It was pointed out in the Introduction that there is no agreed definition of "exercise tolerance ". Few, indeed, have attempted to define it, yet it is widely used in a loose way in clinical discussions. A priori, it might have a definite meaning in either of two senses. First, it could be used to mean the maximum performance in some particular type of exercise of which the subject is capable. One authority at least believes that a maximum performance is the best measure of working capacity (Taylor, 1944). We have seen, however, that it is unsuitable in clinical work because motivation, so important in maximum effort, cannot be relied upon and because maximum effort is not without danger in some types of cardiopulmonary disease. Alternatively, one might use the term "exercise tolerance" to mean the level of exercise at which the smooth integra- tion of the various responses to exercise begins to break down; or, as `?ahlund (1948) put it, the level at which the steady state can no longer be efficiently maintained. The clinical usefulness of this definition depends upon whether this breakdown (if it exists as a clear -cut point) begins to show itself at a level of exercise low enough to be safely determined in patients. Wahlund (1948) placed the limit of efficiency (quite arbitrarily) at 170 heart beats per minute and 30 respira- tions per minute. Henderson et al. (1927) stated that 180 beats per minute is the ort9;nary limit of profitable tachycardia, because stroke output declines beyond this rate. Their conclusions were based upon determinations of heart output by the ethyl iodide method, the accuracy of which it is permissible to question. The validity of these figures is thus open to doubt; and in any event they represent a level of exertion to which a sick person could not reason- ably be taken. Concerning ventilatory efficiency, no cor- responding standards have been laid down, as far as I am aware; but the oxygen extraction or CO2 discharge might possibly be used to define the point of declining efficiency. This investigation has shown, however, that the changes in neither heart rate nor ventilation, in relation to the rate of work, reveal anything in the nature of a limiting point in the majority of healthy subjects, up to moderately heavy exercise. The response of a subject to progressively increasing exercise forms a more or less continuous grada- tion from the very mildest work to the most severe. Nor does it appear possible to predict with any accuracy a per- son's maximum capacity from measurements made during sub - maximal effort. And although the patients studied were too few and too disabled to allow a firm statement to be made, it would be surprising if these conclusions did not apply also to states of disease. It may well be that in the last resort "exercise tolerance" is not measurable, in the senses defined above. On the other hand, it is possible to obtain very useful information about a given subject's response to exercise by comparing his response to a particular test-exercise with the response expected of a normal person of the same age and sex. This approach requires that a normal range of values be established for each variable over a big range of exercise intensities. The individual under test might then be exercised at perhaps three set intensities, accord- ing to his ability, and his ventilatory and other responses compared with the normal at the appropriate values of oxygen consumption. Ventilation, alveolar and expired air composition, respiratory frequency, oxygen debt and, perhaps, heart rate, together with their derivatives, could then be expressed as percentages of the expected values. Limitation of the number of exercises done would make duplicate or triplicate determinations feasible. Perhaps the chief contribution of the present study has been to show that all these variables may be measured simultaneously, and without much difficulty, during the most natural of all forms of exercise.