 3).
Trophic relationships between D. gigas and V. lucetia are evident during the day and night and D. gigas probably feeds on V. lucetia in deeper waters.Alarcon et al. (2004) carried out experimental fishing with semipelagic trawls for lightfish and recorded large catches of jumbo squid.In research surveys of pelagic resources carried out by IMARPE, the jumbo squid catches also coincided with the detection of mesopelagic fish layers.The acoustic biomass estimates and echogram signals of D. gigas and V. lucetia in the water column suggest a close relationship between prey and predator, as the estimated distribution of jumbo squid near and offshore the continental shelf break overlapped with that of V lucetia.During the night, D. gigas was near the surface water surrounded by V. lucetia (Fig. 3).
They were together until the first hours of the day when both V. lucetia and D. gigas moved to mid-and deeper water.They were in the same water layer during the day and returned to the surface water in the first hours of the night.This trophic relationship is reflected in the stomach content analysis of the squid, in which V. lucetia was the main component (Fig. 2).Evidently, jumbo squid prey on other species, such as the Peruvian anchovy (E.ringens) and other Myctophids, which were present in the stomach contents at the same time as V. lucetia.However, their abundance in the ecosystem is probably lower than that of V. lucetia.Moreover that V lucetia and D. gigas migrate to deeper waters during the day (migratory behaviour: Markaida et al., 2005, Gilly et al., 2006) and it is the factor that determines the dominance of V. lucetia in the squid's stomach contents.
Based on the stomach content analysis of D. gigas and the acoustic detection we can infer that V. lucetia was the main component of the jumbo squid's diet in 2007 and 2008 in the Humboldt Current System off Peru.As mesopelagic fish are important components of oceanic ecosystems, they are abundant and have a wide distribution in the ocean (Ahlstrom et al., 1976).These characteristics of the group are evident in V. lucetia, one of the most important fish in the Humboldt Current System of Peru.It is found between 5º and 18ºS and can dominate the total catch in up to 68% of scientific cruises (Cornejo and Koppelmann, 2006).This dominance is expressed in other parts of the ocean.In the eastern tropical and subtropical Pacific this species has the third largest biomass after the northern anchovy Engraulix mordax and the Pacific hake Merluccius productus (Smith and Moser, 1988).The acoustic biomass estimates of V. lucetia and D. gigas made in the Humboldt Current System off Peru since 1998 show that there is a relation between the two species (Fig. 4).When there is a high biomass of V. lucetia, D. gigas has been observed to focus its feeding on this species (Table 3).Like other cephalopods, jumbo squid do not feeding selectively, which causes direct impacts on the biomass of the most abundant species present in the same water layer.Therefore, the importance of the prey species lies in its ability to support jumbo squid predation.
V. lucetia is a species with dynamic development that promotes rapid population growth, which is reflected in a high abundance and wide distribution in the ocean.If trophic relations in the ocean ecosystem are influenced by the distribution and abundance of prey groups and V. lucetia is an important mesopelagic component of this ecosystem, then it can maintain the biomass of different predators, even D. gigas.Due to these trophic relationships, an increase in D. gigas biomass may be the result of a similar increase in biomass in V. lucetia, which promotes a positive trophic effect as well as the development of other species in the ecosystem.

Fig. 2 .
Fig. 2. -Left: Composition by percentage number (%N) and weight (%W)(vertical axis) and frequency of occurrence (%FO, horizontal axis) of the main prey found in the stomach contents of jumbo squid collected in the Humboldt Current System off Peru (March 2007, June-July 2008 and November-December 2008).Right: Percentage of V. lucetia found in the stomach contents of jumbo squid.

Fig. 3 .
Fig. 3. -Acoustic detection of V. lucetia and D. gigas in 2007 and 2008 in the Humboldt Current System off Peru.Transect A, B, C, D and E were completed in March 2007; F, G and H in June-July 2008; and I, J and K in November-December 2008.SD, seashore distance in nautical miles.Vertical movements of jumbo squid and V. lucetia were clearly detected in the echogram of transects F and H.

Table 1 .
-Summary data of jumbo squid collected in the Humboldt Current system off Peru in 2007-2008 Table2.-Frequency of occurrence (FO), number (N), and weight (W) of prey in the stomach contents of jumbo squid during 2007-08, weight in grams.