Feelings are states of the self, and incorporate moods and sensations. Although a person may appreciate precisely his state on a selected dimension, words may fail to describe the exactness of the subjective experience. The paucity of suitable quantitative terms in common speech limits the amount of information whichcan be transferred. Continuous phenomena have to begraded in artificial categories. A digital system is imposed on the observer, when the freedom ofan analogue system would be welcome. An understanding of many problems in clinical research presupposes that it is possible to communicate the desired information from patient to clinician in a wayamenable to measurement. A working party of the British Association defined measurement as' the assignment of numerals to things so as to represent facts and conventions about them'(Stevens 1946). For the measurement of feelings, communication based on a simple visual analogue seems appropriate. Lines, with their boundaries clearly defined as the extremes of the feeling, serve well for marking (Hayes & Paterson 1921).

The limitations of an analogue system are no more than those true of words (Fig 1); even speech contains an assumption that the same language is being spoken in order to communicate information, though this assumption may be far from true. The same word can be used with different meanings, and need not imply that people experi-ence the same feeling. The same amount ofchange may take place, but in only some people will this alter a category term, and then only if the change is from a certain initial value. With verbal scales,