In this connection the important fact is that the various typical fatigue effects exhibited in living substances are exactly reproduced in metals, where there can be question neither of fatigue-product producing fatigue effects, nor of those constructive processes by which they might be removed. We have seen, both in muscles and in plants, that if sufficient time for complete recovery be allowed between each pair of stimuli, the heights of successive responses are the same, and there is no apparent fatigue (see page 39). But the height of response diminishes as the excitation interval is shortened. We find the same thing in metals. Below is given a record taken with tin (fig. 72). Throughout the experiment the amplitude of vibration was maintained constant, but in (_a_) the interval between consecutive stimuli was 1', while in (_b_) this was reduced to 30". A diminution of height immediately occurs. On restoring the original rhythm as in (_c_), the responses revert to their first large value. Thus we see that when the wire has not completely recovered, its responses, owing to residual strain, undergo diminution. Height of response is thus decreased by incomplete recovery. If then sufficient time be not allowed for perfect recovery, we can understand how, under certain circumstances, the residual strain would progressively increase with repetition of stimulus, and thus there would be a progressive diminution of height of response or fatigue. Again, we saw in the last chapter that increase of strain necessitates a longer period of recovery. Thus the longer a wire is stimulated, the more and more overstrained it becomes, and it therefore requires a gradual prolongation of the interval between the successive stimuli, if recovery is to be complete. This interval, however, being maintained constant, the recovery periods virtually undergo a gradual reduction, and successive recoveries become more and more incomplete. These considerations may be found to afford an insight into the progressive diminution of response in fatigued substances. [Illustration: FIG. 72.--DIMINUTION OF RESPONSE DUE TO SHORTENING THE PERIOD OF RECOVERY The stimulus is maintained constant. In (_a_) the interval between two successive stimuli is one minute, in (_b_) it is half a minute, and in (_c_) it is again one minute. The response in (_b_) is feebler than in either (_a_) or (_c_).] #Fatigue under continuous stimulation.#--Fatigue is perhaps best shown under continuous stimulation. For example, in muscles, when fresh and not fatigued, the top of the tetanic curve is horizontal, or may even be ascending, but with long-continued stimulation the curve declines. The rapidity of this decline depends on the nature of the muscle and its previous condition.