Trophic position of Otodus megalodon and great white sharks through time revealed by zinc isotopes

Our study reveals zinc isotopes to be a promising trophic indicator in sharks and other fishes in general, similar to previous studies featuring both terrestrial and marine mammals^{9,10,11,12,13}. We analysed the Zn isotope values for extant sharks spanning captive/aquarium and wild individuals from various localities and found close correspondence with their respective trophic level. Further, Zn concentration and isotope composition suggest preservation of this biological signal in fossil specimens with little diagenetic alteration. A survey of fossil shark teeth spanning the Miocene–Pliocene reveal similar δ⁶⁶Zn values and variation as found in related (e.g., congeneric) extant sharks with similar dentition and ecology. Our δ⁶⁶Zn results indicate high trophic levels for Otodus and perhaps a trophic change in C. carcharias, the great white shark.

Zinc isotopes in extant sharks and teleosts

As with mammals^{9,10,11,12,13}, bioapatite δ⁶⁶Zn values in wild extant elasmobranchs and teleosts show overall lower values with increasing trophic level (Fig. 1; Supplementary Fig. 1). Both δ⁶⁶Zn and δ¹⁵N_coll correlate with FishBase¹⁷ trophic levels, despite differences in species’ geographic origin and tissue types sampled (Spearman’s correlation r = −0.87, p = 5.65E^–16, n = 48 and r = +0.42, p = 1.47E^–5, n = 40, respectively). There is no statistically significant relationship between bioapatite δ⁶⁶Zn values and δ¹³C_coll, but there is one between wild fish δ⁶⁶Zn and δ¹⁵N_coll values from the same tooth or individual (R² = 0.28, p = 6.89E^–4, n = 38; Supplementary Fig. 2). Both proxies thus generally reflect trophic levels. Large apex predatory sharks (e.g., Carcharodon carcharias, Isurus oxyrinchus, and Lamna ditropis) have significantly more negative δ⁶⁶Zn values than lower trophic level teleosts and the plankton-feeding basking shark (Cetorhinus maximus). In particular, the shortfin mako shark (I. oxyrinchus) and great white shark (Carcharodon carcharias), both apex predators^18,19, have much lower δ⁶⁶Zn values than in any previously recorded extant vertebrate species (enameloid up to –0.71 and –0.63‰, respectively). These low δ⁶⁶Zn values are likely due to the larger number of trophic levels in the marine ecosystem than in terrestrial food webs (terrestrial mammal enamel lies typically between 0 and +1.6‰^9,11) and perhaps differences between marine and terrestrial Zn isotope baselines.

Absolute enameloid δ⁶⁶Zn values vary by up to 1.26‰ among the extant species analysed from various oceanographic areas (Fig. 1). Our results also demonstrate large variability in enameloid δ⁶⁶Zn values among extant sharks within the same region; for example, there is a 0.88‰ difference between mean values of Carcharodon carcharias and the bull shark (Carcharhinus leucas; Fig. 1) from KwaZulu-Natal (South Africa, KZN). In contrast, enameloid δ⁶⁶Zn from the same species (e.g., Carcharodon carcharias, Carcharhinus obscurus) demonstrate a low isotopic variability, independent of geographic location (Fig. 1).

We observe uniformity in the enameloid δ⁶⁶Zn values of five gilt-head bream (Sparus aurata) individuals fed on a controlled fish pellet diet in pisciculture cages located offshore of Central Israel, with values within the measurement uncertainty of each other (–0.01 ± 0.01‰). As with δ⁶⁶Zn, the δ¹⁵N_coll and δ¹³C_coll values are distinct from those of wild teleost individuals caught nearby in Haifa Bay, reflecting the artificial pelleted diet of the pisciculture individuals (Fig. 1, Supplementary Figs. 3, 4). Strongly contrasting the homogenous control fed S. aurata δ⁶⁶Zn values, we observe a higher δ⁶⁶Zn variability among (and even within) wild and aquaria elasmobranch individuals (fed with wild-caught fish and cephalopods). For instance, two teeth of a single tiger shark (Galeocerdo cuvier) individual (–0.52 and –0.27‰, Fig. 1) have a variability higher than the total variability among the three KZN G. cuvier individuals. Galeocerdo cuvier is well known for its highly opportunistic prey selection²⁰. Therefore, the δ⁶⁶Zn value of bioapatite is likely highly responsive to an individual’s diet at the time of tissue formation, and as shark teeth form and replace continuously, enameloid δ⁶⁶Zn values can vary among teeth of a single individual. Thus, although fish can absorb Zn via their gills, waterborne Zn absorption appears to have a negligible effect on elasmobranch tooth δ⁶⁶Zn values, in line with Zn incorporated into soft and skeletal tissues in natural environments being predominantly derived from dietary gastrointestinal uptake^7,8.

Carcharhinus enameloid δ⁶⁶Zn values are high relative to sharks with similar bulk δ¹⁵N_coll values, which contrary to the here analysed Carcharhinus species more regularly consume pelagic prey offshore, oceanic and on the continental shelf (e.g., Galeocerdo cuvier)²¹. This discrepancy may relate to Carcharhinus species inhabiting neritic waters where they feed primarily on demersal/benthic, freshwater-brackish-coastal prey^22,23,24,25. While the diet of KZN G. cuvier and Carcharodon carcharias can also include reef-associated or demersal prey, pelagic organisms are typically more important by mass, especially in adult individuals^20,26. Zinc isotope variability among marine organisms and their tissues is largely unknown, currently limiting our ability to identify specific food items based on shark enameloid δ⁶⁶Zn values beyond generally observed trophic level effects. Whether higher Carcharhinus enameloid δ⁶⁶Zn values relate to specific prey species (and trophic level) or general differences in basal organic matter source between a primarily neritic food web compared to a more open marine pelagic food web remains unclear (Supplementary Discussion 1). However, we observe no difference in δ¹³C_coll values that imply a more terrestrial carbon signal in the KZN Carcharhinus species relative to sympatric species, arguing against differences in the basal organic matter source amongst the KZN shark species (Supplementary Fig. 3).

A previous study on Arctic marine mammal bones suggested a higher geographic independence of δ⁶⁶Zn values from baseline variability compared to δ¹³C_coll and δ¹⁵N_coll values¹³. Likewise, fish taxa with similar diet composition, habitat use and/or trophic level, have a similar range of bioapatite δ⁶⁶Zn values regardless of their geographic locality (Fig. 1), indicating that δ⁶⁶Zn may allow worldwide dietary and trophic level comparability with limited marine baseline variation. Further studies will need to expand our knowledge on δ⁶⁶Zn variability in extant marine vertebrates as well as the effects of baseline on marine vertebrate enameloid δ⁶⁶Zn values, especially compared to dentine δ¹³C_coll and δ¹⁵N_coll values. Nevertheless, the high taxa-specific and perhaps baseline-independent δ⁶⁶Zn values suggest δ⁶⁶Zn is an independent indicator of trophic level and an asset for present and past food web reconstructions in the marine realm.

Deep-time zinc isotope preservation in fossil enameloid

Fossil enameloid has δ⁶⁶Zn values and Zn concentrations ([Zn]) in the range of extant elasmobranch species, arguing against significant diagenetic modification (Figs. 2, 3, Supplementary Fig. 5). Fossil shark teeth examined herein are from Germany, Malta, Japan, North Carolina (USA) and Florida (USA) covering the Early Miocene (Burdigalian, 20.4–16.0 Ma), Miocene-Pliocene transition (Messinian-Zanclean boundary, ca. 5.3 Ma), and the Early Pliocene (Zanclean, 5.3–3.6 Ma; Fig. 2; Supplementary Note 2). Importantly, extant and fossil elasmobranch enameloid δ⁶⁶Zn values (–0.71 to +0.28‰ and –0.83 to +0.27‰, respectively) differ from: (1) previously reported values of terrestrial mammal enamel (0 to +1.6‰);^9,11 and (2) sedimentary carbonate δ⁶⁶Zn values of the fossil sites (+0.34 to +0.49‰, Supplementary Fig. 6, Supplementary Table 1). These differences support a preserved biological signal in fossil enameloid.

We observe the same within tooth Zn spatial concentration pattern in extant and Miocene tiger shark (Galeocerdo spp.) teeth, with Zn being more enriched in the outer enameloid than close to the enameloid-dentine junction. If significant diagenetic Zn exchange had occurred throughout the enameloid, this original Zn concentration pattern would not be preserved in the fossil tooth (Supplementary Fig. 7). Additionally, both extant and fossil shark enameloid show the same variation in [Zn] according to their taxonomy, with carcharhiniforms generally having higher [Zn] than lamniforms (Supplementary Fig. 5), again arguing against significant diagenetic enameloid Zn exchange. For the European Miocene sites, δ¹⁸O_P analyses were also conducted on a subset of teeth, where their enameloid appears to be generally well-preserved as suggested by δ¹⁸O_P values demonstrating species-specific relative in-vivo temperature ranges as expected compared to the habitat use of equivalent modern species²⁷ (Supplementary Fig. 8).

To discern the effects of diagenetic Zn alteration, we compare visually pristine appearing enameloid with areas sampled along fractures and dentine of the same tooth (Supplementary Figs. 6 and 9, Supplementary Table 2). Our results demonstrate that the diagenetic Zn exchange in fractured enameloid leads to higher δ⁶⁶Zn values than in the pristine enameloid of the same tooth, whereas we observe no differences in enameloid δ⁶⁶Zn profiles of modern teeth (Supplementary Figs. 9 and 10, Supplementary Table 3). Likewise, the diagenetically more susceptible fossil dentine shows higher δ⁶⁶Zn values as reflected by a significantly higher and more variable dentine-enamel δ⁶⁶Zn offset (+0.78 ± 0.33‰, n = 13) than observed for extant teeth (+0.22 ± 0.1‰, n = 23; Supplementary Discussion 2, Supplementary Figs. 6 and 11). For the fossil enameloid shown here, in-vivo δ⁶⁶Zn values must be at least as low as their current values, indicating limited to no alteration. Consequently, δ⁶⁶Zn analysis of fossil enameloid can enable deep-time dietary reconstructions.

The homogeneity in δ⁶⁶Zn values for the same species or genera independent of locality and geological age is a remarkable observation (Figs. 2, 3), not only limiting the likelihood of Zn diagenetic alteration but also arguing for minimal variability in habitat-specific food web baselines or a strong homogenisation of δ⁶⁶Zn values at low trophic levels. There are still some limitations, such as the absence of reported δ⁶⁶Zn values of marine non-mammalian vertebrates for comparison outside this study, the limited sample size for some species, and uncertainties regarding Zn isotope baseline variability. However, our extensive δ⁶⁶Zn dataset includes not only multiple species from different localities and periods with distinct differences in dietary Zn uptake among extinct elasmobranch species, but also direct overlap in extant and fossil δ⁶⁶Zn values of the same genus and/or lifestyle. This spatial and temporal coherence suggests that it may be possible to use the same interpretative framework on extant and fossil elasmobranch assemblages globally (Figs. 1, 2, 3), and our remaining discussion is based on this assumption.

Zinc isotopes and ecology of Miocene-Pliocene sharks

Absolute and relative δ⁶⁶Zn values among some taxonomic groups show no statistical variation with geologic age and locality (e.g., Carcharias spp., Galeocerdo spp.), indicating relatively stable trophic levels and ecological niches throughout time and space. For example, most extinct elasmobranchs with a slender tearing, grasping tooth morphology (e.g., Carcharias) have δ⁶⁶Zn values that can be directly compared to modern equivalents (e.g., Carcharias taurus, Isurus oxyrinchus, Lamna ditropis). This type of dentition and corresponding tooth morphology are adapted to restrain small, active prey—like fish and cephalopods^28,29,30. However, there are differences among the δ⁶⁶Zn values for these types of elasmobranchs within the Early Miocene of Germany, with Mitsukurina lineata and Pseudocarcharias rigida having higher mean δ⁶⁶Zn values compared to Araloselachus cuspidatus (Fig. 2). Indeed, post hoc Tukey pairwise comparisons draw out A. cuspidatus as distinct from most species for the Germany (Early Miocene) assemblage, including those with a similar grasping tooth morphology (Supplementary Table 4). Our δ⁶⁶Zn values indicate that A. cuspidatus was likely a higher trophic level piscivore than M. lineata and P. rigida, supported by the larger tooth size of A. cuspidatus.

Zinc isotope values within the Galeocerdo lineage show no statistical variability with age nor locality, suggesting tiger sharks occupied a similar trophic level and ecological role in the marine ecosystem since at least the Early Miocene (Fig. 3). Notably, our results imply that the increase in body size from G. aduncus to the modern G. cuvier did not change its overall trophic level, which is in line with the highly similar tooth morphology between the two species³¹.

For Carcharhinus, the Early Miocene Malta assemblage is drawn out as statistically different from extant wild Carcharhinus spp. (Supplementary Table 5). Still, Carcharhinus spp. in both extant and fossil assemblages always have higher mean δ⁶⁶Zn values than other sympatric predatory sharks and are drawn out as statistically different from sympatric shark species in each fossil assemblage (Figs. 2, 3, Supplementary Tables 6–8). Based on similarities in tooth morphology and δ⁶⁶Zn values among extant and extinct Carcharhinus spp., we suggest that extinct taxa also primarily occupied a neritic-coastal habitat feeding upon demersal-benthic prey^22,23,24,25. For the Carcharhinus teeth from Malta, this interpretation is supported by lower δ¹⁸O_P values than sympatric species, indicating a higher water temperature or lower salinity: i.e., a shallow and/or brackish water habitat (Supplementary Fig. 8). Consequently, the uniformly higher δ⁶⁶Zn values of extant and fossil Carcharhinus spp. indicate the consumption of food items distinct from other measured sympatric species already during the Early Miocene and Early Pliocene.

Absolute δ⁶⁶Zn values for Otodus spp., along with values relative to sympatric species, indicate megatooth sharks were apex predators feeding at a very high trophic level (Figs. 2, 3). In all Early Miocene assemblages, mean O. chubutensis δ⁶⁶Zn values are among the lowest compared to sympatric species, including the lowest bioapatite δ⁶⁶Zn value measured to date (–0.83‰). Mean O. chubutensis δ⁶⁶Zn values are as low as extant Carcharodon carcharias (respectively, –0.57 ± 0.18‰, n = 19 and –0.57 ± 0.05‰, n = 4). Noteworthy, Games-Howell pairwise comparisons indicate the lower extant C. carcharias δ⁶⁶Zn values as distinct from most fossil Carcharodon populations, possibly indicating a dietary shift in the Carcharodon lineage (Supplementary Table 9). Early Pliocene values from O. megalodon from Japan also demonstrate very low mean δ⁶⁶Zn values (–0.62 ± 0.11‰, n = 5) that are statistically different from the Atlantic O. megalodon populations sampled from Florida and North Carolina, which have higher mean δ⁶⁶Zn values (respectively, –0.34 ± 0.11‰, n = 11; –0.38 ± 0.11‰, n = 7; Fig. 2, Supplementary Table 10).

Possible explanations for the observed spatial and temporal variability in Otodus and Carcharodon δ⁶⁶Zn values in our study are differences in prey consumption (and trophic level) or baseline variation. Additionally, we cannot rule out other factors such as interpretive limitations due to sample sizes. For example, extant C. carcharias can exhibit some degree of dietary individuality³², yet we only have δ⁶⁶Zn data from two individuals (4 teeth) from two localities. Still, the low δ⁶⁶Zn values in both extant C. carcharias compared to other extant sharks is in line with the generally high trophic level estimates of this species¹⁸. Particularly for O. megalodon from Japan where we have only one species analysed, we cannot exclude the possibility of either differences in δ⁶⁶Zn baseline or regionally different prey species. However, the absence of significant δ⁶⁶Zn differences within many taxa amongst locations and geological ages implies negligible differences in δ⁶⁶Zn food web baselines (Figs. 2, 3). Therefore, the observed spatial and temporal variability in δ⁶⁶Zn values likely demonstrates true dietary differences amongst Otodus and Carcharodon populations both geographically and temporally, with important implications for each species’ feeding ecology and evolution both on a local and global scale.

Otodus and Carcharodon in the Early Miocene are represented by O. chubutensis and C. hastalis, respectively, with statistically significant higher mean δ⁶⁶Zn values from the latter (Figs. 3, 4, Supplementary Tables 11, 12). The mean δ⁶⁶Zn value for all O. chubutensis is the lowest of all mean values recorded in our fossil shark dataset (Fig. 3), suggesting that O. chubutensis could occupy a higher trophic position than C. hastalis. Importantly, differences between δ⁶⁶Zn values of O. chubutensis and C. hastalis do not appear related to a different ratio of juveniles to adults in either species, as our results do not record an ontogenetic diet shift (Supplementary Fig. 12). We observe no correlation between the total body length of Otodus spp., Carcharodon spp. and their respective δ⁶⁶Zn values (Supplementary Fig. 12), likely, because each examined specimen had already surpassed the body size for which ontogenetic dietary shifts, if any, occur.

When including only Otodus spp. from the Atlantic and Paratethys/Tethys regions, we observe a statistically significant difference between O. chubutensis and O. megalodon (Supplementary Table 12, Fig. 4b). During the Early Pliocene, the Otodus lineage represented by O. megalodon shows a considerable increase in the mean δ⁶⁶Zn value for the Atlantic populations, hinting at a reduced trophic position for the megatooth shark lineage in the Atlantic. At the same time, the Early Pliocene C. carcharias remains at the same trophic level as C. hastalis (Figs. 2–4, Supplementary Tables 11, 12). Although the extant sample size is limited, our results are intriguing because the mean δ⁶⁶Zn value for extant C. carcharias places it at a trophic level that would be higher than the Atlantic Early Pliocene O. megalodon (Figs. 3, 4, Supplementary Table 11, 12).

Extant Carcharodon carcharias is a predatory shark whereby larger individuals regularly feed on high trophic level marine mammals³³. Although Neogene Carcharodon and Otodus were likely opportunistic in their prey selection similar to many extant apex predatory sharks³³, fossil evidence of bite marks suggests that both taxa fed largely on marine mammals such as cetaceans (mysticetes and odontocetes) and pinnipeds^{1,2,4,34,35,36,37,38}. However, in the majority of cases, it remains unclear if these feeding events on mammals document active hunting or scavenging and how important each prey taxa were to their overall diet. Early Pliocene C. carcharias and O. megalodon δ⁶⁶Zn data suggest that lower trophic level mammal prey such as mysticetes (and perhaps herbivorous sirenians) may have been an important food item for both species. Mysticetes are filter-feeders and likely to have higher tissue δ⁶⁶Zn values than piscivorous odontocetes or pinnipeds, similar to the higher δ⁶⁶Zn values in the plankton-feeding extant Cetorhinus maximus compared to piscivorous sharks (Fig. 1). Bite marks on Late Miocene–Early Pliocene mysticetes bones from both Carcharodon carcharias and O. megalodon^1,4,34,38 corroborate at least occasional feeding events.

Now extinct small- and medium-sized mysticetes (e.g., Cetotheriidae and various small-sized Balaenidae and Balaenopteridae) were abundant during the Early Pliocene^39,40 and were thus available as prey for large sharks, i.e., Otodus megalodon⁴ and Carcharodon carcharias¹. In contrast, Early Miocene cetacean fossils are dominated by toothed cetaceans, where the Early Miocene European and North American sites sampled in this study lack any mysticete remains^41,42,43,44. The Early Miocene Otodus (and modern C. carcharias) lower δ⁶⁶Zn values (higher trophic level) may partly be related to the lack of lower trophic level mammals (e.g., mysticetes) available as prey. Mysticetes became more abundant following a diversity plateau during the mid-Miocene^39,45. Subsequently mysticetes remains become more prominent in the Late Miocene to Early Pliocene fossil assemblages from North Carolina and Florida studied herein^43,44, where mysticetes were, together with other mammals (e.g., odontocetes), possibly preyed upon by O. megalodon and C. carcharias.

For the Early Pliocene of North Carolina, where we have δ⁶⁶Zn values for both Otodus megalodon and Carcharodon carcharias, our results suggest largely overlapping trophic levels for both species. Feeding at the same trophic level does not necessarily imply direct dietary competition, as both species could have specialised on different prey with similar trophic levels. However, at least some overlap in food items between both species is likely, as also indicated by fossil bite marks^1,4,34,38. Extant predatory sharks typically feed on a wide range of food items³³, and there is evidence for generalist feeding, as well as, in some cases, specialisation at lower trophic levels for extant C. carcharias³². Higher dietary individuality and the opportunistic nature of apex predators are possible explanations for the range of δ⁶⁶Zn values observed in both species (–0.61 to –0.04‰ in Pliocene North Carolina).

The extinction of Otodus megalodon could have been caused by multiple, compounding environmental and ecological factors^46,47, including climate change and thermal limitations⁴⁸, the collapse of prey populations⁴ and resource competition with Carcharodon carcharias¹⁵ and possibly other taxa not examined here (e.g., carnivorous odontocetes). The δ⁶⁶Zn results presented here indicate the potential of trophic change, where we find evidence for a decrease in the mean trophic position from O. chubutensis to O. megalodon in the Atlantic and an increase in trophic position for C. carcharias from the Early Pliocene to its extant form. If these trophic dynamics are accurate, then there is a possibility for the competition of dietary resources between these two shark lineages¹⁵. Our results also support the hypothesis of Otodus size-driven co-evolution and co-extinction with mysticetes⁴, indicated, at least for the Atlantic assemblage, by a shift towards lower trophic level prey from the Early Miocene to the Early Pliocene within the Otodus lineage. In general, our study demonstrates δ⁶⁶Zn to be a powerful, promising tool to investigate the trophic ecology, diet, evolution, and extinction of fossil marine vertebrates.

Source: Ecology - nature.com