Transgenic and knockout mice are increasingly important tools in the study of human genes controlling neural function. Spontaneous and induced mouse mutations are also utilised in the characterisation of novel genes of biological importance. Murine models of human neurological and psychiatric disease require reproducible protocols for the assessment of structural and functional phenotype. We have developed the SHIRPA protocol (Mammalian Genome 8 (1997) 711-713), a three-stage, comprehensive behavioural testing procedure applicable to phenotypic analysis of transgenics and other murine mutants. Behavioural, neurochemical and histopathological procedures are included to systematically assess mouse phenotype and ethological details of the test battery are presented. All tests are conducted in a specific order, with the procedures most sensitive to physical manipulation being completed first. 1. Elevated plus-maze test of anxiety; 2. Primary behavioural observation screen to assess gross phenotypic abnormality; 3. Rotarod assessment of sensorimotor deficits; 4. Open-field test of altered emotionality; 5. Spontaneous locomotor activity test; 6. Hole-board test of exploratory behaviour; 7. Primary observation screen in response to simple challenges; 8. Startle response; 9. Water maze spatial learning procedure. The battery has been used to characterise differences in the behavioural phenotype of inbred strains (Rogers et al, Behav, Brain Res., in press) and has successfully identified subtle phenotypic differences in a number of gene-targeted mice studies, including ApoE knockout (Rogers et al, BPS, Chester, April 1998).