Ecological theory suggests generalists with large ecological niches are more likely to persist through stochastic processes. However, there is growing evidence that many generalist populations are comprised individual specialists (Bolnick et al., 2002). The stable isotope analysis of dental collagen from shark teeth provides multiple measurements per individual and can elucidate the extent of individual and population-level variation (Zeichner et al. 2017; Matich et al. 2021; Shipley et al. 2021). This study of the Suruga Bay shark community reveals a large variation in isotopic niche width. Generally, δ15N values increase with total length, suggesting correspondence of trophic level and size, but there is substantial variation within individuals (Fig. 3). Our results demonstrate that individuals within a species have different isotopic niche widths (Fig. 4), which does not necessarily correspond with the species-level niche width or overlap between individuals within a species (Fig. 5). The variation within individuals and among species indicates a continuum of generalists and specialists with a complex and diverse food web.
Ecological Niche Width with Standard Ellipse Area (SEAc)
We estimated species-level isotopic niche using standard ellipse area (SEAc) to reduce errors associated with sample size. While this metric accounts for sample size bias, we found that species with a greater total length diversity among individuals also had the largest standard ellipse area. This pattern reinforces that many sharks are gape-limited predators and undergo ontogenetic dietary shifts (Heupel et al., 2014), which is supported in this study by the significant relationship between total length and δ15N values (Fig. 3). The largest sharks in this study are the Needle Dogfish, which were near maximum adult size, but did not have the largest total niche width based on standard ellipse area (SEAc=1.96). A telemetric study of Needle Dogfish in Suruga Bay found bimodal depth preference (i.e., 300–400 and 580-620m) but little horizontal movement (Yano & Tanaka, 1986); the limited range of habitat is reflected in its relatively low SEAc. In contrast, the Sharpnose Sevengill shark had the largest length range (78.3–117.5 cm), which spans the size at maturity (95–105 cm for females and 70–85 cm in males) (Tanaka & Mizue, 1977), and largest standard ellipse area (SEAc=3.26). Previous studies found evidence of an ontogenetic shift in Sharpnose Sevengill diet (Barnett et al., 2012; Braccini, 2008), which likely accounts for the large isotopic niche exhibited in this study. The largest size distribution in this study was among Smooth hammerhead (58.4–106.1 cm) but the estimated isotopic niche width (SEAc=1.46) only reflects juveniles since the individuals sampled were substantially smaller than total length at maturity (250–260 cm; Miller, 2016). These comparisons of standard ellipse area among species imply the importance of size and ontogeny on the overall niche width sharks. We expect the actual isotopic niche to be larger for the species that we had limited representation, such as Spinner Shark, Rough Dogfish, and Japanese Velvet Dogfish. An assessment of population-level isotopic niche width is necessary to compare within vs. between individual variation, which is the greatest asset of using shark teeth as a substrate for stable isotope analysis.
Individual Variability with Convex Hull Area (CHA)
We used convex hull area to investigate how individuals parse the species-level isotopic niches. Using convex hull area as a metric for determining niche width is often criticized for biased niche geometry due to outlying data points and inconsistent sample sizes (Shipley and Matich 2020). However, convex hull areas are well suited to shark teeth because the same number of possible data points is controlled by the number of teeth available, and within a shark species the number of series is relatively constant (i.e., 3–5), so sample size bias is of less concern. Through teeth, we can compare convex hull areas of individuals within a species. Individuals that behave as generalists will have larger convex hull areas, whereas individual specialists will have smaller convex hull area.
Our results indicate no relationship between individual- and species- level niche width as evidenced by convex hull area and standard ellipse areas, respectively, which suggests a continuum of ecological strategies spanning specialists to generalists in these adjacent shark communities within Suruga Bay. We chose to highlight the substantial individual-level variation within many species with convex hull area and compare to a species-level summary since this feature is uniquely captured in the stable isotope analysis of shark teeth. For example, individual Sharpnose Sevengill sharks have a convex hull area range of 0.2–5.65 (n = 8) while Smooth Hammerhead have a convex hull area range of 0.02–1.10 (n = 12). The differences in maximum convex hull area for these species indicate the extent of diet variation while the range demonstrates individual-level dietary differences. The range of convex hull area for each species loosely increases with number of individuals and given the wide range of individuals per species in this study, we are cautious to designate species as generalists or specialists with these estimates of convex hull area.
Overlap among individuals within a species with Pianka’s Measure (w)
A key feature of proposed quantitative comparisons of individual vs. population niche width is the extent of overlap (i.e., within vs. between individual component as outlined in Bolnick et al. 2002). We use Pianka’s measure to estimate this proportional overlap between individuals within a population while accounting for multivariate covariance (Pianka 1974; Yeakel et al. 2011). This metric complements convex hull area and standard ellipse area because it demonstrates the similarity among individuals within a population. In other words, Pianka’s measure can help distinguish if a generalist population is composed of specialist individuals behaving differently or individuals are also generalists. The most notable result of Pianka’s measure in this study is a broad range throughout both shark communities and within species that encompasses almost the entire possible range: 0.06–0.94 (Fig. 5). Some species featured in this study had small sample size (i.e., n ≤ 3) and therefore their range of Pianka’s measure was limited.
While the correspondence of increasing δ15N values with total length supports some influence of gape-limited predation among sharks in these ecosystems (Fig. 3), most prominent in our results is the variation among individuals. For example, Needle Dogfish have the largest total length, highest δ15N values, and Pianka’s measure span 0.16–0.71, which suggests a range of foraging behaviors and/or preferences. We note the lack of correlation between Pianka’s measure and convex hull area throughout our dataset, which suggests the extent of individual diet variation is not tightly coupled to niche breadth for an individual or population.
Comparing the ecology of sharks within Suruga Bay’s habitats
Suruga Bay is an ideal location to explore patterns in ecological organization among predators given the proximity of multiple marine habitats. The spatial heterogeneity of this region coupled with the diverse shark community results in resource partitioning as evidenced by differences in stable isotope composition of sharks in the four habitats (Fig. 2, Table 2). Further, the population vs. individual stable isotope variation for habitats within Suruga Bay resemble ecological models with patch formation and emergence of specialist competitors (Levin and Paine 1974).
Epipelagic - We sampled juvenile Spinner Shark, Smooth Hammerhead, and Dusky Shark in the epipelagic zone of Suruga Bay. Based on δ13C and δ15N values, this habitat is well linked to the deep benthic habitat, but statistically different from the coastal and abyssal habitats. This dissociation between the epipelagic and coastal habitats is surprising since their proximity would suggest energy and resource exchange. However, the individuals featured in this study were all juveniles based on their total length (Compagno 2001). Further, it is likely that most individuals were young of the year given their total length compared to published the length at birth for each species (Joung et al., 2005; Choi, 2018; Rosa et al., 2017; Joung et al., 2015). The δ15N values of these juvenile sharks likely reflect a maternal signal given their age (Olin et al. 2011; Tamburin et al. 2019) and incorporation rate of dentin collagen (Zeichner et al. 2017). There is previous evidence for long distance migration in Spinner Shark (Rigby et al., 2020), Smooth Hammerhead (Santos and Coelho 2018), and Dusky Shark (Rogers et al., 2013). Further, studies from Korea and Taiwan on these species report the absence of the smallest size classes (Joung et al. 2005, 2015; Choi 2018). It is possible that the stable isotope compositions in this study reflect adult diet in these other localities. The epipelagic habitat of Suruga Bay may serve as an important nursery for these species throughout the western North Pacific Ocean.
Coastal - The two species from the coastal habitat, Japanese Topeshark and Starspotted Smooth-hound, are endemic to the western Pacific Basin and listed as endangered by the International Union for Conservation of Nature (Rigby et al., 2020; Walls et al., 2021). These two species exhibit a relatively narrow and overlapping range in δ13C values, which are distinct from all other habitats in Suruga Bay except the abyssal zone (Table 2). The Japanese Topeshark exhibits a smaller isotopic niche on the species and individual level based on standard ellipse and convex hull area (Fig. 5) but has higher δ15N values than the Starspotted Smooth-hound (Fig. 2). A previous study focused in the Seto Inland Sea, located southwest of Suruga Bay, found that Japanese Topeshark were only seasonally present and predominantly preyed on benthic cephalopod and fish whereas Starspotted Smooth-hound were present year around and fed on crustaceans and polycheates (Kamura and Hashimoto 2004). Starspotted Smooth-hound are known to have variable diets; a comparison of stomach contents between five localities found differences in prey preference (i.e., mantis shrimp, crab, hermit crab, shrimps, crustacean fragments, and polychaetes) with less diversity in larger individuals (Yamaguchi and Taniuchi 2000). In Suruga Bay, Starpotted Smooth-hound feed at a low trophic level (i.e., low δ15N values; Fig. 2) and the variation within and among individuals suggest it is a true generalist with variable individual diet (i.e., larger convex hull area; Fig. 4) and extensive overlap within the population with Pianka’s measure > 0.75 for four individuals (Fig. 5C). The stable isotope composition of Japanese Topeshark in Suruga Bay reflects their diet of small fishes and cephalopods (Kamura and Hashimoto 2004), which are higher trophic prey than consumed by Starspotted Smooth-hound. In addition, Japanese Topeshark are caught in deeper waters (Yano & Kugai, 1993), which often have food webs enriched in 13C and 15N compared to epipelagic waters (Davison et al., 2013; Mintenbeck et al., 2007). These data contribute to growing evidence of the importance of sharks and fish in transferring energy and nutrients to deeper marine habitats (Carlisle et al., 2021; Davison et al., 2013; Trueman et al., 2014).
Deep benthic - The deep benthic habitat of Suruga Bay is represented by the Shortspine Spurdog and Sharpnose Sevengill Shark. These two species differ in their distribution and stable isotope results provide new ecological insights. The Sharpnose Sevengill Shark is globally distributed in deep benthic habitat up to 1000m depth on the upper slope and has dietary data available from multiple regions (Barnett et al., 2012; Finucci et al., 2020). Previous studies determined Sharpnose Sevengill Shark diet to specialize on fish, crustaceans, and cephalopods in the Mediterranean, central Eastern Atlantic, and southern Australia (Braccini 2008; Barnett et al. 2012). However, in the context of the Suruga Bay shark assemblage, Sharpnose Sevengill Sharks have the largest isotopic niche on both the population and individual-level based on standard ellipse area and convex hull area, respectively, with a high degree of overlap among individuals (Fig. 2, 4, 5). This high degree of individual variation is also supported by two specimens analyzed for compound specific isotope analysis of amino acids; the δ15N values for the baseline are similar, but trophic variation is high (Fujiwara et al. 2021). In contrast, the Shortspine Spurdog has a limited distribution in deep benthic habitats of the Northwestern Pacific Ocean (Finucci et al., 2020; Ziadi-Künzli et al., 2020). The three specimens of Shortspine Spurdog in this study are immature females based on total length (Taniuchi and Tachikawa 1997) and their δ15N values suggests a lower trophic level diet than the Sharpnose Sevengill but similar to the Japanese Topeshark in the coastal habitat (Fig. 2). Further, the Shortspine Spurdog has a constrained isotopic niche based on dentin δ13C and δ15N values; this species has a smaller standard ellipse area and convex hull area than the Sharpnose Sevengill but also has high overlap with Pianka’s measure > 0.50 (Fig. 5). The species within the deep benthic habitat have different population- and individual-level isotopic niches, but individuals within each species have a similar feeding ecology and exhibit a high degree of overlap in their isotopic composition.
Abyssal - The deepest habitat of Suruga Bay is its central trough where the Japanese Velvet Dogfish, Roughskin Dogfish, Rough Longnose Dogfish, and Needle dogfish inhabit (Yano & Tanaka, 1983). The specimens for each species represent a relatively narrow range of total length and are fully mature adults so there is minimal ontogenetic insight from their stable isotope composition. Generally, the diet of Squaliformes is known to be fish and squid, but more specific diet data is sparse given that many stomachs are emptied as specimens are brought to the surface (Yano & Tanaka, 1983). The δ13C and δ15N values from dentin collagen suggest there are differences in ecological niche among species as well as extensive individual-level variation within species. Our results indicate Needle Dogfish to occupy the largest isotopic niche as a species in the abyssal habitat with a standard ellipse area = 2.0, but this may also be a function of the relatively high sample size (N = 8; Table 1). The convex hull areas for individual Needle Dogfish spanned the largest range within this habitat and the extent of overlap among individuals varied widely (Table 1, Fig. 5). Our study only included two specimens of Japanese Velvet Dogfish, but its diet is completely undescribed to date (Rigby et al. 2021). The Japanese Velvet Dogfish had elevated δ15N values, which is likely a result of feeding in the abyssal food web (Mintenbeck et al. 2015; Trueman et al., 2014) rather than high trophic level given that the specimens caught were < 100cm (Fig. 2). This species had the smallest isotopic niche and complete overlap in isotopic composition among individuals (Fig. 5). While these traits could be due to low sample size of Japanese Velvet Dogfish, we also only sampled two individual Roughskin Dogfish. These Roughskin Dogfish specimens produced similar results for species- and individual-level isotopic niche with their standard ellipse area and convex hull area, respectively (Fig. 5), but each specimen had different δ13C and δ15N distributions from dentin collagen and therefore diverged in their overlap. This variation among individuals aligns with a previous result from compound specific isotope analysis of amino acids that indicated similar baselines but different trophic level from two other specimens caught in Suruga Bay (Fujiwara et al. 2021). The final species caught in the abyssal habitat is the Rough Longnose Dogfish, which has reported distribution data but no published record of diet (Compagno 2001; Carpenter and Garilao 2021). The isotopic niche for Rough Longnose Dogfish are similar to the Roughskin Dogfish on the species- and individual-level as well as extent of overlap among individuals (Table 1; Fig. 5). However, the Rough Longnose Dogfish has the lowest mean δ15N values from dentin collagen of all species within the abyssal habitat, which could indicate foraging on prey from a shallower depth. Previous studies featuring deep water consumers and stable isotope analysis demonstrate the importance of diel-vertical migration, which transfers nutrients and links epipelagic, deep benthic, and abyssal marine habitats (Carlisle et al., 2021; Trueman et al., 2014). Although this study presents stable isotope compositions of consumers without prey data, the assessments individual vs. population variation provides an ecological context to evaluate resource use, ecological niche, and food web structure, which are substantial contributions for these species that are largely classified as “data deficient” and/or “vulnerable” (Froese and Pauly 2021).