A number of neural substrates have been proposed to mediate complex learning and memory processes in mammalian organisms. One strategy for testing the involvement of a particular gene in learning and memory is to create a mouse line with a null mutation in that gene. Recently, embryonic stem cell-based gene-targeted homologous recombination techniques have been employed to create a number of such mutant mouse lines that do not express interesting candidate genes. These animals have been examined for impairments in several complex learning paradigms which are known to depend on the integrity of the hippocampus. In this review several complex learning and memory paradigms are described, the techniques to create null mutants are reviewed, and the results of recent studies with null mutants are described. Finally, the limitations for interpretation of behavioral data using null mutants are discussed.