Pain can either signal the threat of tissue-damage (nociception) or can result from tissue damage itself. The physiology and “pathology” of pain, in the second instance, suggest that it can be diagnosed (not “measured”, “assessed”, etc.) on the basis of its association, but not equality, with tissue damage and by its coincidence with changes in behaviour. Pain will be present as part of a syndrome and cannot occur without cause or association. Pain may occur in experiments: a) coincidentally and unrelated to any experimental procedure, b) accidentally, when a procedure goes amiss, and c) as part of the experimental design. Where pain is intrinsic to the aims of an experiment (for example in the testing of analgesics), “escape routes” and “limits” must be formulated for the benefit of experimental subjects. In addition, the empirical value of the experiment must be assessed. There is insufficient information on the use and efficacy of analgesic drugs in animals. These drugs, however, are neither the only nor the most effective means of alleviating physical pain in animals. Other forms of treatment are considered. All depend on sound diagnosis and prognosis. Decisive action against the cause of pain is necessary.