Vaccination is the most effective strategy to control infectious diseases in species with adaptive immunity. In human and in mouse, vaccination typically induces specific memory cells, which can mediate a fast anamnestic response upon infection by the targeted pathogen. In these species, successful vaccination induces a long-lasting protection, long after the titres of specific antibodies and the frequency of specific T cells have returned to steady state. Vaccination is also an important challenge in aquaculture, since alternative treatments are either too costly, or, in the case of antibiotics, are harmful for the environment or may result in dangerous resistances. However, the mechanisms of the long-term protection elicited by vaccines in fish remain poorly understood. Although fish possess typical B- and T-cells expressing diverse repertoires of immunoglobulins and T-cell receptors, many features of antigen specific responses are different from what is known in mouse and in human. Memory is one of the most elusive properties of fish adaptive immunity, and its basis is widely unknown. In this opinion article, we discuss the concept of immune memory in the context of the fish immunity. We illustrate the complexity of this question by discussing the results of experiments showing that protection can be passed through adoptive transfer of leukocytes from vaccinated donor fish to naive histocompatible recipients. Combined with tools developed in Targetfish and in previous projects, such as monoclonal antibodies against Band T-cell markers, we propose that such models of protection transfer provide excellent systems to dissect the mechanisms of B- and T-cell memory in the future.