Feeding habits of billfishes ( Carangaria : Istiophoriformes ) in the Ecuadorian Pacific Ocean

The feeding habits of Makaira nigricans, Kajikia audax, Istiophorus platypterus (Istiophoridae), and Xiphias gladius (Xiphiidae) in the southeast Pacific Ocean were examined in Manta and Santa Rosa, Ecuador. This study describes the diets of these billfish species, evaluates dietary differences between species, and assesses seasonal differences in diet. A total of 274 M. nigricans, 321 K. audax, 267 I. platypterus, and 252 X. gladius were collected between February 2014 and April 2015. The scombrid Auxis spp. was the most important prey for M. nigricans, K. audax and I. platypterus, while the squid Dosidicus gigas was the most important prey for X. gladius. The results of the ANOSIM confirmed significant differences in feeding habits between the members of the family Istiophoridae and X. gladius. Seasonal differences for I. platypterus also were observed. Billfishes are specialist consumers with a narrow niche breadth (Ba: M. nigricans and K. audax=0.1, I. platypterus=0.05, and X. gladius=0.04).


Introduction
Commonly known as billfishes, the large fishes of the families Xiphiidae and Istiophoridae are active predators characterized by a protruding upper jaw that extends considerably beyond their lower jaw (Nakamura, 1985).The longline fishery for large pelagic fishes operates year round in Ecuador and billfishes are one of the most important resources for the local artisanal fishery as well (Martínez-Ortíz et al., 2015).
The most tropical billfish species is the blue marlin Makaira nigricans Lacépède, 1802, which is commonly found in equatorial waters (Nakamura, 1985).This epipelagic species it is the most oceanic of the istiophorids and is found in coastal waters only where the continental shelf is narrow (Nakamura, 1985;Joseph et al., 1988).The striped marlin Kajikia audax (Philippi, 1887) is an epipelagic predator usually found above the thermocline, although the species migrates to deeper waters at night (Nakamura, 1985;Sippel et al., 2011).The sailfish Istiophorus platypterus (Shaw, 1792) prefers a coastal habitat (Nakamura, 1985) and is usually found within 50 m of the surface, but undertakes short-duration vertical movements to 150 m depth (Chiang et al., 2011;Kerstetter et al., 2011).The swordfish Xiphias gladius Linnaeus, 1758 is the sole member of the family Xiphiidae and exhibits the greatest temperature tolerance among billfishes (Nakamura, 1985), performing vertical migrations to more than 900 m depth (Abascal et al., 2010).

e160162[2]
Habitat partitioning is important as it can lead to the consumption of distinct prey in different areas (Shimose et al., 2010); meanwhile, resource partitioning may reduce the competitive pressure between species (Papastamatiou et al., 2006).Prey availability also can be influenced by changes in the primary productivity as well as other biological and physical variables (Olson et al., 2014) that affect the occurrence of billfishes (Shimose et al., 2008;Shimose et al., 2010).
Marine predators provide valuable information about prey abundance and composition (Potier et al., 2007;Olson et al., 2014).Stomach content analyses also facilitate the identification of feeding habits as well as competition or resource partitioning between species that occurs in the same region (Baker, 1966;Vaske et al., 2004;Shimose et al., 2010), ecological aspects that are not commonly considered for billfishes.Traditionally, studies on the feeding habits of billfishes in the Pacific Ocean have focused on a single species (Abitia-Cárdenas et al., 1999;Rosas-Alayola et al., 2002;Shimose et al., 2006;Watanabe et al., 2009;Abitia-Cárdenas et al., 2011;Tsai et al., 2015) or prey group (Rosas-Luis et al., 2016).
Despite their ecological and commercial importance, biological information on these fishes remains scarce, particularly for the Southeast Pacific Ocean.The feeding ecology of billfishes in Ecuador was assessed to: (1) quantify the diet of billfishes, (2) evaluate the possible differences in diet between distinct species and seasons, and (3) describe their feeding strategy.

Study area and sample collection.
The study area is located off the coast of Manta (0°56′59″S, 80°42′34″W) and Santa Rosa (02°13′0″ S, 80°58′0″W) (Fig. 1).One of the most important characteristics in this area is the Equatorial Front located between the Galapagos and the Ecuadorian mainland at about 0°-3°, that divide the cold nutrient-rich waters of the Humboldt Current, as well as its extension, the South Equatorial Current, from warmer, nutrient-poor surface waters in the north (Wooster, 1969).Front hydrographic conditions determine the formation of convergence and divergence areas in surface levels that promote primary production, zooplankton biomass and fish stocks (Jiménez, 2008).Fishermen place their longline gear approximately 40 to 200 miles from the coastline, mainly overnight.This gear consists of a mainline with 120-300 branch lines, each separated by 40-60 m.Branch lines are 11-34 m in length and typically hold a J-shaped hook with a curved shank (Martínez-Ortíz et al., 2015).
Billfishes were sampled from the artisanal fishing fleet between February 2014 and April 2015.Samples were collected principally during morning hours, including almost all of the landed specimens.All billfishes were measured (lower jaw-fork length, L LJF in cm) and the stomach contents were stored in a labeled plastic bag and then preserved on ice for transportation to the laboratory.Samples were frozen at -20°C until they could be analyzed.Stomach content analysis.The stomach contents were identified to the lowest taxonomic level, and prey species were counted and weighed to the nearest 0.01 g.Whole fish and cephalopods were identified following Fischer et al. (1995a,1995b), and Jereb, Roper (2010).Other prey items were identified based skeletons or beaks through comparison with the project's fish skeleton collection or published identification guides (Clothier, 1950;Clarke, 1962;Iverson, Pinkas, 1971;Wolff, 1984;Clarke, 1986).

Sample size sufficiency.
To evaluate if the sample size was sufficient to describe the full diet, a randomized cumulative prey curve was generated using the "vegan" package (Oksanen et al., 2010) in R (R Development Core Team, 2014) based on the lowest taxonomic level identified for each prey (Preti et al., 2012).The mean species accumulation curve (± 2 standard deviations) was plotted from 500 random permutations of the data.The number of samples is assumed to be sufficient to describe the diet when the curve approaches the asymptote (Hurtubia, 1973).When the asymptote was not evident, a straight line to the last four points was fitted and the slope of the line was compared with a line of slope zero, reaching the asymptote when the lines did not differ significantly (Bizzarro et al., 2007).

Data analysis.
To determine the importance of each prey in the diet, the prey-specific index of relative importance (%PSIRI) (Brown et al., 2012) was used according to the equation: %PSIRI = 0.5 %F O (%P Ni + %P Wi ), where %F O is the percent frequency of occurrence (the number of stomachs containing prey i divided by the total number of stomachs, n), and %P Ni and %P Wi are the prey-specific abundances by number or weight, respectively.Prey-specific abundance was calculated with the equation %P , where %A ij is the abundance (by number %P Ni or weight %P Wi ) of prey i in stomach sample j and n i is the number of stomachs containing prey i.The %PSIRI is a modification of the index of relative importance (IRI) (Pinkas et al., 1971).The measure accounts 3   e160162[3] for %F O redundancies in the %IRI, and is additive with respect to taxonomic levels.Thus, the %PSIRI of a family will be equal to the sum of the %PSIRI of all of the species contained in this taxon (Brown et al., 2012).The stomach content index (SCI) was compared between species using the Kruskal-Wallis test; Dunn's test was applied for post hoc comparisons.This index was calculated as SCI = (SCW × 100)/PW, expressing stomach content weight (SCW) as a ratio of the total weight of all stomachs (PW) for each individual billfish to estimate its stomach fullness (Shimose et al., 2006).
The similarity of the diet between different species and seasons (rainy season = December to May; dry season = June to November) was evaluated using analysis of similarities (ANOSIM).Prey were grouped by family in order to reduce the number of prey categories in the samples with zero values, thus increasing the effectiveness of the multivariate analysis (White et al., 2004;Espinoza et al., 2013;Szczepanski, Bengtson, 2014).Data were permutated 999 times for a distribution to determine the p-value of the ANOSIM's R statistic (R = 0 is identical, R =−1 or 1 is most divergent).The SIMPER analysis (Clarke, Warwick, 2001) was used as a post-hoc test to identify the prey taxa responsible for the dissimilarities in the diet of different billfish species.
Levin's standardized measure of niche breadth (B a ) (Krebs, 1999) was calculated by applying %PSIRI converted to proportions of the different prey identified.This measure varies from 0 to 1, where values close to 0 reflect a specialized diet and values close to 1 indicate a generalized diet (Krebs, 1999).Feeding strategy was assessed using the graphical analysis proposed by Amundsen et al. (1996).This procedure, which is a modification of Costello's (1990) graphical method, provides information about prey importance and the predator's feeding strategy by evaluating a two-dimensional plot of prey-specific abundance (%Pi) against %F O , with %Pi = (Σ prey i weight/Σ weight of all prey in the stomachs containing prey i) × 100.
Feeding strategy.The Amundsen graphical analysis suggests that these billfish species are specialist consumers with Auxis spp.being the dominant prey for M. nigricans, K. audax, and I. platypterus, while D. gigas is the main prey of X. gladius.Several prey are located in the upper left corner of the diagram, suggesting a high between-phenotype contribution to the niche width, with these prey being consumed by a few specialized individuals (Fig. 3).The niche breadths calculated for M. nigricans, K. audax, I. platypterus, and X. gladius were narrow (B a =0.1, 0.1, 0.05, and 0.04, respectively).Fig. 3. Prey-specific abundance plotted against frequency of occurrence of prey species.Explanatory axes for foraging patterns are those of Costello (1990) as modified from Amundsen et al. (1996).The two diagonal axes represent the importance of prey (dominant vs. rare) and the contribution to the niche width (high between-phenotype vs. high within-phenotype contribution); the vertical axis defines the predator feeding strategy (specialist vs.

Discussion
The number of different prey species identified for all billfishes examined in this study was lower than that reported for other areas of the Pacific Ocean (Abitia-Cárdenas et al., 1997;Abitia-Cárdenas et al., 1999;Arizmendi-Rodríguez et al., 2006;Letelier et al., 2009), where these predators include other taxa, like crustaceans, in their diet (Shimose et al., 2006;Watanabe et al., 2009;Abitia-Cárdenas et al., 2011).These billfish species exploit a smaller group of feeding resources in the Ecuadorian Pacific Ocean coinciding with their narrow niche breadth.In the eastern Pacific Ocean, M. nigricans, K. audax, I. platypterus and Tetrapturus angustirostris prey are scarcer near the equator, which may be related to lower productivity (Shimose et al., 2010).
The asymptote was not reached in any of the species accumulation curves despite the large sample size; a number of factors contributed to this, including differential digestion rates, regurgitation of the stomach contents as the result of being caught during fishing (Chase, 2002;Hernández-Aguilar et al., 2013), and the daily periodicity of feeding (Shimose et al., 2006).The low SCI values suggest that these billfishes do not feed actively before dawn or at night in this area.Off the coast of Japan, Shimose et al. (2006) observed that the frequency of empty stomachs among M. nigricans tended to be higher in the early morning than at any other time of day, suggesting a daily periodicity in feeding.Furthermore, these low values also may be related to the prey availability in our study area, as prey items varies by area in the Pacific Ocean (Shimose et al., 2010).
Fishes were the main prey group for M. nigricans, K. audax and I. platypterus in our study area; similar results have been reported by other authors in the North Pacific (Abitia-Cárdenas et al., 2002;Shimose et al., 2006), off the coast of Taiwan (Tsai et al., 2015), and in the Ecuadorian Pacific Ocean (Rosas-Luis et al., 2016).Auxis spp.has been reported as an abundant prey in the eastern Pacific Ocean (Olson, Boggs, 1986;Olson et al., 2014); the importance of Auxis spp.for billfishes previously was demonstrated off the coast of Cabo San Lucas, Mexico, for M. nigricans (Abitia-Cárdenas et al., 1999), and off Acapulco, Mexico, for I. platypterus (Hernández-Aguilar et al., 2013).Scombrid fishes, including Auxis thazard and Auxis rochei also have been reported as the most abundant prey group for I. platypterus, M. nigricans, and K. audax in the eastern North Pacific Ocean (Shimose et al., 2010).
The most important prey in the billfishes diet varies by geographic location (Shimose et al., 2010).Thus, for M. nigricans off the coast of southwestern Japan, Katsuwonus pelamis was important and was consumed mainly during the day (Shimose et al., 2006).Scomber japonicus was the most abundant prey for K. audax off the coast of Cabo San Lucas, Mexico (Abitia-Cárdenas et al., 1997), while the Humboldt squid D. gigas was the prey most commonly consumed prey by I. platypterus off Mazatlán, Mexico, (Arizmendi-Rodríguez et al., 2006).Our results demonstrate that Auxis spp. is an important feeding resource in the Ecuadorian Pacific.The Humboldt squid D. gigas was reported previously as a representative prey in the diet of X. gladius in the Pacific Ocean (Ibáñez et al., 2004;Castillo et al., 2007;Letelier et al., 2009).This feeding strategy is related to the abundance of this prey species (Nigmatullin et al., 2001) and to the vertical migrations undertaken by both species.During the day, X. gladius prey on D. gigas in deep waters (Watanabe et al., 2009), whereas at night they consume the same species on the surface when D. gigas migrates there in pursuit of its own prey (Rosas-Luis et al., 2011).
A low trophic overlap was observed between X. gladius and the istiophorid fishes, which may be the result of differences in their horizontal distributions.X. gladius is found in deeper waters than members of the family Istiophoridae, allowing it to exploit other areas as well as shallower waters where it is is found at night (Abascal et al., 2010;Abecassis et al., 2012).Furthermore, horizontal movements in this area (Abascal et al., 2010) as well as the importance of D. gigas in the diet may suggest a close prey-predator relationship.
Although istiophorid fishes consume similar diets throughout the Ecuadorian Pacific, there is low trophic overlap between I. platypterus, M. nigricans, Tetrapturus angustirostris and K. audax in the eastern North Pacific Ocean where the prey consumed varies by zone and size (Shimose et al., 2010).While stomach content analyses indicate that M. nigricans and K. audax consume similar diets in the southern Gulf of California, Mexico, stable isotope analyses show significant differences, suggesting niche segregation (Torres-Rojas et al., 2013).
In the Ecuadorian Pacific Ocean, M. nigricans, K. audax, and X. gladius consume distinct prey species during different seasons; however, we did not observe significant differences in our study.In the northern hemisphere off the coast of Cabo San Lucas, Mexico, the number of food components recorded for M. nigricans varies between summer and fall (Abitia-Cárdenas et al., 1999).Off the coast of Mexico, K. audax shows seasonal and size-related differences in feeding (Abitia-Cárdenas et al., 2011) wherein cephalopods are the most important prey in the summer, but fishes are the primary prey consumed during spring and fall (Abitia-Cárdenas et al., 1997).
The diet of I. platypterus varied seasonally; this contrasts with the results reported for this species off Acapulco, Mexico, where no seasonal differences have been observed in diet (Hernández-Aguilar et al., 2013).This seasonal variation may be influenced by the coastal habitat of this species and the seasonal availability of different prey in the area.In the Ecuadorian Pacific, I. platypterus consumes clupeid fishes like Ophistonema libertate and Etrumeus teres only during the dry season.Ophistonema libertate is abundant in the central region of Ecuador (Patterson, Santos, 1992) and E. teres is mainly distributed inshore, coinciding with the coastal habitat preferences of I. platypterus (Nakamura, 1985).Ophistonema libertate catches increase at lower temperatures (Patterson, Santos, thus, the 7 e160162 [7] seasonal preference for this prey may be explained by its greater abundance in cooler waters and the opportunistic feeding strategy employed by I. platypterus (Rosas-Alayola et al., 2002;Varghese et al., 2013;Tsai et al., 2015).
In conclusion, these specialist predators consume fewer prey species in the Ecuadorian Pacific than do their counterparts in other regions of the Pacific Ocean.Moreover, resource partitioning occurs between members of the family Istiophoridae and X. gladius.As an opportunistic predator, I. platypterus varies its diet based on prey availability during different seasons.

Fig. 1 .
Fig. 1.Map showing the study area.Filled circles represent the sampling localities.