Carbon and nitrogen discrimination factors for elasmobranch soft tissues based on a long-term controlled feeding study

Abstract

The foraging ecology of elasmobranchs (sharks, skates, and rays) is difficult to study because species have spatially and temporally diverse diets. Many diet and habitat preference studies for mammals, birds, and teleosts use stable isotope analysis, but interpretations are limited for elasmobranch studies because taxon-specific isotope discrimination factors from a controlled experiment are unavailable. Trophic discrimination factors for plasma, red blood cells, and muscle were determined from an experiment with leopard sharks (Triakis semifasciata) fed a constant diet of squid over 1000 days. The δ¹³C values for shark tissues at equilibrium with the squid diet did not vary significantly among individuals, but plasma and red blood cell δ¹⁵N values differed significantly among individuals and sampling day. Individual variation of muscle δ¹⁵N averages was observed and likely related to growth. Overall, carbon and nitrogen discrimination factors corresponded to previous studies featuring high-protein diets and carnivorous taxa. The muscle-to-diet discrimination factors from the controlled feeding study were applied to blue sharks (Prionace glauca) and smooth hammerhead sharks (Sphyrna zygaena) caught offshore from Baja California, Mexico. This case study demonstrates the potential of stable isotope analysis to illuminate differences in foraging patterns between elasmobranch species.

Introduction

Elasmobranchs occupy a wide range of habitats as meso- and apex predators in marine ecosystems, but relatively little is known about individual species’ abundance, diet, or movement (Dulvy et al. 2000, 2008). Resolving aspects of foraging ecology within populations or among species will help illuminate trophic structure and food web dynamics, which is especially important in ecosystems affected by the recent sharp decline in shark populations (Baum et al. 2003; Ward and Myers 2005; Myers et al. 2007; Ritchie and Johnson 2009). Elasmobranch dietary studies could benefit from stable isotope techniques, which are routinely used to study the ecology of mammals, birds, and teleosts (Kelly 2000; Vander Zanden and Rasmussen 2001; Crawford et al. 2008; Inger and Bearhop 2008; Newsome et al. 2010). However, the interpretation of elasmobranch ecology from stable isotope analysis is hindered by the lack of specific trophic discrimination factors (i.e., the isotopic difference between consumer tissues and diet; Fisk et al. 2002; Ostrom et al. 1993; Estrada et al. 2003; Domi et al. 2005), a biological parameter that should be calculated from long-term controlled feeding experiments or well-characterized field studies (Martínez del Rio et al. 2009).

Stable isotope analysis is a biogeochemical tool that can supplement dietary and movement data from traditional methods (i.e., observation, stomach/gut content analysis, and tagging), especially for wide-ranging and elusive taxa, such as elasmobranchs. With robust estimates of discrimination factors and tissue turnover rates, ecological studies can use stable isotope ratios of carbon (¹³C/¹²C) and nitrogen (¹⁵N/¹⁴N) to help resolve a consumer’s diet composition and trophic level, as well as ontogenetic shifts in these variables. Furthermore, comparisons among species and populations through time can elucidate patterns of trophic interaction and baseline changes (i.e., Estrada et al. 2003; Layman et al. 2007; Newsome et al. 2007).

Currently, there are no published discrimination factors from a controlled feeding study on elasmobranchs. Although teleosts may share similar habitats and lifestyles to elasmobranchs, physiological differences in osmoregulatory and buoyancy control may lead to differences in trophic discrimination factors between these taxa. In addition, Chondrichthyes, the clade containing elasmobranchs, is the sister taxon to the Teleostomi, the clade containing teleosts and tetrapods such as mammals and birds, which are the main subjects of controlled feeding studies. A previous study has published elasmobranch-specific discrimination factors, but the isotopic values of prey were modeled rather than directly measured (Hussey et al. 2010). In this study, a controlled feeding experiment was conducted and the carbon and nitrogen isotope values of leopard shark (Triakis semifasciata) tissues (plasma, red blood cell [RBC], and muscle) and their diet were monitored for 1000+ days. This extensive data set allowed us to determine trophic discrimination factors and critically evaluate isotopic variation within tissues and among individual sharks. In addition, a case study is presented comparing the results of stomach/gut content analysis with those of stable isotope analysis to assess the foraging ecology of blue (Prionace glauca) and smooth hammerhead (Sphyrna zygaena) sharks caught offshore from Baja California, Mexico.

Overview of stable isotope analysis

Carbon and nitrogen isotope values track nutrient-flow through a food web and vary spatially at the base of the food web in response to differences in productivity, upwelling, and other oceanographic factors (Saupe et al. 1989; Clementz and Koch 2001; Post 2002; Barnes et al. 2009). Isotopes of both of these elements are sorted (or fractionated) during metabolic processes, such that consumer tissues become enriched in the rare heavier isotopes (¹³C and ¹⁵N) relative to prey (Wada et al. 1991; Martínez del Rio et al. 2009). Isotope ratios will be discussed using δ values, where ${\delta^{\text{h}}}{\text{X}} = \left( {\left( {{{\text{R}}_{\text{sample}}}/{{\text{R}}_{\text{standard}}}} \right) - {1}} \right)*{1}000$ and X is the element, h is the high mass number, R is the high mass-to-low mass isotope ratio, and R_standard is V-PDB for carbon and AIR for nitrogen. Units are part per thousand (per mil, ‰). The difference between prey and a consumer’s tissue is referred to as the trophic discrimination factor, with the following notation and relationship: ${\Delta^{\text{h}}}{{\text{X}}_{{{\text{tissue}} - {\text{prey}}}}} = {\delta^{\text{h}}}{{\text{X}}_{\text{tissue}}} - {\delta^{\text{h}}}{{\text{X}}_{\text{prey}}}$ (Martínez del Rio et al. 2009).¹

A review of data for aquatic and terrestrial invertebrates and vertebrates from laboratory and field studies reported average Δ¹³C and Δ¹⁵N values of 0.4‰ and 3.4‰, respectively (Post 2002). Physiological and dietary differences can affect discrimination factors, which depend on the biochemical processes used to metabolize and incorporate nutrients (Gannes et al. 1997; Koch 2007; Martínez del Rio et al. 2009). Nitrogen fractionation occurs during deamination and transamination of amino acids (Macko et al. 1986; Fogel et al. 1997; Gannes et al. 1998) and its magnitude is likely regulated by the concentration of nitrogen lost in excreted waste products (Koch 2007). The processes controlling dietary carbon isotope fractionation are less well understood, but in protein-rich tissues, are likely related to isotopic routing of dietary amino acids during metabolism (Gannes et al. 1997, 1998; Martínez del Rio et al. 2009).

Results from previous feeding studies illustrate that discrimination factors differ between diets in relation to protein quality and quantity (Focken 2001; Pearson et al. 2003; Robbins et al. 2005; Martínez del Rio et al. 2009; Florin et al. 2011). Vertebrate carnivores are rarely kept in captivity for feeding experiments, but published studies featuring carnivorous mammals and teleosts fed a natural carnivorous diet (in captivity or well-known wild individuals) found Δ¹⁵N values for blood (Kurle 2002), collagen (Fox-Dobbs et al 2007) and muscle (Sweeting et al. 2007a) to be greater than the average of 3.4‰ reported by Post (2002) (Table 1). Additionally, there is a growing body of literature featuring omnivores (e.g., bears, spiders, birds, etc.) that suggests carbon and nitrogen discrimination factors increase depending on dietary protein content (Hilderbrand et al. 1996; Oelbermann and Scheu 2002; Pearson et al. 2003; Robbins et al. 2005; Florin et al. 2011). Because most elasmobranchs are carnivores, carbon and nitrogen discrimination factors are likely greater than previously reported average values.

Table 1 Discrimination factors calculated for controlled feeding experiments featuring carnivorous teleosts and mammals. Note that most subjects fed pure carnivorous diets (seals, carnivorous mammals, wolf, and lynx) have greater ∆¹⁵N _tissue-diet values than subjects fed pellets or supplement additives, except sharks that were not extracted of urea (Hussey et al. 2010)

Elasmobranchs retain urea ((NH₂)₂CO) to maintain osmotic balance. The isotopic effects of this physiology are not well understood. Urea isotope values may differ from those of body tissues, so high concentrations of urea in tissues may skew ecological interpretations (Kim and Koch, in press). In addition, the urea cycle in elasmobranchs is different from that in other taxa (reviewed by Ballantyne 1997 and Hazon et al. 2007) and it is unknown if the additional enzymatically-catalyzed step affects urea isotope values in elasmobranchs relative to other taxa. The only urea-related δ¹³C values we found were for human and bovine urine, which were 3–5‰ and 0–3.5‰ greater than diet, respectively (Ivlev et al. 1996; Knobbe et al. 2006). Kim and Koch (in press) reported average 1.3‰ depletion in δ¹⁵N values from elasmobranch muscle containing urea. Additionally, there may be biochemical adaptations specific to elasmobranch physiology that alter isotopic fractionation patterns and ultimately require calculation of elasmobranch-specific discrimination factors.

Case study

Blue and smooth hammerhead sharks are seasonally abundant and frequently caught in artisanal fisheries targeting sharks in the northeast Pacific offshore from Baja California, Mexico (Cartamil et al. 2011). Blue sharks (Prionace glauca) are a cosmopolitan pelagic species with a wide distribution throughout the Pacific. The species is thought to segregate by size and sex with extensive migration throughout the north Pacific and Atlantic (Nakano and Stevens 2008). Recent stomach/gut content analysis revealed cephalopods and red crab to be the primary prey of blue sharks near our study location (Hernández-Aguilar 2008). Stomach/gut content analysis suggests that hammerhead sharks (Sphyrna zygaena) also have a cephalopod-dominated diet (Ochoa-Díaz 2009), but they mostly inhabit the continental shelf along the California and Baja California coast in the northeast Pacific (Casper et al. 2005). In this case study, estimates of dietary contributions from isotopic analysis are qualitatively and quantitatively compared to stomach/gut content analysis results.

Methods

Controlled feeding experiment sampling

The leopard shark (Triakis semifasciata) is a coastal species found from Washington to Mazatlán, Mexico (including the Gulf of California; Carlisle and Starr 2009; Farrer 2009) that is ideal for experimental study; individuals are relatively small (<2 m), locally abundant, and relatively easy to maintain in captivity. Leopard sharks were caught in otter trawls between August–December 2005 in San Francisco Bay, acclimated to captivity, and transported (1–2 sharks at a time) to Long Marine Lab at the University of California, Santa Cruz (UCSC). Two sharks were maintained per polyethylene tank (2.3 m diameter, 1.2 m water depth) with a continuous flow of filtered seawater from the Monterey Bay and fed squid (Loligo opalescens) from Monterey Bay. Data for three control individuals maintained on a constant diet from a larger diet-switching experiment are presented here.

From February 2006 to November 2008, blood and muscle were sampled approximately every 21 days. Sharks were individually placed in a smaller tank (1.2 m diameter, 0.2 m water depth) where an anesthetic, tricaine methanesulfonate (MS-222) was dissolved. Once the shark lost mobility, length and weight measurements and a 0.7 cc sample of blood was taken (25-gauge needle into the caudal vein) and sub-divided into a no-additive Vacutainer^® and 2 hematocrit tubes (capillary glass tubes filled with heparin). To prevent coagulation and collect well-separated plasma samples, the whole blood sample was immediately centrifuged and the plasma was transferred into a glass test tube. Muscle biopsies were taken above the lateral line between the first and second dorsal fins with a histo-cut needle (14 gauge × 16 cm, Avid Medical Supplies). All samples were stored in glass tubes and immediately frozen until isotopic analysis. After sampling, hematocrit tubes were centrifuged and packed cell volume (length of RBC in tube/length of whole blood in tube) was measured to monitor the sharks’ health.

Case study sampling

The blue and hammerhead sharks were caught off the coast of Baja California Sur, Mexico, in July 2004 and September 2005. Artisanal fishermen used long-line techniques to capture individuals. All blue sharks (n = 10; total length [TL] range = 136–260 cm) and one hammerhead (TL = 142.5 cm) were caught offshore from Isla Magdalena (24º 46′N, 112º 15′W) and the remaining hammerheads (n = 10; TL range = 47–143 cm) were caught offshore from Las Barrancas (26º 0′N, 112º 12′W). The average of three TL measurements were taken and ~10 g of muscle tissue were sampled from below the first dorsal of each shark. Samples were kept on ice for up to 5 days until a freezer was available to store the samples. After transport to UCSC, samples were frozen at −20°C until prepared for isotopic analysis.

Stable isotope analysis

All shark tissues were freeze-dried and plasma and RBC samples were not chemically treated before isotopic analysis. Small shark muscle samples (from the feeding study) were packed in Accelerated Solvent Extractor (ASE^®) cells between glass fiber filters (GF/F), lipid extracted with 2 rinses of petroleum ether and urea extracted with 3 rinses of de-ionized water (9 ml solution at 50°C, 1500 Psi, and 60% volume for 5 min; Dobush et al. 1985; Bodin et al. 2009; Kim and Koch in press). Larger muscle samples were lipid extracted with 2 rounds of petroleum ether (~10 ml) and 10 min of sonication and then rinsed with 3 rounds of de-ionized water (~10 ml) and 15 min of sonication to remove urea. All muscle samples were dried in an oven set to 50°C or freeze-dried overnight. The isotopic effects of this lipid and urea extraction method are detailed in Kim and Koch (in press).

All squid fed to the leopard sharks were from a single shipment. Throughout the experiment, 1–3 squid were randomly analyzed from each box fed to the sharks to insure the homogeneity of diet. Each complete individual squid (n = 43) was individually homogenized in a blender, decalcified with 0.5 N hydrochloric acid overnight at 4°C, and lipid extracted with 3 rinses of petroleum ether. Beaks from another 24 squid were dissected, cleaned, and lipid extracted with 5 rinses of petroleum ether (10 mL). All squid samples were dried in an oven set to 50°C, and then further homogenized using a mortar and pestle before isotopic analysis.

All the samples were weighed to 500 μg (±50) into 3 × 5 mm tin boats (Costech Supplies) and analyzed at the Stable Isotope Laboratory at UCSC with an elemental analyzer coupled to a Thermo-Scientific Delta+XP continuous flow, isotope-ratio-monitoring mass spectrometer (CF-IRMS). A gelatin of known C and N isotope composition was replicated within and between runs for corrections. Comparisons of this standard yielded standard deviations of <0.1‰ and <0.2‰ for δ¹³C and δ¹⁵N values, respectively (n = 178). All C:N ratios were also corrected to this gelatin standard and calculated as the atomic ratio.

Data analysis for controlled feeding study

Most elasmobranchs ingest whole prey or large chunks of prey, so discrimination factors were based on the analysis of whole squid samples. However, whole prey isotope values are often unavailable, especially for large prey items, and isotopic values for protein in hard tissues that are preserved in stomach/gut contents (i.e., chitin, keratin, or bone) are often substituted. Although previous studies comparing muscle and collagen have reported similar isotopic patterns (i.e., Sholto-Douglas et al. 1991; Hobson and Clark 1992a, b; Nardoto et al. 2006; Fox-Dobbs et al. 2007), the δ¹³C and δ¹⁵N values of cephalopod muscle and beaks have reported differences of 0.6–1.2‰ and 3.3–5‰, respectively (Hobson and Cherel 2006; Ruiz-Cooley et al. 2006). Whole squid and squid beaks were analyzed to determine a correction factor that converts isotopic data from squid beaks to whole squid.

Isotope values from plasma, RBC, and muscle collected after day 600 were used to calculate trophic discrimination factors. Leopard shark muscle tissue was assumed to be in equilibrium with the squid diet based on the following: 1) calculation of 1σ from all individuals over 3 consecutive sampling days <0.4‰ and 2) the estimate by MacNeil et al. (2006) of 95% muscle turnover in 422 days. Additionally, Logan and Lutcavage (2010) estimated 95% muscle turnover in 555 and 340 days for carbon and nitrogen, respectively, though this estimate is based on assumed discrimination factors and a 46–55 day experiment. Finally, preliminary diet-switching results from leopard sharks indicate that plasma, RBC, and muscle will undergo a minimum of two half-lives in 600 days (Kim et al., unpubl. data).

The sampling of multiple individuals fed a constant diet allowed the determination of individual mean values, estimation of uncertainty, and evaluation of significant differences for prey and shark tissues. For the squid, means and standard deviations (SD) were based on 1–2 squid collected from each box (weight = 2.3 kg) fed to the sharks starting on day 1 in captivity until day 1020, the last sampling day in this data set. Differences between beak and whole squid δ¹³C and δ¹⁵N values were assessed with a Wilcoxon rank sum test, a conservative comparison between two related sample populations that are not normally distributed. For shark plasma, RBC, and muscle, mean isotope values were compared among tissues and isotope systems with a Kruskal-Wallis rank sum test, which is similar to the Wilcoxon test for data containing more than 2 groups. A one-way Analysis of Co-Variance (ANCOVA) assessed if there was a non-random interaction between individual and sampling day on plasma, RBC, and muscle isotope values. If there was no correlation between individual and sampling day, then the significance for each model effect (individual and sampling day) was independently determined with an Analysis of Variance (ANOVA) test. All statistical analysis of these data was performed in JMP^® (version 7.0, SAS Institute Inc. 2007).

A suite of Δ¹³C and Δ¹⁵N values for plasma, RBC, and muscle relative to whole squid were determined for captive leopard sharks fed squid. These values were based on the population mean values (all individuals from days 600–1020 and all squid); SD was propagated for each tissue.

Data analysis for case study

Isotope values between blue and hammerhead sharks were compared using the Wilcoxon rank sum test and ontogenetic relationships between δ¹⁵N values and TL were evaluated with JMP. The diets of blue and hammerhead sharks were estimated from a Bayesian stable isotope mixing model (MixSIR v. 1.0.4; Moore and Semmens 2008). Advantages of this model are 1) the incorporation of uncertainty in discrimination factors and prey sources and 2) the model results are posterior distributions of each source in a potential diet mix. The model parameter details are listed in Appendix A. Each potential dietary mix within the mixing model sums to 100% (Moore and Semmens 2008), but median values of prey contribution distributions may not sum to precisely 100%. The isotope mixing model was performed with two sets of Δ¹³C and Δ¹⁵N values (and SD): the values from this feeding study and those reported by Post (Δ¹³C = 0.4‰ [SD = 1.4] and Δ¹⁵N = 3.4‰ [SD = 1.0]; 2002). The values from Hussey et al. (2010) were not used because the isotopic value of diet was not directly analyzed and urea was not removed from the muscle tissue before isotopic analysis.

Potential prey species were identified in stomach/gut content studies of blue and hammerhead sharks caught in similar locations to those sampled for isotopic analysis (Hernández-Aguilar 2008; Ochoa-Díaz 2009). Potential prey isotope values were found in the literature and unpublished data from samples caught in the same region as the sharks. The prey isotope data were categorized based on taxon and region. Jumbo cephalopod (Dosidicus gigas; Ruiz-Cooley et al. 2006) and coastal cephalopod (caught on the continental shelf—Ancistrocheirus lesueurii, Onychoteuthis banksii, Pholidoteuthis boschmai, Sthenoteuthis oualaniensis, Thysanoteuthis rhombus; Ochoa-Díaz 2009) categories were used in mixing models for both blue and hammerhead sharks. Red crabs (Pleuroncodes planipes; Peckham and Newsome, unpublished data) and sardines (Sardinops sagax, Etrumeus teres; Gendron et al. 2001) were included in the blue shark and hammerhead shark mixing models, respectively, based on stomach/gut content results (Hernández-Aguilar 2008; Ochoa-Díaz 2009). Offshore cephalopods (Ommastrephes bartrami, Berryteuthis anonychus, Octopoteuthis deletron, Histioteuthis dojleini, Taonius pavo; Gould et al. 1997) were included for blue sharks, because sub-adults are thought to migrate to the central Pacific and often caught as bycatch in cephalopod targeted fisheries (Nakano and Stevens 2008). All cephalopod beak data was converted to represent whole cephalopods using the isotopic differences from the squid in the controlled feeding study. The average isotope values and SD for relevant prey categories were entered into MixSIR (Appendix A).

Results

Controlled feeding study

The plasma, RBC, and muscle of leopard sharks had significantly different mean δ¹³C and δ¹⁵N values (Kruskal-Wallis test H_C = 81.9, d.f._C = 2, p_C < 0.0001; H_N = 73.2, d.f._N = 2, p_N < 0.0001; Fig. 1 and Table 2). The C:N ratio (SD) of plasma, RBC, and muscle equaled 1.3 (1.0), 2.2 (0.1), 2.9 (0.1), respectively. The ANCOVA and ANOVA tests for plasma, RBC, and muscle isotope values demonstrated variability attributed to individual and sampling day. The plasma and RBC δ¹⁵N values co-varied for individual x sampling day (p_plasma = 0.0007, p_RBC = 0.0148), but δ¹³C values were not significantly different for these model effects (Table 3). Although the individual model effect for RBC δ¹³C results had a low p-value (0.0547), individual means only varied 0.1‰ throughout the sampling period (Fig. 1b, Table 2), which is within analytical variability. There was no co-variance of individual and sampling day for δ¹³C and δ¹⁵N values of muscle (Table 3). The ANOVA results for individual and sampling day were not significant for muscle δ¹³C values, but the individual effect for muscle δ¹⁵N values had a low p-value (0.0515) and individual means differed (Tables 2 and 3). The growth rates for all three individuals were similar (0.0167, 0.0217 and 0.0255 cm/day). However, total length measurements from sampling days indicated relatively steady growth for two individuals, but one individual (represented in all graphs by “•” symbol) had a reduced growth rate and low hematocrit values (Supplemental Material A).

Fig. 1

Carbon and nitrogen isotope values for (a) plasma, (b) RBC, and (c) muscle for 3 individual leopard sharks (each represented by a different symbol) from January 2006–October 2008. Only data collected after 600 days (shown beyond the vertical dashed line) were analyzed to calculate trophic discrimination factors

Table 2 Plasma, RBC, and muscle isotopic averages across all individuals and sampling days from three leopard sharks fed a constant squid diet over 1000 days. Isotopic averages were compared to whole squid to estimate discrimination factors^a and standard deviations. The statistical significance of individual and sampling day effects are shown in Table 3

Table 3 Plasma, RBC, and muscle mean values and statistical results of ANCOVA and ANOVA tests for individual sharks and sampling day (referred to as “day”) during days 600–1020, when shark tissues were equilibrated to the squid diet. Individual means are listed from the slowest to fastest growth rates and significant effects are indicated with an asterisk (α = 0.05)

Accurate calculation of isotopic discrimination factors requires a diet with a uniform and well known isotopic composition and consumer tissues that are in equilibrium with this diet. The mean δ¹³C and δ¹⁵N values for whole squid (n = 43) and squid beaks (n = 24) were significantly different (Table 2; Z_C = 5.27, p_C = <0.0001; Z_N = −6.77, p_N < 0.0001). The equations to convert beak data to represent whole cephalopods are as follows: δ¹³C_beak- 0.6 = δ¹³C_whole and δ¹⁵N_beak + 4.7 = δ¹⁵N_whole. Isotopic means for each tissue were compared to isotopic means of whole squid and errors were propagated to determine trophic discrimination factors and SD (Table 2).

Case study

Blue shark and hammerhead shark muscle δ¹³C and δ¹⁵N values were significantly different (Wilcoxon rank sum test; Z_C = −2.15, p_C = 0.029; Z_N = 0.0043, p_N = 0.0039). The mean δ¹³C and δ¹⁵N values (SD) for blue shark muscle (n = 10) were −17.3‰ (0.6) and 17.8‰ (0.6), respectively and for hammerhead shark muscle (n = 11) were −16.7‰ (0.4) and 18.9‰ (0.8), respectively. There were linear relationships between blue shark and hammerhead shark δ¹⁵N values and TL (cm; δ¹⁵N_blue = 15.5 + 0.0127× TL_blue [r² = 0.75, d.f. = 8, p = 0.0012] and δ¹⁵N_hammerhead = 17.5 + 0.020× TL_hammerhead [r² = 0.48, d.f. = 9, p = 0.018]; Fig. 2). The smaller hammerhead sharks had a lower mean δ¹⁵N value (by 1.2‰) and a wider variation in δ¹⁵N values than the larger hammerhead sharks (>100 cm TL, Fig. 3b). There was a significant difference between δ¹⁵N values of small and large individuals, but not for δ¹³C values (Wilcoxin Rank Sum test; Z_C = 0.0451, p_C = 0.034, Z_N = 0.91, p_N = 0.81).

Fig. 2

The ontogeny of blue sharks and hammerhead sharks as described by muscle δ¹⁵N values. A linear regression of each species total length by δ¹⁵N values yielded significant results (see Results for equations, r², and p values)

Fig. 3

The qualitative dietary analysis of isotope values from (a) blue sharks and (b) smooth hammerhead shark muscle collected in Baja California Sur, Mexico. Potential prey items were selected based on stomach/gut content analysis from sharks caught in this region (Ochoa-Díaz 2009). Discrimination factors from this study added to the prey isotopic values. Quantitative stomach/gut content and dietary mixing model results are presented in Table 4

Dietary preferences, as dictated by stable isotope analysis, are qualitatively illustrated by adding muscle-to-whole prey discrimination factors from the feeding study to all potential prey data (Fig. 3) and quantitatively evaluated using MixSIR (Table 4). As expected, MixSIR determined different relative contributions of prey based on the discrimination factors used (Table 4). The δ¹⁵N values of blue shark muscle were positioned between those of coastal and offshore cephalopods. Red crabs and jumbo cephalopods were ¹³C-enriched relative to blue shark muscle (Fig. 3a). Mixing model results using different discrimination factors determined smaller median contributions of red crab than stomach/gut content analysis (Table 4). All hammerhead sharks had δ¹⁵N values between those of coastal cephalopods and sardines, similar to jumbo cephalopod (Fig. 3b). The mixing model with discrimination factors from this study and Post (2002) estimated greater median sardine contributions than stomach/gut content results (Table 4).

Table 4 Dietary results for blue and hammerhead sharks from stomach/gut content and isotope mixing model analyses. The discrimination factors used in the two MixSIR analyses were from this study (Δ¹³C = 1.7‰ and Δ¹⁵N = 3.7‰) and Post (2002; Δ¹³C = 0.4‰ and Δ¹⁵N = 3.4‰)

Discussion

Controlled feeding study

• a. Variation among and within individuals

Controlled feeding studies with vertebrates that are difficult to keep in captivity often sample deceased specimens and perform dietary reconstructions to determine prey isotope values (Ambrose and DeNiro 1986; Hobson et al. 1996; Bocherens and Drucker 2003; Fox-Dobbs et al. 2007; Hussey et al. 2010) or sacrifice specimens incrementally throughout their experiment (Pinnegar and Polunin 1999; Trueman et al. 2005). However, these experimental designs may lack accuracy because isotope values of prey are estimated and individuals may vary. In this study, sharks were kept in captivity and tissues were repeatedly sampled for isotopic analysis for 1000+ days. The individual mean δ¹³C values differed by 0.1‰ for all tissues, which is within instrument variability, but individual mean δ¹⁵N values differed by 0.1‰, 0.3‰, and 0.4‰ for plasma, RBC, and muscle, respectively. Despite the small number of subjects, the results of this study illustrate potential individual variability when subjects are fed a constant diet and kept in a controlled environment. Two possible explanations for the variation in individual mean δ¹⁵N values could be differences in either growth rate or urea retention.

Variation among individual growth rates alters the rate or efficiency of nutrient assimilation and therefore, individual δ¹⁵N values (Trueman et al. 2005). All captive sharks were juveniles with similar growth rates and squid consumption, but one individual’s growth rate slowed from days 400–800 (• in all figures and the first individual mean in Table 2; Supplemental Material A). The two healthy sharks in this experiment had an inverse relationship between growth rate and mean muscle δ¹⁵N values, as expected from the results of Trueman et al. (2005). However, the significance of this correlation must be tested with a larger sample size that encompasses adult elasmobranchs, which have lower growth rates after maturity (Araya and Cubillos 2006).

Elasmobranches maintain high concentrations of urea and TMAO, which contain carbon and nitrogen that likely have different isotopic values from tissue. Previous urea concentration measurements for leopard shark plasma revealed large variability (400–650 mM), likely caused by variations between individuals, seasonal water chemistry, or time since last feeding (Wood et al. 2005; Kim and Koch in press). For all individuals and tissues, ANCOVA and ANOVA tests found no significant differences among δ¹³C values, which indicates little variation among individuals and sampling day and suggests the urea and TMAO carbon input has little effect. Although there was a low p-value for individual effects on RBC δ¹³C values, this result is likely an artifact of the one individual with health complications. The statistical results for δ¹⁵N values vary between tissues; plasma and RBC results have significant differences based on ANCOVA tests, but muscle results have no significant differences based on ANCOVA or ANOVA tests. The different responses between these tissues (plasma and RBC vs. muscle) could be related to urea content. Currently, there are no methods to remove urea from blood, although there is a negative correlation between urea concentration and plasma δ¹⁵N values (Kim and Koch, in press). Previous isotopic results from shark muscle indicate an increase in δ¹⁵N values by 0.3–2.2‰ and decrease in δ¹⁵N variation with urea removal techniques (Kim and Koch, in press). Once urea removal techniques for blood are developed, the significant differences associated with individual and sampling day effects may be mitigated for plasma and RBC samples. Future studies using shark plasma and RBC δ¹⁵N values should consider the additional error associated with urea during data interpretation.

• b. Discrimination factors

For shark plasma, RBC, and muscle, the Δ¹³C values are higher than average values but Δ¹⁵N values are similar to average values reported in Post (2002). The quantity and quality of protein an organism consumes relative to its metabolic needs can generate variation in Δ¹³C values (Martínez del Rio et al. 2009). In studies of omnivorous taxa with variable dietary protein, there is a positive linear correlation between protein content and Δ¹³C values (Hilderbrand et al. 1996; Pearson et al. 2003). For all leopard shark tissues, the Δ¹³C values were higher than averages reported in Post (2002) and Sweeting et al. (2007b; Table 1). There are published studies of carnivorous taxa with low Δ¹³C values, which are likely an artifact of high lipid concentration (Kurle 2002) and inaccurate prey value estimates (Bocherens and Drucker 2003; Hussey et al. 2010). Leopard sharks in this study were fed squid, prey with high quality and quantity of protein, and the Δ¹³C value of plasma, RBC, and muscle was greater than previously reported averages in Post (2002).

The Δ¹⁵N values we calculated for elasmobranchs may resemble values for herbivores and omnivores because elasmobranchs retain and recycle nitrogenous waste for osmotic balance. Transamination and deamination lead to ¹⁴N-enrichment in nitrogenous waste (urea, uric acid) relative to body tissues (Macko et al. 1986; Fogel et al. 1997; Sponheimer et al. 2003). Under simple mass balance models, high proportional fluxes of N lost as such waste (rather than as shed tissues or enzymes in fecal matter) lead to ¹⁵N-enrichment of body tissues relative to diet (Gannes et al. 1997; Sponheimer et al. 2003; Koch 2007). However, protein-limited species can decrease urinary nitrogen loss, thereby reducing the nitrogen isotope difference between consumer tissues and diet (Gannes et al 1997; Sponheimer et al. 2003). Elasmobranchs also have a modified digestive physiology—intestinal valves increase the surface area of elasmobranch guts (despite their short length) to maximize nutrient absorption. The specialized physiology of elasmobranchs maximizes nitrogen use efficiency and likely decreases the degree of expression of isotopic fractionation associated with metabolic processes to values more like those for omnivores than carnivores.

• c. Differences in isotope values between whole squid and squid beaks

The ¹³C-enrichment for samples analyzed in this study and by Hobson and Cherel (2006) and the consistent ¹⁵N-depletion in beaks relative to muscle or whole squid across various cephalopod species suggest that these isotopic differences may be an effect of different amino acid compositions between muscle and beak protein (Hobson and Cherel 2006; Ruiz-Cooley et al. 2006). Cephalopod beaks are composed of chitin and contain proteins with high concentrations of glycine and histidine (Broomell et al. 2007). Previous studies analyzing δ¹³C values of individual amino acids in terrestrial invertebrates, terrestrial mammals, and marine birds found glycine to be ¹³C-enriched relative to bulk δ¹³C values (Hare et al. 1991; Fogel and Tuross 2003; O’Brien et al. 2005; Lorrain et al. 2009; studies did not report δ¹³C values for histidine). The relationships between bulk tissue, glycine, and histidine δ¹⁵N values vary to a greater extent and are more difficult to assess because glycine and histidine are considered source amino acids. These amino acids are not affected by trophic fractionation and instead, reflect δ¹⁵N values of primary producers in the food web (McClelland and Montoya 2002). Thus, the relative ¹⁵N-depletion of glycine and/or histidine relative to a bulk tissue sample depends on the consumer’s trophic level. For example, δ¹⁵N values of glycine do not differ from bulk samples in zooplankton (histidine not reported; McClelland and Montoya 2002), but glycine and histidine are ¹⁵N-depleted (~8–13‰) relative to bulk samples in tuna (Popp et al. 2007).

Case study

• a. Ontogenetic shifts

The correlation between blue and hammerhead shark δ¹⁵N values and size (Fig. 2) represent a gradual increase in trophic level (e.g., consumption of larger cephalopods, Ruiz-Cooley et al. 2006) or foraging in habitats with a higher ¹⁵N baseline as animals mature. For example, smaller blue sharks caught near Baja California, Mexico could be recent immigrants from the ¹⁵N-depleted central Pacific near Hawaii and larger individuals could be sub-adults that have been resident in the ¹⁵N-enriched coastal Pacific (Saino and Hattori 1987; Nakano and Stevens 2008). The sampling of more individual blue sharks over a greater size range will clarify potential migrational segregation within the population. Hammerhead sharks do not have large-scale movement patterns on- and offshore, but may migrate latitudinally to continuously occupy temperate waters. The large variation among young hammerhead sharks (TL < 70 cm and umbilical scars) likely reflects opportunistic foraging. Isotopic analysis of more hammerhead sharks >100 cm will help characterize potential ontogenetic dietary shifts in this species.

• b. Isotopic mixing model for blue sharks

The variation in blue shark isotope values likely reflects foraging in coastal and offshore areas that have different baseline isotopic values. Stomach/gut content analysis results suggest that blue sharks have a diet dominated by red crabs and coastal cephalopods (Hernández-Aguilar 2008). The isotope mixing models with discrimination factors calculated in this study and Post (2002) determined a much smaller contribution from red crab. The discrepancy between stomach/gut content and isotope mixing model results may occur because stomach/gut content results only provide a snapshot of diet and cannot account for seasonal variation. Although red crab may dominate blue shark diets in coastal Baja California, Mexico, in February and March (Hernández-Aguilar 2008), their annually integrated contribution to muscle growth may be smaller. The isotope mixing model using discrimination factors from this study estimated a similar contribution of coastal cephalopods as the stomach/gut content analysis (Table 4). The isotope values for red crab, coastal cephalopods, and jumbo cephalopods, the major prey according to stomach/gut content analysis, did not fully characterize blue shark diet. Isotopic analysis of Atlantic blue sharks by Estrada et al. (2003) and MacNeil et al. (2005) demonstrated little seasonal diet variation, but fisheries catch data suggests migratory movement of blue sharks in the north Pacific between coastal and offshore areas (Nakano and Stevens 2008). Blue sharks are frequently caught in Baja’s artisanal fisheries from February–June (Hernández-Aguilar 2008; Cartamil et al. 2011). Because blue sharks are offshore for approximately half the year, offshore cephalopods likely compose a similar proportion of their diet; however, remains are not present in stomach/gut contents from individuals caught nearshore. The mixing model results using the discrimination factors from this study support the extended time blue sharks are not caught in Baja artisanal fisheries.

• c. Isotope mixing model for hammerhead sharks

Stomach/gut content analysis of juvenile hammerhead sharks caught in the eastern Pacific offshore from Baja California Sur, Mexico, revealed a diet dominated by jumbo cephalopod (85%; Ochoa-Díaz 2009). In contrast, isotope mixing models with discrimination factors from this study and Post (2002) suggest smaller contributions of jumbo cephalopods and larger contributions of sardines (Table 4). These differences are likely because cephalopods have beaks made of chitin and are more resistant to digestion than sardine, as evidenced in studies of penguins and grey seals (Wilson et al. 1985; Grellier and Hammond 2006). According to the discrimination factors from this study, coastal cephalopods and sardine dominate hammerhead shark diet, whereas the Post (2002) discrimination factors estimate a diet dominated by coastal and jumbo cephalopods. The analysis of more hammerhead sharks >100 cm would allow an isotope mixing model analysis of ontogenetic dietary switches. Stable isotope analysis may improve our understanding of smooth hammerhead shark diet and habitat preferences, two biological characteristics that are not well known (Casper et al. 2005).

Conclusion

Previous studies suggest using trophic discrimination factors determined from studies featuring similar taxa, diet, and tissue (Gannes et al. 1997, 1998; Martínez del Rio et al. 2009). Although the ∆¹⁵N values from Post (2002) and this study are similar, the difference between average ∆¹³C values (0.4‰; Post 2002) and elasmobranch muscle (1.8‰) may be enough to change some ecological interpretations. Furthermore, it is important to consider the prey tissue analyzed and to apply appropriate discrimination factors. Because the isotopic values of whole cephalopods and their beaks vary, the interpretation of diet and habitat preference is affected by a conversion from beak to whole body or muscle, especially in ecosystems with subtle isotopic variation. Isotopic analysis of blue and hammerhead sharks revealed ontogenetic dietary variation. A comparison of results from stomach/gut content and isotopic analysis illustrated the potential overestimation of prey with hard parts resistant to digestion with stomach/gut content analysis (i.e., chitin and shells).

Appendix A

Table 5 Potential prey data for blue and hammerhead sharks used in MixSIR