TABLE 4. MOMENTARY AUDITORY STIMULUS, HAMMER BLOW

In these results two striking differences between the males and females appear: first, the reënforcement is not so great for the females as for the males; second, inhibition appears earlier and continues longer with the females than with the males. The average reënforcement with simultaneous stimuli is 82.5% for the males against 58.0% for the females. Inhibition begins to appear in case of the females when the interval between the stimuli is .25˝ to .35˝; in case of the males it appears between .35˝ and .45˝. Finally at .90˝ interval inhibition is slightly greater for the females.

Although the exact significance of these facts is unknown, it is not improbable that they are indicative of fundamentally important sex-differences in reaction to sound. The males among frogs are usually the vocalists, although in some species the females also croak. Moreover, in case of the green frog the tympanum of the male is much larger than that of the female. The results presented would seem to indicate that certain sounds stimulate the males to activity, whereas they inhibit activity in the females.

Graphically represented, the results of the momentary auditory stimulus experiments with frogs Nos. 1, 2, 3, and 4 are as follows:

The curves are all plotted by the method which will now be described in connection with Fig. 2. This figure presents the reënforcement-inhibition curves for the males No. 1 (solid line in the figure) and No. 3 (broken line). If in this figure we let the zero-point on the ordinates represent the value of the reaction to the tactual stimulus when given alone, then the value of the reaction to the auditory-tactual stimuli would be represented at some point above the zero-point if this reaction was greater than the tactual reaction (reënforcement), and below the zero-point if the reaction was less than the tactual (inhibition). Since one of our chosen measures of reënforcement and inhibition is the amount, in per cent of tactual reaction, by which the auditory-tactual reaction exceeds or falls short of the tactual reaction, such a curve of reënforcement-inhibition as that of Fig. 2 (solid line) can be constructed at once from the data given in column four of Table 3. Here the auditory stimulus, when simultaneous with the tactual, caused 62% reënforcement, as is indicated in the figure. The figures in the left-hand margin of the curves indicate amount of reënforcement or inhibition in per cent of tactual reaction; those at the bottom of the curves mark the intervals. On the curves dots indicate the intervals used in the experiments. Each of the curves is plotted on the basis of 700 reactions.

In every way comparable with the curves for the males No. 1 and No. 3 in Fig. 2 are those for the females No. 2 and No. 4 of Fig. 3. The similarity of the two curves in each figure is noteworthy. Inasmuch as the conditions of experimentation were the same for all the animals this would seem to indicate sex-differences which are worthy of further investigation. The curves show clearly the greater reënforcement in the males, and the greater inhibition in the females.

Figures 4 and 5 are the reënforcement-inhibition curves for the same series of experiments plotted on the basis of the number of reactions in excess of half that were reënforced or inhibited. As there were fifty pairs of reactions with each frog for each interval, uniform reënforcement would be represented by twenty-five reactions above the base-line; uniform inhibition by twenty-five reactions below the base-line. The number of reactions is indicated by the figures in the left margin; the intervals, by those below the base-line. As an illustration of the application of the method of plotting, the curve for male No. 1 (solid line) of Fig. 4 is constructed from the data of column five of Table 3. With simultaneous stimuli 17 reactions in excess of half, i. e., 17 + 25, or 42, were reënforced; at .35˝ interval .5 of a reaction was the average amount of reënforcement; at .45˝ interval 10 reactions in excess of half, i. e., 35, were inhibited, therefore the curve falls to 10 below the base-line.

Just as Figures 2 and 3 permit of direct comparison of the results of the measurement of the amount of reënforcement and inhibition for males and females, so Figures 4 and 5 make possible comparison in similar fashion of the number of reënforced and inhibited reactions for the sexes. It is to be noted that the two sets of curves, plotted on the bases of amount and number of reaction, agree in all important respects.

Figure 6 is the composite curve of amount of reënforcement-inhibition for the four animals; Figure 7 is the composite curve of the number of reactions reënforced and inhibited.

Summarily stated, the results of the experiments thus far described are: (1) The auditory stimulus of a quick hammer blow produces the maximum amount of reënforcement of tactual reaction when it is given simultaneously with the tactual stimulus; (2) as the interval between the auditory and the tactual stimulus approaches .35˝ the amount of reënforcement gradually decreases; (3) when given .35˝ before the tactual stimulus the auditory is practically without effect upon the tactual reaction; (4) as the interval increases above .3˝ inhibition begins to appear; (5) the inhibitory influence of the auditory stimulus is greatest when the interval is about .45˝; (6) when the interval is as long as .90˝ the auditory stimulus is again ineffective. It thus appears that the reënforcement-inhibition curve of this particular stimulus under the conditions described is representative of a neural process which completes itself, in passing through two phases, a positive phase (reënforcement) and a negative phase (inhibition), in about one second.

b. Prolonged auditory stimulation. The experiments previously described have proved that a momentary auditory stimulus, which when given alone never produces a visible motor reaction, either reënforces or inhibits the reaction to a tactual stimulus which it accompanies or precedes. The experiments now to be described were made for the purpose of ascertaining whether reënforcement and inhibition occur in the same way if the auditory stimulus is prolonged, instead of momentary.

In a trial series of experiments with frog No. 1, one hundred pairs of reactions were recorded for each of six intervals of auditory stimulation. The auditory stimulus was given by the ringing of an electric bell. For all intervals the ringing of the bell continued until the tactual stimulus was given. When the two stimuli were given simultaneously the auditory stimulus was necessarily momentary, as in the foregoing experiments, but for all other relationships of the stimuli the bell rang for a certain length of time before the tactual stimulus was given. The six relations of the stimuli were: (1) simultaneous, (2) bell .2˝ before and until tactual, (3) bell .6˝ before, (4) bell 1.05˝ before, (5) bell 1.5˝ before, and (6) bell 2.0˝ before. The other conditions of these experiments were the same as those previously described, except that the auditory stimulus was here given by the opening of the key which released the pendulum, instead of being given 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 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 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.

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)

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.

Bowditch and Warren[156] 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[157] undertook to investigate 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.[158]

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.

Important in their bearings upon the phenomena of reënforcement and inhibition which we are now considering, are the various studies of refractory period and rhythm of nerve cell and fibre. The existence of a refractory period in neural substance, similar to that demonstrated for certain kinds of muscle by Marey,[159] Englemann,[160] Kaiser,[161] Cushny and Matthews,[162] Woodworth,[163] and many others, has been proved by Broca and Richet.[164]

Broca and Richet found that in the normal dog the refractory period of the nerve substance is too short to be easily detectable, they therefore experimented with animals which were lightly chloralized and kept at a temperature of 30 to 34° (the mean normal temperature of the dog is about 39.5°). Under these conditions a dog, when two identical stimuli (quality and intensity the same) were applied to the cerebral cortex successively, exhibited the following reactions: (1) When the stimuli were separated by .01˝ they reënforced one another (addition); (2) when the interval was .1˝ they inhibited the reaction partially (subtraction).

Concerning this phenomenon Richet writes in his dictionary of physiology (p.5): "Marey showed, in 1890, that the heart of the frog, at certain moments of systole, was inexcitable. Now our experiments prove that the cerebral apparatus, a certain time after the excitation, also ceases to be excitable: it then has a refractory phase, and this refractory phase is much more prolonged than that of the cardiac muscle." In a later publication Richet[165] makes the somewhat startling statement that a refractory period is not exhibited by the nerves of cold-blooded animals. In the tortoise, according to his results, reënforcement occurs so long as the interval between the two stimuli is not greater than 2˝, while for longer intervals each stimulus to all appearances works independently. Richet seems to have generalized from a study of the tortoise. That his generalization is unwarranted seems to me highly probable in the light of the results of this paper, for there are many reasons for supposing that the reënforcement-inhibition phenomena with which we have been dealing in case of the frog are manifestations of the existence of the same process in the nervous system which under somewhat different conditions of experimentation exhibits itself in the so-called refractory period.

The researches of Richet and his students indicate that the time of the process which conditions the phenomena of reënforcement-inhibition is about .1˝. Stimuli given at .1˝ intervals do not interfere with one another. That the process underlying the refractory period and the reënforcement-inhibition phenomena of our experiments is a rhythmic double-phase process is made still more probable by the following results. Horsley and Schäfer[166] found that the rate of response of the monkey to cortical stimulation was 12 per second, and Schäfer[167] discovered that the maximum rate of volitional impulses in man is 10 to 12 per second.

It was shown by Exner that certain movements of the foot of a rabbit could be produced by stimulating either the cortex or the skin of the foot. Simultaneous stimulation of both regions gives reënforcement. Stimulation of the cortex, if given not more than 3˝ before subliminal stimulation of the skin, renders the latter effective. When both stimuli are subliminal each makes the subsequent one effective if the interval between them is not over 1/8˝ (Schäfer[168]). Similarly for the dog Exner[169] proved that cortical and cutaneous stimuli reënforced one another, when both were subliminal, if the interval between them was not greater than .6˝. Cortical and auditory stimuli, and auditory and cutaneous (of the skin of foot) gave similar results.

Physiologists have long been familiar with several aspects of the phenomena of reënforcement and inhibition in the frog, but I know of no detailed study of the significance of the temporal relations of stimuli in this connection. Goltz[170] called attention to the inhibition of the croaking reflex by peripheral stimulation, as well as to several similar phenomena. Nothnagel,[171] Lewisson,[172] and Wydensky[173] further contributed to our knowledge of the interference effects of stimuli in the frog. Wydensky proved that the application of an induced current to a nerve-muscle preparation may result in either contraction or relaxation of the muscle, according to the frequency of stimulation.

More recently Merzbacher[174] has dealt with the influences of complication of stimuli in the frog with the purpose of ascertaining the relations of the sense-organs to the reflex movements of the animal. His first paper is concerned especially with the functional importance of the eye in connection with reflexes. Unfortunately for the demands of this research, he did not attend particularly to the temporal relations of his stimuli. That a visual and a cutaneous stimulus were given either "at the same time or within a short interval of one another" (p. 250) is not the sort of information our problems demand.

According to Merzbacher's very interesting results a visual stimulus reënforces the reaction to a cutaneous stimulus. As the results of this paper show, this is only half a truth, for the two stimuli may either reënforce or inhibit one another's reactions. As Merzbacher observed no evidences of reaction to auditory stimulation he presumably did not attempt to study the influences of the ear in connection with reflexes.

There can be no doubt that the words reënforcement and inhibition as at present used in connection with the functions of the nervous system cover a multitude of widely differing phenomena. We can at once distinguish at least two important kinds of reënforcement or inhibition: first, that which is due to the functioning of special augmentary or inhibitory portions of the nervous system; second, that which is the result of the complication of stimuli. Any and every process in the nervous system may have either a reënforcing or an inhibiting influence upon simultaneous or succeeding processes; doubtless most processes or impulses at various times have both effects. The nervous system is constantly being modified by impulses from many sources, which suppress or strengthen one another according to their relative intensity, their temporal relations, and the motor relations of the portions of the organism which they affect.

The existence of the so-called refractory period in brain cortex and nerve indicates that every stimulus causes certain fundamentally important changes in the condition of the neural substance. These changes we may for convenience of illustration describe as modification of excitability, or of the functional capacity of central or peripheral tissues. Every stimulus causes a portion of the neural substance to pass from its normal state through a condition of increased excitability, which we may designate the positive phase, to a condition of diminished excitability, the negative phase. There is first an increase in the functional capacity of the tissues, then a decrease. If during the course of the change produced by a given stimulus a second stimulus becomes effective its result in reaction is determined by the particular phase of the tissues upon which it intrudes. If the nervous system is in the condition of increased excitability, and the two stimuli act upon sensory regions whose motor connections are not antagonistic, the reaction will be reënforced, as we say, by the previous stimulus; if, however, the second stimulus falls upon the negative phase of the nerve substance, the reaction will be partially or totally inhibited.

The facts which are most prominent as the result of this investigation are, first, that the temporal relation of stimuli is an important condition of certain forms of reënforcement and inhibition; second, that the interference effects of two stimuli cannot be studied to advantage without attention to the relations of the forms of reaction which are appropriate to each stimulus.

V. SUMMARY

1. Motor reactions of the green frog to electric stimuli are inhibited either partially or wholly by photic stimuli. The visual stimulus of a moving object has a like effect. It has been found, furthermore, that the same visual stimulus may either inhibit or reënforce the motor reaction in response to electric stimulation. When the two stimuli are given simultaneously reënforcement occurs, when the visual stimulus precedes the electric by half a second or more inhibition appears.

2. An auditory stimulus, which does not produce any visible reaction when given alone, modifies respiration and the reactions to other stimuli when given in connection with them.

3. The momentary auditory stimulus of a quick hammer blow when simultaneous with tactual stimulation reënforces the reaction to the latter stimulus. This reënforcement, or increase in the amount of reaction, ranges from 50 to 100% of the average reaction to the tactual stimulus alone. When the auditory stimulus is given before the tactual reënforcement occurs in gradually decreasing amount until the interval between the two stimuli reaches .35˝; at this point the auditory stimulus has no apparent effect upon the tactual reaction. As the interval is still further increased inhibition appears and continues for intervals between .35˝ and .9˝. Reënforcement is greatest when the two stimuli are simultaneous; inhibition is greatest when the momentary auditory stimulus precedes the tactual by .4˝ to .6˝. When the interval reaches .9˝ the first stimulus does not affect the reaction to the second.

4. Reënforcement is greater for the males than for the females; inhibition appears sooner and lasts longer in case of the females. This apparently indicates that the males are stimulated to activity by certain auditory stimuli, whereas the females are rendered passive by similar sounds.

5. Prolonged auditory stimulation by means of an electric bell causes reënforcement and inhibition, according to the temporal relations of the stimuli, as does momentary auditory stimulation, with the following differences: The maximum reënforcement occurs when the tactual stimulus is given about .25˝ after auditory stimulation has begun; reënforcement continues for a period of 1.2˝, i. e., when the electric bell continues to ring until the tactual stimulus is given, it reënforces the tactual reaction from simultaneity to 1.2˝. Inhibition then appears, and continues until 1.8˝. Both momentary and prolonged auditory stimulation cause first reënforcement, then inhibition of the appropriate reaction to a tactual stimulus.

6. The reënforcement-inhibition curves for the frog are very similar to those for man.

7. In case of the several pairs of stimuli whose interference effects have been studied reënforcement-inhibition appears. The first stimulus reënforces reaction to the second so long as the interval between them is not more than about .4˝, while it inhibits the reaction when the interval is longer. Whether this reënforcement-inhibition curve as given in the experiments described may similarly be obtained for any and every pair of stimuli, no matter what their relation to reactions, remains to be determined.

8. In connection with the study of the mutual relations of stimuli of which this paper gives an account certain facts concerning the sense of hearing have been discovered. A summary statement of the results on hearing may be found on page 551.

THE TEMPORAL RELATIONS OF NEURAL PROCESSES

Muscle contraction-time, according to the determinations of several investigators, varies about .0035˝.[175] Sanderson states that the time for direct stimulation of the muscle is approximately .0035˝ and for indirect stimulation, by means of the nerve, .007˝. The rate of nerve-transmission in the frog ranges from 25 to 35 metres per second.

Reflex reaction-time, as might be expected, varies widely with the nature of the reaction elicited by a stimulus, the condition of the animal, and the quality and strength of the stimulus. For many of the simple motor reactions of the frog it ranges between 20 and 60^σ.[176] Whether reflex reaction-time is to be sharply contrasted with instinctive and voluntary reaction-times, or whether they indistinguishably merge into one another is a question of considerable interest and importance for the student of the evolution of activity.

Voluntary reaction-time may be as short as 150^σ or as long as life, in an animal capable of profiting by experience as does the frog. It is preëminently the delayed type of reaction-time.

So much concerning the temporal relations of neural processes in the frog being well established, the purpose of the present paper is to call attention to some experimental results which indicate the existence of clearly defined types of reaction, and suggest possible values of reaction-time as a sign of mind.

The specific problems to be considered are: (1) Do reaction-times, in any given animal, range with equal frequency of occurrence from short to long, or are there certain modes (most frequented classes) which indicate definite types of reaction, such, for example, as the reflex, instinctive, etc.? (2) If there is distribution of the reaction-times about one or more modes, what are the types of reaction indicated thereby? (3) Finally, is reaction-time of service as a sign or measure of consciousness?

I wish especially to call attention to the fact that this paper deals with the reactions of the frog, not with animal reactions in general.

REACTIONS TO ELECTRICAL STIMULATION AND TYPES OF REACTION

Two years ago in connection with a discussion of the reaction-time of the green frog to electrical and tactual stimuli,[177] I presented a curve showing the distribution of 277 reaction-times to an electrical stimulus. The curve exhibited two clearly defined modes: one at between 60 and 70^σ and the other at about 160^σ. There was further a group of delayed reactions ranging about 500^σ. This form of distribution was interpreted, at the time, as indicative of three types of reaction, called, respectively, the reflex, the instinctive, and the delayed.

I have since obtained and examined with reference to form of distribution the further data which are presented in this paper. The reactions are all those of the green frog to electrical stimulation. The stimulus was applied by means of wires on the reaction-board on which the frog rested during the experiments. When reaction occurred in response to the electrical stimulus a circuit through the time-measuring apparatus was broken by the release of a delicate spring which had been held in place up to the instant of reaction by the weight of the frog. A Hipp chronoscope, controlled by a Cattell falling screen, served as a time-measuring mechanism. Three intensities of stimulus were used: (1) A current from one Mesco dry cell, (2) from two cells, and (3) from four cells.

Of the reactions whose time was measured there are three series. Series I is constituted by the recorded reaction-times in response to a one-cell stimulus, Series II, those in response to a two-cell stimulus, and Series III, those in response to a four-cell stimulus. The number of reactions, range and mode of each series are as follows:

The distribution of the 481 reaction-times of Series I and II is shown by Figure 1; that of the 256 reaction-times of Series III, by Figure 2. For both of these distribution polygons the reaction-times were arranged in ten^σ classes, beginning with the class 41-50^σ[178] in the case of the combined Series I and II and with the class 61-70^σ in the case of Series III.

Series I exhibits a primary mode at 235^σ. There are no reflex reactions in this series, unless it be maintained that the reflex reactions of the frog may have a reaction-time of over 160^σ, but there are a number of delayed reactions, some of which have reaction-times as long as 798^σ. This intensity of stimulation (one cell) may be said to call forth prompt reactions, which we may provisionally call instinctive, and delayed reactions, which have all the appearances of voluntary acts. There are no reactions which come within the range commonly considered as the reflex range of the frog (20-60^σ), and there are relatively few delayed reactions: almost all centre about the mode 235^σ.

Series II, in contrast with Series I, exhibits a secondary mode at 65^σ in addition to the primary mode at 235^σ. The stimulus-intensity of this series (roughly twice as great as that for Series I) induces a variety of short reaction, which did not appear in the case of the one-cell stimulus, and at the same time fewer delayed reactions. The range of the reaction-times for the two series is about the same, but the lower limits are markedly different.

Observation of the subjects during the experiments revealed two methods of reaction to the two-cell stimulus: a locomotor reaction (jump) which at once removed the animal from the source of stimulation, and a twitch of the hind legs which was instantly followed by the above-mentioned locomotor reaction. The leg reactions constitute the reflex group of Fig. 1, the usual prompt locomotor reactions, the instinctive group, and the slow locomotor reactions, the delayed or voluntary group.

It is to be noted that the instinctive reaction-time mode is the same for the two intensities of stimulation. This apparently indicates that change in intensity of stimulation causes a change in the type of reaction, not merely a gradual change in the position of the mode. For example, the modal reaction-time of 235^σ given by a one-cell stimulus did not shift to 200^σ or lower, as might have been expected, but instead there appeared a new type of reaction. The average reaction-times for the two series indicate a decrease in time with increase in intensity of stimulation, but they give no indication of the really important difference in the two series of reactions. The great importance of the distribution of the data, in addition to the common statistical quantities, is manifest.

Series III, whose reactions occurred in response to a very strong stimulus, differs in several important respects from the other series. Its range is much narrower, only 117^σ. Delayed reactions are lacking, and so also, curiously enough, are the reflex reactions of Series II. Instead of either or both of the modes of Series II, there appears in Series III an intermediate mode at 105^σ.

Our interpretation of these facts is facilitated by results of observation of the reacting subject. The leg reflex which frequently occurred in response to the two-cell stimulus never appeared in response to the four-cell stimulus. This in part explains the lack of the short reaction-time mode of Series II; it does not, however, account for the lack of delayed reactions. The latter fact may be referred to the intensity of the stimulus. Another difficulty in interpretation appears in connection with the intermediate mode, 105^σ. Is this to be considered an instinctive mode, as were those at 235^σ, or a reflex mode? Where is the line between reflex and instinctive action to be drawn? These results very clearly indicate that no line can be drawn, except quite arbitrarily. Reflex reaction-time, in the case of the frog, is continuous with instinctive, yet for any given situation the reflex, instinctive, and delayed (voluntary?) modes are likely to appear, as, for example, in the case of the data of this paper. Our conclusion must be, therefore, that although types of reaction are indicated by reaction-time results, the mode for a given type varies too much in position with different conditions to make it possible to say that a particular reaction-time is that of a certain type.

We may safely say, then, that for any given subject, the muscle contraction-time, nerve transmission-time, and simple sensory reaction-time to the constant stimulus in question being known, we should be able safely to interpret reaction-time records in terms of reaction types. For reflex, instinctive, and voluntary are terms which designate modes of reaction, albeit not isolated classes, for they intergrade.

Whether there are more types of reaction than are indicated by the data of this report does not concern us at present, for the practical as well as the theoretical bearings of our conclusions depend upon the existence of types, and not upon their number.

REACTION-TIME AS AN INDICATION OF CONSCIOUSNESS

Hesitation in reaction is commonly accepted as an important sign of volitional consciousness in man; consequently delayed reactions in lower animals are supposed to be indicative of psychic processes. Granting this much, reaction-time may be used as a sign of consciousness. It cannot be denied that the longer the reaction-time of a given animal the greater the probability that the reaction is conditioned by mental processes. Such a statement, it is true, has a basis neither better nor worse than that of most of our inferences concerning the nature of the actions of our fellow beings. As I have already attempted to show in a discussion of criteria of consciousness in animal psychology,[179] there is no one criterion of consciousness which can be used alone satisfactorily, but instead there are numerous signs of mind each of which has value according to the number and variety of our observations concerning its occurrence in connection with states of consciousness. The more of such signs we discover and learn to evaluate properly in relation to consciousness in its different grades and to one another, the safer will be our inferences concerning the existence of mental processes in animals.

Reaction-time is presented in this paper as an additional sign of mind. Like all other signs it is of value only if used as one of a series of indications of mental life. For if we attempt to judge of consciousness by reference to reaction-time alone, we may be seriously misled, whereas if we use it in connection with docility, variability, neural specialization, and other recognizedly valuable signs, we may be greatly aided in our inference. As in juristic procedure judgment is not based upon one bit of evidence nor even upon the evidence of a single witness, but upon evidence accumulated from all available sources, so in our attempts to judge of the existence of consciousness, it matters not whether the being be human or infra-human, we should make use of all phenomena which are recognized as signs of mind. The chief task of comparative psychology at present is the discovery and evaluation of signs of mind.

Reaction-time data, however, furnish another sign, or, as I prefer to call it in this case, measure of the intensity of consciousness; for variability of the time of reaction as well as its duration is significant. Reflex reaction-time is relatively constant, instinctive varies considerably, and the variability of voluntary reaction-time is extremely large. Degree of variability of reaction-time may be used as an indication of consciousness in the same way that variability in the form of reaction is used. The higher the power of consciousness the greater the variety in form of reaction and the variability of the reaction-time.

Reaction-time studies, as well as introspection and the investigation of animal behavior, indicate the importance of three activity concepts: automatism, instinct, and will. The automatic act is quick and relatively constant in form as well as reaction-time, while all signs lead us to infer that consciousness, when it accompanies the act, is a sequent phenomenon and not a condition of the act. The instinctive act is both slower and more variable in form and time than the automatic: consciousness is indicated as an accompaniment, and apparently it is at times a condition of the act. The will-act is extremely variable, unique in form, and almost without limits of reaction-time, for the conscious organism may react to the present situation in a fifth of a second, a day, or a year. Will is experience in action: it is our name for individually acquired control, and voluntary action is above all consciously conditioned activity.

Reaction-time, with respect to its two aspects of duration and variability, may be used as a sign or criterion of consciousness, for in accordance with the nature of these two sets of facts we classify acts as reflex, instinctive, or voluntary.

THE MENTAL LIFE OF THE DOMESTIC PIGEON

The brain of the pigeon is well developed, although the hemispheres are unconvoluted. When they are removed, the animal retains unaltered its reflex and vital activities, but ceases for a time at least to show evidence of mental life, for example, memory and will.[180] In the normal animal sight and hearing are acute, and touch seems keen, although the claws are not used for grasping and eating, as in the case of more intelligent birds, especially, parrots. There is considerable sensitiveness to temperature changes. Taste, and probably smell, appear to be deficient.[181] The "sense of support" is marked, even in the young.[182]

Since the pigeon seems to dream and also to miss its absent mate, some observers believe that imagery is present. There is certainly local memory, and also capacity to observe. Various intelligent acts have been reported.[183] The remarkable homing habits of the carrier pigeon have received no satisfactory explanation. While Cyon[184] suggests the stimulation of the nasal organs by air currents, Thauzièr[185] holds to the electrical theory; they agree, however, that certain higher psychical processes are probably involved.

Graber's[186] tests indicate that pigeons have no color-preference. Beebe's[187] statement concerning birds in general is peculiarly true of pigeons: "There are few species which do not show the emotions of love and sympathy, and ... one will sometimes pine and die of grief at the loss of its mate." After referring to their patient care of the young, he adds: "Indeed, sympathy is the keynote in the development of the higher mental faculties." These birds communicate, but their language consists of comparatively few sounds. As in many other birds, the play-instinct is highly developed.

II. PROBLEMS AND METHODS

My study of the pigeon's emotional life had for its object certain respiratory "expressions." These were investigated by means of a pneumographic tracing, secured while the animal was comfortably fastened in a shallow nest, partially open below. A small box was placed over the bird, and apparatus was so arranged that the time of giving various stimuli was recorded automatically on the smoked paper of the kymograph drum, below the breathing-curve. A third line indicated rate of drum movement. Although some interesting results were obtained, the chief significance of the research consists in its demonstration of the fact that this method of studying animal mind is valuable.

In the study of association I sought to determine the sense-data which the process involves, its method of formation (with due regard to social conditions), its rapidity, permanence, and modifiability, and also its probable degree of complexity. Material contributing to the subject was secured by observing the behavior of the animal when seeking to obtain food by overcoming such obstacles as labyrinths with wire passages, and latches, when the food was left in view, or by finding it when out of sight. In the latter case it was placed in a box occupying a customary place in a group of exactly similar boxes, or else in a box of color or form unlike the other members of the group and variously arranged, from time to time, with respect to them. When the animals were learning the labyrinth habits, various stimulations were given them; later the character of some of these was altered, and the resulting changes in behavior were noted. After the habits had been thoroughly learned, the birds were given a rest for some weeks, and then tested again under the old conditions. A few trials were arranged with special reference to the study of imitation; the animals here were tested as to their ability to execute simple but unfamiliar acts, after having only seen them performed by an animal previously trained. Throughout the associational tests the animals very seldom received food in their cages; but as they were tested daily and allowed to satisfy their hunger completely at the last test, they were never in a state of "utter hunger"—a condition which most experimenters think best to avoid. A series of tests, given at the conclusion of the investigation, indicated that the odor of the food had not assisted the animals in reaching it.

In the two series of experiments (emotion and association) thirty-five animals in all were used. They were confined in large cages in a fairly well lighted and ventilated room, and were fed wheat, cracked corn, and occasionally fruit, and kept well supplied with fresh water and sand. They generally remained in a healthy condition throughout the tests, especially during the winter. To exclude, as far as possible, the disturbing influence of fear, they were usually handled only after the room had been darkened. As the noise made by the curtains was objectionable, the birds were tested with the room illuminated by incandescent lamps; the light was turned off before the birds were placed in position for the trials, and again before they were removed from the apparatus to the cages. It is generally agreed that an experimenter should be out of sight when giving a test. I am convinced that it is important to avoid being seen by the animals at any time. This involves great inconvenience, especially when one employs the pneumographic method, but better results are thus obtained.

For practical suggestions as to apparatus and methods I am greatly indebted to Dr. Robert MacDougall, at the beginning of my investigation, and to Dr. Robert M. Yerkes, throughout. I also owe much to the researches of Zoneff and Meumann,[188] Thorndike,[189] Mills,[190] Small,[191] and Kinnaman.[192] Porter's[193] interesting study of sparrows was made almost simultaneously with the investigation here reported. Fewer animals were used by him, but in some instances more tests were given.

III. INVESTIGATION OF EMOTION[194]

1. Respiration in general. The normal breathing-curve in pigeons is quite similar in contour to that of the human subject, although the rhythm is more rapid and the pauses are less pronounced. When acoustical, visual, olfactory, or tactual stimuli are given, various modifications appear, for example, quickening, deepening, and minor irregularities. It was noticed that meaningless stimuli (pistol-shots) quickly lose their disturbing influence, whereas the breathing remains sensitive to those of a significant character, such as the noises made by other birds. It was also found that a stimulus which no longer affects the breathing will sometimes occasion disturbance if accompanied by a second stimulus of another order, although of a weak intensity (summation).

2. Respiratory reactions to light. As the easy control of conditions makes vision an excellent field in which to work, light reactions were investigated in detail. Two distinct series of tests were given. One sought to determine the relation between quality of light and reaction; the other, between intensity of light and reaction. Four colors of one intensity and three intensities of one color, respectively, were used. In the first series four stimuli, one for each of the colors, red, yellow, green, and blue, were given daily; in the second series five daily stimuli were given, of the same intensity for any one day, and one minute apart; this made it possible to observe also the effect of repetition. Each stimulus was given at the beginning of a respiration and continued two seconds. When the tracings were studied, various modifications were noted, but special attention was paid to alterations in rate of breathing. In the case of both sets of trials an immediate quickening usually occurred after stimulation, and occasionally shallowing[195] and minor irregularities of contour.

In the first set of tests ten animals were used for twenty-five days. Average results indicated that red and yellow are less stimulating than green and blue. To secure data that would assist in the interpretation of these results, an investigation was made of the animals' color-preference. This was done by recording, at thirty-minute intervals, the position of the birds when confined, singly, in a box one half of which was illuminated (from the side) by light of one color, and one half by light of different color but of the same intensity. A water screen excluded the heat rays. After nine records had been taken the colored glasses were interchanged, and the animal's position relatively to the two colors was observed as before. This was repeated with the other colors until each of the four had been used with each of the other three. There were far more choices of green and blue than of red and yellow, though none of the colors was avoided. It seemed a question of degree of liking, rather than of liking or disliking. As stated, Graber's experiments indicated that pigeons have no color-preference, but his results are probably untrustworthy, since he tested several animals at once and apparently was not careful to change the colored glasses regularly. Putting together our two sets of data (the latter stated first) we have the following comparison: