When testing vaccine-induced protection an effective and reliable challenge method is a basic requirement and we here present a comparative study on different challenge methods used for infection of rainbow trout Oncorhynchus mykiss with Aeromonas salmonicida, a bacterial pathogen eliciting furunculosis. Fish were vaccinated with three different adjuvanted trivalent vaccines containing formalin killed A. salmonicida, Vibrio anguillarum O1 and O2a. These were 1) the commercial vaccine Alpha Ject 3000, 2) an experimental vaccine with water in paraffin oil adjuvant, 3) an experimental vaccine with water in paraffin oil in water adjuvant. Fish were then exposed to A. salmonicida challenge using i.p. injection, cohabitation in freshwater, cohabitation in saltwater (15 ppt) or combined fresh/saltwater cohabitation. Cohabitation reflects a more natural infection mode and was shown to give better differentiation of vaccine types compared to i.p. injection of live bacteria. The latter infection mode is less successful probably due to the intra-abdominal inflammatory reactions (characterized in this study according to the Speilberg scale) induced by i.p. vaccination whereby injected live bacteria more effectively become inactivated at the site of injection. Compared to cohabitation in freshwater, cohabitation in saltwater was less efficient probably due to reduced survivability of A. salmonicida in saltwater, which was also experimentally verified in vitro.