Ontogenetic trophic segregation between two threatened smooth-hound sharks in the Central Mediterranean Sea

Elasmobranchs are among the species most threatened by overfishing and a large body of evidence reports their decline around the world. As they are large predators occupying the highest levels of marine food webs, their removal can alter the trophic web dynamic through predatory release effects and trophic cascade. Suitable management of threatened shark species requires a good understanding of their behaviour and feeding ecology. In this study we provide one of the first assessments of the trophic ecology of the “vulnerable” smooth-hounds Mustelus mustelus and M. punctulatus in the Central Mediterranean Sea, based on stomach contents and stable isotope analyses. Ontogenetic diet changes were addressed by comparing the feeding habits of three groups of individuals: juveniles, maturing and adults. Our results highlighted that the two species share a similar diet based mostly on the consumption of benthic crustaceans (e.g. hermit crabs). Their trophic level increases during ontogeny, with adults increasing their consumption of large-sized crustaceans (e.g. Calappa granulata, Palinurus elephas), cephalopods (e.g. Octopus vulgaris) and fish (e.g. Trachurus trachurus). Our results provide also evidence of ontogenetic shifts in diet for both species showing a progressive reduction of interspecific trophic overlap during growth. The results of this study contribute to improve the current knowledge on the trophic ecology of these two threatened sharks in the Strait of Sicily, thus providing a better understanding of their role in the food web.

Scientific RepoRtS | (2020) 10:11011 | https://doi.org/10.1038/s41598-020-67858-x www.nature.com/scientificreports/ ontogenetic stages. In this way, different species can use similar feeding resources, which decreases diffuse competition, and allows the coexistence of species within a given ecosystem 15−17 . Variability in trophic spectrum may lead to coexistence of several species inhabiting the same area through the exploitation of different prey items 18 . The genus Mustelus (Linck, 1,790) of the family Triakidae in the order Carcharhiniformes includes about 27 valid extant species in the world's oceans 19 . It is probably the most challenging group of elasmobranchs concerning taxonomic aspects because species are often difficult to identify due to their conservative morphology, combined with highly intraspecific variable diagnostic characteristics 20 . Mustelus species, commonly known as smooth-hounds, have a high level of regional endemicity and are commonly distributed on the continental shelf in temperate to tropical waters 21,22 . They are small to medium-sized demersal mesopredator sharks commonly exploited as target species or caught as by-catch by several different types of fisheries (e.g. artisanal, trawlers, recreational) in many oceanic areas 2,23 . Smooth-hounds are all viviparous, with reproductive modes including both placental and yolk-sac viviparity 20 . Some studies have also shown that smooth-hounds and other mesopredator sharks have an important role in marine ecosystems, inducing top-down control on the populations of their prey when their abundance is not consistently reduced by apex predators, such as big sharks 7,24 or by fishing.
In the Mediterranean Sea, there are three smooth-hound species: the starry smooth-hound, Mustelus asterias Cloquet, 1819, the common smooth-hound, M. mustelus (Linnaeus 1758), and the blackspotted smooth-hound, M. punctulatus (Risso, 1827) 20 . The last two species, inhabiting the shelf up to − 200 m deep 25,26 , are morphologically very similar and consequently often misidentified during field work. More so because identification guides or papers report different and sometimes contrasting diagnostic traits. Such species similarity might be the result of a relatively recent speciation event and it is not surprising that in a recent genetic study, the authors found signs of hybridisation events between M. mustelus and M. punctulatus, although hybrid viability is actually unknown 27 . On the other hand, the starry smooth-hound is more easily distinguishable because of its body colouration and evident white spots 28 . Sharing the same area and having similar morphological characteristics, these species could compete for food resources. All the three smooth-hounds are threatened in the Mediterranean region. According to the IUCN's European Red List of marine fish 6 , M. mustelus and M. punctulatus are classified as 'Vulnerable' and M. asterias is evaluated as 'Near Threatened' . However, a recent study has shown a decline in common and blackspotted smooth-hounds of 72% and 78% respectively over three generations (i.e. 60 years), thus suggesting that in the next regional assessment a shift to the 'Endangered' category should be considered 29 . The case of the starry smooth-hound is even worse since it has disappeared from most of the Mediterranean coastal areas suggesting a high risk of regional extinction 29 . The current smooth-hound status is the result of a long fishing exploitation history that began in the neolithic and has rapidly accelerated over the past 50 years due to the expansion of trawling 30,31 . Currently, smooth-hounds are still exploited or caught as by-catch in trawling and artisanal fisheries in the few Mediterranean areas where viable populations are still existing, such as Tunisia 32 , Northern Adriatic 33,34 and Mediterranean Turkish coasts 26,35,36 .
In this study we focused on the trophic ecology of the common (M. mustelus) and blackspotted smooth-hound (M. punctulatus) in the Northern sector of the Strait of Sicily (SoS), an area that probably hosts some of the last viable populations of these species in European Mediterranean waters 29 .
The knowledge about life history traits of the two species is poor and geographically scattered. Information on sexual maturity and reproduction is available for Tunisia 25,32,37,38 and the Northern Adriatic Sea 33 . Age and growth data are available only for the common smooth-hound in the northern Aegean Sea where it appears late-maturing, long-lived and slow-growing 26 . Similarly, current knowledge about prey preferences is limited to studies conducted in the Northern Adriatic for the black spotted smooth-hound, in the Western Mediterranean (Gulf of Valencia, Spain) and Aegean Sea for the common smooth-hound 37,39 and in Tunisia for both species 25,40 . All these studies showed that the smooth-hounds are important predators of coastal crustaceans, cephalopods and fish, and may also have a major impact on important commercial species, such as the spiny lobster and common octopus, with potential repercussions for coastal fisheries.
The main objective of our study is to elucidate the trophic role played by the common and the black-spotted smooth-hounds in the SoS for a better understanding of the possible factors triggering species coexistence. To this aim, diet composition, trophic niches and trophic levels were analysed across ontogenetic stages through a combined approach based on stomach contents and stable isotope analysis.

Material and methods
Study area. The study area covers approximately 34.000 km 2 off the southern coast of Sicily, corresponding to the North sector of the Strait of Sicily (SoS- Fig. 1, see 41 ). It is characterised by high primary productivity and high levels of biodiversity due to the occurrence of complex and diverse benthic biocoenosis 42,43 . Recent studies showed high diversity and biomass of demersal communities over the offshore detritic bottoms of the Adventure bank 43−45 . Given the importance of this area for conservation, several sites have been identified for inclusion in a Mediterranean network of marine protected areas 46 . The off-shore banks are particularly important for elasmobranchs both in terms of species diversity and abundance 48 . In addition, the SoS is one of the most important Mediterranean fishing areas, with a high concentration of fishing vessels from different countries (i.e. Italy, Malta, Tunisia), which exploit a wide range of demersal and pelagic species 49 . From an oceanographic point of view, the SoS plays a key role for the Mediterranean thermohaline circulation 50 and the exchange of water masses between the eastern and western Mediterranean basins. The Atlantic Ionian Stream (AIS) and the Atlantic Tunisian Current (ATC) flow across the SoS 51 . The AIS is associated with a number of well-known semi-permanent features in particular two large cyclonic vortexes; the first one lies over the Adventure Bank and the second, over the Malta shelf, off Capo Passero.  Specimens were immediately placed on ice after capture and taken to the laboratory. Once in the laboratory, sharks were identified, measured to the nearest 1 cm (total length, TL), weighed to the nearest 1 g, and both sex and maturity were registered using the maturity scale for viviparous sharks adopted by the MEDITS survey protocol 52 . Species identification was based on the inspection of dermal denticles using a stereomicroscope 53 . After measurements, stomachs were removed and stored in labelled plastic boxes filled with a 70% ethanol solution. In addition, from the individuals caught during the last sampling year (2017), a small section of dorsal muscle (1 g) of both M. mustelus (n = 58: 36 juveniles, 18 maturing, 4 adults) and M. punctulatus specimens (n = 39: 20 juveniles, 14 maturing, 5 adults) was removed for stable isotope analysis, rinsed with deionised water, oven dried for 3 days at 60 °C, then ground with a pestle and mortar and kept in a dry room until analysis.
Gut content analysis. In order to assess dietary ontogenetic shift of the two Mustelus species, the specimens of each species were grouped in three size classes according to sexual maturity: juvenile (mmus: TL ≤ 80 cm; mpun: TL ≤ 70 cm); maturing (mmus: 81-105 cm TL; mpun: 71-90 cm); and adult (mmus: TL > 105 cm; mpun: TL > 90 cm) 25,38 . To obtain a precise description of the diet, it is important to determine the minimum number of stomachs required (Ferry and Cailliet 1996). The number of both mmus and mpun collected was tested to determine whether sufficient specimens were sampled. Given the rather low number of stomach contents, especially for adults, we decided to pool together the available stomach sample data and calculate the saturation curves by species. The cumulative number of randomly pooled stomachs was plotted against the cumulative number of prey taxa. www.nature.com/scientificreports/ Prey items in each stomach were sorted and determined to the lowest possible taxonomic level. For a more accurate prey identification, all the hard parts were cleaned with distilled water and otoliths were identified using otolith Atlas 47 . Then, all prey items were counted and weighed. For the evaluation of the diet composition of each species and size class, we calculated the following indices: (1) Vacuity Index (VI = number of empty stomachs/ total number of stomachs × 100), (2) relative frequency of occurrence (%F = number of stomachs containing prey i/total number of full stomachs × 100), (3) relative numerical abundance (%N = number of prey i/total number of prey × 100), and (4) relative gravimetric composition (%W = weight of prey i/total weight of all prey × 100). The index of relative importance (IRI) of 54 , as modified by 55 , was used to assess the importance of each prey item in the diet of the two species at each size class: IRI = %F × (%N + %W). Prey species were listed in decreasing order according to their relative IRI contribution and then %IRI was calculated for each prey item as follows: Interspecific differences in diet composition were tested using permutational multivariate analysis of variance (PERMANOVA) 56 and based on the prey biomass from the Bray Curtis similarity matrix were obtained from square root transformed data.
The analysis, carried out with the PRIMER statistical package 6.0 57 , included one factor: "species" (two levels: mmus and mpun).
The effect of year on the diet composition of both species size classes was not analysed due to the limited number of samples available, thus implicitly assuming a constant diet through time.
Stable isotope analysis. Subsamples of dorsal muscles destined to carbon stable isotope analysis were delipidated before conducting analysis. The lipids were removed by rinsing the ground tissue several times with 2:1 chloroform:methanol mixture according to 58−60 . For both species and all size classes, approximately 1 mg of individual muscle sample was weighed in tin capsules, automatically loaded in an elemental analyser (Thermo Flash EA 1,112) for the determination of total carbon and nitrogen, and then analysed for δ 13 C (from delipidated subsamples) and δ 15 N (from bulk subsamples) in a continuous-flow isotope-ratio mass spectrometer (Thermo Delta Plus XP). Stable isotope ratio was expressed, in relation to reference international standards (atmospheric N 2 and PeeDee Belemnite for δ 15 N and δ 13 C respectively), as: where R is the 13 C/ 12 C or 15 N/ 14 N respectively. Analytical precision based on standard deviations of internal standards (International Atomic Energy Agency IAEA-CH-6; IAEA-NO-3; IAEA-N-2) ranged from 0.10 to 0.19‰ for δ 13 C and 0.02 to 0.08‰ for δ 15 N.
The changes in carbon and nitrogen stable isotopic signatures of the two Mustelus species with size increasing was evaluated through linear regressions between δ 13 C (‰) and δ 15 N (‰) as dependent variables and size (TL, cm) as the independent variable. To test for differences in isotopic niche width between the two Mustelus species and among size classes, standard ellipse areas (SEAc and SEAb) were estimated by Bayesian statistics based on carbon and nitrogen stable isotope data. In more detail, SEAc, which is the standard ellipse area corrected for small sample size, was set to contain 40% of the data and provided unique values, while SEAb, which is the Bayesian standard ellipse area, derives from 4,000 posterior iterations and is reported as mode along with 95% credible interval 61 . Differences in SEAb between species and among size classes were tested through pair-wise comparisons, by calculating the probability that the SEAb of one group is larger than that of the other, hence a significant difference was regarded as a probability of at least 95% 61 . SEAc, SEAb, as well as the overlap between SEAc, were estimated with R package SIBER v. 2.1.3 (Stable Isotope Bayesian Ellipses in R) 61,62 .
Trophic positions based on stable isotope data (TP SIA ) were estimated according to the following equation 63 : where δ 15 N and δ 15 N b are respectively the nitrogen isotopic signature of each Mustelus specimen, and that of the baseline, for which we used the mean value of Pagurus prideaux (see results) sampled in the same study area (6.88 ± 0.36 ‰). Δ n is the trophic enrichment expected for each trophic level (2.78 ‰), according to 64 , and λ is the trophic position of the baseline, that was set to 3 as the mean in the literature data 65,66 . Additionally, trophic position for the three size classes of both target species were calculated based on stomach content data (TP SCA ) following the method proposed by 67 : where TLej is the trophic level of each prey category j and Pj is the proportion of each prey category to the diet of the species. Prey categories taken into account were cephalopoda, decapoda, fish, molluscs and other invertebrates, whose relative trophic level was taken from 67 .

Results
We collected 234 specimens of mmus and 98 of mpun, whose sizes ranged between 31.5 and 170.0 cm total length (TL) for mmus, and between 35.5 and 120.0 cm TL for mpun (Fig. 2, Table 1). A higher statistically significant vacuity average index was found in mpun (VI = 34.5%) than in mmus (VI = 23.7%) (Mann-Whitney p = 0.0125). The total number of prey taxa found in the stomach was 41 for mmus (some of the prey were found in different  Fig. 4). In more details, juveniles fed mostly upon portunids and hermit crabs, with Pagurus prideauxi (%F = 13.5; %IRI = 29.0) and Liocarcinus depurator (%F = 10.9; %IRI = 15.6) as the main prey, followed by teleostea as secondary prey (%IRI = 14.7). Maturing specimens appeared to switch their diet to fish (%IRI = 49.3), which were found generally too highly digested in the stomachs to allow for prey species identification, and Anomura (%IRI = 28.6). In this latter group, the most important prey was the hermit crab P. prideauxi (%F = 11.7; %IRI = 27.1), whilst portunids were replaced by the bigger-sized shamefaced crab Calappa granulata (%F = 8.3; %IRI = 12.1). This latter species was dominant in the diet of adult mmus (%F = 22.2; %IRI = 78.9), which was found to feed also on cephalopods, such as Octopus vulgaris (%IRI = 5.2) and fish (%IRI = 5.0) ( Table 2 (Table 3; Fig. 4). The importance of cephalopoda and fish increased with size with the %IRI of the former, increasing from 21.2% to 54.8% in maturing and adult individuals respectively. In this latter size class, O. vulgaris was one the most important prey, with approximately 10% of the IRI. Fish prey also increased progressively in mpun stomach contents from juveniles to maturing and adult individuals (%IRI = 0.6, 16.1 and 31.4 respectively). In most cases, fish prey identification at the species level was impossible due to the advanced digestion state found (Table 3; Fig. 4). The diet of mmus and mpun, based on the analysis of stomach content, was significantly different. In more detail, multivariate analysis of variance (PERMANOVA) based on the prey biomass showed significant interspecific differences (Table 4; Fig. 4). The mean trophic position calculated from isotopic data (TP SIA ) was comprised between about 3.7 and 4.1 in mmus and between 3.6 and 3.9 in mpun (Table 5). Trophic position based on stomach content data (TP SCA ) showed lower variability among size classes, than TP SIA , being 3.7 in all the mmus size classes and varying between 3.4 and 3.6 in mpun (Table 6). Moreover, it should be noted that TP SIA showed a gradual increase across size classes, which was not observed in TP SCA .

Stable isotope values
The isotopic niche of both species and all size classes is represented by the corrected Standard Ellipse Areas (SEAc, Fig. 6). Juveniles of mmus and mpun showed the widest and the narrowest isotopic niche respectively, and only a partial mutual overlap (accounting respectively for the 12% and the 50%, Fig. 6, Table 6). In contrast, the isotopic niche of mmus and mpun maturing specimens showed a high mutual overlap (respectively 72% and 86%), and both were highly overlapped with the niche of mmus juveniles (respectively 83% and 78%). Adult specimens of both species showed a rather narrow niche with a very low reciprocal overlap (respectively 6% and 8%), which was due to the higher δ 15 N values of mmus adults. Only the niche of mpun adults was partially overlapped with the other niches, accounting for about 62% on average (Fig. 6, Table 6). Accordingly, the patterns highlighted by the Bayesian standard ellipse area (SEAb) showed significantly higher values of the isotopic niche width in the mmus juveniles and the maturing specimens of both mmus and mpun, than in the mpun juveniles and the adult specimens of both species, which, in turn, showed the lowest values of the isotopic niche width (Fig. 7).

Discussion
In this study we compared the diet composition, and, for the first time in the Mediterranean Sea, isotopic niches and trophic positions of two threatened smooth-hounds, mmus and mpun, co-occurring in the southern coasts of Sicily (northern sector of the Strait of Sicily, central Mediterranean Sea), across size classes. The study adopted an integrated approach based on the analyses of both stomach contents and stable isotopes. Integrating the two sources of data can be extremely beneficial in studies addressing the trophic ecology of poorly abundant or rare species, where the available number of samples is limited, as for the two smooth-hound of the South of Sicily. Gut content analysis, although useful to determine the composition of the diet, reflects the identifiable food ingested recently, and hence it can underestimate the prey ingested over a longer period. In contrast, stable isotope analysis integrates diet and food assimilation over time (3-4 months, according to 63,68,69 ). Although the different time scale of the two approaches may lead to apparently inconsistent results, the combination of both provides complementary insights that may help in disentangling complex trophic aspects. The analyses of    www.nature.com/scientificreports/ stomach contents revealed that the diet of the two smooth-hound species was composed of benthic and demersal species, mostly crabs, cephalopods and fish, showing that they actively search for their prey close to the bottom and then progressively shift toward bigger-sized species during ontogeny. The diet of mmus showed progressive changes during ontogeny, although not so straightforward, with smaller individuals preying mainly on benthic crustaceans (both anomura and brachiura) and fish. The maturing specimens showed a higher reliance on fish, despite consumption of crustaceans remaining fairly high, and the largest specimens relied on big-sized brachiura (i.e. C. granulata), fish and cephalopods (e.g. O. vulgaris). Prey preference of mpun across size classes showed the juveniles feeding mostly on small hermit crabs (P. prideauxi), maturing specimens broadened their diet to include fish and cephalopods that became the almost exclusive prey of adults. These findings highlight that ontogenetic shift in diet was more evident for mpun, compared with the congeneric mmus, consistent with previous studies carried out in Tunisia 25,40 and the Adriatic Sea 37,39 . A similar pattern in dietary shift during growth was found for mmus in South Africa 70 and confirms that a diet mostly based on crustaceans, mainly anomura, often with fish and cephalopods as secondary prey is a common feature in smooth-hounds, and observed in many different species and marine regions (see 71 and references therein).
Nevertheless, unlike our findings, along the Tunisian coasts, the two species showed a wider trophic spectrum with mmus preying upon several species of crustaceans, fish and cephalopods, along with sipunculids, polychaetes and echinoderms as occasional prey 40 and a similar pattern was highlighted for mpun 25 . Dietary studies on mpun from the Adriatic Sea also reported the occurrence of crustaceans, fish and cephalopods in their stomachs 37,39 with an increased predation of molluscs in adult sized individuals, as already observed in Tunisian waters 25,40 . More similar to our findings, the only study carried out in the Western Mediterranean (Gulf of Valencia, Spain) reported that the common smooth-hound juveniles (TL < 75 cm) prey mostly upon Liocarcinus spp. crabs and the stomatopoda Squilla mantis 72 . Similarly, a study in the Aegean Sea identified crustaceans as the main prey of mmus 73 . Fish, instead, was the main prey of Mustelus mustelus in Lybian waters 74 .
The analyses of the isotopic niches of the two congeneric Mustelus species across size classes well mirrored the results obtained through the stomach contents analysis. Particularly, consistent with the differences in diet composition between the juvenile specimens of the two species, the related isotopic niches were only partially overlapped, with the niche width of mmus wider than that one of mpun. This result mirrors the wider carbon and nitrogen ranges of the former species, compared with the latter, and is most probably driven by the higher diversity of prey consumed (26 vs. 15 prey items for mmus and mpun respectively). Besides the two common dominant prey (P. prideauxi and L. depurator), other prey showed a high relative importance (> 5%) in juvenile mmus, among which other anomura, fish and unidentified organic remains, differing from mpun. In contrast to juveniles, maturing specimens showed wide and highly overlapped niches as well as similar isotopic values. Although in contrast with diet composition results, this isotopic similarity reflects a trophic switch observed in both species that, besides crustaceans, started to include bigger sized-prey with a higher trophic level in their diet, such as cephalopods and fish 37,39,75 . For example, maturing mpun specimens, showed a niche much wider than the one of juveniles and was coupled by an increase of the δ 15 N values, confirming the gradual ontogenetic www.nature.com/scientificreports/ shift toward prey at a higher trophic level (e.g. fish and cephalopods). The niche expansion was followed by a further narrowing, that was evident in the adult specimens of both species and could be due to species-specific trophic habits on a few prey items, that also implies a clear niche segregation 75,76 . The low number of adult samples however does not allow to exclude a sample-size effect on the observed reduction in prey diversity and further studies are required to better explore this pattern.   www.nature.com/scientificreports/ The high δ 15 N values recorded in adult specimens are indeed consistent with predation on carnivorous prey, such as cephalopods and fish, but seem less consistent with the almost exclusive dominance of the crab C. granulata in the mmus stomachs. Although very poor information is available so far about the diet of C. granulata (bivalves in 77 ; crab remains from author personal observations), the shovel-shaped chelae allow Calappidae crabs to break hard shells and carapaces revealing a carnivorous opportunistic foraging behaviour 78 that may imply also a high trophic level. It is worth noting that probably the segregation between adults of both species is not only due to the segregation in diet but also to difference in body sizes since mmus adults are almost 20 cm longer than mpun adults as shown in Fig. 5. The low trophic niche overlap, especially between the juveniles and adults of both species, highlighted here, suggests that resource partitioning facilitates coexistence. This is particularly important for endangered species living in the same area and competing for the same resources in their different  Table 5. Summarized stable isotope data (mean ± standard deviation of δ 13 C and δ 15 N) for M. mustelus and M. punctulatus size classes. Mean ± standard deviation of the trophic position based on stable isotope data (TP SIA ) is also shown along with trophic position based on stomach content analysis (TP SCA ). The number of the individuals taken into account for both methods is also specified.

Species
Size class n SIA Trophic position based on stable isotope (TP SIA ) and stomach content (TP SCA ) data of the two shark species was overall comparable in most cases, but for adults TP SCA values were slightly lower compared to TP SIA . Moreover, the patterns of trophic position across size classes were different: consistently with isotopic niche patterns, isotopic values and TP SIA subtly increased across size classes, especially in mpun, also revealing the highest values in adult mmus, while TP SCA did not. Slight variations between the two methods have already been observed in 80 and 81 and have been linked to the intrinsic differences of the methodological approaches. www.nature.com/scientificreports/ Although the mean trophic position of both species was overall lower than previous results for maturing specimens of Mustelus spp. (4.0 ± 0.3 in 64 , 3.9 and 4.3 in 65 , 4.2 ± 0.6 in 58 ), ontogenetic diet shifts are common features in sharks and mirror an increased preference toward prey at higher trophic levels 76,82,83 . The observed progressive diet shift of the two smooth-hounds, occurring with a reduction in their interspecific trophic overlap during growth, can play an important role in promoting species coexistence, as shown also in other studies 84,85 . Lack of a strong interspecific segregation, as expected for two ecological and morphological similar species, such as mmus and mpun, can be the effect of a reduced trophic competition. Past and current overexploitation has strongly reduced the size of the Mustelus populations in the Mediterranean region, thus reducing any selective pressure that would have induced an interspecific shift in prey preferences. Nevertheless, field studies on Mediterranean marine fish have shown that coexistence of trophic guilds (i.e. group of species with similar trophic preferences) is promoted by species segregation along the depth/habitat dimension, so that species with similar prey preference are generally distributed in different habitats 86 . In the case of the two shark species, a lack of a significant segregation both along the habitat and trophic habits can be an indication that competition is not a critical factor for species coexistence. In the literature, several mechanisms have been described to minimize competition for food resources, such as prey size, species segregation, habitat separation, differentiation in foraging tactics, predator size and morphology 87,88 . Among these mechanisms, a differentiation in foraging tactics could partially explain the results obtained in this study. 85,89 .
It is worth noting that one of the main limits of our work is the sample size particularly for adults, as well as the lack of data on interannual changes in diet composition cannot allow to deeply explore interspecific differences in feeding strategy and how these can promote species coexistence. The results of this study however can reflect the current depleted status of the Mustelus populations where the selective pressure for segregation in diet between the two species is presumably low. Further studies in this direction should be carried out especially if sound conservation measures will be implemented to protect and rebuild Mustelus populations in the SoS.

conclusions
Although trophic interactions among organisms are one of the crucial drivers of ecosystem dynamics, dietary information of wide-ranging predators is often missing. Still today, the knowledge of the trophic role of demersal sharks in the Mediterranean Sea remains limited to the most common species, such as Galeus melastomus and Etmopterus spinax 90 , whereas, there is a poor knowledge about the lesser abundant species. Among these, Mustelus spp. played a very important role in the nectobenthic communities of the Mediterranean continental shelf, as witnessed by the relative abundance found in old surveys and catch records 29 . The SoS is one of the Mediterranean areas still hosting viable populations of smooth-hounds, probably due to the occurrence of large off-shore banks that offer refugia and protection from fishing 41 . Both smooth-hound species here studied are common in this area, including both African (i.e. Tunisia, Libya) and Italian (Sicily) waters, but they are still highly exploited by targeted artisanal fisheries and show an alarming decline 91 .
With this study, we improved the knowledge of both diet composition and trophic niches of mmus and mpun, two endangered sympatric species showing very similar morphological characteristics and apparently sharing the same habitats 53 .
By applying for the first time the complimentary and robust approaches based on stomach contents and stable isotopes, we found that the two endangered sympatric Mustelus species inhabiting the north sector of the Strait of Sicily share similar trophic habits, although showing a progressive interspecific segregation during growth. The main findings highlight a rather generalist feeding behaviour mostly based on benthic crustaceans for the juvenile and maturing specimens of mmus, followed by a trophic niche narrowing in the adult stage. While mpun showed narrow niches during both the juvenile and adult stages, with a progressive ontogenetic shift from benthic crustaceans to cephalopods and fish, interspersed with a niche widening in maturing specimens. and M. punctulatus (mpun): j, juveniles; m, maturing; a, adults. Boxes from dark to light grey represent the 50%, 75% and 95% respectively of credibility intervals while black points represent the mode. Italic letters underneath the boxes indicate homogeneous groups based on pairwise probability test (not significant differences for p < 095 and p > 005). This figure was created using R version 3.6.2 https ://www.r-proje ct.org/ by CA.
Scientific RepoRtS | (2020) 10:11011 | https://doi.org/10.1038/s41598-020-67858-x www.nature.com/scientificreports/ Lastly, this study confirms that trophic shifts during growth play a key role in the foraging ecology of sharks and is a crucial aspect for understanding the ecological role of these predators. Furthermore, by providing quantitative trophic information, this study gives useful insights for defining future conservation strategies in light of the dramatic decline of smooth-hounds in the Mediterranean Sea, by making evident the possible trophic cascading effects that may be induced by the rebuilding of their populations.