Open Access

JSON Export

{"created":"2022-01-31T12:57:49.932908+00:00","id":"lit28732","links":{},"metadata":{"alternative":"Studies from the Yale Psychological Laboratory","contributors":[{"name":"Seashore, C. E.","role":"author"}],"detailsRefDisplay":"Studies from the Yale Psychological Laboratory 4: 27-61","fulltext":[{"file":"p0027.txt","language":"en","ocr_en":"INFLUENCE OF THE RATE OF CHANGE UPON THE PER CEPTION OF DIFFERENCES IN PRESSURE AND WEIGHT.\nBY\nC. E. Seashore, Ph.D.\nAmong those who have experimentally advanced our psychological knowledge of the effects of very slow rates of change in the stimulation of the senses are Heinzmann1, Fratscher2, Preyer3 and Sedgwick4. More or less systematic investigations have been made on changes between the slowest perceptible and the instantaneous ones by Preyer5, Hall and Donaldson6 *, Hall and Motora\u2019, Scripture8, Stern9 and Stratton10.\nResearch on this subject has demonstrated several general facts: (i) The tendency of sensation to vary with the rate of stimulation is not primarily a peculiarity of any particular sense, because it is determined by general mental factors which enter into the perception of weak stimuli of all the senses in a similar manner. (2) The main value of knowledge of the laws here obtained does not lie in the acquaintance with the functioning of the particular sense organs that they furnish, but rather in that\n1\tHeinzmann, Ueber die Wirkung sehr allm\u00e4liger Aenderungen thermischer Reize auf die Empfindungsnerven, Archiv f. d. ges. Physiol. (Pfl\u00fcger), 1872 VI 222.\n2\tFratscher, Ueber continuirliche und langsame Nervenreizung, Jenaische Zeitschrift f. Naturwissenschaft, 1875 IX (n. F. II) 130.\n3\tPreyer, Die Empfindung als Function der Reiz\u00e4nderung, Zt. f. Psych, u. Physiol, der Sinn., 1894 VII 241.\n4\tSedgwick, On the variation of reflex excitability in the frog induced by changes of temperature, Stud, from the Biol. Lab., Johns Hopkins Univ., 1882, 385.\n5\tPreyer, Die Grenzen der Tonwahrnehmung, Jena 1876.\n6\tHall and Donaldson, Motor sensations on the skin, Mind, 1885 X 557.\n' Hall and Motora, Dermal sensitiveness to gradual pressure changes, Am. Jour. Psych., 1887 I 72.\n8Scrii>ture, On the method of minimum variation, Am. Jour. Psych., 1892 IV 577; Ueber die Aenderungsempfindlichkeit, Zt. f. Psych, u. Physiol, d. Sinn., 1893 VI 472.\n9\tStern, Die Wahrnehmung von Helligkeitsver\u00e4nderungen, Zt. f. Psych, u. Physiol, d. Sinn., 1894 VII 249 and 395; Die Wahrnehmung von Bewegungen vermittelst des Auges, same volume p. 321, and Die Wahrnehmung von Tonver\u00e4nderungen, same, 1896 XI i.\n10\tStratton, Ueber die Wahrnehmung von Dr\u00fcck\u00e4nderungen bei verschiedenen Geschwindigkeiten, Phil. Stud., 1896 XII 525.\n27","page":27},{"file":"p0028.txt","language":"en","ocr_en":"28\nC. E. Seashore,\nthey become accessory means by which we may investigate the involved central conditions of sensational, emotional and voluntary reaction. (3) There are probably three stages of this time influence for all senses, (a) The threshold for the preception of instantaneous change is generally lower than the threshold for any gradual change, (b) A gradual change may be so slow that it cannot be perceived in any period during which it may be studied, even though the compared stimulus may be raised to several times the intensity of the standard and in some cases produce fatal results. (c) The variation of sensitiveness in the region between these two extremes of change depends upon several complex central and peripheral conditions with reference to which it must be defined.\nThe present report is upon experiments in two different senses, pressure and muscle sense, in which gradual changes are compared with each other and with instantaneous change. The investigation was in progress in the Yale Psychological Laboratory from October, 1895, to February, 1897.\nI. Effect of the rate of change upon the perception of increase\nIN PRESSURE.\nThe object of this first series of experiments is to determine the law of sensitiveness to increase in pressure when the increase is made at various representative rates of gradual change. The problem requires the following experimental conditions : ( 1 ) an initial, standard pressure over a definite area ; (2) a uniform increase in this at several desired rates without any other disturbance of the original pressure; and (3) the means of varying the rate and amount of increase. After considerable preliminary work I found that these conditions were best fulfilled by a hydrostatic balance. The one used is constructed on the principle that a solid body, gradually immersed in a liquid, loses weight in proportion to the displacement of the liquid. It will be described in parts for convenient reference in the successive series of experiments.\nApparatus A.\ni. The graduated tube. This consists essentially of a vertical glass tube, continued at the lower end by a U-shaped metal tube whicM is terminated by metal nozzles. The vertical tube has an inside diameter of 42ram and is 555\u201cm long. The curved tube has an inside diameter of 25111m an(j radjUs of its curvature is i2omm. There is an outlet at the lowest point of this tube through which the water may be conducted into","page":28},{"file":"p0029.txt","language":"en","ocr_en":"Influence of the rate of change.\n29-\nan escape tube by opening a pinch cock. The free end of the U-tube is adjusted for the insertion of nozzles to regulate the stream of water which shall pass. These are interchangeable brass cylinders inserted in the lower end of a rubber tube which leads from a reservoir of water. In the upper part they have an inside diameter of 6\"m, but through the lower end they have a smaller bore ; the five here used vary in a series according to the standard drill gauge numbers: 60, 45, 30, 15 and 1 with diameters of i.omm, 2.Tmn, 3.3\"\u201c, 4.6\"\"\" and 5.8'\"'\u201d respectively. Thus five different rates of flow may be obtained by using successive nozzles. The purpose of the U-tube is to break and quiet the stream. The straight tube carries a graduated scale of heights.\n2.\tThe balance. A very delicate balance is constructed of a steel rod, diameter 2mm, and length 4IO,,,\",, with the fulcrum at the middle. It is supported on knife-edge bearings and braced by a diamond shaped framework of fine steel wire. Light hooks are inserted at the two extremities to serve as means for the attachment of parts to be balanced.\n3.\tThe float. This is a metal tube suspended from one end of the balance beam inside of the graduated tube. It causes a displacement in the liquid as it is gradually immersed. It is smooth and uniform, in this series 8. imm in diameter and 450\"\"\u201c long and heavy enough to retain a steady vertical position in water. Its bottom ends in a tapering hard rubber point which reduces the friction and upward pressure of the stream.\n4.\tStimulus rod. The float is counterbalanced on the other end of the balance by a similar metal tube which carries the pressure point on its lower end. A tube is chosen because it gives rigidity to the pressure point ; it may also serve as a receptacle for weights. The point is a hard rubber cylinder, of 5\"\u201d\u201d diameter, whose edges are not rounded off but dulled by a light buffing.\n5.\tThe graduated scale. Readings of the pressure are made on a millimeter scale attached to the graduated tube. The zero point of the scale is at the surface of the water when the balance is in a horizontal position and the lower end of the float is immersed to a point just above the tapering end. Since the diameter of the float and the specific gravity of the water are known, the readings of the height of the column of water in millimeters are readily converted into grams of pressure as exerted by the pressure point.\n6.\tThe guide lever. The float end of the balance may be fixed rigidly at the point of equilibrium by means of a spring lever. The place to be stimulated can be brought within a definite distance of the pressure point. Releasing the lever releases the balance which transfers the standard","page":29},{"file":"p0030.txt","language":"en","ocr_en":"30\nC. E. Seashore,\nweight to the point pressed at a definite time and with a regulated momentum.\n7.\tThe inlet and outlet clamps. Ordinary pinch clamps are used for these purposes.\nIn order to test the rate of flow through a nozzle it is necessary to apply some time-measuring instrument to the apparatus. An electric key is fitted up which makes the circuit the moment the water begins to flow. It consists of one of the above pinch clamps furnished with an adjustable make-contact.\n8.\tThe fountain. The water reservoir is placed three meters above the outlet in order to secure an approximately constant flow even when there is some difference between the levels at the two extremities. The reservoir is a shallow vessel holding 160 liters of water the surface of which can without inconvenience be kept within \u00b1 50\"\"0 of a constant point. A rubber hose of 18\"\"\" inside diameter conducts the water in a vertical column to the apparatus. A piece of smaller and more flexible tube is used just above the nozzle where the inlet pinch clamp is applied.\n9.\tThe hand rest. This is a special support to be used when the outer surface of the index finger is to be experimented upon. It is so constructed that the index finger and the thumb may rest upon a support and the other fingers brace themselves firmly and comfortably so as to obtain perfect stillness of the index finger without interfering with its circulation. A wooden cylinder stands on a base board and carries on its top a hard rubber plate 2mm thick. The thumb and index finger rest upon this plate and the other fingers grasp the pillar below, while the forearm and side of the hand rest upon the base.\nExperiments.\nThe rate of change was varied in successive steps while the other experimental factors were kept constant. Five rates were selected such that the slowest was as slow as could generally be perceived upon the present standard and the fastest as fast as the present apparatus and method would admit. The other rates were taken between these two extremes so that the increase in pressure per second upon a standard of 5* by the respective rates was as follows: 0.18s, 1.10s, 2.85s, 4-^3*, and 6.63s. The standard, or initial pressure, of 5s was applied to a circular area 5\"\u201c in diameter on the outer side of the middle of the third phalanx of the right hand index finger. The hand support was so adjusted that the finger in position upon it came as near the stimulus point as it could without touching, or about 0.3\"\".\nThe observer and the experimenter sat on opposite sides of the table","page":30},{"file":"p0031.txt","language":"en","ocr_en":"Influence of the rate of change.\n31\nwith an opaque screen between them. The observer occupied a comfortable position with his finger on the hand support and kept his eyes closed during the trials. By the signal \u201cone\u201d he was warned to be ready ; after \u201c two \u201d the initial pressure was applied and about two seconds after this had been done \u201cthree\u201d signified the beginning of the increase in pressure. Further instructions to the observers were as follows : \u201cThe pressure may increase and it may not ; as soon as you are sure that it has increased, say \u2018 up \u2019 as promptly as possible. Make sure that you have the same degree of certainty in all trials.\u201d This standard of certainty was fixed by a few preliminary trials. If the observer thought that he had not kept the standard of certainty or had suffered any disturbance he was required to call at once for a repetition of the trial. No observer was allowed to see the experimenter\u2019s side of the apparatus until all the experiments were completed.\nTo estimate the distortion due to the order of sequence of the rates, they were taken in rotation in opposite orders by successive observers, and the experiments were begun at different steps in the series in a systematic manner. They were also taken in the double fatigue series, i. e., half the number of trials on each point were made ingoing through the series the first time and then the rest were made by repeating it in the reverse order. A brief rest was made at the middle of the experiment ; the whole lasted about one hour. The results for thirteen observers who tried this experiment are contained in Table I and are represented graphically in Figure 1.\n0.036\t0.220\nFig. i.\nThe horizontal axis in this figure is marked off into parts according to the y data in Table I, i. e., proportional to the part of the initial stimulus by which.the increase was made per second.\nThe lowest point marked on the curve to the left is 0.35\u00ae which is the increment when the change is made most slowly. The same over the dotted line to the right indicates how far the curve must drop as the rate increases to the instantaneous. This result is transferred from the third series of experiments.","page":31},{"file":"p0032.txt","language":"en","ocr_en":"32\nC. E. Seashore,\nThe law discernable in the table is this : Within the limits of the investigation the amount of the least perceptible increment rises with the increase in the rate, i. e., the faster the increase in pressure the larger is\nTable I.\nLeast perceptible increase in a pressure of five grams at different rales.\n\tI\t\tII\t\tIII\t\tIV\t\t\tV\na\t0.18\t\tI.IO\t\t2.85\t\t4-83\t\t6.63\t\n\u00df\t5 55\t\t0.91\t\t0-35\t\t0.21\t\t015\t\n7\t0.04\t\t0.22\t\t0-57\t\t0.97\t\t1-33\t\n\tA\td\tA\td\tA\td\tA\td\tA\td\nA. B.\t3-4\t1.9\t4-5\ti-9\t5-9\t2.6\t6.9\t2.7\t7-4\tI.I\nF. B.\t5.2\t1.6\t4.0\ti-9\t8.8\t2.6\t8.4\t1.6\t7-4\tI.I\nS. P.\t3-1\t2.1\t9.0\t4-8\t16.7\t3-6\t19.5\t5-7\tii.6\t4-9\nG. O.\t6-3\t2.4\t10.6\ti-5\t11.4\t4.6\tJ3-4\ti-5\t16.9\t2.6\nM. A.\t2.8\t1.4\t5-2\t0.7\t7.2\ti-9\t10.4\ti-9\t12.5\t2.8\nA. N.\t2.9\t0.9\t4.6\ti-3\t9.8\t2-3\t139\t3-9\t15.0\t3-7\nE. B.\ti-9\tI.I\t6.9\t3-1\tio-3\t6.7\t10.2\t2.0\t14.0\t2.6\nG. H.\t2.6\t0.9\t4.9\t1.6\t5-4\t0.9\t9.8\ti-3\t11.4\t4-4\nS. K.\t1.8\t0.9\t5-5\t1-4\t7.0\t1.0\t6.0\t1-4\t6.1\t0.9.\nA. H.\t1.8\t1.8\t6.1\t1.8\t6.4\t2.0\t8.9\t0.9\t9.0\t2.8-\nP. P.\t3.1\t3-4\t7.6\t4.0\t11.8\t4-5\t8.4\t4-3\t12.6\t3-8\nM. J.\t2.7\t3-7\t8.2\t3-0\t10.4\t2.1\tn.4\t2.1\t155\t4.2\nA. S.\t4.0\ti-5\t7-7\t3-7\t10.6\t2-3\t11.i\t31\t17.8\t2-9\nAverage\t3-2\t1.8\t6-5\t2-4\t9.4\t2.8\t10.6\t2.4\t12. i\t2.9\nTime\t17.7*\t\t5-9\u2019\t\t3-3s\t\t2.2*\t\t1.88\t\nThe unit of measurement is the gram.\nThe number of measurements in each case is \u00ab = io, of which the median1 is taken.\nInitials, the observers.\nRoman numerals, the different rates.\nA, the increment in grams.\nd, mean variation ; to find the mean variation for the series divide each of these by V k = 3.2.\na, number of grams of increase per second.\n\u00df, time to increase one gram, in seconds. y, part of the initial stimulus to which the increase amounts per second.\nTime, the time represented by the average increment for all observers at each rate.\nthe size of the increment which is just perceptible. These limits include those which we experience most in normal life. But, referring to the curve, near the two ends there must be deflections of the curve in\n1 In the present research I have used the median in all the experiments with the reaction method ; the average has been used with other methods. This is because, by the nature of the experiments, the variation is larger and there are more abnormal records by the former method. For the account of the median and its relation to the average see SCRIPTURE, On mean values for direct measurements, Stud. Vale Psych. Lab., 1894 XI I-","page":32},{"file":"p0033.txt","language":"en","ocr_en":"Influence of the rate of change.\n33\nopposite directions if it is to be extended, i. e., if extended on the left the curve must soon reach an almost vertical direction and the extension of the other end must bend in some way so as to eventually reach the point which represents the increment in instantaneous change. The law of such deflections must be determined by future investigation.\nThe experiment was repeated three times upon one observer, M. J., at intervals of two weeks, each time under similar circumstances. The judgments in the successive experiments were equally unbiased, except in so far as they were influenced by the pressure sensations. The final averages for the three experiments are as follows : Rate I, A 2.8g, do. ig ; Rate II, A 9.1s, d 1.6*; Rate III, A 11.3s, d i.6\u2018; Rate IV, A 13.1s, d 1.2s; Rate V, A 13.7s, d 1.5s. The results indicate that the rate influence is definite and persistent for these trials. This remarkable consistency, as well as the agreement of the thirteen observers above, can only be accounted for by assuming a definite time influence which corresponds to this variation. The general law here found has an extensive application to the whole sensory side of our mental experience and involves some disputed points. I will, therefore, give a brief critical estimate of the apparatus, the method and the conditions adopted. Much of what is here said applies, also, to the following series of experiments.\nCritical estimate.\ni. The apparatus. The only noticeable jarring of the apparatus came from the jarring of the building, and this disturbance was somewhat reduced by placing the experimenting table on sand bags. Furthermore the above experiments were made in the evening when it was comparatively quiet in the building. The variableness in the level of the source and the mouth of the stream caused some degree of inaccuracy. The two levels were placed so far apart vertically that the necessary variation in the level of the mouth would not materially affect the results beyond the degree of accuracy here required. The time for the water to rise through the first ioomm of the graduation tube was to the time to rise through the second ioomm above the zero of the same as 38 is to 39. This error is negligible because it affects all rates similarly and practically equal for proportional increments by the various rates. Since most of the measurements came within the limit of the first 2oomm the rates adopted were determined empirically for the average of this distance. This determination was made to a more than sufficient degree of accuracy by the graphic method of recording time.\nThe adjustment of the zero point of the column of water could be 3","page":33},{"file":"p0034.txt","language":"en","ocr_en":"34\nC. E. Seashore,\nmade with an accuracy of \u00b1 jS4\u201cm which equals about of a gram by displacement. There is again a possible error of =fc by the same unit, in the adjustment of the hand rest. Both these errors affect the standard pressure and can only affect the increment in an indirect way.\n2. Method. The main reasons for using the reaction method here are: (a) It is time-saving, which is a vital point when a long series of records must be taken, (h) It does not carry with it any suggestions as to what may be desired or expected. (<r) It is easier to interpret the results by this method than by any other. The results attained by this method need to be corroborated by other methods and that will be done, but I must here point out some of the sources of error of the method in its present application. The observer reacted with a vocal sound and I, as experimenter, made a sight reaction to that sound. This latter reaction is negligible because my eye followed the reading point as a point of regard. The sensory time in the observer\u2019s reaction should be counted to the record, i. e., the record should include the time fron, the beginning of the physical change to the moment it was perceived as a change, but not the motor element in the reaction. This I tried to eliminate directly in each trial by means of a subjective estimate. After some practice in sight readings of this kind I acquired some skill in estimating equivalents on the scale to the amounts lost by the observer\u2019s reaction at the various rates. I could hear the very beginning of the vocalization of the u in \u201cup.\u201d But the allowance to be made had to vary with the definiteness with which this sound was uttered. I was aware of the common illusion of motion as well as of the difficulty of perceiving two simultaneous impressions in different senses. I found difficulty only in the fastest rate, but even here the possible error would be small in comparison with the whole records and the corresponding mean variations. This method of eliminating the reaction-time is not fully satisfactory, but it is superior to the previous methods that have been proposed for similar purposes.\nDo we want the na\u00efve judgment of the unpracticed but skilled observer? or the discriminative and critical judgment of the experienced observer who is familiar with the conditions and elementary processes upon which his judgment is based and gives his decision after having taken all known factors into consideration ? While the latter is necessary in order to make a detailed analysis of the facts, the former is necessary for the establishment of the facts. Though it involves more than double the labor of the other method, I have made it a characteristic of this research that the facts shall be obtained as they appear without analysis in the common experience of the scientific mind. No one of my observers knew what to expect and they were expressly cautioned not","page":34},{"file":"p0035.txt","language":"en","ocr_en":"Influence of the rate of change.\t35\nto make any guesses with consequent Conscious or unconscious corrections.\nThe observer was directed to react when he could distinguish \u201cchange\u201d from \u201c no change.\u201d In this series I did not have any regular system of control experiments, i. e., trials in which no stimulus was applied. They were interspersed irregularly and by them I satisfied myself in regard to the necessary absence of illusion. The danger of illusion was emphasized in the preliminary trials. If the illusion took place then, the trial immediately following would show the trace of a reacting influence in overcautiousness, and consequent missing of the standard. In such cases the preliminary practice was continued until the observer had settled upon a normal standard.\n3. Conditions of the experiment. The standard pressure may seem light, but it is adapted to the place experimented upon. It was advisable to use a light stimulus which would not produce a deadening effect upon the nerves under long continued pressure. A light pressure on a small area produces a simpler and less disturbing sensation than a heavy pressure, which is liable to produce sensations of strain in parts not directly stimulated. It is also important to avoid pain.\nThe point upon the first finger was chosen for stimulation because it admits of being kept in a horizontal plane when the rest of the body is in a comfortable position. The place is of a good sensitiveness and free from hairs.\nThe slowest rate here used was determined by preliminary experiments in which smaller floats were used. It was then found that the change would not be perceived at all, in the slowest rates, even when it amounted to four or six times the original stimulus. The pressure sensation would either be entirely lost or else it would continue indefinitely to seem as but a fraction of the standard. 6.18* per second was found to be about the slowest rate per second by which the pressure might rise and still be perceptible every time under normal conditions. The fastest rate was taken within safe limits, and rates above that were reserved for a separate test.\nII. Variation in sensitiveness to change as depending on the\nDELAY OF THE STIMULUS.\nIn previous experiments with sight and sound1 I found that when a stimulus near the threshold is delayed beyond a certain time at which it\n'Seashore, Measurements of illusions and hallucinations in normal life, Stud. Yale Psych. Lab., 1895 III 36, 50.","page":35},{"file":"p0036.txt","language":"en","ocr_en":"36\nC. E. Seashore,\n\u00bb\nmay be expected, its threshold is lowered, and, within certain limits, this is proportional to the delay. If this law applies to pressure, a part of the facts established in the first series of experiments will be explained by it. The application of the law of delay to this particular case was tested in the following manner :\nApparatus A was used just as in the foregoing series but with only one rate, namely i.io5 increase per second, which is 0.22 per second of the standard. Counting from the end of 2s allowed for the perception of the original pressure, the increase was not begun until after a delay for the respective sets of trials as follows : I, os ; II, 5\u2019; III, io*; IV, 15s; V, 20s; and VI, 25*. In other respects the general methods and conditions were the same as in the first series. There was no suggestion by which the observer was led to expect the change to be felt at any definite time. He was not aware of the delay ; he only knew that he would feel a gradual change and that he should begin to look for it at the given signal. Nothing was stated as to when it would begin physically. He learned, however, from the preliminary trials that it would take different time-intervals for the change to become perceptible. Hence this is different from the cases of suggestion in which the observer is led to expect the stimulus at a definite moment. Here the conditions of expectation and general preparation were similar to those in the foregoing series. It is a case of suggestion that works through the variation of time-influence in ordinary perception.\nTable II.\nVariation in sensitiveness as resulting from delay of the stimulus. o'\t5'\t10\u00bb\t15\u2019\t20\u00bb\t( 25'\nA d A d A\nA. H.\t7.6\t3-1\t6.7\t2.2\t4.0\nP. D.\t8-3\t3-4\t6.0\t1.6\t4-7\nP. P.\t13.2\t4-5\t51\t1.2\t5-1\nA. B.\tS-4\t1.4\t3-5\t*\u20223\t2.9\nD. S.\t6.6\t3-8\t4.1\t2.8\t5-i\nJ. M.\t5.8\t2.1\t4-3\t1.1\t2.9\nA. S.\t20.0 .\t3-0\t17.1\t2.9\t12.9\nAve.\t9.6\t3-0\t6.7\ti-9\tS-4\nThe unit of measurement is the gram. The number of measurements in each case is \u00ab=10, of which the median is taken.\nThe delay of the stimulus is indicated in seconds at the head of each column.\nd\tA\td\tA\td\tA\td\n2.0\t3-8\t2.2\t3-6\t2.0\t4-1\t1-4\n0.9\t4.0\t1.6\t. 2.6\t1.2\t3-4\t1.4\n2-5\t5.8\t1-9\t3-2\t1.2\t3-5\t0.9\n0.6\t2.1\t0.7\t2.2\t0.6\t1.9\t0.4\nI-S\t5-1\t2.1\t3-1\t2.2\t4.1\t\u00ef-3\n1.2\t2.7\t0.4\t2.7\t0.9\t2.9\t0.5\n3-o\tji-9\t4-1\t13-4\t4-3\t13-1\t3-7\n1-7\t5-\u00ef\ti-9\t4.6\t1.8\t4-7\ti-4\nA, the threshold increment. d, mean variation ; the mean variation for the series is found by dividing by y/n \u2014 3-2-","page":36},{"file":"p0037.txt","language":"en","ocr_en":"Influence of the rate of change.\n37\nThe trials were made in the double fatigue order. Surprise and the \u25a0 disturbance of abrupt transitions were avoided. The control trials were used freely in the preliminary trials but not during the experiment. Table II contains a summary of ten trials on each point by each of seven observers. The figures give the medians, and the average of these is taken for the final summary. There is a remarkable uniformity in the results. The abnormal record of the last observer is accounted for by the fact that his hand was callous.\nThe table shows that the threshold decreases, i. e., the sensitiveness increases, as the delay is extended. This law is most noticeable in the first five or ten seconds and seems to extend only to about twenty seconds.\nThe results may best be interpreted by means of the comparison in Figure 2. The short curve represents the results of the first series of ex-\n\u2022 \u00bb \u2022 \u2014 CO\nFig. 2.\nperiments, showing the relation of the threshold to the time which it took to produce it. The number of seconds is laid off as the abscissa and the number of grams as the ordinate. The longer curve shows the same for the present series. Within the limits of time occupied in the first series (17.7 sec.) the threshold of difference is lowered as the time increases. Beyond that limit the delay does not seem to have any power to lessen the increment. The rate, which is actually the same in both cases, requires the largest increment in the series in which it occupies an extreme point.\nA great part of this variation may be explained by the fact that the longer the time the greater the expectancy will be and the summation of suggestive elements will be on a constant increase. The smaller signs","page":37},{"file":"p0038.txt","language":"en","ocr_en":"3\u00ab\nC. JE. Seashore,\n' of change come directly into the focus of attention, a greater number of them will be noticed and those noticed will be magnified. This implies that the variation of the threshold with the rate is not entirely because of the difference in impression that quick or slow rates of change make, but largely because of the different attitude of mental preparedness which is caused by different time relations. Though the observers tried to be equally attentive all the time there was a semi-conscious reinforcement of attention as time went on. When the signs of change came soon they had to compete with more rival sensational elements than if they came later. According to the conditions of the experiment there could scarcely be any surprise, but we may characterize the different states of mind by saying that in the fast rates the observer was open to conviction, while in the slower changes he more anxiously sought some imaged facts of assurance.\nThe fast rates are probably affected by contiguity with the slow, and likewise the reverse. The faster rates were, perhaps, at a disadvantage. It is probable that if the highest point in each curve had been established separately, without reference to any other rate, it would have been lower in both cases. Yet those would be entirely different conditions. If one rate, or time of change, is taken separately, the time for the change to be felt will be known ; it will be envisaged more definitely ; the attention will be sharply focused at the expected moment of change and no attention-energy will be scattered as above. Both conditions are facts of ordinary experience and it would be interesting to compare them. One form of the latter condition will be taken up in the next series of experiments.\nFatigue makes a light weight or pressure feel lighter. This is true for 5s. How does that affect those curves or their continuation ? Does it have the effect of lessening the standard and thus making the increments proportionally greater during the time extension ? Or, does it work in the opposite way so that the amount of the standard, plus the increment, is constantly lessened, necessitating a longer time to make the pressure feel heavier than at first? Both are true, i. e., there is a certain limit at which the amount lost by fatigue is just equal to the increase at a certain rate. This point lies beyond the lower end of the short curve. Then the weight would feel the same ad infinitum if there were no fluctuations in this limit. If the change is slower, however, several possibilities of sensation-changes are open, but it is not probable that any increase in pressure will be felt before pain sets in and the experiment for pressure must be discontinued. But if the rate of increase is faster than the rate of falling off by.fatigue, the standard will actually","page":38},{"file":"p0039.txt","language":"en","ocr_en":"Influence of the rate of change.\n39\nseem smaller as time is extended, as in the present series, and the increment, which is largely detected by feelings of change, will be felt larger in proportion.\nIII. Threshold for instantaneous increase in pressure.\nApparatus B.\nThe compound pressure balance here described was constructed primarily to serve as a means by which a standard pressure over a definite area might be applied and then increased at any moment without jarring the stimulus point or causing any disturbance except absolute increase in pressure. It consists of two coordinated balances, one of which is apparatus A with the only exception that a tapering rubber point ( T) is substituted for the original pressure point. The other balance consists of a light steel beam 300\"\"\" long on one side of a knife-edge bearing with the balancing mass on the other side. The pressure point is a cylinder inserted at the end of the beam which is supported from a frame by a pair of electro-magnets. This part of the frame is capable of minute adjustment in height. The frame carries a millimeter scale parallel to the beam. A sliding weight on the beam carries a pointer which indicates the position of the weight on the scale ; the change of weight at the pressure point (P) is proportional to the distance over which the weight is from the fulcrum. A light arm projects from the rear end of the cylindrical counterbalance for the purpose of affording leverage for the action of pressure from the other balance.\nWhen the two balances are brought together, the point T in balance A is brought to bear on the leverage arm in the other balance, and can be removed by the guide lever (apparatus A 6) without friction or jarring. By this combination of the two balances we secure a pressure point which keeps a rigid position, a means of retaining the standard pressure constant, and an instantaneous change or a gradual change in pressure at any desired rate.\nTo illustrate the case of instantaneous change, suppose that the balance is set with an initial pressure of 5\u00ae at P. This is supported by the electro-magnets at a definite distance (f/P'\"\") from the surface to be pressed upon. Then if we want to prepare to increase that by, e. g., ig after it has been applied, the weight on the beam is moved from the fulcrum until the scale indicates that the movement is equal to i8 at P. Then balance A is adjusted to press with a force of i8 by T. This i8 counterbalances the i8 just added to P and we have again the standard weight at","page":39},{"file":"p0040.txt","language":"en","ocr_en":"40\nC. E. Seashore,\nP. Opening the magnet circuit always places the standard pressure with the same momentum. To obtain the is increase upon the standard the point T is lifted vertically by a rapid movement of the guide lever (A 6). The I8 is of course transferred directly to the point P without any movement of the beam except through the extra indentation which the is causes on the skin.\nI have only had time to make one of the tests for which this apparatus is intended. This test consisted in finding the least perceptible increase when the change was made instantaneously. The standard pressure and the area were the same as before ; the general method was also the same. At a signal the observer got ready and about two seconds later the point P was applied by releasing the magnets. The increase was made about two seconds after this. The threshold was approached in both directions by steps with a constant difference of 0.2e. The observer simply stated whether he perceived the change or not. Control trials were interspersed irregularly. The average of the complete measurements on each of seven observers is given in Table III. The figures denote the smallest increment above which all were perceived.\nTable III.\nThreshold of instantaneous increase in a pressure of jt>.\nA. N.\tA 0.26\td 0.08\nM. A.\t0-34\t0.07\nJ. M.\t0-34\t0.08\nG. 0.\t0.17\t0.06\nA. S.\t0.64\t0.15\nP. D.\t0-34\t0.07\nP. P.\t0.36\t0.07\n\to-35\t0.09\nThe unit of measurement is the gram. The number of measurements in each case is n\u2014io, of which the median is taken. A, threshold increment.\nd, mean variation ; the mean variation for the series can be found by dividing this by i/n = 3.2.\nThe main value of these results lies in that they establish one end of the curve in Figure 1 as it would terminate if continued. The curve has to fall from its highest point 12.1s to this point 0.35s. This suggests an important problem, namely, in terms of the figure, what is the shape of the curve which must connect these two points? This will be answered for somewhat different conditions in a following series of ex-","page":40},{"file":"p0041.txt","language":"en","ocr_en":"Itifluence of the rate of change.\n41\nperiments. The present threshold meanwhile gives a standard in comparison with which we must interpret all the previous measurements on gradual change. Thus, the slowest gradual increase requires a threshold nine times as high, and the fastest a threshold thirty-five times as high as the threshold for instantaneous difference. In making this comparison \u201cinstantaneous\u201d must be taken in a relative sense (as it always must) according to the above details, and it must be remembered that these two thresholds were found by different methods.\nPsychologically the two judgments of gradual and instantaneous differentiation are not only made under totally different conditions of attention and expectation, but there is also an entirely different grouping of the sensations which form the basis for the discrimination.\nIV. Effect of the rate of change upon the perception of increase IN WEIGHT.\nApparatus C ; Experiments.\nAn apparatus is needed by means of which gradual changes in the weight of a lifted object can be made and measured. Apparatus C, which was constructed for this- purpose, consists of the following parts, used in conjunction with apparatus A :\n1.\tThe weight cell. This is a polished hard-rubber cylinder with a diameter of 2 imm and a height of 75mra. Its own weight, 25\u00ae, may be increased to different amounts by placing weights inside. A silk cord hangs from the bottom, by which weights may be attached.\n2.\tThe arm support. A board base is fixed in such a position that the arm from the elbow may rest upon it in a horizontal position. The weight cell stands on this in a position to be grasped comfortably. The cord from the cell runs vertically through a hole in the board to the stimulus end of balance A.\nApparatus C works on the same principle as apparatus A, the only difference being that the weight is exerted on a lifted cell instead of a pressing point. To illustrate by an example, let it be desired to obtain a measured gradual increase on a standard of 40\u00ae weight in a cell. The cell itself weighs 25\u00ae and we add 15\u00ae by weight placed inside of the stimulus rod, to whose top the cell is attached. Since this rod is counterbalanced by the float, the 15\u00ae are added to the weight of the cell when the observer lifts the cell, say 2mm from the base. At that point the standard will be reached. The water in the graduated tube is then allowed to rise at some definite rate and as the float is immersed weight is transferred to the cell.","page":41},{"file":"p0042.txt","language":"en","ocr_en":"42\nC. E. Seashore,\nThe same general method that was used in the study of pressure was here applied in the study of the so-called muscle sense, the word being used in its widest significance as including all the sensations by means of which we estimate lifted weight. The aim was to determine whether there is any law for muscle sense that corresponds to the law of rate influence that we have found for pressure, and, if so, to observe some of the relations between the two.\nTable IV.\nLeast perceptible increase in a lifted weight of forty grams at different rates.\n\tI\t\tn\t\tIII\t\t\na\t0.18\t\t1.10\t\t6.63\t\t\n\u00df\t5-55\t\t0.91\t\t0.15\t\t\ny\t0.005\t\t0.028\t\t0.166\t\t\n\tA\td\tA\td\tA\td\te\nJ. L.\t3-3\t1.1\tii-5\t4.0\t20.3\t2.1\t0\nv. s.\ti-7\t0.4\t6.1\t1.8\t9-7\t5.0\t0\nM. M.\t5-3\t1.4\t11.0\t2.4\t18.0\t4-4\t0\nj.r.\t12.0\t5-4\t14.9\t4.0\t21.9\t4.9\t0\nw. J.\t2.4\t1.0\t6.6\ti-S\t14.8\t4.9\t43\nA. S.\t4.0\ti-3\t12. i\t3-6\t17-5\t2.4\t43\nF. K.\t3-9\to-S\t12.4\t2.6\t17. i\t5-3\t0\nE. J.\t6.0\t2-3\t13-8\t4-9\t*3-5\t4-9\t0\nC. C.\t4.0\t2.4\t7-3\t3-4\t13-7\t4.2\t0\nAverage,\t4-7\tTi\t10.6\t3-1\t16.3\t4.2\t\nTime,\t26.1*\t\t9.6s\t\t2.5*\t\t\nThe unit of measurement is the gram. The number of measurements in each case is \u00ab=10, of which the median is taken. I, II, III, rates of increase. a, number of grams of increase per second.\n\u00df, time to increase I*.\ny, increase per second as a fraction of the initial stimulus.\nA, threshold increment. d, mean variation; to find the mean variation for the series divide by j/\u00ab = 3.2.\n\u00a3, percentage of the control trials in which illusions occurred.\nThe standard was taken as 40s, because that weight in the given cell is favorable for a distinct feeling of weight and does not cause noticeable fatigue very soon. Only three rates were employed and these were the same as I, II and V in the first series; they gave an increase of 0.188, i.iog and 6.63s per second, respectively.\nThe observer shut his eyes and grasped the cell at the middle between the thumb and the first two fingers in such a way that when he had raised","page":42},{"file":"p0043.txt","language":"en","ocr_en":"Influence of the rate of change.\n43\nthe cell 2\u201d\u2019\" from the base the side of the hand and the little finger rested on the base by their full length and the cell was held upright. The position was comfortable and could easily be retained for the required time. The observer grasped the cell as lightly as possible. As soon as the correct position had been secured the signal was given which meant that except in the control trials the physical change would begin in two seconds. The reaction was made by saying \u201cup\u201d and the results were read as before. Nine persons made the complete experiment, which consisted of ten trials on each rate, exclusive of the control trials. The summary is contained in Table IV.\nFrom one to five control trials were interspersed with each ten regular trials. The column giving the percentage of these trials which resulted in illusions gives an index to the reliability of the discrimination. All but two are perfectly reliable in that they have chosen the standard threshold so high that there is no danger of confusion. The high percentage of errors for the two must, however, not be interpreted to mean that such a percentage of the total number of trials may be considered as illusions, for in each case there were thirty trials in which the regular change was made and only seven in which the stimulus was withheld. All three illusions out of the five possible were in rate I for W.J., and for A.S. two out of four possible were in rate I and one out of two possible in rate II. The value of these percentages depends upon the relation between the number of control trials to the number of regular trials as well as upon the degree to which a trial with no sensation of change was expected. The effect of the rate influence is as marked for these observers as for the others and their mean variation is not excessive. This argues that, since they had nothing but the direct sensation to judge by, the real signs of change must have been present in a much larger proportion of trials than the above percentages would indicate. The error cannot be explained as due to the ordinary premature automatic reaction, for then the mean variation should have been much larger.\nThe figures in the table express the law that, for the three rates investigated, the threshold of perceptible increase in lifted weights is higher in fast rates than in slow rates. This accords with the law we found for pressure. The relation of the two will be discussed later.\nV. Threshold for instantaneous increase in weight.\nApparatus D ; Experiments.\nThis problem requires an apparatus by which the standard weight of a body may be increased instantaneously without causing any other dis-","page":43},{"file":"p0044.txt","language":"en","ocr_en":"44\nC. E. Seashore,\nturbance. This was accomplished by a compound weight balance. The cell ( Cj) and the arm rest (C2) are used as in the foregoing series. A weight-pan is suspended by a silk cord, branching out from the center of the bottom of the cell, 400'\"\"' below it. Midway between the cell and the weight-pan there is a light fibre balance beam 140\"\"\u201c long with its fulcrum at the middle. It is adjusted with one extremity perpendicular over the pan and attached to it by a cord which passes between the branching cords and is so fastened to the pan that changing the support of the pan from this cord to the branching cords, or the reverse, will not cause the pan to shake. At the other end a hook is suspended by a cord 200\"\"\" long. Upon this a series of gram weights was fitted to be hooked firmly. The hook is light\tand makes the weights conven-\niently interchangeable. At first the instantaneous increase was made by means of a weight acting over a pulley by a cord from the hook end of the balance beam. This weight was allowed to drop ioomm with the cord slackened 90\"\u201d\" ; thus the hook end of the balance would be elevated iomm at a definable rate. But the present method required that the change should be made noiselessly ; therefore at the risk of some irregularity, I simply raised the weight by pulling a cord perpendicularly from the hook as quickly as possible.\nThe plan of the apparatus may be explained better by an example. Suppose we want to get an increase of i* on 408 lifted by the cell. The cell weighs 25s and enough weight is added to the pan to make it weigh 15*. Since the 15e are supported by the cell its total weight is 40*. We then place the ig increment in the pan and counterbalance it by Is on the hook (including the weight of the hook); the two weights on the balance, therefore, do not act upon the weight of the cell. The observer grasps the cell and lifts it 2mm ; this makes the balance beam stand in a horizontal position. The grasp is made in such a manner that the whole side of the hand rests upon the support in a position that can easily be retained during a prolonged experiment. He is then lifting the standard weight, but when the hook-weight is raised suddenly its counterbalance is transferred to the support of the pan without any movement of the balance beam. I was able to elevate the hook-weight as quickly as it would start to fall by its own gravity, or faster ; therefore, the increment was transferred to the new support at the rate that it would assume by its own weight when beginning to fall, i. e., the increase was made by simply releasing the support of the required amount without imparting motion to the standard weight. By this method any instantaneous increase may be made without impact. It is evident that by reversing the action of he balance a decrease in weight may be made equally well.","page":44},{"file":"p0045.txt","language":"en","ocr_en":"Influence of the rate of change.\n45\nThis series of experiments was made on the same observers as the fourth series, and under similar circumstances, in order that the two sets of results might be comparable. They were made during the same period, this experiment being made alternately before and after the other. The point to be determined here was the threshold for the perception of a so-called instantaneous increase in the weight of a body lifted by the rested hand. A form of the method of minimum variation was employed. The steps varied by one gram each and were taken alternately ascending and descending the series. In ascending, the steps of change were continued until .the observer had perceived the increment correctly three times in succession ; and, in descending, the series of steps was begun with the highest one of the ascending series or by one above it if this series was low. At a signal the observer lifted the cell to the 2\"\u201d\" limit and as soon as steadiness was attained in this position another signal warned him to watch for an increase, to which he should react by saying \u201cup \u201d at the moment he perceived it. The change was made from two to six seconds after this signal. The observer\u2019s ability to react just at the right time was considered a criterion for his certainty and accuracy. The time of making the change was varied irregularly within this region of four seconds. If he reacted perceptibly before or after the change, due allowance for the reaction-time being made, the fact was recorded as an error or illusion. Of course, failure to react indicated that the stimulus was below the threshold. This time-criterion was chosen in preference to the method of control trials in order to make the experiment short. Instead of concentrating the attention on the moment two seconds after the signal, it here had to be scattered over a time of four seconds. This, presumably, tended to raise the threshold. Upon a definite inquiry each observer testified that he had not perceived any suggestion as to the moment of change except by direct feeling of change in weight.\nThe results are contained in Table V ; the initials of the observers will aid in the comparison of the individual records here with those in Table IV.\nThe A value marks an arbitrary limit. With careful observers it may be considered a pretty safe limit, denoting the point above which we may expect to find all increments perceptible under similar conditions. The A value must be interpreted with reference to the mean variations and the figures in the first sections of the table. According to the table 4. f would be perceptible ahout seven times out of ten, or 70 per cent, of the trials. This percentage is brought down so low because it is an average for different observers and not for successive trials on the same observer. The last column gives the number of times each observer reacted wrongly.","page":45},{"file":"p0046.txt","language":"en","ocr_en":"46\nC. E. Seashore,\nIn comparing this and the preceding table we notice the striking coincidence that the threshold for instantaneous change is the same as the threshold for the slowest gradual change. Between those two points the curve changes to nearly four times that height (see Figure 3). Here the relation between the instantaneous and the very slow increments is very different from the corresponding relations for pressure, where the instantaneous increment is only about one-ninth of the slowest.\nTable V.\nThreshold of instantaneous increase in a lifted weight of 409.\n\tI\u00ab\t28\t38\t4'\t58\t68\t78\tA\td\te\nJ. L.\tI\t4\t6\t4\t8\t10\t\t4.6\t1.1\tI\nV. S.\tO\t0\t6\t8\t10\t\t\t3-6\t0.7\t0\nM. M.\tO\t0\ti\t7\t10\t\t\t4.2\t0.4\t0\nJ. P.\tO\t0\t0\t2\t4\t10\t\t5-6\t0.5\t0\nW. J.\tI\t7\t5\t9\t10\t\t\t3-4\t0.8\t3\nA. S.\tO\t0\t0\t2\t6\t5\t10\t6.2\t0.8\t2\nF. K.\tO\t0\ti\t2\t9\t10\t\t4-9\t0.4\t0\nE. J.\tO\t0\t0\t3\t8\t10\t\tS.o\t0.4\t0\nC. C.\tI\t2\ti\t4\t8\t10\t\t4.9\to-5\t5\n\to-3\ti-4\t2.2\t4.6\tJTi\t9-4\t10.0\t4-7\t0.6\t\nThe unit of measurement is the gram.\nThe number of complete determinations of the threshold for each observer is n = 10.\nThe increments are denoted by the numbers at the heads of the columns.\nThe numbers below these show how many times, out of ten possible, each was perceived.\nA, the average increment above which the next two are correctly perceived, hence the threshold.\nd, mean variation ; the mean variation for the series is found by dividing each of these by (/V =3-2.\nc, total number of errors.\nExperiments were made upon six observers to study the effect of delay of the stimulus. Instead of devoting a special section to them I will make a brief statement here.\nUsing the identical apparatus described above, I found how the threshold varied if the stimulus was withheld after the signal for the number of seconds that the averages in series IV indicate, namely, 2.5\", 9.6* and 26.1s, i. e., the time it required to perceive the change at the three rates of gradual increase. Ten trials were made on each of the four increments : 3g, 4*, 5* and 6s. The results may be stated in a general way in terms of the percentage of those increments which were perceived. These are: for 2.5s, 63\u00b0f0 \\ for 9.6s, 69%; and for 26.1% 67%. The difference is not large enough to indicate any tendency toward a systematic variation.\nThe elements of fatigue and distribution of attention spoken of in the","page":46},{"file":"p0047.txt","language":"en","ocr_en":"Influence of the rate of change.\n47\nsecond series seem to counterbalance each other here. Thus after the 26. i\" delay the standard seems lighter on account of the fatigue ; but an instantaneous increase at that point will not be affected by fatigue, and will, therefore, appear larger in proportion to the standard at that point than at any previous point. On the other hand, at the end of 2.5* there is no noticeable effect of fatigue, but the attention is not yet so strongly focused.\nThe present experiments should be compared with those in series II. Here the change was instantaneous, there gradual. The difference in the results points to the fact that the methods of perceiving change are entirely different in the two cases.\nVI. Verification of series I and IV.\nTo increase the data found in series IV and to verify the law expressed in'the results of series I and IV by means of a different method, the latter series of experiments was repeated by a form of the method of minimum variation. The only alteration necessary in apparatus C was to insert a scale reading in grams instead of the millimeter scale on the graduated tube. The steps of increase in the experiment differed by two grams each, running from 2s to 208.\nI tried different steps at random and considered a determination complete when I had found at least three consecutive steps in which the increment had been correctly perceived and all the steps below this had been tried. I was fully aware of the influence this procedure has upon the mean variation, but it secured a good distribution of attention and served as a sort of control method in that the observer had no means of knowing whether the increment should be a large or a small one as he would if the threshold had been approached by steps taken in regular order. If in the large steps the observer was sure that he felt an increase before the full amount had been reached, he was allowed to signify this in order to save time, but such a reaction was recorded as if only the whole increment had been perceived. No steps higher than 20* were tried. Five complete determinations were made for each rate by each observer.\nAt the signal the observer lifted the 408 cell to the standard position, and at another signal he began to watch for an increase in weight which began two seconds later except in the control trials. A final signal was given at the end of the increment and upon this the observer had to reply immediately by one of three answers, namely: \u201cchange,\u201d \u201cno change,\u201d or \u201cuncertain.\u201d","page":47},{"file":"p0048.txt","language":"en","ocr_en":"48\nC. E. Seashore,\nThe method of manipulating the apparatus may be understood from the explanation in series IV ; the main difference was that here definite amounts of increase were produced and the observer stated whether he perceived them or not, while there the change continued until he reacted.\nThe possible inaccuracy in reaching the exact increment in the fastest rate was \u00b1 0.2s. If a larger error than that was made the trial was repeated. The variation by unsteadiness of the observer\u2019s hand may introduce a possible error of \u00b1 0.2*. These are the only two marked sources of error. The first affects only the fastest rate.\nThe results for the observers are contained in Table VI. The first part of the table shows the numbers of R (change perceived) and U (uncertain) answers that were given out of five trials on each increment. At the right hand end of each line I have omitted all but the first \u201c5\u201d when this is the first of three successive fives. The R and the U trials are recorded separately and the reader may distribute the U trials as he thinks best. A general expression of the results may be gotten by a study of the R trials alone.\nTable VI.\nThreshold of increase in a lifted weight of 40e.\nM. M. .\nF. C..\nN. H.\nB. L.\n\t\t2\u00ab\t4g\t6\u00ab\t8\u00ab\t10\u00ab\nIj\trR\t3\t5\t\t\t\n\ttu\t2\t\t\t\t\n11 j\tfR\t\ti\t3\t4\t4\n\ttu\t\t2\t2\t\t\nIII J\trR\t\t\ti\t3\t3\n1\ttu\ti\t3\t2\t\ti\nI <\tfR\t3\t5\t3\t5\t\n\ttu\t\t\t\t\t\nnJ\t[R\t\ti\t5\t\t\n1\ttu\t\t2\t\t\t\nHI J\t[R\t\ti\t2\t3\t5\n1\ttu\t\ti\t\t2\t\n\u25a01\tfR tu\t4\t4\t4\t5\t\n<*)\t;r tu\t\t2\t5\tS\t4\n1\tr R\t\t\t2\t3\ts\n1111\ttu\t\t\t\ti\t\n\u25a01\tR\tI\t3\t5\t\t\n\tu\ti\t\t\t\t\n\u25a0\u25a0I\t\u2019 R\tI\t2\t\t4\t4\n\t. U\t\t\t3\tI\t\n111 ]\t' R\t\t\ti\t3\t4\n\t;u\t\t\t\t\t\n12\u00ab\tI48\t16\u00ab\tl8*\t20\u00ab\tA 2.8\ta 1.0\n5\t\t7.2\t1.0\n5\t\t10.4\t1.2\n\t\t5-4\t2.2\n\t\t5-6\t0.6\n\t\t8.4\t1.2\n\t\t4.0\t2.4\n5\t\t5-2\t1.0\n\t\t8.0\t0.8\n\t\t4-4\t1.2\n4\t5\t9.2\ti-9\ni\t5\t12.4\t2.6\n2","page":48},{"file":"p0049.txt","language":"en","ocr_en":"Influence of the rate of change.\n49\nJ. r.\nW. J. '\nJ. R..\n\\V. H. \u25a0\nB. H. \u25a0\nR. S.-\nAv.\nofR I\n\u25a0f \"! ra { \u25a0I \u00bb!\n\u25a0{\n\"{\n\"I\n\u25a0\u00bb!\n\u25a01\n\u00bb!\nra 1\n\u25a0!\n\u00bb!\n\u00bb\u25a0!\ni\nii\nhi\nR\nU\nR\nU\nR\nU\nR\nU\nR\nU\nR\nU\nR\nU\nR\nU\nR\nU\nR\nU\nR\nU\nR\nU\nR\nU\nR\nU\nR\nU\nR\nU\nR\nU\nR\nU\n\tI\t3\t3\t5\t\t\t\t\t\t7.6\t1.9\n\t\t\t3\t3\t4\t5\t\t\t\t10.0\t2.4\n\t\t\t\t2\t2\ti\ti\t3\t4\t16.8\t3-0\n\t\t\t\ti\t\tI\t\t\t\t\t\n2\t3\t2\t5\t\t\t\t\t\t\t6.4\t1.9\n\t\ti\t\t\t\t\t\t\t\t\t\n\t\t2\t4\t4\t4\t5\t\t\t\t8-4\t2-4\n\t\ti\t\t\t\t\t\t\t\t\t\n\t\t\t2\t3\t4\t4\t4\t4\t4\t12.0\t3-2\n\t\ti\ti\ti\t\tI\t\t\t\t\t\n2\t2\t4\t4\t4\tS\t\t\t\t\t6.4\t2.2\n2\t\t\ti\t\t\t\t\t\t\t\t\n\t3\t4\t3\t4\t5\t\t\t\t\t8-4\t2-4\n\t\t\tI\t\t\t\t\t\t\t\t\nI\ti\tI\t2\t3\t3\t4\t4\t5\t\t12.0\t3-2\n\t\t2\t\t\t\t\t\t\t\t\t\n2\t4\t5\t\t\t\t\t\t\t\t4.0\t0.8\n\t2\t\t\t\t\t\t\t\t\t\t\n\t4\t4\t5\t\t\t\t\t\t\t5-4\ti-5\n\ti\tI\t\t\t\t\t\t\t\t\t\nI\t\t4\t4\t4\t5\t\t\t\t\t7.6\ti-9\n\t2\ti\t\t\t\t\t\t\t\t\t\n3\t4\t5\t\t\t\t\t\t\t\t3-6\ti-3\ni\t3\t5\t2\t2\t5\t\t\t\t\t10.0\t2.4\n\t\t\t\tI\t\t\t\t\t\t\t\n\t\ti\t4\t3\ti\t5\t\t\t\t13.2\t1-3\n5\t3\t4\ts\t\t\t\t\t\t\t3-6\ti-9\n\tI\tI\t\t\t\t\t\t\t\t\t\n\t4\tI\t5\t4\tS\t\t\t\t\t8.4\tJ-4\n\t\tI\t\t\t\t\t\t\t\t\t\n\t\tI\t2\t4\t4\tS\t\t\t\tXI.2\ti-4\n\ti\t3\t\t\t\t\t\t\t\t\t\n50\t68\t80\t94\t98\t100\t\t\t\t\t4-8\ti-7\n4\t40\t58\t80\t78\t94\t100\t\t\t\t7.6\ti-7\n4\t4\t26\t52\t72\t70\t78\t88\t94\t96\tII.2\t2.o'\nThe unit of measurement is the gram. Number of trials on each point, \u00ab = io. Roman numerals give the rates :\nI,\t0.18\u00ab per sec.\nII,\tX. to* \u201c \u201c\nIII,\t6.638 \u201c \u201c\nR, number of times the reply \u201c change \u201d \u25a0was given in five trials.\nhi, number of times the reply \u201c uncertain \u201d was given in five trials.\nThe difference between 5 and R -f- U will give the number of replies \u201cno change.\u201d\nThe numbers at the head of the columns denote the increments.\nJ, the threshold above which it is probable that all increments would be perceived.\nd, mean variation of J ; to find the mean variation for the series divide by y/~n = 2.2.\n4","page":49},{"file":"p0050.txt","language":"en","ocr_en":"5\u00b0\nC. E. Seashore,\nI have here assumed a standard condition which gives a threshold of the same degree of probability in all rates. This is obtained by the conditions upon which the figures of the \u00e0 column in the second part of the table are based, i. e., the first of these consecutive increments that has been correctly perceived is taken for the threshold. With this as a standard, we may compare the rates with each other and, with certain precautions, the general results with those obtained by the reaction method.\nThere is a remarkable uniformity in the results, considering the delicacy\nFig. 3.\nA,\tResults in Table IV.\nB,\tResults in Table VII.\nC,\tResults in Table VI.\nThe horizontal axis is divided into parts proportioned to the number of grams of in-\ncrease per second, or, which is the same, to the part per second of the initial stimulus by which the change was made. The increments are laid off in grams as ordinates.","page":50},{"file":"p0051.txt","language":"en","ocr_en":"Influence of the rate of change.\nSi\nand difficulty of this discrimination and the lack of practice. Bearing in mind the differences in the conditions of these experiments and those\nTable VII.\nfive experiments on the author under conditions similar to those in the preceding table.\n\t\t\t4g\t6*\t\u00a38\t.ICS\t128\t148\t168\tiS\u00ab\tA\td\n\t\u2022 *1,\tR\tI\t2\t4\t5\t\t\t\t\t5-6\t0.8\n\t\t1 u\tI\t2\ti\t\t\t\t\t\t\t\nA\tII\tR\t\t\t\t3\t4\t5\t\t\t10.8\ti-7\n\t\tlu\tI\t\t4\ti\ti\t\t\t\t\t\n\tIII,\tfR '\t\t\t\ti\t3\tS\ti\t5\t15.6\t1.8\n\t\tlu\t\t\tI\ti\ti\t0\t2\t\t\t\n\tI\tfR\ti\t3\t5\t4\t4\tS\t\t\t8.0\t2-4\n\t\tlu\tI\t\t\ti\t\t\t\t\t\t\nB\tII,\trR\t\t\tI\t2\t3\t5\t\t\til.6\t1-4\n\t\tlu\t\tI\t3\t2\t\t\t\t\t\t\n\tIII\tfR\t\t\t\t\t2\t3\t5\t\t14.4\t1-7\n\t\tlu\t\tI\t3\t3\t3\t2\t\t\t\t\n\tI,\tfR\t\t4\t5\t\t\t\t\t\t6.4\t0.6\n\t\tlu\t4\tI\t\t\t\t\t\t\t\t\nC\tII,\tfR\t\t\t3\tS\t\t\t\t\t8.8\t1.0\n\t\tlu\tI\t2\ti\t\t\t\t\t\t\t\n\tIII,\tfR\t\t\t2\t2\t5\t\t\t\t11.6\t2.0\n\t\tlu\t\t\t2\t2\t\t\t\t\t\t\n\tI\tfR\tI\t4\t4\t5\t\t\t\t\t6.0\t0.4\n\t\tlu\t2\tI\t\t\t\t\t\t\t\t\nD\tII,\tfR\t\t\t3\t5\t\t\t\t\t8.8\t1.0\n\t\tlu\t\t2\ti\t\t\t\t\t\t\t\n\tIII\tfR\t\tI\t3\t4\t5\t\t\t\t9.2\t1.8\n\t\tlu\t\t\t\t\t\t\t\t\t\t\n\t\tfR\t3\t4\t4\t5\t\t\t\t\t5-6\ti-9\n\t\tlu\ti\tI\ti\t\t\t\t\t\t\t\nE-\tII,\tfR\t\t5\t\t\t\t\t\t\t6.0\t0.0\n\t\tlu\ti\t\t\t\t\t\t\t\t\t\n\thi-!\tR\t\t\t3\t5\t\t\t\t\t8.8\t1.0\n\t_\tlu\t\ti\t\t\t\t\t\t\t\t\nAve.\t1\t\t24\t68\t88\t96\t96\t100\t\t\t6-3\t0.7\n% \u25a0\th\t\t\t12\t8o\t88\t100\t\t\t\t9.2\t1.6\nof R\thi\t\t\t4\t32\t72\t80\t92\t84\t100\t11.9\t2-5\nCapitals denote the successive experiments.\nOther notation same as in Table VI.\nin series IV we may obtain some general conclusions from this comparison :\ni- The law of the relation of the increment to the rate of increase established by the previous method in series I and IV is supported and","page":51},{"file":"p0052.txt","language":"en","ocr_en":"52\nC. E. Seashore,\nproved beyond the possibility of a doubt. This applies to series I only indirectly, but the general agreement between the facts and conditions of series I and series IV justifies the deduction in regard to the qualitative statement of the law.\n2.\tThe slowest rate requires the same increment, within o.i8, by this as by the previous method. By the present method the other two rates require smaller increments, somewhat in proportion to the rates.\n3.\tThe proportional differences by the two methods do not point to any notable error in the previous method, but are a good expression for the difference in the mental attitude in the two methods.\nThe foregoing experiment was repeated five times upon the writer during as many successive days. The conditions differ in that I knew just what was going on physically except in one respect, namely, that I did not know whether the trial would be a regular or a control trial. In this series I do not record the result of the control trials because I set my standard so high that I made practically no error. The following conclusions may be drawn from the summary in Table VII.\n1.\tThe general direction of the rate-influence is the same as that found for the other observers.\n2.\tThere is a noticeable decrease in the rate-influence for the fastest rate during the progress of the experiment.\n3.\tA reasonable expression for the difference in mental attitude due to my knowledge of the rate of change is found in the amount by which my threshold differs from that of the other observers.\nVII. Effect of the rate of change upon the perception of very\nRAPID INCREASE IN WEIGHT.\nApparatus E.\nVery rapid rates of change being desired here, it was necessary to use a float of larger diameter, or nozzles allowing a more rapid stream to flow,., than had hitherto been used. The former alternative was adopted, but: with this alternative a change in the mechanism of the rest of the apparatus also became necessary. Apparatus C required that the operator should stop the increase when it had reached the desired height. The rates of change approaching the instantaneous, here required, could only be obtained by an automatic determination of the amount of increase by the apparatus itself. This was accomplished as follows :\nApparatus C was used in all its parts except three, namely : the float, the connection between the cell and the beam, and the rod which counter-","page":52},{"file":"p0053.txt","language":"en","ocr_en":"Influence of the rate of change.\n53\nbalanced the float. A light steel tube of 25\"\"\" diameter was used for a float. The connection with the cell wqs reestablished in the following manner. A balance beam E of the same dimensions as the balance beam A is placed immediately above A and parallel to it. We may name the float end of the original beam Af, and the corresponding end of the upper beam Ef and the other ends respectively Aw and Ew. Ef is vertically over Af and they are joined by two loops of fine wire which interlock at the middle in such a way that when the tension is released and Af and approach each other the loops cause no friction and allow no spring or elasticity; the Ip wer is fixed in an upright position and the upper falls vertically by its own weight. Now, following the same principle as in apparatus D, the standard weight is maintained by means of the compound balance action. The cell is attached to Ew and from the same point a balance pan (pan E) is suspended 400\"\"\" below. In place of the rod at A70 another scale-pan (pan A) is suspended at the same distance as pan E. When the weight in pan E is equal to the weight of the float at zero the two beams stand in a horizontal position and the loops between Af and Ef are just on the point of making contact. A is balanced and exerts no influence on E, and E is balanced and exerts no influence upon the cell which, when lifted, has its own weight plus the weight of pan E. If the weight in pan A be diminished Af will pull on Ef by that amount and this will in turn lift at Ew, which lessens the weight of the cell by the same amount.\nThe balancing of the apparatus depends upon the position of the cell when lifted. A special support for the cell regulates this. It consists of a trap on top of the arm support (C,). This is essentially a slat with one end hinged to the rear end of the arm support. Its front end is held up against an adjustable catch by a weight acting through a cord over a pulley. This cell is placed on the front end of this slat in the same lateral position as before, but it is supported at the standard height to which it should be lifted. When the cell is grasped it is held in the same comfortable position. The experimenter elevates the weight which supports the trap, the trap falls and leaves the cell supported by the hand in a definite position.\nLet me further explain the apparatus by an illustration of how it works. Suppose I desire to produce a measured increase of 2s on a standard of 40e at a very rapid.rate. The cell itself weighs 25s and a weight is placed in pan E, so that the total weight of the cell when held in position, after the trap is dropped, will be 408. Both beams are balanced in a horizontal position and exert no influence on the ce^. Then we take off 2* from pan A. This makes Af pull on Ef","page":53},{"file":"p0054.txt","language":"en","ocr_en":"54\nC. E. Seashore,\nby 2s and this in turn lifts the pan E by 2s, reducing the standard to 38\". This we correct by adding 2s to. pan E and again obtain our standard, 40\u201c. But now pan E is lifting 2% at Ef and the same amount at Ew; if we then gradually restore the balance of A by immersing the float, E being stationary, the amount lifted at Ew will decrease until the zero point is reached and the contact between Af and Ef is broken. The contact is broken at the moment 2* have been added to the weight of the cell making it 42*. The rule determining the size of the increment is that the desired weight must be removed from pan A and placed in pan E. The rate of increase is determined by the size of the stream of water as before.\nThe above is an entirely satisfactory solution of the problem of how to produce quick and accurate changes of weight in a given standard without movement or jarring of the body lifted. The adjustment to the zero point may permit a possible error of\tThe trap holds the cell\nin the correct position, and, after a brief practice, the observer can retain this approximately, but for accidental movements of the hand there may be an error of within\tThere could be no other inaccuracy in the\nsize of the increments, since they were made by actual interchange of gram weights.\nThe rates were timed by the graphic method of recording time. The recording pointer was placed in circuit with a means of making the circuit at the beginning of the increment and breaking it at the completion. The electric clamp key was used as in series I to make the circuit the moment the stream was let on. A platinum contact was built up at Af, such that Af rested upon it and kept the circuit closed as long as Af was heavier than -Aw. The moment the beam A passed through its point of balance and the tension on the interlocking loops was released, Af left its contact and the electric current was interrupted.\nI first measured the time of all the increments (the steps in the series differed by 2s each) from 2g to 20s on one rate to find if the changes were uniform. This rate made the change at the rate of ig in 0.12\u201d. Taking the time to increase 208 as a standard, and calculating the theoretical time for each increment, I found that the empirical results deviated from the theoretical only by an irregular fluctuation no larger than might be allowed for the error of measurement, i. e., there was no systematic error large enough to demaad consideration here, and the increase may be considered as practically constant. Having found this I timed all the rates for the step of 2og, and divided that up proportionately for the other steps. The rates adopted and measured in this manner are as follows :","page":54},{"file":"p0055.txt","language":"en","ocr_en":"Influence of the rate of change.\n55\nI.\th.\tin.\tIV.\tV.\n4-54\t8-33\t33-33\t50.00\t66.66\n0.22\tO.I2\t0.03\t0.02\t0.015\nO.II\t0.21\t0.83\tI.21\t1-54\nHere a denotes the number of grams of increase per second, \u00df the number of seconds to increase one gram, and y the part of the initial stimulus per second.\nExperiments.\nThe fastest rate hitherto used required the largest increment. Referring to the curves in Figure 3, what is the highest limit for these if they be continued to the left, and what form will they assume in returning to the low point that marks the instantaneous increase ? This is the question I have tried to answer by the present series of experiments.\nThe quoted rates are within the limits of sufficiently accurate measurement by the present apparatus. The slowest connects with the fastest of the previous and the fastest approaches the instantaneous.\nThe method of minimum variation was here used somewhat differently from the previous manner. The threshold was determined five times for each rate as follows : the increments differed by 2g each, but, in order to save time, five trials on each step were made in succession. A number of steps in the middle region were tried until a block of records was obtained in which the change had been perceived correctly every time in the largest of the increments tried and no time in the smallest. The few exceptions to this rule may be seen in the records from the fact that the highest number is less than five in those cases.\nThe observer was given a choice of two answers only, namely, \u201cChange\u201d or \u201cNo change.\u201d The merits and demerits of that limitation are well known. At a signal the observer grasped the cell with the hand in position to rest firmly. At a second signal the trap fell and he was to look for the differentiation which might begin about two seconds afterwards. He was required to give his answer as soon as the increment was completed.\nThe results for six observers are contained in Table VIII. rlhe left hand section of the table shows the number of perceived changes out of five possible for each step. There the variation of any single increment may be traced for each observer. All but the first of the successive \u201c5s\u201d are omitted, and all increments above this are counted as perceptible in the respective single determinations of the threshold, A is found as before by taking the average of the single thresholds above","page":55},{"file":"p0056.txt","language":"en","ocr_en":"C. E. Seashore,\n56\nTaule VIII.\nThreshold of increase in a 401 weight at five rapid rates.\n\t\t\t28\t48\t68\t8\u00ab\tIO\u00ab\t128\t148\t168\t\u20ac\tj\td\n\t\tI\t\ti\tS\t\t\t\t\t\tO\t4.8\t0.8\n\t\tII\t\t\t5\t5\t4\t5\t\t\tO\t7.2\t1.8\nF. C. \u25a0\t\tIII\tI\t3\t3\t5\t\t\t\t\tO\t5-o\t1.8\n\t\tIV\t\t3\t4\t5\t\t\t\t\t0\t5-2\ti-4\n\t\tV\tI\t2\t5\t\t\t\t\t\tO\t5-2\t1.0\n\t\tI\t\t\ti\t2\t0\t\u25a05\t\t\tO\t12.0\t0\n\t\tII\t\t\t\t\t\t2\t2\t5\tO\tI4.8\t1.4\nN. H. \u2022\t\tIII\t\t\t\t3\t2\t2\t5\t\tO\t12.4\t1.6\n\t\tIV\t\t\ti\t4\t5\t\t\t\tO\t9.2\t1.0\n\t\tV\t\t\t4\t0\t2\t5\t\t\tO\t10.8\ti-4\n\t\tI\t\ti\ti\t3\t3\t4\t3\t\tO\t12.4\t2.0\n\t\tII\t\ti\t3\t3\t4\t3\t5\t\tO\t11.2\t2.6\nV. S. \u2022\t\tIII\t\ti\t4\t5\t\t\t\t\t0\t6.0\t0.8\n\t\tIV\t3\t3\t4\t4\t\t\t\t\tX\t6.4\t2.0\n\t\tV\t\t4\t2\t3\t5\t\t\t\t0\t7.6\t2.0\n\t'\tI\t\t2\t2\t5\t\t\t\t\tX\t6.9\t1.4\n\t\t11\t\t3\t3\t5\t\t\t\t\tX\t6.0\t1.6\nW. J.\t\tIII\t4\t4\t\t\t\t\t\t\tX\t2.8\t1.4\n\t\tIV\t3\t5\t\t\t\t\t\t\tX\t2.8\t1.0\n\t\tV\t5\tS\t\t\t\t\t\t\tX\t2.0\t0\n\t'\tI\t\t2\t2\t3\t2\t3\t3\t5\t0\t14.0\t1.6\n\t\tII\t\ti\t2\t3\t3\t4\t\t\t\\\t11.6\t1.2\nJ. R.-\t\tIII\t4\t4\t3\t5\t\t\t\t\tX\t4-4\t2.8\n\t\tIV\t4\t2\t4\t4\t\t\t\t\tX\t5-2\t2.6\n\t.\tV\t4\t5\t\t\t\t\t\t\t0\t2.4\t0.4\n\t\tI\t\t3\t3\t5\t\t\t\t\t\\\t5-6\t2.0\n\t\tII\t\t\t\t3\t5\t\t\t\t0\t8.8\t1.0\nW. II.-\t\tIII\t\t\t2\t1\t2\t5\t\t\t0\t11.2\t1.0\n\t\tIV\t\ti\t2\t2\t5\t\t\t\t0\t9.2\t1.0\n\t\tV\tI\t2\t4\t4\t3\t4\t\t\t0\t9.6\t2.4\n\t'\tI\t0\t3\u00b0\t47\t77\t90\t87\t100\t\t237\t9-3\t3-5\n\t\tII\t0\t27\t43\t63\t70\t80\t90\t100\t&\t9.8\t2.6\nAv. <f0\t\tIII\t3\u00b0\t40\t57\t77\t80\t90\t100\t\t\u00c0\t7.0\t3*3\n\t\tIV\t33\t47\t67\t80\t100\t\t\t\t\u00ee3r\t6-3\t1.9\n\t-\tV\t37\t60\t83\t90\t83\t97\t100\t\t2V\t6-3\t3-1\nRoman numerals, rates of change.\t\t\t\t\t\t\t\t<7,\tmean variation :\t\tto find\tthe mean\t\nFigures at the top, increments\t\t\t\t\t\tin grams.\t\tvariation for\t\tthe series divide each\t\t\tby\ne, the numerator gives the number of errors that were made in the number of control trials denoted by the denominator.\nJ, threshold.\n2.2.\nThe average is stated as the percentage of the possible number of correct answers.","page":56},{"file":"p0057.txt","language":"en","ocr_en":"Influence of the rate of change.\n57\nwhich all increments were perceived, on the supposition that, when an increment had been perceived every time, all steps above that would also be perceived. The assumption is quite valid, for the steps are larger and the successive steps would be taken under similar conditions with reference to the larger fluctuations in sensibility. The value of this threshold must also be estimated by its mean variation and by a comparison with the distribution of the figures in the first part of the table.\n\ncf oo\nFig. 4.\nE, Results in Table VIII, p. 56.\nD, Results in Table IX, p. 58.\nI, Results in Table V, p. 46.\nA, Highest point in A, Figure 3, p. 50.\nB, Highest point in B, Figure 3.\nThe horizontal axis is divided on a scale 20 times as large as in Figure 3.\nThe number of experiments is too small to afford a detailed expression of the law of variation, but it is evident that we have found the maximum or turning point in the size of increments. The results as expressed in curve E, Figure 4, show the gradual return to the point of instantaneous increase. This curve shows their relation to the point / denoting instan-","page":57},{"file":"p0058.txt","language":"en","ocr_en":"C. E. Seashore,\n58\ntaneous increase and the point A denoting the highest point in curve A, Figure 3, i. e., the threshold for the rate of i\u00ae in o. 15s. The object of these experiments is to trace the connection between H and /. Since their relative distance is very much magnified in Figure 4, the scale in this figure should be compared with the scale in Figure 3.\nThe above experiment was repeated four times upon the writer, as far as possible under similar conditions. The results are expressed in Table IX and curve D, Figure 4.\nTable IX.\nFour\t\texperiments upon the author\t\t\t\t; conditions\t\tsimilar\tto those in\tthe preceding table.\t\t\n\t\t\t2*\t48\t6\u00ab\t8\u00ab\t10\u00ab\t128\t148\te\tJ\td\n\t\tI\t2\t2\t4\t5\t\t\t\tX\t5-6\t1.2\n\t\tII\t\t\t2\t5\t\t\t\tX\t7.2\t1.0\nA-\t\tIII\t\t\t3\t3\t3\t5\t\t0\t10.4\t1.2\n\t\tIV\tI\t2\t2\t5\t\t\t\tX\t6.8\t1-4\n\t\tV\tI\t3\t4\t5\t\t\t\t0\t5-2\ti.S\n\t\tI\t\t\t\t2\t4\t\t\tVi\t9.6\t1.2\n\t\tII\t\t\t\t3\t5\t\t\t0\t8.4\t1.2\nB\t\tIII\t\t\tI\t2\t5\t\t\t0\t9.2\t1.0\n\t\tIV\t\t\ts\t\t\t\t\t0\t6.0\ta\n\t\tV\t\t2\t4\t\t\t\t\t0\t6.0\to.S\n\t:\tI\t\t\t\ti\t2\t3\t2\t1 7\t14.0\t2.4\n\t\tII\t\t\tI\ti\t5\t\t\t0\t9.6\t0.4\nC\t\tIII\t\t\tI\t3\t2\t5\t\t0\t11.2\ti.a\n\t\tIV\t\t\t3\t4\t5\t\t\t2 \u2022f\t7.2\t1.4\n\t\tV\t\t4\t4\t4\t\t\t\t\u00ce\t6.0\t2.2\n\t'\tI\t\t\t2\t2\t4\t3\t\t0\t14.0\t1.6\n\t\tII\t\t\tI\t2\t4\t4\t\t0\t10.8\t1.8\nD.\t\tIII\t\t\t2\t3\t3\t5\t\t0\t9.2\t2.2\n\t\tIV\t\tI\t3\t4\t4\t\t\tX\t7.6\t2.0\n\t\tV\t\t2\t3\t5\t\t\t\t0\t6.4\t1.2\n\t\"\tI\t10\t10\t3\u00b0\t5\u00b0\t75\t80\t85\t*\t10.8\t2.7\n\t\tII\tO\t0\t20\t55\t95\t95\tIOO\t\u00c0\t9.0\t1.2\nve.\t\tIII\tO\t0\t35\t55\t65\t100\t\t0\t10.0\t0.8-\n\t\tIV\t5\t10\t65\t90\t95\t100\t\tA\t9.6\to-5\n\t.\tV\t5\t55\t75\t95\t100\t\t\t\t5-9\t0.4\nNotation same as in\t\t\t\tTable VIII.\t\t\t\t\t\t\t\t\nThese results agree with the foregoing in that they show that the threshold is lowered as the point of instantaneous increase is approached.\nIn comparing the superposed curves in Figures 3 and 4 it must be remembered that there are important differences in the conditions upon which the results in each curve are based. Therefore, the comparison must not be one of absolute units, but of general directions and tenden-","page":58},{"file":"p0059.txt","language":"en","ocr_en":"Influence of the rate of change.\n59\ncies. Thus, in Figure 3, the curves descend in a decided manner towards the right, and in Figure 4 they descend toward the left, though not with as great regularity. The curves in these two figures may be joined together and then they will express the general law stated in the- next paragraph.\nVIII. Some observations and conclusions.\nThe various conclusions in regard to the influence of the rate of change upon the threshold of change for pressure and muscle sense made under the conditions adopted in the above experiments may be generalized as follows : the threshold for the perception of difference in pressure and lifted weight rises rapidly from the threshold of instantaneous change and soon reaches a maximum from which it falls off gradually until the slowest rate at which the change can be perceived at all has been reached.\nThis is not an absolute law, but it is a well defined tendency. Its value depends largely upon the conformity of the above conditions of experiment to the normal conditions of our every-day experience. I have hoped to obtain the facts undistorted by using the so-called \u201cunconscious\u201d method and making the results partially statistical. I am not going to enter into any polemic with those who have found contradictory results (mainly Stratton), for, like other factors in our perception, the rate influence depends upon its relation to a number of unknown subjective and objective conditions which determine its nature and effect. The subject has just been opened for experiment.\nAny law expressing the influence of the rate of change upon the perception of difference in sensory stimuli must be stated particularly with reference to the following among other factors :\n1.\tThe special sense organ. The law, derived from the above experiments, is indicative of relations that we find to obtain in other senses. Thus, there is a general agreement between these results and those found by Scripture and Stern on sight and by the same authors on sound (see references, p. 27), but there are important differences depending upon the functioning of the particular sense organs. The above may be made a general law of sensation, but it has definable peculiarities for each sense. Compare, e. g., Stern\u2019s results on sight and sound or mine on pressure and weight.\n2.\tThe kind of threshold. The rate influence has mainly been studied in the threshold of change. It is equally important and may be just as well studied in other thresholds, e. g., the threshold of sensation. I have made some experiments upon the least perceptible touch as depending upon the rate of impact of the stimulus. These experiments were made with a modified form of apparatus A. The rate influence was here more","page":59},{"file":"p0060.txt","language":"en","ocr_en":"6o\nC. E. Seashore,\nmarked than that for the threshold of difference in pressure, i. e., the largest stimulus was required near the instantaneous rate, which was much lower than for any gradual stimulus. At very slow rates the perception became very uncertain. Thus, I found it possible to apply a pressure of 48 over an area only i\",m in diameter upon a finger without the observer being able to detect it. I have made similar observations on the thresholds of sight and sound and on the thresholds of sensation, disagreeableness, and pain under electrical stimulation. These last experiments were made by an ordinary slide inductorium and a pair of electrodes. In such experiments the psychological method promises to be of great value for the study of the development of \u00e6sthetical ideas and tastes as depending upon the rate at which the sensory impressions are made, e. g., in approaching the threshold of pleasure or pain. And, what is mainly of theoretical interest in psychology has a very extensive practical interest for education.\nHall and Donaldson and Stern (see references, p. 27) have studied the rate influences in the perception of motion by sight. I have found a marked variation in the perception of tactual and muscular space. It has been customary to take the rate variation into consideration in estimating lifted weights by requiring that the weights should always be lifted to the same height at the same rate. M\u00fcller and Schumann\u2019s1 measurements on this point are valuable.\nI have constructed a dynamometer for measuring active pressure. A beam 300\"\"\" long is supported upon pivot bearings at one end. The other end carries a pointer which moves over an arc graduated empirically from os to iooog. The pointer end of the beam is supported by a steel spring of seventy coils hung vertically. The pressure point is a hard rubber disk 15\u201c\"' in diameter supported by a loop from a point on the beam 75\"'m from the bearing end of the beam. This is a convenient and satisfactory dynamometer. In some experiments I required the observer to press to the standard, 500s, and then reproduce it from memory at various rates of increase in pressure so as to reach the standard in the following times : (1) 2s; (2) the observer\u2019s own time, generally about 5s; (3) 5s; (4) io\"; (5) 15\u201c; (6) 20\"; (7) 25s. The standard was pressed before each single trial ; it was aimed to reach it in 5*. The results show that the slower the pressure increases the more it is overestimated. The 2* pressure is generally underestimated.\nWhat is true for one unit applies also more or less to other units of measurement in.sensation. I made some experiments, e. g., on the double\n\u25a0M\u00f6ller and Schumann, Ueber die psychologischen Grundlagen der Vergleich un gehobener Gesoichte, Archiv, f. d. ges. Physiol. (Pfl\u00fcger), 1889 XLV 37.","page":60},{"file":"p0061.txt","language":"en","ocr_en":"Influence of the rate of change.\n61\nstimulus in pressure with a standard of 5g using apparatus A as in series I with rates I, II and V. The results for five observers were: I, 10.5*; II, 12.1s; and V, 13.4s, i. e., the faster the increase the more the comparative pressure is underestimated. This is in accord with the laws found for dynamometry and for the threshold of pressure.\nThe rate influence is not limited to sensation. It enters our higher and more complex emotional and intellectual experience and activity.\n3.\tThe standard stimulus. The variation with the standard stimulus is, perhaps, best expressed by Weber\u2019s law.\n4.\tThe direction of change. For most stimuli there is a close relation between the threshold for increase and the threshold for decrease.\n5.\tMethod of marking the beginning and the end of change. Any method which leaves the reaction-time in the results is necessarily crude and the methods of elimination are uncertain. Introducing other methods changes fundamental factors in the conditions. Stratton\u2019s experiments and the mine have shown that the results depend to a great extent upon the method. According to Stratton, the law may even be reversed. We may grant that possibility, but that does not detract from the value of our results, for each method reveals a characteristic tendency. It seems to me that Stratton\u2019s reversal of the law obtained by a gradation-method, is not due only to the change in method of recording the end of the increment, but to this in connection with other important factors, such as the special conditions of knowledge of the physical relations and the state of preparedness.\n6.\tKnowledge of the facts. This changes the laws of perception in several ways. Knowledge of the physical facts acts as a suggestion. A conscious or unconscious distortion or correction is liable to creep in. The unconscious corrections are perhaps the most vitiating. Our ordinary experience affords us examples of change in which the physical processes are known and others in which they are not, and our experiences are different in the two cases. This is just what we find in experiment. From this point of view the cases in question must be stated with reference to (1) the state of expectancy or preparedness, (2) the distribution of attention, and (3) the degree of complexity of the discrimination. It is well known what a difference it makes whether a person knows what to expect and when and how to expect it. Such knowledge guides the distribution of attention. Thus, if I must distribute my attentive energy over 25s it will be less potent at any moment of response than if it were sharply focused just for that moment. But on the other hand, during a prolonged uncertainty, expectation rises and the effort of attention becomes greater and greater.","page":61}],"identifier":"lit28732","issued":"1896","language":"en","pages":"27-61","startpages":"27","title":"Influence of the Rate of Change upon the Perception of Differences in Pressure and Weight","type":"Journal Article","volume":"4"},"revision":0,"updated":"2022-01-31T12:57:49.932914+00:00"}