So long as the contraction of the left ventricle lasts and the valves are open, the action of the heart produces a certain pressure in the aorta, but the influence of the intraventricular pressure is zero when the valves are closed. The second phase of the curve Fig. 1, where the pressure is zero, certainly gives the influence of the intraventricular pressure during the diastole, because there is no communication between the ventricle and the arterial system when the valves are closed. The question is whether the rest of the curve can represent the changes of the intraventricular pressure when the valves are open.

The first curves of intraventricular pressure were traced by Chauveau and Marey. These experiments were made on a horse, and they have been repeated since it was discovered that they can be performed also on smaller animals. Besides Chauveau and Marey may be mentioned the names of Fick, Huerthle, v. Frey, Rolleston, Bayliss and Starling. The curves obtained by various observers belong to two types; one shows the so called "plateau," the other does not. Recent experiments have proved that this difference of results is due to a difference in methods. This also is suggested by the fact that different curves have been obtained from animals of the same species. Two methods have been applied lately for testing these curves of intraventricular pressure. The first was devised by Bayliss and Starling. It consisted chiefly in the photographic registration of the movement of the liquid in a manometer tube. The photographic registration is frictionless, and the mass of the moving liquid was so small that vibrations by inertia were fairly excluded for pressures which are not greater than the intraventricular pressure.[65] The second method was used by Porter. The idea of this method was to trace only a part of the curve, not the whole. The writing lever, thus, has in the beginning of the tracing no inertia at all, and the tracing may be overdrawn but is certainly correct in form up to the next point of inflexion of the curve.[66] These tests and the repeated experiments of Chauveau leave no doubt as to the existence of the plateau.

The varying pressure from the heart which produces the pulse wave may be described in this way: The pressure suddenly rises to a maximum and maintains it for a certain time; when the semilunar valves close, the pressure drops as suddenly as it rose, and remains at zero until the valves open again. Such a function can be represented by a curve like Fig. 1, and this is the reason why the complicated action of the heart can be superseded by the compression of a bag without changing the mechanical conditions of the problem. Of course it can not be expected that a schematic curve will show all the details of the real tracing. It is suggested, however, by Frank[67] that many of the small irregularities of the curve of intraventricular pressure are due to vibrations caused by the inertia of the apparatus and that the true form of the curve of intraventricular pressure is very simple. This remark is supported by Huerthle,[68] who tested the apparatus of Marey, Knoll and Grunmach. Marey's tambour was found to be the most exact, but even this instrument produces deformities in the tracings, though the general outlines are exact. This would indicate that the schematic representation of Fig. 1 is a very close imitation of the real form of the curve of intraventricular pressure, although empirical tracings do not show right angles and straight lines. It seems, however, that the undulations of the plateau are genuine, since they are found in the most reliable tracings, and it may be possible to explain them merely on the basis of the physical conditions of the experiment.

The fourth factor of importance is the rate of outflow. We may introduce the following assumption as to the rate of outflow of the blood through the capillaries: The outflow through the capillaries is uniform in the short time of one heart-beat. The fact has been mentioned above that the quantity of outflowing blood must be equal to the quantity of incoming, for any stationary form of the pulse movement; this new hypothesis means that the velocity of the outflow is constant. One might think that this assumption is warranted by the law of Poiseuille that the amount of outflow through a horizontal capillary filled with liquid under constant pressure depends on the fourth power of the radius and on the difference of pressure at the two ends of the tube, and is inversely proportional to the constant of friction and to the length of the tube. This law has been proved mathematically and tested physically only for horizontal tubes and constant pressure. Neither of these suppositions holds for the capillaries of the arterial system. The connection between the hypothesis in question and Poiseuille's law is this. Let us suppose that an artery splits up in a great number of arterioles which go off in every direction. The amount of outflow is then a complicated function, because the law of Poiseuille does not hold for every direction of the capillaries; but it will be equal to the outflow through a tube of certain radius and certain direction in the same time. Our assumption says that the law of Poiseuille holds for this typical but imaginary tube. The essential point of this hypothesis is merely the supposition that the outflow of blood through the capillaries follows α law.[69]

It is possible to show that the graphic registration of a movement under these four conditions must give curves which correspond to the pulse curves in every respect. The action of the left ventricle causes the pulse wave which travels through the arterial system with considerable velocity. This wave expands the arteries and the whole system is filled with blood because the wave arrives by its great velocity at the periphery before the contraction of the ventricle is finished. The increased pressure forces the blood to enter the arterioles, through which it passes at a constant rate. When the valves are closed, the amount of blood decreases uniformly and the volume of the blood contained in an artery can be represented graphically by a straight line of more or less steep descent, as is shown in Fig. 2. Now the walls of an artery have to a high degree the qualities of an elastic body, and, therefore, they are forced back by elasticity after being displaced from the position of equilibrium by the shock of the arriving pulse wave. The movement of a point of the arterial wall, therefore, results from two components: (1) From the movement which it would perform if it were merely forced to remain on the surface of the blood in the artery, and (2) from the movement due to the elasticity of the arterial wall. Both movements have the same direction, because the column of blood is enclosed in a cylinder the radius of which decreases regularly, and the elastic force of the arterial wall is directed towards the centre. The direction of both forces is in the line of the radius, and the resulting movement of these two components, therefore, can be found by simple superposition. Of the first component we know that it can be represented graphically by a straight line.

An elastic force tends always to bring the body back to the position of equilibrium; if the distance is not too great, the force is proportional to the elongation. A physical body is always under the influence of friction, the acceleration of which is opposite to the direction of the movement, and therefore diminishes the velocity. The form of the resulting movement depends on the amount of friction, and, roughly speaking, we may distinguish two types of elastic movements:[70] the first type is a periodic movement, the second an aperiodic. Let us suppose that a body is carried from its position of equilibrium by a sudden impulse, which transmits a certain velocity to the body. Friction and elasticity diminish this velocity, and after a certain time the body attains a maximum elongation, where the velocity is zero. Then the body returns under the influence of elasticity and under the retardation of friction. There are two cases possible, either the elastic force is strong enough to overcome friction and to carry the body over the position of equilibrium, or it is not strong enough. In the first case, it is easy to see, the body repeats the same form of movement on the other side of the position of equilibrium, and the conditions being constant a vibratory movement results as the stationary form. In the second case the body approaches the position of equilibrium asymptotically. The first case may be illustrated by the vibrations of a magnet needle suspended with little friction, the second by the movement of a door which is regulated by a well-working shutter.

These forms of the movement of a body under the influence of elasticity and friction are illustrated in Fig. 3.

Curve 1 shows a movement where friction is so small that it can be neglected; it is, of course, a simple sine curve. Curve 2 shows the effect of friction on vibrations. The period of damped vibrations is greater than in the frictionless movement, but the amplitudes are smaller. The amplitudes of a damped vibration decrease constantly and there is a simple relation between two subsequent amplitudes. The ratio between them is constant, and, therefore, if one amplitude and this constant ratio are known, all the other amplitudes can be calculated. The amplitudes of such a movement decrease as the terms of a geometric series. The dotted line in Fig. 4 represents the rapidity of this decrease. It is obvious that the smaller the constant ratio of two subsequent terms is, the more rapidly will the amplitudes decrease. This ratio depends on friction, and becomes smaller when friction becomes greater. A vibration under heavy friction dies out quickly. Curve 3 shows a movement where friction is too great to allow any vibrations. The body does not acquire a velocity which can carry it over the position of equilibrium, but it approaches this position with ever diminishing velocity.

These are the types of movement which the arterial wall can perform by its elasticity in consequence of the shock of the arriving pulse wave. The mechanical nature of the components on which depends the form of the sphygmographic curve is, therefore, known. The constructions in Fig. 5 show how the resulting movement can be found.

These curves are constructed in this way. The lines AB represent the time of the interval of one heart-beat. The straight line EB represents the decreasing volume of the artery and the curves on AB represent the elastic movement of the arterial wall. Both are synchronous movements, and a line perpendicular to AB gives the corresponding points. The points of the resulting movement are found by arithmetical addition of the two ordinates. The results of these constructions prove that the curves show the dicrotic elevation only if the elastic force is great enough to make a vibratory movement possible. Aperiodic movements do not produce this elevation. The friction is always great for the movement of the walls of an artery, and there are only the two possibilities, of a vibratory movement which dies out quickly, and of an aperiodic movement. This accounts for the fact that the dicrotic elevation may be missing sometimes, and that in other cases several secondary elevations may be seen, the number of which, however, is always limited, and their relative height rapidly diminishing. It may be remarked that the length of the lines AB seems essential to the form of the resulting curve. Curves I and III differ very much in the length of the lines AB, while the lines AE are equal and the vibratory movements are only slightly different. The resultants, nevertheless, seem to differ very much. It is easy to see that a different speed of the recording drum will have an effect on the tracings which is similar to that of a change in the length of the lines AB in the constructions. This is one more reason why mere inspection of the curves cannot give a satisfactory result.

These constructions show that the sphygmographic curves must show great variations, since the amount of blood pumped into the system, the elasticity of the arteries and friction of the surrounding tissues are subjected very likely not only to individual but also to local and temporal variations. But under given conditions only a certain form of the pulse wave is possible, and this form does not change so long as these conditions do not change. The sphygmograms in Fig. 6 show some of the typical forms of the pulse curve.

No. I shows the influence of high arterial tension, and No. II of low tension. The first corresponds to No. II in Fig. 5, the second to Nos. I and III. Nos. IV and V of Fig. 5 show the effect of great friction and small elasticity. The constructions differ in the form of the elastic movement; the position of equilibrium is reached with different velocity in both cases. The resulting movements differ slightly in the form of the catacrotic phase. Both forms may be seen in No. III of Fig. 6. This sphygmogram was taken from an artery with low tension, and this form of the sphygmographic curve is well known as characteristic of the "soft" pulse. If the artery has lost to a large extent the qualities of an elastic body, and if the outflow is very rapid, the pulse curve shows nothing but the slight elevation of the travelling wave; No. IV in Fig. 6 shows a curve of this character.

This theory explains many surprising facts which resisted every attempt at explanation. The anacrotic part shows a steep ascent, because it is due to the sudden arrival of the blood wave. It seems that an interruption in the descent may be seen only in abnormal cases. The sphygmograms of twelve normal individuals were observed regularly by me during more than a year without once discovering an anacrotic elevation.

The hemautographic curve of Landois is produced in this way. The form of this curve depends on the velocity of the escaping jet of blood. The velocity of the blood flow depends on the resistance of the arterial system in the sense that the velocity decreases when the resistance increases. When the arterial wall is in the negative phase of vibration the lumen of the artery is smaller, and, therefore, the velocity smaller. This is confirmed by the actual tracings of the velocity of the circulation by Marey.

It is also obvious that the dicrotic elevation never can arrive before the primary wave, because the arterial wall cannot perform elastic vibrations before it is expanded by the impulse of the arriving blood wave. Neither is it surprising that the "dicrotic wave" seems to travel in the same direction and with a velocity equal or almost equal to the velocity of the pulse wave. Such a difference can be produced only by a difference in the time of the vibrations of the arteries at different points of the body. The time of one vibration is necessarily very short, and the length of this interval depends on the circumstances which determine the elasticity of the arterial wall and the friction. These conditions may be subjected to local variations. If, therefore, the time-interval between the primary and the secondary elevation is measured at two different points (e. g., at the carotid and at the radialis) a difference of time may be found. Starting from the supposition that the dicrotic elevation is due to a wave travelling in the blood, one could attribute this difference of time to a velocity of the "dicrotic wave" which is slightly different from the velocity of the primary wave. The fact that the dicrotic elevation appears later in places farther from the heart was interpreted as a proof that the wave travelled out from the heart. No theory which assumes that the dicrotic elevation is due to a wave travelling in the blood can give a reason why two waves of the same form and origin should travel through the same liquid at different velocities.

At this point a theory must be mentioned, which was brought forward recently, because it is based on measurements of the velocity of propagation of the dicrotic wave. This theory is connected with Krehl's theory of the function of the valves. The blood, according to Krehl, enters the aorta through a small opening, and expanding in a large space it produces fluctuations and eddies, which would close the valves if they were not kept open by the blood which streams through under high pressure. They must, therefore, close at the moment when the aortic pressure is equal to the intraventricular pressure. This occurs shortly after the moment indicated by the beginning of the decline of the intraventricular pressure curve. Now the second sound of the heart is heard somewhere in the descending part of the cardiogram[71] and the measurements of Huerthle[72] have shown that the second sound is heard 0.02" after the beginning of the descent of the cardiogram. This seems to indicate that the second sound of the heart is in a temporal relation to the closure of the valves. Many theories of the origin of the sounds of the heart agree on this one point that the second sound is due to a noise in the muscles. It therefore may be supposed that the second sound is due to the tension of the valves when they close or shortly afterwards. The problem now would seem to be to find an elevation in the descending branch of the curve of intraventricular pressure, or in the tracings of the apex beat, which could be attributed to the closure of the valves. It was taken for granted that the curves of intraventricular pressure and those of the apex beat were identical. In many of these tracings an elevation was found which may be called "the wave f." This elevation is not found in all the tracings, and its position seems to be rather variable. Edgren[73] remarks that the wave f was always found near the abscissa no matter whether the preceding decline of the curve was great or small. In some of Chauveau's tracings the wave f is missing or indistinct,[74] in others it is very well marked and approximately in the middle of the descending branch of the curve.[75]

Edgren made experiments on the temporal relation of the wave f and of the dicrotic wave, which to avoid misunderstandings he calls the "wave f´." His experiments were made as follows. A sphygmogram from the carotid and a cardiogram were taken simultaneously, the points of the writing-levers being in the same vertical line. The wave f´ appeared a little after the wave f. The length of this interval could be calculated by measuring the distance between these waves, as the speed of the drum was known. From this was subtracted the time of propagation of the dicrotic from the heart to the point where the instrument was fixed. In this way it was found that the time between the appearance of the wave f and of the wave f´ was equal to the time of propagation of the dicrotic wave from the heart. Edgren concluded that the dicrotic wave is in close temporal relation to the closure of the valves.[76] To this comes the supposition that the wave f´ is due to a change of pressure proceeding from the heart. The wave f´, therefore, could be attributed to the tension of the valves.[77] Edgren and Tigerstedt are the chief exponents of this theory.

In so far as this theory assumes that the dicrotic elevation is due to a wave travelling from the heart to the periphery,[78] it is open to all the arguments against a theory of the central origin of the dicrotic wave. Against the more special assertion that the dicrotic elevation is in connection with the closure of the valves, the following facts must be mentioned. We grant that the tracings of the apex beat may be directly substituted for the curves of intraventricular pressure, although this is by no means obvious, since one tracing gives the form of the pressure changes and the other the effect of the shock of the heart against the wall of the chest. It is, furthermore, not proved that the wave f is due to the closure of the valves and that the waves f and f´ correspond to each other so closely as Edgren's experiments seem to indicate. His measurements of the length of lines were made with an exactitude of 0.1 mm., but his computations were carried to the third decimal place of a second. The third decimal is generally inexact and the second in a large number of cases. Experimental evidence, furthermore, directly contradicts the statement that the dicrotic elevation corresponds to the wave f. Fredericq[79] traced pressure curves in the ventricle and in the aorta, and determined the points of equal pressure in both curves. He thus found that a point near the beginning of the descent of the curve of intraventricular pressure corresponds to the dicrotic. His experiments are rather conclusive against the theory in question, since the wave f is very well marked in these tracings of Fredericq. The following facts, however, are fatal for the theory that the closure of the valves causes the dicrotic elevation: The dicrotic wave disappears in diseases like atheroma and arteriosclerosis which do not impair the function of the valves, but affect the elasticity of the arterial wall, and it is not affected by valvular insufficiency. The independence of the dicrotic from the function of the valves is conclusively proved by v. Kries, who found the dicrotic elevation in the femoral artery of an animal whose heart was replaced by a valveless bag.

All these facts, on the contrary, can be understood easily in the light of the theory that the sphygmographic curve gives the movements of the arterial wall, which movement is conditioned by the decreasing amount of blood in the artery, and the elastic vibrations of the wall around a variable position of equilibrium. In some cases the conditions of the problem are rather simple, and admit an analytic treatment, the results of which fit closely to the experimental facts. This part of the theory, however, has merely physiological interest, and therefore is discussed in a separate paper. It may be mentioned at this point that this theory of normal dicrotism is essentially identical with the theory of abnormal dicrotism as stated by Galen. He believed that the second beat of the pulsus bis feriens was due to an elastic vibration of the arterial wall. "Ex eodem genere sunt dicroti; nam arteria in occursu quasi repellitur, moxque redit.... Neque enim tum arteria contrahitur, sed quasi concuteretur, occidit; cuius delapsum a primae distentionis termino nulla dirimit manifesta quies, ut animadvertitur in contractione: sed simulatque attolli destitit, recidit atque ita paulisper vibrata, mox occurrit iterum."[80] Galen, however, is mistaken in his view, and in his observation that sometimes three or more pulse-beats may be felt with the finger. No form of the pulse is known where three or more beats may be felt for every heart-beat, and the actual tracings exclude the possibility of this observation for the pulsus bis feriens. The pulsus bis feriens is due to an increase of the frequency of the heart-beats. If the new pulse wave arrives before the vibrations of the arterial wall have had time to subside, the new wave and the already existing vibration may interfere in such a way as to produce this abnormal pulse form.

The form of a single wave of the sphygmographic curve may be influenced by changes in the following conditions:

(1) The pulse wave may have an initial form which cannot be represented by the schematic curve in Fig. 1. This may be due to an irregularity of the function of the ventricle. The action of the heart has an influence on the length of the waves, which length is determined by the rapidity of the heart-beats. This influence has been mentioned before. A change in the rapidity of the heart-beats has no great influence on the form of the catacrotic part of the curve so long as the impact of the new pulse wave does not arrive before the vibrations of the arterial wall have had time to subside.

(2) Differences of the elasticity of the arterial wall affect materially the form of the catacrotic part of the sphygmographic curve. It has also some influence on the height of the curves, because the amplitude of elastic vibrations depends on the elastic force for a given force of the shock. The degree of elasticity of the arterial wall is subjected to individual variations, and it depends in a given subject on the state of innervation of the wall.

(3) The surrounding tissues have a certain influence, since their resistance determines the friction opposing the vibration. This accounts for the fact that merely local conditions, such as a change of the position of the arm or the adjustment of the instrument, may change the form of the pulse curve. For instance, if the sphygmogram is taken from the a. radialis the instrument is placed between the styloid process of the radius and the tendon of the flexor carpi radialis. In the neighborhood of this place are two venae comites and a superficial branch of the median or radial vein. A change in the position of the arm will have a certain influence on the circulation in the veins, and influence the turgor of these vessels. Increased turgor increases the friction, and thus produces the different forms of the tracings.

(4) The changes of the turgor of the artery, moreover, cause a general rise or lowering of the curve. This symptom is essentially ambiguous for the turgor of the artery may be changed as well by an increase or decrease of the amount of blood pumped into the arterial system as by a decrease or increase of the amount of blood which passes through the capillaries.

The influences mentioned under (1) may be seen in tracings taken from cases of cardiac insufficiency, and have merely pathological interest. All the other influences, however, can be observed in the curves which are traced for psychological purposes. The changes in the general rise or fall of the curves are not so very hard to observe,[81] and for the observation of the rapidity of the heart-beats it is only necessary to trace a time-curve and count the number of beats or measure the length of every single beat. Also the changes of the height of the waves can easily be measured. This has been done conscientiously by several observers. It is by far harder to see the changes in the form of the catacrotic branch, and only a few keen observers have seen them. These changes of the pulse curve under the influence of feelings were proved as facts by experiments, but their interpretation was doubtful. With the exception of the rapidity of the heart-beats, which could easily be observed in some other way, all the symptoms of the influence of feelings on circulation are ambiguous. A difference in height of the single waves may be due to a change in the amount of blood which is pumped into the artery, but it also may be due to a change in the amplitude of the vibrations of the artery. The form of the catacrotic part of the sphygmographic curve may be changed by a different state of innervation of the arterial wall, but it also may be due to an increase or decrease of the friction of the surrounding tissues. The general rise or fall of the curve may indicate a change in the amount of blood which leaves the left ventricle, but it also may indicate a change in the amount of the capillary outflow.

The problem, nevertheless, is fully determined, and a solution is suggested by the constructions in Fig. 5. The form of the resulting movement depends, first, on the length of the line AB, secondly, on the length of the line AE, and thirdly, on the nature of the elastic movement. An elastic movement is determined if three constants are known, one of which is the amplitude, the second the friction, and the third the elasticity. Only AB can be measured directly, and there remain four unknown quantities to be determined. Four measurements must be sufficient for this purpose. It is obvious, however, that not any four measurements will do, but a method can be devised by which it is possible to determine each one of these four quantities. The problem can be solved in every case provided that the sphygmogram is trustworthy enough to justify the work. The length AB is proportional to the time of one heart-beat, and the length of the line AE is proportional to the amount of blood pumped into the arteries. The successful analysis of the pulse curves, therefore, shows changes of the action of the heart and makes it possible to distinguish them from the changes at the periphery.

Besides the length of the heart-beats there are invariably these four quantities which must be determined by the analysis of the pulse curves: Amount of incoming blood, amount of outflowing blood, elasticity of the artery, and friction of the tissues. These quantities depend on the action of the heart, the peripheral resistance, and the state of innervation of the artery. It is not possible to discuss here the bearing of this theory and of the facts which may be connected with it, on the different views of the localization and operation of the centres which control these functions. Anatomical and physiological evidence, however, leaves no doubt that the function of the heart and the innervation of the arteries and capillaries are under the control of nervous centres. It may be supposed, therefore, that changes of the pulse curve like those due to the influence of feelings are the effect of the function of these centres. It is to be expected that the detailed analysis of the pulse curves may give some indications as to the nature of this influence, for it may be observed how the function of these centres changes under the influence of mental processes.

A complete analysis of the physiological accompaniments of a feeling process must give a description of the changes in the function of the heart and the system, besides a description or at least enumeration of the other changes which can be observed. By a number of such investigations material for a general theory of physiological accompaniments of feelings may be obtained, which would not be void of interest for the psychology of feelings. Such a theory must contain the answers to the following questions: (1) How do the physiological reactions depend on the sense-stimulus? (2) How many possible circulatory reactions are there? (3) What is the location and interdependence of the respective physiological centres? The first question cannot so far be answered in general, but it will be possible to give a general answer when a greater number of systematic investigations on the effect of sense-stimuli have been carried on. Papers like those of Mentz may settle the question for certain sense-stimuli. From the results which have been obtained so far it comes out clearly that the reaction does not depend merely on the nature of the stimulus, but that it depends largely on the psychical and physiological state of the subject. The answer to the second question may be given readily, but it seems advisable to give it in connection with an experimental investigation. It may be said, nevertheless, that the number of typical reactions is rather limited. The third problem, by its nature, cannot be definitely answered before the location of the respective centres is ascertained and their interdependence explained.

It is, finally, a merit of this theory of the pulse curves that it shows how the form of this curve may depend on central processes. The problem of the mysterious influence of mental processes is thus reduced to the analysis of merely physiological conditions. The theories on the nature of this influence are so numerous that they may well be called innumerable, and they vary from accepting a direct influence of ideas on the circulation to considering the body as a sounding-board which by every sensation is shaken in all its parts. Each one of these theories is also a theory of feelings, and a more or less exact description of these changes has been often taken for a descriptive psychology of feelings. The example of the sounding-board is taken from one of those papers which expound the theory that bodily changes follow directly on the perception, and that our sensation of these facts is the emotion. Every one of these bodily changes, whatsoever, is perceived, acutely or obscurely, the moment it occurs. This theory is defended by the argument that if we try to abstract from consciousness all the sensations of our bodily symptoms, we find we have nothing left behind. This argument, which may be found in almost every paper that deals with this theory, is remarkable, because it sometimes is referred to processes of every description, and thus comes into contradiction with psychophysical parallelism which excludes the acceptance of psychical states which have no physical correlate. This theory, as will have been noticed, is the theory of feelings expounded by James, Lange, Ribot, and others. It is widely accepted, and may be found also in books of popular or semi-popular nature. Two observations must be made against this view:

First, a perception of a bodily change which is felt in the moment the change occurs exists only in the theory, every real process needing a certain time. This point of the theory may be improved by admitting that the afferent process lasts as long as any other of the physiological processes of this kind. Either assumption, however, is contradicted by the experimental evidence supplied by Lehmann that the physiological changes occur after the beginning of an emotional state.

Secondly, if the theory refers only to those bodily changes which we know, it certainly is not true, for emotional states are sometimes observed without it being possible to find with modern instruments any bodily accompaniments. If the theory refers to bodily changes of every description, it is certainly true, or, better, it is beyond all attack because it becomes identical with psychophysical parallelism. In this general form this theory of feelings is as good as no theory at all, because it refers to mental states of every description.[82]

This conception of emotional states of mind as perceptions of bodily sensations would hardly have been promulgated, if the authors had tried to base it on experiments performed in the laboratory. An emotion but not the feeling-tone of a simple sensation may be mistaken for the sum of bodily sensations. It is, furthermore, remarkable that the promoters of this theory do not make a clear distinction between sensation and feeling. They introduce an emotional element by calling the perception of bodily changes a feeling of these changes. Only in this way do they succeed in building up emotional states of mind out of elements which are seemingly sensational. This does not succeed if the word feeling is replaced by the word sensation. The failure of this theory is due to two facts, first to the starting from a philosophical doctrine, and second to the lack of a precise distinction between feeling and sensation. It cannot be doubted after the above discussion how a definition of this difference may be given which holds for every empirical investigation.

A sense-stimulus produces a complex of nervous and central processes. Among these is a certain group of processes which manifest themselves by changing the innervation of the heart, the blood-vessels, the lungs, and certain muscles. Another group is formed by those nervous and central processes which are more or less immediate effects of the sense-stimulation. The first group of processes is referred subjectively to an emotional state of mind, and the second to a cognitive process; the first group of processes is the physiological accompaniment of feelings, the second that of sensations. The relative independence of the first group from the second group is warranted by the fact that the same processes are observed as accompaniments of ideational processes. A strict limit between these two groups of processes can be drawn when the central processes are better known, because to the first group belong all those processes which are found to be accompaniments as well of sensational as of ideational processes. In different sensations the emotional process may be more or less marked, and in others the cognitive process may be prominent, but it seems that feelings are an invariable accompaniment of the sensation. This suggests the definition of feelings as psychic processes, the physiological accompaniment of which are central processes which depend largely on the state of the organism, and which manifest themselves by changes in the innervation of the heart, the blood-vessels, the lungs, and muscles. The impossibility of directly comparing the sensations of different subjects is recognized, and it is also impossible to compare feelings, because in either case we are dealing with psychic processes.

THE MUTUAL INFLUENCE OF FEELINGS

The basis of appreciation was a numerical scale[83] as follows:

• 1. Very disagreeable.
• 2. Disagreeable.
• 3. Slightly disagreeable.
• 4. Indifferent.
• 5. Slightly agreeable.
• 6. Agreeable.
• 7. Very agreeable.

The color series began with the one hundred thirty-six colors as put out by the Milton Bradley Co. This series consists of ninety pure spectrum colors, ten whites, blacks, and grays, and thirty-six broken spectrum colors. The colors were exposed at the back of a semicircular black-lined box for about two seconds. The subject was seated at a convenient distance, about three and a half feet, from the colors. In order to have a constant light, all experiments were conducted in a dark room with an electric light suspended over the subject's head. The whole series was used for ten times in order to get the range of judgments. Then twenty-eight colors, covering as fully as possible the range from 1 to 7, were selected for further experiment in combination.

At the same time a series of thirty-six touches, from velvet to sandpaper, was being employed as the colors were. From this number fourteen were finally selected.

Similarly, by using a reed box, with reeds ranging from 128 to 1024 vibrations per second and separated from each other by four vibrations, from a much larger series twenty-seven tone-combinations were finally selected.

Moreover, from time to time, each selected series was given alone; and on the basis of these readings, averaging from thirty to forty, the "standard" for each stimulus was made. Tables I to III give a brief description of the stimuli and also the "standards" for each of two subjects, F. and M.

TABLE I. COLORS