Abstract. Stereotypies are repetitive, invariant behaviour patterns with no obvious goal or function. They seem to be restricted to captive animals, mentally ill or handicapped humans, and subjects given stimulant drugs. In this respect they are abnormal, although possibly the product of normal behavioural processes. Stereotypies are often associated with past or present sub-optimal aspects of the environment, and have been used as a welfare indicator. It has been hypothesized that stereotypies have beneficial consequences which reinforce their performance, although other means, such as positive feedback, may equally explain their persistence. Empirical evidence links them with lowered awareness of external events, and reduced arousal and distress. However, as most of this evidence is correlational it remains uncertain that the stereotypies are themselves the cause of coping. Furthermore, they are heterogeneous in source of origin, proximate causation and physical characteristics, and they change over time in important respects, becoming more readily elicited by a wider range of circumstances. Therefore the properties of one stereotypy are not necessarily those of another.