SOURCE: "A Criticism of Pavlov's Concept of Internal Inhibition," in the Psychological Review, Vol. 44, No. 4, July, 1937, pp. 297-312.

[In the following essay, Wenger points out a flaw in Pavlov 's theory of conditioned response concerning the notion of "internal inhibition. "]

The concept of the conditioned response, with or without Pavlovian terminology, has become an important consideration in any contemporary theory of the learning process.¹² The major phenomena discovered in the laboratories of Pavlov (18) and Bekhterev (2) have been verified by many other workers. However, Pavlov's interpretations of some of these phenomena have not met with general acceptance. The specific concept bearing the brunt of attack has been that of 'internal inhibition.' About it Razran (19), Guthrie (11), Beritoff (3), Wendt (24), Winsor (28), Chappell (5), Lashley (17), and others have had something derogatory to say.

The term 'inhibition' has found general use in two different contexts. First, it has been used in a descriptive sense to designate a condition of an organism characterized by various degrees of response decrement. Decrement during a competing reaction or Pavlov's external inhibition illustrates this behavioral use of the term. A second connotation of 'inhibition' involves an assumed process or substance affecting the neurones with a resulting decrement in response. In this context, the term is not only descriptive, but explanatory, and it is here that Pavlovian 'internal inhibition' must be classified.

Pavlov regards it as some neural condition which at times irradiates over the entire cortex, and which at other times becomes concentrated in one limited cortical area. It holds an important place in his theories for he uses it to explain experimental extinction of a conditioned response, and to account for negative, differential, trace, and delayed conditioned reflexes. Furthermore, the phenomenon of disinhibition, or the reinstatement of an extinguished or partially extinguished response, is supposed to occur through a temporary removal of internal inhibition as a result of some distracting stimulation. Of still further interest is Pavlov's belief that sleep is the result of a general irradiation of internal inhibition.

The present paper is a criticism of this concept. It will attempt to show (1) that the behavior which probably led Pavlov to conceive of a gradually irradiating cortical inhibition may be explained more parsimoniously, and (2) that such behavior is not necessarily a phenomenon of learning but instead is an artifact due to certain experimental conditions.

The first suggestion of the tenability of such a point of view presented itself during an investigation of external inhibition and disinhibition of a conditioned galvanic skin response in human adults (25). The subjects reclined on a cot in a relatively dark environment, and a masking tone was applied through a set of head phones. The situation, therefore, was conducive to sleep. In the report of the study the following statement occurs:

The subjects of this experiment reported somnolence during the extinction series. Although they had been instructed to remain as quiet as possible, their efforts to remain awake resulted in many slight movements and adjustments in posture and a tendency toward sporadic deep inspiration. Almost invariably, an increment in the conditioned responses followed. The proprioceptive sensations arising from these movements and inspirations are believed to have resulted in autonomous disinhibition, and their frequent occurrence to have been directly responsible for the prolonged extinction series. This interpretation suggests that at least part of what is usually termed extinctive inhibition may be explained in terms of a decreased level of tonicity.³

Observations made during a conditioning experiment with newborn human infants in the Iowa Infant Laboratory (26) have redirected attention to the problem. In work with infants the same general picture of sporadic activity described in the foregoing paragraph may be seen. Frequent periods of sleep alternate with irregular breathing and states of activity. Thus there existed conditions of tactual and proprioceptive stimulation comparable to those referred to in the above experiment.

In the first ten days of life, which is the time infants were available, it was found that conditioning did not reach 100 per cent response. Thus there was not opportunity to study extinction or disinhibition of conditioned responses as such. It was with some surprise, therefore, that behavior similar to these two phenomena was noted. Although it has been long known that infant responses are quite variable, it seemed unusual that a subject, when no sign of self-stimulation was apparent, should suddenly not respond after having given several successive responses to the conditioned stimulus. These states of no response seemed to occur when the subject was very quiet, and were frequently followed by sleep. Of equal interest was the observation that conditioned responses first seemed to appear when the infant was very slightly active, and also that it was under conditions of very slight activity that the conditioned response reappeared after having disappeared. The latter fact strongly suggested spontaneous or autonomous disinhibition by tactual or proprioceptive stimulation. Yet, it was occurring here in infants not only before conditioning was complete but before any specific attempt had been made to build up internal inhibition, i.e., to extinguish the response. Two interpretations of-these observations are possible. One, in terms of Pavlov's concept of irradiated internal inhibition, would hold that both response decrement and muscular relaxation are the result of a development of cortical inhibition, the dissipation of which would be accompanied by the termination of muscular relaxation and the reinstatement of the conditioned response. He states, " . . . in all cases of internal inhibition, . . . drowsiness and sleep are met with continually ... " (18, p. 251), and further, "The development of inhibition with its ultimate expression in the form of sleep is due to functional fatigue of the cellular structures of the cortex" (18, p. 259).

The other interpretation has not been expressed so succinctly. It is peripheral rather than central in nature. According to it, the magnitude of response of an organism at any moment depends to a certain extent upon the state of tension of the skeletal musculature at that moment. This interpretation will be amplified throughout the paper. According to it, decrement of the conditioned responses in these infants occurred when other conditions had induced relaxation of the skeletal musculature, and when relaxation had been interrupted, the response reappeared.

There is abundant evidence to support such an interpretation. Within the past two decades a considerable amount of work has been done (1) on the effect of changes in muscular tension upon behavior, and (2) on the effect of changes in behavior upon muscular tension. Dodge (6) G. L. Freeman (9), Bills (4), and Jacobson (16) have figured prominently in these investigations. The results have been surprisingly consistent and have produced a body of fact which has been summarized by Freeman (10) in his recent text on physiological psychology. The main generalizations are:

1. Augmented tension, up to a certain limit, results in response increment; beyond that limit it results in response decrement.

2. Decreased tension results in response decrement.

3. Mental work results in tension increment.

4. Cessation of mental work results in tension decrement.

Freeman therefore has been led to make the statement that,

. . . all response contains two aspects, the exteroceptive-phasic influence and the proprioceptivetonic influence. . . . Tonic contraction precedes and underlies all phasic contraction (10, 458-59). [It] . . . may be called the postural substrate—a general neuromuscular background which operates to facilitate, inhibit, or otherwise modify phasic responses (10, p. 451).

Illustrations of these generalizations are numerous, and have already been summarized by Bills (4) and Freeman (7). A few are cited here. Concerning response decrement, Dodge (6) has stated that during voluntary relaxation there is an inhibition of reflex processes. Jacobson and Carlson (15) have found that the amplitude of the knee jerk is decreased during conditions of relaxation. A familiar example of response decrement is the decreased speed and extent of movement under conditions of rest or relaxation. Recently, Bagchi (1) has experimentally demonstrated with the electromyographic technique that latency of voluntary movement during relaxation is two to four times greater than that during normal posture. Concerning response increment, Tuttle (23) has shown that the amplitude of the knee jerk is increased about ten times over the passive state when the subject is attending to mathematical symbols, and Freeman (8) has completed the picture by showing that increments in quadriceps tension occur during mental arithmetic. Therefore under a condition of hypertension the amplitude of the knee jerk is increased. A familiar example of response increment following tonus increment is the 'handgrip' technique, long used by neurologists in order to augment weak reflexes.

If this collation of experimental results pertained only to response decrement during muscular relaxation, there would be no basis upon which to contradict Pavlov's interpretation. Since, however, it also includes examples of response increment following experimentally induced tension increment, there is excellent ground for maintaining that the state of the peripheral musculature per se in some way affects the responses of an organism. Regardless of the causation of a given state of muscular tension, and regardless of how it serves to modify phasic responses, it becomes possible to proceed from a premise about peripheral conditions. Such a premise might be stated as follows:

(a) If, in an organism functioning at a level of maximum magnitude of response, there occurs a reduction in tension in the skeletal musculature, there will likewise occur a reduction in magnitude of response.

From this premise it may be predicted that

(b) Any condition or set of conditions inducing relaxation of the skeletal musculature in an organism functioning at an optimum level of response, will also induce a decrement in response.

Now we know from Pavlov that relaxation and sleep did occur in his experiments. To quote a passage already partially cited, " . . . in all cases of internal inhibition, . . . drowsiness and sleep were met with continually. In the case of extinction of a conditioned reflex some animals even at the first extinction showed not only a disappearance of the secretory and corresponding motor reaction, but also a great dullness as compared with the normal state of the animal before extinction" (18, p. 251). The conclusion may therefore be drawn that

(c) At least part of the decrement in response noted in Pavlov's experiments on extinction of conditioned responses may be attributed to relaxation of the skeletal musculature.

Here, then, is the logical exposition of the interpretation which was suggested but not analyzed during the experiment upon disinhibition (25).

Since the response decrement which occurs with general relaxation and drowsiness is not confined to any specific response or pattern of responses, the above deduction might be amplified.

(d) If there exists a generalized relaxation of the skeletal musculature there will exist a generalized decrement in response.

It is not surprising, then, to find Pavlov saying:

. . . experimental extinction of any single conditioned reflex results, not only in a weakening of that particular conditioned reflex which is directly subjected to the extinction (primary extinction), but also in a weakening of other conditioned reflexes which were not directly subjected to extinction (secondary extinction). . . . Sometimes secondary extinction reaches a profound degree, involving even the unconditioned reflexes (18, p. 54).

Nor is it surprising that he felt the need of an irradiating general inhibition to explain the phenomena he witnessed.

Now if this peripheral explanation of response decrement in some extinction experiments is accepted, certain questions arise:

1. How is muscular relaxation itself to be explained?

2. Does muscular relaxation account for all of the response decrement in experimental extinction?

The answer to the first question is considered to be without the scope of the present paper. Let it suffice to say that certain factors influencing muscular tension, as well as some of the nerve pathways over which they operate, are known. Since the latter are fairly well localized, there is small need to postulate a vague "internal inhibition' irradiating over the cortical hemispheres. The second question, however, is essential for the present discussion, and it has already been answered in the negative, for there are repeated examples in Pavlov's own work wherein the response decrement has become specific and is accompanied by no reduction in other responses. It must be concluded, therefore, that

(e) The construction of the original premise is correct, not only in form but in fact. Only part of the response decrement occuring in experimental extinction may be attributed to reduction in the postural substrate.

It follows then that:

(f) Experimental extinction is a two-fold process when muscular relaxation is known to have occurred.

And further,

(g) Experimental extinction of a conditioned response should be possible without the general decrement induced by muscular relaxation.

Since experimental extinction is involved in establishing delayed, trace, and differential conditioned responses in that in the first two the immediate response to the conditioned stimulus becomes extinguished and in the last the responses to all but a specific conditioned stimulus become extinguished, it becomes possible to predict that,

(h) The development of delayed, trace and differential conditioned responses should be possible without the general decrement induced by muscular relaxation.

Data already exist supporting these predictions. Wendt (24), in a recent experiment, has presented evidence purporting to show that extinction of a conditioned response is an active process, explainable in terms of competition between reactions systems. His experiment, the subjects of which were monkeys, is briefly described in his own words:

Each animal is trained to respond to the onset of a tone by opening a drawer to get a piece of food. The animal is confined in a cage with a barred aperture (grille) through which it can reach out to grasp the handle of the food drawer. The grille can be closed by a roller curtain attached just beneath the grille so that raising of the curtain closes the grille. . . . The situation confronting the animal is this: At approximately one minute intervals a ready signal is given and five seconds later the curtain is opened, giving the animal access to the drawer. It eventually learns to come to the grille when the curtain opens, waits there in a listening attitude until the tone occurs, then promptly opens the drawer and takes out a piece of food. The tone occurs either eight seconds or sixteen seconds after the curtain opens, the delay interval being changed in an unpredictable order. If the animal responds before the onset of tone, the drawer is empty. It therefore learns to inhibit response during the delay period (pp. 263-264).

It will be seen that the training involved learning to respond to the drawer for food, then learning to inhibit that response until the tone was heard. According to Pavlov this would involve both extinctive inhibition and the inhibition of delay, both of which are aspects of internal inhibition. Furthermore, we should expect an irradiation of internal inhibition during the period of delay which should result in observable muscular relaxation and a decrement in other responses. On the contrary, Wendt found that his subjects, if anything, exhibited greater activity during this period. In no case did he find relaxation, sleep, or any general decrement in other responses.⁴ He did find a change in activity, which continued to vary during training, but the change always involved some more or less specific response appropriate to the situation and to the species involved. The important consideration is that the monkeys learned to inhibit immediate response to the drawer by substituting some other activity. He holds that this substitute activity must be regarded as a competing response which has become dominant over the anticipatory conditioned response, and thus that extinction and delay are explainable in terms of the principle of reciprocal innervation.

Now, how shall these disparate results be reconciled? Both Pavlov and Wendt worked under well-controlled conditions. There was, however, a major difference in their experimental techniques. Pavlov's animals were held stationary in harness while Wendt's were free in a cage. As Wendt has already pointed out, the characteristic behavior of a confined animal in a controlled environment is quiescence after a short period of adjustment, while the behavior of a monkey free in a cage results in continuous self-stimulation from proprioceptive and tactual end organs. Wendt has suggested that perhaps something in Pavlov's experiments facilitated the pattern of relaxation and sleep. He notes three factors (1) restraint of the animal, (2) absence of varied stimulation with its possibility of varied reactions, and (3) repetitive monotonous stimulation. Actually, the first factor is the important one for, in controlled experiments on extinction, stimuli with the exception of those of proprioceptive and tactual origin are eliminated and the presentation of the conditioned stimulus by design is repetitive and therefore monotonous since other stimuli are absent. The point is that in Wendt's experiment muscular relaxation never appeared because the stage was not set for it, while in Pavlov's experiments the stage was set for muscular quiescence, and therefore for relaxation.

It has been experimentally demonstrated, then, that experimental extinction, trace, and delayed responses can be established with little or no indications of irradiation of inhibition and the predictions may be regarded as substantiated. The question immediately arises, how are these phenomena to be explained if not by internal inhibition? Razran (19) has suggested an interpretation in terms of changes in dominance of certain neural patterns. Wendt's interpretation is not essentially different. He believes that his subjects succeeded in inhibiting the anticipatory response to the food drawer when some substitute response apparently had become dominant. Regardless of how that shift in dominance occurred, it offers an opportunity to explain tentatively the inhibition of a given response in terms of the principle of reciprocal innervation of a competing reaction system.

Wendt needlessly belabors his interpretation, however, when he claims that relaxation and sleep also are to be interpreted as competing reactions. It is not intended to suggest that this is never possible. There are many indications for believing that these patterns may be conditioned to certain stimuli and some of the instances described by Pavlov should probably be so interpreted. Nevertheless, there are at least three reasons for questioning any generalized conception of relaxation and sleep as competing reactions.

In the first place, when one thinks of a competing reaction, he thinks of a positive muscular reaction. Wendt indicated his realization of this implication when he cited McDougall's references to contracture of the eyelids and to certain postures favoring sleep, as possible muscular components of the competing reaction of sleep. Relaxation however is, to the best of our knowledge, the fundamental prerequisite of sleep, and it is a passive muscular reaction. So far as is known, it is not dependent upon peripheral innervation but instead upon reduction of that innervation. Even if there is a sleep center, its mechanism will probably be found to be a reduction of stimulation at the nuclei controlling muscular tension.

In the second place, even though it is admitted that sleep may become conditioned to normally distracting stimuli, in the usual analysis relaxation and sleep, except in the greatly fatigued or intoxicated organism, is rather a fragile process during its onset. Instead of it competing with other reactions, the latter typically compete successfully with it.⁵

In the third place, if we agree on the ground of passivity and fragility that the competitive potency of relaxation and sleep is questionable, it seems theoretically premature to cast these responses in the mold of other more dominant reaction patterns and speak of only one form of inhibition. Since we know that experimentally induced increments in muscular tension will facilitate response, and conversely that experimentally induced decrements in tension will inhibit response, we have both logical and parsimonious reasons for recognizing an inhibition due to reduction of proprioceptive facilitation as well as an inhibition due to a competing reaction.

Summarizing the foregoing discussion several established principles and one postulate appear.

Principle 1.—Decrement in muscular tension usually will be accompanied by a general decrement in response.

Principle 2.—Increment in muscular tension usually will be accompanied by a general increment in response.

Principle 3.—The inhibition of a conditioned response may occur without the accompaniment of muscular relaxation and therefore without the phenomenon of general response decrement.

Principle 4.—When muscular relaxation accompanies the inhibition of a conditioned response, decrement in response will be generalized, and the total process must be regarded as two-fold.

Postulate 1.—There are at least two forms of inhibition: (a) Reduction of proprioceptive facilitation and (b) a shift in dominance to a competing reaction.

These premises may be utilized in a further analysis of behavior.

A. (From 3 and 4.) Muscular relaxation and its accompanying generalized decrement in response is not necessarily related to the inhibition of a conditioned response and when it occurs it must be regarded as an experimental artifact as far as the conditioning process itself is concerned.

B. (From 1 and 2.) During the development of inhibition of a conditioned response which is being accompanied by muscular relaxation, any extraneous stimulus producing an increment in muscular tension will be followed by an increment in the conditioned response.

This prediction has already been realized in the data which Pavlov refers to as 'disinhibition' (cf. footnote 5).

C. (From Principle 4 and Deductions A and B.) Disinhibition is not necessarily a law of the conditioned response per se.

It is only known to operate upon that phase of inhibition which has been shown to be an experimental artifact. Disinhibition, therefore, is a law of response as influenced by muscular tension. Whether it also is a law of the pure conditioned response remains to be determined.

D. (From Principles 1, 2, and 4.) After the extinction of a conditioned response which has been accompanied by muscular relaxation, a reinstatement of a more normal level of muscular tension will be accompanied by at least a partial reinstatement of the conditioned response.

This prediction also has been realized in the findings of Pavlov and others. It is referred to as 'spontaneous recovery' and is supposed to occur within a short time (varying according to the 'depth of inhibition') after the extinction process. In terms of the present interpretation the length of the time interval becomes unimportant. Importance attaches instead to the changes in tension of the skeletal musculature during the interval.

E. (As in C above.) Spontaneous recovery is not necessarily a law of the conditioned response per se.

Like disinhibition, it is only known to operate upon that phase of inhibition relating to muscular relaxation. It also is a law of behavior as relating to muscular tension, and its operation in the pure conditioned response remains to be determined.⁶

Other deductions relate to the effect of drugs upon the conditioned response.

F. (From Principle 1.) The extinction of a conditioned response should be facilitated by administration of a drug known to effect muscular relaxation.

G. (From Principle 1.) The extinction of a conditioned response should be impaired by administration of a drug known to interfere with muscular relaxation.

Switzer's work (22) concerning the effect of caffeine on extinction already affords support of the latter deduction.

Another deduction takes the form of an explanation of the inhibition which is purported to adhere to the conditioning process itself. It has been reported by Platonov (cf. Razran's review, 20) and more recently by Hovland (12), that a conditioned response immediately after the conditioning process is occasionally lesser in magnitude than it is some time later. Since it seems that monotony of stimulation is one factor which effects muscular relaxation, and since there is a certain monotony in the conditioning process especially if the stimulations occur at short intervals, it is possible to assume that some relaxation of the skeletal musculature may occur during the conditioning process itself. The logical deduction follows:

H. (From Principle 1.) If, during the conditioning process, there occurs a decrement in muscular tension there will occur a decrement in the conditioned response.

If this deduction is the proper explanation for the inhibition in question then the conditioned response so inhibited should be subject both to disinhibition and spontaneous recovery. Only the latter has yet been demonstrated.

A final inference would question the validity of a smooth curve of experimental extinction. It proceeds from the assumptions that (a) learning is not a smoothly progressive process but is interrupted by numerous plateaus and regressions, and that (b) muscular relaxation, if uninterrupted, is a smoothly progressive process. A slight degree of relaxation reduces the postural substrate. The facilitative effect upon higher nervous centers is thereby reduced and cortical and bulbar stimulation of the peripheral musculature are in turn decreased, again diminishing the proprioceptive bombardment upon the higher centers. The latter assumption would, of course, explain parsimoniously the smooth curve of extinction which Pavlov and others have reported, but leave the true nature of the curve unpredicted. If Wendt is right in maintaining that extinction of one response involves learning of another, the curve of pure extinction may be the inverse of the learning curve of the newly dominating response.

It is apparent that the picture of inhibition presented in this paper is radically different from that presented by Pavlov. If the present analysis is tenable, much of the work of Pavlov and his pupils is questionable, for their researches on inhibition were accompanied to such an extent by drowsiness and sleep that Pavlov, at least, regarded the whole as one process. Probably most of the data concerning inhibition present an erroneous picture as far as the conditioning or learning process itself is concerned. In short, the work on inhibition of conditioned responses must be repeated under conditions of known stable muscular tension before many laws of conditioned responses can be recognized and applied to the more complex processes of human behavior.

NOTES

¹ The writer is indebted to Drs. Orvis C. Irwin, Charles Van Riper, Eliot H. Rodnick, and Carl Iver Hovland, for their critical reading of the manuscript.

² The substance of this paper was reported at the summer meetings of the A.A.A.S. held in Minneapolis, Minnesota, May, 1935.

³ Recently the data have been re-read and consistent though variable increases in skin resistance have been found during extinction in eighteen of the twenty subjects. In the light of a recent experiment by Wenger and Irwin (27) which demonstrated that increases in palmar skin resistance are related to muscular relaxation, these data confirm the report of the subjects that much relaxation occurred during extinction. In order to determine whether the same would be true of a seated subject one adult was conditioned and readings were taken of resistance during conditioning and extinction. During conditioning, resistance maintained a nearly constant level. During extinction it rose. Therefore relaxation also occurred in the seated subject.

⁴ It should, perhaps, be stated that Hull (13) has reported some evidence of irradiation of inhibition to another conditioned response. Trace conditioned galvanic responses were established in human adult subjects and during the period of delay between the presentation of the conditioned stimulus and the onset of the conditioned response, the stimulus for a previously conditioned lid closure was given. According to Pavlovian theory, a decrement in amplitude of lid closure would have been expected. On the contrary, results from ten subjects showed little or no indication of a weakening of the conditioned lid reaction during the period of delay of the trace conditioned response. However, the mean latency of the lid responses was found to have increased 17%. Dr. Eliot H. Rodnick (21) informs the writer that he has repeated Hull's experiment under slightly different conditions and has attained a 19% decrement in amplitude and a 15% increment in latency.

Under the conditions of the experiments the response decrement and latency increment can hardly be explained in terms of muscular relaxation. However, it is not necessary to revert to 'irradiation of internal inhibition' for an explanation. One needs only to refer to Wendt's application of the principle of competition between reaction systems. If this explanation is valid, the 'set' to perform some other response should inhibit the interpolated lid response just as much as did the conditioned delay stimulus which Rodnick used.

⁵ It is pertinent here to cite from Pavlov (18, p. 50): ".. . to obtain a smooth curve of extinction of a conditioned reflex it is necessary to maintain the unreinforced conditioned stimulus rigidly constant in character and strength; the environing experimental conditions also must remain absolutely constant . . . the effect of an extra stimulus of even small intensity is a temporary weakening, not of the reflex, but of the progress of experimental extinction."

⁶ Humphrey (14) has reported disinhibition (which he terms 'dehabituation') and spontaneous recovery in experiments on negative adaptation in controlled environments. If, as has been postulated, these phenomena are the result of tension increment, it should be possible to find that tension decrement had occurred previously. Unfortunately, Humphrey's data do not provide the necessary information. The literature was searched and paradoxically the only pertinent data were found in Pavlov (18) where the 'investigatory reflex' of a dog to a whistle was negatively adapted. Pavlov says (p. 255) "Special experiments . . . showed that the disappearance of the investigatory reflex is based on the development of inhibition, and is in all details analogous to the extinction of conditioned reflexes." Even more specifically (p. 256), "The inhibition of the investigatory reflex invariably leads to drowsiness and sleep (even more easily than the inhibition of conditioned reflexes)."

It is thus found that under certain conditions negative adaptation results in relaxation, which, among other observations, seems to have led Pavlov to state that it also is mediated through an irradiation of internal inhibition. It is known, however, in common experience, that adaptation of a response may occur in active and uncontrolled environments. Therefore it seems that negative adaptation, like extinction, must be considered as a twofold process when it occurs in a controlled environment, and that the relaxation which there is observed must likewise be regarded as essentially artificial to adaptation as such. Perhaps both 'dehabituation' and 'disinhibition' as well as 'spontaneous recovery' might more accurately be described as 'de-relaxation.'

REFERENCES

1. Bagchi, B. K., An electromyographic study with respect to speed of movement and latency, disparate and reciprocal innervation, attention and relaxation, Psychol. Monog, (to be published).

2. Bekhterev, V. M., General principles of human reflexology, New York: International Publishers, 1932, pp. 467.

3. Beritoff, J. S., On the fundamental nervous processes in the cortex of the cerebral hemispheres, Brain, 1924, 47, 109-148.

4. Bills, A. G., The influence of muscular tension on the efficiency of mental work, Amer. J. Psychol., 1927, 38, 227-251.

5. Chappell, M. N., Inhibition, facilitation, learning, Psychol. Rev., 1931, 38, 317-331.

6. Dodge, R., Conditions and consequences of human variability, New Haven, Conn.: Yale University Press, 1931, pp. xi, 162.

7. Freeman, G. L., Mental activity and the muscular processes, Psychol. Rev., 1931, 38, 428-449.

8.——, The spread of neuromuscular activity during mental work, J. Gen. Psychol., 1931, 5, 479-494.

9.——, Studies in muscular tension [Collected papers], 1929-1933, pp. 182.

10.——, Physiological psychology, New York: Ronald Press, 1934, pp. 579 + x v.

11. Guthrie, E. R., The psychology of learning, New York: Harper & Bros., 1935, pp. viii, 258.

12. Hovland, C. I., The effects of varying amounts of reinforcement upon the generalization of conditioned responses, Psychol. Bull., 1935, 32, 731-732.

13. Hull, C. L., The alleged inhibition of delay in trace conditioned reactions, Psychol. Bull., 1934, 31, 716-717.

14. Humphrey, G., The nature of learning in its relation to the living system, New York: Harcourt Brace, 1933, pp. vii, 296.

15. Jacobson, E. and Carlson, A., The influence of relaxation on the knee jerk, Amer. J. Psychol., 1925, 73, 324-328.

16. Jacobson, E., Progressive relaxation: a physiological and clinical investigation of muscular states and their significance in psychology and medical practice, Chicago: University of Chicago Press [C. 1929], pp. xiii, 428.

17. Lashley, K. S., Nervous mechanisms in learning, in 'The Foundations of Experimental Psychology' (Ed. by Carl Murchison), Worcester: Clark University Press, 1929, pp. 907.

18. Pavlov, I. P., Conditioned reflexes: an investigation of the physiological activity of the cerebral cortex, Oxford University Press: Humphrey Milford, 1927, pp. 430-xv.

19. Razran, H. S., Theory of conditioning and of related phenomena, Psychol. Rev., 1930, 37, 25-43.

20.——, Conditioned withdrawal responses with shock as the conditioning stimulus in adult human subjects, Psychol. Bull., 1934, 31, 111-143.

21. Rodnick, E. H., Characteristics of delayed and trace conditioned responses (to be published).

22. Switzer, St. Clair, The effect of caffeine on experimental extinction of conditioned reactions, J. Gen. Psychol., 1935, 12, 78-94.

23. Tuttle, W. W., The effect of attention or mental activity on the patellar tendon reflex, J. Exper. Psxchol., 1924, 7, 401-419.

24. Wendt, G. R., An interpretation of inhibition of conditioned reflexes as competition between reaction systems, Psychol. Rev., 1936, 43, 258-281.

25. Wenger, M. A., External inhibition and disinhibition produced by duplicate stimuli, Amer. J. Psychol., 1936, 48, 446-456.

26.——, An investigation of conditioned responses in human infants, in 'Studies in Infant Behavior,' III, Univ. Iowa Stud., Stud. in Child Welfare, 1936, 12, No. 1, pp. 207 (pp. 7-90).

27.——and Irwin, O. C, Fluctuations in skin resistance of infants and adults and their relation to muscular processes, in 'Studies in Infant Behavior,' III, Univ. Iowa Stud., Stud. in Child Welfare, 1936, 12, No. 1, pp. 207 (pp. 141-179).

28. Winsor, A. L., Observations on the nature and mechanism of secretory inhibition, Psychol. Rev., 1930, 37, 399-411.