Enhancing the complexity of the environments of captive animals is often referred to as environmental enrichment, and aims to have positive effects on the animals' well-being. Such enrichments may have consequences both for so-called 'normal' behaviour and for the pathophysiology of the animals in question. The effects of a lack of environmental complexity, including social isolation, on home cage behaviour and on pathophysiology in rats is considered in this review. Several preference tests on rats - choice tests and operant tests - indicate a preference for bedding, nesting material and social contact. Contradictory research results concerning the need for gnawing objects per se are more difficult to interpret, and it is argued that excessive gnawing may be indicative of primary frustration and hence reduced welfare. One disadvantage of providing environmental enrichment to laboratory animals is a possible increase in subject variability, resulting in the need to use a greater number of test animals. However, this increased variability seems to be inconsistent and is not very well documented. It is argued that in cases where the behavioural benefits of environmental enrichment justify the use of more animals, better welfare should be more highly valued than a reduction in the number of animals used.