Spatial and Ontogenetic Patterns in the Trophic Ecology of Juvenile Bull Sharks (Carcharhinus leucas) From the Northwest Gulf of Mexico

A combination of stomach content and stable isotope (δ13C, δ15N, and δ34S) analyses were used to characterize and examine spatiotemporal and ontogenetic trends in the feeding ecology of juvenile bull sharks (Carcharhinus leucas) captured in estuaries throughout the northwest Gulf of Mexico (GoM) between 2013 and 2016. Shark diets were dominated by fish prey taxa [>98% index of relative importance (%IRI)], and of those identified to the family level, two families comprised greater than 50% IRI, Mugilidae (mullets: ∼32%) and Sciaenidae (drums and croakers: 27%). Clupeidae (herrings: 14%) and Ariidae (sea catfishes: 15%) also contributed substantially to the diet of juvenile sharks, though consumption of Ariidae increased as consumption of Clupeidae decreased in juvenile sharks larger than 893 mm Fork Length (FL) (∼1 year old). Values of δ15N increased significantly with shark size, indicating a shift toward larger or higher trophic level prey with increasing shark size. Latitudinal and temporal trends in δ13C and δ34S suggest isotopic variation occurred in correspondence with shifts in primary producer assemblages and environmental drivers of sampled estuaries. These results highlight the importance of teleost prey resources along the freshwater-marine continuum in the diet of juvenile bull sharks, as well as the utility of natural tracers in tracking ontogenetic trends in feeding ecology.


INTRODUCTION
Bull sharks (Carcharhinus leucas) are top predators in the Gulf of Mexico (GoM), particularly in coastal habitats, where their ability to osmoregulate across a wide range of salinities allows them to inhabit brackish estuaries otherwise inaccessible to elasmobranchs with narrower salinity tolerances . Adult female bull sharks move seasonally into estuarine habitats to pup (give birth), where juveniles may remain for several years before transitioning to higher salinity habitats (Heupel et al., 2010). As the most abundant shark in estuaries of the northwest Gulf of Mexico (nwGoM) (Froeschke et al., 2010;Plumlee et al., 2018), and perhaps across the GoM, bull sharks may play an important role in structuring estuarine food webs via direct or indirect effects (Papastamatiou et al., 2009). Previous studies of bull shark diet have included stable isotope, stomach content, and fecal metabarcoding analyses (Matich and Heithaus, 2014;Tillett et al., 2014;van Zinnicq Bergmann et al., 2021), and highlight the cosmopolitan nature of their diets. Studies of bull shark stomach contents across their geographic range have shown juveniles primarily consume teleost prey from families Ariidae, Clupeidae, Mugilidae, Sciaenidae, and Carangidae, with limited consumption of benthic invertebrates (e.g., Callinectes spp.) (Clark and von Schmidt, 1965;Snelson et al., 1984;Thorburn et al., 2004;Cottrant et al., 2021). With increasing size, bull sharks expand their diet to larger, vertebrate prey such as elasmobranchs, birds, marine reptiles, and marine mammals (Tuma, 1976;Snelson et al., 1984;Cliff and Dudley, 1991;Werry et al., 2011). Such dietary shifts are thought to be driven in part by morphological changes (e.g., increased gape width, swimming speed) that allow young sharks to exploit a wider range of prey. Dietary ontogenetic shifts may also be facilitated by an ontogenetic habitat expansion from freshwater/estuarine habitat toward marine/coastal habitats, in which larger prey are more abundant (Werry et al., 2011;Belicka et al., 2012;Daly et al., 2013). Nevertheless, the reliance of bull sharks on estuaries during the juvenile stage may make them particularly susceptible to the impacts of anthropogenic alterations of these habitats, and it is difficult to predict the ecosystem level effects of these alterations without a clear understanding of how bull sharks interact with surrounding ecological communities. The trophic dynamics of bull sharks in relation to estuarine communities in the nwGoM are not well understood, with only one study published to date (Cottrant et al., 2021). Therefore, studies of top predator trophic interactions are instrumental in management efforts directed at the conservation of these taxa and ecosystems alike. Such conservation efforts have recently taken ecosystem-based approaches, which rely upon detailed information regarding species interactions, particularly in the context of food webs (e.g., predator-prey interactions) (Crowder and Norse, 2008).
The dietary composition of a consumer species or population can be better understood through the taxonomic identification of prey remains from the stomachs of those consumers (Cortés, 1997). Quantification of the abundance, mass or volume, or frequency of occurrence of prey items in consumer stomachs may further be used to estimate the relative contribution or importance of each prey taxon to consumer diets (Hyslop, 1980). However, a key limitation of stomach content analysis is that it only provides a "snapshot" of the diet of each individual sampled. For elasmobranchs, the time between feeding bouts may exceed digestion and gastric evacuation times (Cortés, 1997), which results in a substantial proportion of stomachs that are empty, or contain highly digested (i.e., degraded) prey. This approach is unsuitable for gathering dietary information over time scales greater than several days, unless a very large number of stomachs is collected in a time-stratified sampling design (Cortés, 1997). In contrast, analyses of dietary tracers such as stable isotopes (hereafter "isotopes") of carbon, nitrogen and sulfur from consumer tissues provide a time-integrated representation of the feeding ecology of a consumer (Hussey et al., 2012). While values of δ 15 C and δ 15 N are typically utilized as indicators of primary production sources and trophic position, (respectively), δ 34 S values are known to vary along the freshwatermarine continuum (Fry and Chumchal, 2011), and vertically in the water column. Benthic sulfate-reducing bacteria produce characteristically low values of δ 34 S that are in turn assimilated by other benthic fauna, and as these isotopes undergo little trophic fractionation, values can therefore provide an indication of the degree to which a consumer forages in benthic vs. pelagic food webs (Peterson, 1999).
In this study, we combine stomach content analysis and dietary tracers (stable isotopes of carbon, nitrogen, and sulfur) to investigate the feeding ecology of juvenile bull sharks in estuarine nurseries from the nwGoM. Dietary composition and stable isotope data were used to investigate the influence of size (fork length; FL), capture location (Estuary), and capture date (Year, Season) on the diet and trophic interactions of juvenile bull sharks. This research contributes to our understanding of the dietary requirements of juvenile bull sharks, as well as their broader role as a potential top predator in estuarine food webs.

Sample Collection
From May 2013-October 2016, 246 juvenile bull sharks (549-1183 mm FL) were collected in 7 estuaries along the nwGoM (Sabine Lake: n = 49, Galveston Bay: n = 67, Matagorda Bay: n = 30, Aransas Bay: n = 69, Corpus Christi Bay: n = 29, Upper Laguna Madre Bay: n = 1, and Lower Laguna Madre Bay: n = 1) (Figure 1). Sharks sampled were incidental mortalities occurring in Texas Parks and Wildlife Department gill net surveys (Martinez-Andrade et al., 2009) conducted in the spring (April-June) and fall (August-November). Prior to processing, each shark was stored frozen (−20 • C). Upon processing, sharks were thawed, measured (FL), sexed, and stomachs and ∼100 g of epaxial (hereafter "muscle") tissue were removed. Stomachs were preserved in 10% formalin for 48 h and subsequently stored in 70% ethyl alcohol. Upon removal, muscle tissue was stored (−20 • C) until subsequent stable isotope analysis (TinHan, 2020). Stable isotope analyses were performed on tissue from a subset of 138 sharks selected to maximize the balance of sample sizes across years, seasons, estuaries and shark sizes (Supplementary Table 1). Due to lack of replication, stable isotope analyses were not performed for sharks from Upper and Lower Laguna Madre Bays.

Stomach Content Analysis
The contents of preserved stomachs were separated across three sieves (mesh widths: 1.27, 0.14, and 0.05 cm), rinsed and the combined wet weight of all contents was recorded. All contents were identified to the lowest possible taxon, sorted, weighed (to the nearest 0.001 g), and where possible, the number of individual prey items was recorded. Three metrics of dietary composition were estimated from stomach contents: percent frequency of occurrence (%FO), percent number (%N), and percent weight (%W). %FO was calculated as the number of stomachs containing each prey taxon, divided by the total number of stomachs containing prey. %N was calculated as the number of individual prey items in each taxon, divided by the total of all individual prey items in each stomach. %W was calculated as the mass of each prey taxon divided by the total mass of all prey taxa by stomach. While the proportion of sharks with empty stomachs was recorded, these samples were not considered informative for examining dietary composition and they were therefore omitted from dietary metric calculations. To obtain estimates of variability around %N and %W, these were calculated for each sample (Bizzarro et al., 2007). A composite dietary index, the index of relative importance (IRI) (Pinkas et al., 1971) was similarly, calculated for each prey category in each sample and converted to %IRI: where i represents each prey taxon in sample stomach j, and n represents the total number of prey taxa in J. Due to the preponderance of fish in bull shark stomachs, dietary metrics were also calculated at the family level (or order, where familylevel identification was not possible) for fish prey, under the assumption that identified fish were representative of all fish prey (unidentified and identified). Relationships between each dietary composition metric (%N, %W, %IRI) and predictor variables [shark size (FL), capture location (Estuary), and capture year (Year)], were tested using permutational analysis of variance (PERMANOVA; function adonis2 in R package vegan). Dietary composition metrics (%N, %W, %IRI) were square root transformed, and analyzed against FL and Estuary. Ontogenetic trends in stomach contents were examined by pooling sharks (for which stomach contents were analyzed) into size classes. The number (N = 2) and ranges (small: 549-892, large: 893-1183 mm FL) of these classes were defined by an equal-variance hierarchical clustering model (R package mclust). Spatial patterns were examined by pooling sharks into two regional groups based on location of capture: (1) northern (Sabine Lake, Galveston Bay), and (2) southern (Matagorda, Aransas, Corpus Christi Bays) estuaries. Shark size was also compared against the binomial probability of a stomach being empty or containing prey (logistic regression), as well as percent fullness (mass stomachcontents /mass shark × 100) (linear regression).

Stable Isotope Analysis
Muscle tissue samples were oven dried at 60 • C, and lipid extracted using non-polar petroleum ether solvent in a Dionex 350 Accelerated Solvent Extractor (ThermoFisher Scientific Inc.) following (Plumlee and Wells, 2016). Lipid-extracted samples were then homogenized with a ball/vial grinder (Wig-L-Bug; Sigma-Aldrich Co., LLC), and ∼1 mg of material was transferred into 5 × 9 mm tin capsules for analysis at the University of California, Davis Stable Isotope Facility. Samples were then analyzed for 13 C: 12 C and 15 N: 14 N isotope ratios on a PDZ Europa ANCA-GSL elemental analyzer and PDZ Europa 20-20 isotope ratio mass spectrometer (IRMS; Sercon Ltd., Cheshire, United Kingdom). Analysis of 34 S: 32 S isotope ratios was conducted on an Elementar vario ISOTOPE cube and SerCon 20-22 IRMS (Sercon Ltd.). Results are reported in delta notation, relative to standards Vienna Pee Dee Belemnite (carbon), air (nitrogen), and Vienna Canyon Diablo Troilite (sulfur), using the following equation: where R is the ratio of heavy to light isotopes of element X in the sample or standard reference material.

Stable Isotope Analysis
All three isotopes varied significantly with shark size (FL), with the greatest effect seen in values of δ 15 N (F 1,128 = 27.38, p < 0.001), which increased with FL. Values of δ 13 C and δ 34 S showed weaker, negative relationships with shark size (δ 13 C: F 1,128 = 10.11, p = 0.002; δ 34 S: F 1,128 = 6.99, p = 0.009) (Figure 3 and Table 3). The disparity in δ 13 C values between northern estuaries (Sabine Lake, Galveston Bay) and southern estuaries (Matagorda, Aransas, Corpus Christi Bays) increased from 2013 to 2016 (Figure 4A), though differences across years in δ 13 C and δ 15 N were not significant (δ 13 C: p = 0.112; δ 15 N: p = 0.056). Slight variation in δ 34 S was observed with respect to year (δ 34 S: F 3,128 = 4.95, p = 0.003; Table 3), and variation among individuals in Galveston Bay increased from 2013 to 2016 ( Figure 4C). Values of δ 13 C were significantly different between seasons (F 4,128 = 6.28, p = 0.014), with δ 13 C values being slightly higher in the spring than in the fall, though this trend was not consistent among years and estuaries (Supplementary Figure 1). Values of δ 13 C exhibited a clear latitudinal trend among estuaries (F 4,128 = 31.53, p < 0.001), with values increasing from north to south. Significant, albeit slight, differences were observed in δ 15 N values among estuaries (F 4,128 = 0.08, p < 0.001), and values in sharks from Galveston Bay were elevated relative to those from other sampled estuaries (Galveston Bay, mean ± SD: 17.5 ± 1.7 ; all other estuaries: 15.4 ± 1.2 ; Figure 4B). Though δ 34 S values varied significantly among estuaries (p = 0.003), no latitudinal trend was apparent ( Figure 4C). There was a significant negative relationship between δ 13 C and δ 15 N (r 2 = 0.17, p < 0.001), though δ 34 S values were uncorrelated to δ 15 N or δ 13 C (p > 0.05; Figure 5 and Table 4).

DISCUSSION
In this study, we characterized the diet and trophic dynamics of juvenile bull sharks from the nwGoM using a combination of stomach content and stable isotope analyses. Asymptotic prey richness was not reached in identifications of prey to the lowest possible taxon, potentially owing to the heavily digested state of most prey items, and the subsequent difficulty of identifying such items to genus or species. However, juvenile bull shark diets consisted almost exclusively of teleost prey, which is in accord with previous dietary studies of juvenile bull sharks from across their range (Sadowsky, 1971;Tuma, 1976;Snelson et al., 1984;Werry et al., 2011;Tillett et al., 2014;Trystram et al., 2016). A small proportion of analyzed stomachs also contained invertebrates and vegetation, and while it is possible invertebrates were specifically targeted for consumption, the majority of stomachs containing vegetation also contained fish or invertebrates. This suggests vegetation might have been incidentally ingested while foraging among habitats containing aquatic plants or algae. One coastal shark species (bonnethead, Sphyrna tiburo) is capable of digesting and assimilating seagrass (Leigh et al., 2018); however, it is unclear whether consumption is incidental, and whether other elasmobranchs are also capable of omnivory. Nevertheless, these items were neither abundant nor frequently encountered in this study, and are therefore not expected to be an important component of the diet of juvenile bull sharks in Texas. Elasmobranchs were similarly, rare, and only present in just three of the 171 stomachs that contained prey. Identifications of fish prey to family (or higher) approached asymptotic prey richness, which suggests that, although there may have been specific genera or species overlooked in our sampling, greater sampling was not likely to identify additional prey families. Just two fish families (Mugilidae and Sciaenidae) contributed to greater than 50% of the diet of juvenile bull sharks, though Ariidae and Clupeidae were also important components of their diet. In contrast to %IRI, the relative contribution of Clupeidae to bull shark diet increased when expressed in terms of %W and %FO. Similarly, the relative contribution of Ariidae decreased when using these metrics, rather than %IRI. This may be due to inherent differences in the digestibility of ariid v. clupeid prey, as well as the presence of conspicuous structures in ariids by which they are easily identified (barbed first-dorsal and pectoral fin spines), whereas clupeids were usually only observed in stomachs with relatively undigested contents. Recent work by Cottrant et al. (2021) utilizing both stomach content analyses and a literature review identified these four families as the most commonly encountered in juvenile bull shark diets from San Antonio Bay, Texas, and among the most common in juvenile and adult bull shark diets published elsewhere. Biases introduced by digestion rates or identifiability may therefore be minimal, or are more likely consistent across studies. Bull sharks spend the first several years of life in coastal nurseries, during which they may more than double in size. Over this period, increases in gape width and swimming speed (Lauder and Di Santo, 2015) concomitant with body size and changes in tooth morphology (Cullen and Marshall, 2019), may confer an advantage to their foraging success, leading to ontogenetic shifts in diet. When examining overall dietary metrics calculated for smaller vs. larger bull sharks in the present study, the importance of piscivore/invertivore ariid prey (Motta et al., 1995;Mendoza-Carranza, 2003) supplanted detritivore/planktivore clupeids (Castillo-Rivera et al., 1996;De Brabandere et al., 2009) and piscivore/invertivore sciaenids (Sheridan, 1979;Willis et al., 2015) in the diet of larger sharks. Nevertheless, changes in dietary composition were not necessarily accompanied by a shift toward higher trophic level taxa, and the observed shift was minor (<20% IRI). In fact, detritivore/planktivore mugilid (Cardona, 2016) prey remained the dominant fish family in the diets of both size classes of sharks. Though bull sharks are expected to increase their consumption of large-bodied or higher trophic level prey (e.g., chondrichthyans, marine mammals) with ontogeny (>∼150 cm) (Cliff and Dudley, 1991;Werry et al., 2011;Tillett et al., 2014), no such shift was observed. While it is possible bull sharks in the nwGoM simply do not exhibit a shift to these larger prey taxa, it is more likely due to focused sampling of juvenile sharks in this study. The lack of larger, marine prey reported in the diet of bull sharks elsewhere (e.g., Carangidae, Scombridae, Elasmobranchii; (Clark and von Schmidt, 1965;Cliff and Dudley, 1991) may also be explained by the limited seasonal sampling in this study, as it is not known what prey juvenile bull sharks exploit when, during the winter months, they vacate or reduce their use of estuaries in the GoM. Ontogenetic shifts in diet were not found to be statistically significant when testing for differences among individual-based estimates of dietary metrics. The opportunistic nature of bull shark feeding and likely patchy prey distributions across Texas estuaries may have driven the high individual variation in stomach contents, and obscuring ontogenetic patterns. Although a greater number of samples may not be necessary to characterize the overall diet of sharks in this region, larger sample sizes may yet be required to clearly describe ontogenetic shifts in diet. Across the size range of sharks sampled (549-1183 mm FL), ontogenetic shifts in trophic level and food web occupancy along the freshwater-marine continuum were indicated by significant shifts in δ 13 C (decrease), δ 15 N (increase) and δ 34 S (decrease) values with shark size. However, the sizes of ingested prey were not estimated here, and it remains unclear whether increases in δ 15 N values with ontogeny reflect shifts in prey size or prey trophic level, growth/metabolism related shifts in nitrogen fractionation rates (Wyatt et al., 2018;Marcus et al., 2019), or a combination thereof. Though sharks exhibited ontogenetic decreases in their consumption of clupeids (pelagic planktivores), FIGURE 5 | Biplots of δ 34 S, δ 15 N, δ 13 C from muscle tissue of juvenile bull sharks collected from Texas estuaries (estuary indicated by filled circle color) and robust linear regression trendlines for estuaries in which isotope relationships were significant (p < 0.05). and corresponding increases in consumption of ariids (benthic piscivores and invertivores), these shifts were relatively small, and may be associated with the lack of significant shift in δ 34 S with ontogeny. Ontogenetic changes in diet may also occur in tandem with habitat shifts, where sharks expand or shift their dietary niche to include prey encountered in new habitats (Grubbs, 2010).
Salinity is considered to be a key factor driving δ 13 C and δ 34 S values in aquatic food webs (Peterson, 1999), and this was supported in the present study by the observed spatiotemporal trends in both isotopes. Latitudinal gradients in stable isotopes were identified, likely driven by corresponding shifts in primary producer assemblages, and environmental factors influencing the isotopic composition of carbon, nitrogen, and sulfur available to primary producers and higher-level consumers (i.e., sharks). Estuaries in south Texas are dominated by seagrasses (e.g., Thalassia testudinum, Halodule wrightii), while northern estuaries (Sabine Lake and Galveston Bay) contain limited abundances of these primary producers (Adair et al., 1994). Seagrasses are enriched in 13 C relative to other estuarine primary producers (Fry and Parker, 1979), which may explain the increase in δ 13 C values when moving from north to south. Increased rates of precipitation occurring in northern Texas estuaries presumably increase the inflow of freshwater depleted in 13 C, which may serve to further decrease δ 13 C values in northern estuaries, compared to those in the south. Though δ 13 C values in sharks from southern estuaries were consistent across years, between 2013 and 2016, values decreased and became more variable in sharks from northern estuaries, which may have been driven by increased freshwater input in northern estuaries in (Texas Water Development Board, 2020. Similarly, a positive relationship between δ 13 C and δ 34 S values in the present study was observed exclusively in northern estuaries. This is unsurprising, as low values of both isotopes are often associated with primary producers or food webs occurring in freshwater habitats (Peterson and Fry, 1987;Fry and Chumchal, 2011). We saw no clear latitudinal gradient in δ 15 N values, though sharks from Galveston Bay had consistently elevated δ 15 N values compared to other estuaries in each year. The size range of sharks did not differ significantly among estuaries, so sharks in all estuaries were presumed to be of similar ontogenetic status. It is possible that sharks in Galveston Bay simply fed at a higher trophic level than those from other estuaries, though we detected no significant differences in dietary composition among estuaries, which may be due in part to the limited taxonomic resolution of stomach contents identification. Nevertheless, elevated δ 15 N values in Galveston Bay sharks may be alternatively explained by increased levels of urbanization and anthropogenic nitrogen ( 15 N-enriched) input into Galveston Bay compared to other estuaries on the Texas coast (Holt and Ingall, 2000;Barcenas, 2013). Though limited in geographic and trophic scope, some comparisons can be made with published isotope values from other estuarine consumers in the nwGoM. In a characterization of trophic relationships in a sub-bay of the Matagorda Bay estuary, among the highest δ 15 N values reported by Akin and Winemiller (2008) ranged from 13.2 to 14.7 , and belonged to small-to large-bodied teleosts (e.g., Gulf menhaden (Brevoortia patronus), red drum (Sciaenops ocellatus), gafftopsail catfish (Ariopsis felis), and alligator gar (Atractosteus spatula)). Bull sharks from Matagorda Bay in the present study had δ 15 N values of ∼15-17 , placing this species approximately one trophic level above other estuarine consumers. In addition to bull sharks, blacktip sharks (Carcharhinus limbatus) are among the most commonly encountered shark species in nwGoM estuaries (Plumlee et al., 2018). Published values of δ 13 C (∼−16 ) and δ 15 N (∼16.5 ) from blacktip sharks captured along the northern and central Texas coast (Plumlee and Wells, 2016;Matich et al., 2021) compared with those in the present study (δ 13 C: ∼−17 ; δ 15 N ∼16 ) suggest that bull sharks occupy similar or higher trophic positions as blacktips. Given the similarity in trophic positions implied by similar δ 15 N values, slightly diminished values of δ 13 C in bull sharks may indicate this species exploits 13 C depleted freshwater/brackish food webs to a greater extent than blacktips. This is supported by direct comparisons of bull and blacktip shark habitat use (Matich et al., 2020), and may exist as a form of niche partitioning in which bull sharks exploit their tolerance for low salinities to mitigate potential competition with blacktips or other stenohaline piscivores (Dwyer et al., 2020).
Carcharhinid sharks develop their offspring in utero where they are nourished via yolk-sac placenta and, at parturition, the isotopic composition of the offspring reflects that of their mother (Hussey et al., 2010). Post-parturition, sharks retain maternal isotope ratios during a turnover period (555-786 days for δ 13 C in white muscle; Logan and Lutcavage, 2010) during which their tissues become increasingly reflective of their diet. As the majority of sharks included in our analyses were YOYs, the observed isotopic values are expected to represent maternally derived isotope ratios, if only partially (Olin et al., 2011;Belicka et al., 2012). This may provide an alternative explanation for slightly elevated δ 13 C values in smaller (presumably younger) sharks, if these values decreased with age as juveniles begin to integrate the isotopic composition of their prey. However, it becomes difficult to explain the significant increase in δ 15 N values with ontogeny if maternal signatures are indeed present in the tissue of YOY sharks in this study. Adult female bull sharks almost certainly occupy a higher trophic position than their juvenile offspring, thus neonatal sharks should exhibit an overall decrease in δ 15 N values with age (and by extension, size), as maternal δ 15 N values eventually begin to reflect the juveniles' prey. Estimated turnover rates of δ 15 N in white muscle of C. plumbeus range from 340 to 457 days (Logan and Lutcavage, 2010), in which case YOYs (which comprised the majority of sampled sharks), should still exhibit maternal δ 15 N values. The apparent lack of maternal influence in δ 15 N values may be explained if adult females occupied food webs depleted in 15 N while carrying offspring, after which, juvenile sharks began foraging in 15 N-enriched food webs in nursery habitats. In such a case, losses of the 15 Ndepleted maternal signal would be difficult to differentiate from increases in δ 15 N values arising from ontogenetic increases in trophic position.

Future Directions
Given the patterns of spatiotemporal variability of isotopes in this study, future studies seeking to estimate trophic position of consumers in nwGoM estuaries should perform time-matched collections of a representative suite of primary producers across the freshwater-marine continuum. Sampling efforts including a greater proportion of subadult bull sharks would provide better representation of ontogenetic shifts occurring later in the life of this coastal elasmobranch. Nevertheless, where high levels of overlap exist in the isotopic niche of diverse estuarine food webs (e.g., Winemiller et al., 2007), the application of isotope mixing models may not effectively estimate prey-specific contributions to the diet of a top predator such as the bull shark. Direct examination of stomach contents coupled with stable isotope analysis represents a powerful approach for identifying ontogenetic shifts in dietary composition or among disparate aquatic food webs, and such results may readily incorporate into ecosystem-based models and their attendant management strategies.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The animal study was reviewed and approved by Texas A&M University Institutional Animal Care and Use Committee.