The trophic state of the environment, and thus the availability of food, is generally considered a key factor influencing foraging behavior (16). Surprisingly, however, it does not appear to be the key factor affecting species TNW, the most general aspect of foraging behavior. Although the smallest average TNW was observed in the reservoir with the highest trophic state, there was considerable variability among species TNW sizes and among sites. The combination of environmental heterogeneity, species richness, intraspecific or interspecific competition, and predation pressure will influence TNW more than the site trophic state itself (31).
As we hypothesized, the largest TNW was achieved by the apex predator, the European catfish. This can be explained by the theory of generalism of large-bodied apex predators, with their high energy requirements and minimum foraging constraints they utilize a larger percentage of the TNW (9). In contrast, the TNW of the other species were relatively similar to each other. The Northern pike had the smallest TNW, being an obligate predator feeding exclusively on moving prey (32). The trophic position of Northern pike at all study sites is just below that of European catfish, which appears to control the species through both competition and predation (33, 34). A significant influence of European catfish was evidenced by the finding that the smallest TNW of all species and at all sites for Northern pike was observed at site 4, where catfish biomass was highest (35). In contrast, the TNW of Northern pike was greatest at sites 2 and 3, where the European catfish biomass was lowest (35). Thus, considering the ecological factors, the results prove that the Northern pike is a foraging specialist (12). The TNW size of the other species (Eurasian perch, rudd) was between the European catfish and the Northern pike. Eurasian perch has the role of a mesopredator with a wide range of prey (fish, zoobenthos, zooplankton) (22). Similarly, rudd can feed on a relatively wide range of food sources, from various species of macrophytes to zoobenthos (36), and at larger sizes can rarely switch to predation (10). Due to sufficient food availability, grazers are not pushed to become foraging specialists, in contrast to obligate predators, which are expected to have high requirements for learning specific hunting skills (14).
The most surprising results emerged from the INW in relation to fish body mass. We assumed decreasing INW with increasing body mass. This strategy minimizes the time required to search for and handle prey and is consistent with optimal foraging theory (14), but it could not be clearly demonstrated across the entire size spectrum of any of the species. The trend was clear in smaller individuals of apex predator, obligate predator and mesopredator. All individuals were adults, but with significantly different body mass, typical of ectotherms with indeterminate body growth (37). European catfish and Northern pike clearly exhibited the Apex predator effect, in which feeding behavior aimed at increasing specialization suddenly changes to generalism after reaching a certain threshold of body mass. This effect was particularly evident when all individuals from all study sites were considered. The thresholds of body mass of European catfish and Northern pike were about 10 and 5 kg, respectively. The effect was also partially observed for the mesopredator (Eurasian perch), but only three very large individuals were caught (2 × 0.9 kg; 1 × 1.9 kg). INW size seemed to increase rapidly for these largest individuals. However, due to the limited data set, the result was not statistically significant. Future studies could focus on the INW of Eurasian perch or other smaller predators at sites where they are the largest individuals or where large-bodied apex predators are not present, so that perch has the role of the apex predator (33), and the peak predator effect can be observed.
Thus, generalism above a certain threshold of body mass may be determined by the attainment of the highest trophic position. As an apex predator, a species has no limitation in its foraging behavior, has high nutritional requirements (9), and does not need to minimize foraging time thanks to low predation risk (14). The second, less likely reason could be the fact that specialists tend to be more vulnerable to environmental changes than generalists, which are able to adapt better (38). This theory can be applied also at the individual level. Older and larger individuals are expected to have repeatedly encountered critical periods of food shortage, and this 'die or adapt' selection pressure (39) forces them to expand their food spectrum. Our results bring forth a fundamental question regarding the differences in the ways ectothermic and endothermic species rise to the position of apex predator. Endothermic vertebrates, mammals and birds, are already born as apex predators (34) thanks to parental care and common group formation (40), whereas ectothermic vertebrates with indeterminate growth must reach the apex predator position via lower trophic positions (41). Only some species, and only a limited number of individuals, can reach this position (8, 42). Thus, for ectothermic species, apex predator status is not directly linked to the species but rather to the size of the individuals.
The size of INW of the herbivorous generalist rudd follows the opposite trend from what we expected. INW size increased with body mass, but the reason is not entirely clear. Ectotherms with a tendency to herbivory often switch from omnivory to herbivory over time (43), so INW size should decrease. Rudd have a pronounced seasonal variability of diet depending on temperature (11). Sampling was done at the end of summer, so SIA should mainly reflect the growing season with abundant plant food and ideal temperature for cellulose digestion (See SI Appendix for isotopic half-lives of rudd tissues), rather than the cold period when the rudd diet changes significantly (11). A possible explanation could be the ability of large rudd individuals to digest a more diverse plant diet (for instance with low levels of toxic substances) that smaller individuals cannot digest due to the physiological limitations (43), or the occasional piscivory (10).
The four described trends of INW were combined in the increasing order of the fish trophic position. This theoretical INW model for the entire food web suggests that the INW of herbivores is small, it increases in generalists, decreases in predators, and is greatest in apex predators. It would be prudent to corroborate these results with a similar study design using similar trophic position of different fish species in a different ecosystem, or even with endothermic vertebrate species. The latter model could include more than four species in the food web with fewer individuals per species.
When multiple species coexist in an ecosystem, their partitioning of food resources is usually specialized (21). According to our observations, unlike TNW, the extent of IS is significantly influenced by the site trophic state. This is particularly evident in apex and obligate predators and is consistent with the theory that individuals are forced to become more specialized when resources are limited (16, 25). The trend is more evident at the apex predator position, likely due to overall 90% energy loss at each level in the food web (44). Therefore, predators, especially apex predators, face greater resource limitation than species at lower trophic positions and must respond more strongly to site trophic decline. In general, exposure to more demanding conditions leads to greater inductive specialization (25).
Contrary to expectations, the degree of IS did not increase with trophic position (45). This trend was partially observed only at oligotrophic sites 1 and 2. In contrast, the trend was slightly opposite at mesotrophic sites 3 and 4. There, the IS of the apex predator was extremely low. Thus, the overall trend of IS apparently changes at the position of the apex predator, similar to the INW closely associated with IS (25, 46). Although these results are intriguing, we are well aware of the limitations of the selected study sites, where two oligotrophic and two mesotrophic sites have very similar TP concentrations. Accordingly, it is necessary to approach these findings with caution. Present study shows the direction; however, further research will be needed into this important yet so far largely neglected phenomenon we have highlighted above.
The degree of IS in relation to tissues differing in isotopic half-life indicates whether the IS is short-term (seasonal) or long-term in nature. The degree of IS of European catfish (apex predator) and Eurasian perch (mesopredator) was generally low and even lower in the long term. Switching between food sources is thus very common during the season depending on availability, but this observation needs to be confirmed by replicating the results in more studies. This behavior is fully consistent with the Optimal Foraging Theory, which states that individuals use the most optimal and available resources. When the previously available resource becomes scarce, individuals switch to the most readily available food source (20). Both European catfish and Eurasian perch are characterized by high food plasticity (7, 9). Low IS with instability over time is probably a general trend for apex predators and mesopredators, consistent with their role in the ecosystem (8, 33). In contrast, Northern pike as an obligate predator generally has a high IS. It is obviously a long-term pattern, as signals from different tissues with very different isotopic half-lives were similar. The low variability of tissues with similar isotopic half-lives was likely due to sporadic consumption of low abundance food sources that were not present at all in tissues with the long half-life (47). The long-term specialization has also been observed in other typical obligate predators, seabird Northern gannet (Morus bassanus (48)) or the loggerhead sea turtle (Caretta caretta (13)), both with extensive migration. Thus, it seems that food specialization at the individual level is a typical feature of obligate predators, but not of apex predators or mesopredators.
Contrary to our expectations, the isotopic signals of rudd did not show short-term IS. Rudd commonly switch between omnivorous and herbivorous feeding depending on water temperature (11), thus, a significant difference in isotopic signals was expected. The uniform signals revealed by our data were probably caused by sampling at the end of the growing season, when all tissues already reflected plant nutrition from the warm part of the year. Variability among individuals could be caused by i) the largest individuals, probably due to their physiological abilities, exploiting a more diverse diet (as mentioned above in the context of INW), ii) a probable specialization of some individuals on nutrient-rich foods (flying insects or benthos (43)), or iii) by a difference between males and females, which are forced to use more nutrient-rich food due to gonad formation (17). However, sex was not determined, and we plan to take this factor into account in future studies. In order to obtain the entire variability, it would be advisable to collect tissue samples several times a year, especially when the water temperature drops below 15°C (11).
Our results show that site trophic state has no significant impact on TNW, but a significant impact on IS, especially on the IS of species in higher trophic positions. INW is significantly influenced by trophic position of the species and its magnitude reaches two maxima in the positions of generalist and apex predator. The findings were presented on four freshwater fish, but they may be generalized to other food webs with similar trophic positions, in both terrestrial and aquatic ecosystems.