We determined the diet preference of roach (Rutilus rutilus), perch (Perca fluviatilis), and bream (Abramis brama) before and after a fish removal program in Lake Finjasjön, Sweden. The biomass of roach and bream was reduced to 33 and 10%, respectively, after the program. The predominate diet change following this major reduction in fish biomass was an increased use of benthic organisms by most size-classes of roach. Also, bream shifted to benthos at an earlier ontogenetic stage. These diet shifts were most probably attributed to the drastic reduction in biomass of the benthivorous bream, resulting in underexploited benthic invertebrates. In order to test if alterations in diet choice were reflected in the composition of stable isotopes of consumer tissue, we determined the temporal changes in the 15N/14N ratio (δN) of potential fish prey as well as in fish tissue. No temporal trends were found for δN of fish, possibly due to the high temporal variability in δN of zooplankton. However, minimum and maximum δN values of major food organisms (chironomids and zooplankton) were generally reflected in the δN of both small roach and perch one sampling occasion later (3 months) and in piscivorous perch (exclusively feeding on small fish) two sampling occasions later. Hence, the stable isotope composition could be followed through food links, providing that the consumer mainly fed on one specific food item. However, several other predictions regarding connections between diet and stable isotope composition were not corroborated. We conclude that stable isotope analysis of consumer tissue cannot replace traditional methods of diet determination, but might well provide complementary data.