Herbivory is more prevalent in the tropics than at higher latitudes. If differences in ambient temperature are the direct cause for this phenomenon, then the same pattern should be visible in a seasonal gradient, as well as in experiments manipulating temperature. Using (15)N stable isotope analyses of natural populations of the copepod Temora longicornis we indeed observed seasonal differences in the trophic level of the copepod and a decrease in trophic level with increasing temperature. In a grazing experiment, with a mixed diet of the cryptophyte Rhodomonas salina and the heterotrophic dinoflagellate Oxyrrhis marina, T. longicornis preferred the cryptophyte at higher temperatures, whereas at lower temperatures it preferred the non-autotrophic prey. We explain these results by the higher relative carbon content of primary producers compared to consumers, in combination with the higher demand for metabolic carbon at higher temperatures. Thus, currently increasing temperatures may cause changes in dietary preferences of many consumers.

Nutritional imbalances between predator and prey are the rule rather than the exception at the lower end of food webs. We investigated the role of different grazers in the propagation of nutritionally imbalanced primary production by using the same primary producers in a threetrophic-level food chain and a four-trophic-level food chain experimental setup. The three-trophic-level food chain consisted of a classic single-cell primary producer (Rhodomonas salina), a metazoan grazer (the copepod Acartia tonsa) and a top predator (the jellyfish Gonionemus vertens), while we added a protozoan grazer (Oxyrrhis marina) as primary consumer to the food chain to establish the four-trophic-level food chain. This setup allowed us to investigate how nutrient-limitation effects change from one trophic level to another, and to investigate the performance of two components of our experimental food chains in different trophic positions. Stoichiometry and fatty acid profiles of the algae showed significant differences between the nutrient-depleted [no N and no P addition (-P), respectively] and the nutrient-replete (f/2) treatments. The differences in stoichiometry could be traced when O. marina was the first consumer. Copepods feeding on these flagellates were not affected by the nutritional imbalance of their prey in their stoichiometry, their respiration rates nor in their developmental rates. In contrast, when copepods were the primary consumer, those reared on the -P algae showed significantly higher respiration rates along with significantly lower developmental rates. In neither of our two experimental food chains did the signals from the base of the food chains travel up to jelly fish, our top predator.

It is a well accepted fact that nutrient limitation of plants affects the growth and survival of herbivores, generally leading to lower performance of herbivores feeding on nutrient stressed plants. The effect of plants' growing conditions on predatory organisms, feeding one trophic level up, has been much less studied, and there is a general consensus that such effects would be small as herbivores often show relatively strong homeostasis with respect to their nutrient content. Here, we challenge this view, and show from several examples that despite the fact that herbivores buffer much of the variance in nutrient stoichiometry of their food, effects of growing conditions of the primary producers can travel up the food chain. We discuss the implications of these findings, and argue that phosphorus limitation of secondary consumers might be more common in marine than in freshwater systems.