Feeding behavior and trophic interaction of three shark species in the Galapagos Marine Reserve

Study area

The Galapagos Islands are located 960 km from mainland Ecuador in the Pacific Ocean. The waters associated with this island complex form a marine reserve delineated by a “baseline” linking the outer edges of the islands to a distance of 74 km, creating a protected area of about 138,000 km² (Heylings, Bensted-Smith & Altamirano, 2002) (Fig. 1). The archipelago’s unique oceanographic setting is believed to be largely responsible for the sporadic colonization of the islands, which led to the evolution of the divergent species that can be observed today in the archipelago’s ecosystems (Ryan et al., 2006; Nims et al., 2008). It is the Cromwell and Humboldt ocean currents that bring most nutrients to the region, generating, around the islands and seamounts, areas of continuous upwelling that induce phytoplankton and zooplankton blooms (Palacios et al., 2006; Schaeffer et al., 2008), hence increasing species richness and diversity in the region.

Sample collection

This research was performed under the research permits: PC-13-12, PC-52-13 and PC-38-16; and was carried out following the protocols of ethics and animal handling approved by the Galapagos National Park and the Ecuadorian laws.

Sampling was done on board a boat seized in July 2011 by the Galapagos National Park authorities and the Ecuadorian Navy while illegally fishing sharks in the northwest region of the Galapagos Marine Reserve (0°10′18″N, 89°21′58″W) (Fig. 1). The fishing boat “Fer Mary I” from Manta, Ecuador was equipped with a 370-hook longline fishing gear and six 8-m fiberglass boats with outboard motor, to check the longline for sharks. A total of 380 sharks found in the boat’s hold were seized. For each specimen, the species and sex were identified and age group was determined based on body size. Samples of muscle tissue were taken from a total of 91 adult sharks belonging to the three species mentioned above (A. pelagicus, P. glauca or C. falciformis) (Table 1). All remaining shark material was then destroyed, as required by Ecuadorian laws.

The samples were washed with distilled water, placed into 20-ml Eppendorf tubes and then frozen to −20 °C. To confirm that all sharks were adult, the total length (TL) and precaudal length (PCL) of each specimen were measured. Total length could not be measured in A. pelagicus, however, because the upper lobes of caudal fins had been cut off; it was thus calculated using the allometric equations proposed by Carr et al. (2013).

Sample processing

All muscle tissue samples were rinsed with deionized water to eliminate residues that could alter their isotopic signature, and placed in glass vials previously treated for 24 h with a chromic acid mixture prepared from sulfuric acid and potassium dichromate. They were then dried in a desiccator at 80 °C for 12 h to remove all moisture. A microwave-assisted extraction protocol (MAE) was applied (Microwave oven model: 1000-W MARS 5x, CEM, Matthews, USA) using 25 ml of a 1:1 chloroform/methanol solution (Tieszen et al., 1983) and dried again. The samples were homogenized with an agate mortar to obtain a very fine powder, of which ∼1 mg was weighed by means of an analytical microbalance with a precision of 0.001 mg and transferred into a tin capsule for isotopic analysis.

δ¹³C and δ¹⁵N stable isotope ratios were determined by a PDZ Europa 20–20 continuous-flow isotope-ratio mass spectrometer (Sercon Ltd., Cheshire, UK) at the Stable Isotope Facility of the University of California at Davis (CA, USA). The results, expressed in parts per thousand (‰), were obtained using the following equation: δ¹³C or δ¹⁵N = 1000([R_sample∕R_standard] − 1), where R_sample and R_standard are the ¹³C/¹²C or ¹⁵N/¹⁴N ratios of the sample and the standard, respectively. The standards used were Pee Dee Belemnite (PDB) for δ¹³C and atmospheric N₂ for δ¹⁵N.

Data analysis

Data were tested for normality and homoscedasticity using the Shapiro–Wilk and Levene test, respectively. The statistical significance of differences in δ¹³C and δ¹⁵N values was determined using parametric or non-parametric tests, and reported when P < 0.05. All statistical analyses were performed using the software Statistica 8.0.

The bayesian method SIBER (Stable Isotope Bayesian Ellipses in R) was used to define the isotopic niche space among the three species, as a measure of their isotopic resource use area at the population level. This method is based on the two-dimensional isotopic space of δ¹³C and δ¹⁵N, and assessed using Bayesian analysis of standard ellipses; that unlike the Euclidean methods (e.g., convex hulls), can incorporate uncertainties such as sampling biases and small sample sizes into niche metrics (Layman et al., 2007). We used Monte Carlo simulations to correct the bivariate ellipses (δ¹³C and δ¹⁵N), surrounding the data points in the 95% confidence interval for the distributions of both stable isotopes (Jackson et al., 2011). These corrected standard ellipse areas (SEAc) represent the isotopic niche width and the overlap parameters (Jackson et al., 2011). Furthermore, we calculated the magnitude of the isotopic overlap among the three species of sharks based on 100,000 posterior draws of the SEAc parameters (Jackson et al., 2011).

The trophic position of each shark species was estimated by the formula proposed by Post (2002): TL = λ + ((δ¹⁵N_predator∕δ¹⁵N_base)∕Δ_N); assuming that the δ¹⁵N of a consumer and other components of the food web provides information on a species’ trophic level. The values used were the trophic position of the base species (λ), the δ¹⁵N value of zooplankton occurring in the region (as base), value previously reported by Páez-Rosas et al. (2012), and the isotopic fractionation factor (Δ_N) for marine predators in general, established by Hobson et al. (1996).

Isotopic comparison of species and sexes

The mean estimated δ¹³C and δ¹⁵N values in the muscle tissue of A. pelagicus were −16.59 ± 0.44‰ and 11.89 ± 1.32‰; in the tissue of P. glauca, −16.66 ± 0.05‰ and 13.46 ± 0.92‰; and in the tissue of C. falciformis, −16.82 ± 0.25‰ and 13.99 ± 1.42‰, respectively (Table 1). The C/N ratios of the samples ranged from 2.8 to 3.2, and were thus within the theoretical range established for the assimilation of protein from a predator’s diet (McConnaughey & McRoy, 1979); therefore the isotopic values reflects the diet of these predators (Table 1). There were no significant interspecific differences in δ¹³C values (Kruskal–Wallis test, p = 0.09). The δ¹⁵N values, in contrast, were significantly different between shark species (Kruskal–Wallis test, p = 0.01), the δ¹⁵N of A. pelagicus differing from those of P. glauca and C. falciformis (multiple comparisons of median ranks, p < 0.05) (Fig. 2).

The mean δ¹³C and δ¹⁵N values of A. pelagicus, P. glauca and C. falciformis are shown by sex in Table 1. None of the three species showed significant differences in δ¹³C and δ¹⁵N values between the sexes (A. pelagicus: Mann–Whitney U test, p = 0.22 and 0.23, respectively; P. glauca: paired t-test, p = 0.81 and 0.13, respectively; and C. falciformis: Mann–Whitney U test, p = 0.49 and 0.76, respectively) (Fig. 3). The comparison by sex of δ¹³C values between the species showed that the females of P. glauca (mean −16.65‰) had higher δ¹³C than those of A. pelagicus (−16.72‰) and C. falciformis (−16.77‰), but these differences were not significant (Kruskal–Wallis test, p = 0.09) (Table 2). The δ¹⁵N values of the females of A. pelagicus (12.04‰), in contrast, significantly differed from those of the females of P. glauca (13.76‰) and C. falciformis (14.09‰) (Kruskal–Wallis test, p = 0.01) (Table 2). With regards to males, the only significant difference in δ¹⁵N values was between A. pelagicus (12.65‰) and C. falciformis (14.26‰) (Kruskal–Wallis test, p = 0.01) (Table 2).

Trophic level, niche breadth and niche overlap

The average estimated trophic level was 3.8 for A. pelagicus, 4.1 for P. glauca and 4.4 for C. falciformis, making these shark species tertiary or quaternary predators in the food web.

The corrected standard ellipse area (SEAc) in SIBER showed that A. pelagicus and C. falciformi s could be exploiting different types of habitat unlike P. glauca (Table 3). These results suggest that feeding habitat of P. glauca may be limited to a specific area, as opposed to the other species. The Bayesian ellipse of C. falciformis (SEAc = 1.14‰, 95% credibility interval of 0.82–1.46‰) and A. pelagicus (SEAc = 1.55‰, 95% credibility interval of 0.93–2.17‰) have a minimal overlap (Table 3 and Fig. 4), confirming different resource use patterns for these two groups of sharks. In contrast, the Bayesian ellipse of P. glauca (SEAc = 0.73‰, 95% credibility interval of 0.28–1.18‰) is overlapped in large part with the ellipses of the other two species (Fig. 4). The overlap area (0.33%) of the Bayesian ellipses from C. falciformis and P. glauca represented the 30.1% of the ellipse surface of the former and the 46.5% of the ellipse surface of the latter. Conversely, the overlap area (0.26%) of the Bayesian ellipses from A. pelagicus and P. glauca represented only the 19.7% of the former and the 48.3% of the ellipse surface of the P. glauca (Fig. 4).

When taking sex into account, a significant isotopic overlap was observed between males and females in A. pelagicus (1.00) and C. falciformis (0.69). While males and females of P. glauca, showed only a small isotopic overlap (0.36) (Table 4 and Fig. 5). In all cases, with the exception of P. glauca, the Bayesian ellipses of males are larger and encompass most of the Bayesian ellipses of females of the same species (Fig. 5), suggesting a higher diversity of foraging strategies in males compared to females in A. pelagicus and C. falciformis.

Interspecific differences

It is widely known that resource use by wildlife species may vary according to a number of factors, e.g., body size, sex or age. Depending on the availability of resources (e.g., space, food) in the ecosystem, this may lead to interspecies or within-species resource segregation (Bolnick et al., 2003; Matich, Heithaus & Layman, 2011).

In this study, no differences were observed in the foraging habitats among the three shark species. The δ¹³C values (between −17.2 and −16.2) indicate an oceanic/pelagic foraging strategy, consistent with what was previously reported in various studies conducted in the coasts of Ecuador, Colombia, Chile and Mexico (Polo-Silva et al., 2013; Hernández-Aguilar et al., 2015; Estupiñán Montaño et al., 2017a; Estupiñán Montaño et al., 2017b; Klarian et al., 2018). Spatial variation in δ¹³C values of predators can be partly explained by differences in the isotopic composition of primary producers, which are the main energy suppliers for the food web (Goericke & Fry, 1994; Pancost et al., 1997). Indeed, different levels of primary production, concentration of dissolved CO₂, macro and microalgae composition and phytoplankton growth rate, among other factors, create a coast–ocean gradient in C-isotopic composition, with a decrease in δ¹³C from the coastal to the pelagic zone (France, 1995; Newsome et al., 2007).

These variability factors are present in the marine ecosystem of the Galapagos Islands which, despite being located in tropical (oligotrophic) waters, are characterized by highly productive local conditions due to the “island-mass effect” (Palacios, 2002), the combination of oceanographic currents flowing there generating high productivity around the islands (Feldman, 1986; Palacios et al., 2006). A δ¹³C-enriched isotopic signature, which was not observed in this study, would thus only be expected if the three shark species had foraged close to the insular shelf.

The δ¹⁵N values revealed differences in the trophic position occupied by the three shark species, A. pelagicus being a lower trophic-level predator compared with the other two species. The consumption of prey from different environments (epipelagic or mesopelagic) and different trophic levels is reflected by δ¹⁵N values that differ between populations exploiting similar habitats (Takai et al., 2000). These differences in δ¹⁵N values may be due not only to differential consumption of prey, but also to isotopic differences at the base of the food web (Newsome et al., 2007). Such isotopic differences manifest themselves at higher trophic links through bioaccumulation of δ¹⁵N from prey to consumer and the resulting isotopic enrichment along the food chain (Minagawa & Wada, 1984; Newsome, Clementz & Koch, 2010).

In Ecuadorian waters, Alopias pelagicus has been shown to feed on squid and small fish (Polo-Silva, Rendón & Galván-Magaña, 2009; Polo-Silva et al., 2013), with a preference for epipelagic squids like Sthenoteuthis oualaniensis (Galván-Magaña et al., 2013). The reported prevalence of these preys suggests that A. pelagicus forages in epipelagic waters, possibly at night when squid are more likely to be caught (Markaida & Sosa-Nishizaki, 2010). While P. glauca is known to be mainly teuthophagous in the North Pacific (Kubodera, Watanabe & Ichii, 2007; Hernández-Aguilar et al., 2015), recent studies carried out in South Pacific mention that its diet is based in pelagic fish, reaching even to consume marine mammals that are dead as dolphins (Klarian et al., 2018). This shark has a greater diving capacity than the other two species, allowing it to explore various habitats along the water column (Carey, Scharold & Kalmijn, 1990; Roper, Sweeney & Hochberg, 1995). It even undertakes daily vertical migrations to depths of over 600 m in order to feed on mesopelagic cephalopods occupying a high trophic level, like Vampyroteuthis infernalis and Ancistrocheirus lesueurii (Kubodera, Watanabe & Ichii, 2007; Galván-Magaña et al., 2013).

The highest δ¹⁵N values were found in Carcharinus falciformis, making it the highest trophic-level predator of the three species. This result agrees with other trophic studies that found this shark to have a certain preference for high-energy oceanic prey like fish of the Scombridae family (Duffy et al., 2015; Estupiñán Montaño et al., 2017b). Consistent with these references and the isotopic values observed in this study, it could be assumed that C. falciformis spend most of its time in surface waters at a depth of ∼50 m, conditions that are usually favorable for the aforementioned preys (Filmalter et al., 2010).

Intersexual differences

A dietary study of populations of A. pelagicus occurring along the coasts of Ecuador found that their diet varies with stage of sexual maturity, but that male and female adults feed on the same prey, mostly juvenile squids of the species D. gigas and S. oualaniensis (Polo-Silva, Rendón & Galván-Magaña, 2009). Isotopic analyses of adult specimens caught by the fishing fleet at the Ecuadorian ports of Manta and Santa Rosa showed that male adults have higher δ¹⁵N values than females. Generally in most shark species the females are bigger in size compared to males, this condition allows them explore more feeding zones. While the males normally stay close the bottom to search protection and easier prey to catch (Compagno, 1990; Polo-Silva et al., 2013). These characteristics could influence that the males consume larger prey associated with the bottom, habitats that are enriched in ¹⁵N, so this behavior can explain this difference.

The females of A. pelagicus had δ¹³C and δ¹⁵N values that were slightly more negative and with smaller variance than those of males. This suggests that females adopt a foraging strategy that is more oceanic and focused on specific prey, while males exploit different ecosystems (oceanic–coastal), with a higher diversity of foraging strategies. Isotopic variability within a population is known to reflect, up to a certain point, dietary heterogeneity among individuals, their high mobility among feeding areas allowing them to consume different prey (Estrada et al., 2003; Jaeger et al., 2009).

The populations of P. glauca in the North Atlantic, is known to exhibit spatial segregation of the sexes, the males tending to frequent coastal waters more intensively than females, which are completely oceanic (Carey, Scharold & Kalmijn, 1990; Vandeperre et al., 2014). This may influence the level of consumption/abundance of their main prey (Henderson, Flannery & Dunne, 2001; Hernández-Aguilar et al., 2015). This differential behavior of males and females could explain what was observed in this study, since despite not observing statistical differences; the females had higher δ¹⁵N values than males. In light of the above studies, the spatial heterogeneity in δ¹⁵N values between sexes can be explained to some extent by the Galapagos Islands’ oceanographic setting, which allows high levels of primary productivity—not characteristic of the region—to be maintained throughout the year (Banks, 2002; Palacios et al., 2006).

With regards to the feeding behavior of C. falciformis, both isotopes showed no intersexual difference, which suggests that both sexes exploit similar prey and foraging habitats. Dietary studies conducted on the eastern Pacific Ocean has reported fish of the Scombridae family and squid as the species’ main prey (Duffy et al., 2015; Estupiñán Montaño et al., 2017a; Estupiñán Montaño et al., 2017b). Other authors have found males of C. falciformis to be more active than females, undertaking vertical migrations in the evening to complement their diet with squid and rest there at night (Compagno, Dando & Fowler, 2005). Such a behavior may explain the high variability observed here in the δ¹³C and δ¹⁵N values of males.

Trophic level, niche breadth and niche overlap

One method to estimate trophic levels is to use the isotope fractionation occurring in the food chain. The formula proposed by Post (2002), used in this study, allowed A. pelagicus and P. glauca to be classified as tertiary predators, and C. falciformis, as a quaternary predator (Mearns et al., 1981). The trophic levels (TLs) obtained for A. pelagicus (TL = 3.8) and P. glauca (TL = 4.1) agree with the diet previously reported for both species, based mainly on small or juveniles squid and fishes of intermediate trophic levels (Polo-Silva et al., 2013; Hernández-Aguilar et al., 2015). A higher trophic level was estimated for males in A. pelagicus, which may be associated with them having a greater ability to approach coastal areas, where they can feed on fish that are larger and occupy a higher trophic level (Rosas-Luis et al., 2017).

The highest trophic-level predator was C. falciformis (TL = 4.4), which agrees with the feeding habits previously reported for this species, consisting in targeting large carnivorous fish such as tuna, black skipjack and supplement their diet with other prey like cephalopods (Duffy et al., 2015; Estupiñán Montaño et al., 2017b). Such as feeding behavior could explain the species’ heavy isotopic signature and allows to put forward the hypothesis that these sharks feed on other high trophic-level predators occurring in the region, e.g., Galapagos sea lions, which have δ¹⁵N values ∼1.5‰ lower (equivalent to prey-to-predator fractionation) than those of C. falciformis (Páez-Rosas & Aurioles-Gamboa, 2014). This would explain the origin of the attack marks (shark bites) that are observed in the pinnipeds of Galapagos Islands in different regions of the archipelago.

Sharks have long been classified as opportunistic predators, feeding on the resources available at a given place and time (Calow & Tytler, 1985). In recent years, it has been shown for some shark’s species that although their food spectrum includes a large variety of prey, the largest part of their diet is made up of three or four prey species, which may change with the seasons (Polo-Silva et al., 2013; Loor-Andrade et al., 2015). The trophic breadth of the three shark species was calculated based on isotopic niche breadth, resulting in a generalist type of feeding strategy for A. pelagicus and C. falciformis, and a specialist strategy for P. glauca. It is possible that given the great diving capacity that P. glauca (Roper, Sweeney & Hochberg, 1995), this species is only using a specific habitat, being more selective in the use of the resources that exist in that space.

This result about A. pelagicus differs from those of other studies that reported this species as a specialist predator based on stomach content analysis (Polo-Silva, Rendón & Galván-Magaña, 2009; Polo-Silva et al., 2013). Several studies mention that A. pelagicus shows some variability in trophic niche breadth over its life, depending on its stage of sexual maturity and the resources available in the environment (Compagno, 1990; Gerking, 1994). It is thus likely that these sharks tend towards a specialist feeding strategy as their energy requirements increase (e.g., for reproduction, gestation, etc.).

The narrower trophic niche breadth estimated for P. glauca suggests that this species exploits specific prey and environments, consistent with the results of Hernández-Aguilar et al. (2015) and Klarian et al. (2018) who found that populations of P. glauca off the coasts of Baja California and Chile show a high degree of specialization, as evidenced by the reduced number of food components that are well represented in their diet.

This study allowed C. falciformis to be classified as a generalist predator in the Galapagos Marine Reserve. It has been shown, however, that when there is a lower diversity of prey in the environment, these sharks tend to be more selective in choosing them (Cabrera-Chávez-Costa, Galván-Magaña & Escobar-Sánchez, 2010; Filmalter et al., 2010). Other authors have described this shark as a specialist consumer, because despite feeding on a wide food spectrum, a large proportion of its diet is made up of a limited number of prey types (Duffy et al., 2015; Estupiñán Montaño et al., 2017a). This may explain to some extent the great variability in δ¹⁵N values observed in this study.

Population-level approaches have often used trophic position and niche breadth to understand interactions within the food web and the role of predators in the ecosystem (Bearhop et al., 2004). Isotopic niche analysis showed have a minimal overlap between A. pelagicus and C. falciformis, which suggests that both species feed on prey from different trophic levels and different feeding areas (Jackson et al., 2011). In contrast, the small (but non-significant) overlap between the isotopic ellipse of P. glauca and those of the other two shark species reveals the consumption of a similar combination of prey and use of the same feeding areas (Matich, Heithaus & Layman, 2010). P. glauca would thus be more susceptible to competition for food resources (e.g., squid) and even feeding areas (Newsome et al., 2007; Graham, Spalding & Sheppard, 2010).