This work was designed to test the following hypothesis: in the presence of predation more species will coexist in the prey community than in the absence of predation. The hypothesis may also be stated in terms of the evenness with which individuals are distributed among species. This hypothesis was tested using the communities that occur within the water-filled pitchers of the northern pitcher plant, Sarracenia purpurea L. The predator in this system is the larval stage of the mosquito, Wyeomyia smithii (Coq.). The prey community consisted of ciliated and flagellated protozoans as well as rotifers. Four approaches were used to gain information about this predator prey system: (1) observations on the densities of the predator in the field, (2) observations on the correlation between predator density and protozoan community structure in the field, (3) field experiments in which both the predator density and initial protozoan community structure were controlled, and (4) observations on the development of protozoan community structure in the long term absence of predation. The results show that there is considerable variability of predator densities in both time and space and that the protozoan community responds to these differences. The field experiments show that for the pitcher plant system the hypothesis must be rejected: species numbers drop monotonically on the average as predator densities increase. On the other hand evenness defined as H/Hmax does increase as predator densities increase. The above patterns hold except during larval diapause during the fall. These results are discussed within a general framework of factors which appear to be important for determining how predators affect prey community structure: the intensity of predation and its timing in space and time; the importance of competition in the prey community; the pattern of predation; the degree of correlation between species population parameters; the importance of space as a limiting resource. On the basis of a consideration of these factors it is concluded that the pitcher plant system has very few characteristics that would cause the system to respond to predation with enhanced species numbers. The most important characteristic leading to decreased species numbers appears to be a low level of interaction that exists among the protozoan species. The increase in evenness in response to predation can best be explained by a shift in the factors that control species numbers from species interactions to rates of reproduction.