One of the tasks of the international Organisation for Economic Co-operation and Development (OECD) is to produce standardized guidelines for harmonizing national regulations. The Paris-based OECD's environmental directorate is currently preparing a draft guidance document on the recognition, assessment and use of clinical signs as humane endpoints for experimental animals used in safety evaluation (OECD Series on Testing and Assessment, N.19, ). Experts and interested people are invited to submit comments through national coordinators.

Comments, including ours, have been delivered to OECD in the past few weeks. But we are concerned that some important scientific and ethical issues may not be receiving the attention they deserve. On other occasions we have pointed out some shortcomings in guidelines for the care of laboratory animals. For example, the recent US National Research Council guide is more concerned with dogs than with rodents, possibly because of the influence of veterinarians, who are trained to care for pet animals, in preparing the guidelines1. Yet rodents are by far the most commonly used experimental species of mammal.

The current OECD document risks overlooking some behavioural features of laboratory animals that could help in improving their general level of welfare. To take one example, behavioural observations aid in judging the vitality of the newborn rodent pup. The degree of spatial dispersion at the nest site provides an assessment that is otherwise difficult in newborns before their eyes are open, but under current guidelines is vital to the declaration of the ‘moribund’ state which compulsorily precedes euthanasia. In our experience, the ultrasonic calling pattern is also highly informative.

Furthermore, refined ethological techniques (some devised in our laboratory) provide improved, lower-suffering methods for measuring the reaction of rodents toward painful stimuli2. These experimental protocols, protecting both data quality and psychophysical welfare, received substantial approval at the third international congress on alternatives in animal use, held in Bologna last summer3.

The care and maintenance of laboratory animals has to take into account the evolutionary history of the species under study. This is particularly true when the subjects are laboratory-maintained male mice or female rats, living all their lives in unnatural social settings which fairly frequently result in fighting4. The need to consider behavioural and species-specific characteristics is also relevant when primates are used as experimental animals. The European Federation of Primatology ( ), through its primate expert group, is currently preparing a series of recommendations on minimal cage-sizes and specific environmental enrichment techniques, to be proposed to the Council of Europe.

There is still much work to do, and we urgently need more reliable behavioural data. This is particularly true for animals used for neuroscience research, where there is possibly most potential for stress and overt suffering. The ethologist's perspective will also be of great use given the explosion in use of transgenic mouse strains, some of which have unexpectedly been found to be hyperalgesic or to have abnormal levels of fear and inter-male aggressive behaviour.