on decerebrate cats. Sherrington describes in the decerebrate cat amongst other motor reactions some mimetic movements simulating expression of certain affective states, but he could never evoke such mimesis as might, had the cerebrum been present, have been indicative of pleasurable sensation. Pagano reports that in the dog, deprived of its cortical functions, he got mimetic expressions of both fear and anger1. And Goltz's dog also showed expression of anger without any signs of other emotions. Goltz argued from this that the other emotions seemed to demand higher nervous organisation than did anger. Observations on infants seem to confirm this view, for there do not seem to be any clearly defined expressions indicative of pleasure in the new born babe-laughing aloud, for example, did not occur in one of my children until the eighth week, whereas expressions of displeasure were certainly present at birth, and reflex starting at loud noises occurred during the first week. In reference to Goltz's argument as to anger Sherrington states: "Be that as it may, the retention of its expression by Goltz's dog indicates that by 'retrogradation' the complex movement of expression has in certain emotions passed into a simpler reflex act.... The impulsive movement has in this case become an automatic reflex no longer necessarily combined with the psychical state whence it arose "2.

The nerve centres of mimetic expression are said by some to have their seat in the optic thalamus, but as Sherrington points out various grades of mimetic expression exist, and some seem phylogenetically much older than others; observations on hemicephalic children and Sherrington's experiments in which part of the optic thalamus was removed point to the conclusion that in some reactions neither the forebrain nor midbrain are necessary. It seems justifiable therefore to make the speculation, as Head and Riddoch have done, that the 'flexionreflex' and the 'mass reactions' in cases of traumatic lesion of the spinal cord are of the same nature as the mimetic expressions in decerebrate animals, that they are pseudaffective reflexes of a much earlier stage of evolution. And it might even be possible to speculate further, in conformity with Pierce Clark's views on epilepsy that the fit is a regressive phenomenon, and so ascribe the epileptic convulsions to the re-awakening of certain pseudaffective reflexes, much in the same way that Babinski's extensor response reasserts itself when cut off from the inhibition of the cerebral cortex by lateral sclerosis, though in order to uphold such a view it would be necessary to assume that the stimulus

1 L'Année Psychologique, 1914, p. 483.

2 C. S. Sherrington, The Integrative Action of the Nervous System, p. 267.

which sets up the reaction is also responsible for removing the inhibition.

These considerations point to the conclusion that muscular expression is no indication, at least as far as the coarser emotions are concerned, of the subjective feeling experienced, for it would seem to be justifiable to make the assumption that decerebrate animals do not in fact experience any such feeling. It is at least certain that the amount of muscular expression is intimately related to the strength of the cortical control which is brought into play. Given the same strength of subjective feeling the person with no powers of inhibition may give vent to an explosive outburst, whilst the one with a higher development of self control will give no outward expression at all, and in the same person the discriminative and inhibitory functions of the cortex vary from time to time, depending largely on physical factors, such as fatigue and the influence of hormones, endocrines, drugs, alcohol, etc. But it is also true that the muscular expression depends on the strength and duration of the impulse, so that Spencer insisted on "the general law that feeling passing a certain pitch habitually vents itself in bodily action and that nerve force takes first the most habitual routes." That is, that it expresses itself first in the voice and muscles of the face and hand. Spencer's law however would seem to be a one-sided explanation as it does not take into account the question of control, unless we agree to accept McDougall's drainage theory of inhibition. For we recognise that the first functions in the central nervous system to become deranged in mental disorders are those that have developed last from the evolutionary point of view, the higher centres involving the discriminative and inhibitory functions are affected first, and it would seem that Spencer's 'habitual routes' are the last to develop as methods of expression and are less under the influence of control.

It would appear then that the term 'emotional' as popularly used is incorrect. The persons to whom we ascribe this term are really persons who have little or no control over their muscular expression-they should rather be called 'impulsive.' The converse question as to whether emotion can be subjectively experienced without any muscular expression we can answer in the affirmative from our own introspection. And as Shand says " "the arrest of an instinct is that which most frequently excites the emotion connected with it-there is no anger so intense as when the blood boils and all the sudden energy that comes to us cannot vent itself on our antagonists"1.

1 Op. cit. p. 189.

This brings us back to Drever's conception that emotion only occurs when an instinct is checked and tension arises. More recently Carver has expressed a similar view and defined emotion "as the subjective. experience which develops when gratification of the instinctive impulse is held in check by higher level control"1. He quotes in support the observation of Romanes that the lower the mental development of the organism, the more fixed and definite becomes the instinctive reaction and the less becomes the manifestation of emotion. The truth of Romanes' statement depends on the meaning given to 'manifestation of emotion.' If by manifestation of emotion is meant only visceral expression then it seems to be perfectly true, for experimental results show that the lower the mental development the greater is the muscular expression and the less the visceral expression, provided that we are justified in regarding the psycho-galvanic reflex as an indication of the visceral expression of emotion.

All therefore that we are entitled to say is that the muscular expression of emotion cannot be treated as a separate problem divorced from the question of control. It certainly seems as if the processes of inhibition and facilitation are the fundamental problems. So that when we return to our original question as to whether the so-called emotional or hysterical persons really impulsive persons-subjectively experience any emotion, we can only answer, on the evidence of muscular expression and the behaviour of decerebrate animals, that it is quite as justifiable to assume that they do not as to assume that they do. And we are left only with the visceral expression which is not under voluntary control to answer our question.

We have then to ask-Are we justified in assuming that the amount of visceral reaction is an indication quantitatively of emotional excitement and subjective experience? This question is better left for discussion in connection with my experimental results. I would like to be able to maintain that we are so justified, but there are difficulties in applying this assumption. It seems certain at any rate that the emotion gains in intensity from the consciousness of visceral sensations and a way out of the difficulty might be to restrict the term emotion so as to make it include only the central psychical excitement together with the sensations from visceral expression, and to exclude muscular expression as being due to instinctive impulse.

This seems to be the most profitable view to take as an explanation 1 A. Carver, "The Generation and Control of Emotion," Brit. Journ. of Psychol. 1919, X. 52.

of the facts, and I would therefore accept with some reservations the definitions of emotion as given by Drever and Carver. But in order to be a little more explicit I will anticipate my experimental results and will provisionally describe emotion as "a subjective feeling, consisting of central excitement and consciousness of peripheral sensations, occasioned by situations which powerfully oppose or facilitate the aim of any instinctive impulse."

(2) Experimental Results.

For the purpose of studying the visceral expression of emotion I have been experimenting with the so-called psycho-galvanic reflex. Before deciding as to how far we are justified in assuming that we can measure the intensity of an emotional reaction by this method it is necessary to set out the results of my experiments.

There are two methods of obtaining the psycho-galvanic reflex, that originally reported by Féré in 1888 due to apparent change of resistance in the skin when a weak electric current is passing through the body, and that reported by Tarchanoff due to production of an electromotive force when no external current is being used. And of these two methods it has been demonstrated by other investigators that the use of the current by Féré's method is the more reliable for measuring emotional reactions1.

I am indebted to both Professor Waller and Dr F. L. Golla for help and instruction in the technique of this method, and it is with Féré's phenomenon that all my experiments have been concerned.

The technique I employed originally was that described by Waller2, in which the subject is placed in the fourth arm of a Wheatstone bridge and his resistance determined by balancing out in the third arm. I have used a Paul's ‘Unipivot' galvanometer, which has an internal resistance of 830 ohms, and as electrodes either saline baths or zinc discs, covered with lint soaked in saline, which are bandaged on to the palm and dorsum of the hand. I soon found that no reliable work could be carried out without photographic records, and for this purpose I adopted a method suggested by Professor A. V. Hill for the photographic recording of the heat production of muscles3.

As a source of light for the galvanometer I use a 4-volt lamp, and as the light is reflected from the mirror of the galvanometer it is passed 1 F. L. Wells and A. Forbes, "On certain electrical processes in the human body and their relation to emotional reactions," Archives of Psychology, 1911, no. 16, p. 25.

2 Proc. of the Royal Society, 1919, B. vol. 91, p. 17.

3 A. V. Hill and W. Hartree, "Photographic recording of the heat production of muscles," Journ. of Physiology, 1920, LII. no. 5, p. lxx.

through a wide reading lens and brought to a focus, about the size of a pin's head, on to a kymograph, to which sensitive bromide paper is attached. The kymograph is adjustable as to speed and worked by clockwork, and has a small signal lamp attached to its framework. The lamp allows for a streak of light being projected on to the bromide paper for use as a stimulus signal, and this signal lamp is worked synchronously with the contact for the stimuli which can be given electrically, e.g. the ringing of a bell, flashing a light, a motor horn, releasing a weight, and giving a faradic shock. Time marking is indicated by a shutter working on an electro-magnet connected with a Brodie's clock, by means of which the light can be cut off at regular intervals, so that the time is indicated on the galvanometric curve itself.

This method allows of a continuous photographic record being taken over a period of half an hour, and by the use of an additional apparatus for rolling the paper off cylinders on to the kymograph, the record may be continued for 2-3 hours, or more. I am now using this method of getting simultaneous records of reactions from different parts of the body by reduplicating my Wheatstone bridge and galvanometer circuits.

Reproductions of photographs taken in this way are shown in the figures attached. A rise in the curve indicates a decrease in the apparent resistance of the skin-the method of time marking is indicated in the interrupted curves, the interruptions having been made every second, and the stimulus signal is indicated by the mark below the curve.

In a recent number of Brain1 I have given a lengthy review of the work of previous investigators on this subject and of the problems involved. It has been definitely established that galvanometric deflections can be produced by certain psychical, physical and physiological causes, and it is agreed that the genuine galvanic reflex occurs only after a latent period of an average of two to three seconds following the stimulus, and that the most marked reactions are obtained either from the palms of the hands or the soles of the feet. There is still a controversy as to the physiological factors responsible for the reflex, and until these are ascertained we shall not be able to make much progress as to its clinical value. It is claimed by some observers that the reflex is due to the activity of the sweat glands, and those who maintain this view have reported experiments to show that the reflex can be abolished by the subcutaneous injection of atropine. It may be that these observers were dealing with secretory currents, for Waller has shown quite clearly that atropine does not abolish the much larger reaction obtained by Féré's 1 E. Prideaux, "The Psychogalvanic Reflex: A Review," Brain, 1920, XLII. p. 50.