Current advances in the ability to assay adult aquatic vertebrate behaviour are potentially very useful to aquatic toxicologists wishing to characterise the effects of pollutants on behaviour, cognition or neuro development. This review considers two specific challenges faced by researchers wishing to exploitthese technologies: maximising reliability and validity. It will suggest two behavioural procedures, withthe potential for automation and high-throughput implementation, which can be used to measure socialcohesion and anxiety, two areas of interest in behavioural aquatic toxicology. In addition, the reviewwill make recommendations about how these procedures (and others) could be carried out to maximise reliability and validity.