by the turning of a key in the course of the pendulum swing. This method of giving the auditory stimulus as the pendulum was released was found unsatisfactory because of the irregularity of the magnetic release; at one time the pendulum would start immediately, at another time there would be a delay of as much as .1"

The reënforcement-inhibition curve plotted on the basis of the 1200 reactions in this series is presented in Fig. 8. Before stopping

9" 10" 11" 12" 13" 14" 15" 16" 17" 18" 19" 2.0" FIG. 8. Reënforcement-Inhibition curve for prolonged auditory stimulation, based upon amount of reaction. Frog No. 1.

to consider the important features of this curve we should note the results of certain more accurate experiments with prolonged auditory stimulation.

With two animals, No. 2, a female, and No. 3, a male, fifty pairs of reactions were taken for nine different intervals (see Table 5) of auditory stimulation. Each of the curves of Figures 9 and 10 is therefore based upon 900 reactions. The conditions for these experiments were the same as those for the momentary stimulation series, save that the electric bell took the place of the electrically actuated hammer, as the mechanism for auditory stimulation.

The important facts exhibited by the results of these prolonged auditory stimulation experiments in contrast with those with momentary auditory stimulation are: (1) That whereas for the momentary auditory stimulus of a hammer blow the reënforcement is greatest for simultaneous stimuli, in case of the prolonged stimulation with the electric bell, reënforcement increases during an interval of .25" of auditory stimulation. Hence, the two conditions of stimulation give

us different types of reënforcement-inhibition curve. For the momentary stimulus the maximum reënforcement appears at simultaneity, and for the prolonged stimulus at .25′′; (2) that the transition from reënforcement to inhibition occurs at 1.2" in the prolonged stimulation curves, while in the momentary stimulation curves it occurs at .35"; (3) that the maximum inhibition which appears in the curves under discussion at about 1.5′′ is less in comparison with

FIG. 9. Reënforcement-Inhibition curves for prolonged auditory stimulation, based upon amount of reaction. Female No. 2 Male No. 3. . .

the amount of reënforcement than that of the momentary stimulation curves; (4) that the auditory stimulus becomes ineffective when the interval during which it continues before tactual stimulation is 2.0". The curves of Figures 8, 9, and 10 are then representations of a neural process which passes through a positive and a negative phase in about 2". The effect of prolongation of the auditory stimulation interval

is to lengthen the period of reënforcement; the period of inhibition
shows little modification.

For the purpose of showing in greater detail the nature of the results
of this work the data from which the curves of Figures 9 and 10 were

constructed are presented in the accompanying Table 5.

FIG. 10. Reënforcement-Inhibition curves for prolonged
based upon number of reactions. Female No. 2

auditory stimulation,
Male No. 3.

Having now presented the results of my own investigation I wish to call attention to certain of their relationships to the work of other investigators, and to discuss briefly their significance.

TABLE 5. PROLONGED AUDITORY STIMULATION (ELECTRIC BELL)

Frog No. 2. Female. Weight usually 25 grams.

VI. DISCUSSION OF LITERATURE AND Results

The literature on reënforcement and inhibition is large, and even that portion of it which deals especially with the importance of the temporal relations of stimuli in connection with reënforcement and inhibition is so extensive that it does not seem worth while to attempt to give a systematic résumé of it for the purposes of this paper. I shall therefore call attention merely to those investigations which have contributed directly to the solution of the problems with which we are now concerned.

1

Bowditch and Warren 1 discovered that knee-jerk in the human subject is reënforced when an auditory, a visual, or a tactual stimulus precedes the tendon blow by .1" to .5", whereas the same stimuli have an inhibitory influence when they are given from .5" to 1.0" before the tendon blow.

At the suggestion of Bowditch, Cleghorn 2 undertook to investigate

1 Journal of Physiology, vol. 9, pp. 60, 61, 1890.

2 American Journal of Physiology, vol. 1, p. 336, 1898.

the influence of complication of stimuli upon voluntary movements. In this research graphic records taken in connection with an ergograph indicated (1) that "a sensory stimulus applied just as the muscle was beginning to contract (voluntarily) caused an increase in the height of the contraction, and (2) that the relaxation following a contraction with intercalated sensory stimulus is quicker and more complete than when no stimulus is given (p. 344). Cleghorn did not give special attention to the significance of the temporal relations of the stimuli which he employed, and his work was limited to the phenomenon of reënforcement of voluntary action by reason of the appearance, during the progress of his research, of an excellent paper on the interference of stimuli by Hofbauer.1

Hofbauer covered thoroughly the ground which Cleghorn had planned to work over. The ergographic method was employed also by Hofbauer in his very careful study of the interference of impulses in the central nervous system of man. It was noticed that while the subject was rhythmically contracting a certain group of muscles in response to some prearranged signal (e.g., the sound of a metronome) the report of a pistol caused the contraction which immediately followed it to be much greater than the average of the rhythmic series, while the next contraction was correspondingly less than the average. It thus appeared that the sudden sound caused, first, reënforcement of the voluntary movement, then, inhibition. The reënforcement is greatest, according to Hofbauer, when the voluntary movement occurs immediately after the pistol report. When the report precedes the metronome signal by .2" reënforcement is still marked, but thereafter it decreases rapidly in amount, until finally at .5" inhibition appears. When the interval between the two stimuli is 1.0" the first stimulus has practically no effect upon the voluntary movement in response to the second. (Hofbauer, p. 558.)

What Bowditch and Warren, not to mention other students of the subject, have described for reflex action in man, Hofbauer, Cleghorn, and others have shown to hold true also of voluntary movements. Unfortunately my own investigation was completed up to the point of the writing of this paper before I read Hofbauer's work, so I have not followed methods of dealing with my data which would make our results directly and easily comparable. But, whatever may be the relations of our results in detail, there can be no doubt that what he has demonstrated for man is true in its important aspect of the reënforcement-inhibition phenomena for the frog.

1 Arch. f. d. ges. Physiol., vol. 68, p. 546, 1897.