INTRODUCTION
Studying the feeding habits of fish allows us to understand aspects of their biology, ecology, physiology, and to better understand their functional role in the trophic web of aquatic ecosystems (Hyslop, Reference Hyslop1980; Valente, Reference Valente1992; Blaber, Reference Blaber1997; Wootton, Reference Wootton1998; Hajisamae et al., Reference Hajisamae, Chou and Ibrahim2003). These studies are fundamental for the description of trophic spectra and allow us to make inferences on trophic ecology, highlighting aspects of niche overlap, and therefore of intraspecific and interspecific interactions (Graham & Vrijenhoek, Reference Graham and Vrijenhoek1988).
In pelagic ecosystems one of the most representative species (because of its abundance and wide distribution) is the sailfish Istiophorus platypterus (Shaw, 1792), which forms part of what are called billfish. This term has been widely accepted to designate large fish from the families Xiphiidae and Istiophoridae, which are characterized by an extremely elongated upper mandible (Nakamura, Reference Nakamura1985). For sailfish, its main anatomic characteristic is its enormous dorsal fin, which gives it its common name.
Currently in Mexico, billfish are reserved for sport fishing, with sailfish in particular representing the most important resource in this group, constituting up to 90% of the sport-fishing catch (Arias, Reference Arias2007; Santana, unpublished data). Sailfish generate a large direct and indirect economic spillover from the collateral services (transportation, accommodation, feeding and taxidermy).
In the research on sailfish in the Mexican Pacific waters, the work on feeding habits has been made mainly along the northern coast of Mexico (Evans & Wares, Reference Evans and Wares1972; Eldridge & Wares, Reference Eldridge, Wares, Shomura and Williams1974; Galvan, Reference Galvan1999; Rosas et al., 2002; Arizmendi et al., Reference Arizmendi, Abitia, Galvan and Trejo2006), and only sparsely in the southern region (Romero, unpublished data). Despite its status as one of the most important pelagic species from ecological and economic viewpoints, to date there are no studies on the trophic biology of the sailfish, specifically for the waters off Acapulco in southern Mexico.
Our present study analysed the trophic spectrum of sailfish in the Acapulco region to generate information that could be compared with other areas in the Pacific to determine whether the trophic behaviour has variations in the Mexican waters.
MATERIALS AND METHODS
From March 2008 to December 2009, weekly sampling of the landings of sailfish from the sport fishing fleet that operates in the Bahia de Acapulco, Guerrero, Mexico (16°48′54″ to 16°51′55″N, 99°51′03″ to 99°54′16″W: Figure 1) were made. The postorbital length (PL, cm) (eye-fork length) and total weight (kg) were determined for each organism, and the stomach was then extracted. The determination of sex was made by macroscopic observation of gonads, which were also collected and weighed.
Fig. 1. Location of the study area. Semicircle around Bahia de Acapulco, Mexico, denotes the fishing area.
During stomach-contents analysis diet items (food components) were separated and identified to the lowest taxonomic level possible, depending on the state of digestion. For fish that showed a minimal digestion we used the keys by Allen & Robertson (Reference Allen and Robertson1994), Fischer et al. (Reference Fischer, Krupp, Schneider, Sommer, Carpenter and Niem1995) and Thomson et al. (Reference Thomson, Findley and Kerstitch2000). For fish in an advanced digestion state, vertebral characteristics were used following the identification keys of Clothier (Reference Clothier1950), Monod (Reference Monod1968) and Miller & Jorgensen (Reference Miller and Jorgensen1973). For cephalopods, the mandibular apparatus was used for identification using the keys of Wolff (Reference Wolff1982, Reference Wolff1984) and Clarke (Reference Clarke1986).
To determine if the number of stomachs analysed was representative of the sailfish diet, we drew the curve of accumulated diversity (Estimate Swin820). For this, we considered the index of relative importance (IRI) of each prey in the stomachs, estimating the value of the Shannon–Wiener (H′) diversity index for each stomach. The coefficient of variation (CV) was calculated to obtain a quantitative estimation of the number of stomachs that would be adequate and representative of the diet. When this CV is low or equal to 5% (0.05), the number of stomachs examined is adequate to represent the diet (Jimenez & Hortal, Reference Jimenez and Hortal2003; Rodriguez et al., Reference Rodriguez, Moreno, Abitia and Palacios2009).
Diet analysis was made by means of the frequency of occurrence (%FO = number of stomachs containing prey i/total number of full stomachs × 100), percentage of numerical abundance (%N = number of prey i/total number of prey × 100), and percentage of weight (%W = weight of prey i/total weight of prey × 100) (Hyslop, Reference Hyslop1980). Once these values were obtained, we calculated the IRI (IRI = %N + %W * %FO) (Pinkas et al., Reference Pinkas, Oliphant and Iverson1971), which incorporates the previous methods to evaluate the importance of each item in the trophic spectrum of the species (Liao et al., Reference Liao, Pierce and Larscheid2001). This index was expressed as percentages according to the method described by Cortes (Reference Cortes1997).
In addition to the general trophic spectrum, we integrated and compared diets by sex (males and females), and by sexual maturity (immature and mature) using the size at first maturity of 150.2 cm PL, estimated by Cerdenares (Reference Cerdenares2011) for Istiophorus platypterus close to the study area. For the seasonal climate analysis, diet data were grouped into warm and cold seasons, according to anomaly values of the sea surface temperatures, taking as warm months those that were above average and cold months those that were below average (Figure 2). The sea surface temperature records were obtained from monthly satellite images of the MODIS-AQUA sensor (level 3) at a resolution of 4 × 4 km (http://oceancolor.gsfc.nasa.gov/). The data processing was done using Matlab v.7.1., December to April were the cold months, and May to November were the warm months. The water temperature during the study was between 27.7 and 31.5°C.
Fig. 2. Values of surface temperature by month (dashed lines) in Bahia de Acapulco during the years 2008 and 2009. Cold season (black bars) and warm season (grey bars). The primary y-axis shows temperature (Celsius) and the secondary y-axis shows anomalies with respect to the general average.
To determine whether there were differences in diet by sex, season, and sexual maturity, we used a randomized permutation method of the similarity matrix (ANOSIM, PRIMER 6 v. 6.1.6). The global similarity range R (0 ≤ R ≥ 1) was used as a measure to compare degrees of separation. When R is close to zero, Ho is true, and there are no separations among the analysed groups (Clarke & Warwick, Reference Clarke and Warwick2001).
The diet breadth was calculated using the Levin standardized index (Bi) according to the method described by Hurlbert (Reference Hurlbert1978). For this analysis, we used the absolute values of the numerical method. This index has values from 0 to 0.59 when the species uses few prey resources and prefers certain prey. It is then characterized as a specialist predator, whereas Bi values over 0.6 indicate that the predator uses all resources without selection and is therefore a generalist.
The Costello plot modified by Amundsen et al. (Reference Amundsen, Gabler and Staldvik1996) was used to confirm the feeding strategy of I. platypterus.
RESULTS
A total of 561 sailfish were examined. Of these 254 had food contents (45%) and 307 (55%) were empty. The 254 organisms containing food were 101 to 201 cm postorbital length and 16 to 40 kg total weight. The CV in the cumulative trophic-diversity curve indicated that the curve reached an asymptote at about 91 stomachs, which were sufficient to represent the diet of sailfish (Figure 3).
Fig. 3. Randomized cumulative prey curve for sailfish. H, Shannon–Wiener diversity (H′); Max, maximum diversity; Min, minimum diversity; CV, coefficient of variation.
A total of 25 food components were identified, comprising five cephalopods, 19 fish, and unidentified organic matter (UOM). The total weight of stomach contents was 103865 g, of which fish was 86.18% (89520.27 g), followed by cephalopods at 7.84% (8148.77 g), and the UOM at 5.97% (6196.1 g). The most important prey by weight was the scombrid fish Auxis thazard (Lacépède, 1800) at 63.4% (65,850 g), followed by the cornetfish Fistularia commersonii (Rüppell, 1838) at 6.5% (6749), and the UOM at 5.97% (6196.1 g) (Table 1).
Table 1. Absolutes and percentages of numerical (N), gravimetric (W), frequency of occurrence (FO) and index of relative importance (IRI) of the prey of sailfish (Istiophorus platypterus) caught in 2008 and 2009 off the coast of Acapulco, Guerrero, Mexico.
A total of 1246 prey organisms were quantified, corresponding to 170 cephalopods (13.64%) and 1076 fish (86.36%). By number, A. thazard was 27.44% (342 organisms), followed by the fish F. commersonii at 19.34% (241 organisms), the oceanic puffer Lagocephalus lagocephalus (Linnaeus, 1758) at 8.03% (100 organisms), and the cephalopod Octopus spp. at 7.95% (99 organisms) (Table 1).
By frequency of occurrence, fish were the most important, occurring in 92.1% of stomachs, followed by cephalopods (31.5%), and the UOM (19.69%). The most important items were the fish A. thazard (49.21%), the UOM (19.69%), and the fish Caranx caballus (Günter, 1868) (14.17%) and L. lagocephalus (11.42%) (Table 1).
According to the IRI, fish were the most relevant prey at 86.19%, followed by cephalopods at 7.9%. The most important prey items were the fish A. thazard at 63.04% and F. commersonii at 6.64%, followed by the UOM at 5.91%, and the cephalopod Octopus spp. at 4.58% (Table 1).
Of the 254 stomachs, 114 were female (44.8%), 138 were male (54.41%), and two were unidentified (0.79%). In the female stomachs 21 items were identified, belonging to 10 families, 13 genera and 13 species. The IRI showed that the most important prey was the fish A. thazard (87.99%), followed by the UOM (2.94%), the fish F. commersonii (1.85%) and the cephalopod Octopus spp. (1.31%) (Figure 4A). In male stomachs 25 prey items were identified, belonging to 13 families, 17 genera and 17 species. The IRI showed that the fish A. thazard was the most important prey (85.79%), followed by the UOM (3.44%) and the fish F. commersonii (3.41%) (Figure 4B).
Fig. 4. Variation of prey species consumed by sailfish, determined with the index of relative importance: (A) females; (B) males; (C) immature; (D) mature; (E) cold season; (F) warm season. Auxis thazard (At); Fistularia commersonii (Fc); unidentified organic matter (UOM); Octopus spp. (Os); Balistes polylepis (Bp); Caranx vinctus (Cv); Lagocephalus lagocephalus (Ll); unidentified cephalopods (UC).
The ontogenetic analysis, using 150.2 cm PL as the size at first maturity for females, yielded stomachs from 18 immature and 96 mature organisms. The low number of immature female stomachs is because the fishing fleet focuses their efforts on catching larger fish.
In the immature organisms 20 items were identified corresponding to 9 families, 13 genera and 13 species. The IRI showed that the most important prey was the scombrid fish A. thazard (85%), followed by the triggerfish Balistes polylepis (Steindachner, 1876) (9%) and the jack Caranx vinctus (Jordan & Gilbert, 1882) (1.68%) (Figure 4C).
In the mature organisms 25 items were identified, corresponding to 13 families, 17 genera, and 17 species. The IRI showed that A. thazard was the most important prey (90.21%), followed by the UOM (3.42%) and the fish F. commersonii (1.42%) (Figure 4D).
For the cold season, 67 stomachs with food were analysed and 18 items were identified, belonging to 8 families, 11 genera, and 11 species. The IRI showed that the most important prey was the fish A. thazard (76.18%), followed by the UOM (12.91%), the fish F. commersonii (3.46%) and the cephalopod Octopus spp. (3.34%) (Figure 4E).
For the warm season, 187 stomachs with food were analysed and 24 items were identified belonging to 13 families, 16 genera and 16 species. According to the IRI, the most important prey was A. thazard (90.75%), followed by the fish B. polylepis (1.92%) and the fish L. lagocephalus (1.87%) (Figure 4F). The difference between the numbers of stomachs analysed between seasons is because the species has an affinity for warmer water; its abundance was therefore highest during the warm season.
The analysis of similarity of diet composition by sex (ANOSIM, PRIMER 6 v. 6.1.6) showed no differences by gender (R = 0.006, P = 0.001), by sexual maturity (R = 0.035, P = 0.001), or by season (R = 0.044, P = 0.001).
In general, sailfish can be characterized as specialist predators, given that the estimated diet-breadth value was low (Bi = 0.22). This same pattern was consistent for males and females (Bi = 0.23 and Bi = 0.26), immature and mature organisms (Bi = 0.26 and Bi = 0.22) and for the warm and cold seasons (Bi = 0.19 and Bi = 0.18). The Costello plots based on the number of prey (%N) showed that the most frequent and abundant prey was the fish A. thazard, which confirms that sailfish are specialist predators (Figure 5).
Fig. 5. Prey specific abundance (N) as plotted by numerical values against frequency of occurrence (FO%) of prey species in the diet of the sailfish.
DISCUSSION
In studies of the stomach contents of fish and of other marine organisms (e.g. dolphins and whales), biases caused by the different retention times and degradation of consumed food can occur. These biases are common when contents include prey species such as cephalopods, for which the mandibular apparatus (beak) is usually the only vestige found in the stomachs, because cephalopod soft muscles are digested and evacuated rapidly (Robertson & Chivers, Reference Robertson and Chivers1997).
In our study the occurrence of squid beaks and fish in an advanced state of digestion were recorded frequently. This could be because the sailfish captures occur habitually between 0800 and 1400, and landings occur between 1400 and 1800. Evidently the time-period from capture to sample processing allows the digestion processes to continue, also allowing the degradation of organic matter, because gastric enzymes have already been excreted by the predator (Abitia et al., Reference Abitia, Galvan and Muhlia1998).
According to this, the only apparent way to avoid this time bias would be to analyse organisms that have just been captured, to obtain only fresh prey. However, this could limit greatly the sample size, and for predators such as the sailfish that are sampled from the catch made by sport fishermen, this could not be done because of problems inherent to biological sampling and because the fishermen want to take their catch back and weight them. Most studies on fish-feeding habits have biases in the calculation of the prey relative importance, caused by different digestion rates (Hyslop, Reference Hyslop1980; Robertson & Chivers, Reference Robertson and Chivers1997).
The sailfish diet in the southern Mexican Pacific was dominated by teleost fish and to a lesser extent by cephalopods. Sailfish fed preferentially on epipelagic organisms from the oceanic area and to a lesser extent on organisms from the coastal waters. The most common preys were the oceanic species Auxis thazard (which comprised over 60% of the diet) and the coastal fish Fistularia commersonii in a lower proportion (%).
In the general diet, separated by sex, maturity stage, and climatic season, A. thazard made up >60% of the biomass. This ichthyophagous trophic behaviour was corroborated by the Levin's amplitude index, which gave low values (Bi < 0.3). The sailfish in the Acapulco region can therefore be categorized as specialist predators.
This mostly ichthyophagous behaviour was also reported in the first published studies for the species. For example, Jolley (Reference Jolley1977) mentions that about 85% of items consumed by Istiophorus platypterus were fish, and most recently Pimenta et al. (Reference Pimenta, Lima, Cordeiro, Tardelli and de Amorim2005) corroborated this behaviour, reporting that sailfish off Rio de Janeiro, Brazil fed predominantly on A. thazard and Selar crumenophthalmus (Bloch, 1793).
In recent studies conducted in the north and central Mexican Pacific, Rosas et al. (Reference Rosas, Hernandez, Galvan, Abitia and Muhlia2002) reported that the species has a generalist behaviour because it feeds on 78 prey items, although according to the IRI only four of them comprised over 80% of the diet: Dosidicus gigas (Orbigny, 1835) (37.3%); Auxis spp. (19.7%); Argonauta spp. (16.9%); and Scomber japonicus (Houttuyn, 1782) (7%).
Arizmendi et al. (Reference Arizmendi, Abitia, Galvan and Trejo2006) for the Sinaloa coast (northern Mexican Pacific) concluded that despite the high diversity of prey items consumed (62), the sailfish had a low diet breadth (Bi = 0.02) caused by the dominance of a low number of prey, and were therefore considered specialist predators that fed mainly on the cephalopods D. gigas and Argonauta spp.
Romero (unpublished data) reported that sailfish off the coast of Oaxaca (southern Mexican Pacific) can be categorized as specialists, caused by a low diet breadth as estimated by the Levin index (Bi < 0.6). In total the trophic spectrum was made up of 32 prey items, of which the most important were the fish A. thazard, Vinciguerria lucetia (Garman, 1899) and Euthynnus lineatus (Kishinouye, 1920). It is therefore necessary to specify the trophic behaviour of this fish, because, as has been mentioned, it can feed on a wide diversity of prey. However, the dominance of a reduced group of food components is evident and it can therefore be categorized as a specialist predator because it feeds on A. thazard and related species but takes advantage of the most abundant prey (Gerking, Reference Gerking1994).
According to the information generated for the Mexican Pacific, one difference for the feeding habits of sailfish was detected. Off the coasts of the States of Guerrero and Oaxaca (southern Mexican Pacific), an ichthyophagous trophic behaviour was characterized, with the scombrid fish A. thazard making up the largest prey biomass, whereas off the coasts of the States of Jalisco, Colima, Sinaloa and Baja California Sur (north and central Mexican Pacific) a teutophagous behaviour has been reported, with the cephalopod D. gigas as the most important prey (Figure 6).
Fig. 6. Comparison of trophic spectrum (index of relative importance (IRI)) of main prey found in different studies of sailfish in the Mexican Pacific (MP) area. North and central MP: (1) Arizmendi et al. (Reference Arizmendi, Abitia, Galvan and Trejo2006); (2) Rosas et al. (Reference Rosas, Hernandez, Galvan, Abitia and Muhlia2002) and south MP; (3) present study; (4) Romero (unpublished data).
This implies that independently of the prey items consumed by sailfish, they show the characteristic behaviour of preying on school-forming prey. This strategy occurs in fish from the Istiophoridae family, which possess large muscular mass, high metabolic rates and a reduced coelomic cavity and stomach, and therefore need to feed constantly. The consumption of this type of prey increases the possibility of prey capture, reducing the energy spent feeding, and maximizing the consumption and storage of energy (Abitia et al., Reference Abitia, Muhlia, Cruz and Galvan2002; Vaske et al., Reference Vaske, Vooren and Lessa2004).
The ANOSIM test showed that there were no significant differences in the diet between sex, maturity stage, or seasonal climate. Because the analysis was made with data for which most prey were identified to the lowest taxonomic level and with a representative sample size, we consider that the analysis was optimal and reflects clearly the fact that only one prey, the scombrid fish A. thazard, contributed to over 80% of the dietary relative importance for males and females, mature and immature organisms. The same trophic coincidences occur among males, females, and different-sized organisms in the study by Arizmendi et al. (Reference Arizmendi, Abitia, Galvan and Trejo2006), who reported a high trophic overlap of up to 0.96 calculated with the Morisita–Horn index.
There were no significant differences in the stomach contents of the sailfish between the warm and cold periods of 2008–2009, and the higher consumption of fish over cephalopods was maintained. Though the sailfish capture data in the Acapulco region revealed that the species is always present in the area, and is more abundant during the warm period (spring–summer), there were no differences in prey preference between the warm and cold seasons. We can infer that in the southern Mexican Pacific region there is stability in prey abundance caused by the oceanographic characteristics of the area.
Finally, we conclude that because of the dominance of a reduced number of prey and despite the consumption of a high diversity of prey items, the trophic behaviour of the sailfish I. platypterus in the Mexican Pacific can be characterized as specialist.
ACKNOWLEDGEMENTS
We are grateful to the Consejo Nacional de Ciencia y Tecnología and Instituto Politécnico Nacional for funding this work (SEP-CONACYT 60376). S.B.H.A. thanks the Programa Institucional de Formación de Investigadores (PIFI-IPN) for funding. L.A.A.C. and M.A.M. thank the COFAA-IPN and EDI-IPN for fellowships granted, and X.G.M.S. thanks the Consejo Nacional de Ciencia y Tecnología (CONACyT) through its Programa Apoyos Complementarios para la Consolidación Institucional de Grupos de Trabajo (Modalidad, Retención). We thank Dr Ellis Glazier for editing this English-language text.
