Two immunization trials were performed in rainbow trout, using a formalin killed Aeromonas salmonicida suspension. In the first experiment, 200 g fish were administered orally or intraperitoneally. Circulating antibodies were elicited in injected animals only, but no protection was found when they were challenged with a highly virulent strain. Rough and smooth bacteria have provided similar agglutinins titers. In the second study, fingerlings were vaccinated by injection, but experimental disease was reproduced through IP or IM inoculations. Protection was effective in the second case. In the IP injected group the mortality was the same in vaccinated fish and controls, and carriers were detected among survivors 3 months later. The surprising fact is that serum agglutinins did not always provide a protection. This last one depends on different factors such as pathogenicity of strains and route of penetration of the challenge bacteria. So, experimental challenge has to be improved.