PSYCHO BIOLOGY EDITED BY KNIGHT DUNLAP Johns Hopkins University VOLUME I BALTIMORE, MD. 1918 F If, I CONTENTS NUMBER 1— JULY, 1917 Announcement 1 The retention of habits by the rat after destruction of the frontal portion of the cerebrum. S. I. Franz and K. S. Lashley 3 Action of some opium alkaloids on the psychological reaction time. D. I. Macht and Schachne Isaacs 19 On cerebral motor control: the recovery from experimentally produced hemiplegia. Robert Oden and S. I. Franz 33 The effect of delayed feeding upon learning. John B. Watson 51 Internal secretion in learning. (Discussion) Knight Dunlap 61 NUMBER 2 — SEPTEMBER, 1917 Continuous stimulations versus continuous shock in the phototactic response. S. J. Holmes 65 The effects of cerebral destruction upon habit-formation and retention in the albino rat. H. S. Lashley and S. I. Franz 71 The effects of strychnine and caffeine upon the rate of learning. K. S. Lashley 141 The stop-watch and the association test. Knight Dunlap 171 NUMBER 3 — NOVEMBER, 1917 On the motor functions of the cerebral cortex of the cat. Joseph Duerson Stout 177 Relative values of reward and punishment in habit formation. J. D. Dodson 231 NUMBER 4 — JANUARY, 1918 A classification of groups. Carl W. Bock 277 A synchronous motor kymograph. Knight Dunlap 319 Dunlap's method for the mean variation. Buford Johnson 325 Action of some antipyretic analgesics on psychological reaction time. D. I. Macht, S. Isaacs and J. Greenburg 327 NUMBER 5— MARCH, 1918 Some notes on the auditory sensitivity of the white rat. Walter S. Hunter. . . 339 A simple maze: With data on the relation of the distribution of practice to the rate of learning. K. S. Lashley 353 NUMBER 6— MAY, 1918 Methods of studying controlled word associations. Mildred West.Loring. . 369 Word-lists for adjective and noun reactions. Mildred West Loring 429 Methods of using balanced magnet chronoscopes. Knight Dunlap 445 Discriminative responses to visual stimuli. H. M. Johnson 459 Index 495 iii ANNOUNCEMENT PSYCHOBIOLOGY is established for the publication of research bearing on the interconnection of mental and physiological functions. It will include in its volumes therefore, not only investigations of what is sometimes called " psychological physiology" but also investigations in pharmacology, physiol- ogy, anatomy, neurology and psychiatry in so far as the results of these investigations have explicit bearing on problems of mental life, or mental factors are included in the essential conditions of the investigation. We would emphasize the fact that in spite of the increasing tendency in the several sciences represented, especially in psychology, to seek practical results, and formulate as quickly as possible conclusions having immediate application, it has been our purpose to establish another journal devoted to pure sci- ence, in which the ideals of those whose primary interests are in sound scientific progress shall be represented as fully as pos- sible. In pursuance of this purpose, we urge contributors to pay attention to the details of method and technique which alone can give their conclusions value, and we set no arbitrary limit of length. Papers submitted will be accepted in so far as space permits, if they conform to the following standards. (1) They must be in the proper field of the journal. (2) They must possess sufficient importance. (3) They must be adequately written. Unnecessary length and brevity are alike excluded by this last criterion. Manuscript may be sent directly to the editor-in-chief. When an article lies in the special field of one of the associate editors, it may be sent to him. PSTCHOBIOLOQY, VOL. I, NO. 1 THE RETENTION OF HABITS BY THE RAT AFTER DESTRUCTION OF THE FRONTAL PORTION OF THE CEREBRUM S. I. FRANZ AND K. S. LASHLEY From the Government Hospital for the Insane and the Psychological Laboratory of The Johns Hopkins University Much has been written regarding the neurology of learning and especial attention has been directed to the cerebrum. Com- paratively little evidence has been adduced to show what cere- bral elements are used in the formation of habits, although recent experimental investigations show that the frontal posi- tions of the cerebrum are utilized by monkeys, dogs, and cats.1 In only those animals with a highly developed brain is there a distinct differentiation of the frontal (as an association area) from the central (so-called motor and sensory-kinesthetic) area, and in fact the possibility of the histological differen- tiation of numerous areas of the brains of many of the lower animals is slight. The relatively simple and homogeneous character of the cerebral cortex in the rodents makes their cerebral physiology worthy of study, and there is the added advantage that the animals acquire habits rapidly and much information is at hand regarding their normal reactions. At the same time, on account of their low cost and ease of housing, many different experiments on the brain may be made which are not possible with animals having larger and more highly developed brains. Such experiments on rats may be expected to give results of at least suggestive value respecting the functions of corresponding parts of more highly evolved brains. Thus, if it is found that these animals can acquire habits after the removal of certain small or large parts of the cerebrum, 1 For most of the evidence: S. I. Franz, The Frontal Lobes, New York: Science Press, 1907, pp. 64. 3 4 S. I. FRANZ AND K. S. LASHLEY but not after the removal of other parts, or if they can retain but can not acquire habits after certain cerebral destructions, there will exist a basis for further extensive and intensive work on the so-called higher animals. The present work was under- taken with these objects in view. Several questions were posed, although the facts to answer only a few parts of these questions are now available. Some of the questions are: Do rats retain habits of recent formation after the destruction of certain cerebral regions? Do they re- tain habits of long standing, or those in which there has been an overtraining or over-learning? Can rats learn after the removal of the whole cortex? If learning and retention are possible after destruction of parts of the cortex, how much and what parts of the brain are necessary for, 'and what parts are normally used in the formation and the retention of habits? At the present time there are available results of experiments in which the frontal portions of the brain have been destroyed, and in which there have been destructions of two-thirds or more of the whole cortex (that of the cerebral convexity), but only those experiments dealing with the effects of frontal de- structions will be reported here. When the experiments were undertaken there was available a large number of rats which has been trained in a simple maze for other purposes and it was decided to use them in preliminary tests. The maze was built after the pattern of the Yerkes dis- crimination box (fig. 1). It consists of a starting compartment (a) leading by a sliding door to a central alley (6), which at its outer end offers the alternatives of the cul de sac (c) and the alley (d) leading directly to the food (e). A maze of this character had decided advantages for the training of large numbers of animals but is not altogether suitable for tests on retention on account of the speed with which it is learned and the degree of probability that any given correct trial is the result of chance. Incidental observations on the behavior of the animals in the maze are therefore of great importance for the determination of the retention of the habit. Two activities of the animals in the maze are to be especially RETENTION OF HABITS BY THE RAT 5 noted, as their characteristics are evidences of learning or lack of learning, and of retention of the maze habits. These activi- ties relate to the reactions at the door of the starting box and to the shortening of the path to the food. When first introduced into the starting compartment of the maze the rat sniffs at the wire cover, sides, and corners of the compartment and pays no particular attention to the door. When the door is first raised he almost always stands erect and sniffs at its lower edge before venturing into the first alley. With practice his reactions be- come centered on the door; he tries to push it up or sniffs at the FIG. 1. SIMPLE MAZE a Starting compartment; e, food. The dotted line shows the path taken by well trained animals, keeping close to the right-hand partition and cutting close around the corner. crack under it. The moment that the experimenter touches the door to open it the rat turns with his head in the right front corner of the starting compartment and as soon as the door is raised high enough to admit his body crawls out into the alley. This behavior is noted in the records of the different animals as " normal orientation to opening door." In his first trials in the maze the rat spends much time in sniffing at the wire cover, the walls, and particularly the corners of the maze. The trained rat can go from the starting compartment to the food in 1.2 seconds. The minimum time on the first trial for any of the sixty rats that have been trained has been eight seconds and 6 S. I. FRANZ AND K. S. LASHLEY the modal time is about thirty seconds, most of which is spent in exploratory sniffing. With practice these exploratory move- ments disappear and the animal runs to the food without a pause. Many animals come to follow the path marked by the dotted line in figure 1. That is, they keep close to the right- hand wall of the middle alley and keep close to the end of the partition in rounding the turn. This cutting down of excess distance and absence of exploratory sniffing are characteristic of the later stages of learning and when they appear in retention tests are therefore conclusive evidence for at least a partial retention of the motor habits 'of the maze. In training, ten successive errorless trials were taken as evi- dence for learning (rarely more than six errors are made in the hundred trials following the achievement of this record) . Some of the rats were then given an overtraining of from one to two hundred trials before the destruction of the frontal lobes. Others were operated upon on the day following that on which learning was completed. The operations were performed under ether anesthesia, and at the end the cut scalp was closed with sutures and was covered with a cotton and collodion dressing. In some cases a transverse opening about 4 by 8 mm. was made in the skull just back of the front o-parietal suture and the frontal area of the brain was destroyed by passing a narrow scalpel diagonally forward to the region of the olfactory bulbs and thence cutting out to the sides of the cranial cavity. In other cases two small trephine holes were made in the region of the suture and a spear-pointed needle was inserted through these, pushed through the frontal area and drawn to the sides to cut away the frontal regions. Owing to the small operative field it is not possible to determine the exact extent of the lesion at the time of operation but the possibility of using a large num- ber of animals and of later determination of the extent of the destruction of tissue makes it possible to obtain records of some animals in which the exact lesion desired has been produced. Most of the animals have been kept for two weeks or more after the operation and in many cases the absorption of the clot RETENTION OF HABITS BY THE RAT 7 has progressed to such an extent that it seems advisable to wait until histological examinations of the brains can be made before describing the lesions in detail. Fourteen animals have been operated upon for destruction of the frontal lobes and of these eight have been autopsied. In these the gross lesion has been in every case as extensive as that indicated in figure 2, and in three of the animals has extended back so as to involve the anterior two-thirds of the cortex. Brief records of the animals studied are given below. When- ever possible, fifteen trials in the maze were obtained from each animal on the day following the operation. The time con- FIG. 2. DIAGRAM OF THE EXTENT OF THE LESION IN RAT HI 9 , AS DETERMINED BY GROSS DISSECTION sumed in each of these trials in going from the starting compart- ment to the food dish and the number of errors, either of enter- ing the cul de sac or of turning back upon the true pathway, was recorded. In the following records the total time consumed in these fifteen trials and the total number of errors are compared with the total time and errors of the first fifteen trials made by the same rat in its first training in the maze. The time and number of errors of the rat's first trial in the maze at the begin- ning of training are also compared with those of the first trial of the retention tests. In addition to this observations are reported on the general behavior of the animals in the maze. 8 S. I. FRANZ AND K. S. LASHLEY Animals tested for retention without overtraining Experiment 1. G2 $ . Ninety-four days old at the beginning of training. Learning was completed in 54 trials at 10 trials per day. No overtraining. Operation through two trephine holes, followed by extensive hemor- rhage. Retention was tested twenty-five hours after the operation. The animal was constantly irritated by inflamation of the nasal sinuses but was otherwise in good condition. Orientation in starting compart- ment was normal. In every trial except the second the animal kept close to the right-hand wall of the middle alley and cut close to the end of the partition. On the second trial she turned into the entrance of the cul de sac but did not advance more than two inches. Various tests suggest that she was anosmic. A comparison of the records of learning and retention follows. LEARNING RETENTION ( Total time 188 seconds 54 seconds First 15 tnals |En.org 5 1 (?) f Time 15 seconds 2.6 seconds First trial | ^ 0 0 Experiment 2. Gl 9 . Ninety-four days old. Learning was com- pleted in 23 trials with 2 trials per day. No overtraining. Operation through large transverse opening, followed by consid- erable hemorrhage but with recovery of motor coordination within half an hour. Retention was tested twenty-six hours after the opera- tion. The rat was very weak, falling over when attempting to make quick turns or to scratch the dressing on her head. She oriented in the starting compartment and gave no evidence of exploratory snif- fing in the maze. On the first trial she turned into the cul de sac and wandered back and forth for a few seconds, then went directly to the food. A second error was made on the eleventh trial. The other trials were made correctly but at a rather slow rate. The rat made frequent long stuporous pauses and spent a good bit of time also in scratching at the dressing on her head. On the following day she was given 20 trials in the maze and in every case reached the food without error and in less than four seconds. The records of learning and re- tention follow. RETENTION OF HABITS BY THE RAT 9 LEARNING RETENTION f Total time 117 seconds 159 seconds First 15 trials < -, \Errors 8 2 fTime 10 seconds 45 seconds First trial < g^T 1 1 Experiment 3. HI 9 . Ninety-two days old. Learning was com- pleted in 21 trials with 2 trials per day. No overtraining. Operation through two trephine holes, with little hemorrhage. Retention was tested twenty-two hours after the operation. The rat oriented in the starting compartment, ran promptly and without exploratory smelling, and never explored the maze in a way compar- able to that of normal animals in their first trials. She entered the cul de sac on the first, fifth, eleventh, and thirteenth trials but turned back promptly before reaching the end of it. On the errorless trials she followed the path marked in figure 1. The records of learning and retention were the following. LEARNING RETENTION f Total time. . . 85 seconds 64 seconds First 15 trials < ,-, 1 .terrors 2 4 f Total time .. 10 seconds 8 seconds First tnal \Errors 1 1 Animals tested after overtraining Experiment 4- Glcf . Seventy-three days old. Learning was com- pleted in 21 trials with 10 trials per day. Training was continued for 170 trials. Destruction of frontal lobes through large transverse opening, fol- lowed by little hemorrhage. Retention was tested twenty hours after the operation. The rat oriented correctly in the starting compart- ment and advanced promptly when the door was opened. On the first trial he turned into the cul de sac and stopped with his head in the first corner, then backed out and went directly to the food. He made a second error on the fourteenth trial. The other trials were correct but were delayed by a peculiar reaction. When he reached the 10 S. I. FRANZ AND K. S. LASHLEY first corner after passing the turn he would pause with his nose close in the corner (but without apparent sniffing), then back away and turn down the alley to the food. Tests made by pulling his vibrissae while he was eating indicated that these organs were lacking in tactile sen- sitivity. The records of learning and retention were the following. LEARNING RETENTION ( Total time 1018 seconds 92 seconds First 15 trials |Em)rs 21 2 fTime... 34 seconds 12 seconds First tnal (Em)rg 1 1 Experiment 5. Fl 9 . Seventy-five days old. Learning was com- pleted in 90 trials with 10 trials per day. Training was continued for 120 trials. Operation through a single small trephine hole on the left with little hemorrhage. Retention was tested twenty-six hours later. The rat oriented in the starting compartment, and in the majority of trials fol- lowed the path indicated in figure 1. On the third and fifth trials she retraced a part of the direct pathway to the food and on the tenth trial she swerved so that her head was in the cul de sac but she never ven- tured entirely off the direct pathway. The records of learning and retention were the following. LEARNING RETENTION , ( Total time... 640 seconds 44 seconds First 15 trials < * 1 -terrors 19 2 f Time . . . 15 seconds 4 seconds First trial < ^ \ Errors 2 0 Experiment 6 '. F2cf. Sixty-nine days old. Learning was completed in 24 trials with 2 trials per day. Training was continued for 170 trials. Operation through large transverse opening with very severe hemor- rhage. Retention was tested twenty-three hours after the operation. The rat was active and oriented correctly in the starting compartment. In four trials the rat returned to the starting compartment after ad- RETENTION OF HABITS BY THE RAT 11 vancing for his own length into the middle alley, but he did not leave the pathway once in fifteen trials. The records of learning and re- tention follow. LEARNING RETENTION ( Total time . . . 93 seconds 68 seconds First 15 trials < „ ^Errors 7 o /Time... 13 seconds 6 seconds First trial < ^ 1 Errors ... .... 1 0 Experiment 7. G3cf. Seventy-three days old. Learning was com- pleted in 24 trials with 10 trials per day. Training was continued for 170 trials. Operation through two trephine holes with little hemorrhage. Re- tention was tested twenty hours after the operation. He oriented correctly to the opening of the door, ran quickly and followed the path of figure 1 in all but the first and second trials. In the first trial he entered the cul de sac but turned back without exploratory sniffing. In the second trial he put his head into the entrance of the cul de sac, but did not enter it. He appeared to be anosmic. The records of learning and retention follow. LEARNING RETENTION f Total time 359 seconds 55 seconds First 15 trials { * 1 rL/rrors 11 1 fTime.. 11 seconds 8 seconds First trial { ± 1 -terrors 0 1 Experiment 8. F2 9 . Seventy-five days old. Learning was com- pleted in 25 trials with 2 trials per day. Training was continued for 200 trials. Operation by transverse opening with severe hemorrhage. Re- tention was tested twenty-six hours after the operation. The rat was very weak and her movements were as a rule, slow and hesitating but without marked pauses at the entrance of the cul de sac. On the first trial she explored the cul de sac quickly, but without exploratory snif- fing, and had some difficulty in finding the food in the dish. On the 12 S. I. FRANZ AND K. S. LASHLEY tenth and thirteenth trials she again entered the cul de sac but did not go to the end. On the other trials she followed the most direct path and cut close around the end of the partition. The records of learning and retention were the following. LEARNING RETENTION ( Total time « 203 seconds 115 seconds First 15 trials |Errorg 7 3 fTime... 60 seconds 18 seconds First tnal \Errors 1 1 Experiment 9. Dl 9 . Sixty-nine days old. Learning was completed in 48 trials with 10 trials per day. Training was continued for 200 trials. 9 The frontal lobes were destroyed by a transverse incision. Re- tention was tested twenty-four hours after the operation. The rat oriented correctly in the starting compartment, attempting to lift the door by thrusting her nose under it. Two errors were made on the fifth trial. The others were run correctly and by the shortest possible path. There was no exploratory sniffing. The records of learning and retention follow. LEARNING RETENTION f Total time 566 seconds 57 seconds Fireti5triais{ErL ......:::::::::::.:. 23 2 fTime 18 seconds 1.8 seconds First trial [^ 1 0 Experiment 10. Blcf. Sixty-seven days old. Learning was com- pleted in 70 trials with 10 trials per day. Training was continued for 200 trials. The frontal lobes were destroyed by a transverse incision. Re- tention was tested twenty-eight hours after operation. On the first trial the rat explored the maze hurriedly without pausing to sniff. In the later trials he usually paused and swayed back and forth at the end of the first passage but entered the cul de sac only once, on the RETENTION OF HABITS BY THE RAT 13 seventh trial. He oriented in the starting 'compartment. The records of learning and retention follow. LEARNING RETENTION f Total time. . . 1797 seconds 137 seconds First 15 trials < „ (Errors 33 2 [Time... 32 seconds 12 seconds First trial < „ 1 Ji/rrors . ... 1 1 Experiment 11. F3cf. Sixty-nine days old. Learning was com- pleted in 60 trials with ten trials per day. Training was continued for 200 trials. The anterior third of .the cortex was destroyed. Little hemorrhage. Retention was tested twenty-two hours after the operation. On the first trial the rat turned into the cul de sac and ran half way to the end, then turned back and went directly to the food. He made no other error in the fifteen trials of the test, and followed the most direct route to the food. His records for learning and retention follow. LEARNING RETENTION f Total time. . . 428 seconds 50 seconds First 15 trials < " ^Errors 1] 1 f Total time. . . 175 seconds 11 seconds First trial < * ^Errors 4 1 Experiment 12. G4cf. Seventy-three days old. Learning was completed in 16 trials with 2 trials per day. Training was continued for 200 trials. Frontal lobes were removed by a transverse incision in the region of the fronto-parietal suture. Retention was tested twenty hours after the operation. The animal was very weak and spastic. He reacted promptly to the maze, however, orienting in the starting com- partment and never hesitating at the turn in the maze. He had some little difficulty in finding the food, pushing under instead of above the edge of the dish. After 10 trials he began to show evidence of fatigue so the remaining trials for retention were postponed for two days when he had regained almost normal strength. The records of learning and re- tention follow. 14 S. I. FRANZ AND K. S. LASHLEY LEARNING RETENTION ( Total time 135 seconds 51 seconds First 15 trials < ± ^Errors 2 o Clime... 20 seconds 2 seconds First trial < ^ } Errors 1 o Experiment 13. G2cf. Sixty-nine days old. Learning was com- pleted in 24 trials with 2 trials per day. Training was continued for 200 trials. Frontal lobes destroyed by a transverse incision. Retention was tested twenty-two hours after the operation. The animal was very weak and spastic. He had great difficulty in finding the food and gnawed at the edge of the dish as much as at the bread. Nevertheless he followed the direct path to the food with never a suggestion of reac- tion to the entrance to the cul de sac. During the trials there were many stuporous pauses (one of fifty seconds duration, which accounts for the long time consumed in the fifteen trials recorded below) so, after five trials the rat was returned to the home cage and the tests continued two days later when he had recovered strength. The records of learning and retention follow. LEARNING RETENTION , ( Total time 134 seconds 123 seconds First 15 trials < £~* ^Errors 3 o fTime 21 seconds 8 seconds First trial < ±*™ ^Errors 1 0 Experiment 14- F4cf. Sixty-nine days old. Learning was com- pleted in 18 trials with 2 trials per day. Training was continued for 120 trials. Frontal lobes destroyed by a transverse incision through a small trephine hole on the left. Retention tested 20 hours after operation. The animal had developed a left hemi-paresis and failed to leave the starting compartment of the maze in an hour on each of three consecu- tive days. Autopsy showed an extensive clot over the orbital sur- face of the right hemisphere, extending back to the pons. RETENTION OF HABITS BY THE RAT 15 SUMMARY OF RESULTS OF EXPERIMENTS The records of time and errors have been summarized in table 1. From the averages it appears that the rats which were not overtrained required 29 per cent less time for the first 15 trials after the destruction of the frontal lobes and made 53 per cent TABLE 1 The time required for reaching the food and the number of errors made by rats in learning the maze and in the retention tests after the destruction of the frontal lobes Animals without overtraining- TRAINING RETENTION NUMBER First 1 5 trials First 1 5 trials Time Errors Time Errors G29 . 188 5 15 54 1(?) 2 6 G19 117 8 10 159 2 45 HI 9 85 2 10 64 4 8 390 15 35 277 7 55.6 Animals overtrained Glo* 1018 21 34 92 2 12 Fl 9 F2d".... 640 93 19 7 15 13 44 68 2 o 4 6 G3tf 359 11 11 55 1 8 F29 203 7 60 115 3 • 18 Dl 9 . . . . 566 23 18 57 2 1 8 Bid1 F3c?.... 1797 428 33 11 32 175 137 50 2 1 12 11 G4tf 135 2 20 51 o 2 G2c? 134 3 31 123 o 8 5408 137 399 791 13 82.8 fewer errors than they did in learning the maze. This in itself is evidence for a partial retention of the habit. When consid- ered in connection with the data on their behavior in the maze it shows that there was little if any loss that can not be accounted for by the distracting effects of the head bandages and the gen- eral shock effects of the operation. None of the animals showed 16 S. I. FRANZ AND K. S. LASHLEY the exploratory sniffing at cracks and corners which is so char- acteristic of the untrained rat in the maze. All were tested im- mediately after the retention tests by being placed in a strange cage with food and all spent at least thirty seconds in exploring the cage before pausing at the food, so that the lack of explora- tory activities in the maze must be looked upon as due to re- tention of the habit and not to a general sluggishness resulting from the operations. The three rats which were not over- trained oriented in the starting compartment and two regularly followed the path marked in figure 1. The abnormality of be- havior of the third (Gl 9 ) was probably due to loss of sensi- tivity of the vibrissae. The animals which were overtrained required 87 per cent less time for the first fifteen trials after operation and made 90 per cent fewer errors than in their initial learning. This, in addi- tion to the data on individual behavior in the maze shows that there was practically no loss of the habit resulting from the de- struction of the frontal lobes. There is an apparent difference in the amount of retention be- tween animals which were over-trained and those which were trained only until they had learned the problem. This differ- ence is probably not so great as is indicated by the averages because the long time spent by the non-overtrained group is the result of the inclusion of the rat Gl 9 which spent a great deal of time in trying to remove the dressing from its head. Only one animal did not show evidence of the maze habit after removal of the frontal portions of the brain. This animal showed such an amount of muscular weakness, or apathy, "that the running of the maze was not attempted by it even after the fash- ion of an untrained animal. With this exception the tests gave indisputable evidence of the retention of the habit after the frontal portions of both hemispheres has been excised. More- over, the evidence is more compelling because of some obvious behavior disturbances in a number of the animals. Thus, it has been reported of the second animal, Gl 9 that, although the time for running the maze after the operation was greater than in the training series its other behavior relating directly to the RETENTION OF HABITS BY THE RAT 17 maze was "retained. The time variation (lengthening) was due entirely to changes in its physical condition other than those necessarily related to its maze activities. That this is so will be realized when it is remembered that the delays were made up of periods of scratching its head-dressing and of long stuporous or apathetic pauses. In the fourth animal the sensibility of the vibrissae was decreased, perhaps they were anesthetic, and the short times for running the maze after the operation are especially noteworthy. The twelfth and thirteenth animals were weak and spastic, and exhibited abnormal reactions in connection with the food dish, but both managed to find the correct path quite promptly. The time for the first fifteen trials of the thirteenth animal, G2 0.656"* r.v. 10.0% 6.6% 3.3% 14.8% 16.1% Time 36' 50' 53' 39' 44' After in- jection < second. . M. m.v. r.v. (23) 0.300" 0.025" 8.0% (22) 0.158" 0.008" 5.7% (24) 0.136" 0.008" 6.3% (19) 2.300"g 0.362"* 15.7% (19) 3. 702" 2 0.592" ^ 16.0% After in- Time 57' 77' 80' 60' 66' jection -f V>irrl M. m.v. (23) 0.280" 0.022" (23) 0.167 0.012 (26) 0.140" 0.007" (21) 2.401" 0.535" (21) 3. 578" 8 0.654"* r.v. 8.0% 7.2% 5.5% 22.2% 18.3% After in- jection Time M. m.v. 84' (15) 2.018" 0.244" 88' (14) 3. 591" 8 1.084"* fourth. . r.v. 12.1% 30.2% Key to tables: M. = mean or average reading in seconds; m.v. = mean variation; r.v. = relative variation. The line marked Time, indicates the time after injection of the drug when the readings were begun. Figures in parentheses indicate the number of readings made. Owing to the expense of publication, it is impossible in this paper to report in tabular form all the data obtained in the pres- ent investigation. Only the chief results will therefore be de- scribed here. Tables 3 and 4, however, are printed as examples of the manner in which the data were studied. In table 3 the ACTION OF OPIUM ON PSYCHOLOGICAL REACTION TIME 25 effect of 4 mgm. of morphin on one of the subjects is tabulated. In the first series of readings the normal reaction times are re- corded, the first line indicating the average of a large number of readings; the second line the mean variation, and the third line the relative variation. The normal having been established, the drug was injected and four series of readings were made in the same order as above, as indicated in the table. In table 4 the effect of the same dose of morphin, 4 mgm., in the form of narco- phin, i.e., in combination with 8 mgm. of the inert alkaloid nar- cotin, in the same subject is given. Here the number of errors are also given. TABLE 4 Dr. Macht, July 20, 1916; Narcophin 12 mgm. (= Morphin meconate 4 mgm.) at 3.15 p.m. SERIES SOUND TOUCH (VOICE REAC- TION) LIGHT ONE PLACE FIGURE X3 + 4 ONE PLACE FIGURE X4 — 3 Before in- 1 jection... | M m.v. r.v. 0.278" 0.020" 7.2% 0.248" 0.026" 10.4% 0.162" 0.011" 6.7% 1.154" 0.183" 15.8% 1.682" s 0.321" § 19.1% ~ After in- jection first Time M. m.v. 15' 0.308" 0.023" 57' 0.258" 0.033" 62' (18)0.150" 0.015" 45' 1.298" 0.198" 49' 1.686" | 0.386" S r.v. 7.6% 12.7% 10.0% 15.2% 22.8% After injec- t i o n second.. . Time M. m.v. r.v. 80' (19)0.300" 0.030" 10.2% 109' 0.238" 0.025" JO. 5% 114' (19)0.136" 0.008" 6.1% 99' 1.432" 0.239" 16.6% 104' 1.644' 0.491" 29.9%' NORMAL DATA As might have been expected, the normal reaction time varied widely in the different subjects. The normal reaction time for simple sound ranged from 0.172 to 0.340 of a seco'nd, the commonest figure lying between 0.250 and 0.300 of a second. The simple reaction time for touch ranged from 0.120 to 0.280 of a second, the most frequent reading lying between 0.140 and 0.190 of a second. 26 DAVID I. MACHT AND SHACHNE ISAACS The simple reaction time for light ranged from 0.112 to 0.220 of a second, the commonest figure lying between 0.120 and 0.160 of a second. The reaction time for the addition test ( + 17) varied from 1.154 to 3.010 seconds, the commonest figure being a little over two seconds. The reaction time for the multiplication and addition tests was much longer, the figures ranging from 2.456 to 6.810 sec- onds, the commonest figures being between three and five seconds. It will be seen that of the simple reactions, that for light was the shortest and that for sound was the longest. EFFECT OF MOEPHIN It was found that the effects of a morphin injection depends on the size of the dose and manifests itself in one or more of three ways. In the first place, the absolute reading of the reaction may be affected. In the second place, the mean variations in the readings may be greatly increased or decreased. In the third place, in the case of association reactions, there may be an effect noted on the accuracy with which the subject performs mental problems. After small doses of morphin (4 to 6 mgm.) there was noted a distinct primary effect, which consisted in a stimulation or a shortening of the reaction time, or in a decrease in the mean varia- tion of the readings, or sometimes in both, and furthermore in a lesser number of errors made in the computation of mathematical problems. This primary effect of morphin generally lasted half an hour or more and was followed by a secondary stage charac- terized by a depression, as indicated by the prolongation of the reaction time and greater variations in the readings or both. After very small doses of morphin, however, the depression was sometimes lacking. After larger doses of morphin (8 to 15 mgm.), however, the primary stage of stimulation was very short and could be easily overlooked unless the readings were begun very soon after the injection of the drug. Depression, on the other hand, was ACTION OF OPIUM ON PSYCHOLOGICAL REACTION TIME 27 the predominant picture as could be seen by the prolongation of the reaction time readings and greater variations in the same and also, in case of associations, by a greater number of mistakes. Although the two stages of morphin action above described were not always marked, a careful analysis of all the experiments indicated that they were present in almost all the cases. The primary stage of quickening or stimulation, in our opinion, probably corresponds to the stage of euphoria or well-being so well known to the pharmacologist and which occurs after small doses of opiates. It is this euphoria or sense of well-being which probably is responsible in a great measure for the greater accuracy in mathematical calculations, especially in subjects with a nervous temperament; inasmuch as the narcotic action of the drug is just sufficient to "take the edge out" of the subject's anxiety. The primary stage of increased efficiency noted agrees well with the results of some other tests of mental efficiency produced by opium per os or by mouth described by Mtinsterberg (12). The ordinary therapeutic doses of morphin (8 to 15 mgm. or | to J grain) are generally too large to make the primary stage of quickened reaction in normal individuals very noticeable and it is for this reason often overlooked. It may be remarked in this place that although the nauseat- ing effect of morphin, so commonly met with, occurred in several of the subjects, the nausea had apparently no effect on the read- ings. Indeed, in the individual who was most markedly affected in this way (Dr. Kiang) the reaction time, if anything, was quickened. EFFECT OF COMBINATIONS Three experiments were made with injections of narcotin hydrochloride alone and have already been described. No defi- nite change in the reaction time was produced by the drug. Three experiments were made with a combination of mor- phin and narcotin in the ratio of one to two, by administering the drug called narcophin, which is a mixture of morphin and narco- tin meconates. In two of the experiments there was a definite increase in narcosis and corresponding prolongation or depres- sion of the reaction time noted as compared with morphin alone, 28 DAVID I. MACHT AND SHACHNE ISAACS being very marked in one case, but of a lesser degree :'n the other. In the third case the narcosis was, i anything, less than that produced by morphin alone. Seven experiments were made with a combination of all the opium alkaloids in the form of pantopon, a mixture of hydro- chlorides containing 50 per cent of morphin. In four of the ex- periments there was a very marked increase in narcosis and pro- longation of the reaction time produced by pantopon as compared with that produced by the same amount of morphin when given alone. The greater narcosis in these cases was also shown by the greater number of errors in the association problems. In two other cases the greater narcosis was also present but not in so marked degree as in the preceding two, and in one experiment the result was doubtful. On analyzing all the experiments with pantopon and narco- phin, we may summarize by saying that out of ten experiments five showed a marked increase in narcosis and prolongation of the reaction time; three experiments showed also a definite but not so marked a prolongation of the reaction time as compared with morphin alone; one subject gave doubtful results although his accuracy was markedly affected in regard to association prob- lems; and in one case the reaction time was quickened* by the combination more than it was by morphin alone. It was also noted that in all experiments both with morphin alone and with its combinations, the simple reactions were less affected by the drugs than the association tests, thus showing that the narcotics exerted their influence especially upon the higher functions of the brain. Tables 5 and 6 present a summary of most of the experiments performed in the present investigation. In these tables the figures for reaction time are expressed in terms of per cent as compared with the normal readings in each experiment. Only, the primary and the maximum effects of each drug are shown in order to enable us to better analyze the effects of the drugs. The tables also indicate the time after the injection at which the readings were made, and in table 6 the run of errors is indicated by arrows. 1« ui rikraq SSuipBa.1 Snip jo uoi^oafui ia:>jtj SutpBai jo a'raij, II ui aSuBqo LOOCOCO ui aSuBqo sSuipuai -CO»OO?'^tlCO . S — i.T-iTHOOiOOOOi'-iOOOiOOO'-HOOO CD IO CD CO O Oi ^ c5 co 10 iO CO UOI^BUBA ueaui ui aSu'eqb ui aSu^qo 1.1 2 Snap jo uopoafui je Suip^aj jo 'a'unj, UOI^BUBA UB3UI ui aSu^qo ui aSuBqo Snip jo uopaafui e _ _ 3uip«ai jo e'uiij, *H »-i r- 1 i— I r— I CO »— I i i i i 00 i in s s E s s to iO<£>OOOO s i-iO i DQDQCQCQCQG ^ c a 29 ii ££ I: II n eo IS | i 30 DAVID I. MACHT AND SHACHNE ISAACS •§•8 1 T ~ T > 1 1 t t T T T- t T ! T T T TS» rHC^OO*OT i— iC^OSTfl^COlO't^'l 1 * T i CO , |> 05 CO •* T T T T"» •-I (M rH t T i ui 83m?qb |1 g 8 J2 £ 8 £5 88S§SS§S;£S ui a3uijqo II S §iO !>• i— i O O5 i-( CO O5 3iup jo uorpafui g ^ o co »o oo oo Suipeai jo auiix '§ tjC- O5Tf §05(M001>C005 rHrH ui § '-ICS OOiOOOOOiOOlOO 3njp jo uoi^oafui jo arajx %%&%$% O •— ^ »O O5 CO l •So 1 T t T T T- O _ ^ ^, ^H | N A ^^ rH ^^ C^-l C^ T^ C"l tt «t TTt'T TH CO t1"* '-HOOO t T =T T T T T O O Tt< O "H CO O S ui a3uBqo O O c^8 ui eSuBqo Snap jo uoi^oafui SuipBaj jo eraix 8 S2 3 UOI^BIJBA UB8UI O CO O5 (M CO i-H ui a3uBqo 1,1 2 3iup jo uopoafui I g co aa^jB SuipBaj jo eraix | 'g ^ iO O O rJH CO CO O T-H i— I C^ i— I T— I i-H i— I ISM|I s a s| e SON COOOOOi-^COOOi— IT-H s a II ft 10 3 - C s ft ~ m m c3 03 ^3 ^! ^ft^ft O> CQ fl d ^^ ft ft 3 3 £ £ ,£3 X! ^3 ^3 ^ft^ft^ft^ft^ CQ CQ O2 QQ QQ d ^ d d d d d 3 o I.I.1II1I tn JH ?7 EH M tL T 00000 ll^ggg^^fS d d cJ rt sas sss ACTION OF OPIUM ON PSYCHOLOGICAL REACTION TIME 31 It is interesting to note that in the case of both narcophin and pantopon, much less nausea was produced by those combi- nations than by the same amount of morphin which they con- tain, when given alone. This agrees perfectly with numerous other experiences of one of the authors (M.) recorded else- where (13). SUMMARY A careful analysis of all the experiments performed lead the authors to the following conclusions: 1. The effect of morphin alone and in combination with other opium alkaloids depends upon the dose used and may be mani- fested by a change in the mean reading, a change in the mean variation of the readings, or by both of these; and in case of association tests, by the number of errors made in performing a mathematical calculation. 2. After small doses of morphin, there is generally a primary stage of stimulation or quickened reaction time; this may or may not be followed by a secondary stage of depression, as indicated by narcosis and prolongation of the reaction time. 3. After larger doses of morphin, the primary stimulation stage is very short and may be overlooked, whereas the secondary or stage of depression is predominant. 4. From the experiments made with combinations of morphin with other opium alkaloids in the form of narcophin and pan- topon, it appears that morphin given in such a form is more narcotic and correspondingly more depressant to the psychic functions than when the same dose of morphin is administered to the same subject by itself. REFERENCES (1) ABEL: Physiological aspects of the liquor problem. 1903, ii, 1-169. (2) DODGE AND BENEDICT: Psychological effect of alcohol. Carnegie Insti- tute, Washington, 1915. (3) HOLLINGWORTH : Arch, of Psychol., no. 22, April, 1912. (4) LOEWALD: Kraepelin's Psychologische Arbeiten, 1896, i, 489. (5) POFFENBERGER : Arch, of Psychol. Am, Jr. of Psych, 1914, xxv, 82. (6) DIETL AND VINTSCHGAU: Pfliiger's Archiv, 1877, xvi, 316. 32 DAVID I. MACHT AND SHACHNE ISAACS (7) KRAEPELIN: Cited by Wundt, Physiol. Psychol., 1903, iii, 446. (8) MACHT: Journ. Pharmacol. and Exper. Therap., 1915, vii, 339. (9) MACHT, HERMAN AND LEVY: Proc. Natl. Acad. of Sciences, 1915, i, 582; and Journ. Pharmacol. and Exper. Therap., 1915, viii, 1. (10) DUNLAP: Jour, Exp. Psychol. 1917, ii, June. (11) DUNLAP: Psychological Review, 1913, xx, 154. (12) MUNSTERBERQ: Beitrage 2, Exp. Psychol. 1892, iv, 121, (13) MACHT: Trans. Assoc. of American Physicians, 1916. ON CEREBRAL MOTOR CONTROL: THE RECOVERY FROM EXPERIMENTALLY PRODUCED HEMIPLEGIA1 ROBERT OGDEN AND SHEPHERD IVORY FRANZ From the Physiological Laboratory of the George Washington University, and the Government Hospital for the Insane Attention has recently been directed to the possibility of re- covery of voluntary muscular control in human cases of cerebral hemiplegia.2 The results which have been reported are so dif- ferent from those which have been predicted by neurologists that the whole matter of cerebral control again comes to the fore as a problem of intense practical as well as theoretical interest. It has long been believed that if improvement in motor ability does not occur in man within a period of two years following the cerebral accident the paralysis is permanent. The series of cases which have been reported show that this is not true, be- cause even in cases of paralysis of eight or more years' duration, considerable improvement follows suitable remedial measures of the nature of exercise, including massage. It is well known that an animal which has had its so-called motor cortex destroyed or the pyramidal fibers cut on one side shows a condition similar to that of the human apoplectic hemi- plegia. It is also known that, even though the hemiplegia be complete, recovery of (voluntary) motor function takes place. The beginning of this recovery in the dog comes in a day or two, and after a few weeks the dog can use the legs on the paralyzed side apparently as well as those on the non-paralyzed side. The animal does, however, use the legs of the non-paralyzed side 1 This investigation was made possible by a grant to one of us (F.) by the Carnegie Institution of Washington, and we beg to express our sense of obliga- tion for the assistance thus given. 2 S. I. Franz, M. E. Scheetz, and A. A. Wilson: The possibility of recovery of motor function in long-standing hemiplegia. Jour, of Amer. Med. Assn., 1915, vol. 65, 2150-2154. 33 PSYCHOBIOLOQY, VOL. I, NO. 1 34 ROBERT OGDEN AND SHEPHERD IVORY FRANZ in preference to those of the paralyzed side even though the latter be " recovered." And also, when the animal is under the influence of certain toxic agents, such as alcohol or ether, the previously paralyzed limbs exhibit motor disturbances even though prior to the administration of the alcohol or ether the animal appeared to be perfectly normal in a motor way. The recovery in the monkey and ape is less rapid than in the dog, although after nine to twelve months it may not be possible on casual inspection to notice any motor disturbances.3 The effect of suitable exercises in the long-standing human hemiplegics suggested that if the paralyzed segments of an ani- mal with an experimentally produced hemiplegia were ade- quately dealt with the recovery would be more rapid and more complete than if the animal were permitted to recover by itself. The suggestion was tested and the results of the observations are given in the subsequent paragraphs. Four male monkeys (macacus rhesus) about a year and a half old were successfully used as subjects. One other animal died too soon after the operation to make the results of value. The hemiplegia was produced under ether anesthesia by the destruc- tion of the motor cortex with an electric thermocautery. The extent of the motor area was determined by faridization (bi- polar), and the area destroyed corresponded with the electrical delimitation of the motor zone. After the destruction different parts of the destroyed area, and beyond, were stimulated to see if the electrical stimulations would then produce move- ments, and in one case in which movements were obtained the area of destruction was extended and the part already cauter- ized was gone over again with the cautery. To destroy the motor zone lying concealed within the central fissure the white hot cautery was pushed about 6 to 8 mm. into the brain substance and carried close to and parallel with the fissure. The lesions which were produced were different from those produced by cerebral hemorrhage in man (apoplexy) in that they 3 For a general account of the phenomena of recovery from cerebral paralysis in the dog and monkey see Luciani: Human Physiology (Trans, by F. A. Welby), vol. 3, pp. 581 ff. (dog) and pp. 586 ff. (monkey). CEREBRAL MOTOR CONTROL 35 were cortical, and entirely so as far as this could be done. It will be remembered that most of the hemiplegia-producing cerebral hemorrhages in man are in the lenticulo-striate region, the cere- bral insult producing in those cases interferences with or de- structions of the pyramidal fibers. The lesions also differed from many in man in that in man a hemorrhage may act by pressure to produce a temporary alteration in conductivity or irritability which simulates the effect of destruction. This altera- tion in conductivity or irritability may partly or wholly dis- appear when the clot becomes organized or is absorbed. Since there is no regeneration of the cellular elements the experi- mental destruction may be regarded as the more complete and the more satisfactory. Information regarding the methods of dealing with the ani- mals are given in the brief accounts of the observations which follow. The post-operative management of the cases differed in order that the value of different procedures might be determined. Each animal was operated upon under a general anesthetic, and the later feeding and care was the same for all, except in those cases when it became necessary to feed the animal by hand. Experiment 1 Monkey 1. Before operation this animal appeared to be right- handed, although the observations were not sufficient in number to make this perfectly certain. It was normal and lively. The left motor cortex was cauterized May 30, 1916. The animal was then ob- viously hemiplegic on the right side, the paralysis extending to the face as well as to the arm and leg segments. The right arm and leg were flaccid and the right side of the face drooped. The animal could not feed itself with the right hand and arm. Management and results. The left (normal) arm was strapped to the trunk by means of a jacket so constructed that the arm could not be used for any of the important operations of feeding and climbing. The left leg could not be hampered in the same manner although it was thought to be desirable. The object of the restriction was to compel, if possible, the animal's use of the corresponding paralyzed segment. In addition to this passive method of treatment, efforts were made to get 36 ROBERT OGDEN AND SHEPHERD IVORY FRANZ the animal to move the paralyzed segments. The flaccid arm, for ex- ample, was dealt with in the following manner. The animal was held by a strap attached about the waist and the dorsal surface of the right hand was struck with a strap; this appeared to "anger" the animal and he endeavored to escape from the irritation (by the use of shoulder and arm muscles), and to lift the arm and hand to grasp the irritating stimulus; subsequent to the attempts to get the animal to move the arm muscles, the muscles which in human hemiplegic cases are those most difficult to recover (the extensors) and the nerves of those muscles were stimulated by friction and tapotement from five to ten minutes, the duration of the treatment depending upon the conditions of the involved muscles. The treatment of the leg was as follows: The ani- mal was strapped to a table, the right leg was held and the sole of the foot was struck to cause the animal to withdraw it; friction was also applied to the nerves and to the muscles; reactions similar to those of the arm segment were obtained from the leg, the animal attempting to es- cape from the 'stimulus by drawing up and by abducting the leg. At the same time the animal was led around by its strap, and in this way the animal was encouraged to use its paralyzed leg in walking and its paralyzed arm for support as it went about the room. At first little or no reaction was obtained from the stimulating treatment, but soon the application of the stimulus brought about slight appropriate or adequate responses, and after a few days the responses to the stimuli were almost equal to those of a normal animal. Soon also the animal began to use the arm for grasping food and in carrying it to the mouth, and the arm and the leg were used, but of course awkwardly at first, for climbing and holding. At the end of fourteen days the animal could use its leg and arm very well, and three weeks after the operation the monkey was able to pick small objects from the floor and to convey them to his mouth; he was able to use the two legs, both individually and together, very well, and there was no observable disturbances in walking and climb- ing beyond that to be expected from an animal which had one arm (the left normal one) rendered useless by the restricting jacket. In three weeks the monkey's movements on the right side were as accu- rate, precise, and forceful as those of a normal animal, and when the left strapped-down arm was liberated it was found to be less accurate than the right (disuse phenomenon). About two months later this animal was observed to catch with the right hand a fly that had alighted CEREBRAL MOTOR CONTROL 37 in the monkey's cage. The coordination and quickness for the per- formance of this act will readily be appreciated. Summary. By preventing movement of the normal arm and then "compelling" the animal to move the paralyzed segments, and by me- chanical stimulation of the peripheral nerves and of the muscles, in three weeks the animal recovered from its paralyzed condition to such an extent that the movements on the paralyzed side were judged to be normal. Experiment 2 Monkey 1. A week after this annual had thoroughly recovered the use of its right side a second similar operation was performed on the right hemisphere. The whole of the right stimulable cortex was de- stroyed under asepsis and general anesthesia, June 26, 1916. The paralysis involved the left side of the face as well as the arm and leg, and the paralysis was typical of the upper neuron type. Management and results. The right (recovered from paralysis) side of the body was not restrained and the left half of the body was not given any special treatment. In this respect the animal was given the chance to recover by itself without interference. The animal usually lived in a cage (90 by 58 cm., and 114 cm. high) by itself so that it would not get the stimulus of combat, etc., with another animal, but it was let out into a large room for exercise each day for periods varying from one to four hours. Some forced exercise of the newly paralyzed parts could not be prevented, for it was necessary to compel the animal to come close for observation and for testing, and its solitary living had made it somewhat timorous although during the period of its former paralysis it had been handled with relative ease. The animal, there- fore, cannot be said to have been entirely without some of the treat- ment which had been given to it following its first hemiplegia, although this kind of treatment was given as little as possible. Even though the animal had received a small amount of forced exercise it has remained paralyzed and apparently without much capa- bility of using its left arm and hand (December- 24, 1916). It can walk and jump; it climbs on the wire netting of its cage, it uses the left arm for a prop, and with the left hand takes hold of its strap when the latter is pulled upon. It tends to fall towards the left side, when it jumps it does not always reach the cage or box which it apparently attempts to reach, when it climbs over its cage the right arm and hand 38 ROBERT OGDEN AND SHEPHERD IVORY FRANZ are used for pulling and the left is apparently used only for support. When food is given, even though the food be close to the left hand, the animal always reaches for the food with the right. Unlike a normal monkey which grasps and holds food with both hands and feet, this animal uses only the right hand and the right foot. When compelled to stand the animal holds the left arm limp at its side, the right grasps the strap to support itself. When standing the toes of the left foot are spread, the great toe is at times doubled under the foot, and the leg is used uncertainly. When excited, as by some special stimulation or when a stick is pointed at him, the monkey will jump away and in the excited condition the left arm and leg appear to be used to much better advantage than in the unexcited condition. This may be due to the predominance of reflex activity at these times. If swung from his strap above the floor he also attempts to grasp the strap with his left hand, but only a slight amount of force is necessary to disengage that hand, although the right hand holds very firmly and cannot be easily removed from the strap. Summary. This experiment with a hemiplegic animal without special management and treatment shows that the animal may remain for a period of six months or more without very much improvement in voluntary control. This is in direct opposition to the results obtained with the right side of the same animal which, under treatment, recovered in three weeks. Experiment 3 Monkey 2. The left motor cortex was destroyed June 2, 1916, the operation being similar to those of the first two experiments. The animal then exhibited an upper neuron paralysis, involving the face and the upper and lower extremities. The right arm was at first completely useless, the right leg was limp. In coming out from the effects of the anesthetic the animal immediately used the left arm. Management and results. The unparalyzed side of the animal was not restrained, and in this respect the experiment was the same as in experiment 2. The animal did, however, receive " general" massage of the affected limbs, the parts being rubbed daily and the muscles being carefully kneeded. No special effort was made to get the animal to use the paralyzed segments, and the stimulation exercises like those in experiment 1 with monkey 1 were not carried out. The treatment (general, instead of special types of, exercises and massage) was carried out regularly for twenty-six days, and at the end of that time the CEREBRAL MOTOR CONTROL 39 monkey used the left hand exclusively for all operations. The right hand showed marked wrist drop, there was very little strength in either flexion or extension of the fingers, and the whole arm segment had not advanced much towards recovery during the period of the treat- ment. The leg showed a similar condition. There was a dragging of the foot when the animal crawled or attempted to walk over the floor, and the foot and leg could not be used with any facility for climbing or other kinds of operations which a normal monkey performs. It was evident, however, that gome improvement was taking place, and that there would be a recovery in time seemed to be a justifiable conclusion. Summary. General massage for twenty-six days of the paralyzed segments of an hemiplegic monkey did not bring about a recovery of motor ability, although there was some evidence of returning function, much more than that found in monkey 1 after six months' "laissez faire" treatment. Experiment 4 Monkey 2. At the time of the second operation on this animal it was in the condition just described. The second operation was the cauterization of the right motor cortex on June 28, 1916. This resulted in a complete paralysis of the left side with characteristic flaccid condition of the arm, face and leg. Management and results. The right (not completely recovered paralyzed) arm was bound closely to the body and only the left arm could be used by the animal for the purpose of feeding and climbing about its cage. In addition, active movements of the extensor muscles were invoked by mechanical stimulation, and massage was used for the muscle groups and for the nerves. The conditions of treatment were the same as in the first experiment with monkey 1. The results of this treatment for twenty-six days were evidenced by great activity on the part of the animal, by its ability to use the newly paralyzed segments, and the movements could not be said to be different from those of a normal animal. The movements are accu- rate and of good force, and the animal dominated monkey 3 which had been in the same cage with him for some time. He now uses his legs very well in walking, he jumps more accurately than the other two animals which still survive, and he is very much more active. There is an apparent preference for the use of the left hand in feeding, but when food is withheld until the animal uses the right hand for grasping it, it is seen that the right is used apparently equally well. At the 40 ROBERT OGDEN AND SHEPHERD IVORY FRANZ present writing the animal appears in all respects to be normal, there having been a continued betterment of the right side since the special exercises on that side were stopped. Summary. This animal with hemiplegia was given special exercises with massage and it was compelled to use the paralyzed segments; voluntary ability to move the paralyzed segments returned in twenty- six days, the recovery being present and apparently permanent five months after the operation. Experiment 5 Monkey 3. On June 2, 1916, the left cerebral motor cortex was cauterized as completely as possible. The paralysis was the same as in the previous experiments in that there was an evident complete hemiplegia of the upper neuron type of the whole right side. Management and results. The left arm was strapped to the body of the animal so that movements of the paralyzed right side would be necessary for feeding and climbing. No other kind of treatment was given, the animal being permitted to recover " spontaneously." After twenty-six days the amount of recovery was slight. Some movement of the paralyzed arm and leg was possible, but the animal was obviously incompetent on the right side. There was a characteristic wrist drop and there was some atrophy and an extreme weakness of the right arm. The leg was moved more than the arm, but it also was weak and the movements were uncertain and rather gross in nature. This was the condition on June 28, after which time active treatment of the Tight arm and leg was instituted, the treatment consisting in daily muscle and nerve stimulation by vibratory digital means, and in the stimula- tion of the animal by the special method already described. This treatment was continued for four weeks and in that period of time all evidence of the paralysis had disappeared, and the leg and arm had regained their normal power and precision. Summary. The normal arm of a paralyzed monkey was restrained but no special treatment of the paralyzed segments was given for a period of about four weeks, and this management did not bring about a return of motor function. During the next four weeks the nerves and muscles were stimulated and the animal was encouraged by speciai stimulations to use the arm and leg. During the second month the treatment brought about a complete return of motor function so that the animal's movements became normal. CEREBRAL MOTOR CONTROL 41 Experiment 6 Monkey 3. After the animal had recovered its normal motor ability on the right side following the destruction of the left cortical motor area, the right cerebral motor area was cauterized (July 28, 1916). This produced a left hemiplegic condition similar to those in the other experiments. Management and results. Both arms and legs were permitted to be free to move, but the left paralyzed arm and leg were carefully mas- saged without, however, giving individual attention to the special muscle groups and to the nerves as had been done in the second part of the preceding experiment. This animal continued to use the right arm almost to the exclusion of the left, although both may now be used when it is necessary, the left more awkwardly than the right. The animal moves well, climbs and jumps, it has been seen to pick over its cage companion for parasites (?), but all of its movements are more awkward than those of monkey 2 which is in the same cage with it. There has been a slight deterioration of the right side in that the right hand can- not be used as well as at the end of the special training period, and at the present writing it exhibits a slight wrist 'drop on the right side, but a marked wrist drop on the left. When it handles food, which it usually takes in both hands at one time, it is noticeable that there is considerable weakness on the left, there is also a marked awkwardness. During December, 1916, this animal was noticed to have convulsions. One began December 18, 1916, at about 2.00 p.m., and the animal was under observation during the convulsive attack. The animal had been feeding, and was holding two bananas in his hands. The food was suddenly dropped, and the monkey tried to get it from the floor, but not being apparently able to do this with the hands, he lowered his head to where the bananas had fallen. In that position a series of clonic movements began. The animal fell to its side, and the convulsive movements were noticed to be especially (or entirely, it could not be said with certainty) of the right side, but the face area was not ap- parently involved. The leg was more active than the arm, although the arm shook all over and the fingers were also in alternate contraction and relaxation. When the convulsion had partly subsided the animal tried to crawl over a partition (about 35 cm. high) which separated the cage into two parts. This attempt was unsuccessful at first, but the monkey continued to try until it succeeded. Success was finally at- tained only with great effort of the left arm and leg, and the right 42 ROBERT OGDEN AND SHEPHERD IVORY FRANZ corresponding segments were dragged over. The animal returned to its food about a half minute after it had successfully negotiated the par- tition. For about four or five minutes subsequently the right side could be used only with great awkwardness, but at the end of that time there was an apparent return to its former ability. There was a facial cyanosis for fully half an hour subsequent to the convulsion. Another similar convulsion occurred the same evening, about 8 hours after the first one. Others have been noted repeatedly both by one of us and by an assistant, and they have appeared to be of the same char- acter as that described. The monkey has not been able to use the right hand as well as he did previous to the occurrence of the convulsions, and some of the apparent deterioration in the proper use -of that hand may reasonably be ascribed to the unknown convulsive-producing condition.4 Summary. In this experiment although general massage was given to the paralyzed segments there was less recovery than in those cases in which special attention was paid to the individual muscles and nerves. The recovery has been sufficient to enable the animal to feed himself, and to perform other necessary acts, but not sufficient to make the finer kinds of movements; the muscles remain weak. This case is complicated with a unilateral epileptiform condition, which may have been the reason for a slight deterioration in the use of the right hand. 4 Since the above was written another convulsion has occurred in the presence of one of us, and its characters have been noted. The monkey had been eating a piece of carrot for about three minutes, when the food which he had been hold- ing was dropped to the floor, the right hand was clenched to make a fist, then there was a tonic flexion of the forearm on the arm, and this was followed by a slow tonic abduction of the arm to about 75 degrees from the normal position. A sudden relaxation then occurred, followed by a series of clonic movements in the whole of the arm area, and at this time the monkey cried several times. From this time the convulsion was purely clonic, the right leg following the arm, and in a few seconds the left side followed the right in a series of severe and extensive movements. The convulsion ended in 27 or 28 seconds, with a gradual lessening of the rate and of the extension of the movements, and as soon as the clonic movements had stopped the animal took up the piece of car- rot which it had been eating previous to the attack. The convulsions have been coming at longer intervals, and they have been observed chiefly after the animal begins his morning meal. CEREBRAL MOTOR CONTROL 43 Experiment 7 Monkey 4- On July 25, 1916, the left motor cortex was destroyed in the manner previously described. This brought about a right-sided hemiplegia, with flaccidity of the arm and leg. Management and results. The movements of the left arm were pre- vented by tightly bandaging that arm to the trunk. The special treatment consisted in the stimulation of the extensors of the arm, the shoulder muscles, and the muscles of the leg by friction, by the stimu- lating of the corresponding nerves, and by the irritating exercises to FIG. 1. PART OP THE SUPERIOR SURFACE OF THE BRAIN OF MONKEY 4 The obviously destroyed motor area is indicated by horizontal lines, and the apparently abnormal post-central area indicated by vertical lines. C, central fissure, partly indeterminate and indicated by the dotted line; S, fissure of Sylvius. About natural size. cause the animal to make defensive and offensive movements. The recovery was rapid, after the first few days the improvement being marked. After three weeks of the treatment (August 18) it was not possible to notice any difference in the activities of the two sides of the body, the right arm being used as well as the left for such operations as feeding and climbing, and the right leg being perfectly controlled and coordinated in walking, running, climbing, and jumping. This animal had been sent from Washington to New Hampshire, and on August 20 it was noticed to have coryza. This developed like an influenza with pulmonary symptoms, and the animal died three days later. The brain was removed and the accompanying diagram illus- 44 ROBERT OGDEN AND SHEPHERD IVORY FRANZ trates the condition of the left hemisphere. The microscopic examina- tion has not been made at this writing, but the gross appearance shows that most, if not all, of the arm and leg areas of the left cortex had been destroyed. The precent'al area which showed destruction is shown in the diagram by horizontal lines. Behind the central fissure there is anobher area (post-central and intermediate post-central, ac- cording to Campbell's histological differentiation) which appears to be affected. This area may have been involved because of changes in the blood supply in the application of the cautery to the precentral cor- tex, although it is not possible to determine the matter until the brain has been examined histologically. It also appears from the gross examination that some of the uppermost part of the leg area may have escaped destruction because of its proximity to the longitudinal sulcus, but this also cannot. be definitely determined until the results of the microscopical examination are available. It has previously been noted that the cautery was pushed into the brain so as to destroy the parts not usually accessible, and it may be that the superficial normal appear- ance will not be borne out by further examination. Summary. Monkey 4 was made hemiplegic on the right side, and after three weeks, treatment of the arm and leg muscles, and by com- pelling the animal to use the right side, it became able to use the right arm and leg as well as a normal animal. The brain showed ex- tensive destruction of the precentral area on the left, with a possible complication of the post-central area on the same side. GENERAL SUMMARY AND DISCUSSION The seven cases of hemiplegia in the four animals were treated in different ways in order to determine some of the conditions favorable to the recovery of voluntary motor function. The second experiment shows that motor recovery after the pro- duction of an hemiplegia does not result if the animal is left to its own devices, and this management (or lack of management) it is almost unnecessary to remark is what is given to most human paralytic cases. Even though the animal be prevented from using the sound (unparalyzed) segments there is little difference in the improvement from that in which no treatment is given unless in addition to the limitation of the possibility of move- ment there be added some extra stimulation to the muscles CEREBRAL MOTOR CONTROL 45 and nerves of the paralyzed side (experiment 5), although the recovery is rapid in such a case if treatment by muscle stimu- lation and nerve vibration be directed to the involved parts and if special stimulation exercises be given to the animal which will provoke the animal to move the paralyzed segments (final part of experiment 5). The method of treatment recommended by neurologists, general massage, does produce a slight amount of improvement but not to an extent to enable the animal to use the arm and hand properly for such ordinary operations as feed- ing and climbing, although these activities may be carried out after such treatment in an awkward manner. When, however, efforts are directed to the special nerves and muscles, and when the sound side of the animal is restrained so that movements of climbing and feeding must be made, if at all, by the use of the paralyzed segments the improvement is rapid and the recovery is practically complete (experiments 1, 4, and 7). One fact that stands out prominently is that recovery from the hemiplegic state may be very rapid. It has long been known that an hemiplegic monkey left to its own devices will after a considerable period of time recover the ability to use the arm and leg, but this period is one of months and is well illus- trated in the one of the experiments described (experiment 2), where the animal after six months has not recovered to any great extent the ability to use the paralyzed left side. The rapid recovery of the animals used in experiments 1, 4, and 7, and in the last part of experiment 5, is suggestive, and perhaps conclusion-compelling, that the continued paralysis of animals, and by analogy the persistence of motor incapacities in man, is due to lack of management rather than to a real inability.8 The results also suggest a reconsideration of the whole prob- lem of cerebral motor control, and especially that of cortical motor control. It has long been believed and taught that the cerebral cortex is necessary for the production of a voluntary 6 Each of the authors has in preparation a report of a series of cases of paraly- sis in man which will be published shortly, both showing that considerable im- provement may result from properly directing the attack against certain muscle groups and their related nerves. 46 ROBERT OGDEN AND SHEPHERD IVORY FRANZ movement. While it would be too venturesome to say from the experiments on the monkeys that the power of purely " volun- tary" movement was recovered, the experiments on man which have previously been cited, and those which will later be pub- lished, show conclusively that such " voluntary " movements may be produced even though the paralysis has been what neurologists call " residual/7 and in some cases even when it has persisted for a decade or more. It is, however, reasonable to suppose that not all of the " recovered" motor ability of the monkeys is of the nature of reflexes of a complicated type, and if we conclude that only a few of the recovered movements are " voluntary" it is sufficient to cause us to hesitate to accept the generally accepted view of cortical motor function. Here also may be cited the results which have been reported by von Monakow regarding the pyramidal fibers, for he finds that after the complete destruction of the motor cortex there is approximately from 25 to 33 per cent of the pyramidal fibers intact, or rather undegenerated. This fact would point, as- suming the pyramidal fibers to be purely motor, to the conclu- sion that other parts of the brain normally send impulses to the anterior horn cells, and that the control of the body muscula- ture is not entirely from the so-called precentral region, and it may be not entirely cortical. There remains from this anatom- ical argument the question of " voluntary" and "involuntary" movements, but this is more completely answered by the re- sults of the present series of experiments as well as by the re- sults of the experiments with human paralytics to which refer- ence has been made. The results are of interest in another direction, in that they place in the hands of the experimenter the means for the rapid recovery of motor function so that the " vicarious" functions of other cerebral parts may be investigated. If there is a delay in the recovery for periods of six to twelve months the pos- sibilities of experimentation are greatly reduced. With the possibility of producing such rapid recoveries as we have de- scribed in this paper there is opened up the means of investigat- ing certain motor functions which were not feasible previously CEREBRAL MOTOR CONTROL 47 on account of the long delays, in which there are numerous chances of intercurrent affections taking off some of the animals. Some of these problems have been planned and it is hoped that results will be obtained for a future article using at least two of the animals which have been reported upon here. It is be- cause of this that the brains of the animals now alive have not been taken out and illustrated, but it is expected that a full report upon them will be made at a subsequent time. The illustrations which are reproduced are selections of photographs of the animals at different stages. Some of the original photographs were small, and they were enlarged. To make the illustrations stand out well, the backgrounds on some of the negatives were " blocked out" and prints made from them in that condition. In other cases after prints had been made the figures of the animals were cut out. From the photographs without backgrounds the illustrations have been made. Al- though some of the fine detail at the edges of the prints are lost, the principal characteristics have been retained unchanged for there was no retouching of the negatives beyond the changes in the background. PLATE 1 Fig. A. Monkey 1, immediately after the second operation, after the recovery of the right side, showing the paralysis on the left. Note the manner in which the left hand is turned under and the left leg outspread. Fig. B. Monkey 1, after the second operation. The left arm is now being used as a prop as the animal sits upright. Figs. C. and D. Monkey 1, after second operation, showing the ability of the animal to use the right (recovered) hand for holding a strap which it is trying to chew. In both illustrations the animal is shown after the strap had been pulled upon so that the animal was irritated. Note the apparent helplessness of the left arm and leg, both of which members are flaccid. Fig. E. Monkey 2, after the second operation. The left arm of the animal, which has been just paralyzed, was accidentally caught in the strap which was pulled upon to get the animal to sit upright. The right arm is bandaged to the body. Note the utter helplessness, there being no effort to get the left arm out from the restraining strap. Fig. F. Monkey 2, after the second operation. Note that the animal has now recovered the ability to use the left hand and arm to a certain extent, since it holds the strap. Fig. G. Monkey 2, after the second operation. Note that the animal now uses its left hand and arm with apparent ease to support itself in walking. Fig. H. Monkey 2, 26 days after the second operation. The animal was pho- tographed in the act of attempting to take hold of the irritating strap. Note that the left arm is now used in a normal manner. Fig. J. Monkey 3, after the second operation. Note that the right side has recovered, but that the newly paralyzed left is badly used. The left arm shows wrist drop, and the small toe on the left foot is turned under. Fig. K. Monkey 4, after first operation, 26 days. Note that the right arm is used well in holding to the strap and that the right leg takes a normal position when the animal is sitting. 48 CEREBRAL MOTOR CONTROL ROBERT OGDEN AND SHEPHERD IVORY FRANZ PLATE 1 49 THE EFFECT OF DELAYED FEEDING UPON LEARNING V JOHN B. WATSON From the Psychological Laboratory of the Johns Hopkins University In the Psychological Bulletin of February, 1916, I gave the results of some experiments on the effect of delayed feeding upon rapidity of habit formation. The experiment there summarized was carried out in the psychological laboratory of the Johns Hopkins University in the winter and spring of 1915. Six male rats approximately one hundred days of age and six male rats approximately sixty days of age were required to learn the simple problem of entering the food box shown below (fig. 1). The problem box and restraining cage were of the same gen- eral character as figured in my book.1 Two modifications of the problem box there figured were essential: the first was to pro- vide a means of keeping the animal from going back into the restraining cage after it had solved the problem by entering b. This was accomplished by means of a thin metal shutter held down by a light string. The shutter was made slightly larger than the opening. It was held open by the experimenter through the aid of a string passing to the outside of the restraining case. After the animal's tail had cleared b the string was released and the spring closed the shutter. The time was taken from the moment the animal passed through a the opening into the re- straining cage until its body had cleared b. The interval of time the animal spent in the under-floor space, passing through c, etc., was not recorded. The second modification was necessi- tated by the fact that a means had to be provided for restraining the animal from getting its food until a definite time interval had passed. This was accomplished by making a cylindrical 1 Behavior, Henry Holt and Company, 1914, pp. 94 ff. 51 PSYCHOBIOLOGY, VOL. I, NO. 1 52 JOHN B. WATSON food box d, 5 cm. in diameter and 8 cm. high. This was sup- plied with a lid e perforated with several 1 mm. holes to allow the possibility of olfactory stimulation. A small vertical rod was screwed into the center of the lid. The rod passed up through a hole in the wire mesh of the problem box and restraining cage (the sleeve / was needed to keep the animals from attempting to push through the hole in the mesh). By means of this rod the experimenter could allow the animal to get food at the desired FIG. 1 time. The food cup had a raised bottom below which was cut an inside screw thread. A disc supplied with the same sized threads was screwed into the wooden floor of the problem box. The metal food box could thus be fastened securely into this floor plate. This device afforded a firm support for the food box and also made its removal for cleaning quite easy. The firm support was found to be necessary because when the animals were not fed immediately they attacked this box with the utmost vigor. EFFECT OF DELAYED FEEDING UPON LEARNING 53 The twelve animals were divided into two groups. Six of these had to learn the problem by the usual method of immediate feeding. Six others were required to learn by the method of de- layed feeding. The animals chosen were laboratory pets and exceedingly gentle. All twelve were bred from stock which had been in the laboratory for a long time.2 The twelve animals were distributed at random into the two groups before the first trial. The only precaution taken was to see that each group con- tained three young animals and three of the older ones. The early trials (table 1) show that the initial ability (after the first trial, which need not enter our records because of the fact that it was not until after the first solution of the problem that any difference in procedure was introduced) of the two groups was not very different. All twelve animals were allowed to get their food in the box before the first trials were given. During this preliminary habituation stage the opening at b was closed so that the explorations of the animals were confined to the inside of the problem box proper. A hinged door in the top of the box permitted the animals to be lifted in and out. In the regular trials the problem box was banked up with sawdust on all four sides to a height of four inches. At all times (and with both groups) during the tests the lid to the food box was left on. After a given animal, working by the immediate feeding method, had scratched away the sawdust and entered 6, the door was closed and the food box immediately opened. It was always opened by the time the animal could pass up through c. The rats were allowed to eat for five seconds and then they were lifted out and taken back to their living cages. Only one trial per day was given. Exactly the same method was adopted for the group working by the delayed feeding method except that in this case the lid to the food box was held down for thirty seconds. The behavior of the animals working under the ordinary con- ditions offers nothing worthy of comment. The behavior of the group whose feeding was delayed for thirty seconds presents an unusually difficult problem to those who hold that the getting • 2 1 wish to thank Dr. Helen Hubbert for supplying me with these animals. 54 JOHN B. WATSON of the food, following usually immediately upon the completion of the last (" successful") act, stamps in that movement. The successful act was to dig away the sawdust in exactly the right place and enter b. But these animals after entering b ran im- TABLE 1 Showing complete records of the twelve animals used in the experiment. One trial per day given TRIALS DELAYED FEEDING AVERAGE DE- LAYED FEED- ING AVERAGE IM- MEDIATE FEEDING IMMEDIATE FEEDING AVERAGE BY 3*8 DELAYED FEEDING AVERAGE BY 3*8 IMMEDI- ATE FEEDING 60 days of age 100 days of age 6 0 days of age 100 days of age 1 2 3 1 2 3 1 2 3 1 2 3 1 400 1190 160 360 177 25 385 923 2545 656 345 265 1680 40 2 40 100 25 92 55 40 58 111 125 169 50 46 145 135 3 155 40 20 11 42 240 84 89 193 93 18 105 85 40 176 374 4 40 15 11 25 29 63 30 24 27 25 10 25 55 6 5 57 7 5 10 7 25 18 36 33 15 12 22 7 129 6 36 4 25 5 20 12 17 23 7 7 65 26 4 30 22 28 7 27 12 7 18 7 5 13 9 7 4 13 12 6 11 8 12 27 3 15 5 8 12 13 5 5 46 7 8 5 9 25 10 5 9 3 5 9 8 4 21 10 4 5 5 11 10 10 20 3 11 7 3 12 9 15 42 19 7 15 4 5 11 30 10 20 7 4 10 13 9 18 6 5 13 6 3 12 20 4 16 3 4 9 8 13 13 11 6 6 5 9 10 11 13 3 7 3 7 6 20 8 9 5 5 4 20 6 13 14 14 9 10 7 2 15 9 7 6 10 3 7 3 12 15 15 30 10 6 2 3 11 24 13 70 47 7 3 5 9 13 16 10 15 8 5 3 3 7 8 7 25 2 4 4 3 17 5 9 3 10 3 o 6 9 12 8 25 3 4 4 18 3 8 23 7 5 8 6 17 6 3 5 3 3 7 8 19 18 16 10 7 4 10 5 6 8 4 8 3 3 20 3 25 8 11 2 9 5 6 K u 8 6 4 3 21 10 15 7 10 2 8 6 6 6 3 10 6 4 9 5 22 2 12 .12 K c 3 A 6 6 5 6 3 6 6 12 23 10 9 7 12 3 10 8 7 4 13 5 11 6 3 24 8 9 7 7 4 L 6 7 7 18 4 7 4 4 7 7 25 8 11 c 7 ^J 4 6 8 10 21 4 K O 4 3 26 27 6 7 6 9 c t. 4 7 6 7 7 6 7 5 9 7 9 14 r tj e O 3 3 7 4 6 7 I I/ 7 u • mediately to the food box. Finding it closed they became fran- tic. They would fight the rod, tear at the box, then they would leave the neighborhood of the food box momentarily, pass back through c to the under-floor space, and then return to the food EFFECT OF DELAYED FEEDING UPON LEARNING 55 box. Often considerable pressure had to be exerted upon the rod to keep the lid on. Their acts were very rapid — too rapid for the recording of accurate notes. Under such conditions thirty seconds is a long time and a great tax upon the experi- menter. There is a strong tendency on his part to let the ani- 2 46 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 <0 mals have the food a bit too early. For this reason the experi- mental technique here is not ideal and automatic control for all the variable factors should have been introduced. Notwithstanding this weakness in technique some interesting light is thrown upon this general problem. Below I give sepa- 56 JOHN B. WATSON rately the complete individual records of all twelve animals, the averages by trials, and the average of each succeeding three trials. The curve is plotted from the daily average. After the above records had been obtained by the one trial a day method, it occurred to me that a disturbance might result if the animals were given two trials per day. I then gave the test for three days, each animal receiving two trials in immediate succession. For convenience in comparing the records, table 2 is plotted as a continuation of the graph of table 1. The graph of table 2 begins at A., figure 2. TABLE 2 Showing effect of giving two trials daily DELAYED FEEDING GROUP il ^ IMMEDIATE FEEDING GROUP § M H 0 TRIALS 60 day animals 100 day animals W Q « * 0 60 day animals 100 day animals 1 2 3 1 2 3 |1 E^ 1 2 3 1 2 3 28 5 5 9 5 2 3 5 4 14 13 19 6 3 3 29 5 3 9 6 2 4 5 5 5 6 5 5 4 5 30 3 4 9 5 2 7 5 5 6 5 5 6 2 3 5 5 31 4 5 5 3 3 3 4 5 9 5 4 4 2 5 32 3 3 5 6 2 4 4 4 3 3 3 4 2 7 33 2 3 7 2 2 2 3 4 5 3 3 3 5 5 3.7 4 It will be seen that there was a steady continuation of the learning process and that no evil results followed from this short test. Attention is called to the fact that animals 1, 2, and 3 of the immediate feeding group were slow on the first trial on the first day. This was undoubtedly due to the fact that some wild rats ran over the problem box during the night. These three records were not included in the averages. It occurred to me finally that it would be desirable to see what would happen if the feeding methods were reversed for a few trials. Accordingly three of the animals from the delayed feeding group (the three sixty day animals) were forced to run for eight trials receiving their food immediately at the end of each trial. On the other hand three of the animals from the immediate feeding group (the three sixty day animals) were forced to work by the delayed feeding method. EFFECT OF DELAYED FEEDING UPON LEARNING 57 Taole 3 gives the results of the first eight trials (table 3). The graph of table 3 begins at B, figure 2. It will be seen from the averages of these few trials that nothing significant appeared. Looking at these records we see that no matter what our cri- terion or standpoint may be the fact remains that the delaying of the feeding for thirty seconds after the solving of the problem did not alter the learning process. I think it extremely unlikely that two groups working by the same method would ever show a more closely similar time record. It may be that experimenta- tion with a larger number of animals and with automatic con- trols might show a significant difference. It is quite possible that a longer period of delay might give far different results. TABLE 3 Showing effect of reversing the method of feeding TRIALS DELAYED FEEDING GROUP WITH FEEDING RELATIONS REVERSED 0 8 .M AVERAGE OS IMMEDIATE FEEDING GROUP WITH FEEDING RELATIONS REVERSED 60 day animals 60 day animals 1 2 3 1 2 3 34 6 3 10 6.3 3 3 5 35 4 3 5 4.0 4.3 7 3 3 36 4 5 5 4.6 4.3 5 3 5 37 4 11 3 6.0 5.3 7 6 3 38 3 2 5 3.3 3.3 4 4 2 39 2 3 5 3.3 4.3 6 4 3 40 6 3 6 5.0 5.3 3 6 7 41 2 3 5 3.3 4.3 4 3 6 It would of course be desirable to have some system of recording the number of random movements and the type of these move- ments. This ought to be done especially during the delay per- iod. So far as the experimenter can state, the animals were working as vigorously during the thirty seconds delay as at any other time and they were displaying the same type of movements, viz., exercising their instinctive and habitual repertoire. I assume that what psychologists mean when they say that a movement is stamped in by reason of the " satisfaction" which it brings and that another movement is stamped out by reason 58 JOHN B. WATSON of the " dissatisfaction" it brings could be stated somewhat as follows: The getting of the food produces metabolism, increased circulation, changes the tone of the organism, supplies nutrition to the blood stream, etc. Now the whole system of neuro-mus- cular arcs exercised throughout the solving of the problem does not share equally in this bettered condition of the organism. The particular arc last functional (the one employed in the successful act) by reason of the fact that 'activity in it has not completely died down' will reap the greatest benefit from the bettered func- tional condition of the organism. .This in the long run would tend to favor the successful arc at every trial. I do not mean to say that any of them would state it in just this way. As a matter of fact none of them has ever given even a fair presenta- tion of just what they do mean. Swift has given the nearest approximation to this statement. Thorndike has advocated the satisfaction and dissatisfaction theory most consistently but he has not attempted to give even a crude physiological basis for his views. While I have no solution to offer I cannot help but see in the experiment which I have just reported a serious objection to any such formulation. I think we may assume without exaggeration that there are from ten to fifteen complex acts performed during the thirty seconds delay. All of these acts come after the so- called successful act, i.e., the scratching away of the sawdust at the proper place and the entering of b. As an interesting specu- lative point one should consider the average total time of the solving of the problem. After the fourth trial the average time drops below thirty seconds. Thus from the fourth trial on more useless movements occur in any given trial after the problem has been solved than occur prior to the actual solution of the problem! Why should not the neuro-muscular arcs used in executing these later random movements be the ones to share in the bettered physiological condition of the organism? I offer this experiment not as throwing any conclusive light on the learning process but as opening up the possibility of carry- ing out experimental work upon the fixation of arcs in habit and as showing the very great need there is for such study. The EFFECT OF DELAYED FEEDING UPON LEARNING 59 control of habit is one of the most vital problems in every system of psychology. The answering of this question ought to enable us to attack the problem of habit control in a far more scientific manner than is now possible. DISCUSSION INTERNAL SECRETION IN LEARNING KNIGHT DUNLAP The Johns Hopkins University Non-psychological writers, and many psychologists, assume that emotion has a direct influence on action, and in particular, that pleasure (or satisfaction) and pain (or dissatisfaction) are instrumental in the formation of habits. Certain writers, how- ever, have objected to this assumption on various theoretical grounds, the most important of which, at the present time, is the alleged fact that no detailed mechanism is discoverable or has even been suggested, by which the effects of emotion on action might be mediated. This omission I have had it in mind for several years to supply by a hypothesis which seems to offer grounds for experimental test: but as I shall be able to carry out the tests on only a few points, it seems proper to outline the hypothesis for the consideration of others, who may be ex- perimentally interested in the matter. The hypothesis I have in mind is a Iqgical outcome of the view to which some of us were earlier forced concerning the emotions, namely, that in the important bodily changes which are com- monly called " expressions of emotion" (and which I, following Lange, would insist are the emotions), the activities of certain, probably all, of the endocrine glands play a part. Five years ago I assumed that this would be found true of the " major emo- tions" (or what are sometimes called " emotions" as distinguished from " feelings"), as Cannon has so admirably shown. At the present time, I have no hesitation in adopting it as a working hypothesis for the " feelings" of pleasure and pain, and all other definite affects. There are certain cases in which the effect of pleasure in "fixing" a reaction can be explained by the immediate repeti- 61 62 KNIGHT DUNLAP tion of the act, or by the dwelling in thought on the act, which is probably physiologically equivalent to a repetition. These cases come under the heading of repetition or frequency, which, so far as I know, no one doubts to be an important factor in association or habit-formation. The cases which cannot be subsumed under this head are the ones in which the efficacy of pleasure has been challenged, and these are the ones with which we are at the present moment concerned. If pleasure (to neglect pain for the moment) is directly con- nected with a change in internal secretions, and if internal secre- tions may act on the nervous mechanism (both of which condi- tions are possible), we have in this aspect of pleasure a possible means of influencing habit formation. Since the reaction which we suppose to be " fixed7' by pleasure precedes rather than fol- lows the pleasure; or at least precedes the hedonic secretory effects — as we must allow a time interval for the secretion to be carried in the blood stream to the effective locality — the influ- ence of whatever hormone is involved is retroactive, i.e., it will act on a pathway over which discharge has occurred in such a way as to make discharge over that pathway more ^probable in the future than it was before. In brief, the obviously suggested theory is that the nervous discharge leaves an arc or certain important pbints in the arc in such a condition chemically, that a certain substance (hormone) may a few moments later "fix" it. Artificial as this theory sounds at first, I believe it is worth putting to the test. I might point out familiar observation, and data from the experimental investigations of learning, which fit this working hypothesis. There are many pertinent cases. I believe how- ever that this procedure would not be legitimate, since the ob- servations which support my hypothesis were not made with this hypothesis in view, and hence the strong backing they afford may be more apparent than real. The important thing is that experimenters should hereafter keep the hypothesis in mind, and observe specifically in future work the data which have direct bearing on it. The particular efficacious homone which is liberated in pleasure INTERNAL SECRETION IN LEARNING 63 is conjectural. It can hardly be adrenalin, for this, as Cannon's experiments seem to show, is the endocrinic correlate of excite- ment, which is not conducive to habit formation, but rather to the breaking down of habits. It must be a secretion, which like adrenalin, is discharged directly into the blood (not indirectly through the lymph channels), by which it is carried to the " centers'7 in which habits are formed, i.e., in which the critical synapes lie: unless indeed the route may be still more direct as seems hardly possible, even from the pituitary body. It may be, however, that the secretion is not formed in a " gland" proper, but in some tissue whose primary function is not secretion. The effects of pain, in preventing the fixing of the preceding acts, may not be so specific as are the effects of pleasure. It is possible that adrenalin or some other active principle is the negat- ing agent here, but it is also possible that the effects are produced by the setting up immediately of more powerful reactions which disturb the interconnections left by the preceding algesogenic reaction. By "pain" is here meant the affective content usually (and properly) described by this term, ignoring the unfortunate psychologists' confusion between this and certain specific sen- sations. The implications of the theory which admit of experimental verification, or the reverse, are numerous. Those in which I have been most interested in are the following. 1. Actions performed shortly before the reaction which pro- duces the " isatisf ying" result, and actions immediately follow- ing it, would be fixed, along with the act itself. The normal pause in activity following the " satisfying" reaction (where the reaction itself is not immediately repeated) is probably a useful phenomenon. 2. If an animal, in solving a simple "problem" makes a short series of reactions, including a number of "wrong" acts and terminating with the correct (satisfying) act the probability of repetition of the "wrong" acts is as great as that of the "right." But after the solving several times, the probability of the" right" act becomes greater than that of any "wrong" act unless a 64 KNIGHT DUNLAP " wrong" act has been in every series. In that case, the animal should eventually repeat the "right" act uniformly preceded by the "wrong" one. 3. If the apparatus is so disposed that satisfaction is not given to the animal until several "wrong" acts have been done after the act which really makes the satisfaction available, learn- ing will be made especially difficult, unless the animal is able to make a conceptual analysis of the problem. 4. In a problem involving the necessity of a definite series of actions for its solution, and allowing the performance of "wrong" acts at various points, the elimination of wrong acts will follow no law except that of probability, based on a num- ber of repetitions. Of the wrong acts which have been done up to a certain point in the learning, those which have been done the fewest times will be eliminated soonest. 5. In the human animal, in spite of the non-emotional fac- tors which enter most learning problems, the hormone factor should be discernible. CONTINUOUS STIMULATIONS VERSUS TRANSITIONAL SHOCK IN THE PHOTOTACTIC RESPONSE S. J. HOLMES From the Zoological Laboratory of the University of California In a paper by Miss K. W. McGraw and the writer (1) a de- scription is given of several experiments which were designed to ascertain whether the phototactic response is due to the con- tinuous stimulating effect of light or to the stimuli »hat result from changes in the intensity of light. The ordinary photo- tactic movements of animals afford an opportunity for both these forms of stimulation to come into play. An animal going toward or away from the light naturally deviates more or less from a straight path and hence subjects its photosensitive sur- faces to more or less frequent changes of light intensity. In order to separate the effect of continuous stimulation from that of the shock of transition the attempt was made to eliminate the latter so far as possible and thus to observe the influence of constant stimulation alone. In one set of experiments insects with one eye blackened over with asphalt varnish were placed within a small circular enclosure in the center of a cylindrical container whose sides were lined with uniformly colored paper. Light was admitted through the center of the upper side which was also uniformly colored, so that in whatever way the insect turned the amount of light entering the eye would be approximately the same. In general, insects that were positively phototactic performed circus move- ments toward the normal eye when placed in the enclosure, while negatively phototactic insects turned about in the reverse direction. The movements of the insect doubtless produced some changes in intensity of photic stimulation, but they were slight; nevertheless the behavior of the insects within the en- closure was practically the same as when they were outside and 65 PSYCHOBIOLOGY, VOL. I, NO. 2 66 S. J. HOLMES exposed to the varying influence of lights and shadows from a multitude of objects. In other experiments insects were held above an easy-running horizontal disk or turn-table which they could rotate with their feet. Light was made to fall upon the insect from one side. If the insect attepted to turn toward the light the movements of its feet would rotate the disk in the reverse direction. Butter- flies proved to be very convenient forms to work with since by grasping them with their wings held together above the back they could be easily held above the disk so that their feet could move it in either direction. Under these conditions several species "of butterflies were found to rotate the disk quite con- stantly away from the illuminated side in their efforts to turn toward the light. They would also promptly change the di- rection of rotation when the light was carried from one side of the insect to the other. Some species of Diptera gave results quite parallel to those obtained with the butterflies. As was pointed out, the slight movements of one's hand, although these were minimized by the use of a hand rest, and the movements of the head of the insect due to its own activities in rotating the disk, would produce some fluctuation in the amount of stimulation received by the eye, but these changes were necessarily small. The general outcome of both sets of experiments was therefore favorable to the view that the con- tinuous stimulating influence of light is a potent factor in effect- ing orientation. In a paper on the reactions of Euvanessa antiopa to light Mr. W. L. Dolley (2) has questioned the conclusion just expressed, and has described some experiments whose results in his opinion justify a different interpretation. Dolley's work was done under the supervision of Prof. S. O. Mast whose views regarding the so-called "continuous action theory' ' he apparently shares. The apparatus that Dolley employed, consisting of an easy- running horizontal disk above which the insect could be held, was in principle the same as ours. The wings of the butterfly, however were held in a clamp, thus getting rid of one source of movement. In other respects the body and especially the head CONTINUOUS STIMULATIONS VERSUS TRANSITIONAL SHOCK 67 were free to move and naturally changed in position somewhat as the insect became active. By observing the shadow of the head when the insect was moving its legs it could be seen that the head did in fact move up and down. Whenever the insects were active, however, "they attempted to turn toward the func- tional eye, never in the opposite direction/' . The results of the experiments, so far as the nature of the orienting stimulus is concerned, are regarded as not conclusive, "for the moment the animals become active, and before they attempt to turn, there is a change in the position of the eye owing to the vertical movements of the head, and this, no doubt, results in changes in the luminous intensity on the various ommatidia. Thus, it is evident that the attempt on the part of the animal to turn toward the illuminated eye may be due to stimuli dependent upon the time-rate of change of intensity." So far as the but- terflies responded to light at all, their behavior, while rather spasmodic, was in essential agreement with what was described by Miss McGraw and myself, only as the creatures jiggled somewhat during the experiment it was held that the conclusions drawn in our paper were "not justified." In order to test the question still further and to obtain results which would permit of only one possible conclusion, an apparatus was devised which would hold the eyes of the insect in a firmly fixed position. If the eyes could be maintained in a perfectly constant relation to the light the element of transitional shock or differential sensibility would of course be eliminated. To secure this end a piece of wood was fashioned into the shape of an L with a long and a short arm. The wings of the butterfly could be clamped on either side of the upper arm which was held horizontally, while the body lay below with the head pointed toward the shorter arm. Into the shorter arm a nail with a flat circular head was driven a short distance from the end. The butterfly was so placed that the front of its head abutted against the head of the nail. By means of a quick drying glue the butter- fly's head was cemented to the head of the nail, and left there until the glue dried. With the wings firmly clamped above, and the head firmly glued to the head of the nail, the activities of the 68 S. J. HOLMES insect could only produce a limited movement of the body, but no movement of the head. The wooden L was then clamped upon an iron frame so as to bring the insect in a horizontal posi- tion, and the feet of the insect were allowed to rest upon a light, horizontal disk capable of easy rotation. Thtf light employed rested upon a firm stand so that no complication could arise from that source. Care was taken to see that the glue was firmly set and that the head made no motions during the strug- gles of the insect. Everything was made as rigid as could well be, so that one could be assured that eye and light must remain in a constant relation, however much the insect might move its legs or even its body. Under these conditions, which would seem to insure the con- stant stimulating influence of light, how did the butterflies react? Specimens of Vanessa caryce were chosen for experi- mentation since they were easily obtainable. The first individual selected was found to rotate the disk quite consistently away from the light. When the light was placed on the other side of the insect the disk was rotated in the reverse direction. Fre- quently the movements of the insect were soon discontinued, but when the foot or body was touched with a needle the rotation of the disk was generally resumed. The light was changed from one side of the insect to the other fifteen times and each time the disk was unmistakably rotated away from the light. After a rest in the dark for about an hour the light was changed from side to side twelve times with the same result as before. The insect which now gave evidence of exhaustion was liberated; the glue fastening its head to the apparatus was found to be well hardened. Other butterflies of the same species were experimented with in the same manner and gave results equally decided. The responses of the butterflies were not quite so vigorous as in the experiments originally described. This is probably due to the unnatural stimulation caused by the head being firmly glued to a foreign object, and also to confinement in a fixed position during the time the glue was being dried. Nevertheless the results made it abundantly evident that the butterflies with the CONTINUOUS STIMULATIONS VERSUS TRANSITIONAL SHOCK 69 head firmly fixed showed a marked tendency to rotate the disk away from the eye receiving the most light. Inasmuch as the conditions of the experiment were such as hardly to leave any room for fluctuations of light intensity to play any part, the only reasonable interpretation of the results is that they are due to the continuous action of the light. REFERENCES (1) HOLMES, S. J. AND McGRAw, K. W. : Jour; Animal Behavior, iii, 367-373, 1913. (2) DOLLEY, W. L.: Jour. Exp. Zool., xx, 357-420, 1916. THE EFFECTS OF CEREBRAL DESTRUCTION UPON HABIT-FORMATION AND RETENTION IN THE ALBINO RAT K. S. LASHLEY AND S. I. FRANZ Department of Psychology of the Johns Hopkins University and the Government Hospital for the Insane INTRODUCTION It has long been known that the cerebrum is anatomically complex and it is recognized that its functions are diverse. There is also the appreciation of the failure to correlate the known anatomical facts with the functional observations. There are many reasons for the last. The variation in cerebral struc- ture in different species of animals is one of the greatest of the difficulties in the correlation of the anatomical and the func- tional, for two animals with obviously dissimilar brain struc- tures may appear in the main to be functionally similar, and conversely two animals with obviously dissimilar activities may have quite similar cerebral structural arrangements. The di- versity of structures in the same brain has been more recently emphasized. Differences in the arrangements or in the method of combination of cells and fibers have been described as points of major importance, and many cortical areas, a score or more, have been described as different organs. They have also been assumed to represent equal functional differences. On the func- tional side, especially when man has been under consideration, the diversity has been exaggerated, and this has probably been carried over into modern neurology from the mental faculty aspect of GalPs phrenological system, although the anatomical part of Gall's work is totally discarded. The differences of emotions, hate, joy, fear, anger, of sensations such as color, tone, and the taste and smell qualities, of movement and of will, of desire, of interest, and of all mental states, have led to assump- 71 72 K. S. L ASHLEY AND S. I. FRANZ tions of uniqueness by the uncritical in psychological analysis. There followed the further belief that each one of the supposedly different mental states must have a corresponding anatomically distinct counterpart. When these beliefs are looked at critically they are recognized to be logical from the standpoint of any of the current brain- mind hypotheses. But this logical conclusion applies only to the general and not to the particular way in which the differences are looked at. It is doubtless true that there are many arrange- ments of cells and fibers in different parts of the cortex, and these differences in arrangement may mean differences in function. That great differences in function are not necessary consequences of anatomical differences may be illustrated, but of course not proven, by an analogy. It is well recognized that a house may be built of wood, or of bricks, or of stone, and that it may be one or two or more stories in height, be narrow or broad, or with large windows and doors, or with Venetian blinds, or painted blue, or with a gable roof, or with any other characteristics. But, whatever these anatomical characteristics may be the struc- ture remains a place of habitation. The differences in structure do not indicate fundamental functional differences of the col- lections, they are modifications of only certain minor elements, and point only to minor functional differences. The function of the habitation may, therefore, differ slightly because of its com- binations of internal and external structures, for it may be more comfortable in winter, or be conducive to cleanliness, or the reverse. It still remains a house with the general functions which all such structures have. The windows let in light, the doors permit entry and exit, the walls protect, and the furnish- ings give comfort, to different degrees, but the same general functions pertain to the cottage, to the mansion and to the apartment house. The functional similarities are more pro- nounced than the structural differences. Equally with mental states and the cerebral anatomical struc- tures. We may discover by careful analysis that although the mental states differ they are not. so divergent as a superficial observation would indicate. There are many similarities, many CEREJBRAL FUNCTION IN THE RAT 73 common elements, in apparently dissimilar mental conditions. We also find that the cerebral structures do not differ widely in fundamentals. Many of the cells have the same general mass and appearance, but in one place may differ in number from those in another location. Their staining qualities are similar, the neurofibrils are much the same, the most marked differences are the direction and the length of the axones and the number of collaterals. Consequently, it should be obvious that great care should be taken that points of minor difference are not made the basis for major deductions. At the same time care should be exercised that apparent simi- larities are not accepted as equalities and that a certain condi- tion should not too hastily be assumed to point exclusively to one explanation. But this mistake is one of the most common in dealing with cerebral functions. Von Monakow has empha- sized the distinctions which must be made in certain cases, and has pointed out some of the difficulties of interpretation. In one direction these difficulties are to be appreciated by the care- ful consideration of the temporary and the residual defects fol- lowing a cerebral accident. But there are many more. For example, after the destruction of certain parts of the cerebrum of an animal, if a beam of light is thrown into the eye, or if an object crosses the visual field, and the eyes turn in the proper direction one cannot be certain whether the animal sees or the reaction is a reflex without vision. For cases such as that just mentioned we have a means of more exact determination. This is the method of training which can be used both prior and subsequent to any experimental destructions. We may set such conditions to an operated ani- mal which exhibits a reaction to light stimuli that the animal will be able to form a habit provided the cerebral sensory, asso- ciational (if they exist), and motor elements be intact. If a necessary part of the sensory-motor mechanism be out of order such a habit can not be acquired. By varying the end condi- tions, either the sensory or the motor or both, it is possible to deduce the effects of destructions if they be more than tem- porary, and only in this way is it possible to determine accu- 74 K. S. LASHLEY AND S. I. FRANZ rately the character of a certain loss. Such a defect might apparently be motor. If this should be so we have reason to believe that a required movement could not be initiated from the cerebrum by any form of stimulus, whether light or sound or touch. If, however, the defect be sensory, let us say visual, the problem changes. A stimulus of that particular kind would not serve to produce a habit of any character, whereas other sensory stimuli (sound, touch, smell, etc.) could be used as the means to produce habits. By changing the conditions both on the sensory and motor sides it is also possible to discover what may be termed an " association" loss, although this is in some particulars more difficult than the sensory and motor losses. After a cerebral accident in man there is also the possibility of determining the degree of retention, the characters of the retained habits, and the capability of formation of new habitual modes of reaction, i.e., behavior. In the clinical examinations of man chief, and frequently exclusive, use is made of the speech habits, although the exclusive employment of speech as the mode of reaction has certain disadvantages. They are not always obvious and they appear to be little appreciated. One of these is that an inability may be reported by speech which by other means it is possible to discover does not exist. This fact is recognized in certain well developed cases which in the past have been called hysterical. It is now admitted that in many other individuals similar " hysterical" symptoms may exist without there being easily recognized signs. Diverse mental states, such as dementia, great joy, or even a lack of understanding, may result in variations in speech habits which mask other effects. This being so, it is clear that to accept as final evidence of a defect a negation (by speech) is similar to the conclusion that an animal sees if the eyes are turned towards the stimulating light. The experimental method should be extended and the habit reaction method especially should be applied to the determination of defects in man after cerebral lesions. The limitation of clinical studies on cerebral functions to the consideration of speech reactions as the motor side of learned reactions can in many cases result only in superficial CEREBRAL FUNCTION IN THE RAT 75 knowledge. The adoption of methods looking toward the deter- mination of other habit forms will result in more complete information. Although it is commonly believed that habit formation is due to the functioning of cerebral parts, and especially those parts which are called cortical, it is by no means right to conclude that all habit reactions are due to cerebral, or more restrictedly cor- tical, activities. Leaving aside those reactions which must pass by way of subsidiary stations, like the thalamus, as part of the total reaction, it is doubtless the case that certain habits are carried out by the exclusive use of non-cortical, and exclusively also perhaps by non-cerebral parts. This is shown in the reactions of the so-called lower animals, and it has been contended that for the execution of some long-established habits in higher animals only non-cerebral parts are needed. There is reason to believe that in the brainless frog certain simple habits may be acquired. One of the urgent needs at the present time both in neurology and in the study of learning is the establishment of the parts played by the different portions of the nervous system. With these facts in hand it will not only be possible to under- stand something of learning and forgetting, but at the same time we shall be better able to appreciate that inadequately named condition which is called cerebral vicarious functioning. Two main problems were in mind when the following work was begun. One was the effects of different lesions upon habit formation and retention, the second was the determination of the parts needed for habit formation. These problems are identical in some particulars, but they can also be considered to be independent. It may be that in a normal animal a habit is formed by the interaction of certain nervous elements, but that when any of these elements have been destroyed or even inter- fered with other elements may take their places. Or, it may be that in the process of learning many different elements are used at first, but that the number decreases as the habit is acquired. The scope of the experiments described here has been limited to a determination of the relation of the frontal pole and dorsal 76 K. S. LASHLEY AND S. I. FRANZ convexity of the cerebrum to the formation and retention of habits which involve chiefly responses to tactile and kinesthetic stimulation. After operation to destroy the frontal pole of both hemispheres (section III) or the entire dorsal convexity of the cortex (section V) the animals were trained on the in- clined-plane box or maze and their rate of learning was compared with that of normal individuals. In other experiments (sections I, II, and IV) normal animals were trained upon the same prob- lems and after the destruction of various areas of the cortex were tested to determine the degree of retention of the habits pre- viously established and their ability to reform the habits in the cases where these had been lost. In the carrying out of the operative procedures excessive hemorrhage was not infrequently encountered. To avoid this as much as possible the operations were performed rapidly. The sizes of the brain and skull are such that there is no space for packing to check a hemorrhage, and it was thought best to operate quickly even if some animals died because of the result- ing hemorrhage from the cerebral lesion, rather than have an equal or greater number of deaths from hemorrhage from the longitudinal sinus. This was possible because of the number of animals which were available. Most of the cerebral operations were carried out through small skull openings, and partly be- cause of this the lesions differed in all cases. Slightly different positions of the heads of two animals make the views of the small operative fields dissimilar, variations in the sharpness of the section knife make distinct variations in the ease, depth and completeness of the intended sections, and even slight variations in the relations of the brain to the skull sutures give difficulties that are not easily overcome. The complete removal of the top of the skull and the consequent full view of the superior surface of the brain would have been a more satisfactory pro- cedure in some ways. The difficulty of controlling the hemor- rhage from the skull is however added to the difficulty of con- trolling the hemorrhage from the cerebral section. At the same time even though the top part of the skull be removed without any ill effect other difficulties persist. The point of insertion of CEREBRAL FUNCTION IN THE RAT 77 the knife can be more accurately determined, but slight varia- tions cannot be avoided and these would make the lesions dis- similar. In addition, the quantity and the extent of the hemor- rhage cannot be controlled, and this hemorrhage acts as a super- imposed lesion. In some cases the lateral ventricles were found to be filled with blood, and in other cases clots were found at the base and even at the cerebro-cerebellar junction. Mention may also be made here of the variations in the location of the stimulable areas (the motor areas), report of which will later be made by one of us (L). Whatever operative technique is used the destructions are varying, and the most that can be expected in these small animals is to operate on sufficient numbers, and to select and compare those cases which have relatively similar lesions. This we have done. In all cases the lesions will be described and the comparison of sets of animals will be made. When the experimental part of the work was completed the animals were killed and their brains were removed, fixed in 10 per cent formaldehyde, dehydrated, and cut in celloidin. Hori- zontal sections 80 M. in thickness were cut and every tenth section was stained in toluidin blue and mounted in balsam. Each brain was given an arbitrary number and the examination and description of the lesions were made without knowledge of the behavior data, except in the case of the extensive lesions which were recognizable in the sections. For reconstruction of the lesions serial outline sketches of the sections were made under the projection microscope to show the extent of the lesion visible under low powers of the microscope. The sections were then examined under higher power and the parts which seemed certainly involved in the lesion, either through actual destruction of the cortex or severance from all connection with the descending fiber tracts, were filled in on the sketches. The plane of the sections was determined for each brain by reference to the corpora quadrigemina and knee of the corpus callosum, measurements of the extent of the lesion were made, and these were reduced to the dimensions of the diagram used in case the brain varied from the average. Finally the extent of the lesion was marked on the diagram in the planes of the sections and the areas so outlined were blocked in. 78 K. S. LASHLEY AND S. I. FRANZ The brains were first turned over to a technician to be sec- tioned. As a result two brains were lost and several others so badly prepared that accurate reconstructions were impossible. Six brains were sectioned by the technician, the remainder by one of the writers. In the descriptions of the lesions reference is made to well defined land-marks, particularly the limits of the ventricles, the divisions of the corpus callosum, the gyrus hip- pocampus, and the subcortical ganglia. The relative positions of these structures are shown in the serial sections, figured in plate I, figure 22. I. THE RETENTION OF A SIMPLE KINESTHETIC-MOTOR HABIT AFTER DESTRUCTION OF THE FRONTAL POLE OF THE CORTEX In experiments described earlier (1) a number of animals were trained in the simple maze (fig. 1). An attempt was then FIG. 1. THE SIMPLE MAZE a, starting compartment; b, middle alley; c, cul de sac; d, alley leading to food. made to destroy the frontal portion of the cortex by transverse incisions through small openings in the parietal bones and as soon as the animals had recovered sufficiently from the opera- tion, usually within forty-eight hours, they were tested for retention of the habit. In thirteen of the fourteen animals that survived the operation there was demonstrated some re ten- CEREBRAL FUNCTION IN THE RAT . 79 tion of the habit, evidenced either by relatively few errors in the first trials of the retention tests or by survival of individual peculiarities of reaction to the maze. Details of these experi- ments have been given in the earlier paper. The extent of the cerebral lesion has now been determined for each animal and a report of the findings will be given in the first section of this paper. The experiments are here given the same numbers as in the first article and page references for each experiment are given to facilitate reference to the original description. A brief estimate of the degree of retention is included for each animal with the record of the total number of errors made during the first fifteen trials of training and the first fifteen of the retention tests. This is followed by the description of the lesions. In the diagrams prepared for this section of the paper we have figured only so much of the lesion as is obvious, either from the total separation of the injured parts from the remainder of the brain, the complete degeneration of the cortex, or the unmistak- able section of the descending tracts. In every case it is prob- able that the lesion affected a much more extensive area than that indicated but it has seemed best to restrict the description to the obvious destructions. After certain of the animals (experiments 8, 9, 10, 11, 12 and 13) had recovered from the effects of the operation and had given evidence of the retention of the habit a second operation, designed to destroy the temporal or occipital regions, was per- formed. The description of the second lesion is included with that of the first, in the animals which survived; the effects of the operation will be described in experiment 15. Experiment 1 (page 8). Almost perfect retention. Errors: learning 5, retention 1. Lesion (plate II, fig. 1). There is a transverse cut extending diagonally across both cerebral lobes. From above the knee of the corpus callosum on each side it passes out along the forceps of the callosum to the external capsule and thence to the cortex, 1.5 mm. behind the base of the olfactory bulb. Below, on the left, the lesion extends to the floor of the lateral ventricle, on 80 K. S. LASHLEY AND S. I. FRANZ the right to the base of the lobe behind the olfactory bulb. Practically all the cortex of both hemispheres lying in front of the knee of the corpus callosum is destroyed by the lesion. Experiment 2 (page 8). Partial retention. Errors: learning 8, retention 2. Lesion. Not enough sections of the brain were preserved for an accurate determination of the extent of injury. Apparently the lesion was superficial, extending over the antero-dorsal sur- face of the frontal poles. Experiment 3 (page 9). Partial retention. Errors: learning 2, retention 4. Secondary evidence of retention. Lesion (plate II, fig. 3). The brain was imperfectly sectioned. Right hemisphere. There is a transverse section of the hemi- sphere just in front of the forceps of the corpus callosum, extend- ing from the inner margin of the mesial cortex through the lateral cortex. Left hemisphere. The lesion is similar to that on the right but involves the mesial cortex as well. The posterior extent of the lesion could not be determined accurately. The destruction was certainly not less than that shown in the figure. Experiment 4 (page 9). Partial retention. Errors: learning 21, retention 2. Lesion (plate II, fig. 4). Right hemisphere. There is a complete transection of the anterior pole in a vertical plane passing through the knee of the corpus callosum. Left hemi- sphere. The cut is in the same plane as that on the right but penetrates only from the mesial surface to the external capsule and ventrad to the fibers of the peduncle. Experiment 5 (page 10). Perfect retention. Errors: learn- ing 19, retention 2. Lesion (plate II, fig. 5). Right hemisphere. There is a small injury on the mesial surface above the knee of the corpus callosum. Left hemisphere. There is a small cylindrical cyst extending from the dorsal cortex through the knee of the corpus callosum and the cerebral peduncle, severing the fibers coming from the frontal pole. Experiment 6 (page 10). Perfect retention. Errors: learn- ing 7, retention 0. CEREBRAL FUNCTION IN THE RAT .81 Lesion (plate II, fig. 6). The operation was by a single median opening. There is an extensive destruction of cortex at the opening in the skull and a transverse cut extending ventrad from it past the knee of the corpus callosum, destroying all the cortex lying between and in front of the forceps of the callosum. Experiment 7 (page 11). Retention nearly perfect. Errors: learning 11, retention 1. Lesion (plate II, fig. 7). Right hemisphere. There is a transverse cut extending forward from over the knee of the cor- pus callosum, through the forceps to the fibers of the peduncle. There is a large cyst in the mesial cortex of the frontal pole. The destruction is probably limited to the dorso-medial surfaces of the lobe. Left hemisphere. The lesion extends from the tip of the frontal pole along the course of the fibers to the end of the forceps, then turns laterad along the external capsule for a short distance and finally out along the course of the radial fibers to the cortex. Experiment 8 (page 11). Partial retention. Errors: learn- ing 7, retention 3. The brain was so damaged in sectioning that nothing could be made of the extent of the lesion. Experiment 9 (page 12). Nearly perfect retention. Errors: learning 23, retention 2. Lesion (plate II, fig. 9). Right hemisphere. There is a sec- tion of the hemisphere from above the knee of the corpus cal- losum diagonally forward to the base of the olfactory bulb. The lesion is superficial, destroying chiefly the mesial cortex in front of the callosum and above the olfactory tracts. Left hemisphere. There is a complete transverse section similar to that on the right but extending around onto the orbital surface. Experiment 10. (page 12). Nearly perfect retention. Errors: learning 33, retention 2. Lesion (plate II, fig. 10). Right hemisphere. First operation. There is a transverse incision extending ventrad in a plane pass- ing through the end of the forceps of the corpus callosum to the base of the olfactory bulb, destroying all the frontal pole, which is filled by a large cyst, and cutting a part of the olfactory tract. Second operation. There is a small incision extending down PSYCHOBIOLOQY, VOL. I, NO. 2 82 K. S. LASHLEY AND S. I. FRANZ from the dorsal convexity through the fornix and internal cap- sule and penetrating the thalamus in the region of the lateral thalamic nucleus. Left hemisphere. First operation. The lesion is similarly placed to that on the right but more super- ficial, passing just under the cortex through the end of the for- ceps to the base of the olfactory bulb. It narrows at the base and does not involve the olfactory tracts. The entire pole exclusive of the mesial cortex is destroyed, however. Second operation. There is a transverse lesion extending down from the dorsal convexity just in front of the hippocampus, cutting through the fornix and penetrating the crux cerebri for one half its width. Experiment 11 (page 12). Perfect retention. Errors: learn- ing 11, retention 1. Lesion (plate II, fig. 11). Right hemisphere. First operation. There is a transverse section of the hemisphere just back of the knee of the corpus callosum and extending diagonally laterad to the cortex along the forceps of the callosum and ventrad to the floor of the lateral ventricle. The ventricle is distended until it occupies one-third of the width of the hemisphere. Sec- ond operation. There is a superficial incision extending from above the anterior end of the lateral ventricle to the posterior border of the corpus callosum and laterad along the external capsule to the level of the thalamus. There is a large cyst replacing the fibers of the external capsule in the occipital lobe. Left hemis- phere. First operation. There is a transverse lesion just behind the knee of the corpus callosum extending downward to the level of the fornix. Internally it extends deeper, pentrating the olfac- tory fibers in the region of the anterior olfactory nucleus. Sec- ond operation. There is an incision passing in from the cortex to the external capsule near the posterior limit of the first lesion, thence extending backward along the external capsule for an in- determinate distance. The external capsule is largely replaced by a cyst extending into the occipital pole. The frontal poles of both hemispheres were completely severed from the rest of the brain by the first operation. Experiment 12 (page 12). Retention perfect. Errors: learn- ing 2, retention 0. CEREBRAL FUNCTION IN THE RAT 83 Lesion (plate • II, fig. 12). Right hemisphere. First opera- tion. There is a transverse incision just in front of the forceps of the corpus callosum severing all of the anterior pole, except the mesial cortex, down to the olfactory tracts. Second operation. There is a lesion over the dorsal and orbital surfaces with com- pletely degnerated cortex in the regions indicated in the figure. Left hemisphere. First operation. There is a lesion extending from the end of the forceps of the corpus callosum to the mesial cortex and ventrad into the olfactory fiber tracts. All of the mesial co/tex anterior to the knee of the callosum is destroyed but little injury has been done to that of the lateral face of the anterior pole. Second operation. There is a lesion over the dorsal and orbital surfaces of the hemisphere extending along the fibers of the external capsule over the area indicated, with complete degeneration of the overlying cortex. Experiment 13 (page 12). Retention nearly perfect. Errors: learning 3, retention 0. Lesion (plate II, fig. 13). Right hemisphere. First opera- tion. There is a complete transverse section of the anterior pole through the end of the forceps of the corpus callosum to the level of the olfactory tracts. Second operation. There is a lesion extending from the first backward along the external capsule to the posterior border of the hippocampus, narrowing forward below at the level of the lateral ventricle. All the cor- tex overlying the lesion is completely degenerated. Left hemi- sphere. First operation. Transverse section of the anterior pole extending downward to just above the level of the olfac- tory tracts. Second operation. There is a longitudinal incision extending backward in the external capsule from the first lesion to the posterior edge of the hippocampus, covering the entire orbital surface down to the olfactory cortical funiculus. All the overlying cortex is degenerated. Experiment 14 (page 13). No retention. The animal de- veloped hemiparesis and was stupor ous. Lesion. An invasion of the thalamus by an extensive clot was determined by gross dissection. Of the fourteen animals described in this series thirteen gave 84 K. S. LASHLEY AND S. I. FRANZ some evidence of the retention of the habit. An exact estima- tion of the individual degree of retention is precluded by the Simplicity of the habit and the different extents to which the animals showed shock effects of the operation. Since all the animals with lesions restricted to the frontal regions retained the habit wholly or in part and since the lesions were for the most part incomplete the first question that arises is whether or not any particular area in the frontal region was left intact in all the animals. The combined extents of the lesions in this series are shown in figure 2. The entire frontal pole of each hemisphere extending down to the olfactory tracts was destroyed in one or another of the animals. No one part of the frontal pole remained undestroyed in all. It seems, then, that no particular part of the frontal pole of the rat's cortex is necessary for the FIG. 2. TOTAL EXTENT OF INJURIES TO CORTEX IN ANIMALS DESCRIBED IN SERIES I All parts of the cortex anterior to the knee of the corpus callosum have been destroyed. retention of the maze-habit; there is no specialized region con- cerned with the maze-habit which has a uniform position for all animals. Furthermore, in some of the animals (experiments 1, 3, 4, 10 and 11), there was an almost complete destruction of all the cortex above and in front of the knee of the corpus callo- sum so that it seems very probable that no part of the frontal region of the rat's brain is concerned with the retention of the maze-habit. A number of attempts have been made to find some correla- tion between the extent of lesion in these cases and the accuracy of performance in the retention tests but no such correlation seems to exist. Both animals with extensive and those with slight lesions made perfect records in the retention tests CEREBRAL FUNCTION IN THE RAT 85 and others with almost identical lesions made several errors. This lack of relation between the extent of frontal lesion and the degree of retention provides further evidence that the func- tioning of the maze-habit is independent of the frontal region of the cortex. II. THE RETENTION OF THE MAZE-HABIT AFTER FRONTAL AND TEMPORAL OR FRONTAL AND OCCIPITAL LESIONS Experiment 15. After retention had been tested in the animals described in experiments 8, 9, 10, 11, 12 and 13 they were allowed to rest for two weeks and then were retrained for twenty-five trials, showing practically no loss of the habit in this time. They were next subjected to a second operation. In three, an incision was made from the locus of the first operation backward FIG. 3. TOTAL EXTENT OF THE LESIONS IN ANIMALS OF SERIES II, WHICH RETAINED THE MAZE-HABIT AFTER THE SECOND OPERATION on both sides for a distance of 5 mm. In the other three the scalpel was passed almost horizontally backward through the cortex of the dorsal convexity from the region of the first lesion to the tentorium. Three of these animals survived, one after, occipital lesion (11) and two after temporal lesion (10 and 12). The lesions have been defined in the descriptions of these ani- mals and figured in plate II, figures 10, 11, and 12. The com- bined extent of the lesions is shown in figure 3. All three animals showed perfect retention of the habit after the second operation. They were each given fifteen trials in the maze on the day following the second operation, and ten trials on the second day. Number 10 averaged 2.5 seconds per trial in these twenty-five trials and made only one error. Number 12 averaged 2.6 seconds and made one error in the same number 86 K. S. L ASHLEY AND S. I. FRANZ of trials. Number 11 was slower and required an average of 9.0 seconds per trial but did not make a single error or react in any way to the entrance to the cul de sac in twenty-five trials. After the first operation all these animals had shown a per- fect retention of the habit, in spite of a practically complete destruction of the frontal poles of both hemispheres. The sec- ond operation probably did little additional harm to number 10 beyond cutting the fimbria on both sides. In number 12, how- ever, large areas on the dorsal and orbital surfaces of both hemi- spheres were destroyed. In number 11 practically all the dorsal and lateral surfaces of the cortex overlying the gyri hippocampi and between them and the frontal poles was destroyed. The effects of the first operation, taken in conjunction with the other experiments of the same series, indicate rather clearly that the persistence of the habit was not conditioned by the functional integrity of the frontal pole. The second operation resulted in the elimination of not less than half of the cortex of the dorsal convexity and the combined destruction resulting from all the operations performed on these animals, certainly includes not less than two thirds of the cortex, leaving only the occipital regions, the ventral surface, and the gyri hippocampi intact. We are justified in concluding therefore that no part of the cortex in front of the caudal end of the corpus callosum and above the level of the floor of the lateral ventricles is con- cerned with the retention of simple kinesthetic-motor habits. Experiments with the inclined-plane box The simple maze offered some disadvantages for a study of retention owing to the fact that it did not require a reaction that was sufficiently well defined to be certainly recognizable in the retention tests. It seemed best therefore to use some more com- plex habit in the later experiments for the sake of getting a more clearly defined series of activities and also with the possibility that the more complex habit, involving different types of reac- tion, might reveal a selective effect of the cerebral lesion upon certain types of activity. The inclined-plane box (2) was CEREBRAL FUNCTION IN THE RAT 87 finally selected as combining a fairly specific reaction to a definite series of stimulating objects with relatively complex kinesthet- ic-motor habits. In order to complicate the kinesthetic habits somewhat and allow of greater individual variation than is possible in the usual type of this problem-box, the design shown in figure 4 was adopted. The problem-box consists of a wire covered rectangular frame- work having a door (D), held open by a pair of flat brass springs, compressed between the door FIG. 4. THE INCLINED PLANE Box A, door of restraining cage; B, inclined plane; C, catch extending down to upper edge of door, D, door leading into food-box. and its frame. On top of the box, above the door is a board (B), supported in the middle by a metal fulcrum and weighted so that the end C is slightly heavier than the other. From the board an adjustable metal rod (C) extends through the top of the box to the door and serves as a latch to hold it closed. A pressure of from 50 to 75 grams at the free end of the plane (B) is required to tilt it and release the door. This, the food- box, is enclosed in a larger restraining cage having a door at A and is pushed against the side of the restraining cage opposite 88 K. S. LASHLEY AND S. I. FRANZ the door so that only three sides are exposed. The animal, introduced into the restraining cage through the door (A) must climb to the top of the food-box, push down the outer end of the plane (B) for a distance of half an inch until the door springs open, then climb down and enter the door to get food. A momentary pressure is not sufficient to depress the plane and the animal must maintain the full pressure while the plane is moving down. The fact that he must climb upon the food-box makes possible a variety of ways of approaching the plane and the development of individual peculiarities of reaction. The methods of training were those generally employed in experiments with similar apparatus. The animals, unless much weakened by loss of blood, were fed only in the food-box after the conclusion of the day's tests. Five trials were given daily. Records were made for each trial of the time required by the animals to reach and push down the plane, and separate records of the time taken in going from the plane to the food after the plane had been tripped. Extensive notes were made of individ- ual peculiarities in the path followed to the plane, of the method of pushing it down, and of the path followed from the plane to the food. III. THE RATE OF LEARNING AFTER INJURY TO THE FRONTAL POLE Experiment 16 As a preliminary test of the ability of the animals to form complex habits after the destruction of the frontal pole of the cortex a group of six animals which had been used to experiments on the maze (experiments 1, 2, 3, 4, 6 and 7) and had fully recovered from the operation was trained on the inclined plane, described and figured above. Their behavior when placed in the restaining cage was in all respects normal and they learned the problem at a practically normal rate. Their rate of improvement in the average time required for tripping the latch is compared in figure 5 with that of an equal number of normal animals trained under identical conditions. The actual time consumed in learning was somewhat less than that required by normal animals; their methods of learning were the CEREBRAL FUNCTION IN THE RAT 89 same and their final efficiency was as great as that of any nor- mals. Their quicker time on the first trials is probably due to the fact that they had been handled a great deal and were more accustomed to the experimenter than were the normal animals. 2000 1500 1000 500 FIG. 5. A COMPARISON. OP THE RATE OP LEARNING THE INCLINED PLANE Box BY NORMAL ANIMALS AND BY ANIMALS AFTER INJURY TO THE FRONTAL AREAS Six animals are included in each group. The average time in seconds per trial is plotted. Normal rats : . Operated rats : . The extent of the lesions in these animals is shown in plate II, figures 1, 3, 4, 6 and 7. The combined extent of the lesion is shown in figure 6. Practically every part of the anterior pole of the cortex in front of the knee of the corpus callosum FIG. 6. TOTAL EXTENT OF LESIONS IN ANIMALS WHICH LEARNED THE INCLINED PLANE Box AFTER OPERATION 90 K. S. LASHLEY AND S. I. FRANZ was destroyed in one or another of these animals without inter- fering in any way with the formation of the relatively complex habit. In numbers 3, 6 and 7 the lesions were small but in numbers 1 and 4 there was a loss of practically the entire frontal pole of both hemispheres without a correlated reduction of the learning ability. It seems quite certain, therefore, that no particular part of the anterior third of the cortex of the white rat is necessary for the formation of a complex kinesthetic-motor .habit and, from the condition in numbers 1 and 4, that the entire frontal region of the cortex may be dispensed with without any reduction of the ability to learn. IV. RETENTION OF THE INCLINED-PLANE HABIT AFTER CEREBRAL LESIONS Earlier experiments with the higher mammals indicate that learning may take place in the absence of parts of the cortex which are normally functional in habit formation. Thus, Franz (3) has found that the destruction of the frontal lobes in the cat and monkey is followed by the loss of recently formed habits but that the habits can be immediately reacquired and that the rate of learning is practically the same as before the loss of these areas. The ability of the animals described in series III to learn the inclined plane problem does not, there- fore, preclude the possibility that the frontal region of the cor- tex is normally functional when present. To test this a series of normal animals' were trained on the inclined plane box, then operated upon to destory the frontal areas, and tested for re- tention of the habit. Seventeen animals survived the opera- tion and were tested thoroughly for retention or relearning of the habit. In three animals the lesions were restricted to the dorsal and temporal regions, first to test the relation of these parts to the functioning of the habit, second, as a control of the effects of operative shock. No final criterion of learning was adopted, but the animals CEREBRAL FUNCTION IN THE RAT 91 were not operated upon until they had several times tripped the plane and gone to the food without excess movements. This required from 25 to 45 trials. A part were overtrained for a time about equal to that required for learning, receiving a total of from 75 to 115 trials. Operation without injury to the brain Experiment 17. Two trephine holes were made in the skull of an old female just back of the fronto-parietal suture and these were enlarged with bone forceps to a diameter of about 5 mm. on either side. The scalp was then sewn up without further injury to the brain than was necessary in making the openings. In training the animal had been given 100 trials on the inclined box. Her average time per trial was for the first five trials; to plane, 1025 seconds; to door, 26 seconds. For the last five trials the averages were: to plane, 9.4 seconds; to door, 2.0 seconds. Retention was tested on the second day after the operation. The average time per trial was: to plane, 46 seconds; to door, 9.2 seconds. Her reactions were unhesitating. The path fol- lowed from the door to the plane was direct and the plane was the only object in the restraining cage which excited more than a momentary reaction. In every trial the animal sprung the plane by standing beside it and pushing down upon its outer end with her fore feet. The slight increase in time over the last trials preceding the operation is due to a slower rate of movement and not to any increase of exploratory movements. Autopsy showed small adhesions of the convexity of the hemi- spheres at the region of the trephine openings. The extent of the lesions is shown in plate III, figure 17. After opening of the skull and exposure of an area of the cortex as great as that involved in the greater number of the experiments here described, this animal showed perfect retention. 92 K. S. LASHLEY AND S. I. FRANZ Temporal lobes destroyed Experiment 18. A large opening, 2 by 5 mm. was made on each side of the skull just back of the front o-parietal suture of a medium sized female rat, 140 days old. Through these the scalpel was passed to destroy the temporal lobes. The animal had been trained on the inclined-plane box for 115 trials. The average time per trial for the first five trials was: to plane, 2719 seconds; to door, 61 seconds. That for the last five trials was: to plane, 34 seconds; to door, 2.2 seconds. Retention was tested on the second day after the operation, when the animal was active and in good condition. The aver- age time for the nine trials given on this day was: to plane, 22 seconds ; to door, 1 1 seconds. Her specific reactions were wholly confined to the plane and door. On different trials she ap- proached the plane from different directions, but always tripped it in the same way, by pushing down on its outer ^end with her fore feet, and in every trial except the first she went directly from the plane to the door of the food-box. Extent of lesions (plate III, fig. 18). Right hemisphere. There is a large lesion on the dorsal convexity including almost all the cortex dorsal and lateral to the gyrus hippocampus, ex- tending from near the longitudinal fissure, over the convexity and down over the temporal lobe to the level of the posterior horn of the lateral ventricle, following the course of the external capsule. Left hemisphere. The lesion on the dorsal convexity is simi- lar to that on the right. It extends downward around the antero-lateral face of the gyrus hippocampus and thence for- ward through the lateral face of the corpus striatum. After a bilateral lesion destroying most of the cortex lying dorsad and laterad to the gyri hippocampi the motor habit previously established was retained. Experiment 19. In the skull of a small female, 148 days old, two trephine holes were made about three millimeters back of the fronto-parietal suture. The scalpel was passed back from these to destroy the temporal lobes. The animal had been CEREBRAL FUNCTION IN THE RAT 93 trained on the inclined-plane box for 75 trials. The average time per trial for the first five trials of learning was: to plane, 916 seconds; to door, 85 seconds. The average for the last five trials was: to plane, 2.8 seconds, to door, 2.2 seconds. Retention was tested on the second day after the operation. At this time the animal was active and showed no abnormal symptoms. She was given ten trials on this day with the fol- lowing average time per trial: to plane, 51 seconds; to door, 14 seconds. The animal's behavior toward the problem box was in all respects normal. Her exploratory movements were restricted to the plane and in seven of the trials she tripped the catch by pushing the plane down with her fore feet, a stereotyped method which she used before the operation. FIG. 7. TOTAL EXTENT OF LESIONS IN THE CONTROL ANIMALS (EXPERIMENTS 17, 18 and 19) Lesion (plate III, fig. 19).' Right hemisphere. There is an extensive lesion of the cortex extending from the anterior border of the hippocampus to the knee of the corpus callosum and narrowing rapidly as it extends laterad on the orbital surface. It is continued over the orbital surface as a lesion of the external capsule around the upper half of the corpus striatum with degen- eration of the overlying cortex. Left hemisphere. There is a lesion similar to that on the right over the dorsal convexity but it is broader over the tem- poral region and extends downward only to the upper level of the lateral ventricle. This animal, after the destruction of the cortex over the temporal and a portion of the orbital surface of both hemispheres, showed no deterioration of the habit previously formed. The perfect retention of the habit by these three animals 94 K. S. LASHLEY AND S. I. FRANZ shows that any loss of the habit shown in other animals is prob- ably not the result of operative shock but must be ascribed to the actual brain injury acting either locally in the operative field or extending by hemorrhage or intracranial pressure to other regions. The extensive destruction of the temporal sur- faces of the cortex in experiments 18 and 19 shows that these areas play no important part in the retention of the habit. Fig- ure 7 shows the combined extent of the lesions in these animals. Frontal region destroyed: animals showing retention after operation Experiment 20. Two trephine holes were made through the calvarium of a large female rat, 140 days old, and a transverse incision was made through the frontal pole of each hemisphere. The animal had been trained for 35 trials on the inclined plane box. The average time required for the first five trials was: to plane, 123 seconds; to door, 42 seconds. The average time required for the last five trials was: to plane, 13 seconds; to door, 3 seconds. Retention was tested twenty-four hours after the operation but the animal was stuporous and did not react to the problem box. On the following day she opened the food-box three times. Her movements were slow and much time was spent in scratch- ing at the dressing on her scalp. There were, however, few random movements and except for the diversions incident to this scratching, she kept closely to the direct path from the door to the plane and from the plane to the food. Her method of tripping the catch before the operation had been to walk to the back of the plane and, standing either on or beside it with her hind feet, to reach out to its outer end with her fore feet and push down, gradually throwing her weight forward on her fore feet. This same method was used throughout the trials following the operation, but the movements involved were carried out with greater inaccuracy than before. In some trials she went from the plane to the door without making any attempt to push the former; in others she pushed, but did not wait for the catch to spring. CEREBRAL FUNCTION IN THE RAT 95 During the first days of the retention tests she frequently leaped into the air, whirled through 180 degrees, and snapped her jaws repeatedly, or at other times spun about in a similar manner and bit her own tail. Another movement which oc- curred frequently was a sort of scampering, a series of short leaps which did not carry her forward more than a few inches. These movements, as well as the inexactness of the reactions to the plane served to prolong the time and make it less comparable with the time of learning. Training was continued for 45 trials after the operation. The average time for the first five trials of the retention tests was: to plane, 63 seconds; to door, 23 seconds. The rate of improvement was more uniform and at first more rapid than that observed during learning, but the time required for springing the catch was not reduced so low as it had been before the operation. The rate of learning and of improvement in the retention tests are shown in figure 8, where the total time required for successive groups of five trials is plotted. Lesions (plate III, fig. 20). Right hemisphere. There is a transverse lesion on the anterior covexity, passing forward along the fibers of the callosum and through the end of the forceps of the callosum to the base of the olfactory bulb, completely sep- arating the frontal pole. Left hemisphere. The lesion is quite similar to that on the right but does not extend quite to the olfactory bulb and pre- serves the greater part of the mesial face of the frontal pole. After nearly complete destruction of the frontal poles of both hemispheres this rat retained her previous method of tripping the catch. There ivas, however, some inaccuracy of movement indicat- ing a partial loss of the habit. Experiment 21. The frontal pole of the cortex was destroyed in a small female, 130 days old. She had been trained for 40 trials on the inclined-plane box. The average time required for the first five trials was: to plane, 618 seconds; to door, 74 sec- onds. The average time required for the last five trials was: to plane, 5.2 seconds; to door, 1.8 seconds. Retention was first tested on the day following the operation. 96 K. S. LASHLEY AND S. I. FRANZ The animal was stuporous and did not react to the problem situation. On the following day she was active and gave evi- dence of hunger. She wag given five trials on this day and five trials per day thereafter for 8 days. The average time per trial for the first five -trials of the retention tests was: to plane, 81 seconds; to door, 42 seconds. In the tests on the first day she used no uniform method of springing the plane; she once pulled 28 trial* S0 FIG. 8. RATE OF IMPROVEMENT IN TRAINING AND RETENTION TESTS. AVERAGE TIME PER SUCCESSIVE FIVE TRIALS is PLOTTED FOR EXPERIMENT 20 THE Learning Retention it up from behind, once jumped up on it from in front of the food box, once walked straight across it from the back of the box, and twice walked out on it from the rear. After the fourth trial she adopted this latter method and used it uniformly in suc- ceeding trials, varying the method only rarely by pushing down the plane with her fore feet. In the learning trials she had uni- formly climbed up from the rear of the box and tripped the catch by pushing the plane down with her fore feet. CEREBRAL FUNCTION IN THE RAT 97 The curves of the learning and retention tests are given in figure 9. It is evident that very much less time was required for the solution of the problem at the beginning of the retention tests than at the beginning of learning, but that improvement was relatively slower. Apparently the plane retained a certain stimulatory value, much like that which it acquires in the early stages of learning, but did not at first call out the appropriate reactions for getting food. FIG. 9. RATE OF IMPROVEMENT IN TRAINING AND RETENTION TESTS FOR EXPERIMENT 21 Arranged as figure 8 Lesions (plate III, fig. 21). Right hemisphere. There is a transverse lesion extending from just behind the knee of the corpus callosum laterad to the cortex and diagonally ventrad through the base of the olfactory bulb. The greater part of the mesial face of the frontal pole is intact but there is a probable destruction of all the antero-lateral area. Left hemisphere. There is a transverse lesion extending from the anterior horn of the lateral ventricle along the forceps of PSYCHOBIOLOGY, VOL. I, NO. 2 98 K. S. LASHLEY AND S. I. FRANZ the corpus callosum to the external capsule and downward to the floor of the ventricle. It probably involves very little de- struction. There is also a lesion extending into the mesial cortex just back of the knee of the corpus callosum, destroying a small area. There is a partly absorbed clot in the third ventricle. In this animal three-fourths of the cortex on one frontal pole and about one-eighth on the other were destroyed. The habit was partly retained. The reactions after the operation resembled those ap- pearing in the early stages of learning before a stereotyped mode of solving the problem has been acquired. Experiment 22. Transverse incisions were made in the frontal areas of the cortex of a large male, 140 days old, through two trephine holes in the temporal regions. The animal had been trained for 40 trials on the inclined plane box. The average time per trial required for the first five trials of training was: to plane, 330 seconds; to door, 38 seconds. The average time per trial for the last five trials was: to plane, 6.6 seconds; to door, 3.2 seconds. Retention was tested on the day following the operation but the animal was stuporous and did not attempt to climb upon the food box. On the third day he climbed upon the box but got down immediately and remained quiet in a corner for 20 minutes. On the fourth day after the operation he climbed on the box and fell off, striking the back of his head and lying stunned for 30 seconds. When he recovered he climbed up on the box, walked out on the plane from the rear and then went directly to the food. The remaining trials were made quickly. The average time for the first five trials, exclusive of the one in which the fall occurred was: to plane, 45 seconds; to door, 5.6 seconds. The tests were continued for 30 trials, which are compared with the first practice series in figure 10. In the first tests he made fre- quent trips from the plane to the door and made few pauses at other parts of the restraining cage. In later practice he acquired the habit of jumping up from the door to the back of the plane, then walking out to the free end of the plane. This was not the most frequent method that he had employed during learning. CEREBRAL FUNCTION IN THE RAT 99 Like the animal in experiment 21 he seemed to retain a gen- eral response to the region of the plane but to have lost the specific reactions used in springing it. Lesion (plate III, fig. 22). Right hemisphere. There is a vertical transverse lesion extending downward from the anterior horn of the lateral ventricle to the cerebral peduncle, without sec, . . , trials FIG. 10. RATE OF IMPROVEMENT IN TRAINING AND RETENTION TESTS FOR EXPERIMENT 22 Arranged as figure 8 involving this, however. On the antero-dorsal convexity there is a lesion corresponding to the trephine opening, with the cortex cephalad to it somewhat degenerated. The deeper lesion in- volves little destruction, as the cut follows the course of the fibers. Left hemisphere. The lesion on the dorsal convexity is somewhat larger than that on the right. There is ,a large cavity 100 K. S. LASHLEY AND S. I. FRANZ in the region of the external capsule opposite the corpus striatum with degeneration of the fibers and cortex laterad to it, and invasion of the cerebral peduncle in the region receiving fibers from the mesial surface of the frontal pole. Bilateral destruction of the cortex above the knee of the corpus callosum with degeneration on the orbital and mesial surfaces of the left hemisphere was followed by a partial loss of the habit in- cluding the definite mode of depressing the plane. Experiment 23. The frontal areas of the cortex were tran- sected through two trephine holes in a small male rat, 149 days old. He had been trained previously on the inclined-plane box for 100 trials. The average time for the first five trials of practice was: to plane, 332 seconds; to door, 48 seconds. The average time for the last five trials was: to plane, 10.2 seconds; to door, 2.2 seconds. During the training he developed an absolutely stereotyped method of tripping the plane. He ran first to the door, jumped up from there upon the food box in such a way as to alight upon the free end of the plane, turned and jumped down to the floor in front of the door and went in to the food. Retention was first tested on the second day after the opera- tion. The animal was fairly active, sniffed at the door of the food ,box and stood up and sniffed the plane repeatedly but made no effort to get upon the box in thirty minutes. On the fifth day after the operation he first showed marked activity comparable to that of a normal animal. For the first four trials he climbed to the top of the box always from the door, but gave no specific reaction to the plane, in each case pushing it down apparently by chance and using a different method each time. On the fifth trial he went first to the door, jumped up upon the lower end of the plane, then turned quickly and jumped down to the door. This was repeated five times before he alighted on the outer end of the plane. During the next ten trials he invariably jumped up from the front of the box to the plane, and then down to the door again, repeating this per- formance as often as fifteen times in a single trial before he alighted on the free end of the plane. This was due in part to CEREBRAL FUNCTION IN THE RAT 101 the fact that he was too weak to jump the full distance readily and usually jumped from near the side of the restraining cage, catching its wires if he fell short. In the repeated failures to open the door by jumping up at this place a new method of working the plane was gradually evolved; he came to scramble up over the lower end of the plane, turn and push down on its free end until the door opened. The average time per trial for the first five trials of the re- tention tests was: to plane, 176 seconds; to door, 13 seconds. The relations of the practice curves of learning and retention are practically the same as for those shown in experiments 20, 21, and 22. Lesion (plate III, fig. 23). Right hemisphere. There is a transverse incision extending cephalad from above the hippo- campus through the lateral ventricle and caudal end of the corpus striatum to the peduncle. The internal capsule and the fibers of the peduncle are uninjured, so that the lesion probably involves little more than the superficial areas through which the knife passed. This includes the greater part of the cortex above the corpus callosum and a narrower band extending down over the orbital surface. Left hemisphere. There is a transverse lesion extending from above the anterior border of the hippocampus to the base of the olfactory lobe, transecting the corpus striatum and completely severing the portion of the cortex lying cephalad to it. Destruction of all of the left frontal region and the temporal region of the right was followed by the retention of the stereotyped mode of reacting to the problem box but with inaccuracies of ad- justment which resulted in the acquirement of a new mode of response. Experiment 24. The frontal poles of both hemispheres of a medium sized female rat, 147 days old, were incised through large trephine holes, well back of the fronto-parietal suture. She had been trained previously for 75 trials on the inclined- plane box. The average time per trial required for the first five trials of training was: to plane, 2422 seconds; to door, 23 102 K. S. LASHLEY AND S. I. FRANZ seconds. The average time for the last five trials was: to plane, 15.4 seconds; to door, 2.4 seconds. Retention was tested on the second day after the operation. She reacted promptly to the problem-box situation, never pausing except at the door and at the plane. She tripped the plane by one of two methods; either she jumped up from in front of the door so that her hind feet struck the free end of the plane, or, this method failing, she turned and pushed the plane down with her fore feet. These were the methods used before the operation. The average time required for the first five trials of the retention tests was: to plane, 37.6 seconds; to door, 9.6 seconds. Lesion (plate III, fig. 24). Right hemisphere. There is a small lesion on the antero-dorsal convexity corresponding to the extent of the opening of the skull. From this a transverse cut extends ventrad through the union of the forceps of the corpiis callosum with the external capsule to the base of the olfactory bulb. Its effects are limited to the lateral face of the cortex where there is a narrow degenerated area along the edges of the cut, widening on the orbital surface. Left hemisphere. The lesion is almost identical with that on the right but less extensive on the orbital surface. Destruction of the antero-lateral faces of the cortex in this animal resulted in practically no loss of the habit. Frontal region destroyed: animals showing questionable retention after operation Experiment 25. The frontal poles of the cerebrum of a small male rat, 143 days old, were destroyed by incisions through two large openings in the parietal bones, 5 mm. back of the fronto- parietal suture. The animal had previously been given 45 trials in the inclined-plane box. The average time required for the first five trials was: to plane, 575 seconds; to door, 40 sec- onds. The average time for the last five trials of training was: to plane, 9.6 seconds; to door, 2.8 seconds. Retention was first tested on the fourth day after the opera- tion as the animal was quite stuporous up to this time and had to be fed by hand. On this and the following day he moved CEREBRAL FUNCTION IN THE RAT 103 slowly about the floor of the restraining cage but made no at- tempt to climb up to the plane in more than an hour. On the following day he was more active but gave marked fear reac- tions, squeaking and jumping blindly against the sides of the cage when touched and biting at the experimenter's hands or at his own feet when he was picked up. This behavior persisted for four days, then he became more normal and on the eighth day after the operation he solved the problem box five times. On the following day he was spastic and on the tenth day after the operation died. His average time per trial on the five suc- cessful trials was: to plane, 138 seconds; to door, 50 seconds. At no time in the retention tests did the plane seem to have any more intense stimulating value than other parts of the problem situation. The only evidence for retention is the short time required for solving the problem, and this may be merely a chance variation. Lesion (plate III, fig. 25). Right hemisphere. There is a lesion over the dorsal surface of the gyrus hippocampus, extend- ing cephalad along the roof of the lateral ventricle, then ventrad through the knee of the corpus callosum, transsecting the corpus striatum and extending to the base of the hemisphere. A large cyst has been formed filling the greater part of the frontal pole and the anterior half of the corpus striatum is degenerated. Left hemisphere. There is a large pit on the dorsal con- vexity, extending down around the gyrus hippocampus and a transverse incision extending from this through the anterior horn of the lateral ventricle to the base of the olfactory bulb, completely severing the anterior pole. After practically complete destruction of the anterior third of both hemispheres this animal succeeded in solving the inclined- plane problem in considerably less time than that usually required by normal individuals and in kss than one fourth of the time which he took for an equal number of trials in the initial training. Experiment 26. The frontal poles of both hemispheres of a small male rat, 143 days old, were incised through large open- ings in the parietal bones. He had been trained on the inclined- plane box for 40 trials. The average time required per trial 104 K. S. L ASHLEY AND S. I. FRANZ for the first five trials was: to plane, 721 seconds; to door, 35 seconds. The average time per trial for the last five trials was: to plane, 5.4 seconds; to door, 2.4 seconds. Retention was tested first four days after the operation. The animal was still stuporous and on this, and the following day made no attempt to get upon the food box. On the seventh day after the operation he ran actively back and forth from the door to the back of the food box and tripped the plane twice by climbing up over it from the door. On the following day he was stuporous but on the ninth day he solved the problem five times quite rapidly. Thereafter he became progressively worse and died on the fourteenth day after the operation. The average time per trial for the first five trials of the re- tention tests was: to plane, 179 seconds; to door, 26 seconds. This is less than one fourth of the time required for the first five trials of training and indicates some degree of retention although no specific habits of reaction to the plane which had been noted before the operation were observed to persist. Lesion (plate III, fig. 26). Right hemisphere. There is a transverse lesion extending diagonally forward and downward from just above the hippocampus through the lateral ventricle to the base of the olfactory bulb. The corpus striatum is cut in two transversely. The entire frontal pole is probably non- functional. Left hemisphere. The lesion is very similar in form to that on the right but penetrates less deeply toward the mesial sur- face and probably leaves the mesial cortex functional. After extensive destruction of the antero-lateral regions of both hemispheres this animal gave no indication of the retention of specific habits of reaction to the plane. However, even though he was stuporous for much of the time, he required only one fourth as much time for the retention tests as for corresponding tests in ini- tial learning. CEREBRAL FUNCTION IN THE RAT 105 Frontal pole destroyed: animals in which the habit was lost but reacquired after operation Experiment 27. The frontal poles of the cortex were incised in a small female rat, 145 days old, through two trephine holes at the frontal-parietal suture. The animal had been given 30 trials on the inclined-plane box. The average time required for the first five trials was: to plane, 396 seconds; to door, 23 sec- onds. The average time per trial for the last five trials was: to plane, 24 seconds; to door, 4.0 seconds. The animal was very weak from the operation and could not be tested for eight days. On the ninth and tenth days she was very spastic and remained motionless in the restraining cage. On the eleventh day she first tripped the catch and on the following days im- proved considerably in the directness of her approach to the plane but never reached her previous efficiency. Throughout the period of retraining she remained spastic and lost weight rapidly. During the first trials of retraining there was ne,ver any indication of the retention of any specific mode of reacting to the problem box. The average time per trial for the first five trials of the retention tests was: to plane, 599 seconds; to door, 65 seconds. Lesion (plate III, fig. 27). Right hemisphere. There is a transverse section passing over the anterior face of the gyrus hippocampus, through the lateral ventricle to the base of the hemisphere, separating all parts of the cortex in front of the an- terior horn of the lateral ventricle. The corpus striatum is completely destroyed and the lesion extends caudad from it along the external capsule to the hippocampus, with degenera- tion of all the cortex laterad to it. Left hemisphere. The lesion is less extensive, passing ven- trad just behind the knee of the corpus callosum to the base of the olfactory bulb and out diagonally through the anterior end of the corpus striatum to the cortex, severing the anterior pole. The destruction of the anterior poles of both hemispheres, of part of the orbital surface of one hemisphere, and of all of one and part of the other corpus striatum was followed by a persistent spas- 106 K. S. LASHLEY AND S. I. FRANZ tidty and by a complete loss of the motor habit. The habit was reacquired with, normal rapidity at first but a normal degree of proficiency was not attained. Experiment 28. The frontal poles of the cortex were destroyed in a large male, 140 days old. He had been trained on the in- clined-plane box for 30 trials. The average time per trial was for the first five trials: to plane, 2594 seconds; to door 224. That for the last five trials was: to plane, 22.0 seconds; to door, 1.6 seconds. Retention was tested on the day following the operation but no approach to normal activity was obtained until the third day. He then tripped the plane three times, requiring an aver- age time of: to plane, 493 seconds; to door, 79 seconds. For the following week he did not get upon the food box in a total of six hours spent in the restraining cage. He then became active again and eventually learned the problem, requiring about the same time as in the initial practice to reduce his aver- age'time to less than 30 seconds. In the early trials of the re- tention tests there was no uniformity in the method of tripping the plane and there was never any indication of the retention of a specific habit of reaction. Lesion (plate III, fig. 28). Right hemisphere. There is a lesion of the dorsal convexity over the gyrus hippocampus and extending forward around the anterior surface of the hippocam- pus, along the roof of the lateral ventricle and ventrad in front of the knee of the corpus callosum through the peduncle, sever- ing the frontal pole. The entire lobe in front of the corpus striatum is degenerated and filled by a large cyst. Left hemisphere. The lesion is similar to that on the right, but extends farther back along the external capsule, with de- generation of a part of the orbital cortex. Only the outer half of the peduncle is injured so that the mesial surface of the lobe is probably functional. After almost total destruction of both anterior poles of the cortex in front of the corpora striata this animal completely lost the habit of the inclined-plane box, but acquired it again in about the same time as was required for initial learning. CEREBRAL FUNCTION IN THE RAT 107 Frontal pole destroyed: animals which lost the habit after opera- tion and failed to acquire it again Experiment 29. The frontal poles of both hemispheres were incised in a large male rat', 155 days old. He had been trained for 85 trials on the inclined-plane box. The average time per trial for the first five trials was: to plane, 2311 seconds; to door, 51 seconds. That for the last five trials was: to plane, 8.9 seconds; to door, 2.8 seconds. The animal was tested daily for a week after the operation but did not once get on top of the food box. His activity was equal to that of a normal animal but was confined to a few ster- eotyped movements. He spent a great deal of time in climbing up the sides of the restraining cage and pushing against the wire top with his nose. This was gradually replaced by a rapid pacing around the food box from the door to the back and to the door again, interspersed with long pauses in front of the door of the food box, which continued day after day without varia- tion. There was no indication of the retention of any part of the habit. Lesion (plate III, fig. 29). Right hemisphere. There is a lesion through the lateral ventricle, passing laterad through the posterior third of the corpus striatum to the cortex and ventrad through the peduncle. There is a second lesion over the frontal pole with complete degeneration of the cortex and formation of a cyst. It is probable that the section of the cerebral peduncle destroyed the lower connections of all the frontal region. Left hemisphere. There is an extensive superficial lesion over the convexity of the frontal pole with a transverse cut ex- tending laterad from the lateral ventricle to the cortex in front of the corpus striatum, following the course of the fibers and probably doing very little damage. After a complete section of one cerebral peduncle, destruction of the frontal pole on that side, and destruction of half the cortex of the frontal region on the other side this animal gave no evidence of retention of the habit, and failed to reacquire it within the limits of the experiment. 108 K. S. LASHLEY AND S. I. FRANZ Experiment 30. The frontal poles of the cortex of a medium sized male rat, 140 days old, were transsected with a spear- pointed needle, introduced through small trephine holes in the parietal bones just behind the fronto-parietal suture. The animal had been trained for 25 trials on the inclined-plane box. The average time per trial for the first five trials was: to plane, 3382 seconds; to door, 39 seconds. That for the last five trials was: to plane, 43.0 seconds; to door, 10.4 seconds. During training the animal developed an easily recognizable method of springing the plane. He regularly placed his right fore foot on the end of the plane and kept it there while he thrust his nose under the plane. In this position his weight was supported largely by the right fore foot and the end of the plane was pulled down. Retention was tested first on the second day after the opera- tion while the animal was still weak and stupor ous. He moved about slowly, smelling in the corners of the restraining cage and once crossed the top of the food box but gave no specific reaction to the plane. He was tested daily for thirty minutes there- after for fourteen days. For the first three days he was very active and quite wild, squeaking and jumping whenever touched; later he became tame but rarely moved away from the door of the restraining cage. Lesion (plate III, fig. 30). Right hemisphere. There is a lesion of the cortex above the dorsal and lateral surfaces of the gyrus hippocampus with prolongation of the lateral ventricle along the external capsule into the occipital lobe with probable degeneration of all the fibers in this region. There is a complete separation of the hippocampus from the external capsule with probable degeneration of the fibers in the capsule. There is a transverse lesion also extending forward along the corpus cal- losum, around the knee to the floor of the lateral ventricle, and laterad through the anterior end of the corpus striatum to the external capsule. This lesion does not penetrate far enough to injure the peduncle so that the ventro-lateral face of the frontal pole is probably intact. Left hemisphere. There is a transverse lesion extending CEREBRAL FUNCTION IN THE RAT 109 from above the gyms hippocampus, through the anterior end of the fornix and the corpus striatum to the base of the olfactory bulb, completely severing the anterior pole. Both corpora striata were injured but show no sign of degen- eration except along the edges of the cut. With almost complete destruction of the frontal poles of both hemispheres and a probable degeneration of much of the cortex on the orbital and occipital regions of the right hemisphere this animal showed a complete loss of the habit and a failure to relearn the problem. Experiment 31. The frontal poles of both hemispheres were incised in a small female rat, 148 days old. She had been given 100 trials on the inclined-plane box, requiring as an average time per trial for the first five trials: to plane, 591 seconds; to door, 67 seconds. For the last five trials these averages were : to plane, 5.4 seconds; to door, 3.2 seconds. Retention was tested on the second day after the operation. The animal was very weak, but ran about actively in the re- straining cage and twice crossed the plane without giving any specific reaction to it. On the following day she spent thirty minutes climbing up to the top of the restraining cage and fall- ing back to the floor. She ran across the top of the food box several times and in each case pitched off on her head without making any attempts to catch the side of the box. For the next two days her activity in the restraining cage was greater than that of a normal rat, but was restricted wholly to attempts to climb around the sides of the restraining cage just below the top, with repeated heavy falls. On the sixth day after the op- eration she died. Lesion. The brain of this animal was lost. Experiment 32. The frontal pole of the cerebrum was tran- sected through two large trephine holes in the parietal region of a small female rat, 146 days old. She had been trained on the inclined-plane box for 55 trials. The average time per trial for the first five trials was: to plane, 945 seconds; to door, 145 seconds. That for the last five trials was: to plane, 7.2 seconds; to door, 2.6 seconds. 110 K. S. LASHLEY AND S. I. FRANZ Retention was first tested on the second day after the opera- tion. The animal moved about the floor of the restraining cage for a few minutes, then settled down and remained motion- less. The same thing happened on the next three days. On the sixth day she became very spastic and during the tests passed through a series of prolonged tonic spasms, exhibited in arching of the back, retraction of the head, and gradual exten- sion of the legs and feet so that she came to stand only on the tips of her toes. This appeared after every rapid movement and persisted for sometimes as much as two minutes. On the fol- lowing day she was still more spastic and the spasms intervened at every attempt to walk. On the ninth day after the opera- tion she had partly recovered and succeeded in tripping the plane four times. The average time for these four trials was: to plane, 1117 seconds; to door, 70 seconds. On the following days she again became spastic and showed no further improve- ment in two weeks. Lesion (plate III, fig. 32). There is a transverse lesion com- pletely separating the frontal poles of both hemispheres from the remaining cortex, along a plane extending from the antero- dorsal face of the hippocampus, through the forceps of the cor- pus callosum to the base of the olfactory bulbs. There is a separation of the left corpus striatum from the external capsule and a partial invasion of the nucleus by large blood vessels. There is a similar degeneration of the anterior end of the right corpus striatum. After complete destruction of the frontal poles of both hemi- spheres and. partial destruction of the cerebral nuclei this animal gave no evidence of retention of the habit. Experiment 33. Operation on the frontal poles of the cere- brum in a large female rat, 142 days old. She had been trained for 120 trials on the inclined-plane box. The average time per trial for the first five trials was: to plane, 2970 seconds; to door, 28 seconds. That for the last five trials was: to plane, 34 seconds; to door, 2.2 seconds. Retention was tested on the second day after the operation. The animal was seemingly in good condition, ran about actively CEREBRAL FUNCTION IN THE RAT 111 and seemed fairly normal in behavior. She was placed in the restraining cage and ran about actively for five minutes but did not get up to plane. She seemed suspicious and explored with neck extended. She then settled down at the back of the box and remained motionless for an hour. On the two days follow- ing this behavior was repeated. There was never any specific reaction to the situation and after the first five minutes there was no normal exploration; she did not get on top of the food box in three hours on three consecutive days. She also was abnormally wild and tried to escape when picked up. Lesion (plate III, fig. 33). Right hemisphere. There is a lesion extending diagonally forward from above the hippocampus along the course of the external capsule over the corpus striatum to the base of the olfactory bulb, severing all the fibers laterad to the forceps of the corpus callosum but leaving the mesial surface of the pole intact. The lesion is filled by a large clot. Left hemisphere. The lesion extends somewhat farther back than that on the right, passing down through the anterior third of the corpus striatum into the lateral half of the cerebral pe- duncle. The mesial surface of both hemispheres is probably intact. After destruction of the anterb-lateral pole of both hemispheres this animal gave no evidence of retention. Seventeen animals are reported in this series. The extent of injury and the degree of retention is given in the following table. LESION RETAINED HABIT LOST HABIT RELEARNED HABIT Slight injury on dorsal convexity '. 1 0 Destruction of cortex on temporal and orbital surfaces 2 0 Partial destruction of frontal region 5 (l?j) 3 0 Complete destruction of frontal region " v-1- 'K (1?) 4 2 After partial or complete destruction of the frontal poles of both hemispheres some of the animals retained the habit, others lost it, and of these some reacquired it after training and others 112 K. S. LASHLEY AND S. I. FRANZ failed. What is the relation of these differences of behavior to the extent of the cerebral destructions? The figures of the brains are grouped on plate III for comparison. The first fact evident from the figures is that in every case where there was a clear retention of the habit the destruction of the anterior pole of one or both hemispheres was incomplete. In animals which showed a loss of the habit there is apparently a greater destruc- tion of cortex, so that in most if not all cases the whole of both frontal poles is involved. It is impossible to say how sharp is the distinction between these two groups when the extents of the lesions are compared. The maximum extent of the lesion has not been determined accurately in any case and the figures are based almost wholly upon the gross lesions. It is probable, however, that these are at least indices of the true extent of injury. Among the animals which showed no retention there was a practically complete destruction of the frontal poles of both hemispheres in experiments 27, 30 and 32, with almost complete destruction in experiment 28. In two of the remaining a part of the mesial cortex was preserved ; the extent of the lesion was not determined in the third. One animal which showed doubt- ful retention (experiment 25) was also found to have a complete destruction of the frontal region of both hemispheres. The only evidence for retention in this case was the solution of ,the problem in less than the normal learning time. There was no recognizable reaction to the plane which persisted after the operation and the successful movements were made as though by chance. The long average time (50 seconds) required by the animal to reach the food^ after tripping the plane gives fur- ther evidence that there was little or no retention of the habit. Allowing for the questionable nature of the retention in this last animal, we are justified in concluding that the complete destruction of the frontal region of the rat's brain results in the loss of the complex inclined-plane habit. After partial destruction of the frontal pole five animals re- tained the habit, one gave questionable evidence of retention and two showed no retention of the habit. Of the latter, one CEREBRAL FUNCTION IN THE RAT 113 (experiment 29) developed an abnormal stereotyped reaction to the situation and the other (experiment 33) was stuporous during most of the tests so that the two animals scarcely pro- vide data from which reliable conclusions can be drawn. With the exception of these two and the questionable case described in experiment 26, the animals with partial injuries to the frontal pole gave clear evidence of retention of the habit. Among them the one described in experiment 20 showed the most extensive lesion, only the left mesial region being left intact (plate III, fig. 20). In the others the lesions were more restricted, leaving the frontal pole of one or other hemisphere intact. This sug- FIG. 11. TOTAL EXTENT OF LESION IN ANIMALS OF SERIES IV WHICH RETAINED THE HABIT gests the question whether or not there is some particular part of the frontal region concerned with the retention of the habit. Figure 11 shows the combined extent of the lesions in the animals retaining the habit. Every part of the frontal region was destroyed in one or another of the operations. It seems, then, that although some part of the frontal region must remain intact if the plane-box habit acquired by the nor- mal animals is to be retained, the particular part preserved is immaterial. The different parts of the frontal region are, to adopt a term from experimental embryology, equipotential in the functioning of the habit. V. HABIT FORMATION AFTER DESTRUCTION OF LARGE AREAS OF THE CORTEX In spite of many dogmatic assertions concerning the function of the cerebrum in so-called associative memory we have not a single published account of any experiments which give con- P8YCHOBIOLOGY, VOL. I, NO. 2 114 K. S. LASHLEY AND S. I. FRANZ elusive evidence that the cerebrum is or is not necessary for habit formation. Burnett (4) failed to get decerebrate frogs to learn a simple maze but it is possible either that the operation resulted in such a disintegration of the animal's other habitual reactions that the incentive for learning the maze was no longer adequate, or that the maze presented a too complex habit and that a simpler habit might still have been acquired. The experiments certainly do not justify the author's sweeping conclusion that learning is not possible in the absence of the cerebral cortex. The statement in the introduction that learning is possible in the decerebrate frog is based upon unpublished results on the facilitation of the crossed reflex of the hind leg.1 Goltz (5) made an attempt to train his decerebrate dog but gave up quickly for fear that the training methods would result in the animal's death. Rothman (6) reported the acquirement of new motor coordinations in his decerebrate dog but gave no details of the experiment. It is not possible to be certain that the changes in behavior noted by these authors were not concomitants of recovery from operative shock rather than true examples of habit formation. It may be that no complex habits are acquired in the absence of the cerebral cortex but a fundamental point in the problem of the physiology of learning is involved in the possibility of the formation of simple habits wholly by the mechanisms of the spinal cord and brain stem. Is there any fundamental difference between the organization of the cerebrum and that of lower centers such as to give the former special functions which are lacking to the more primitive portions of the nervous system, or are the cerebral and spinal functions alike save for the possi- bility of greater complexities of reflex connections within the cortex? Failure or success in obtaining habit formation in 1 Some may object to. this as an example of learning, but it is undoubtedly true that any modification of an animal's behavior due to repeated stimulation (exclusive of fatigue phenomena) is properly called learning. The distinction between " associative memory" and other types of acquired reactions is by no means so clear as the exponents of this as a criterion of consciousness would have us believe. CEREBRAL FUNCTION IN THE RAT 115 de cerebrate animals will furnish significant evidence for this problem. As yet we have not been able to effect a complete decerebra- tion of the rat but by the destruction of large and varied areas we have obtained evidence, perhaps not yet conclusive, that the cerebral cortex is not functional in the formation of simple habits. A number of animals were operated upon to destroy the cor- tex over the dorsal, temporal, and frontal regions and such parts of the orbital and occipital surfaces as could be reached from above without lesion to the thalamus. Six of these survived and all have been trained successfully in the formation of simple habits. A record of the experiments is given below. In the diagrams of the brains of this series of animals (plate II, figs 34 to 39) the solid black areas represent the parts from which the cortex has been completely absorbed, the stippled areas those in which there was degeneration or evident loss of function through destruction of fibers. Experiment 34- A large opening was made in the calvarium of a large adult male rat and a scalpel (curved on the flat) was drawn around the frontal and temporal lobes on each side and thrust backward to the tentorium. On the day following the operation the animal began to walk about. He was very spastic, with back arched constantly, and occasionally remained motion- less for long periods in contorted positions. He would not eat, but drank milk when it was placed in his mouth and when his short vibrissae were touched with a pipette filled with milk he grasped the end of it with his teeth. He would not do this in response to a stick unless it were wet with milk. This reaction suggests the retention of smell. He gave no detectable reactions to light, to light contact, or to aromatic odors which are avoided by normal rats. He reacted to loud sounds, to heavy contact or diffuse pressure, and to, protopathic stimuli. On the second day he ate when his lips were pressed against the food, but when. his movements carried him away from the food he continued to gnaw at the edge of the food dish for some time. During the ten days that he was kept alive he never 116 K. S. LASHLEY AND S. I. FRANZ learned to find food by himself and had to be fed, chiefly with milk given from a pipette. Training. He was fed five or six times a day by placing the end of a pipette between his teeth and squirting milk into his mouth. He almost immediately began to reach for the pipette as soon as it was brought in contact with his short vibrissae. An attempt was made to train him to inhibit this reaction. He was first given a taste of milk, the pipette was then wiped off and brought in contact with his short vibrissae. He took hold of it but, getting no milk, let go immediately. This was re- peated every five seconds for fifty-five times with the following results. He reacted every time in the first five trials, four 5 4 3 2 1 0 \ § 8 i 8 I M \ \ 8 9 3 o CO £ \ ^ \ is \ ^ \ \ £ \ 'fe \ \ \ •g \ \ \ \ \ V \ / \ \ \ \ \ \ 5 25 50 75 100 FIG. 12. RATE OF LEARNING TO REJECT EMPTY PIPETTE The number of times that the pipette was taken in successive groups of five trials is plotted. times in the second five, once in the third, and not once in the fourth and fifth. After a 60 second interval ten more trials were made. He reached for the pipette the first time but not again. After another 60 second interval he reached once again, then failed to reach for nine trials; so after a third interval. After the fourth interval of 60 seconds he did not reach once in ten trials. Milk was then placed in his mouth twice, the pipette was wiped and offered again. It was grasped five times in the first five trials, once in the second five, and not thereafter. Milk was given again and after this he reached for the pipette four times in the first five trials, four times in the second, and not in the third and fourth. Quite similar results were obtained on the following day. The complete series is given in figure 12. CEREBRAL FUNCTION IN THE RAT 117 On the following days he became more and more deteriorated and was finally killed as he could no longer be induced to eat. An attempt was made to establish a conditioned reflex to con- tact with the foot. The animal was fastened in the container de- signed by Watson and the foot was fastened to a pointer in con- tact with a kymograph. The dorsal surface of the foot was then stimulated first by rubbing with the uncharged electrodes, then electrically. The first effect of electric stimulation is a quick jerk of the foot followed by an increased tonic contraction of the muscles for standing so that the animal, with repeated stimu- lation, rises gradually until he stands on the tips of his toes and maintains that position for as long as three minutes. In the apparatus the animal's reactions became more and more violent until, after the fourteenth stimulation he began to bite his thigh and had to be taken out. For five minutes there- after, when his right hind foot was touched he reacted by turn- ing his head to the foot and squeaking. Fifteen minutes later no reaction to contact with the foot was given. Lesion (plate II, fig. 34). The brain of this animal was one of those improperly sectioned so that an exact delimitation of the destroyed area has not been possible. As nearly as can be determined there is a total destruction or degeneration of all the cortex lying above and laterad to the third and lateral ven- tricles, a separation of the frontal pole from all underlying struc- tures at the level of the cerebral peduncles, a great shrinkage of both corpora striata, and a destruction of the medial portion of the occipital lobes. This animal had probably the greatest destruction of cerebral tissue of any reported in this study. His instinctive exploratory reactions were so limited that training on the ordinary laboratory problems was impossible. He nevertheless showed an ability to form very simple habits quite rapidly. Experiment 35. Deep transverse and horizontal incisions were made in the cortex of a large male rat, 115 days old. He recovered rapidly from the operation and gave no evidence of motor disturbances or of other abnormality of behavior. Training in the simple maze (fig. 1) was begun six days after 118 K. S. LASHLEY AND S. I. FRANZ the operation. He required only thirty trials to learn to go to the right. He was then required to go to the left and the earlier habit was completely overcome in twenty trials. The records of time and errors are given in table 1. TABLE Rate of learning for rat described in experiment 85. In this and the following tables the average time per trial and the total number of errors are given for successive groups of five trials TO RIGHT TO LEFT Time Errors Time Errors 43.0 3 28.0 4 35.5 3 8.0 1 16.5 1 7.5 0 8.5 0 6.5 0 7.5 0 6.5 0 He was killed 30 days after the operation. Microscopic examination of the brain showed (plate II, fig. 35) : Right hemisphere. A transverse lesion beginning above the posterior end of the corpus callosum and extending forward almost horizontally to the anterior pole, above the level of the corpus callosum. There is a superficial lesion of the lateral surface of the occipital lobe resulting from a knife cut which separated the cortex from the underlying fibers. A similar lesion extends over the orbital surface, separating nearly half of the lateral surface of the cortex from the underlying fibers and de- stroying the fibers of the external capsule back to the level of the hippocampus. The lateral ventricle was not reached by the injuries and the subcortical ganglia are intact. Left hemisphere. The lesion on the dorsal convexity is simi- lar to that on the right. There is a transverse lesion completely severing the frontal pole just in front of the caudate nucleus. A lesion on the orbital surface following the course of the ex- ternal capsule to the anterior end of the hippocampus has com- pletely destroyed a lens-shaped area on the orbital surface. The cerebral nuclei are intact. CEREBRAL FUNCTION IN THE RAT 119 After the destruction of somewhat more than the anterodorsal third of the cortex, together with a large area on the right occipital lobe, this animal showed no detectable abnormality in behavior or in the rate of habit formation. Experiment 36. A transverse opening 10 by 8 mm. was made in the calvarium of a small female rat, 115 days old, and trans- verse and longitudinal incisions were made in the cortex to as great a depth as possible. Recovery was rapid and without appreciable motor disturbance. She ate alone from the first and seemed practically normal in behavior except for a com- plete indifference to other animals of the same 'or opposite sex. Training in the simple maze was begun 12 days after the operation. The habit of going to the right was perfected in fifty trials. She had, apparently, an initial preference for the right side of the maze. Training to break up this and to estab- lish the habit of going to the left was then started and one hun- dred and twenty trials were required before ten successive error- less trials were made. The records of time and errors are given in table 2. TABLE 2 Rate of learning for rat described in experiment 36 TO RIGHT TO LEFT Time Errors Time Errors Time Errors 86.0 0 137.0 6 27.0 3 97.2 2 29.0 1 53.0 5 29.0 0 35.0 3 43.0 3 76.2 2 65.5 2 37.0 1 62.0 1 16.5 2 17.0 1 52.0 2 18.5 1 21.5 1 51.0 2 31.0 4 12.5 1 263.0 3 34.0 3 23.0 1 19.0 0 12.5 1 20.5 2 12.5 0 42.0 4 23.0 1 24.5 2 9.5 0 15.0 0 8.5 0 The rat was autopsied 40 days after the operation. Micro- scopical examination of the brain shows (plates I and II, fig. 36) : 120 K. S. LASHLEY AND S. I. FRANZ Right hemisphere: Destruction of the entire corpus callosum including the knee and all cortical tissue above it and above the hippocampus. A transverse lesion extends laterad from the anterior end of the lateral ventricle along the forceps of the corpus callosum, through the external capsule to the cortex and ventrad to the cerebral peduncle. The fornix is severed with a slight injury to the adjacent thalamus. The lateral ventricle is much enlarged but the corpus striatum is seemingly intact. Left hemisphere. The corpus callosum is destroyed as on the right with absorption of all cortical tissue above its level. There is a diagonal lesion from the anterior horn of the lateral ventricle to the base of the olfactory lobe, probably involving all the cortex laterad to it. Fornix destroyed and lateral ventricle enlarged until it occupies half of the horizontal area of the lobe. Corpus striatum and thalamus nearly completely destroyed. The animal shows destruction of the antero-dorsal half of the cerebrum without marked loss in ability to form simple habits. Experiment 37. A transverse opening, 11 by 6 mm. was made in the calvarium of a large male, 200 days old, just back of the frontal-parietal suture and through it a transverse frontal and longitudinal temporal incisions were made through the cortex, followed by great hemorrhage. Killed 30 days after operation. The animal recovered quickly, moving about and reacting to other animals within four days after the operation. He was fed by hand for the first four days, then learned to find food for himself but would not eat with the experimenter near. An abscess developed on his neck (infection of the cerebral glands) but cleared up in 20 days. He showed fewer effects of the opera- tion than any other rat in this group, being the only one which showed any responses to other animals. This did not extend to normal sexual reactions. Training in the simple maze was begun twelve days after the operation. At this time the animal showed no motor distur- bance and was normal in his reactions to the experimenter. He was first trained to go to the right in the maze and required fifty trials for learning. He was then required to go to the left CEREBRAL FUNCTION IN THE RAT 121 for food and again learned in fifty trials. The average time re- quired and the total number of errors made in successive groups of five trials are given in table 3. Microscopical examination of the brain showed (plate II, fig. 37): TABLE 3 Rate of learning for the rat described in experiment 37 TO BIGHT TO LEFT Time Errors Time Errors Time Errors Time Errors seconds seconds seconds seconds 143.0 4 9.5 0 31.0 4 11.0 0 143.0 4 14.0 1 17.5 1 12.0 0 15.0 0 14.5 1 16.0 2 21.0 1 42.0 4 8.0 0 18.5 1 19.5 0 19.0 1 11.0 0 15.5 2 12.0 0 Right hemisphere. Destruction of the greater part of the convexity of the cortex from the occipital portion of the corpus callosum to the knee, involving the greater part of the corpus cal- losum and exposing the lateral ventricle and extending around the anterior face of the caudate nucleus to the cortex, separating the anterior pole to the level of the olfactory fibers. The fibers of the anterior pole have been absorbed, leaving a large cyst. The longitudinal incision extends backward from the transverse lesion to the middle of the caudate nucleus, then along the course of its fibers, through the external capsule to the cortex, producing, apparently, very little destruction. Left hemisphere. The lesion on the dorsal convexity is simi- lar to that on the right. The transverse lesion passes just in front of the knee of the corpus callosum and diagonally forward to the base of the olfactory lobe, severing the entire anterior pole to the level of the olfactory tracts. There is invasion of the anterior end of the caudate nucleus by hemorrhage. The longitudinal incision seems to have passed along in the lateral ventricle and to have done little damage. There is an extensive superficial lesion extending over the occipital lobe. 122 K. S. LASHLEY AND S. I. FRANZ Experiment 38. Transverse frontal and longitudinal incisions were made in the cerebral cortex of an old male rat, through a circular trephine hole 8 mm. in diameter. After the operation he was very weak and emaciated rapidly. He was fed by hand for five days, then recovered the ability to eat alone. From the day after the operation he showed a pronounced spasticity of the left side, carrying his head rotated about 60 degrees to the right and walking with the left hind leg extended stiffly and only the tips of the toes set down on the ground. During the first few days he fell over at every third or fourth step and always made a complete rotation to the right in getting to his feet again, his behavior resembling that of an animal after uni- lateral destruction of the semicircular canals. In the cage the animal was much more restless than any of the others in the group. He ran almost constantly around an oval path, bounded by the sides of the cage, and always in the same direction, clock-wise. On one day, 35 days after the operation, his activity was observed nearly continuously for 12 hours. During this time he ran around the same path at an average rate of twenty devolutions per minute, with a total of not more than one hour's interruption for rest and eat- ing; a total distance of not less than seven miles. During the 40 days that he was kept under observation he showed very little improvement in motor control beyond that acquired in the first week after operation. Habit formation. Training in the simple maze was begun 12 days after the operation. He was active, although his move- ments were very badly coordinated. On the first trial in the maze each day he had great difficulty in walking and his first passage of the central alley was usually a series of somer- saults. After the first trial coordination was usually better. After sixty to seventy trials in the maze he learned to tra- verse it with a few falls but this was accomplished by support- ing himself against the sides of the alleys and not by improve- ment in walking in a straight line. He always had great diffi- culty in turning around to his left and when he finally learned to go to the food in the left alley of the maze he accomplished CEREBRAL FUNCTION IN THE RAT 123 the turn at the end of the central alley by holding to the end of the partition with both left feet and pushing his body around with the right. He learned to turn to the right for food after forty trials. The food was then placed in the left alley. The previous habit of turning to the right, reinforced by his motor difficulty in turning to the left made the new habit very difficult for him to acquire and eighty trials were required before he made ten in succession without error. The records of time and errors are given in table 4. TABLE 4 Rate of learning for rat described in experiment 88 TO RIGHT TO LEFT Time Errors Time Errors Time Errors 486.0 1 615.0 100 332.0 30 53.0 2 55.0 12 68.0 6 302.0 5 216.5 41 58.0 7 45.0 1 44.0 35 19.0 0 59.0 2 141.0 12 19.0 1 195.0 0 86.0 25 20.0 2 17.5 0 383.0 91 15.0 2 14.5 1 112.0 20 11.5 0 408.0 31 10.0 0 He was killed 40 days after the operation. A small abscess filled with pus was found over the cranial opening but did not extend into the cortical tissues. There was a deep pit in the cortex, completely exposing the lateral and third ventricles. Microscopic examination showed a complete destruction of the convexity of the cortex including the whole of the corpus callo- sum to the knee (plates I and II, fig. 38). The lesion extends from just behind the corpus callosum diagonally forward to the base of the olfactory lobes, separating the entire frontal pole on either side. The longitudinal incisions extending down over the temporal and orbital surfaces of the cortex to the level of the ventricles with degeneration of neighboring areas. The right corpus striatum is degenerated and infiltrated by large blood 124 R. S. LASHLEY AND S. I. FRANZ vessels. The right half of the fornix is severed and there is an extensive clot filling the third ventricle. There was in this animal a destruction of fully half the cortical tissue, including the frontal pole and nearly all of the dorsal con- vexity and embracing all of the excitable area and the area described as motor by the histologists ; degeneration of the right corpus striatum and fornix ; permanent spasticity confined to the left side] no loss of learning ability beyond that due to motor incoordination. Experiment 39. Transverse frontal and longitudinal incisions were made in the cortex of a small male rat, 126 days old. He was very weak after the operation and grew emaciated rapidly. A hemiparesis appeared immediately after the operation, giving him some difficulty in walking and making it impossible for him to sit up on his hind feet. In the home cage or in the maze when his nose came in contact with the food he at once swung his head to the left, away from the food and against the side of the food dish, which he then gnawed. In order to eat without constant help he had to be placed in a large dish of food. He then usually twisted his head around to the left until his nose was under his left hind leg and ate in that position, gradually straightening out if left to himself. This behavior persisted until the animal was killed, twelve weeks after the operation. During the first three weeks after the operation he had an almost constant erection of the penis and masturbated at frequent intervals. There was a gradual recovery from this but a seem- ing deterioration in other respects as he became more and more inactive and stuporous. He was trained to go to the right in the simple maze 11 days after the operation. His paresis made it easier for him to turn to the left and he showed little improvement in accuracy after two hundred trials. The passage-way to the left was then blocked and he was given one hundred trials with only the direct way to the food open. He soon ceased to pause at the blocked entry and ran directly to the food, but when the left passage was again opened he at once entered it and did no better than before. In all, four hundred and fifty trials were given but he only once made five successive trials without error. An examina- CEREBRAL FUNCTION IN THE RAT 125 tion of the records of the first two hundred trials (table 5) shows, however, that he at first made some improvement both in time and in the number of errors and that this improvement was per- manent. This is evidence for learning, though of a simple type. In the later part of training he developed a stereotyped reaction to the maze which must be classed as a habit, though it did not contribute to the getting of food. In the trials, day after day, he would first turn into the entrance of the right alley, advance for twice his body-length, then whirl about quickly, run to the TABLE 5 Rate of learning in rat described in experiment 89 To right TIME ERRORS TIME ERRORS TIME ERRORS 149.0 3 19.5 2 63.0 8 159.0 6 25.0 3 39.0 2 88.0 4 62.0 2 48.0 6 111.0 5 47.0 1 37.0 6 111.0 7 60.0 6 28.0 1 99.0 6 28.0 2 19.0 0 70.0 4 66.5 7 69.0 8 95.0 6 44.0 3 76.0 5 96.0 6 48.0 5 31.0 3 58.0 3 37.0 3 23.0 2 75.0 6 37.0 3 49.0 4 59.0 5 79.0 3 19.0 1 65.0 5 59.0 3 29.0 4 62.0 2 61.0 4 33.0 4 end of the left alley, turn back to the right alley and go directly to the food. Records of time and errors for the first two hun- dred trials are given in table 5. The rat was killed 71 days after the operation. Microscopical examination of the brain showed (plates I and II, fig. 39) : Right hemisphere. There is a destruction of tissue over the dorsal convexity of the cortex down to and including the corpus callosum. There is transverse lesion extending diagonally for- ward just behind the knee of the corpus callosum to the base of the olfactory bulb, completely cutting off the frontal pole. The for nix and gyrus dentatus are completely destroyed. The cor- 126 K. S. LASHLEY AND S. I. FRANZ pus striatum is cut through in sagittal section for its full length and the lesion filled by a large clot which extends down into the cerebral peduncle, cutting through its fibers obliquely. The anterior portion of the caudate nucleus is replaced by a clot, the remainder is much shrunken. The antero-lateral surface of the thalamus shows indications of degeneration. Left hemisphere. There is destruction of all tissues above and including the corpus callosum. A transverse lesion extends diagonally from behind the knee of the corpus callosum to the base of the olfactory bulb, severing the anterior pole. The lateral ventricle is collapsed and all the cortex laterad to it is replaced by scar tissue. The external capsule lies free in a cyst laterad to the corpus striatum and, in the occipital lobe, is re- placed by a large cyst. The corpus striatum is degenerated. The fornix is largely destroyed. The thalamus is intact. In this animal the destruction of all the cortex above and in front of the corpus callosum and laterad to the left lateral ventricle with partial degeneration of the right and complete degeneration of the left caudate nuclei, destruction of the fornix and injury to the thala- mus on the right was followed by inability to solve the simple maze but did not preclude the formation of simple habits as evidenced by the gradual appearance of stereotyped modes of response to the maze. The -course of learning in these different animals is compared with the average of normal individuals trained under the same conditions in figure 13. The animals described in experiments 35 and 37 were apparently normal in behavior and learned in about the same amount of practice as is required by normal rats. The animals in experiments 36 and 38 learned to go to the right quite readily but had a great deal of difficulty in readjusting to the problem when they were required to overcome the first habit and learn to turn to the left. The rat in experiment 39 was un- able to learn the problem in more than three hundred trials. However, even this animal showed marked improvement in the time required for running the maze during the early part of practice. In experiments 38 and 39 a part of the difficulty experienced by the animals was clearly due to a persistent hemi- CEREBRAL FUNCTION IN THE RAT 127 paresis which made it easier to turn to one side than to the other. They did show, however, some indications of abnormality in learning not due to the motor difficulty. The development of stereotyped errors in experiment number 39 and the large num- ber of errors of the same type made by the rat in experiment 500 4-00 300 200 100 100 0 tTS ZOO trials. normal ey.p. 35 exp.39 7S 100 trials FIG. 13. A COMPARISON OF THE RATE OF LEARNING THE SIMPLE MAZE BY NORMAL ANIMALS AND BY ANIMALS AFTER DESTRUCTION OF THE ANTERIOR AND DORSAL SURFACES OF THE CORTEX 38 in the second training test (table 4) are without parallel in the behavior of normal animals. They indicate, perhaps, an abnormal predominance of kinesthetic-motor chain reflexes and a difficulty in inhibiting them in new situations. This is suggestive of the perseverations sometimes seen in human defectives. 128 K. S. LASHLEY AND S. I. FRANZ In every one of the animals clear evidence of some degree of habit-formation was obtained. Whether or not the possible complexity of habit was limited by the destruction in these cases was not definitely determined by the experiments. No attempt was made to train the animals in more difficult tasks and until experiments can be carried out on a more extensive scale we can conclude only that somewhat more than the anterior half of the cerebral cortex in the white rat (fig. 14) is unneces- sary for the formation of simple habits. FIG. 14. THE TOTAL EXTENT OF THE COMBINED LESIONS IN THE ANIMALS WHICH LEARNED THE SIMPLE MAZE AFTER OPERATION. SERIES V Discussion To the destruction of what structural elements may the loss of the plane-box habit be ascribed, and to what the reten- tion in the cases of partial destruction? The lesions in many cases involve the motor area and the loss of the habitual reac- tion might be considered as primarily the result of the motor disturbances arising from injury to this region. However, such facts as we have at hand concerning the motor area in the rat do not support this view. The cyto- and myelino-architecture of the br.ain of the rat has not been worked out but Isenschmid (7) has given a full account of the arrangement and composition of the cell-layers of the different regions of the cortex of the mouse and the similarity of the brains of the two species is probably sufficient to justify the application of his results to the rat. He finds three principle areas, fronto-medial, dorso-lateral, and sub-occi- pital, distinguished by the relative thickness of the cell-layers and the size of the elements, and within these distinguishes a number of subordinate areas. Those which have come within the scope of the present experiments are shown in figure 15. CEREBRAL FUNCTION IN THE RAT 129 The majority of writers agree that there is a frontal field, variously located in the region I which has not the characteristics of the motor areas, but no statements are made as to its rela- tion to the frontal lobes of higher mammals. The stimulation experiments which we have carried out indicate that even the extreme pole of the cortex is excitable, that there is no silent area corresponding to the frontal lobes of higher mammals, but the small size of the brain and the difficulty of preventing some spreading of the current would make it difficult to detect such an area if it existed. There is no experimental evidence concerning the function of these fields, although a part has been interpreted as motor, FIG. 15. THE CHIEF CYTOARCHITECTURAL AREAS OF THE BRAIN OF THE MOUSE. AFTER ISENSCHMID another as sensori-motor. According^ to Isenschmid, who has summarized the literature of this subject the area described as motor differs considerably in the descriptions of different authors. Upon purely histological grounds Cajal includes an area extend- ing over a, a\, and b, in figure 15. Dollken (8) includes the areas I, k, a, ai, 6, and the anterior portions of q and d Brod- mann (9) includes k and a part of field I. Without physiological verification conclusions as to the func- tion of histologically differentiated parts can have little value. The stimulable area has been found to correspond most closely to the anatomical field described by Dollken. The excitable areas thus far determined are shown in figure 16. They em- PSTCHOBIOLOGT, VOL. I, NO. 2 130 K. S. LASHLEY AND S. I. FKANZ brace the entire frontal pole and the areas I, k, a, ai, and a part of the anterior portions of q and d (fig. 15) including most of the area described by Dollken but not extending so far over the parietal surface as his figures indicate. In the operative experiments this entire motor area was fre- quently destroyed completely without serious interference with the animal's ability to form kinesthetic-motor habits. After purely cortical destruction we have never found any motor disturbance which has persisted for as much as eighteen hours after the operation; in particular there is no indication of localized disturbances in the front or hind limbs resulting from FIG. 16. THE DISTRIBUTION OF THE STIMULABLE AREAS IN THE BRAIN OF THE RAT the destruction of corresponding regions of the stimulable area. In those cases where motor disturbances were observed there appears to have been always a very extensive injury to the cor- pus striatum. (The material is being reexamined with reference to this point.) It seems clear, then, that the loss of the habits was not due directly to motor paralysis. In experiments 27, 30, 32, and 33 there was destruction of the frontal pole including at most only the stimulable • area for the head and neck and, with this, complete loss of the habit. So it seems almost certain that in these cases the loss of the habit was due to the interruption of some other cortical reflex path- ways than those included in the excitable area. The effects of CEREBRAL FUNCTION IN THE RAT 131 the operations upon retention can not be looked upon as the result of a motor insufficiency, nor, probably, as primarily the result of the destruction of the excitable area. Previous work (10) on the function of the frontal portion of the cortex in higher mammals (cat and monkey) has given evidence that this part of the cortex is normally utilized in habit-formation, but that in its absence some other portion of the brain may usurp its function. Further, there is evidence that when the habits have been practiced for a long time by normal animals the preservation of the frontal lobes is no longer a condition of their correct functioning; they come to be carried out, probably, at a lower level of nervous organization. It was the primary purpose of this study to determine the amount of practice necessary for retention of a habit after ablation of the part of the cortex originally functional in its performance. This purpose has been largely defeated by the great variation in the extent of destruction in different animals. The results, how- ever, verify for the rat the findings for higher mammals and suggest some additional possibilities of interest. The experi- ments of series I, II, and V indicate either that very simple habits such as that of turning in the simple maze may be ac- quired by normal animals without the utilization of the frontal portion of the cortex at any time during learning, or that the cortex very rapidly ceases to function in the performance of the habit. There is a little evidence, noted in the first paper, that the animals of series I which have been long overtrained retained the habit a little more perfectly than those which were trained to perfect performance. The irregular conditions of training incident to the use of the animals in other experiments and the great variation in the extent of the lesions makes the justice of such a conclusion questionable. In the more complex habit of the inclined plane box the presence of some portion of the frontal pole is evidently a condition for the performance of the habit where this has been acquired with the frontal pole intact. Again, owing to the small number of animals tested and the variation in extent of the lesions the data on the relation of the amount of practice to the functions 132 K. S. LASHLEY AND S. I. FRANZ of the cortex is inadequate. From experiments 29, 31, and 33 in which animals were overtrained it seems that an amount of overtraining equal to three or four times that required for learn- ing is insufficient to reduce the habit to lower brain levels. The difference in the complexity of the maze and inclined-plane box habits and the apparent failure of the latter to be reduced to subcortical levels within the limits of these experiments suggest that the amount of practice necessary for the assumption of the cortical functions by subcortical ganglia is proportional to the complexity of the habit. Until further evidence is accumulated this can not be considered more than a probable assumption. The data presented in section IV shows that complete destruc- tion of the frontal pole results in the loss of the plane-box habit, whereas the loss of the temporal and parietal regions is without effect upon the habit. This indicates that conduction path- ways involved in the performance of the habit pass through the frontal pole, but an attempt to localize these more accurately brings to light a complexity of function which has not before been suggested. Partial destruction of the frontal pole did not always result in the loss of the habit. An attempt to find a correlation between the part destroyed and the retention or loss of the habit revealed the further fact that no single part of the frontal pole escaped destruction in all the animals which retained the habit (fig. 14). It seems then that for retention some part of the frontal pole must be preserved but no particular part seems necessary. This is, perhaps, what might be expected if we abandon the purely diagrammatic concept of the reflex cortical arc as a single chain of neurones and consider that for the per- formance of even a simple movement a number of such arcs are required. With a vast number involved in complex habitual acts, it is very improbable that all would be projected on a re- stricted area of the cortex; the experimental results suggest rather that in their cortical relations these arcs are widely dis- tributed over the frontal pole so that a partial lesion results in only the partial destruction of the arcs involved in the perform- ance of any simple act. The experiments of series V have shown that the rat may form CEREBRAL FUNCTION IN THE RAT 133 simple habits after the complete destruction of all the cortex of the frontal, temporal, and parietal regions and the greater part of that on the orbital surfaces. This includes all the regions to which the function of habit formation has been ascribed and leaves only those which have been thought to have visual, audi- tory, and olfactory functions. The destruction of cortical tissue has not been extensive enough to prove that learning may take place wholly at the level of the sub-cortical centers but the evi- dence at hand is sufficient to justify more extensive experiments upon this point. The ability of the animals to form habits after the loss of those parts of the brain which are normally used in learning, the reestablishment of motor control after the loss of the stimulable area of the cortex and of the corpus striatum, and the seeming equipotentiality of the different parts of the frontal pole in the functioning of complex habits go far toward establishing the complete functional interchangability of all parts of the cerebral cortex. SUMMARY Rats were trained after destruction of various parts of the cerebral cortex including the frontal, temporal, parietal, and a large part of the orbital surfaces and the influence of the cerebral destruction upon their ability to form and retain kinesthetic- motor habits was tested. It was found that: 1. The habit of turning correctly in the simple maze may be retained after the destruction of any part or all of the cortex lying in front of and above the knee of the corpus callosum and after the destruction of any part of the temporal and parietal regions. 2. The maze-habit may be acquired after the destruction of all the cortex included within these areas, and after the destruc- tion of one, perhaps both, of the striate nuclei. 3. The more complex habits involved in opening the inclined- plane box are retained after destruction of the temporal regions of the cortex. 4. The complete destruction of the frontal regions of the cortex results in the loss of the inclined-plane box habit. 134 K. S. LASHLEY AND S. I. FRANZ 5. The partial destruction of the frontal region does not abolish this habit. 6. So long as the destruction of the frontal pole is not complete the habit is retained, apparently irrespective of what part of the frontal region has been destroyed. 7. No marked motor disturbances appear after the complete destruction of the stumulable areas of the cortex but in certain cases marked hemiparesis seemed to result from the destruction of the corpus striatum. REFERENCES (1) This journal, 1917, i, 1-15. (2) WATSON, J. B.: Behavior, New York, 1914. (3) FRANZ, S. I.: On the functions of the cerebrum. American Journal of Physiol., 1902, viii, 1-22. Archives of Psychol., New York, 1907, no. 2, 1-64. (4) BURNETT, T. C.: Some observations on decerebrate frogs with especial reference to the formation of associations. Amer. J. of Physiol. 1912, xxx, 80-87. (5) GOLTZ, F.: Der Hund ohne Grosshirn. Arch. f. d. ges. Physiol., 1892, li, 570-615. (6) ROTHMAN, M.: Demonstration des Hundes ohne Grosshirn. (Ber. iiber d. V. Kongress f. exper. Psychol.) Leipzig, Barth, 1912, pp. 256-260. (7) ISENSCHMID, R. : Zur Kenntnis der Grosshirnrinde der Maus. Abh. d. Konigl. Preuss. Akad. d. Wiss., 1911, 1-46. (8) DOLLKEN, I.: Beitrage zur Entwickelung des Saugergehirns. Lage und Ausdehnung des Bewegungscentrums der Maus. Neurol. Centbl., 1907, xxvi, 50-59. (9) BRODMAN, K. : Vergleichende Lokalisationslehre der Grosshirnrinde. Leip- zig, 1909. (10) FRANZ, S. I. : Opus cit. PLATE I FIG. 22. Serial horizontal sections, 1.6 mm. apart, of the brain of the rat de- scribed in experiment 22. The positions of the structures used as reference points in the descriptions of the lesions are noted on the figures. C.ex., external capsule; c.st., corpus stratium; /., fornix; f.c., forceps of the corpus callosum; h., hippocampus; n.c., anterior olfactory nucleus; n.L, lateral thalamic nuc- leus; o., base of olfactory bulb; s.p., septum pellucidum; v.l., lateral ventricle. FIG. 38. Serial sections of the brain of the rat described in experiment 38. The method of reconstruction is shown in this and plate II, figure 38. FIG. 36. Horizontal section through the brain of the rat described in experi- ment 36. FIG. 39. Horizontal section of the brain of the rat described in experiment 39, showing a probable complete destruction of the left hemisphere. Is g 135 PLATE II FIGS. 1 to 14. The extent of the lesions in animals which retained the habit of the simple maze after cerebral destruction. The numbers of the figures corre- spond to the numbers of the experiments under which the animals are described. In figures 10, 11, and 12 the striated areas show the extent of the second operation. FIGS. 34 to 39. The extent of the lesions in animals which formed simple habits after operation. The solid black areas are those in which the cortex was completely absorbed. The dotted areas are those in which the cortex was proba- bly not functional as a result of the destruction of its descending fibers. 136 CEREBRAL FUNCTION IN THE RAT K. 8. LASHLEY AND S. I. FRANZ PLATE II 137 PLATE III FIG. 17: Lesion in the control animal in which there was no operation on the brain. FIGS. 18 and 19. Lesions in the control animals after operation on the tem- poral lobes. FIGS. 20 to 24. The extent of the lesions in the animals which retained the inclined-plane box habit after operation. FIGS. 25 and 26. The extent of the lesion in the animals which showed a doubt- ful retention of the habit. FIGS. 27 to 33. The extent of the lesion in the animals which lost the inclined- plane box habit after operation. 138- CEREBRAL FUNCTION IN THE RAT K. S. LASHLEY AND S. I. FRANZ PLATE III 139 THE EFFECTS OF STRYCHNINE AND CAFFEINE UPON THE RATE OF LEARNING K. S. LASHLEY Department of Psychology of the Johns Hopkins University Among the hypotheses which have been advanced to account for the reintegration of conduction paths in learning three stand out as rather definitely opposed to one another in the neural processes which they imply. The hypothesis suggested by Ladd and Woodworth ('11) assumes inhibition of successive activities as the fundamental process which results in the selec- tion and fixation of random activities. The second hypothesis assumes nervous reinforcement as the fundamental process by which successive acts become linked together in habit-formation. In its vaguest forms this hypothesis refers to pleasure and pain as the reinforcing agents without any attempt to analyze the way in which they act. Its most concrete expression is that given by Angell ('09), though the diagram which he gives has little correspondence with physiological facts. The third hy- pothesis depends chiefly upon the chance spreading of nervous excitation (Watson, '14 a), or the simultaneous activation of two afferent pathways in such a way that the final common path of one is able to divert the discharge of the other and so bring about a permanent connection between itself and this afferent path (Max Meyer, '11), (Pawlow, '14). If Sherrington's explanation of the spreading of nerve impulses as a result of fatigue of the final common path be accepted, the mechanism implied in this hypothesis is relatively independent of inhibi- tory and reinforcing activities of the nervous system. The studies of Verworn and of Sherrington have shown that the primary action of strychnine upon the central nervous system is the reduction of inhibitory processes and their final conversion into processes of excitation. A means is thus pro- 141 142 K. S. LASHLEY vided for inducing functional changes in nerve conduction wfiich may have a direct bearing upon the hypotheses mentioned above. It may serve, at least, to distinguish the first, the " inhibition theory," from the others, since, if this theory be true, the reduc- tion of inhibitory processes should retard learning, while the resultant increase in the ease of spread of excitatory processes might be expected to accelerate learning, either by increasing the variety of activities available for selection by repetition (third hypothesis), or by increasing the effects of reinforcing agents such as are presupposed by the second hypothesis. Thus, if strychnine retards learning, it should lend some support to the first hypothesis, if it accelerates it, it will give evidence in favor of the second or third. The interpretation of the results may be complicated, how- ever, by the locus of attack of the drug. Experimental work has located this definitely only in the intercommunicating cells of the dorsal horn of the spinal cord, but the involvement of the entire body in the convulsions of acute strychnine poison- ing indicates a similar reduction of inhibitory processes in the midbrain, or perhaps even in the cerebrum. The reduction of stimulus threshold found by Amantea ('15) after application of strychnine to restricted parts of the cortex also shows an in- creased excitability of the cortex and probably closely similar action of the drug upon cortical and spinal cells. This evidence seems sufficient to justify the interpretation of any effects of the drug found as due to its general reduction of synaptic re- sistances. It was with essentially the above problem in mind that I began a study of the effects of strychnine upon habit-formation. Since caffeine is generally stated to have a physiological action similar to that of strychnine but confined to the cerebrum and since it is so generally used as a stimulant it was included with strychnine in this study of the effects of the central excitants. The three hypotheses outlined above, while most widely accepted, by no means exhaust the possible explanations of the learning process and the results reported in the following pages have, perhaps, complicated rather than simplified the EFFECTS OF STRYCHNINE UPON RATE OF LEARNING 143 problem. Accumulating evidence of the dependence of the direction of nerve conduction upon the relative refractory phases of the neurons involved (Forbes and Gregg/ 15) and the growing acceptance of the all-or-none law of nerve action make it seem less and less probable that any of the current accounts of the neurological basis of habit is in any way adequate to ac- count for the facts. The presentation of the problem in this form has nevertheless seemed of value in spite of the possible falsity of all the hypotheses which it is designed to test, if only as a means of emphasizing the strictly physiological nature of the problem of learning and the necessity for a more concrete formulation of this problem than has hitherto been made. METHODS IN THE PRESENT EXPERIMENTS For testing the effects of the central excitants, albino rats, approximately sixty-five days old, were used. They were trained in the circular graphic maze (Watson, '14 b), time, distance, errors, and number of trials required for learning being recorded. The technique of training was essentially that de- scribed by Hubbert ('15) with the difference that each animal was given five trials per day and that the training was continued until three successive errorless trials were made during one day's practice. Administration of drug. The drugs given were strychnine sulphate (Powers and Weightman) and caffeine (pure alkaloid, Merck) in aqueous solution. They were injected subcutane- ously in solutions of such strength that 0.10 cc. contained the desired dose. The minimum lethal dose of strychnine for an 150 gram rat was determined at the beginning of the experi- ments as about 0.50 mgm. This is usually fatal in about fifteen minutes but its effect varies somewhat with the concentration of the solution and with individual differences in the animals, one rat being killed by 0.10 mgm., others surviving 0.50 mgm. without convulsions. For study, two concentrations of the drug were used, 0.10 mgm. and 0.05 mgm. per 0.10 cc. of solu- tion. These correspond by weight to doses of about 0.8 and 0.4 144 K. S. LASHLEY grains, respectively, for a 150 pound man, or sixteen and eight times the maximum therapeutic dose. The caffeine was administered in doses of 0.50 mgm. and 1.00 mgm. These were chosen arbitrarily as corresponding by weight to 4 and 8 grain doses for man. The drugs were adminis- tered each day exactly ten minutes before the beginning of training. This time was chosen as the interval after which the first convulsion appears, following the minimum lethal dose of strychnine. Since Goldscheider and Flatau ('98) have found changes in the cells of the dorsal horn within three minutes after the subcutaneous injection of strychnine the ten minute in- terval seems ample time to assure the full effect of the drug during training. The absence of any easily recognizable effect of the caffeine made it impossible to note its time of action so the same interval between the administration of the drug and the beginning of training was used with it as with strychnine. A Luer hypodermic syringe graduated in hundredths cubic centimeters was used for injection. Owing to the small quantity (0.10 cc.) of fluid injected some variation in the size of the dose from day to day was unavoidable. With care, however, it was possible to keep this well within 10 per cent of the total quantity injected.1 Method of training. The course of training was as follows. On three consecutive days the rat was confined for fifteen, ten, and five minutes respectively in the feeding compartment of the maze and was given no food except what was eaten there. On the fourth day training was begun. For the first trial the rat 1 The technique devised for injection has proved so satisfactory that I be- lieve it can be substituted advantageously for that used in many types of work. The usual methods of confining the animals for injection were found to excite them to an extent which threatened to interfere seriously with the experiments and it seemed necessary that the experimenter have both hands free to manipu- late the animals. The hypodermic was therefore fastened by a clamp to project over the edge of a table above which was attached a long lever bearing upon the plunger of the syringe and extending below the table-top so that it could be moved by the experimenter's knee. His hands were thus left free to hold the animal without exciting it, and with careful handling, the animals rarely showed any sensitivity to the injection. With practice it is possible to control the quantity of solution injected by this method to within 0.005 cc. EFFECTS OF STRYCHNINE UPON RATE OF LEARNING 145 was pushed out of the starting compartment and the door was closed behind him to avoid the long delay which usually occurs there. In all later trials he was allowed to take his time in starting unless the delay exceeded ten minutes. When he reached the food compartment he was not allowed to retrace the path but was returned to the starting box immediately, getting rarely more than one or two bites of food. After the fifth trial he was confined in the food compartment and allowed to eat for five minutes; then was not fed again until the comple- tion of the next day's practice. If the rat did not reach the food compartment within one hour after starting, he was returned to the home cage, without food, and the trial was continued on the following day. If the food compartment was not reached within less than three hours the animal was discarded. This occurred in the case of only one rat. Throughout the experiments the animals were fed exclusively on bread and milk and of this only so much as was eaten during the five minutes' confinement in the feeding compartment of the maze. On this diet a few lost weight rapidly but the average loss of all the animals trained was only 5.1 grams, or about 4 per cent of their average weight at the beginning of the experi- ments. There was no apparent relation between the loss or gain in weight and the drug administered. Criteria of learning. In computing the results the number of trials preceding the first run without error and also the number preceding three successive errorless runs on the same day have been used as the most dependable criteria of the respective abili- ties of the groups.2 In these experiments and probably in all others where differ- ently treated groups of animals are studied, the rate of running is modified by the differential treatment so that the time con- sumed in learning, when considered alone, is not reliable as a criterion of the amount of practice. There is no evidence that the distance traversed is similarly influenced and this may be 2 1 have dealt with the relative reliability of these two criteria in another paper (Lashley, '17). 146 K. S. LASHLEY taken as the more trustworthy index to the amount of practice. Since the total distance traversed in learning is greatly influenced by the number of trials, irrespective of whether or not errors are made in all the trials, the total distance traversed during learning is not an accurate index to the number of errors made. The total distance in excess of the direct path through the maze is, on the contrary, closely correlated with the num- ber of errors made, and may perhaps be looked upon as the most accurate measure of error available since we can in no other way evaluate the distances to which the animals penetrate the blind alleys before returning to the direct path. The relative values of the number of trials and the number of errors made as indices of the effort consumed in learning can- not be determined at present. It is not necessary that a blind alley be entered in order that the alternative entering of the correct path be fixed in the habit. On the other hand it is possible that the failure to obtain food in the blind alley is an important factor in the fixation of the movements required for running over the true pathway. The data on excess distance are given therefore as contributory evidence, the value of which cannot be determined at present. Records were made, also, of the time taken on each trial and of the total distance traversed, thus giving an index of the relative activity of the different groups as determined by their rate of running, and also of the approximate number of errors made by the different rats. The rdle of chance in determining the results. In every experi- ment where the course of learning is studied in groups of sub- jects trained under diverse conditions some control of chance variations in the subject and in the conditions of training is necessary. In very few of the recorded experiments of this type, whether performed on man or animals, have enough sub- jects been used to make the differences found significantly greater than their probable errors, computed by the usual formulae. Consequently, the results are suggestive rather than conclusive. The significance of the results based upon few subjects may be increased by a careful control of possible variable factors EFFECTS OF STRYCHNINE UPON RATE OF LEARNING 147 other than those which are intentionally varied in the experi- mental procedure. Careful selection of the subjects, and rigorous duplication of the experimental conditions may go far towards the elimination of chance variations in the results. The ideal control of small numbers of subjects is to train all at first in exactly the same activity and then with the results of this training as an index of individual variability, to introduce experimental modifications of the training methods in teaching a second activity. But where learning is relatively slow, as in the white rat, where the age of the subjects may be an important item, and where there may be unknown seasonal differences in the rate of learning, the demands upon the time of the experi- menter become such that he cannot alone supply all the requisite controls. As I have pointed out in a previous paper (Lashley, '17), such experiments should be either carried out by the co- operation of several experimenters, or the subjects should be trained in the simplest possible tasks so that as many as possible may be trained. In the present series of experiments the controls employed seem to cover most of the variable factors save that of individ- ual differences in the animals. They were, briefly, the follow- ing: (1) Genotype. The animals trained were all taken from large litters, one animal from each litter being assigned to each of the differentially treated groups. (2) Sex. The animals chosen from any given litter were all of the same sex so that each ani- mal in any group was controlled by a sibling of the same sex in each of the other groups. (3) Age. Except in the second experiment, all animals were beween sixty-five and seventy days of age at the beginning of training. In the second experi- ment they ranged from seventy-five to eighty-five days. Train- ing was begun with all members of the same litter on the same day. (4) Daily variations. For convenience in injecting the drugs all members of each group were trained at about the same time of day, morning, afternoon, or night, but on succes- sive days the order in which the different groups were taken up was varied so that there was no constant difference in the time PSYCHOBIOLOGY, VOL. I, NO. 2 148 K. S. LASHLEY of day during which the different groups were trained. Hub- bert has shown, further, that the rate of learning is not influ- enced by the time of training. (5) Seasonal variations. No seasonal variation in the learning ability of the rat has been demonstrated, but to control possible variations the experiments were condensed into the shortest possible time and correspond- ing members of all groups were trained at the same time. (6) Differences of weight. Where differences in the weight of sib- lings existed their distribution to the different groups was left to chance. The actual differences in the average weights of the groups were very slight. Individual differences in the rats other than those noted above were beyond control and may have played some part in determining the results obtained. A further argument, how- ever, in favor of the validity of the differences found between the different groups as indices of the effects of the drugs is the internal agreement between the results of the different experi- ments which will be considered later. In the training of animals the personal equation of the ex- perimenter may be influential in deciding the rate of learning. I have sometimes thought that my methods of handling the animals influenced their behavior in the maze, but I have not been able to get any definite evidence of such an influence. Throughout the experiments I made every effort to keep the treatment of all the animals the same and to detect any in- voluntary favoring of one group or another. Furthermore, the most significant check upon the personal equation is the fact that the results obtained were wholly unexpected and by no means agreeable to my preconceived notion of the probable effects of the drugs. THE EFFECT OF THE DRUGS UPON THE GENERAL ACTIVITY OF THE RATS It is generally stated that large doses of strychnine increase the activity of animals, making them restless and increasing the extent of their reactions to stimuli. I have not found any EFFECTS OF STRYCHNINE UPON RATE OF LEARNING 149 concentration of strychnine which notably increases the general activity of the rat. After the minimum lethal dose the ani- mals crouch on the floor of the cage and remain unusually quiet, unless stimulated, until the onset of convulsions. At the same time the irritability to auditory stimuli is very much increased, and the slightest noise will make them leap high in the air. The dose of 0.10 mgm., like the lethal dose, seems to reduce spontaneous activity and to lower the threshold to auditory stimuli.3 After five to ten minutes a coarse tremor, which may be so marked as to resemble a scratching reflex, is visible when the rat raises one foot from the floor. At the same time a marked incoordination of movement appears. The animals are unable to leap accurately or to prevent themselves from falling from the experimenter's hands, and in walking their feet are set down heavily so that the sound may be heard at some dis- tance. When they are frightened their movements become badly incoordinated and even convulsive. The computation of the rate of running in the maze shows that they move slowly. Besides this they frequently show a char- acteristic slowing of movement and hesitation when approach- ing the turns in the maze, so that their progress is a series of quick dashes, alternating with pauses ten seconds or more in length at the turnings of the maze. This behavior may be present even when errors are no longer made. This description applies to the majority of the animals in group A, experiment 1 (table 1). The animals in the second experiment (group E) rarely showed any such activities, and my notes on them reiterate from day to day "no perceptible effect of the drug." Tremor was sometimes noted but only two of the sixteen showed hesitation at the turns of the maze. The animals used in the second experiment were bred during the winter and were older and much larger than those used in the first experiment. Since in the two experiments the dose was not regulated to the weight of the animals it is almost certain that it was relatively smaller for the group E than for 3 Probably the excitability to other stimuli is also increased, but this is more difficult to determine. 150 K. S. LASHLEY the group A and that the physiological effects were, on this account, less. The dose of 0.05 mgm. had a much less pronounced effect on the animals than the larger one. Except for an occasional tremor they seemed in no wise different from those which had not received the drug. Effects of the caffeine, in any quantity, were difficult to de- tect. In general, its administration seemed to be followed by some increase in timidity, exhibited in attempts to escape from the experimenter's hands, to run out of the food compartment of the maze at the experimeter's approach, and to hide in the corners of the maze. This did not seem to be present after the second or third day's practice. No tremor or inaccuracy of movement was noted after injection of caffeine. There is some, though not very certain evidence of the es- tablishment of tolerance to the drugs. After daily injections for two weeks no symptoms were noted in any of the animals following injections of 0.10 or even larger doses of strychnine, and the animals receiving caffeine seemed normal in behavior. There were no deleterious after effects from the use of the drugs. In some cases 0.10 mgm. of stryphnine was given daily for three months. At the end of this time the animals were well nourished and healthy. Practically all were mated, later, and produced large healthy litters. THE INFLUENCE OF THE DRUGS UPON THE AMOUNT OF PRACTICE REQUIRED FOR LEARNING Program of experiments. A preliminary experiment was carried out upon four groups of animals receiving respectively 0.10 mgm. of strychnine, 0.05 mgm. of strychnine, 0.50 mgm. of caffeine, and 0.10 cc. of distilled water. The animals were trained until a record of three successive errorless runs was reached, were then kept without practice for four weeks, and were finally retrained without the drugs to test the retention of the habit. When they were again able to thread the maze without error, their speed and accuracy under the influence EFFECTS OF STRYCHNINE UPON RATE OF LEARNING 151 of the different drugs was tested. The experiment gave fairly clear results for the larger doses of strychnine and caffeine, but as a test of its validity two other experiments, one with strych- nine and one with caffeine, were carried out. The program of the different experiments, with the number of animals trained, is given in table 1. TABLE 1 Synopsis of experiments. Data are given separately on three experiments in each of which the animals were trained simultaneously, with control of age, sex, genotype, etc. NUMBEK OP RATS AGE IN DAYS Experiment 1. June-September, 1915. Learning, re-learning, and efficiency tested A Strychnine, 0.10 mgm. 10 60-70 B Strychnine, 0.05 mgm. 9 60-70 C Caffeine, 0.50 mgm. 9 60-70 D Water, 0.10 cc. 10 60-70 Experiment 2. October-December, 1915. Learning tested E Strychnine, 0.10 mgm. 16 70-90 F Water, 0.10 cc. 16 70-90 Experiment 3. February-April, 1916. Learning tested G Caffeine, 0.50 mgm. 6 60-70 H Caffeine, 1.00 mgm. 6 60-70 I Water, O.lOcc. 6 60-70 • J Strychnine, 0.10 mgm. (after training) 6 60-70 The effects of strychnine on learning. Three groups of ani- mals, A, B, and E, were trained after injection of strychnine and control groups, D and F, receiving only water, were trained at the same tune. The number of trials required by each of the animals in these five groups for reaching the standard of three successive errorless runs is shown in figure 1. The aver- age numbers of trials required for the five groups to meet the two criteria of learning are given in table 2. Groups A, B, and D were trained at the same time and under the same conditions. Among them it appears that the group 152 K. S. LASHLEY receiving the smaller dose of strychnine is not significantly different from 'that receiving water only. Judged by the standard of the first errorless run it is superior, by that of three errorless runs it is inferior. In neither case is the differ- ence enough greater than its probajble error to be significant; there is no indication that the strychnine had any effect upon HgO o.icc St. o.osmy. n! (71 !~F / /I \ \A/ n 7? mr ~] •St. o.i oma d 1 HzOo.)C£. St. o.iomcf. — ; / ; 17 / / //I/ 71 . C7I — n 710 / / / I/I/ / / 10 20 30 40 50 60 70 80 90 100 110 120 Trial FIG. 1. A comparison of the number of trials required for learning by animals trained after injection of strychnine and normal animals (receiving injections of water only) . The squares represent each one animal ; the diagonal lines indicate females. For each group the number of animals making each score is shown on the ordinates, the number of trials required for learning on the abscissae. The dotted lines show the averages of the groups. the learning of this group. The heavier dose of strychnine, on the contrary, seems to have reduced the number of trials required for learning. A comparison of group A with the con- trol group D shows, as a probable result of the administration of the strychnine, a reduction of 36 to 44 per cent in the num- ber of trials required to meet the two criteria of learning. Both EFFECTS OF STRYCHNINE UPON RATE OF LEARNING 153 differences are at least three times their probable errors and therefore almost certainly the result of the differential treat- ment of the two groups. In the second strychnine experiment (groups E and F) the superiority of the drugged over the nor- mal animals is less pronounced, amounting to only 18 per cent for the first errorless run and 11 per cent for three successive errorless runs. The differences also are not significantly greater TABLE 2 A comparison of the number of trials required for learning by rats after injections of strychnine and of water GKOUP DRUG NUMBER OF TRIALS PRE- CEDING FIRST WITHOUT ERROR PER CENT OF WATER CONTROL NUMBER OF TRIALS PRE- CEDING THREE SUCCESSIVE WITHOUT ERROR PER CENT OF WATER CONTROL D A B Water, 0.10 cc. Strychnine, 0.10 mgm. Strychnine, 0.05 mgm. 23.0 =* 2.4 14.8 ± 1.2 20.4 ± 2.3 100.0 64.3 88.7 43.5 ± 3.8 24.5 ± 1.4 48.6 ± 5.4 100.0 56.3 111.7 F Water, 0 . 10 cc. 19.6 ± 1.5 100.0 35.1 ± 1.8* 100.0 E Strychnine, 0 . 10 mgm. 16.1 ± 1.7 82.1 31.0 ± 2.4 88.3 Differences D-A D-B F-E 8.2 =»= 2.7 2.6 ± 3.4 3.5 =*= 2.2 19.0 =*= 4.1 5.1 =±= 6.6 4.1 =*= 3.0 * In a previous discussion of these data (Lashley, '17) a mistake was made in computing the probable error of this average. The error was large but does not significantly affect the conclusions of that paper. than their probable errors, so that from these data alone the second experiment lends little support to the first. As has been brought out in the earlier discussion, however, the drugged animals in this experiment showed in other re- spects a lesser effect of the drug than did those used in the first experiment. The records of the behavior of the animals of group E in the maze give only two rats which showed tremor after injection of strychnine and these two animals required fewer trials for learning than any others in the group. It is 154 K. S. LASHLBY probable, then, that when the greater weight and age of the animals used in this experiment is taken into account, the group corresponds rather to group B of the first experiment, which also showed no tremor resulting from the drug and no super- iority over the controls. The number of trials as a measure of the amount of practice resulting in a given degree of efficiency is probably less depend- able than the number of errors made during practice. The distance traversed during training in excess of the shortest path through the maze expresses the amount of practice fairly accurately and the average excess distance has been computed TABLE 3 A comparison of the distance traversed in excess of the true pathway during learning of the maze by rats after injections of strychnine sulphate and of water GROUP DRUG EXCESS DISTANCE PER CENT OF WATER CONTROL meters D Water, O.lOcc. 266.8 100.0 A Strychnine, 0.10 mgm. 159.8 59.8 A Strychnine, 0.05 mgm. 265.7 99.6 F Water, 0.10 cc. 211.8 100.0 C Strychnine, 0.10 mgm. 146.1 68.9 for each of the groups trained. These averages are given in table 3, and the individual records in figure 2. In the first experiment there is practically no difference between the group receiving 0.05 mgm. of strychnine and the control group. In both experiments the groups receiving the larger dose of strych- nine, 0.10 mgm., traversed a shorter distance in excess of the true pathway than did the control groups. In the first experi- ment this saving amounted to 41.2 per cent and in the second experiment to 31.1 per cent. This difference between the groups in the second experiment is greater than that revealed by the number of trials. It is probable that the excess dis- tance, or number of errors represents more accurately than the EFFECTS OF STRYCHNINE UPON RATE OF LEARNING 155 number of trials the actual amount of practice required for the attainment of a given degree of proficiency and, hence, that the results of the first and second experiments are more nearly in accord than could be determined from the data on the num- ber of trials alone. The results may be judged in yet another way and one which gives a still more reliable basis for comparison than the average St oiomcj. n Jj 7l/n:. Fh 71 n (71 ih fTTTTT 4/i |- AAA V n ! n oa ~7 IJZ 2ZZZ aja too 200 300 400 FIG. 2. A comparison of the total excess distance over the shortest path through the maze traversed during training by rats under the influence of strychnine and by normal animals. Arranged as figure 1, except that the ab- scissae represent meters. number of trials required for learning. The method by which the animals were selected for training makes it possible to com- pare each animal with a sibling of the same sex trained under identical conditions. Table 4 gives the results of such a com- parison of the rats with their individual controls. Of the ani- mals which received 0.05 mgm. of strychnine 60 per cent were superior to their controls. Eighteen out of 25, or 72 per cent 156 K. S. LASHLEY of those receiving the larger dose of strychnine were superior to their controls. This comparison of individuals eliminates the excess weight given to extreme variates by the computation of averagers and forms a basis for an estimation of the regularity of action of the drug. The results are consistent with those obtained from the averages of trials and distances and lend additional support to the view that the superiority of groups A and E over their controls is the result of the action of the drug and not merely a chance variation. The evidence from various sources, when massed together, seems dependable. The animals receiving the smaller dose of strychnine were not affected by the drug. The larger animals showed little effects of the drug either upon their general tonic TABLE 4 The proportion Qf rats receiving strychnine which learned the maze in fewer trials than their individual controls GROUP DRUG NUMBER LEARN- ING IN FEWER TRIALS THAN NUMBER LEARN- ING IN MORE TRIALS THAN PER CENT SUPERIOR TO THEIR CONTROLS THEIR CONTROLS CONTROLS B Strychnine, 0.05 mgm. 6 4 60 A Strychnine, 0.10 mgm. 8 1 88 E Strychnine, 0.10 mgm. 10 6 62 condition or their rate of learning, with the exception of two which showed tremor and learned more rapidly than the other members of the group. The smaller rats, after the large dose of strychnine, showed a fine tremor and learned the maze in considerably less time than was required by the controls. The evidence points to an acceleration of learning resulting from the administration of strychnine, but only when it is given in doses large enough to produce observable alterations in muscular tonus. Time of action of the drug. To test whether the acceleration is due to the immediate effects of the drug or to a general altera- tion in metabolism another series of rats was trained in which the strychnine was administered daily five minutes after train- ing instead of before,. The results of this test are given in table EFFECTS OF STRYCHNINE UPON RATE OF LEARNING 157 5. Apparently the strychnine so given had a retarding effect upon learning, but while the difference between the strychnin- ized animals and their controls is fairly great, the number of animals is so small that no dependence can be placed upon the difference. In this case the dose was graduated to the weight of the animals, 0.10 mgm. of strychnine to each hundred grams weight of the animal so that the immediate effects of the drug resembled those found in the first experiment. The test suggests that the acceleration of learning found for groups A and E is the result of the immediate action of the drug. The influence of caffeine on the rate of learning. The tests of the effects of caffeine were made in the same way as those with strychnine. Three groups of rats were trained after doses TABLE 5 A comparison of the number of trials required for learning by rats after injection of water before training, and of 0.10 mgm. of strychnine after each day's practice GROUP DRUG NUMBER OP TRIALS PRECED- ING FIRST ERRORLESS RUN PER CENT OF WATER CONTROL NUMBER OF TRIALS PRECED- ING THREE SUCCESSIVE WITHOUT ERROR PER CENT OF WATER CONTROL I J Water, 0.10 cc. Strychnine, 0.10 mgm. 14.3 ± 2.2 18.0 ± 1.4 100.0 125.8 30.0 ± 2.8 42.5 =t 4.2 100.0 141.6 of 0.50 mgm. and 1.00 mgm. of the drug and others receiving only water were trained at the same time. The average num- bers of trials required for learning, as estimated from the first errorless run and from three successive errorless runs, are given in table 6. In groups G and H the dose was regulated to the weight of the different animals on the basis of the above doses for an hundred gram rat, so that these groups received relatively more of the drug than did group C. In every case the rats receiving caffeine required a greater number of trials for learning than did their water controls. The differences are quite large, from one and one half to three times as many trials being required by the caffeinized animals as by normal ones. The differences between the groups are at least three times their probable errors and hence significant. 158 K. S. LASHLEY The drug seemed to affect all the animals that received it in the same way. Of the 21 animals trained after injection of caffeine only one required as few trials for learning as did his individual control. The results of the comparison of indi- vidual records are given in table 8. Further, no animal that was given caffeine learned in fewer trials than the average of the control animals, as appears from figure 3. n HtOa.iocc. [7 A \ BE I/I 1 !/ Ca. OLSomq. 171 17 =1 r 71 171 171 r HzO O.IOC.C. \T\ r 7J7I F Co.. auoma. 17 \A | 0L n Ca.ioom I seconds 5 8 seconds 4 5 4.5 6.5 5 1.5 4 3.0 2 5.5 3 4.5 3 3.3 0 4.8 5 1.5 5 2.3 1 3.5 0 3.6 5.0 4.5 1.5 1.5 1.5 0.0 Retention 2 2.4 0 4.0 1.0 Retraining 3 3.5 2 1.9 2 2.4 0 4.0 1.5 1.0 1.0 0 FEMALES GEN. AV. AV. T. No. 91 No. 95 No. 99 Av. E. W. Av.T. W. Av.T. W. Av.T. 3 7 seconds V 7 5 seconds 6 7 seconds 5.3 6.3 5.0 6.2 seconds 5 5 3 0 3.5 4.2 2.9 2.6 5 5 4 3 7 3 2 2 0 1.1 3.4 3.2 2.0 2.0 6.3 2.0 3.2 2.1 5 4 4 4 3 4 3 2 0 2.3 1.8 1.5 1.1 1.9 2.9 1.8 1.7 1.8 5.0 4.6 3.66 2.33 3.33 2.33 1.66 1.33 0.0 5.0 4.6 2.8 2.0 2.6 1.4 1.0 0.8 0.0 2.3 3.1 2.1 1.9 1.9 4.2 1.9 2.4 1.9 Retention 3 2.0 2 2.8 2 2.3 2.33 1.8 2.5 Retraining 4 1 0 2.4 2.0 2.0 3 0 3.6 1.6 3 0 2.0 1.9 3.0 0.33 0.0 2.6 0.4 0.4 0.0 2.4 1.9 2.0 4.0 P. E. 6.8 254 REWARD AND PUNISHMENT IN HABIT FORMATION 255 jects trained with an electric shock of one hundred and fifty units and the curve marked forty-eight hours is based on the average number of errors for subjects trained with hunger of forty-eight hours, and-so-on for the other curves. Figure 4 shows the relative values of different strengths of electric shock, the relative values of different degrees of hunger and the relative values of both electric shock and hunger in the learning process. TABLE 8 Results with hunger of thirty-one hours MALES NO. OF No. 94 No. 96 No. 100 SERIES Av "P w. Av. T. W. Av. T. w. Av. T. AV. .Hi. seconds seconds second* A 6 5 5 5.3 B 4 7 7 6.0 1 5 1.0 4 1.0 5 1.8 4.66 2 4 0.6 5 0.6 4 1.1 4.33 3 6 0.7 5 0.8 3 1.0 4.66 4 2 0.8 1 0.8 3 0.9 2.0 5 3 0.6 0 1.0 4 0.9 2.3 6 3 0.7 4 0.8 2.3 7 2 0.9 4 1.0 2.0 8 4 0.6 2 1.2 2.0 9 4 1.1 2 1.0 2.0 10 4 0.9 0 1.0 1.3 11 4 1.0 1.3 12 2 0.8 0.66 13 3 1.0 1.0 14 0 0.8 Retention 1 3 0.5 7 1.4 4 0.9 4.6 Retraining 1 3 0.7 2 0.8 3 0.8 2.6 2 3 0.6 3 0.8 1 0.7 2.3 3 3 0.5 4 0.9 0 0.7 2.3 4 0 0.6 2 0.8 0.6 5 0 0.9 0.0 256 JOHN D. DODSON TABLE 8— Continued FEMALES NO. OF SERIES No. 93 No. 97 Av.E. GEN. AV. AV. T. W. Av.T. W. Av.T. seconds seconds seconds A 5 5 5.0 5.2 B 3 6 4.5 5.4 1 3 1.0 3 1.0 3.0 4.0 1.1 2 4 0.5 4 0.6 4.0 4.2 0.7 3 1 0.7 4 0.5 2.5- 3.8 0.7 4 3 0.7 5 0.6 4.0 2.8 0.8 5 2 0.7 2 0.6 3.0 2.6 0.7 6 2 0.7 2 0.6 2.0 2.2 0.7 7 1 0.6 1 1.2 1.0 1.6 0.9 8 2 0.7 0 0.6 1.0 1.6 0.7 9 2 0.6 1.0 1.6 0.9 10 1 0.7 0.5 1.0 0.9 11 0 0.7 0.0 0.8 0.8 12 0.4 0.8 13 0.6 1.0 14 0.0 0.8 Retention 1 5 0.8 5 0.8 5.0 4.8 0.8 Retraining 1 0 0.7 6 0.7 3.0 2.8 0.7 2 6 0.6 3.0 2.6 0.6 3 5 1.1 2.5 2.4 0.8 4 3 1.0 1.5 1.0 0.8 5 2 0.7 1.0 0.4 0.8 6 0 1.0 0.0 0.0 1.0 P. E. 6. SUMMARY OF TABLES Table 13 contains the summary results of the experiment proper. It gives (1) the average number of trials and the probable error made by males, by females and the general average and probable error for both males and females in the formation of a perfect habit in case of four strengths of electric shock and four degrees of hunger, and the average time for choice in each TABLE 9 Results with electric shock of seventy-five units NO. OF SERIES MALES AV. E. No. 102 No. 106 No. 108 No. 110 W. Av. T. W. Av. T. W. Av. T. W. Av. T. A B 1 2 3 4 5 1 1 2 A B 1 2 3 4 5 6 7 8 9 1 1 2 3 4 seconds 5 5 7 5 seconds 6 6 seconds 5 6 MC •'!'/••>• 5.75 5.5 4 3.2 4 2.1 1 3.0 3 5.5 0 4.0 5 3 1 0 1.8 4 2.2 2 5.2 1 3.5 0 4.7 4.2 2.7 1.9 4 5 0 4.2 2.7 2.0 4.25 3.5 0.75 0.75 0.0 Retention 5 2.7 0 8.0 0 1.1 2 3.9 1.75 Retraining 4 3.0 0 4.0 0 2.3 1.0 0.0 FEMALES GEN. AV. AV. T. No. 101 No. 103 No. 105 No. 107 No. 115 Av. E. W. Av. T. W. Av. T. S'-C- onds W. 6 7 Av. T. W. Av. T. W. Av. T. 5 6 sec- onds 6 5 sec- onds 2.6 2.5 1.9 5.6 1.6 1.9 2.2 4 8 sec- onds 6 4 sec- onds 5.4 6.0 5.55 5.33 sec- onds 7 4 2 0 2.6 5.0 6.5 1.3 5 4 1 0 2.5 3.0 2.4 5.4 5 5 5 4 3 3 0 6 6 3 5 5 4 5 2 0 2.3 2.0 3.5 2.3 4.0 3.7 3.2 3.9 2.8 4 1 1 0 3.3 2.4 3.3 1.4 5.4 4.0 2.4 1.8 1.6 1.4 1.0 0.4 0.0 4.88 3.77 1.66 1.33 0.88 0.77 0.55 0.22 0.00 2.82 1.1 1.2 1.7 4.0 1.1 2.7 3.9 2.8 Retention 2 5.4 3 4.5 2 1.7 4 2.5 2 3.0 2.6 1.88 3.6 Retraining 6 2 1 0 2.0 2.6 4.3 4.8 6 2 0 2.1 1.8 2.1 1 4 0 1.6 1.5 1.3 3 2 0 2.7 1.4 1.8 1 0 2.6 1.9 4.4 2.2 0.2 0.0 2.4 1.1 1.1 0.0 2.8 2.2 2.4 2.8 P. E. 2.9 257 PSYCHOBIOLOGY, VOL. I, NO. 3 258 JOHN D. DODSON TABLE 10 Results with hunger of twenty-four hours MALES NO. OF SERIES No. 104 No. 112 No. 114 No. 116 AV. E. W. Av. T. W. Av. T. W. Av. T. W. Av. T. seconds seconds seconds seconds A 6 5 7 5 5.75 B 5 4 5 5 4.75 1 5 3.3 4 1.5 7 1.7 4 4.2 5.0 2 5 2.7 4 3.2 5 1.7 4 3.2 4.5 3 6 2.7 4 1.4 4 1.4 6 2.4 5.0 4 6 3.1 6 1.9 1 1.4 5 2.0 4.5 5 4 1.1 5 1.4 3 1.3 3 1.6 3.25 6 4 1.2 4 1.1 4 1.2 3 1.8 3.25 7 4 0.8 2 1.1 3 1.2 4 1.9 2.5 8 4 1.6 2 1.2 3 1.3 2 1.4 2.75 9 2 1.6 5 1.7 6 1.6 5 1.6 4.5 10 1 0.9 1 1.1 1 1.4 1 1.1 1.0 11 1 0.8 1 1.2 0 1.1 5 2.0 1.75 12 0 0.9 3 1.1 3 1.5 1.5 13 4 1.0 3 1.8 1.75 14 3 1.0 4 1.2 1.75 15 5 1.8 4 2.2 2.25 16 1 1.3 2 1.3 0.75 17 0 1.0 2 1.1 0.5 18 2 1.0 0.5 19 0 1.0 0.0 Retention 1 3 1.1 4 1.5. 2 1.6 6 1.1 3.75 Retraining 1 1 1.2 2 1.5 4 1.2 1 0.9 2.0 2 3 1.2 3 1.0 4 1.2 3 0.9 3.25 3 2 1.2 1 0.9 4 1.2 2 0.8 2.25 4 1 1.3 1 x 1.1 2 1.1 3 1.0 1.75 5 0 1.0 2 1.2 2 1.0 2 1.0 1.5 6 0 1.0 1 0.8 0 1.0 0.25 7 0 0.8 0.0 REWARD AND PUNISHMENT IN HABIT FORMATION 259 TABLE 10— Continued NO. OF SERIES FEMALES GEN. AV. AV.T. No. 109 No. Ill No. 113 No. 117 No. 119 Av. E. W Av. T. W. Av. T. W. Av. T W Av T. W Av T sec- onds sec- onds sec- onds sec- ond's sec- onds sec- onds A 5 4 5 5 4 4.6 5.11 B .1 6 5 5 6 6 5.6 5.77 6 4.0 6 2.2 6 5.5 4 4.3 6 3.4 5.6 5.33 3.3 2 4 2.0 7 2.1 3 0.7 5 1.9 6 1.3 5.0 4.77 2.1 3 3 3.5 4 1.7 1 1.2 3 2.3 3 4.0 2.8 4.33 2.2 4 2 6.0 5 3.0 4 1.5 2 2.2 5 1.8 3.6 4.0 2.5 5 5 1.6 4 2.5 1 0.9 1 2.1 3 1.5 2.8 3.0 1.4 6 3 1.6 2 2.9 3 0.8 2 3.0 5 1.4 3.0 3.11 1.7 7 2 1.1 2 1.9 2 1.0 1 1.9 3 1.1 2.0 2.22 1.3 8 2 1.9 6 1.7 1 1.1 3 1.5 4 1.4 3.2 3.0 1.4 9 1 1.2 4 1.7 1 0.7 4 2.0 3 2.0 2.4 3.33 1.4 10 1 1.2 1 0.6 0 0.6 1 0.8 1 1.0 0.8 0.88 1.0 11 0 0.7 2 0.9 0 0.6 3 1.2 1.0 1.33 1.0 12 3 1.3 4 3.0 1.4 1.44 1.5 13 4 1.1 2 1.2 1.2 1.55 1.2 14 1 1.9 0 0.9 0.2 0.88 1.2 15 3 1.7 0.6 1.33 1.9 16 4 1.9 0.8 0.77 1.5 17 2 1.4 0.4 0.66 1.2 18 4 1.1 0.8 0.66 1.2 19 1 1.2 0.2 0.11 1.1 20 1 1 0 1.0 0.0 0.0 1.0 Retention 0 1.0 3 1.2 5 1.0 2 1.4 6 1.0 3.2 3.44 1.2 Retraining 4 1.5 1 0.8 4 2.0 3 1.0 2.4 2.22 1.2 2 5 2.0 1 0.7 5 1.8 3 1.0 2.8 2.77 1.2 3 5 1.8 0 0.7 5 1.8 2 0.9 2.4 2.22 1.3 4 4 1.2 3 1.4 4 0.7 2.2 2.0 1.0 5 3 1.1 3 1.7 2 1.2 1.6 1A4 1.0 6 2 1.0 2 1.1 1 1.0 1.0 o.ee 1.3 7 0 0.8 0 1.0 0 1.0 0.0 O.OC 0.9 P. E. 7.6 TABLE 11 Results with elec.tric shock of sixty units MALE FEMALES NO. OP SERIES No. 118 No. 121 No. 123 GEN. AV. AV. T. Av. E. w. Av. T. W. Av. T. W. Av. T. seconds seconds seconds seconds A 5 3 3 3.0 3.66 B 4 8 6 7.0 6.00 1 4 1.2 4 1.5 3 .3 3.5 3.6 1.5 2 6 1.1 2 2.7 4 .6 3.0 4.0 1.8 3 5 1.7 1 3.1 4 .3 2.5 3.6 1,3- 4 .4 1.7 1 2.8 5 .5 3.0 3.3 2.0 5 4 3.0 2 4.5 4 .6 3.0 3.33 3.0 6 2 4.5 0 3.6 3 .9 1.5 1.66 3.3 7 2 2.3 4 2.1 2.0 2.0 2.2 8 1 4.3 2 1.9 1.0 1.0 3.1 9 0 3.6 1 2.0 0.5 0.33 2.8 10 0 1.8 0.0 0.0 1.8 P. E. 5.7 TABLE 12 Results with hunger of forty-one hours NO. OF SERIES MALES FEMALES GEN. AV. AV. T. No. 124 No. 126 Av. E. No. 125 No. 127 Av. E. W. Av. T. W. Av. T. W. Av. T. W. Av. T. sec- onds sec- onds seconds seconds A 7 2 4.5 6 3 4.5 4.5 B 5 5 5.0 6 5 5.5 5.25 1 5 1.1 4 1.1 4.5 3 1.0 6 1.2 4.5 4.5 1.1 2 5 0.7 5 0.7 5.0 2 0.8 4 0.8 3.0 4.0 0.7 3 4 0.5 5 0.6 4.5 5 0.5 2 0.6 3.5 4.0 0.6 4 4 0.6 3 0.5 3.5 4 0.6 1 0.6 2.5 3.0 0.6 5 2 0.5 6 0.5 4.0 2 0.5 1 0.5 1.5 2.75 0.5 6 2 0.5 3 0.5 2.5 2 0.5 0 0.5 1.0 1.75 0.5 7 4 0.5 2 0.5 3.0 3 0.6 1.5 1.75 0.5 8 0 0.5 1 0.6 0.5 1 0.8 0.5 0.5 0.7 9 1 0.6 0.5 1 0.8 0.5 0.5 0.7 10 0 0.7 0.0 0 0.7 0.0 .0.0 0.7 P. E. 5.6 260 TABLE 13 Summary of results of experiments Original training AVERAGE NUMBER OF TRIALS GEN. AV. P. E. TOTAL AV. T. Males P.E. Females P.E. seconds Reward hours 24 137.5 11.6 122.0 11.6 128.89 7.6 1.5 31 86.6 13.6 85.0 7.1 86.0 6.0 0.8 41 80.0 4.8 70.0 9.7 75.0 5.6 0.66 48 146.7 7.0 125.0 9.2 136.67 5.9 1.2 Punishment in units 60 80.0 9.0 70.0 8.8 73.33 5.7 2.3 75 30.0 2.3 46.0 4.9 38.88 2.9 2.4 115 40.0 4.8 63.33 9.7 54.00 6.8 2.6 150 64.0 6.3 50.0 2.3 57.77 3.9 3.6 Retention tests NUMBER OF INDIVIDUALS MAKING PERFECT RECORD AV. ER. GEN. AV. TOTAL ER. AV. T. Males Females Total Males Females seconds Reward • 24 0 1 1 3.75 3.2 3.44 1.2 31 0 0 0 4.6 5.0 4.8 0.8 48 1 0 1 3.2 3.7 3.44 0.9 Punishment 75 2 0 2 1.75 2.6 1.88 3.6 115 1 0 1 1.0 2.33 1.8 2.5 150 1 1 2 2.4 1.25 1.88 2.5 Retraining AVERAGE NUMBER OF TRIALS GEN. AV. TOTAL AV. T. Males Females seconds Reward 24 50.0 40.0 44.4 1.1 31 30.0 26.0 28.0 0.9 48 20.0 25.0 22.2 0.6 Punishment 75 2.5 20.0 12.2 2.5 115 15.0 13.3 14.0 2.6 150 20.0 2.5 12.2 2.9 261 262 JOHN D. DODSON case ; (2) it gives the number of males and the number of females and the total number of subjects which made a perfect retention test after twenty-one days, the average number of errors made by males and the average made by females and the general average for both males and females in the retention test, and the average time for choice; (3) it gives the average number of trials required by males, the average number of trials by females and the general average for both males and females for the relearning of the habit, and the average number of choices in the relearning process. DIFFERENCES IN LEARNING IN MALES AND FEMALES As may be seen in table 13 there is a difference in the rate of learning with males and females but this difference is neither consistent nor conclusive. In all cases with hunger the average number of trials required for perfecting the habit is less for females than for males but with electric shock the average with seventy- five units and one hundred and fifteen units was less for males. The retention tests show no difference in retention of males and females when trained with different degrees of hunger but a difference in favor of the males four to one when trained with electric shock. The retraining results are even less consistent than in case of training. But having observed rather closely the behavior of all subjects the experimenter would hesitate to say that there is no sex difference. This difference, however, is not necessarily a difference in the capacity of the two sexes to profit by experience but probably a difference in physiological make-up which causes the most favorable condition of learning to vary in a characteristic manner for the two sexes. Females on the whole seemed more anxious for food than the males in all four sets of experiments. This may account for the difference for the rat$ of learning with hunger. INTERPRETATION OF CURVES OF RELATIVE VALUES OF HUNGER If the reader will examine the curve of learning (fig. 4) for different degrees of hunger he will find that there is a constant increase in the rate of habit formation up to forty-one hours of EEWARD AND PUNISHMENT IN HABIT FORMATION 263 AW 130 120 110 100 / \ / V \ / \ \ / \ / 90 80 70 60 50 40 V - — . •••^^ 7 V — ^ ~J \ \ ^ — ^, — ^- — — — — — - \ ^ ^ — \ — • ^^~ oO 20 10 0 60 Units 75 115 ISO 24Hrs, 41 4S FIG. 4. CURVE OF RELATIVE VALUES OF HUNGER AND ELECTRIC SHOCK Abscissa represents the strength of stimuli and the degrees of hunger; ordinate represents the average total number of trials required for the perfecting of the habit. The upper curve represents the relative values of tne four different de- grees of hunger and the lower the relative values of the four strengths of electric shock. 264 JOHN D. DODSON hunger but a sudden decrease from forty-one to forty-eight hours of hunger. The first part of the curve needs no explanation, for the difference in the rate of habit formation when a subject is putting its whole energy in the accomplishment of the act and when it is more or less indifferent towards the performance is fairly well established. Rats which pass from the door between the entrance chamber and nest box in sixty-six hundredths of a second were making about the greatest speed possible for such animals. The scratching reflex or any other distracting in- fluence seldom interfered with their choosing. But the rats which took one and five-tenths seconds had time to scratch occa- sionally or explore the entrance chamber. But the cause of the rapid decrease in the rate of learning for animals of forty-eight hours of hunger is not so apparent. Nor can this decrease be accounted for in terms of poor physical condition of the subjects. At the end of the series of experiments all these animals were in good physical condition though they had somewhat less flesh than their mates which had been trained with electric shock. Had one who did not know their physical condition been observing the manner of choice of this group of rats he would have immediately come to the conclusion that they were not hungry. They neither rushed to get food nor ate eagerly when they had reached it. Their, behavior was very much the same as that of subjects which had gone for only twenty-four hours with- out food. It is true that the carriage of the rat was somewhat different. Dr. J. A. Carlson's careful experiments on hunger in man and dogs account for the behavior of this group of animals in a most satisfactory manner ( 1 ) . Dr. Carlson has demonstrated rather conclusively that the sense of hunger is due to " certain types of contractions in the empty or nearly empty stomach. That these contractions stimulate nerves in the sub-mucosa or muscularis." He demonstrated experimentally that these con- tractions persist almost constantly after the first day of hunger in man : and in young dogs until a short time before death from starvation. In describing his own experience for a period of five days of hunger he says, REWARD AND PUNISHMENT IN HABIT FORMATION 265 The sensation of hunger was almost constant after the first day of starvation — somewhat more severe during the first two or three days. The most severe sensations were at periods of gastric contractions. Appetite ran practically parallel with sensation of hunger. It increased during the first two or three days and diminished on the fourth and fifth days. Instead of an eagerness for food there was an almost indifference to food despite the persistent hunger call of the empty stomach. The writer could not give a more exact description of the behavior of these rats after forty-eight hours of hunger than the above. As the average time of choosing indicates, this group of subjects was slightly more active than the group which was trained with twenty-four hours of hunger but they were not eager for food. Still their behavior was indicative of some disturbing factor: they assumed the hump of a starving animal. They largely abandoned it immediately after eating but assumed it again as the hunger period advanced. This disturbing element was due, no doubt, to the continued contractions of the empty stomach. It seems probable from Dr. Carlson's description of his sensations during a starving period, that these rats had sen- sations of hunger but were not eager for food. As may be seen from next to the last column, table 7, and from the curve of learning representing this column, this disturbing factor was so great that it seemed most probable even as late as the hundredth trial that these animals would never finish the learning process. But from the hundredth trial on the animals were more eager for food. There are two possible explanations for this change: (1) As the subjects grew older they were better able to stand long periods of starvation; (2) the organism doubtless tended to adapt itself to its conditions. Dr. Carlson says, "That the young and growing individual experiences greater hunger than the adult or aged individual is common knowledge." This being the case it is most likely that the period of eagerness for food ends earlier in the young than in the adult individual. When one examines the retraining of the group of animals trained with forty-eight hours of hunger he finds this most interesting fact, that this group retrained more rapidly than the other groups. That it made better average time 266 JOHN D. DODSON in choosing than any other group used throughout the experi- ments is, also, to be noticed. During the retraining these sub- jects showed great eagerness for food. These facts go to sub- stantiate the indications in the latter part of the training series: viz., That rats as they grow older are better able to undergo long starving periods and that the period of eagerness for food is extended; and that the organism may tend to adapt itself to its conditions. While the first of these is sufficient to account for the facts it is very probable that the second entered into the situation. • VALUES OF DIFFERENT STRENGTHS OF ELECTRICAL STIMULI The experimenter determined roughly the minimum stimulus to which the white rat would respond with the control box used in this experiment. Six males and six females of the same age as the subjects for the experiment were used for this purpose. With eighteen units these rats gave no observable signs of response but with twenty-five units all animals tested gave slight movements indicative of feeling the shock. Thus it is fairly certain that with well developed rats the threshold of sensation for subjects of seventy-eight days of age is between eighteen and twenty-five units with the apparatus used in this experiment. This is from ten to fifteen units below the threshold for the experimenter. As may be seen from table 13, subjects trained with an electric shock of sixty units perfected the learning process on an average of 73.33 trials; subjects trained with seventy-five units on an average of 38.89 trials; subjects trained with one hundred and fifteen units on an average of 54 trials; and those trained with one hundred and fifty units perfected the process on an average of 57.77 trials. Thus seventy-five units proved to be the most favorable strengths of stimulus for the learning process in rats of seventy-eight days of age. This is also a very favorable strength of stimulus for animals of fifty-six days of age. Individual differences were very marked with the two weaker strengths of stimuli. All subj ects reacted to a stimulus of seventy- five units very vigorously with the exceptions of numbers 105 and REWARD AND PUNISHMENT IN HABIT FORMATION 267 107. Number 105 would get out of the electric box fairly rap- idly but 107 did not seem much disturbed by the shock and took its time in getting off the electric plates. The experimenter is convinced that the difference in time taken for 107 to perfect the habit and the time taken for 1 10 to perfect the same habit is not primarily a difference in the learning capacities of the two sub- jects but a difference in favorableness of the conditions of learning. With a stimulus of one hundred and fifty units probably number 107 would have even surpassed 110 in the learning process. There is little doubt that a stronger stimulus would have been more favorable to the learning of both 107 and 105. Both of these subjects were slightly lighter than the other rats of this group, but the runt of the entire series was trained with hunger of twenty-four hours and surpassed any of the other animals trained with this degree of hunger. The fairly weak stimuli were more favorable to the observation of individual differences than stronger stimuli, as the stronger stimuli called forth vigorous reaction in the least sensitive subjects. How may we account for the increase in the number of trials required for the habit formation as the strength of stimulus in- creased from about seventy-five units up to one hundred and fifty units? One hundred and fifty units is far below the point of injury to the subject. The only thing to account for this dif- ference that was observable to the experimenter was less nerv- ousness on the part of subjects trained with seventy-five units than subjects trained with the stronger stimuli. Subjects trained with seventy-five units approached the electric box more cautiously and sometimes put their noses into the dark box, then withdrew and entered the light box, while subjects trained with stronger shock would rush into one of the boxes seemingly trying to escape from the situation by running over the grill. Thus it seems that the primary cause for the differences in length of time required for rats to perfect the habit of always choosing the light box when trained with a rather weak stimulus and when trained with stronger stimuli is due to the disturbing factor of excitation. 268 JOHN D. DODSON , COMPARISON WITH EARLIER RESULTS Do the above results agree with the results found in previous experiments on the relative values of different strengths of stimuli in habit formation? Were one to examine the results given by the different experiments without taking into consideration the nature of the subjects used he would conclude that the results are almost diametrically opposed. But when one takes into con- sideration the differences in the natures of the animals it seems that the results point to a common principle. As everyone who is acquainted with the dancing mouse knows, this animal is not especially sensitive to its environment. It dances in the presence of danger with the same indifference to its environment that it does in its cage. It enters an electric box where it may receive a strong shock almost as readily as it does a box where there is no form of punishment. Ordinary changes in its environment affect its behavior very little. On the other hand the rat is extremely sensitive to its environment. The slightest movement in its presence may call forth the native tendency of flight with the suddenness of a simple reflex. 'Subjects trained with electric shock had to be forced through the door between the nest box and the entrance chamber, and here in the presence of the electric boxes the primary motive for choice seemed to be to escape from the situation. If, as all these experiments indicate, there is a point of interference as the strength of stimulus is increased this point should be reached much sooner with the rat than the dancer. And the interference due to excitement will appear much earlier in the series of difficultness with the rat than with the dancer. Thus the most favorable strength of stimulus for habit formation in the rat should be weaker than the most favorable strength of stimulus for the dancer and interference due to excitation should be more noticeable in the rat than in the dancer. SUMMARY OF FACTS TO BE EXPLAINED The above results present certain obvious facts which need interpretation. Why should rats of similar heredity and en- vironment perfect a like habit in so widely different lengths of REWARD AND PUNISHMENT IN HABIT FORMATION 269 time when trained with different degrees of hunger? Why should subjects of similar heredity and environment perfect a like habit in so widely different lengths of time when trained with different strengths of electric shock? Why should subjects with the same heredity and similar environment perfect a like habit in so widely different lengths of time when trained with hunger and when trained with electric shock as motives? Subjects trained with the most favorable condition of hunger perfected the habit in very nearly one-half the number of trials that subjects trained with the least favorable conditions of hunger did. Subjects trained with a shock of seventy-five units perfected the habit in slightly more than one-half the number of trials that subjects trained with one hundred and fifty units did. Subjects trained with the most favorable electric shock perfected the habit in about thirty-nine trials while it took subjects trained with the most favorable condition of hunger seventy-five trials. SOME LAWS OF LEARNING WHICH HAVE BEEN SUGGESTED Thorndike in his Educational Psychology gives three primary laws of learning (8). (1) Exercise. To the situation, " a modifiable connection being made by him between a situation S and a response R," man responds origi- nally, other things being equal, by an increase in the strength of that con- nection. To a situation, "a modifiable connection not being made by him between a situation S and a response R, during a length of time T," man responds originally, other things being equal, by a decrease in the strength of that connection. Corollary to the first part of the law of exercise: the degree of strengthening of a connection will depend upon the vigor and duration as well as the frequency of its making. (2) Effect. To the situation, ' 'a modifiable connection being made by him between an S and an R and being accompanied or followed by a satisfying state of affairs" man responds, other things being equal, by an increase in the strength of that connection. To a connection similar, save that an annoying state of affairs goes with or follows it, man responds, other things being equal, by a decrease in the strength of the connection. 270 JOHN D. DODSON (3) Readiness. By original nature a certain situation starts a be- havior series: this involves not only actual conduction along certain neurones and across certain synapses, but also the readiness of others to conduct. Watson criticizes the conception that pleasure tends to stamp- in desirable acts and pain to stamp-out the undesirable acts, and offers two principles, " recency" and " frequency" as possible explanation for the mechanical process in learning. He says (10) , It is our aim to combat the idea that pleasure or pain has anything- to do with habit formation or that harmfulness or harmlessness has any thing more to do with the situation. Again, We may confess at once that we have no new principles to offer in solving the problems involved in learning, but we hope that by stating our problem carefully and by clearing away the misconceptions referred to, we shall be able to show in a convincing way that the mechanical principles with which we are already familiar and which can experi- mentally be shown to act in the way we maintain are sufficient to yield the solution of those problems. We shall call these principles (1) fre- quency and (2) recency. Holmes says (5) : Profiting by experience in an animal of primitive type of intelligence we conceive, then", to take place as follows : The creature is endowed with the capacity for responding to beneficial stimuli by aggressive, out- stretching movements, and to injurious stimuli by movements of with- drawal, retreat and avoidance. All these are matters of pure instinct. Given the power of forming associations between responses, the animal acquires new habits of /action by repeating those responses which arouse instinctive acts of a congruous, and discontinuing those responses which arouse instinctive acts of an incongruous kind. Peterson has recently suggested " completeness of response" as a fundamental principle in the explanation of the learning process. He says (7). REWARD AND PUNISHMENT IN HABIT FORMATION 271 That the animal in the performance of an act is constantly in a state of muscular tension due to mutually reinforcing, mutually inhibiting tendencies and that these tensions are released only when the proper reactions have been made and the desired act been performed. Haggerty gives a physiological interpretation of the learning process in the following law (3) : A physiological state is not self-contained but tends to radiate to other physiological states both those which form with it a series of states like a habit chain and also to other physiological states which have never formed a series. Frequency. That the frequency of repetition of a desired act is of value in perfecting the habit may hardly be successfully denied, but that it is a dynamic factor may be doubted. Its importance varies with the nature of the habit and the motive used for promoting the learning process. That it fails to offer anything like a complete explanation for habit formation is shown by the following facts: (1) It took rats trained with a shock of seventy-five units an average frequency of 24 right choices to 14 wrong while it took rats trained with twenty-four hours of hunger a frequency of 85 right choices to 42 wrong to per- fect the same habit. (2) It required subjects trained with a shock of sixty units an average frequency of 51 right choices to 23 wrong choices while it took subjects trained with seventy-five units an average frequency of 24 right to 14 wrong choices to perfect the same habit. (3) It required subjects trained with forty-one hours of hunger an average frequency of 45 right choices to 23 wrong while it required those subjects trained with twenty-four hours of hunger an average frequency of 85 right to 42 wrong choices to perfect the habit. Thus we see that the proportion of right to wrong choices is greater in all cases where it took the subject a greater number of trials to complete the learn- ing process. Recency. The importance of recency as a factor in the for- mation of a habit, varies like that of frequency, with the nature of the habit. But it does not help to explain the differences in the results in this experiment. The recency in all series of ex- 272 JOHN D. DODSON periments was the same and as to the recency between trials there was practically no difference. Vigor. The importance of the vigor with which an animal performs an act has been underestimated by some students of behavior. The more nearly the^whole active organism is directed towards the accomplishment of the act the more rapidly will the act be perfected. The subjects which chose most quickly and made the greatest effort to reach the food learned in about one half the time that it took for those subjects which did not seem anxious to get to the food. This is evidently an important factor in accounting for the difference in the time taken for animals trained with twenty-four hours hunger and animals trained with forty-one hours hunger to perfect the same habit. It also has its bearing in the interpretation of the difference in the average num- ber of trials taken by animals trained with sixty units and those trained with seventy-five units. The former stimulus was too weak to keep the subjects up to their greatest efficiency. The directing of all energy in a single channel means efficiency in acquiring any habit. Animals trained with the more favorable conditions were not often interfered with by the scratch reflex and like inhibitory processes. Satisfyingness and annoyingness. Thorndike tells us "that improvement is the addition or subtraction of bonds or the addi- tion or subtraction of satisfyingness and annoyingness.7' But how the satisfaction of eating of food after an animal has perform- ed an act can lap back and in some way stamp in the act is not very easy to understand. Likewise it would be a hazardous science that would say that the satisfaction an animal gets from JM-j eating after forty-one hours of hunger is more effective in stamp- ing in a desired act than eating after forty-eight hours of hunger, or that the annoyingness of an electric shock of one hundred and fifty units is less effective in stamping out an undesirable act than that of seventy-five units. The above illustrations are sufficient to show that the principles of satisfyingness and annoyingness are of no significance in an explanation of the results of this experiment. Other principles. Though the principles of congruity and in- congruity, completeness of response and the law of irradiation REWARD AND PUNISHMENT IN HABIT FORMATION 273 are suggestive and are doubtless of importance in the explanation of the learning process, the writer is unable to give a satis- factory interpretation of this experiment on the basis of one or all of these principles. The seeming simplicity of the mechanical principles of a simple type of learning grows into complexity when one attempts to account for habit formation under different conditions. The writer does not hope to give a set of principles which will explain the mechanics of all types of habit formation, nor does he care to add another guess as to the physiological changes which take place, but desires some kind of interpretation for the results obtained in the above experiments. Any part or even all of the principles which have been mentioned do not give a satisfactory explanation of the facts. There are at least two other factors which seem of importance. Native tendencies. Something of the importance of the native tendencies of the organism in the learning process has been recognized in a very general way by a number of students of behavior, but the writer here refers to specific tendencies. The specific native tendency with which the learning process is linked seems of vital importance in determining the length of time for perfecting any habit. Subjects trained with the most favorable strength of electric shock perfected the same habit in slightly over half the time that subjects trained with the most favorable degree of hunger. Most probably this is due not to a difference in the values of pleasant and unpleasant stimuli but to the fact that in one case the process is tied-up with the tendency of flight and in the other it is tied-up with the food seeking tendency. The tendency of flight is a very strong tendency and takes pre- dominance over the food seeking tendency when both are stim- ulated. The rat which is seeking food, on the approach of an enemy takes to flight and ceases the search of food for the time being. Were these comparisons between sex and flight tend- encies the results might be very different. The stronger the pull or drive of the tendency with which an act is linked-up the less likely is the individual to be attracted from the performance of the act. The subject which is very hungiy is seldom inter- 274 JOHN D. DODSON fered with by the scratch reflex while the subject which is only slightly hungry is frequently interfered with by it; but the sub- ject trained with electric shock is less frequently interfered with than is the very hungry subject. This means that the native tendency of flight holds the subject up to a more efficient performance of the act than does the food seeking tendency. Disintegration. The physiological process which takes place in the nervous system in the learning process is not definitely known but whatever it is it may be interfered with, or there may be a tearing down process taking place along with the building up process. It is in this rather broad sense that the writer uses the term disintegration. The factors which cause disintegration may vary from a minimum to a point where the tearing down process is equal to .the building up process. Subjects trained with a shock of one hundred and fifty units learned less rapidly than subjects trained with a shock of seventy-five units, doubtless because of disintegration due to too great excitement of the situ- ation. That is, the disintegrating and integrating processes were more nearly equal in the former case than in the latter. Subjects trained with twenty-four hours hunger would not in- frequently be headed directly towards the light box when the scratch reflex would predominate over the food seeking tendency and the animal would stop and scratch and then go in the direc- tion which it might be headed regardless of right or wrong. Subjects trained with forty-eight hours of hunger were more active than subjects trained with twenty-four hours of hunger but learned less rapidly, probably because of the interfering effect of the strong contractions of the stomach. The physiological disturbance during the first 80 trials was so great that it seemed that the animals would never finish with forty-eight hours of hunger, but about this time the subjects seemed to adapt them- selves to the condition. RELATION OF RATE OF LEARNING TO RETENTION Our results show no marked difference in the three groups tested for retention in the relative values of different degrees of hunger for the retention of the habit. Of the nine subjects trained with forty-eight hours hunger one made a perfect retention test; of REWARD AND PUNISHMENT IN HABIT FORMATION 275 those trained with twenty-four hours hunger one made a perfect retention test ; and of those trained with thirty-one hours hunger there was no perfect retention test. Likewise groups trained with electric shock showed no special difference in retention. Two subjects for each of the strengths of electric shock, one hundred and fifty and seventy-five units, made perfect records for reten- tion and one subject for the strength of one hundred and fifteen ' units of electric shock made a perfect retention test. These \ facts indicate that subjects trained with electrical stimuli retained 4 better than subjects trained with hunger. Thus it seems that the time required for the formation of a habit has little to do with the retention of the habit but the strength of the native tendency with which the habit is linked is of some importance. RETRAINING The retraining series are on the whole in harmony with the train- ing series. Animals trained with twenty-four hours of hunger re- learned the habit on an average in 44.4 trials; those trained with thirty-one hours retrained on an average in 28 trials; subjects trained with forty-eight hours retrained on an average in 22.2 trials; subjects trained with seventy-five units of shock retrained on an average in 12.2 trials; those trained with one-hundred and fifteen units retrained on an average in 14 trials; and subjects trained with one hundred and fifty units retrained on an average in 12.2 trials. The fact that subjects trained with forty-eight hours of hunger relearned more rapidly than any of the other groups is due to the increased age and adaptation of the organism to long periods of starvation . Just as one should expect, there is no significant difference in the time taken for the three groups trained with electric shock to relearn the process, for the retraining time was too short. CONCLUSIONS 1. In case of hunger the rapidity of learning increases as the hunger increases but the maximum hunger is reached in rats of seventy-eight days of age some where between forty-one and forty-eight hours. After the maximum hunger is reached there is a rapid decrease in the rate of learning as the period of starva- tion is increased. 276 JOHN D. DODSON 2. With a discrimination problem of the difficultness used in this experiment the rate of learning increases as the strength of stimulus increases from the threshold up to about seventy-five units, which is a comparatively weak stimulus, and gradually decreases as the strength of stimulus is increased beyond this point. 3. The electric shock is more favorable to the learning process in the white rat than is hunger in case of a simple discrimination problem. 4. The important factors in accounting for the differences in rate of learning in this experiment are: vigor of performance, frequency of repetition, native tendency with which the process is linked, and the disintegration due to interfering tendencies. 5. The time taken for forming a habit has but little to do with its retention but the tendency with which it is linked may be of considerable importance. REFERENCES (1) CARLSON, J. A.: The control of hunger in health and disease. 1916, Univ. Chicago Press. (2) COLE, L. W. The relation of strength of stimulus to rate of learning in the chick. Journal of Animal Behavior, 1911, i, 111-124. (3) HAGGERTY, M. E. The laws of learning. Psychol. Rev., 1912, xx, 411-422. (4) HOGUE, MILDRED AND STOCKING, RUTH. A note on the relative values of punishment and reward as motives. Journal of Animal Behavior, 1912, ii, 43-50. (5) HOLMES, S. J. Studies in animal behavior. Boston, 1916, R. G. Badger. (6) MARTIN, E. G. A quantitative study of faradic stimulation. Amer. Jour, of Physiol., xxii. (7) PETERSON, J. Completeness 6f response as an explanatory principle in learning. Psychological Rev., 1915, xxiii. (8) THORNDIKE, E. L. Educational psychology, briefer course. Teachers College, N. Y. (9) WASHBURN, MARGARET F. The animal mind. 2d Edit., N. Y. 1917. Macrnillan Co. (10) WATSON, J. B. Behavior — an introduction to comparative psychology. New York, 1914, Henry Holt & Co. (11) YERKES, R. M. The dancing mouse. New York, 1907, Henry Holt & Co. (12) YERKES, R. M., AND DODSON, J. D The relation of strength of stimulus to rapidity of habit formation. The Journal of Comparative Neu- rology and Psychology, 1908, xviii, 457-491. A CLASSIFICATION OF GROUPS1 CARL W. BOCK Introductory and historical 277 I. The general series: a-G, b-G, c-G w-G, x-G, y-G, z-G 281 1. Groups as stable activities 281 2. The coefficients a, 6, c, etc 285 3. The coefficient w 295 4. The coefficient x 297 5. The coefficient a 298 6. The coefficient y 300 7. The coefficient z 312 8. Other analytical methods 313 II. Conclusions 318 INTRODUCTION The present study is based upon the work of Swindle2 who described and defined certain very simple and characteristic movement complexes to which he gave the name groups. As a provisional definition of a group he says as follows: "The simplest conceivable instinctive movement (Bewegungsinstinkt) is the result of the capability of an organism to react so many times with a given member of the body (Koerperglied) until a definite number of movements have been made" — that is to say, until a group of similar movements have been beat in a particular tempo, amplitude, and direction. 1 Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Graduate School of the Ohio State University under the title, "The Association of Voluntary Movements." The writer wishes to take the opportunity of expressing his obligations to his observers, namely, his wife, Dr. Karl Dallenbach of Cornell University, Dr. Geo. F. Arps of Ohio State University, and to Miss Ruth Miller, graduate student of the latter institution; to Dr. A. P. Weiss who read the manuscript and whose constant help and guidance made this work possible; more particularly, how- ever, to Dr. G. F. Arps whose liberality of thought and action and constant interest and attention may not be overestimated. 2 Swindle, P. F. Zeitschrif t fiir Psychologic u. Physiologic der Smnesorgane, 1915, Bd. 49. 277 PSYCHOBIOLOGY, VOL. I, NO. 4 278 CARL W. BOCK In amplification of the above, Swindle's observations on certain animals may well be cited. He observed that certain owls (Glaucidium whitelei Lws.) made characteristic periodic to and fro movements with the tail. More concretely stated, the owls moved the tail to and fro a certain number of times and then stopped; after a pause of variable duration they made similar movements and rested, and so on throughout longer periods of observation. Other animals behaved similarly, some using this, others that member of the body, or limb, for the performance of these group movements. By definition any series of periodic movements of the same body members is called a group, and any single to and fro move- ment, a beat or an element of a group. Moreover, if the group consists of 5 elements or 8 elements, or 23 elements, etc., it is respectively known as the 5-group, 10-group, 23-group, etc., to distinguish it from groups which contain any other number of elements. The term group had however another justification than that of merely defining or describing a series of similar periodic move- ments. It had a functional or behavioral justification which was based upon the fact that certain groups recurred very fre- quently in the same animal, and indeed far more frequently than would be expected, were the frequency of recurrence only a matter of chance. In consequence of the apparent stability of these activities the term group carries with it the implication- that groups are entities of some kind, — functional units of perhaps the same order as the more traditional functional units commonly called reflexes, instincts, and habits. Indeed, Swindle identi- fies groups with instincts and habits accordingly as they are inborn or acquired. The observers whose data constitute the basis of the present study were human beings. They were instructed by the writer to beat periodically on the button of a tambour as long as they de- sired to do so : to rest as long as they had previously beat : and finally, so to alternate beating and resting until instructed to cease. No limitations were placed on the periodicity of their A CLASSIFICATION OF GROUPS 279 beating nor upon the duration except that the periodicity of any given group was to remain constant.3 This procedure gave as its data (1) an alternate series of groups and rests, and (2) a series of discrete but contiguous dura- tions, i.e., durations of groups, and durations of rest periods. Generalized the data would appear as follows: (1) Group, Rest, Group, Rest, G, R, G, R, G, R, etc. (2) G-Dur., R-Dur., G-Dur., R-Dur., G-D, R-D, G-D, R-D, etc. Quantitatively expressed, the above two series may be ex- pressed : (3) a-G, R, b-G, R, c-G, R, w-G,R, x-G, R, y-G, R, z-G (4) a-sec., b-sec., c-sec., d-sec., e-sec., f-sec., g-sec., etc. where a, 6, c, d, etc., in (3) stand for simple numbers, each de- scribing and defining the group to which it has reference from the point of view of the number of elements contained in it; and where the corresponding coefficients in (4) qualify the dura- tions of both groups and rests in seconds. Inasmuch as the rest periods are not otherwise quantitatively described in series (1) and (3), they may be entirely disregarded in these equations or series, and the latter may be written as follows: (5) a-G, b-G, c-G, d-G, w-G, x-G, y-G, z-G The above series only describes groups as to the number of elements or beats contained in them; it says nothing regarding the other characteristic properties of groups, such as their peri- odicities, or their durations; it disregards entirely the rests and their durations, but it is the fundamental series of this paper, which seeks to answer the following questions: (a) What groups are beat by any observer? (b) Is a general classification possible? Our observers were seated in one of the standard tablet chairs, the right fore-arm resting on the tablet, the index finger crooked 3 It goes without saying that observers were not permitted to count either silently or aloud while beating. 280 CARL W. BOCK and resting upon the button of a cardio-tambour. The said tam- bour was fastened to the extreme front of the tablet in a con- venient place for tapping, and was connected to a more compli- cated recording tambour by a rubber tube, as ordinarily in such cases. Records of the excursions of the pointer of the re- cording tambour were made upon smoked paper which was stretched between two horizontal drums. One of the latter TABLE 1 Observer R 1.) 44, 24, 44, 47, 22, 46, 52, 48, 68, 32, 104, 47, 7, 45, 10. 2.) 36, 54, 27, 36, 96, 115, 174, 130, 49, 12, 17, 16, 24, 52, 45, 21, 18, 130. 3.) 23, 28, 42, 45, 46, 46, 46, 22, — , 19, 44, 68, 44, 44, 70, 143, 190, 10, 10. 4.) 22, 115, 22, 22, 45, 24, 32, 8, 93, 156, 106, 99, 47, 31, 79. 5.) 77, 46, 22, 76, 393, 18, 22, 22, 46, 156, 7, 7, 7, 19, 22, 46, 68, 46, 46, 8. 6.) 20, 43, 45, 22, 94, 6, 7, 7, 7, 7, 6, 7, 50, 3, 3, 3, 3, 45, 45, 22, 46, 46,46, 46, 46. 7.) 44, 10, 49, 68, 50, 48, 48, 46, 46, 46, 46, 68, 71, 22, 22, 46, 46. 8.) 22, 75, 70, 46, 46, 70, 115, 96, 108, 71, 71, 22, 46, 46, 46, 46, 46, 46, 46, 46. 9.) .46, 70, 46, 117, 22, 70, 70, 333, 22, 22, 22, 22, 22, 70. 10.) 29, 31, 31, 15, 15, 22, 22, 22, 21, 22, 47, 46, 70, 22, 22, 22, 46, — , — , 75, 22, 47, 60 11.) 46, 46, 47, 45, 46, 8, 8, 46, 22, 22, 22, 22, 22, 22, 22, 22, 96, 94, 46, 98, 47, 46, 46. 12.) 48, 136, 118, 94, 100, 47, 174, 94, 70, 46, 46, 94, 94. 13.) 22, 22, 22, 23, 22, 46, 46, 46, 24, 22, 46, 45, 142, 146, 95, 60, 66, 46, 70, 95. 14.) 46, 94, 94, 68, 95, 46, 46, 47, 416, 48, 94. 15.) 22, 22, 46, 94, 46, 68, 94, 94, 94, 48, 46, 46, 49, 94, 70, 32, 44. 16.) 43, 3, 3, 3, 3, 3, 3, 3, 3, 13, 12, 12, 10, 10, 14, 20, 22, 22, 44, 74, 46, 46. 17.) 47, 47, 46, 133, 94, 79, 94, 118, 95, 94, 70, 46, 46. formed part of an electrically driven kymograph, the other was simple and revolved upon a fixed axis which was supported on two stands by clamps. The two drums were placed on sepa- rate tables which were some 15 to 20 feet apart. The recording tambour as also the time signal were fixed in front of one of the drums to a platform movable in the direction of the axes of the drums. Consequently it was possible by simply shifting the plat- form and thereby the recording apparatus to secure from 4 to 6 complete turns of a continuous series of groups, which, with a distance of 15 to 20 feet between the tables, amounted to about A CLASSIFICATION OF GROUPS 281 100 feet of groups and rests, requiring from thirty minutes to as much as five hours to secure, depending upon the speed of the drums. Observers were given no other instructions than those given above. In other respects, and in so far as it was compatible with beating groups, they were left to their own devices, free to perform such other activities as studying their lessons, talking, singing, reading, or thinking, and free to have whatever ideas occurred to or in them, of whose nature no record was made or desired. i. THE GENERAL SERIES: a~G, b~G, c~G, w~G, x~G, y-G,z~G L Groups as stable activities. In order to show that groups are stable or recurrent activities it is only necessary to show that the coefficients a, 6, c, d, etc. are identical, or, that some of them are, and in numbers sufficient to exclude chance identities. In its ideal, i.e. extreme form, the above general series would reduce itself to one of the following forms: f (6) a-G, a-G, a-G, a-G, a-G, a-G, etc. \ (7) c-G, c-G, c-G, c-G, c-G, c-G, etc. [(8) etc. This would imply, what rarely happens except under extremely well controlled conditions, and certainly never under the condi- tions under which our observers worked, that the same group was always beat. Table 1 contains a record of 17 series of groups beat on approxi- mately seventeen succeeding days by observer R. A glance at this table will show that, considering any given series, there are groups which recur several times within that series; and consid- ering the 17 series, that the above groups likewise occur and recur in practically all series. These facts are more conven- iently shown in table 2, which constitutes a table of the fre- quencies of the several different groups beat by this observer for each of the seventeen days of experimentation and which give also (lower row) the total frequencies of all groups for the total of the seventeen days. 282 CARL W. BOCK 91* — _ «c — — c?c — 06\ — — til — — CM 9fl — — CM «H CH — ZH 9CI — Cfl — — 0£! *sl « IZ" oO H CM *o . — — •30 £fr — Zf — — 9* oa fvj Z£ — — r> l< — •M n « — — . 11 — — 91 — — n — — — — v}~ n •— fM n — — »o •*r CM IO «Ni vO t— Os •0 CM CM V n — — CM oz — — fvl 6i — — CM «i — — CM ii — — 91 — Cl CM CM ¥\ — — Cl — tl — cv •*> 01 — •xi CM >O 9 — — CM m I — *•"> «O o\ 9 — — C •f OO 21 — esi *J ^r *o vO r- OO o\ 0 I± ^JL >Q s: to ^£ t— ^ A CLASSIFICATION OF GROUPS 283 Considering the latter first, it is evident that there are several groups which are beat with frequencies above the average fre- quency of all other groups, because: (a) the total number of groups beat = 303 (b) the total number of different groups beat = 69 (c) the average frequency = 4.4 The numbers 47, 62, 13, and 18 which represent the frequencies of the four groups 22, 46, 70, and 94 give the following percent- age frequencies: 15.5 per cent, 22.1 per cent, 4.2 per cent and 6 per cent. These percentages are, with one exception, quite above the average frequency of all groups, and still greater than 2.4 per cent which represents the average frequency of all groups, the above four excepted. Clearly there are certain groups that are beat more often than others and more often than they would be beat, were their distribution a matter of chance. Graphi- cally this is shown on graph 1 from which it appears that the groups 3, 7, 22, 46, 70 and 94 have maxima that are perceptibly greater than the average for all groups. It will have to be asked and answered how often a group must be beat or must recur in any given series before it can be said that its frequency of recurrence is significant. Clearly, but only generally, this is the case when its frequency rises appreciably over the average frequency of all other groups. In this it is assumed that on a chance basis all groups are equally likely to occur. If certain groups occur more frequently and particularly when these same groups recur on different days of experimenta- tion, or indeed in different observers, then such recurrence must have some significance. On this principle the groups 22, 46, 70 and 94 are significant, whatever the nature of the significance, and speaking generally again, the 303 groups of observer R may presently be divided into the two classes, those that are signifi- cant and those that are insignificant by the frequency criterion. Obviously no hard and fast line can be drawn between the members of the two classes of groups, and consequently the fre- quency principle cannot be considered as the sole or as an abso- 284 CARL W. BOCK lute measure. Moreover this criterion would exclude from con- sideration those groups whose frequencies fall below that of the average of all groups and which include groups that are very sig- nificant on the bases of other criteria. For instance the group 416 occurs but a single time; yet it can be predicted that if the observer beats a group in the 400 's, it ought to be precisely the 416 and no other 400 group. The frequency of this group is given by the fraction 1/303. As a matter of fact the terms significant and insignificant have only a present convenience and no real justification at all except relatively speaking. Organisms do not behave on the chance basis; all their activities are significant, only some are for present purposes more so than others. There are many factors which tend to maximize the difficulties of a rigorous application of the frequency criterion, of which but one can properly be dis- cussed at this time. Assuming that the group 416 (vide table 2) is significant for present purposes and of the same order of sig- nificance as the 22 group or the 46 group of the same observer, it might be asked how it happens that the latter groups are beat so frequently and the former only once? The answer is simple enough. The frequency criterion can only apply where or when all groups have an equal chance of recurrence, which is not in accord with either fact or theory in the present instance. In the first place, the duration of the 416 group is relatively much greater than the durations of the groups 22 and 46. In the second, the 416 is a much longer group. Theoretically it ought to be expected that animals or humans that have a relatively large repertoire of groups, some short, others of medium size, and still others that are long, will beat the medium sized groups most frequently. This is of course on the assumption that the organism in question has received no special training. A long group, or one whose duration is long, has not the same chance of recurrence, just because it takes longer for its performance, and consequently the same measure does not apply for these groups as apply for the shorter groups, and particularly for the medium sized groups. A CLASSIFICATION OF GROUPS 285 In the light of the above the general series, (5) a-G, b-G, c-G, d-G, e-G, w-G, x-G, y-G, z-G can be transformed into the series, (6) a-G, b-G, a-G, b-G, a-G, w-G, x-G, y-G, z-G in which the coefficients a, 6, etc., represent groups of the order of the 22, 46, 70, and 94 of observer R, and the coefficients w, x, y, and z represent groups that are for the present insignifi- cant on the basis of the frequency criterion. The stability of the groups 22, 46, 70, and 94 is established by virtue of the possi- bility of the above transformation. 2. The coefficients a, 6, c, etc. The coefficients a, b, c, d, etc., of general series (5) have been shown to be recurrent or stable in the previous section. Substituting for a, &, etc., in (6) the particular values as found for observer R by the frequency criterion, a = 22; b = 46; c = 70; d = 94. The facts are these: a certain individual R beats 17 series of groups on seventeen consecutive days, and of the 303 groups beat, amongst which there are 69 different groups, he beats four groups, the 22, 46, 70, and 94 relatively very frequently. The question to be asked is this; have the abnormally high frequencies of the four groups 22, 46, 70, and 94 four causes or less than four causes? A glance at graph 1 will reveal a certain regularity or even periodicity in the maxima for the groups 22, 46, 70, and 94. It is true that the maxima for these several groups do not have even approximately the same values, which is explicable by the discussion of a previous section and also by certain facts about to be discussed. To the apparent periodicity of the above four maxima, the following other singularities may be added: consid- ered day by day (table 2) these four groups do not have their maxima on the same days. Consider the 22 group ; for whatever reason, observer R beats the 22 group with increasing frequency 286 CARL W. BOCK as the experiment progresses up to and including the eleventh day. On the twelfth day he does not beat the 22 group at all, although he beats it again on the thirteenth day and on suc- ceeding days. Apparently the factor which makes for the pro- duction of the 22 group, and with progressively increasing fre- quency, ceases for some reason to be operative on the twelfth day. The factors that make for the production of a group in general are either environmental or internal, but the former cannot enter into the discussion because the general conditions 46. GRAPH I. under which the observer worked were practically identical throughout the seventeen days of experimentation. Moreover the observer's general health remained the same, and indeed there is no way in which the observer's acts on the twelfth day are explicable except on the assumption that some very radical change had taken place within his organization, which is con- trary to fact. Consequently it must be assumed that ob- server did beat the 22 group, appearances to the contrary not- withstanding, on the twelfth day. On this day observer beat A CLASSIFICATION OF GROUPS 287 the 94 group four times, one of the maxima for this group which is not beat on the thirteenth day. Similarly, the maxima for the 70 group occurs one day later, i.e., on a different day from that of the 46 group on which day the frequency of the latter group approaches a minimum. On the tenth day a maximum of 7 obtains for the 22 group while the 70 group is beat but once and the 46 group twice. On the fifteenth when the 94 group has a maximum, all other groups practically have minimums. Apparently an inverse correlation of some kind obtains be- tween the frequency maxima of these four groups, and in this case some kind of a relationship must exist among the four groups. Consider finally the four numbers 22, 46, 70, and 94: the following numerical relations obtain: 22 = 22 46 = 22 + 22 + 2 70 = 46 + 22 H- 2 94 = 70 + 22 + 2 These relations are the counterpart of the apparent periodicities previously referred to on the graph. If the above numerical relations correspond to actual facts it would mean that 22-G = 22-G 46-G = 22-G + 22-G + 2-G 70-G = 46-G + 22-G + 2-G 94-G = 70-G + 22-G + 2-G and it would explain the periodicity of the maxima that obtain on graph 1 for the frequencies of the above four groups as well as the otherwise inexplicable fact that an organism can beat a certain group with progressively increasing frequency, or perform an activity of any kind in the same manner and suddenly for no apparent reason cease to perform it at all. The explanation is that the observer does not cease to beat the 22 group — he merely beats it several times in succession, hiding thereby its identity as such in the guise of an apparently new or different group. The above equations explain also why the frequencies of such groups as the 70 and 94 is relatively so small. They do not represent frequencies of different groups but rather the fre- 288 CARL W. BOCK quency of the 22 group, and since the 70 and 94 groups each contain several 22 groups, they must recur less frequently, other things being equal. Fortunately the proofs for the above equations were found in the records of the observer, but for reasons to be explained, only by chance. Before the records in question (containing the proofs) were made, during the last four days of experimentation, the facts as stated above were well known to the writer, who searched in vain among those records that had been made up to this time for some. objective evidence which might support the conclu- sions to which considerations of frequency, inverse correlations, and numerical relations pointed. At present, after three years' experience the writer prefers not to refer to records for proof except under very particular circumstances because he knows that a record cannot normally contain the facts of analysis un- less accidentally or incidentally, — not because the facts of numer- ical analyses do not correspond to behavioral actualities, but be- cause they do not necessarily correspond as would be expected by ordinary habits of thought. The writer will call attention to the facts and reasons of the above at its proper time. Consider figure 5 (first part) a fac-simile of one of the many 46 groups. Figure 5 shows 46 beats of varying amplitudes given in a prac- tically constant tempo. Numerical and other reasons point to the conclusion that the 46 group above consists of two 22 groups and a 2 group. Examination of the figure will show that there exist in this figure absolutely no evidence for the assump- tion. It cannot, except very arbitrarily be said, that here (on say the 22d beat) the first 22 group ends, and there (on say the 23d beat) the second 22 group begins, and that the last two beats constitute the above 2 group. Not only is it impossible to see in this figure the desired relationship, but any other rela- tionships are equally invisible. The above 46 group corresponds to the generality of the 46 groups which observer R made and yet it can otherwise be shown beyond all possibility of doubt that the relation, 46 = 22 + 22 + 2 A CLASSIFICATION OF GROUPS 289 must correspond in some way and somehow to actualities of function. Consider figure 1: A close examination of figure 1, reading it from right to left, i.e., in the order in which it was made, will reveal 21 beats of varying amplitudes followed by the 22d beat whose amplitude greatly exceeds those of its immediate neighbors. These first 22 beats may be said to constitute a 22 group. The 23d and 24th beats are relatively small and are followed by 22 beats of greater amplitude, the 22d, being again clearly larger than the 21 preced- ing beats. Clearly a 2 group and two 22 groups are outlined in the 46 beats which together constitute the 46 group. The FIG. 1 technique of this delineation is by the method of " amplitude variation," but no particular significance would have attached to such variations did they not so closely correspond to predic- tion, or had they occurred but once. Both the latter facts in- dicate that the variations are not accidental, and this being so, there must exist some logical grounds for their occurrence, which cannot be other than the grounds contained in the prediction. In his work Swindle noted and commented on the fact that the final beat of a group was often accented, that is, larger than the others. He referred to the fact by the term final accent. He sought to identify the accentuation with the occurrence of the 290 CARL W. BOCK subjective positive after-image, and did actually re-discover the so-called positive after-image of long duration because he had noted the final accentuation of groups. Specifically he noted that certain animals often accompanied the final beat of a group by a simultanous beat with another member of the body. Thus certain bears that made peculiar up and down movements of their bodies accompanied the last of the movements by simul- taneously slapping the walls of the cage against which they sup- ported themselves with one of their front paws. This slap, or movement corresponds, so Swindle believes, to the final accent as shown in the above figure except that in the figure the extra movement is shown as made by the same body member and not by an extraneous one. The writer has made similar observa- tions. Generally the last beat is accented. Ordinarily he could not observe however that the final beat of a group was accom- panied by any particularly pronounced movement of any other body member. As a matter of fact every beat of the finger was accompanied by slight movements of every other part of the body and very close observations showed that these slight movements were not necessarily mechanical effects of the moving arm or finger, but actual physiological contractions. These movements were given in the same tempo as that of the main activity, and they could be well felt by simply grasping the limbs in question by the hand and feeling the contractions. And generally the final accent of the mam activity was accompanied by a slight in- crease in the amplitudes of the accessory-movements. The writer wishes here merely to call attention to the principle of the final accent and such other observations which he made in this connec- tion as well as those of Swindle. They are theoretically very important and their discussion will be referred to in a separate section of this study. The fact of the final accent establishes the significance of the discussion of figure 2, and therefore the con- stitution of the 46 group above. Consider figure 2: Figure 2 is entirely similar to figure 2 except that the separate components of the 46 group are not as well delineated as in the preceding figure. Consider figure 3. A CLASSIFICATION OF GROUPS 291 Figure 3 differs from the preceding figures in one essential manner — a small rest period ensues on the completion of the first 22 beats of the group which is followed by 22 more beats. This interval is not of the kind called for by the instructions to the observers; it occurred or was made without the knowledge ' FIG. 2 FIG. 3 of observer R, and under external conditions that cannot be held accountable for it. Whatever the causes for the interval, it is significant that it occurs after the completion of a 22 group and that it is followed by another 22 group. As it stands figure 4 represents a 44 group, or two 22 groups plus the small interval. The interval measures at its base exactly 3 mm. The intervals 292 CARL W. BOCK between any four beats and between every four beats measure 3 mm. Between four elements or beats, there can be inserted 2 beats and three of the smaller intervals that normally occur between any two successive elements of a group.* Conse- quently the above intervals is precisely long enough to permit the interpolation of 2 beats given in the same tempo as that of the other beats of this group; the interpolation made in the above figure corresponds to a 46 group and is in accord with all preceding figures as well as with the prediction made with respect to the composition of the 46 group. Consider figure 4. FIG. 4 Figure 4 is similar to figure 3 with the exception that a 46 group and a 22 group are concerned plus a small interval with precisely the same implications as above; two beats and only two can be interpolated, thus making a total of 70 beats which agrees with the predictions made in regard to the composition of the 70 group. Consider figure 5. Figure 5 is an example of the same kind as the preceding. The small interval permits the interpolation of a 2 group given in the tempo of the other beats but there is a significant difference between the intervals, which significance attaches also to all the other intervals of the same kind already mentioned and to be given. Relatively the intervals are all equivalent in that they permit the interpolation of exactly 2 beats, yet absolutely the intervals are not equivalent because they have different durations. A CLASSIFICATION OF GROUPS 293 FIG. 5 Figure 6 agrees with the preceding figures except respecting the particular groups concerned. The groups are the 22 and 71 (70 + 1) plus the usual interval for the interpolation of a 2 group. It is therefore an attempted synthesis of a 95 group (94 + 1). FIG. 6 Figure 7 corresponds to two groups 46 and 71 (70 + 1) plus an interval into which but one beat may be interpolated. Con- sequently it represents an attempted synthesis of a 118 group, a group which is beat as such twice by this observer. The above are examples selected from 15 occurrences of the same kind in the work of observer R. In a certain sense they must be regarded as quite accidental because they occur rela- tively infrequently and consequently they cannot always be relied upon, except under pre-arranged conditions, to furnish objective evidence and proof for conditions that might be indi- PSTCHOBIOLOGY, VOL. I, NO. 4 294 A CLASSIFICATION OF GROUPS 295 cated by other circumstances. However since they have been found, they do in the present instance furnish the necessary proof for the conclusions that analysis has pointed out, and warrant the assumption that the 46, 70, and 94 groups are compounds of a single group, the 22, and that the several maxima of these groups are therefore conditioned by a common factor and not by four separate factors. The general series (6) thereby suffers the following further modification: (7)a-G,(2a+fc)-G,(3a+2/b)-G,(na+(n-l)/b)-G . . . w-G, x-G, y-G, z-G. 3. The coefficient w. A glance at table 2 and graph 1 will reveal the interesting fact that there occur in the vicinity of the maxima for the groups 22, 46, 70, and 94, groups that are beat relatively more frequently than the average of all other groups, but yet not nearly with the abnormally high frequencies of the above four groups. Their occurrence as such is not significant; their occurrence with the said frequencies is not significant; but their occurrence with their particular frequencies in association with the groups 22, 46, 70, and 94 is significant and raises the question whether the high frequencies of the associated groups is in any manner conditioned by the frequencies of the groups to which they are associated? The groups to which reference is made here are the 24, 44, 45, 47, 48, 68, 71, and 95. Considering only the 46 group and its associated groups 44, 45, 47, and 48, the facts are as follows: the group 46 occurs with abnormal frequency; the four neigh- boring (in the numerical sense) groups 44, 45, 47, and 48 occur in far greater numbers than the average of all other groups, but yet with not nearly the exaggerated frequencies of the group 46. Each of the four groups 22, 46, 70, and 94 have one or more such exceptional neighbors, and with the exception of the groups 3 and 7 of which special mention will presently be made, no other group has a frequency that lies above the average. The question that logically follows is this: why are groups like the 44, 45, 47, and 48, that have relatively high frequencies asso- ciated with groups like the 46 group that has an abnormally high frequency? 296 CARL W. BOCK When it was shown that the groups 22, 46, 70, and 94 were stable activities by virtue of their recurrence, it was assumed that this implied that the said groups were functional units, or behavioral entities of some kind, such units as reflexes, etc. Now it is well known of these more traditional units that they vary within very wide limits from time to time in their temporal and spatial attributes and in the number and sequence of their ele- ments, as well as in their amplitudes and directions. Such variability is to be expected. Groups like the 46 above do in- deed vary. Of the 62 different 46 groups no two are exactly alike except in the number of their elements from which they all derive their common name. In their other attributes they vary very widely. Particularly do they vary with respect to the amplitudes of their individual beats. The latter are obviously minor variations which, consequently, ought to occur, as they do, very frequently on grounds of probability. But variations of 1, 2, 3, or more beats cannot be said to be minor variations and therefore would not occur as often as amplitude variations, not necessarily because they are essentially different kinds of variations but because they imply greater variations of the same kind. Thus a variation of one beat from the typical 46 group implies theoretically only that the 46th beat has an am- plitude 0 which is a greater deviation from the 46 group than 45 beats with a normal amplitude plus one beat with an am- plitude of about one-third of the average of the other beats would be. Since amplitude variations do occur very frequently and since variations of one beat or more are of the same nature as these, differing in degree only, and since they therefore imply relatively large variations of amplitude, it is to be expected on purely a priori grounds that a typical group, as the 46 group, will have associated with it groups like itself and of the same species whose frequencies will in general be the greater, the greater the frequency of the type group; and the less, the farther the indi- vidual members of the same species are numerically removed from the type group. This is obviously the case in the present instance where the above conditions- are sufficiently fulfilled as on graph 1. A CLASSIFICATION OF GROUPS 297 Accordingly groups like the 44, 45, 47, and 48 will, be consid- ered as varieties of the group species 46, and it will be predicted that in any similar investigation such varieties will occur. The coefficient w thereby suffers the following transformations: Wi = a =1= n, w2 = (2a + k) ± n; ws = (3a + 2k) =t n; W4 = (na + (n— l) k) =*= n. 4. The coefficient x. The general series (6) contains the un- defined coefficient x whose nature it is the purpose of this sec- tion to discover. The simple facts upon which the discussion is based are as follows: an organism beats a certain group a very often; it was also shown that he beats groups of the order 2a, 3a, 4a, etc. The question to be asked is whether any other multiples of a might be expected to occur which cannot be dem- onstrated to be such by an appeal to simple frequency criteria? The following numerical equations obtain between those groups that have been shown to be multiples of the group 22: 22 = 22 46 = 22 + 22 + 2 70 = 46 + 22 + 2 94 = 70 + 22 + 2 Continuing these equations with their numerical implications, the following expressions can be obtained: 118 =94 + 22 + 2 262 = 238 + 22 + 2 142 = 118 + 22 + 2 286 = 262 + 22 + 2 166 = 142 + 22 + 2 310 = 286 + 22 + 2 190 = 166 + 22 + 2 334 = 310 + 22 + 2 214 = 190 + 22 + 2 358 = 334 + 22 + 2 238 = 214 + 22 + 2 382 = 358 + 22 + 2 416 = 382 + 22 + 2 In the above equations the numbers on the left hand side repre- sent groups that might be expected to be beat by this observer in view of the fact that he beats the 46, 70, and 94 groups. Of these, the following groups are beat with the indicated fre- quency by this observer: 298 CARL W. BOCK 118 2 times, 333 (334-1) 1 time, 142 1 time, 416 1 time. 190 1 time, It is hardly necessary to comment on the remarkable accuracy of the above predictions and it is believed that the concurrence that obtains between prediction and fact is sufficiently great to warrant the belief that the groups 118, 142, 190, 333, and 416 are of the same order as the 46, 70, and 94, and therefore of the nature of the 22 group. It should be pointed out however that frequency criteria cannot be invoked in the cases of these groups to establish the said relations partly because they are very large groups and partly also for reasons later to be considered. Since the general series suffers no radical modification by the above discussion, no transformations will be made; the general co- efficient x may be taken to stand for all multiples of the type group a that can be predicted as in the above. 5. The coefficient a. With the exception of the general co- efficients y and 2, all other coefficients have been shown to be definable in terms of the coefficient a. It is interesting to ex- amine whether a can be expressed in terms other than itself. A rigorous application of the mathematical law of no exception would lead to the conclusion that such groups as the 22 might be analyzed and expressed in terms of groups smaller than itself in much the same manner as the groups 46, 70, and 94 have been expressed in terms of the 22 group. The application of the above principle however carries with it the necessity for giving to its results a very clear meaning which may not be simply conventional or arbitrary, but which must have some relation to, and which will square with, the morphology and physiology of the organism which has produced the groups in question. In consequence thereof analyses may be carried out only so far as they are possible of interpretation, the limits and meanings of which we are about to examine. The facts or data on which an analysis of the 22 group might be possible are these: the hypothetical components must neces- sarily be smaller than the 22 group; of the smaller groups which A CLASSIFICATION OF GROUPS 299 observer R beat, there must be some definite reason or the selection of a given set of groups as the components, i.e., he selection may not be arbitrary; with a possible exception there are no objective proofs to guide in tVe selection of these compo- nents; the frequency criterion is the sole remaining guide, wh:ch points out the groups 3, 7, 8, and 10 as componential possi- bilities of the 22 group because of their recurrence. The groups 3, 7, 8, and 10 have relatively high frequencies. This cannot be due to chance and inasmuch as all other groups whose frequencies are above the average have been shown to be related to the 22 group there can be little question that the latter FIG. 8 is related in some way to one or more of the groups 3, 7, 8, and 10. A bit of objective evidence which may be offered in support of the conclusion is the fact that the 416 group which has been shown to be a multiple of the 22 group, is made up of fifty-seven 6 groups and two 7 groups. A portion of the 416 group is shown in fac-simile in figure 8. Whether or not the coefficient a or as in the present example, the 22 group, can be reduced to lower terms is in a certain sense not important. The most important fact about the 22 group is that it is obviously a functional unit and a functional entity of a very definite kind in terms of which many other groups may be expressed. If reducible, the same importance would attach 300 CARL W. BOCK to the then resulting group as now attaches to the 22 group, and since for present purposes analysis must cease at some point, it might as well be at the 22 group as elsewhere. The impor- tance that attaches to the 22 group, as it stands at present for the reasons cited, does not depend upon the fact whether the 22 group is or is not an ultimate functional unit, if such exist; it depends solely on the fact that the 22 group is a functionial entity and that it is related to other groups in very definite ways. For that reason the question as to its composition will be left open for the present with this additional remark, that it is without doubt, or was at some time a complex activity, compounded of smaller units whose precise nature is at this time indeterminable. 6. The coefficient y. The general series (6) contains a co- efficient y which represents a class of groups whose description and analysis have not been possible thus far. It is the purpose of this section to consider this class of groups and to this end the data of observer S will be studied (table 3). The question must first be raised and answered on what as- sumptions it is possible to attempt a description and definition of a general series of groups such as series (6) by means of an appeal to the data of several different observers, defining cer- tain of the coefficients of this series on the basis of the groups of one observer and others on that of different observers. The question contains the implication that different observers beat si ch radically different groups or classes of groups that no gen- eralizations are possible, or otherwise it ought to be possible to completely define every coefficient of the general series (6) by the data of a single observer, such as R. It is not the same to say that different observers beat differ- ent groups and different observers beat different classes of groups. Different observers do beat different groups, though they often also beat similar or even identical groups. On probability grounds this is to be expected considering the large number of possible observers. But the statement that observers beat dif- ferent classes of groups contains the contrary to fact implica- tion that members of the same species differ very widely in struc- ture and function, i.e., that they differ more than they are alike. A CLASSIFICATION OF GROUPS 301 Since members of the same species are more alike than they are different, it must be true, or expected, that the laws governing their habits of behavior, and the combination of these, and their succession will generally be the same, particularly when good reasons can be assigned for apparent deviations from the above assumption. TABLE 3 Observer S 1.) 24, 33, 57, 73, 76, 45, 11, 55, 52, 38, 24, 35, 40, 5, 38, — , 21, 113, 11, 48, 78, 23, 79, 198, 53, 77, 66, 20, 131, 22, 37, 10, 143, 80, 20, 11, 36, 32, 38. 2.) 33, 27, 74, 87, 83, 45, 128, 23, 258, 27, 256, 39, 54, 240, 15, 176, 124, 58, 22, 154, 57, 20, 248. 3.) 33, 162, 32, 56, 180, 53, 55, 282, 105, 55, 238, 41, 220, 207, 37, 100, 95, 30, 31, 97, 338, 146, 144, 337, 199, 18, 47, 264, 73, 126, 237, 32, 176. 4.) 44, 51, 480, 380, 170, 153, 98, 207, 435, 120, 176, 209, 404, 45, 84, 81, 85, 39, 80, 105, 58, 32, 141, 120, 110, 18, 141, 174, 17, 161, 103, 982. 5.) 64, 112, 62, 173, 75, 83, 279, 118, 65, 148, 144, 47, 144, 144, 192, 142, 38, 163, 134, 59, 137, 156, 115, 27, 93, 277, 90, 132, 111. 6.) 57, 103, 89, 29, 125, 145, 246, 65, 256, 61, 197, 96, 163, 368, 176, 212. 7.) 33, 129, 50, 84, 80, 89, 180, 94, — , 68, 137, 193, 57, 156, 96, 80, 59, 167, 60, 194, 190, 124, 34, 110, 43. 8.) 65, 100, 35, 116, 162, 46, 162, 120, 171, 90; 38, 63, 233, 42, 75, 35, 147, 92, 94, 102, 127, 151, 156, 47, 128, 97, 235, 34, 45. 9.) 28, 69, 90, 131, 41, 107, 255, 47, 96, 87, 183, 60, 169, 49, 141, 156, 188, 94, 299, 167, 57, 225, 52, 95. 10.) 92, 101, 113, 57, 196, 190, 47, 337, 57, 254, 103, 77, 277, 60, 245, 15, 41, 60, 240, 41, 67, 46, 99. For purposes of convenience the second record of observer S is reproduced below in the order in which it was made or beat. 33, 27, 74, 87, 83, 45, 128, 23, 258, 27, 256, 39, 54, 240, 15, 176, 124, 58, 22, 154, 57, 20, 248 The following facts are of interest in connection with the work of observer R. Observer S beats the 27 group twice and its double, the 54 once; the 128 group and its double, the 256 group, and also the group 258 which differs from the double of the 128 by two beats, making it very likely a variety of the 128 or 256 group species; the 22 group and the multiples 176 and 154; the 124 group and its double, the 248 group. In this respect then, the present observer does not differ from ob- 302 CARL W. BOCK server R at all, except that he does not seem to beat groups of type a and their multiples as frequently as R. This difference however is very significant as will appear hereafter, and it is precisely such differences which make generalizations or group classifications on the basis of a single observer difficult. Consider the first four groups of the above series of groups, namely: 33, 27, 74, 87, i and the corresponding numerical equation: (1) 33 + 27 + 27 = 87. and the next four groups of this series, 83, 45, 128, 23, and the corresponding numerical equation: (2) 83 + 45 = 128. and the following seven groups: 258, 27, 256, 39, 54, 240, 15, and the corresponding numerical equations: (3) (9 X 27) + 15 = 258 (4) 39 + 54 +39 + 54 + 54 = 240 (5) 27 + 27 = 54 (6) 39+15 = 54 (7) (16 X 15) = 240 and finally the series: 176, 124, 58, 22, 154, 57, 20, 248, and the corresponding numerical equations: (8) (8 X 22) = 176 (9) 58 + 22 + 22 + 22 = 124 (10) 57 + 20 + 57 + 20 = 154 (11) (7 X 22) = 154 (12) 124 + 124 = 248 (13) 57 + 57 + 57 + 57 + 20 = 248 A CLASSIFICATION OF GROUPS 303 The question to be answered is this: what relations exist be- tween any of the above short series of groups and the corre- sponding numerical equations? It must be noticed first that all the numbers which occur in equations (1) to (13) occur also in that part of the series to which they correspond, and that no other number which does not occur in the corresponding part of the series occurs in the equations. Two possible conclusions can be drawn from these correspondences; either we are con- cerned with a remarkable set of coincidences or the above nu- merical relations have some significance. The first conclusion is untenable from the fact that such numerical relations as be- tween the coefficients of groups are by far too common to be regarded as accidents, and secondly from the fact that in prin- ciple the above equations do not differ from similar equations obtained in the case of observer R except that in the latter case the equations involved the addition of identical coefficients whereas here they involve the addition of non-identical coeffi- cients. From this it is obvious that the general coefficient y may be taken to represent groups whose coefficients are ob- tainable by the addition of two or more different groups which latter groups must occur in the temporal vicinity of the groups whose components they may be said to be. An extension of the arithmetical principles involved in the above equations is possible when they are obtained not only as between neighboring groups of the same record, but also be- tween any groups at all within the said record. Below are given in categorical form a list of the more important equations ob- tainable from among the several groups of the series under dis- cussion. With certain exceptions, the equations do not contain any numbers that are not found as such among the coefficients of the groups of the above series. The exception mentioned above consists of the use of the half of the coefficient of groups in the case of even numbered groups, and approximate halves or so-called physiological halves of odd numbered groups. The only justification for the use of said numbers in the making of analytical equations as above comes from the fact that it is so generally possible to do so, and from the obvious relation that 304 CARL W. BOCK exists as between halves and the multiples of a number. Thus we may consider the 46 group the physiological double of the 22 group, or we may consider the 22 group the physiological half of the 46 group. A partial list of such equations follows: 74 = 33 + 27 + 27/2 (14) = 27 + 27 + 20 = 54 + 20 = 39+15 + 20 87 = 33 + 27 + 27 = 33 + 54 = 33 + 39+15 = 57+15+15 = 27 + 45+15 = 27+15+15+15+15 83 = 33 + 27 + 23 = 33 + 33 + 33/2 = 23 + 45+15 = 23+23 + 22+15 = 23+15+15+15+15 45 = 15 + 15 + 15 = 23 + 22 128 = 83 + 45 = 74 + 54 and all substitution products. 23 = 23 258 = (128 + 1) + (128 + 1) = 83 + 23 + 23 + 83 + 23 + 23 = 45 + 45 + 45 + 45 + 45 + 33 = 33 + 33 + 33 + 33 + 33 + 33 + 33 + 27 = 27 + 27 + 27 + 27 + 27 + 27 + 27 + 27+27+15 and all substitution products. 27 = 27 A CLASSIFICATION OF GROUPS 305 256 = 83 + 45 + 83 + 45 = 128 + 128 = 54 + 74 + 54 + 74 and all substitution products. 39 = 39 , 54 = 27 + 27 = 39+15 240 = 39 + 54 + 39 + 54 + 54 = 33 + 27 + 33 + 27+33 + 27 + 33 + 27 = 176 + 27+15 + 22 = 124 + 83 + 33 = 124 + 39 + 23 = 124 + 74 + 27+15 = 128 + 58 + 54 = 128 + 27 + 45 + 20 + 20 = 83 + 83 + 74 = 74+ 87 + 57 + 22 = 74+ 87 + 39 + 20 + 20 = 45+ 45 + 45 + 45 + 45 and all substitution products. 15 = 15 176 = 22 + 22 + 22 + 22 + 22 + 22 + 22 + 22 = 154 + 22 = 124 + 15 + 15 + 22 = 57 + 22 + 45+15 + 22+15 . = 58 + 58+15+15+15+15 = 128 + 33+15 = 33+ 33 + 33 + 33 + 22 + 22 = 33+ 33 + 33 + 57 + 20 = 33+ 33 + 45 + 45 + 20 = 27+ 27 + 27 + 27 + 27 + 27 + 27/2 = 27 + 27 + 27 + 27 + 33+15 + 20 = 74+ 57 + 45 = 74+ 87+1J5 = 39 + 39 + 39 + 39 + 20 = 39+ 54 + 83 and all substitution products. 306 CARL W. BOCK 124 = 248/2 = 58 + 22 + 22 + 22 = 15 + 15+15+15+15 + 20 + 20 + 22 = 39 + 39 + 23 + 23 = 45+ 57 + 22 = 20 + 20 + 20 + 20 + 22 + 22 = 33 + 33 + 58 , = 27+ 27 + 27 + 20 + 23 = 74+ 27 + 23 = 58+ 33 + 33 154 = 22 + 22 + 22 + 22 + 22 + 22 + 22 = 57+ 20 + 57 + 20 = 33 + 33 + 33 + 33 + 22 = 27+ 27 + 27 + 27 + 23 + 23 = 58 + 58+15 + 23 = 45 + 45 + 22 + 22 + 20 = 74+ 33 + 27 + 20 = 74 + 27 + 23+15+15 = 124 + 15+15 = 87 + 45 + 22 = 39+ 27 + 39 + 27 + 22 = 58 + 22 + 74 = 58 + 57 + 39 248 = 124 + 124 = 154 + 54 + 20 + 20 = 154 + 74 + 20 = 124 + 74 + 20+15+15 = 176 + 20 + 20 + 22 = 128 + 45 + 45+15+15 = 128 + 33 + 33 = 87+ 87 + 74 = 57 + 57 + 57 + 57 + 20 = 57 + 57 + 57 + 22 + 20+15 and all substitution products. Close scrutiny of the above tables will lead to the conclusion that it is generally possible to express any of the coefficients of the larger groups in terms of combinations of the coefficients of A CLASSIFICATION OF GROUPS 307 the smaller groups; and that it is not only possible to find one combination of coefficients to satisfy any of the larger groups but frequently very many, the coefficients being so intimately related numerically. To discuss thoroughly one of the above groups, consider the 240. Before it was beat by observer S, he beat the two groups 39 and 54; then he beat the 240 and after it the 15. Now the sum of two 39 groups and three 54 groups equals 240, a number which agrees precisely with the coefficient of the 240 group, and from this it might be concluded (and under conditions prop- erly though not necessarily so) that the 240 group is actually a succession of 39 and 54 groups each the requisite number of times. And the relation between the 54 and 39 groups itself would strengthen this conclusion for, together with the 15 group which the observer beat immediately after the 240, the 39 equals the 54 (39 + 15 = 54) indicating a relationship between these two groups which enter apparently into combination to form the 240, the more so since 240 is also a multiple of 15. Consider also the 154 group: it is evidently a multiple of the 22 group beat immediately before i.e., 7 X 22 = 154; but fol- lowing it are the groups 57 and 20 which in the combination 57 + 20 + 57 + 20 also equals 154. A question of consider- able significance arises, for it may at once be asked how it is possible for a group to have more than one set of components, particularly where there is no apparent relation between the several members of the two or more sets? Thus, while seven times 22 is equal to 154, it is difficult to see how this same 22 group can be a component seven times of the groups 57 and 20 which are also components of the 154 groups. Without enter- ing into a discussion of this question at this time, it may be said that it is for this reason that the writer does not expect to find his analytical results objectified in the records themselves, for, of the many analytical possibilities, which ought one neces- sarily find? are some of these possibilities actualities and others not? and if so, what criteria are to be used in establishing the identity of the right combination? 308 CARL W. BOCK Below are given in a convenient form a few other examples of the same kind from the records of observer T and S. In every case the group to be analyzed is starred and below it under the horizontal line are given the several sets of components found for the group above the line: Resume of record 8, observer S 97* 47 + 32 + 18 338* 97 + 144 + 97 47 + 32 + 18 47 + 47 + 32 + 18 47 + 32 + 18 146* 73 + 73 144* 47 + 47 + 32 + 18 337* 146 H- 47 + 144 73 + 73 47 + 47 + 32 + 18 199* 73 + 126 47 + 47 + 32 18* 47* 264* 199 +18+47 126 + 73 47 + 47 + 32 73* 18 + 18 + 18 + 18(?) 126* 47 + 32 + 47 32* A CLASSIFICATION OF GROUPS 309 176* 144 + _32 47 + 47 + 32 + 18 Resume of record 8, observer S 65* 100* 35 + 65 35* 116* 35 + 46 + 35 162* 116 + 46 35 + 46 + 35 46* 120* 90* 38* 63* 233* 90 + 38 + 63 + 42 42* Resume of record 4, observer T (To be read from left to right.) 12,* 71* 81* 22* 29* 12 + 59 22 + 59 12 + 47 47 + 12 110* 47* 103* 29 + 81 59 + 22 + 22 22 + 59 47 + 12 12 + 47 PSYCHOBIOLOGT, VOL. I, NO .4 310 CARL W. BOCK 262* 103 59 + 71 50 + 22 + 22 47 +12 12 + 59 12 + 47 Resume of record 7, observer T 8,* 14* 14* 38* 16* 174 89* 8+14+16 10 X 16 + 14 94* 103* 10 X 8 + 14 6 X 16 + 12 All observers whose records have been cited above, observed while they were beating a "feeling" that they should or could cease beating at certain times, for some reason however not doing so. Observers were instructed to accent the beats at which the above feelings were perceived by tapping a little harder. These accents would accordingly divide the whole group into a number of suc- cessive groups whose limits were conditioned by, or accompanied by, the said mental states or " feelings." It was desired to as- certain whether or not these mental states had any behavioral correlates or whether the instructions to accent would as it were "salt out" groups that were known to be, or which could be shown to be, functional entities in the sense in which this term is used in the present study. In oase it happened that the accents, thus determined, coincided with or conditioned groups that were entities, it would mean that the above "feelings of being able to stop" had objective correlates, and that they existed because the organism in question at other times did actually stop beating when the last element of this or that group had been made. As will be seen from the examples to be cited, the facts agree with expectations, but certain factors make it impossible gen- erally to make use of this technique for the purpose of salting out groups. The instructions to accent a beat or beats, when the said mental states obtained, themselves act as a stimulus for having these same mental states at times where they would not have been had, had no instructions been given. The result is that, while at first a certain modicum of success will attend A CLASSIFICATION OF GROUPS 311 the use of the technique, presently the number of accented beats increases so that not only is it impossible to use the data obtained in this way, but it " spoils" an otherwise good observer whose habits become continually modified by the instructions given. This condition no doubt obtains generally, particularly in intro- spections, and it was for this reason that the writer not only did not require introspections, assuming them to have value, but hesitated to mention the word introspection lest the habits of his observers might become too radically changed either by actual instructions or by casual references to it. Observer Q beat the following series of groups under conditions as above: 166 81 225. 148 170 The 166 group was accented on the 83 beat and on the 137th beat. The 83d beat divides the 166 group into two equal parts, namely, two 83 groups. Thus the present observer agrees with others cited in this work. The 137th beat is the 54th beat of the second hypothetical 83 group, thereby dividing this group into the two groups 54 and 29. The 81 group following, which differs only by 2 from the two 83 components of the 166 group is accented on the 54 beat, i.e., thereby separating out another 54 group, and dividing it into the groups 54 and 27 which latter group equals 54/2. The 225 group was accented only on the 54th beat, and the rest of the series is unanalyzable. The same observer beat the following series: 66 86 161 117 42 The group 66 was accented on the 43d beat, i.e., on the 86/2 beat, which 86 group follows the 66 group. The group 161 was accented on the 57th beat, thereby dividing it into the groups 75 and 86. The 86 groups actually preceded the 161. The 117 was accented on the 75th beat, dividing it into the groups 75 and 42. The 75 was shown to be a similar component of the pre- ceding group, and the 42 group actually follows the 117 group. From the above it would seem possible to conclude that the coefficient y of the general series (6) is definable in terms of 312 CARL W. BOCK the sum of two or more coefficients of the type a1} a2, a3, etc., where aiy o2, a*, etc., are of the order a and definable as this one has been defined. Thus a peculiar difference obtains between the observers R and S. Where R beats but a single type group a, (22) and its multiples or physiological multiples, and beats these relatively very frequently, S beats several of the type a, as aif a2, a3, etc. but does not beat any of these with nearly the high frequencies obtainable in R. Whether or not the coefficients aij 0,2, 0,3, etc., can themselves be defined in lower terms, or in terms of a coefficient a is insignificant for present purposes. It is only significant that such coefficients exist, and that the groups they represent are functional entities in terms of which a large number of other groups can be defined. As previously stated, there can be but little question that if a present relation does not obtain between the above type groups, a genetic rela- tionship does obtain which would mean that these groups perhaps once had a common ancestor functionally considered. 7. The coefficient z. The general coefficient 2' will usually represent all groups that are not capable of analysis as indi- cated in the previous sections of this study. It may be inter- esting to point out why it will not always be possible to analyze the totality of any given series of groups, and why there will be found certain individuals whose groups will be generally not1 capable of analysis. The chief reason why this must be so is that the environ- ment even in the best controlled experiments will play its cus- tomary role in modifying behavior of any kind and particularly the habits under present discussion. The present writer took only the most general precautions to insure a more or less con- stant environment for the reason that it is generally quite im- possible to realize even the faintest approximation to this ideal ex- cept where observers are totally devoid of all senses and memory, which would make them worthless for our purposes, but more so because such conditions, whether realized either wholly or in part, are so unnatural and so unusual that they themselves constitute one of the gravest sources of stimulation, and con- stantly condition corresponding modifications of their habits, A CLASSIFICATION OF GROUPS 313 a very undesirable goal in investigations like the present one. The habits in the possession of an organism have generally been acquired under every-day conditions and they are there- fore least liable to change under the conditions under which they were developed than under any other conditions. Since an observer cannot, in general, be entirely removed from sources of environmental stimulation, it must happen now and then, depending on the one hand on the degree of the con- stancy of the environment or on the "habitualness" of the environ- ment, and on the other on the constitution of the organism itself, that groups will be beat whose nature depends essentially on the environment and not on the organization of the organ- ism, which, if it could be accurately measured and defined, would establish the identity of every group of any observer. 8. Other analytical methods. It occurred to the writer to ex- tend the methods so generally successful in the consideration of a single series of groups of any given observer to a comparative study of several records of the same observer for the purpose of ascertaining whether the same general relations obtained for the same group in different series. For this purpose, records of the same observer (beat on different days) which contained a com- mon group were selected and compared. The comparisons were effected between the groups that immediately preceded or fol- lowed the group common to the several series. In records 3 and 10, observer S beat the following series: / 57 196 47 337 57 354 \ 199 18 47 264 73 126 The common group in the two records or parts of records is the 47. If a simple relation obtains between any pair of groups that have about the same temporal positions with respect to the common group 47, or any common group, or any groups that are known to be related, then the significance of the relations that obtain for the groups of any single series is thereby strength- ened because the corresponding probability problem becomes more restricted. Another condition or another set of conditions 314 CARL W. BOCK enter into the problem to which the coefficients of the series must conform. Subtract the number 264 from the number 337 and a number 73 obtains. But this number is identical with the coefficient of the group 73 which follows the 264 group. From this it appears, that while observer S follows the 47 group by the 337 in record 3, he also follows this group by the 337 group in record 10 with this difference only, that in 10 he beats the 337 group in two parts, while in 3 he beats the 264 and 73 groups in strict succession. Records 3 and 4, observer S, give the following series: 18 47 264 73 126 237 18 141 174 17 161 103 The following equations indicate some of the relations here found : (1) 141 = 47 + 47 + 47 (3) 174 = 73 + 47 + 18 + 18+18 (2) 264 = 174 + 73 + 17 (4) 161 = 126 +17+18 The complex may therefore be written in the convenient form: 18* 47* 264* 73* 126* 174 + 17 + 73 47 + 73 + 18 + 18 + 18 18* 141* 174* 17* »161* 47 + 47 + 47 47 + 73 + 18 + 18 + 18 17 + 128 + 18 73 + 17 + 17 + 18 + 18 + 18 The following series are taken from the Records of Ob. T. : 8 12 16 12 13 25 56 16 72 36 12 38 16 174 89 89 94 16 108 — 15 13 16 25 11 A CLASSIFICATION OF GROUPS 315 Analyzed, the following relations are found: 8* 12* 16* 12* , 13* 25* 56* 8 + 8 16* 72* 36 * 12 + 13 12* 12 + 38* 12+ 12 16 + # 12 + 8 8 + 86 174* X 12 12+ 12 + 12 89* 12 X 13 + 13 8 + 8 108 + 25 + (13 X 3) (5 X 13) + (2 X 12) 94* 16* 108* (5X13)4- (2X12) 56 + 38 8+8(9X12) 15* 13* 16* 25* 11 8 + 8 12 + 13 The following series are taken from the records of S: 33 162 32 56 33 129 50 84 33 27 74 87 24 33 57 73 76 Analyzed, they take the following form: 33* 162* 32* 33 + 129 57 + 24 + 24 + 24 33 + 24 96 32 + 32 + 32 33* 129* 50< 57 + 24 + 24 + 24 33 + 33/2 33 + 24 96(1) 32 + 32 + 32 27 + 27 + 27 + 24 + 24 33* 27* 74* 33 + 27 + 27/2 24* 33* 57* 73* 24 + 33 33 + 27 + 27/2 (1) The 96 group appears further on in the first series and therefore cannot be shown. 316 CARL W. BOCK In records 2, 7, and 9, observer S beat the following series whose common group, the 57, does not appear for want of space : 48 33 96 258 129 86-7 27 50 183 (2) 84 60 39 80 169 (2) 89 49 240 182 + 1 142 + 27 54 134 + 48 256 (2) These groups were objectively analyzed into the indicated groups by the method of accents; without instructions. Analyzed, they take the following form: 33* 129* 50* 33 + 96 33 + 33/2 48 + 48 33/2 + 27 + 86 27 + 27 + 32 54 48* 96* 87* 183* 48 + 48 1 + 86 182 + 32 + 27 + 27 48 + 134 33 48 + 86 96 87 48 + 134 + 1 84 + 50 60 + 84 + 39 33 + 27 258* - 27* 256* 129 + 129 138 + 128 33 + 96 45 + 83 74 + 54 48 + 48 134 + 50 + 27 86 + 48 33 + 27 + 27 A CLASSIFICATION OF GROUPS 317 (Continued.) 84* 80* 89* 27 + 57 15 + 50+15 39 + 50 33~T 24 33 + 33/2 33 + 33/2 60* 169* 49* 33 + 27 80 + 89 33 + 33/2 15 + 50 + 15 50 + 39 39* 54* 240* 39 +15 80 + 80 + 80 27 + 27 60 + 60 + 60 + 60 89 + 54 + 49 + 48 The examples cited above seem to indicate that there exists a quasi-stability in the order or succession in which groups are beat at different times, which is manifested in the fact that groups which have the same or about the same place in the temporal order of the series with respect to a group common to the several records can be expressed in terms of each other, or in terms of common groups, or by the sum of two or more groups of the same set, or by the sum of two or more groups of different but neighboring sets. This may be perhaps inter- preted that the above relations merely express the simple fact that an observer like S beats several typical groups of the order aJt a2, eB} a4, etc., and their many combinations, and tends to beat them from day to day without their suffering any radical modifications of any kind, whether from environmental causes or internal causes, thereby giving to these series of groups the appearance of associations akin to the associations of the sub- jective field, and doubtless of the same kind. Thus the several different series, oriented with respect to a common group, may be regarded as a number of different modifications which one of them, or an entirely different one have from time to time suf- fered from various causes. Genetically there are several inter- esting problems contained herein which are capable of experi- mental treatment, but which the writer will defer for later consideration. 318 CARL W. BOCK CONCLUSIONS Quite apart from attempts at interpretation, and considering groups only from their behavioral or functional aspects, inde- pendent of the organisms which have produced them, or the morphology and physiology of these, the following specific con- clusions are warranted from the facts of the preceding pages: 1. The totality of all group activities beat by organisms may be represented in a generalized form, as follows: a-G, b-G, c-G, d-G, e-G, w-G, x-G, y-G, z-G 2. The coefficients of this series may generally be substituted for, a. By one, several, or very many groups of the order «i, a2, a3, a4, etc., these being functional units of varying degrees of stability, the degree varying inversely with their numbers, and themselves being generally incapable of further description except in a hypothetical genetic sense. 6. by simple multiples of such type groups of the order «i, «2, 03, etc., and their halves, or by physiological multiples or halves. c. by varieties of such type groups whose frequencies will gen- erally be in inverse ratio to their deviation from their type form. d. by the sums of such type groups, or by the sums of their multiples, each taken any number of tunes. e. by modifications of the above kinds of types of groups due to environmental factors, or to internal causes, in which case such modified groups are not generally capable of analysis with- out a complete description of the modifying factors. 3. Generally the same analytical methods are applicable to the group activities of all observers, who differ only in the number of type groups of the order a\ 02 0s which they beat, and their respective frequencies. A SYNCHRONOUS MOTOR KYMOGRAPH KNIGHT DUNLAP From the Psychological Laboratory of the Johns Hopkins University The instrument herein described was constructed in the physics workshop of the Johns Hopkins University from my specifications, and has been in successful operation. It will probably be put on the market, as soon as arrangements can be made therefor, by C. H. Stoelting. The instrument consists essentially of a horizontal cast iron base, My on which are mounted a ten-pole synchronous motor, of the type used in the chronoscope previously described1 by me ; a drum, L, with magnetic clutches, FI and F2, of the type de- signed by me and used on the chronoscope; and a worm-wheel, E mounted on a short shaft. The motor is mounted with the armature, C, in the horizontal plane, a worm, D, is cut on the armature shaft, which is of steel, and this worm engages the worm-wheel, E. The shaft of the worm wheel carries the core of the clutch-magnet FI, the core having at its outer end an expansion GI, with flat annular external surface. The armature, and worm wheel and shaft are intended to rotate continuously, the drum remaining at rest until electric current is sent through the windings of the clutch-magnet, FI, from the terminals Oi, whereupon the expansion of the core, Gi, adheres to the disc, HI, mounted on the drum-shaft, and the drum-shaft and drum rotate with the worm-shaft. Upon the breaking of the circuit through the windings of the clutch-magnet, the drum comes to rest, without disturbing the rotation of the armature and the worm-shaft. If desired, current may be applied to the breaking clutch- magnet, F2, at the moment in which the current through FI is 1 Journal of Experimental Psychology. 319 320 KNIGHT DUNLAP interrupted, causing the drum to be positively arrested and rigidly held, through the attraction of the disc H2 to the non- rotating magnet core Gz. The worm-wheel on the present instrument has 150 teeth, and hence when the motor is driven by a 25 vibration fork, the drum makes one rotation in one minute, if the fork and motor are in simple step. However, by starting the motor at double speed (either speed may be secured by a twist given to the armature- shaft by the thumb and finger) it may be made to run at that rate, two poles passing for each current-interruption, and the drum makes one rotation in thirty seconds. Forks of different rates may be used, up to 100 vibrations per second, although ordinary forks above 50 in frequency do not make good contact. Driven by the 60 cycle alternating current, without fork, but with suitable resistance, the drum makes one rotation in twelve and a half seconds. If a rectifier, of the "Tungar" type be used instead of resistance, the drum makes one rotation in twenty-five seconds. The motor runs well on alternating current of frequencies of from 15 to 120 per second. At the slower speeds: with 25 and 50 vibration forks, or with the 60 cycle A. C. with rectifier, or 25 cycle A. C. without rectifier: the method of starting the motor described above (thumb and finger) is the best. For the higher speeds; with 60 cycle A. C. without rectifier: a different method must be used. A single layer of adhesive tape (electricians, or surgeon's, is wrapped smoothly around the armature shaft, and the motor is started by drawing the fingers smartly across the shaft. This method of starting is easier if the current is off the motor-field, and is put on by closing a switch at the moment when the right speed is obtained. In any case, the starting is a " knack'7 which should be readily learned, after which the process is simple. With any method of motor drive, any speed witnin reasonable limits may be obtained by using a worm wheel of the requisite number of teeth. Each size of wheel requires, of course, a specific location of the motor on the base, so that the kymo- graph should be built for the speed of drum required. It is A SYNCHRONOUS MOTOR KYMOGRAPH 321 entirely feasible, however, to design a base with slots for bolting the motor to it, and mount a series of worm-wheels on the horizontal shaft, so that by sliding the proper wheel into position, and moving the motor into its proper position for that wheel, any of the speeds corresponding to the series of worm-wheels may be employed. The kymograph may also be designed to run with a vertical drum, but in this case the drum could not be so simply removed and replaced as in the present instrument. The drum-shaft rotates on bearings Ji and /2, which I have designed especially for this mechanism. Each bearing consists of three screws with polished hardened ends on which the shaft rests. By adjusting these screws, which are provided with lock-nuts, the drum-shaft may be lined up accurately with the worm-wheel shaft. When the current is off the clutch magnets, the drum may be lifted out of the bearings without the un- fastening of any catch, and it is as easily replaced. The writing levers used with the present kymograph are mounted on a heavy iron block, with planed self-cleaning grooves sliding on planed guides on the base (not shown in the cut). For work where records occupying not more than one drum- rotation are required, this is the ideal system. The block may be slid to any desired lateral position, which it will keep without fastening: and it may be slid to the left far enough to remove the writing from the drum, and allow the latter to be lifted out. A mechanical shift of the carrier, for spiral records on the drum, could be arranged. The instrument can also be designed to use continuous paper. For the satisfactory driving of a small synchronous motor of the type used in this kymograph, platinum contacts should not be used on the fork. Dental wire, of gold hardened by alloy with platinum, is best for the moving part of the contact, attached to the fork tine, and a plate or stud of the same alloy should be used for the fixed part of the contact (although platinum is satisfactory for this part). I have found that a short section of large gauge wire, set in the end of the adjusting screw, is a satisfactory arrangement for the fixed part of the contact. 322 KNIGHT DUNLAP A SYNCHRONOUS MOTOR KYMOGRAPH 323 i_J2 324 KNIGHT DUNLAP The practical advantages of this kymograph are: 1. Simplicity of construction. There are few parts, and these are of simple and strong design. 2. Comparative noiselessness. If desired, the master fork may be placed in another room, or in a sound proof box. The motor itself makes little noise when operated on a fork-interrupted current. Driven by A. C., the noise is a little greater. 3. Ease of operation. The motor being in continuous rota- tion, the drum may be started and stopped at will, by manipu- lating a switch anywhere in the room or in another room. 4. Economy of writing surface. Full speed, and complete stop are attained by the drum in a very small fraction of a second, so that the usual waste of surface due to the " picking up" in speed of the usual spring or motor drive is eliminated. 5. Simplicity of time measurement. The rate of the fork being known, the rate of the drum is known and is invariable. The clutch is so efficient, that in tests which I have made, a fork-tracing taken several times around the drum has given but a single line, the writing point following the same path on the first, second, third, and following tracings. In making this test, a recording fork must be used which is in exact tune (snychronous or even multiple) with the fork driving the motor; or a magnetic recorder may be driven from the same fork. In the latter case there may be slight temporary disturbances of the superposition due to changes in the latency of the magnet, but these may be detected, as may also aberrations due to change in amplitude or in form of the vibration. Knowing the rate of the drum, and its diameter, the relation between seconds and millimeters may be established, and hence no time-line on the drum is needed in most cases, since the records may be read by the aid of a millimeter scale. DUNLAPS METHOD FOR THE MEAN VARIATION BUFORD JOHNSON From the Psychological Laboratory of the Bureau of Educational Experiments In the Psychological Review for March, 1913, Knight Dunlap described a simple method developed by him, of obtaining the mean variation of a series of values, together with the necessary operations for carrying out the process on a calculating machine. In the derivation of the formulae there is one obvious misprint in writing (SP + PM) for (SP - PM), but this is correctly printed when the same expression is used in the succeeding line. The formulae derived are as follows: MV = (SP - PM) +1/2N = (RM - 2R) -- 1/2 N, when N = total number of measures M = average or mean M V = mean variation P = number of terms greater in value than the average R = number of terms less in value than the average SP = sum of the terms greater in value than the average 2R = sum of the terms less in value than the average These same formulae are reproduced, with a slightly different symbolization, in Whipple's1 Manual of Mental and Physical Tests, Part I: Simpler Processes, which was published in 1914. The following substitutes are used: N+M forP N-M ior R 2+M forSP S_M for 2R .5 for 1/2 1 Whipple's Manual of Mental and Physical Tests, Parti: Simpler Processes, p. 22. Dunlap developed the method itself and not merely the application of the calculating machine technique, as Whipple's footnote (p. 21) might imply. We are not acquainted with any previous development of the method. 325 PSYCHOBIOLOGY, VOL. I, NO. 4 326 BUFORD JOHNSON Two alternative rules, based on the formulae, are given by Dunlap. In cases where all of the original terms employed are either positive or negative in sign, either of these two rules may be used. The second is a simpler one for operation on the machine. However, there are cases in which some of the terms averaged are positive and some are negative, the average being based on the algebraic sum of the terms. In these cases there may be difficulty in the application of rules (1) and (2), A third rule should be added which makes the procedure in such a case unmistakable. The procedure indicated in this third rule will avoid all danger of errors which may otherwise be serious. The three rules which satisfy all the cases arising are as follows : 1. Add together the terms greater than the average; from the sum subtract the product of the number of terms (P) so added, multiplied by the average (M) ; and divide the remainder by half the total number of terms in the series (1/2 N). 2. Add together the terms which are less than the average: subtract the sum from the product of the number of terms (R) so added, multiplied by the average (M)] and divide the re- mainder by half the number of terms in the total series (1/2 N). 3. In cases where the average is based on the algebraic sum of positive and negative terms, if this average is positive, compute the mean variation in accordance with Rule (1) from the positive terms which are numerically greater than the average. If the average is negative in sign, compute the mean variation in ac- cordance with Rule (2) from the negative terms which are numerically greater than the average. ACTION OF SOME ANTIPYRETIC ANALGESICS ON PSYCHOLOGICAL REACTION TIME D. I. MACHT, S. ISAACS AND J. GREENBERG From the Pharmacological Laboratory of the Johns Hopkins University In a preceding communication by Macht and Isaacs pub- lished in this Journal (1), the authors reported the results of their investigations on the effect of opium and its principal alkaloid, morphin on the psychological reaction time. Opium and mor- phin are the most effective analgesic or pain-relieving drugs at the disposal of the pharmacologists. There is, however, another class of drugs which is also very efficient in the relief of pain, es- pecially of a neuralgic character. This is the group of so-called antipyretics which derive their name from the other very inter- esting property which they possess, namely that of reducing tem- perature in cases of fever. This group of drugs includes a large number of substances widely employed by physicians and laymen for the relief of headaches, neuralgias, rheumatic pains, etc. Following the investigations on the effects of opium and morphin on psychological reaction time, it was interesting to inquire into the effect of the antipyretics in this respect. Accordingly, the present research was undertaken. The only previous work of importance on the subject worth considering is that of Munsterberg (2). That author reported some experiments on the effect of three antipyretics — quinin, antipyrin, and phenacetin — on some mental efficiency tests. The tests employed by him were the reproduction of consonants or digits read to the subject, the counting of letters in a given text within a given perio,d of time, the time taken to name ten colors presented in a row, etc. Munsterberg found marked impairment following antipyrin and quinin. From the pharmacological point of view, however, the doses of the drugs employed in his experi- ments were entirely too large or toxic. Thus, for instance, the 327 328 D. I. MACHT, S. ISAACS AND J. GREENBERG dose of antipyrin administered by him to his subjects was 1 gram or 15 grains; whereas the ordinary therapeutic dose of that drug employed at present is a quarter of a gram or about 4 grains. In the present investigation, the doses of all the drugs employed were therapeutic doses. METHOD The method of investigation was the same as that followed in the opium experiments. The reaction time was measured by means of an improved chronoscope devised by Prof. Knight Dunlap, which is a far more accurate and convenient instrument than the old Hipp instru- ment. The apparatus is described by Professor Dunlap else- where (3). It consists essentially of a synchronous motor, run on a tuning fork vibrating fifty times per second, and registering the time in units of 2o- or 1/500 of a second, the dial-hand of the chronoscope being controlled by an electro-magnetic clutch. The simple sound reaction was obtained by the experimenter calling out a word or number into the speaking disc which started the chronoscope and the subject responding with a set answer as soon as possible through another speaking disc, thus stopping the clock. The results were then recorded in terms of 2o- or 1/500 of a second. It is needless to state that the subject and experi- menter were separated by a curtain in order to prevent their seeing each other. The simple touch reaction was obtained in a similar manner. The experimenter touched the hand of the subject behind a cur- tain, the pressure of the touch starting the chronoscope going. The subject responded as soon as he perceived the touch sen- sation by pressing a bulb or touching a key which immediately stopped the clock. The simple light reflex was tested by the experimenter's press- ing a key and thus lighting an incandescent lamp behind a white screen, the subject responding by pressing another key which extinguished the light and stopped the chronoscope. In order to determine the more complex reaction time or asso- ciation reaction time, various devices were tried, such as response to certain words (nouns and adjective, subject and predicate, ACTION OF ANTIPYRETIC ANALGESICS 329 etc.), but none of these were found satisfactory for the purpose in view. The most convenient and satisfactory method was finally found to be the calculation of a mathematical problem. Two sets of problems were submitted to the subjects in all experi- ments. In one series the subject was requested to add 17 men- tally to a two-figure number, such as those given in the following table (table A) and to announce the sum as quickly as possible through a telephone arrangement which breaks the circuit and stops the clock. In the second series a more difficult task was given to the subject. The experimenter in this case announced a two-figure number and the subject was required to multiply the same by 3 and add 4 to the product, and then announce the result through the speaking disc, thus recording the reaction time (table B). TABLE A TABLE B Exercise: Add 17 mentally Exercise: Multiply by 8 and and respond add 4 o,nd respond 22 35 29 64 71 58 76 27 68 41 55 46 46 75 48 23 33 38 24 69 64 67 79 33 51 54 54 65 32 28 47 22 59 47 58 53 73 26 75 67 29 43 49 39 56 31 68 52 48 66 35 61 34 57 71 36 62 42 , 37 28 78 24 66 73 25 55 45 44 39 27 34 31 74 79 21 25 52 45 78 74 77 21 51 77 49 63 26 43 44 23 72 59 72 37 32 63 53 61 41 56 76 65 38 42 36 64 57 62 330 D. I. MACHT, S. ISAACS AND J. GREENBERG In each test twenty numbers were generally employed at each sitting. This method furnished quite a complicated association test and at the same time eliminated as much as possible mem- ory and habituation or familiarity. The subject in every case was expected to go through the mathematical process in his mind and not to rely on his memory at all. Great attention was paid in the association tests to the number of errors made, and these were recorded for comparison of the normal reaction time with that obtained after the administration of a drug. SUBSTANCES STUDIED After the normal simple and complex reaction times were es- tablished in any one experiment, the subject was given a drug, and the reaction time was then again measured repeatedly at intervals in order to determine the effect of the substance. The effect of the drug was indicated by changes either in the mean reading or by changes in the mean variation between the readings, or by both. In case of association tests the number of errors made in computations, before and after the drug was administered, was also taken into account. Inasmuch as the antipyretic drugs do not lend themselves well to administration by injection, all the drugs were given by mouth under the supervision of Dr. Macht. In testing simple reactions to sound, touch and light, the num- ber of readings taken were generally from twenty to fifty or more in each series. In testing the association time, twenty problems were submitted by each method. An average reading was com- puted with the help of a calculating machine, thus saving an enor- mous amount of time, and the mean variations were also com- puted by means of an adding machine, in accordance with Dun- lap's method (4). The experiments were performed for the most part on the authors themselves, and occasionally on other subjects. About forty experiments were made in all, each lasting from two to five or more hours. The drugs studied were the following: quinin, acetanilid, acetphenetidin (phenacetin), antipyrin, phenyl salicy- late (salol), acetyl-salicylic acid (aspirin), and pyramidon. In ACTION OF ANTIPYRETIC ANALGESICS 331 order to ascertain whether these drugs produced a synergistic ef- fect or not, the following combinations were also studied: ace- tanilid and salol, phenacetin and salol, acetanilid and phenacetin, aspirin and salol, and antipyrin and aspirin. The doses of the drugs never exceeded those employed by conservative therapeutists. SUMMARY OF EXPERIMENTS Owing to the expense of publication, it is impossible in this paper to report in tabular form all the data obtained in the pres- ent investigation. An analysis of the results obtained will there- fore be only given, and a few tables illustrating the method of carrying on the experiments. Tables 1, 2, and 3 exemplify three of the experiments which have been performed. As will be seen, upon examining them, they indicate the drug and the dose em- ployed, the reaction time readings before the administration of the drug, and several series of readings after the drug had been taken. In each case the mean reading (M.), the mean variation (m.v.), and in case of a complex reaction the number of errors, committed are indicated. Furthermore, after the administra- tion of the drug, the relative change in the mean readings and the relative change in the mean variations have been also computed and expressed in terms of percentages of the normal. In table 4, an attempt was made to summarize the results of all the ex- periments. In this table the drugs used and their dosage are indicated and the most striking or maximal effect of the drugs is expressed in terms of percent of the normal reading. Further- more, the time at which that effect was noticed is indicated and in the case of the mathematical problems the run of errors is expressed. 332 D. I. MACHT, S. ISAACS AND J. GREENBERG Greenberg; July 12, TABLE 1 1917, at 12.50; acetanilid — 7 grains SOUND LIGHT TOUCH +17 X3+4 I Normal be- m ... 172 140 156 1486 2504 gan at 12 20 mv 10 8 11 339 791 Errors 3 II Began at m 162 140 168 1356 2180 12 50 then* mv 10 15 11 235 394 dinner Errors 1 Relative change in m . 94% 100% 107% 91% 86% Relative change in mv 100% 187% 100% 67% 49% I After drug Time , 1 hr. 1 hr. 2' 1 hr. 38' 1 hr. 41' 1 hr 44' at 1 50 m 180 154 194 1256 2212 mv 11 9 16 173 533 Errors 4 Relative change in m 104% 110% 135% 87% 88% Relative change in mv . ... 110% 112% 145% 51% 67% II After drug Time 2 hr. 6' 2 hr. 8' 2 hr. 10' 2 hr. 10' 2 hr 15' at 2 56 m 190 140 170 1316 2196 mv 15 21 19 250 494 Errors 3 2 Relative change in m 110% 100% 108% 88% 86% Relative change in mv 150% 271% 172% 73% 62% III After Time 2 hr. 32' 2 hr. 34' 2 hr. 35' 2 hr. 37' 2 hr. 40' drug at 3 22 m 184 142 160 1354 2334 mv 12 11 13 246 727 Errors 1 3 Relative change in m Relative change in mv 107% 120% 101% 150% 102% 111% 91% 72% 93% 91% ACTION OF ANTIPYRETIC ANALGESICS 333 TABLE 2 Dr. Macht; July 19, 1917, at 11.05 a.m.; pyramidon — 6 grains SOUND LIGHT TOUCH +17 X3+4 Before at m. 156 142 138 2088 2816 10.40 a.m. m.v Errors 11 7 10 333 2 692 2 I After drug Time 35' at 11.40 m 158 140 144 1870 2718 m.v. . 11 11 7 232 591 Errors 1 0 Relative change in m 100% 98% 104% 89% 99% Relative change in m v 100% 156% 70% 69% 85% II After drug Time 1 hr 1' at 12 06 m 174 150 152 2028 3168 m v 20 9 17 322 705 Errors 0 2 Relative change in m. 111% 105% 110% 97% 113% Relative change in m.v 181% 128% 170% 96% 110% III After. Time 1 hr. 55' drug at 1.00 m 208 152 152 1968 2658 m v 18 11 11 377 737 Errors. 0 2 Relative change in m 133% 107% 110% 94% 94% Relative change in m.v 163% 157% 110% 113% 106% IV. After Time 3 hr. 4' 3 hr. 7' 3 hr. 11' 3 hr. 13' 3 hr. 17'" drug at 2.09 m ... 170 136 150 2282 2838 (after m.v 17 10 12 426 639 lunch) Errors 3 2 Relative change in m. . 109% 95% 108% 109% 100% Relative change in m.v 154% 142% 120% 128% 92% V. After drug at 3 10 Time m. . 4 hr. 5' 180 4 hr. 8' 140 4 hr. 10' 140 4 hr. 12' 2014 4 hr. 16' 3102 m.v 25 12 13 257 765 Errors 0 0 Relative change in m 115% 98% 101% 96% 114% Relative change in m.v... 229% 145% 130% 77% 110% 334 D. I. MACHT, S. ISAACS AND J. GKEENBERG Isaacs; July 27, 1917, TABLE 3 at 2.58 p.m.; antipyrin—600 mgm. ,SOUND LIGHT TOUCH +17 X3+4 Before m 168 132 144 1484 4786 m v 13 8 9 320 1437 Errors 1 9 I After drug Time 15' 16' 18' 20' 22' at 3 13 m 170 124 140 1412 4994 m v 15 15 15 214 1636 Errors 1 5 Relative change in m 101% 93% 97% 95% 104% Relative change in m.v. 115% 187% 166% 66% 113% II. After drug at 4 12 Time m 1 hr. 14' 162 1 hr. 16' 136 1 hr. 18' 162 1 hr. 20' 1620 1 hr. 23' 5036 m.v Errors 18 12 14 426 3 1786 2 Relative change in m Relative change in m v 96% 136% 103% 150% 112% 155% 109% 133% 105% 124% ANALYSIS A careful study and analysis of all the data obtained has led the authors to the following conclusions. The results obtained with antipyretics are quite different from those found after mor- phin or opium. No primary stage of stimulation or shortened reaction time was noted after administration of antipyretics, ex- cept possibly after small doses of quinin. It was found that in all cases the ordinary doses of antipyretics produced either very little effect on the reaction time, or if affecting it at all, they al- ways impaired it as indicated by the prolongation of the mean readings, by the increase in the mean variations of readings, or by both. The most powerful or depressant drug in this respect was found to be pyramidon. This is not entirely surprising inas- much as pyramidon in the author's (M.) experience and in the experience of many physicians is one of the most efficient anal- ACTION OF ANTIPYRETIC ANALGESICS 335 gesics — its effect coming closer to that of the narcotics than that of most other antipyretics. It was furthermore interesting to note that when the antipyretics exerted an influence on the re- action time, the simple reflexes or reactions to sound, light, and touch were more prolonged or impaired than the more complex association tests. Of the three simple reactions that of touch was more generally retarded than of sound or light. The asso- ciation tests were also depressed or impaired, but usually in a distinctly lesser degree than the simple reactions or reflexes. Thus, for instance, the absolute readings in case of the mathe- matical calculations were sometimes even actually improved and the depressant effect of the drug revealed itself only through the greater number of errors committed. Experiments with combinations of the various antipyretics gave results which could be explained by a simple summation or addition of the individual effects of the components. No so- called synergism or potentiation of one drug by another was observed. The curious difference in the effect of the drugs on the simple reactions as compared with that on the more complex ones is the direct opposite of the results obtained after administration of opium or morphin. In the latter case, the simple reactions were always less impaired than the more complex association t.ests. A comparison of the findings obtained with the two groups of analgesics, the opiates or narcotics, and the antipyretic anal- gesics, seems to point to some lower synapse as the seat of action of the coal tar derivatives. Tfi t 00 CO | 1 CXO ^ t 1 O i— i I-HTH iOCO(M t t t ttt (MOO5OO (M(M(MO T-H tttt t CO t ^ ' 10 I M ^ OLQIO cooodcoiM odoodi— l o OCMOOCN »T3l>.CNO-3d Suip^aj ueara ui aSuiJqo aAi^aw a; ooo oooooo r "*I>.COCOCX| i— i ui 83uBqo CO O i— i !>• -COOOCNIl>'COOtv-dCO(MOl>-^OCOiO?D»O r-HO'-ldd CSOrHdrHdOOOl^OrHr-lOOQOddO uaqAV auijx OOO OOOO UOI^BUBA U 83 UI t>-O^OO r-ltOOO Suip^ai I I; lO'^t1'— iJT—T— lOTfGOOir- I'^r'OsCOCOi— (^Od Oi-^'-ldOdddOOOOCOdddrHO uaqM auiij, C5COO-— IT^TJH-^IO^O'— ii— l CSjTj> b.2 .2 .2 .2 .2 .2 m m g § g iSTSTS^.^oCJa) - - - § S ^ ^ -s . - - - ''''"'' ~"~ ••bb-'bJD-'bb| bio • • bfi • bi) ; ; iilillill^lllllllll lSt2CS«5Ss»2C±'i2S*Sif2?S'as»2 336 2->3->2 T T CO t (M TH 7 TH 1 »o TH 7 TH 1 | o § § oo o o £2 £ oo 8 § § 8 § CO O TH O TH 3 o s £ TH o 28 TH TH TH 00 CO V TH s 0 o TH is TH o o I TH o>o «... 10 10 1O >O (N (N 3 4 5 6 7 8 50 32 40 28 36 58 42 60 54 100 44 58 52 58 54 42 50 56 150 26 44 60 42 56 54 64 56 200 52 52 72 50 54 52 56 50 250 48 52 65 68 60 54 56 64 300 38 44 65 53 64 58 54 50 350 24 52 50 42 52 48 48 52 400 58 48 45 50 57 54 60 64 450 34 42 54 42 44 36 74 48 500 62 48 28 48 48 36 56 56 550 50 56 80 60 60 52 50 34 600 46 56 74 60 60 40 54 56 640 48 47 70 56 50 50 54 58 tion. The controls are therefore not conclusive, although the indication is that tone sensitivity is absent. The following controls were used with rat 3 : Control 1. The noise stimulus ordinarily used only for turning to the left was used for all trials. This substituted a certain "thud" for the "tone and thud" previously used as a cue for turning to the right. No tone therefore was given. All other conditions normal. The re- sponse was counted correct if it coincided with the order of presentation. Control 2. The supports of the tuning fork were fastened to tripods resting upon the floor and in some cases upon two adjacent tables. This control ruled out the possibility of contact cues due to the transmission of vibrations through the table to the discrimination box. Control 3. Cotton was forced between the prongs of the fork. Strik- ing the fork now produced a faint tone of high pitch as well as a noise of new quality. The same stimulus from striking the fork was given in each trial. The reactions were counted as correct if they fitted the series of presentations. Control 1 caused a breakdown in the rat's high percentage of correct responses. When this control was alternated with nor- mal conditions on intervening days, correct responses did not 342 WALTER S. HUNTER exceed 60 per cent, although the normal conditions gave 73 to 80 per cent correct response. Clearly something was eliminated that had served to determine the rats responses. The results, however, do not prove tone sensitivity because the "thud" caused by the free fork was different from the "thud" of the fork when held and may have served as the basis of the original discrimination. Control 2 did not disturb the accuracy of the rat's responses. Control 3 was given one day with an accuracy of 50 per cent; two days, with 70 per cent; and one day each with 60, 50, and 40 per cent. In this control the stimulus used for each trial was the same. The results of this control indicate, therefore, that a high percentage may be made without using an auditory cue. They are also partly in harmony with the view that the original response was influenced by the differences in the noises involved. These three controls have not determined the exact basis of the response during the normal tests. Inasmuch as the percent- age of correct responses was not stable during those tests, and was not particularly high, it seems probable in the light of these controls and of other published work, that rat 3 was not reacting to tone stimulation. ii The experiments on audition to be reported in this section were made in the summer of 1914 and again in the spring of 1915 at the University of Texas. In the course of some correspond- ence, Professor Carr of the University of Chicago described some of the work then being carried out at that university by Mr. Harry Wylie on transfer of training. At that time the present writer was making a study of auditory sensitivity and of habit interference as published in the 1915 and 1917 articles listed above. The methods and apparatus employed were peculiarly adapted to the study of Wylie's method dealing with transfer between sense fields. The method of inter-sense transfer was therefore employed with the result that valuable data bearing upon tone sensitivity were secured. This data was not made public until Wylie presented his report at the 1915 meeting of AUDITORY SENSITIVITY OF THE 'WHITE RAT 343 the American Psychological Association (2), when a verbal report of the present finding was made. The apparatus used was the T-shaped discrimination box of the earlier experiments, figured above. A hole about 1 inch in diameter was cut in the box at the point marked 0 in the figure. Outside of the apparatus and directly in front of this hole was placed a 10 candle power light enclosed in a covered box. This light could be turned off and on as desired. The hole in the box was closed with a square of clear glass over which a layer of bond typewriter paper was placed. The result was a diffused light of good intensity (somewhat less than 10 candle power) . The first rats used were rats 46 and 47 of the 1915 paper. These rats were about six months old at the time of the present observations. They had been trained to discriminate a piston whistle tone of 3906.17 d.v. from silence by running to the right when the tone was given and to the left for silence. Controls had been used which indicated that the rats would substitute for the whistle either: "rush of air" sound made by the whistle with- out the tone; a "rush of air" sound made with the experimenter's lips; or handclapping of medium intensity. In other words, the rat would run to the right when either of the above stimuli were given, and to the left for silence. On the other hand when the following tonal stimuli were substituted for the whistle the rat failed to discriminate them from silence: one 1280 d.v. fork; a 1152 d.v. and a 1280 d.v. fork sounding simultaneously; and the normal whistle when it was removed to a distance, thus eliminat- ing the usual noise accompaniment. Transfer was therefore possible without training from one noise (that of the whistle) to various others; but not possible from the noise in question to tonal stimuli. It was at this point that the suggestion derived from Wylie's work was applied. The rat was placed in the apparatus and stimulated with a light in place of the whistle in order to see if a transfer would be made between the two sense fields in spite of the fact that the previous transfer within the auditory sense field had not occurred. The results for the two animals are given in table 2. The rats were given the regular whistle test in a dark 344 WALTER S. HUNTER room on the first day in order to be certain that the darkness would not disturb the reaction. On the following day the rat was placed in the apparatus with the room dark and the light at 0 turned on. A correct response consisted either of running to the right for the light or to the left for darkness. As the table indi- cates, rat 46 failed to make the substitution, making only 5 < or- rect reactions out of 10 on each of two days with the light. This was mere chance. Rat 47, on the other hand, made the substi- tution on each of three days, making 9, 8, and 9 correct reactions respectively. TABLE 2 Data on light substitution TRIALS TKIALS CORRECT Rat 46: Normal whistle in dark 10 9 Light 10 5 Light 10 5 Normal whistle in dark 10 8 Rat 47: Normal whistle in dark 10 9 Light... 10 9 Light 10 8 Light 10 9 Normal whistle in the dark 10 9 These results indicate clearly a greater similarity, for the be- havior of rat 47, between noise and light than was found between noise and tone. This is harmonious with the apparent fact of insensitivity to the tones as indicated in the earlier papers. Five rats, 1, 7, 9, and 11, had been trained in 1915 to discrimi- nate handclapping from silence, using the same apparatus, by running to the right for handclaps and to the left for silence. (A description of this work is published in the 1917 article.) No direct tests were made on their ability to substitute tones for the noise. It is legitimate to assume that, like all of the many rats heretofore tested, they would have failed. They were tested however upon their ability to substitute a tactual and a visual stimulus for the noise. For the tactual stimulus a piece of coarse AUDITORY SENSITIVITY OF THE WHITE RAT 345 sandpaper was laid in the alley at the point indicated in dotted lines in the figure. (The rat was to run right for the contact and left for its absence.) The visual stimulus was the same used with the two rats above described. The results secured are given in table 3. It will be seen from this that rat 1 failed to substitute the light for the noise. The sandpaper gave 70 per cent correct reactions, which is slightly better than chance but which does not demonstrate clearly that the substitution was made. Rat 7 clearly failed on the sand- paper, but gave slight evidences of substituting the light. The same statements are true of rat 9. Rat 11 failed to substitute the sandpaper. On the other hand, the data on light indicate a clear-cut and all but perfect substitution similar to that in the case of rat 47 described above. It is perhaps a safe conclusion that light and noise are more similar for the behavior of rat 11 than are noise and tone, al-. though no direct tests of tone sensitivity were made. The justi- fication for such a conclusion would rest upon the lack of evi- dence for tone sensitivity in the white rat, and upon the behavior of rat 47 above. With all of these rats careful controls indicated that they were not dependent upon the series of presentations> but were guided, normally, by auditory stimuli. in In the 1917 paper, tests were described in which three rats, 7, 15, and 23, acquired the habit of running to the right for hand- claps and to the left for silence. They were then trained to run to the left for a buzzer and to the right for silence. At the close of this second training, when retested on the first habit, no one of the three fell below 80 per cent correct responses for 30 trials. Rat 27 also acquired the two habits serially. Rats 7 and 27 were continued in the experiments, to be described in this sec- tion, for the purpose of determining additional features underly- ing their reactions. Each of these rats could respond correctly in either of the two following three ways: run right for handclaps (h.c.) and left for 346 WALTER S. HUNTER TABLE 3 Data on the substitution of light and sandpaper CONDITION OF TEST NUMBER OF TRIALS NUMBER CORRECT Normal handclaps 10 Sandpaper 10 Handclaps 10 Sandpaper 10 Sandpaper „ 10 Sandpaper 10 Sandpaper 10 Handclaps 20 Light , 10 Handclaps 10 Handclaps 10 Handclaps 20 Light 10 Handclaps 10 Handclaps 20 Normal handclaps 20 Sandpaper 10 Handclaps 10 Sandpaper 10 Handclaps 10 Handclaps 10 Sandpaper 10 7 •{ Handclaps 30 Light 10 Handclaps 10 Light 10 Handclaps 10 Handclaps 10 Light 10 Handclaps 20 Normal handclaps 20 Sandpaper 10 Handclaps 10 Sandpaper 10 Handclaps 10 , Handclaps 10 Sandpaper 10 Handclaps 20 Handclaps 10 Light 10 Handclaps 10 Handclaps 20 7 8 8 7 7 7 17 5 5 7 18 6 6 16 20 5 10 6 7 8 6 26 7 10 5 7 8 7 16 16 5 10 5 6 10 5 14 8 7 10 17 AUDITORY SENSITIVITY OF THE WHITE RAT 347 TABLE ^—Continued CONDITION OP TKST NUMBER OP TRIALS NUMBER CORRECT f Light 10 J Handclaps 10 Light 10 [ Handclaps 10 Normal handclaps 10 Sandpaper 10 Handclaps 10 Sandpaper 10 Sandpaper -. . . . 10 Sandpaper 10 . Handclaps 10 Sandpaper 10 Handclaps 30 Light 10 Handclaps 10 Light 10 Handclaps 10 Light 10 8 8 6 8 8 5 8 8 6 6 8 5 24 10 10 9 9 10 silence; run left for buzzer and right for silence; or run right for h.c. and left for buzzer. Each day's work was begun with either of the auditory stimuli. This determined the right or left turn and silence was then treated accordingly. Controls were now instituted in an attempt to determine just wherein the difference between the two noises lay. Illness in the animals finally pre- vented an entirely satisfactory solution. The data are presented in their present fragmentary form because it will probably rarely happen that an experimenter will have rats possessed of these two habits, thus making possible the extension of the work. The work on auditory sensitivity in the white rat has indicated an insensitivity to tones in the lower portion of the scale and also to some as high as the pitch 4096 d.v. on the fork. It has also demonstrated that the rat can hear noises. In the case of rats 7 and 27, the opportunity was offered to extend these previous observations on noise. Are noises for the rat grouped into the three classes of continuous, intermittent and beat noises as Hensen claims for man? The buzzer used in the tests with rats 7 and 27 gave a continuous sound and the handclaps constituted an inter- 348 WALTER S. HUNTER mittent noise of about 150 interruptions per minute. The ex- perimenter had been using handclaps for a noise stimulus for many months and had reduced the clapping to an automatism which varied but slightly in rate. The following are the controls used in the analysis of the two habits described. (In the formation of the habits, punishment and reward were used. In the controls, punishment was never used for fear of breaking up the association.) K = turn right for h.c. and left for buzzer. h.c. = turn right for h.c. and left for silence. B = turn left for buzzer and right for silence. Control 1. Hissing through the teeth, turn left. Control 2. Hissing through teeth, turn right. Control 3. "Rush of air" sound with lips, turn right. Control 4. "Rush of air" sound with lips, turn left. Control 5. Buzzer sounded for two seconds, at intervals of one sec- ond, turn left. Control 6. Buzzer sounded normally, turn right. Control 7. Metronome beating 120 per minute, turn right. Control 8. Metronome beating 120 per minute, turn left. Control 9. Metronome beating 200 per minute, turn right. Control 10. Metronome beating 200 per minute, turn left. Control 11. Metronome beating 160 per minute, turn right. Control 12. Metronome beating 176 per minute, turn right. Control 13. Motor knocking 180 per minute, turn right. Control 14. Motor knocking 232 per minute, turn left. Control 15. Motor knocking 140 per minute, turn left. Control 16. Motor knocking 240 per minute, turn right. Control 17. Motor knocking 250 per minute, turn left. Control 18. Motor knocking 500 per minute, turn left. Control 19. Beats 120 per minute with 512 v.s. forks, turn right. Control 20. Beats 142 per minute with 512 v.s. forks, turn right. Control 21. Beats 174 per minute with 512 v.s. forks, turn right. Control 22. Beats 3840 per minute with 512 v.s. and 576 v.s. forks, turn left. Not all of these controls could be used with each rat, and the limitations of time prevented their being exhaustively applied. The buzzer used was an ordinary commercial one and was sus- AUDITORY SENSITIVITY OF THE WHITE RAT 349 pended above the apparatus in a manner which prevented the transmission of vibrations directly to the box. When the metro- nome was used, it was placed upon a ledge just to the back and above the apparatus and entirely separated from direct connec- tion with the experimental table. The forks used (mounted upon resonance boxes) were suspended in the same manner as the buzzer. In controls 13 to 18 a strip of metal was fastened to the spindle of a motor in such a manner that the rotations of the motor caused it to strike at a definite rate against a card- board. This gave an excellent noise of medium intensity and of volume and quality different from the handclaps. The motor (a silent one) was placed upon a nearby table. The results secured are summarized in table 4. Rat 7, who was tested with the beats from the forks, ignored them and would not class them either with the continuous or with the inter- rupted noises. Hissing and the "rush of air" sound were sub- stituted readily for the buzzer, i.e.j at least 80 per cent correct reactions were made going to the left for these noises and to the right for silence. The metronome beating at 120 and at 176 were substituted for the handclap successfully; whereas the rat persistently refused to run to the right for the metronome at 200 and to the left for silence. Two 512 v.s. forts mistuned to give beats of 142 and 174 per minute were ignored as were the beats from the forks 512 v.s. and 576 v.s. sounding together. The evidence suggests that the rat was insensitive to the beats, which is very interesting inasmuch as beats are periodic variations in inten- sity and the rats seem insensitive to tone. The evidence also indi- cates that the transition between continuous and intermittent noises occurs in the region of 200 interruptions per minute. Rat 27 ran to the left for hissing, and for the motor knocking at the rate of 500 per minute. It refused to go to the left for the motor knocking at .the rate of 232 per minute and for the metro- nome at 200 per minute (right for silence). The motor rate of 250 was neither clearly rejected from nor clearly included with the continuous sounds. The rates of 120 and 180 per minute were definitely substituted for the handclaps, whereas hissing, the buzzer, and the metronome at 160, 200 and the motor at 240 were TABLE 4 Controls used with rats in analysis of buzzer-h.c. habit RAT 7 RAT 27 Control Number correct in 10 Control Number correct in 10 K 9 K 9 1 9 2 7 5 3 of 7 1 7 K 18 of 20 B 6 5 9 B 9 3 4 1 8 4 9 B 8 7 8 2 6 9 5 B 7 K 9 K 7 7 8 K 9 9 5 K 35 of 40 K 9 1 8 K 10 1 8 20 6 2 4 7 7 B 8 7 8 7 8 9 4 9 4 K 10 6 4 20 6 16 5 19 7 17 10 K 10 17 7 19 6 9 7 19 5 K 10 19 5 9 8 K 9 10 6 12 9 K 7 9 5 K 9 K 10 h.c. 16 of 20 22 5 11 5 22 4 11 6 B 8 11 19 of 30 9 5 B 9 10 6 h.c. 7 . h.c. 10 9 6 • B 7 h.c. 9 1 9 9 4 h.c. 16 of 20 9 6 h.c. 9 11 7 h.c. 18 of 20 9 6 h.c. 9 9 31 of 50 350 AUDITORY SENSITIVITY OF THE WHITE RAT 351 clearly rejected from the class of interrupted noises. Because of the rat's refusal to class the metronome at 160 with the intermittent noises, the results are not so clear cut as might be desired. However, as with rat 7, the indication is that the frequencies around 200 are the critical ones. It seems probable that the discriminations here made by the two animals were discriminations of " qualities" rather than of "rates." The difference between the noise group of hissing, buzzing and "rush of air" and the noise group of handclaps and slow metronome beats is the qualitative difference of continu- ous and intermittent pointed out by Hensen. The indications are, therefore, that for the white rat there are two classes of noises — beats are apparently not sensed. REFERENCES (1) HUNTER, W. S. : The auditory sensitivity of the white rat. Jr. Animal Behav., 1914, 4, 215-222; and 1915, v, 312-329. HUNTER, W. S., and YABBROUGH, Jos. U. : The interference of auditory habits in the white rat. Jr. Animal Behav., 1917, vii, 49-65. BARBER, A. G. : The localization of sound in the white rat. Jr. Animal Behav., 1915, v, 292-311: (2) WYLIE, H. H. : Some experiments on transfer of training. Psych. Bull., 1916, xiii, 78-79. A SIMPLE MAZE : WITH DATA ON THE RELATION OF THE DISTRIBUTION OF PRACTICE TO THE RATE OF LEARNING K. S. LASHLEY The Johns Hopkins University and the University of Minnesota I. A TEST OF THE VALIDITY OF RESULTS OBTAINED WITH THE SIMPLE MAZE The use of complicated mazes for the study of learning in ani- mals originated in an effort to determine the limits of educability of the lower mammals. The problem of habit-formation in animals soon expanded, however, to include the more general problem of the mechanism of learning and this led to a search for a method of comparing individual differences in learning ability and average differences for groups trained under diverse con- ditions. The application of the complicated maze to the deter- mination of individual differences followed as a matter of course, since it provided a practicable technique for the train- ing of animals. But in none of the pioneer work, nor, indeed, in any of the studies thus far recorded, has there been a thorough test of the applicability of the maze technique to the particular problems studied by its aid. This is true also of the various problem boxes and other apparatus used for training animals in complex habits. The study of habit formation in animals has now advanced to a stage where accurate work is possible and where statistical methods can and should be employed for the evaluation of data. Many problems require the comparison of the rates of learning of numerous animals trained under diverse conditions, as in studies of the action of drugs. If the animals are trained in complex habits the experiment frequently requires an expenditure of time out of proportion to the results obtained, and the necessary 353 354 K. S. LASHLEY restriction of the data to a few subjects leaves the experimental results inconclusive. A simpler technique is therefore desirable; one by which the subjects may be trained rapidly and data gathered upon a large number within a reasonable time. The training of larger numbers might, perhaps, be accomplished by the use of automatic training and recording apparatus but the appara- tus of this character that has been devised is bulky, expensive, and not altogether dependable. Further, many animals seem to be much more disturbed by mechanical contrivances in the train- ing box than by manipulation in the hands of the experimenter and the preliminary training necessary to accustom them to moving doors, and the like, nearly counterbalances the advan- tages of automatic training. The alternative method for gather- ing larger amounts of data involves the training of the animals in some easily acquired habit for which no great expenditure of tune will be required by any one animal. A widely accepted objection to the use of simpler habits arises, however, from the possibility that these may fail to demonstrate individual differences which would appear in more complex forms of learning. This is perhaps true where the primary interest of the study is in the comparative intelligence of the animals but where the problem is primarily that of the mechanism of learning, the nervous changes involved in the reintegration of conduction paths, the phenomena resulting from the complexities of the habit only serve to confuse the data and make it impossible to distin- guish between the effects of simple reintegration and those result- ing from the simultaneous formation of a number of habits. This confusion is illustrated by the results of studies of the effects of the distribution of practice upon the rate of learning. Such results, as I have shown in previous articles (Lashley, '15 and' 17) are due at least in part to the interference of simultaneously formed habits and, as the present study indicates, are probably always due to the complexity of the habit and not to any funda- mental character of the learning process. The failure of a simple habit to reveal a difference which appeared when more complex habits were studied should indicate, therefore, that the difference found was due to some factor introduced by the formation of A SIMPLE MAZE 355 several habits simultaneously, rather than that training in the simple habit failed to reveal any fundamental differences in the processes of learning. In view of the need for a simpler technique it seemed that a study of the adaptability of some simple habit to a comparative study of the rate of learning in diverse groups of animals would be of value. Since the rat is the animal most studied and perhaps best fitted for laboratory study it was chosen for a test of tech- nique. Various attempts were made to adapt the methods of study of the conditioned reflex to the rat, but without much FIG. 1. THE SIMPLE MAZE a, Starting compartment; c, cul de sac; d, food compartment; e, position of food dish. success, and finally a simple maze was adopted as offering a pos- sible technique for comparative work. It was modeled after the Yerkes discrimination box, offering a choice between a single cul de sac and an alley leading to food (fig. 1). The questions to be answered by the study were: When dif- ferentially treated animals learn the simple maze are differences in the rate of learning as evident as in the formation of complex habits? If such differences appear, are they of the same character as those revealed in complex habits? How great an economy of the experimenter's time is made possible by the use of the simple maze? PSYCHOBIOLOGY VOL. I, NO. 5 356 K. S. LASHLEY It seemed advisable to use some conditions of training which have been tested already in complicated mazes, and of these the use of practice periods of diverse length was selected as a condition whose general results are well established and whose details are capable of wide variation. Twenty-five rats were trained in the simple maze with 10 trials per day and 24 with 2 trials per day. Entering the cul de sac or turning back along the true path was counted as an error and ten successive trials without error were required as a criterion of learning. The general precautions to secure uniformity in the groups compared were those which I have described in an earlief study of the circular maze (Lashley, '17b). The conditions or feeding were varied somewhat within the groups: details of this will be considered in the second part of this paper. The average number of trials required for learning by the ani- mals having 2 trials per day was 21.5 ±0.81. The average number for the animals having 10 trials per day was 57.8 ±4.70. The difference is 36.3 ±4.76 trials in favor of the group practicing with fewer trials per day; a saving of about 60 per cent as a result of the distributed practice. This result is in accord with the data obtained by Ulrich ('15) in his study with the circular maze and indicates that the simple maze is as well adapted to bring out group differences of this character as is the more complicated apparatus. When the animals had learned the maze they were used in operative experiments. Portions of the cerebral cortex were destroyed and retention of the maze habit was tested after the operation. In this type of work also the simple maze proved useful (Lashley and Franz, ;17). With respect to the time consumed in training, the simple maze offers a great advantage over the more complicated apparatus. The average time consumed by each of the 49 animals in actual practice in the maze was 7.4 minutes. The average time required in practice by a group of 32 normal animals trained previously in the circular maze was 47.1 minutes. That is, more than six times as much time must be spent in training animals in the cir- cular maze as in the simple maze, yet the results of the former method are in no way more reliable than those of the latter. A SIMPLE MAZE 357 The saving of time is probably of less importance than these figures suggest, for a great deal of time is consumed in placing the animals in the starting compartment, in waiting for them to enter the maze, in recording their behavior, and in caring for them between training periods, so that the time spent in actual training is only a small per cent of that demanded by experi- mental work. Even so, the item of 40 minutes saved in the training of each animal is not be to neglected. It seems then that the simple maze offers a dependable method of comparing individual differences in learning ability, not inferior to that provided by training in more complex habits, and that it affords a significant economy of the experimenter's time. II. DATA ON THE EFFECTS OF THE TEMPORAL DISTRIBUTION OF PRACTICE The relative simplicity of the habits that must be formed for the accurate running of the simple maze has made it possible to distinguish certain types of behavior which seem correlated with the effects of the temporal distribution of practice and may con- tribute something to our knowledge of the way in which long practice periods retard habit-formation in the rat. The experi- ment was originally planned to test the effects of a variety of methods of training other than variations in the distribution of practice. The unexpected result appeared that the diverse methods of training produced different effects, conditioned by the distribution of practice, and it is in these different effects that the chief interest of the experiment lies. For training the animals were divided into four groups with which diverse methods of training were used. With one of the groups (A) the customary methods of training were employed. Food was placed in the end of one alley (fig. 1, e). The animals were confined in the end of this alley and fed there on three days preceding training. During training they were allowed to reach the food at every trial; that is, to correct errors made in the cul de sac. 358 K. S. LASHLEY The second group (B) received similar training except that the animals were allowed to explore the entire maze, without food, for 20 minutes on the day before training was begun. The third group (C) was treated like the first except that the animals were not allowed to correct their errors but were returned immediately to the starting compartment if they entered the cul de sac or turned back along the correct path. Each time that they were returned to the starting compartment a trial was re- corded, whether or not they had reached the food. The fourth group (D) was trained like the first except that a dish of food, covered with fine wire netting so that the food could not be reached, was placed in the end of the cul de sac. Each of these groups was subdivided into two, one of which received ten, the other two trials per day. A summary of the groups is given in table 1. TABLE 1 Conditions of training for animals tested in the simple maze GBOUP VARIATIONS FROM CUSTOMARY METHODS OF TRAINING NUMBER OF ANIMALS Two trials per day Ten trials per day A B C D Normal 10 5 5 5 9 5 5 5 Preliminary exploration Errors not corrected Screened food in cul de sac Training was continued until the animals made ten successive trials without error. In some cases training was continued be- yond this point for as much as 200 trials. None of the animals so trained made more than 5 per cent of errors in these later trials, so that the criterion may be looked upon as indicating practically perfect learning. The average numbers of trials per day required by the different groups are given in table 2. Under all the various conditions of training the animals which were trained with only two trials per day learned more rapidly, per unit of practice, than those which were trained with ten trials per day. The average difference for all the groups resulting from the different distributions of practice A SIMPLE MAZE 359 TABLE 2 The average number of trials required for learning the simple maze by animals trained under diverse conditions GROUP AVERAGE TRIALS AT 2 PER DAT NUMBER OF RATS AVERAGE TRIALS AT 10 PER DAY NUMBER OF RATS DIFFERENCE (TRIALS) A 21.1 0.8 10 51.7 4.3 9 30.6 4.4 B 19.2 1.9 5 29.4 3.6 5 10.2 4.1 C 22.6 1.7 5 44.6 2.7 5 22.0 3.2 D 25.2 2.4 5 110.4 9.7 5 85.2 9.9 All 21.5 0.8 25 57.8 4.7 24 36.3 4.8 is 36. 3 ±4.8, and as this is nearly eight times its probable error indicates a real effect of the distribution of practice. But among the groups trained by different methods the effects of distribution of practice vary enormously, from a minimum of 53 per cent to a maximum of 338 per cent increase in practice required for learning following equal lengthening of the practice periods. The relatively slight individual variation within the groups, indicated by the small probable errors, makes the dif- ferences significant. The methods of training most effective in producing variations from the rate of learning determined in the customary way were preliminary exploration of the maze without food and presence of screened food in the cul de sac. The significant differences found are : TWO TRIALS PER DAT TEN TRIALS PER DAT A-B 1.9 4.1 2.0 2.5 22.3 58.7 5.6 10:6 D-A (C and A are practically identical.) The influence of these different methods of training was in the same direction for both concentrated and distributed practice, but only in concentrated practice were marked effects produced. The simplicity of the movements involved in traversing the maze made it possible to record the behavior of the animals in detail and to distinguish characteristic differences in the behavior of the different groups which seem to be correlated with the rate 360 K. S. LASHLEY of learning. The variations in behavior relate chiefly to two instinctive modes of response to the maze problem. 1. When given food in a somewhat unfamiliar environment the rat will almost invariably explore the neighborhood of the food be- fore eating. In the circular maze the exploration usually includes the food compartment and the alley surrounding it. If the animal is not restrained after reaching the food it is almost certain to go through this exploration on the first trial of each day's practice. Once the exploration is completed the animal will go directly to the food in the succeeding trials. The same tendency appears in the simple maze with an extension of the area explored to include almost the entire maze. This is shown by the following analysis of the path followed by the animals during their first and second trials in the simple maze. The data are taken from 52 animals with which training was begun in the simple maze. Three of these were discarded be- cause of illness; the remaining 49 are those described above. Of the 52, 25 avoided the cul de sac on the first trial. This is one less than the expectation from chance, since only two alternative paths were offered. Of these 25, chance should have led one half to enter the cul de sac and one half to avoid it on the second trial and, if the principle of recency were an important factor in learn- ing, more than half should have gone directly to the food. In- stead of this, however, 17 of the animals entered the cul de sac on the second trial and only eight went directly to the food. This result seems to indicate an instinctive tendency to varied activity, or to a thorough exploration of the environment. The tendency to explore the maze becomes much more pro- nounced if the animals are frightened at any time during training and this leads to additional errors and an apparently increased learning time. 2. The second factor of importance in prolonging the learning process in concentrated practice is emotional disturbance in the food compartment. Unless the animals have been handled a great deal they will give avoiding reactions when the experi- menter attempts to pick them up and readily learn to avoid places where they have been caught. The following behavior is A SIMPLE MAZE 361 typical of the animals in the groups which showed marked re- tardation in concentrated practice. The animal reaches the food and begins to eat. The experi- menter puts his hand into the food compartment to transfer the rat to the starting box. The rat retreats and is caught. On the next trial he advances toward the food, pauses, extends his head upward toward the place whence the hand came before, makes several false starts, and finally advances timidly to the food, giving evidence of increased tonus and readiness for flight. Under the usual conditions of training (group A) this behavior rarely appears on the first trial, becomes marked on the third to fifth, and disappears by the end of the practice period. It seems as though several trials were required to set up the association be- tween the food compartment and the avoiding reaction, in each day's practice, and several more to fatigue the conditioned emo- tional reflex so established. This is shown clearly by the follow- ing average tunes required for successive trials on the second day of training by the animals in group A. Trial 123456789 10 Seconds : 18 22 57 37 28 15 13 13 17 16 The increased time in the second to fifth trials is characteristic of many of the records and is correlated with the flight reactions near the food compartment. The chief importance of these two factors in favoring distributed practice lies in the fact that their effects are shown in the later trials of each day's practice. When only two trials are given there is not time for the summation of the exploratory and flight impulses and so fewer errors due to these causes are made. That they were important factors in determining the effects of dis- tribution of practice is shown by the differences in the behavior of the animals in the three groups which differed in the number of trials required for learning. The rats in group B, which had been allowed to become thoroughly familiar with the maze, rarely gave the flight reactions in the food compartment and there was slight indication of a summation of emotional disturb- ance such as appeared in group A. Familiarity with the maze 362 K. S. LASHLEY reduced the tendency to emotional disturbance. This is shown by the following averages times for successive trials in group B (ten trials per day) on the second day of practice. Trial 123456789 10 Seconds 38 22 20 17 22 8 5 5 5 13 The time of successive trials falls steadily throughout the practice period. The rats in group D, on the contrary, showed a much more marked tendency to flight reactions in the food compart- ment than did those in the other groups. They would frequently advance to the food dish, then turn, run quickly into the cul de sac, and gnaw for some time at the wire cover of the food dish there. Apparently owing to the presence of the screened food in the cul de sac, as offering the stimulus of food without associ- ated handling, these animals frequently persisted in their avoid- ance of the food compartment and in their efforts to get at the food in the cul de sac throughout the practice period. This is shown by the following average times for successive trials in the second day's practice. Trial 123456789 10 Seconds 14 34 33 11 27 22 35 20 36 154 The prolonged time toward the end of the day's practice is char- acteristic of the early days of practice of this group. The summation of the exploratory impulse and of emotional disturbance thus seems to be characteristic of the groups showing marked retardation of learning in concentrated practice. Fur- ther, the extent to which such summation occurred varied with the different conditions of training and resulted in corresponding variations in the amount of retardation. It thus seems clear that a great part of the retardation resulting from concentrated practice is due to this summation of instinctive reactions. Even in group B, which showed the least loss of efficiency in ten trials practice per day, there was some indication of a tendency for the animals to avoid the place where they had been caught and it seems probable that a large part of the retardation shown by this group is due also to the interfering effect of this emotional factor. The A SIMPLE MAZE 363 greater retardation shown by the other groups (in excess of 53 per cent) is due, almost certainly, to this one factor of summation of instinctive reactions. III. THE BEARING OF THE RESULTS UPON THEORIES OF THE NEUROLOGICAL BASIS OF LEARNING Experimental work upon the effects of the distribution of effort in learning has given uniform results for practically every proc- ess studied. Within limits as yet undetermined concentrated practice is less efficient than distributed. But no satisfactory explanation of this seemingly universal phenomenon has yet been advanced. In an earlier paper (Lashley, '15) I have listed seven different possible explanations for the superiority of distributed practice found in archery, between which it is not possible to choose on the basis of the existing evidence, and the list then given was certainly not exhaustive. The universality of the phenomenon might be taken to indicate that it is due to some fundamental process in the formation of new functional connections in the nervous system and this is the view which seems to be most generally held. For example, Starch says ('12) : "Why are shorter and more numerous periods economical? The main reason, no doubt, is the well known fact that a period of rest after newly formed associations gives them a chance to become settled and fixed." Colvin ('11) makes much of this hypothesis, also, and gives it the rank of a general law that "it takes a certain amount of time for associations to fix themselves," and this is used to explain not only the effects of distribution of practice but also retroactive inhibition and the facts included under Jost's law. From the standpoint of neurological theory the truth or falsity of this hypothesis is of extreme importance. If there is a gradual strengthening of associations during periods of non-practice there is implied a continuation of chemical changes within the nerve cells, initiated by the passage of a neural impulse through new channels and persisting for hours or even days without the in- fluence of continued impulses. If, on the contrary, no such 364 K. S. LASHLEY gradual fixation occurs, the problem of the neuro-chemistry of learning is simplified by admission of the hypothesis that the effects of the passage of the neural impulse upon later conduc- tivity are direct and immediate. This hypothesis is more in accord with such facts as are known concerning the alteration of conductivity in regions of decrement (Lucas, '17), where the learning process may, perhaps, be located ultimately, and with the generally established facts of the deterioration of function through disuse. The experimental evidence upon which the belief in a gradual fixation of associations is based is far from convincing. It con- sists primarily of the facts expressed in Jost's law, of occasional records of improvement in complex functions during periods of non-practice, and of the data upon the effects of distribution of practice. All of this can be explained equally well by other hypotheses and, in view of the extreme importance of the point for physiological explanation, we should be careful not to accept the assumption of a 'gradual setting' of new functional connec- tions until some real evidence is advanced in support of it. In studies of the mechanism of learning the processes of adjust- ment and of fixation must be distinguished as absolutely inde- pendent variables. The former is, in lower animals and prob- ably in primates also, solely the production of varied random activity, through which the first adjustment to a new situation is brought about; the latter is a process by which the recurrence of certain of the random activities in future trials is rendered more probable. Slow improvement in any function may result either from difficulty in discovering efficient methods of per- formance or from failure to fix as habits the methods which have been hit upon by chance-. In an earlier paper (Lashley, '17a) I have shown that a part of the superiority of short over long practice periods is due to the fact that distributed practice per- mits of greater variability of response and hence greater probabil- ity of discovering new and successful modes of attacking the problem, than does concentrated practice. The influence of the distribution of practice is here exerted upon the process of adjust- ment to the new situation and not upon that of fixation of the random acts. A SIMPLE MAZE 365 In the study cited it was not possible to determine the extent to which this one factor was responsible for the effects of the dis- tribution of practice. The present study makes it possible to estimate the importance of conflicting habits under different distributions of practice a little more accurately, although the conflicting reactions are of a somewhat different character from those dealt with in the first study. The retardation in concen- trated practice in group D amounted to 338 per cent, that in group B to only 53 per cent of the effort required for learning in distributed practice. The difference, 285 per cent, is clearly due to conflicting habits which affect the efficiency of performance and not the formation of associations. The animals of group B gave some evidence of avoiding reactions in the neighborhood of the food dish, and since a slight exaggeration of this reaction was able to increase the retarding effects of concentrated practice to 145 per cent in group A, we are justified in assuming that a large part of the 53 per cent retardation found in group B is due to the same factors which were effective in groups A and D. When allowance is made for the influence of stereotyped reactions upon random activity and for the establishment of habits which actively interfere with efficient performance, there may yet remain a slight reduction in efficiency in concentrated practice which is due to the influence of the distribution of effort upon the process of fixation of new functional nervous connec- tions. Such a remainder can be demonstrated only by a proc- ess of elimination of agents which modify behavior in problem- solving or efficiency of performance and this can be done only by a more complete control of experimental methods than has yet been undertaken. In the existing evidence there is no reason for the belief that any such remainder will be found, at least in maze-learning. The agents acting upon other processes than those of fixation are adequate 'to account for all the effects of concentrated practice revealed by experiments and there is no foundation for the assumption that there is any "gradual process of fixation" which is influenced by the distribution of practice. Whether or not the concept is more applicable to other forms of learning, such as those involving implicit activity, is a matter for 366 K. S. LASHLEY experimental investigation. The existing studies demonstrate the superior efficiency of distributed practice but give no clue to the reason for it. The inferences drawn from studies of the rat in the maze can not be extended to embrace learning of other types without further study. But the studies with the maze are the only ones in which definite evidence as to the mode of action of the dis- tribution of practice has been obtained. The experiments re- ported here and in the previous study seem to demonstrate that the greater part, if not all, of the effect of concentrated practice in maze-learning is due to the development of habits which inter- fere with efficient performance, either by limiting trial movements or by causing actual avoidance of the correct path. In both cases the interference is with the process of adjustment and not with that of fixation. In other forms of learning there are many agents which such as fatigue and loss of interest in long practice periods which may interfere with efficient performance and so prolong the apparent learning tune, while with verbal habits the possibility of practice outside of the experimental practice periods has not been alto- gether eliminated where short practice periods were used. There is thus a possibility that in all forms of learning the results of distribution of practice are due, not to any fundamental principle in the fixation of nervous integrations, but to wholly incidental factors arising from the particular conditions of training in each case studied. The evidence obtained with the maze lends some probability to this view; sufficient, at least, to preclude the use of such a blanket explanation as the gradual ' set ting' of new ner- vous connection before the influence of other factors has been investigated. SUMMARY 1. The simple maze, including a single cul de sac, provides as reliable an index to the rate of fixation of habit in differentially treated groups of animals as does the more complicated circular maze. A SIMPLE MAZE 367 2. The use of the simple maze makes it possible to train larger numbers of animals and so gain a better control of individual variations. 3. In the formation of the maze-habit distributed practice is more efficient than concentrated. 4. This is due to factors which arise from the particular methods of training used, peculiar to the maze problem, and not to the influence of the time relations upon the process of fixation of new functional nervous connections. 5. The same is probably true of all cases where the distribution of effort has been found to influence the rate of learning. There is no reliable evidence for a gradual "setting" of the nervous connections formed during learning. REFERENCES COLVIN, S. S.: The learning process. New York, 1911. LASHLEY, K. S. : The acquisition of skill in archery. Carnegie Inst. of Washing- ton. Pub. no 211, 1915, 105-128. LASHLEY, K. S. : A causal factor in the relation of the distribution of practice to the rate of learning. Journ. Animal behav., 1917a, vii, 139-142. LASHLEY, K. S. : The effects of strychnine and caffeine upon the rate of learning. Psychobiology, 1917b, i, 141-169. LASHLEY, K. S. : AND FRANZ, S. I. : The effects of cerebral destruction upon habit- formation and retention in the albino rat. Psychobiology, 1917, i, 71-139. LUCAS, KEITH: The conduction of the nervous impulse. London, 1917. STARCH, D. : Periods of work in learning. Jour. Educational Psychol., 1912, iii, 209-213. ULRICH, J. L. : Distribution of effort in learning in the white rat. Behav. Monogr. 1915, ii, no. 10, 1-51. METHODS OF STUDYING CONTROLLED WORD ASSOCIATIONS MILDRED WEST LORING From the Psychological Laboratory of the Johns Hopkins University [The work reported here by Dr. Loring supplies the necessary ground for a series of investigations in which it is proposed to use the asso- ciation method, and especially the method of controlled associative recall for the study of psychobiological problems. In addition to the data presented in this paper, Dr. Loring has accumulated a magazine of 10,888 words which have been selected on an experimental basis and classified, and from which future selections for any desired type of work may be made with great efficiency. Such a comprehensive list is essential if working lists are to be selected in scientific way, avoiding the associative sequences and preferences of the compiler. In accumula- ting and testing this list a labor of considerable magnitude has been performed, by methods which can scarcely be improved upon. The experimental work reported was done in the light of a compre- hensive survey of the literature of the association reaction from Francis Galton's in 1879 to authors of 1916, in which year the paper was com- pleted. The historical survey and word lists are omitted from present publication, but the bibliography is included.] Early in the history of the word association method it was ob- served that this method had distinct possibilities for practical application, especially in pathological fields, both because of its simplicity and because of its intrinsic diagnostic character. Very soon it began to be applied to the insane, the feeble-minded, and the delinquent with the hope that different types of association could be determined for each of these and their subgroups, and so provide one more aid to a critical diagnosis. The method was later used with great success in the detection of guilty knowledge. And with normal subjects there has been some study of the effect 369 PSYCHOBIOLOQY, VOL. I, NO. 6 370 MILDRED WEST LORING of age, sex, environment and the like on the character of the associations, and lengthy efforts on the part of many authors to find satisfying classifications of the associations. But it will be noted though that comparatively little investigation has been done where the chief end in view was an examination into the actual technic of the method. The work of some of the earlier German investigators, it is true, does show considerable effort to regulate the length and type of stimulus words, but many have neglected these factors entirely. Many, too, have been content to use the stop watch to measure reaction time, and it may be for this reason that important variations in reaction time with differ- ent types of stimulus words and controls have been overlooked, inasmuch as the watch method introduces into the total reac- tion time three reaction times instead of one. The unrelia- bility of the Hipp chronoscope without considerable modification and standardization, and its operating difficulties, doubtless caused many to abandon a careful determination of the reaction time. A second fact which the history of the word association method discloses is that most of the conclusions have been drawn from free associations. In some of the early work continuous asso- ciations were studied but the obvious disadvantages of this method prevented its extended use. The controlled word asso- ciation method, on the other hand, has had little investigation, except by some of the German investigators, probably because of the prevalent opinion that free associations represent more truly the natural course of ideas in the individual. And yet the controlled word association method has an advantage worthy of consideration. Because the free association method leads necessarily to a heterogeneous mixture of types of response, nearly every investigator has been forced to expend a great deal of effort in classifying the responses. Nevertheless they have reached no uniformity in their classifications. This difficulty is avoided in the controlled word association method, where the stimulus words are uniform, and the response words by virtue of the instruction to the subject are likewise uniform, so that the problem of classification is obviated. STUDYING CONTROLLED WORD ASSOCIATIONS 371 The purpose of this investigation then was to consider the fol- lowing points in the procedure of the controlled word association method, to determine to what extent these factors must be taken into account in a precise use of this method. 1. Does the length of the reaction time in controlled associa- tions differ for the normal and inverse order of nouns and adjectives in the English language? That is, will the reaction time be longer or shorter when the stimulus words are adjectives to be responded to with nouns, than for the opposite, when the stimulus words are nouns to be responded to with adjectives? Is there any such relation between the verb-object and the verb- subject associations? 2. Does the length of the reaction time vary systematically, if the stimulus words are nouns and the response words are adjec- ^ * tives, according to the logical categories in which these nouns fall? 3. Does the length of the reaction time vary according to the number of syllables in the stimulus word, for the adjective-noun, noun-adjective, verb-object, and verb-subject associations? 4. Does the length of the reaction time vary according to the (' ; position of the accent in the stimulus word for the adjective- noun and noun-adjective associations? 5. If the same lists of adjectives and noun stimulus words are given for three successive days, requiring noun and adjective re- actions, respectively, and with no additional instruction, will there be a systematic change in the reaction tune from day to day? 6. Is there any variation in the length of the reaction time for normal and inverse directions of controlled double associations? If the stimulus word is an adjective, for an association to be made first to a noun (not spoken) and then to a verb as response word with this noun as its subject, will the reaction time be longer or shorter than when the control is reversed and the stimulus word is a verb to associate back through a noun subject to an adjec- tive modifying this subject? A consideration of these questions will indicate whether such factors are important in a precise use of the controlled word association method. Without any apparent knowledge of their r61e many investigators have drawn conclusions from variations 372 MILDRED WEST LORING in reaction time which might well be within the normal limits of variation for the kind of stimulus word used, the length of the word, the position of its accent, and the type of control required. APPARATUS The Johns Hopkins chronoscope, designed by Dr. Dunlap, was used in this experiment to record the reaction times, and in con- nection with it Dunlap voice keys of the small model (97). The auditory method of presentation was employed, both stimulus and response words being spoken. The chronoscope is essentially a synchronous motor driven by a 50 D. V. tuning fork. The motor has 10 poles so that the armature rotates five times per second. Attached to the shaft of the armature is an electro- magnet which rotates with the shaft; anterior to this magnet is a fixed magnet facing the rotating one. A light soft iron disc lies between the two magnets, attached at its center to a light shaft perpendicular to it which passes through a brass bearing to the anterior face of the clock, where it is attached to the index- hand. This sliding shaft moves back and forth according as the iron disc is attracted to the rotating or fixed magnet. When the master key of the voice key circuit (140) is closed, current flows first through the fixed magnet, which is in the branch of the circuit of the stimulus voice key, causing the disc to be attracted to this magnet, and then also through the rotating magnet. Since the current through both magnets is equal, the disc remains in the initial position. Speaking into the stimulus voice key breaks the current through the fixed magnet so that the disc is attracted to the rotating magnet. It then rotates with the ar- mature shaft, causing the hand of the clock to turn at the rate of the armature, 5 rotations per second. Speaking ^into the reaction voice key breaks the current through the rotating magnet so that the disc jumps back again to the fixed magnet, and the hand stops. A spur gear on the shaft of the index-hand meshes with a larger cogwheel on the dial, serving as a rotation-counter. The dial itself is divided into 100 units, so that each unit measures 2 sigma. In using the chronoscope it is only necessary to set the hand at zero, press the master key before speaking the stimulus word. STUDYING CONTROLLED WORD ASSOCIATIONS 373 keep it down until after the response word is spoken, and then read the reaction time directly. The superiority of this chronoscope lies in the fact that it has no significant error, is extremely simple in operation, and runs continuously and noiselessly. It can run neither slow nor fast by the smallest fraction of time, else the motor gets out of step with the fork and stops. The only possible elements of error lie in (a) a possible change of vibration rate of the fork due to temper- ature changes, which can be obviated by enclosing it in a box with a thermostat, but which for this work is a negligible factor, (b) a possible error in the divisions of the dial, and (c) in a possible difference in reluctance of the disc between the magnets, in pass- ing in opposite directions, due to a possible difference in the strength of the two fields. This last was found to be negligible in the chronoscope used. The chronoscope will run continuously if care is taken to keep the fork contact properly adjusted. By experimenting, platinum wire was found impracticable for this; it burns up too quickly and has not enough spring to give the optimum length of period of contact for the motor to " catch." A gold alloy wire was tried, the wire used in dental work, and found extremely satisfactory. The gauge of the wire and its length are important, but these two factors must be determined empirically. As the wire burns back, the contact must be readjusted to keep the optimum period of contact for the motor. Only when this fails does the motor stop. Adjusting and cleaning the contact about twice a day when it is being used continuously all day has usually been found sufficient. The motor was run on the following voltage and amperage: Voltage: 10.0 volts, closed fork contract, but not vibrating. Voltage: 30.0 volts, broken fork contact. Voltage: 25 A volts, fork running on optimum contact for motor; motor dead. Voltage: 25 .4 volts, fork running, motor running. Amperage: 1.3 amperes, closed fork contact, but not vibrating. Amperage: 0.3 amperes, fork running on optimum contact for motor; motor dead. Amperage: 0.3 amperes, fork running, motor running. 374 MILDRED WEST LORING The motor is started by hand. It is equipped with a strobo- scope if a higher rate is required, but for the low rate of 5 rota- tions per second a slight twist of the axle which soon gets to be a knack causes the motor to get in step with the fork. The experimenter and subject sat on opposite sides of a small table with a large black curtain stretched between so that the subject could see neither the experimenter nor any of the appa- ratus. The experimenter sat with the master key and the chro- noscope on his right and the fork on his left, each on separate stands, so that neither could affect the voice keys, which are so sensitive that the passing wagons or a moving chair in an ad- jacent room stops the clock. With this arrangement too, the experimenter could record results and regulate the fork contact without moving from his chair. The subject was given his in- structions in regard to the type of response required and then the list of words. The experimenter signalled the subject before each word by saying " ready." Since the motor was practically noiseless, the only possible distracting noise was the low hum of the fork, which might have been eliminated by enclosing it in a padded box, or placing it in another room. This was a constant, however, throughout the experiment and apparently was not noticed to any appreciable extent except by one subject who worked at night when the building was absolutely quiet and all outside noises were at a minimum. EXPERIMENT I The general plan of the whole problem was to secure a large number of stimulus words — nouns, adjectives and verbs — to be given to a group of subjects, in order first of all that eliminations might be made of unsuitable stimulus words on the basis of the reactions to them. This was called experiment I. Final con- clusions were not to be based on these results because the subjects were few in number and also because the lists of stimulus words contained many words unfitted for the association experiment. The primary object then was to detect these unsuitable words by means of the reaction time and the response word, and to omit STUDYING CONTROLLED WORD ASSOCIATIONS 375 them in a later experiment. In experiment II these revised lists were given to a new group of subjects to make possible a more careful examination into the problems of technic under consideration. To this end a complete survey of an abridged Standard Dic- tionary of 300,000 words was made and all one, two, and three syllable adjectives, nouns, and verbs (transitive and intransitive separately) were listed. A verb having both a v.t. and v.i. mean- ing was classed as v.t., so that the two verb lists comprised verbs which, respectively, can take objects and which cannot. A large number of all these words were necessarily omitted. These fell into three classes, (a) technical words, such as modulus, titrates, (b) unfamiliar and archaic words, such as moil, bosky, (c) obviously vulgar words. In this last list are included only such words as actually occur in the dictionary, and not words having their vulgar meaning only in a subtle and secondary sense. Separate consideration of these was made later. This made twelve lists, which altogether totaled 10,888 words, with the words of each in alphabetical order. To get them in random order, each list was cut up so that one word was on each slip. The words of each list separately were put in a box, shaken thoroughly and drawn out one by one for relisting into groups of 40 words each. When the words were in final form there were found to be the following number in each class. ADJECTIVES NOUNS VERBS (TR.) VERBS (INTB.) 1 syllable 262 1806 792 219 2 syllable 3 syllable 973 791 2643 1726 1028 365 231 52 Totals. 2026 6175 2185 502 These in groups of 40 each, were now ready for use. Two groups of subjects were chosen, one to work on adjectives and nouns, the other on both classes of verbs. The associations were controlled as follows, 376 MILDRED WEST LORING STIMULUS RE8PON8K Adjectives Nouns Verbs (tr.) Verbs (intr.) Nouns Adjectives Nouns (objects) Nouns (subjects) As the words at this time were not yet completely catalogued, the several groups of words were given in serial order, the adjective- noun group of subjects going through the whole series of adjec- tives, in the successive order of one, two, and three syllables, and on their completion going through the whole series of nouns in the same way. The verb group likewise first did the transitive verbs and then the intransitive verbs. As noted above, the various classes of stimulus words were in lists of 40 each in this preliminary work, and four lists were given per hour on a twelve minute schedule — that is, twelve minutes were allowed to a list, any tune left over being given up to a rest period, during which the subject was allowed to do anything he pleased, converse with the experimenter, walk around or sit quietly. It was usually spent in irrelevant conversation. The remainder of the hour was allowed for preparation of material, breakdown of apparatus, and other details. Each day 160 words were given. In this experi- ment all subjects worked one hour per day for three days each week, as far as possible at the same hour of the day. The one exception was subject I who reported only twice a week. He fell far behind the others of his group, and for this reason subject IV was secured to supplement his work. It will be observed there- fore that the results of these two are not capable of intra-com- parison to the extent of the others; subject I finished all adjectives, part of the one syllable nouns, and all the two syllable nouns; subject IV did part of the one syllable nouns, and all the two and three syllable nouns. Subjects were instructed as to the required type of response word, and told to speak the first word occurring to them that fitted this requirement as quickly as possible, even though the response word was not exactly precise. Subjects were frequently reminded STUDYING CONTROLLED WORD ASSOCIATIONS 377 not to inhibit reactions that they might think foolish or vulgar; that the emphasis in this experiment was being put on the reac- tion tune, not the reaction word, so they must feel free to react without inhibition. As thorough a spirit of informality as pos- sible outside the working period was encouraged. All reaction tunes over twelve seconds were rejected arbitrarily as failures. The results for adjectives and nouns will be considered first, the results for verbs later, as the former were investigated in greater detail than were the verbs. Four subjects were used in the adjective-noun experiment, to give three complete sets of reactions for all the adjectives and nouns. All the adjectives were completed by each of subjects I, II, and III, three syllable nouns by subjects II, III, and IV, two syllable nouns completely by subjects I and IV, about half each by subjects II and III, one syllable nouns completely by subjects II and III, almost all by subject I and the rest by subject IV. The exact relation of the numbers of words done by each will be seen in the number of cases given in tables 1, 2, 3, 4, 5, 6, 7, re- membering that these figures represent the number of successful responses, not the entire number of stimulus words given to the subject. Subjects I, II, and IV were men students, a Senior, a Sophomore and a graduate respectively. Subject III was a woman graduate in the Department of Psychology. Results for the adjective-noun and noun-adjective association The results for adjectives and nouns in experiment I are shown in tables 1, 2, 3, 4, and 5, and in figures 1, 2, 3, and 4. Table 1. Here are shown the number of cases, the average re- action times, and the mean variations for each of the six groups of stimulus words, one, two, and three syllable stimulus adjec- tives and nouns, for each of four subjects. There are no results on three syllable nouns for subject I nor on any adjectives for subject IV. The order in which these words were given is to be kept hi mind. This was serially through one, two, and three syllable adjectives, and the nouns then in the same way. For subjects II and III the results are uniform. There is a definite 378 MILDRED WEST LORING I I I ! I II I I I i I I I ! ,11 STUDYING CONTROLLED WORD ASSOCIATIONS 379 MILDRED WEST LOBING STUDYING CONTKOLLED WORD ASSOCIATIONS 381 382 MILDRED WEST LORING increase in reaction time with the number of syllables for both nouns and adjectives, and a uniformly larger reaction time for any group of nouns over adjectives of the same number of sylla- bles. For subject IV no conclusions can be drawn; the reaction time for nouns of one, two, and three syllables is practically con- stant. Since this subject did none of the adjectives no compar- ison is possible between the reaction times for adjectives and nouns. The vocabulary of subject IV was very wide; there were only two or three failures during the whole series due to un- familiarity with the word. This may indicate that the increase in reaction time with the number of syllables is not dependent TABLE 1 Reaction times for unselected adjectives and nouns* (totals) SUBJECT I II III IV ONE SYLLABLE TWO SYLLABLE THREE SYLLABLE Adjectives Nouns Adjectives Nouns Adjectives Nouns 164 211 168 > % 282 248 218 1 > S 404 598 442 414 I > S 240 284 176 1 > •< > S 749 755 728 Jj > 3 192 302 268 1 > 3 828 554 370 1190 1461 1617 591 918 901 907 1324 1396 1328 2522 1649 817 2527 1744 2210 1750 1718 470 630 414 412 1268 1592 1380 1581 1608 1674 2542 2018 1692 * Adjectives and nouns refer to the stimulus words in all tables. on the actual increase in length of time for the stimulus word to be spoken, but rather because words increase in complexity of meaning and strangeness with their length. This would ac- count for the increase in reaction time with the increase in the number of syllables for subjects II and III, and for the con- stancy in reaction tune for subject IV whose interests were chiefly literary. The results for subject I are peculiar but thor- oughly explainable. It will be noticed that for the adjectives, which were completed before the nouns were begun, there was a steady decrease in reaction time in progressing from one syllable to two syllables and then to three syllables, the averages being respectively 1912, 1324, and 1268 sigma. The STUDYING CONTROLLED WORD ASSOCIATIONS 383 results for nouns, in regard to the increase in reaction time with increase in the number of syllables, agrees with those for subjects II and III. The reverse order for adjectives is to be explained by the fact that at the beginning of the work the subject was not in good health. This was not realized by the experimenter until the subject was well started in the work, and then it was thought interesting to see the effect of improvement in health on the reaction time. It will be observed how slow his reaction tune is for one syllable adjectives as compared to the other subjects, and how much it dropped for the two syllable adjectives. It was in the middle of this group of words that his poor health, which manifested itself in extreme nervousness, suddenly improved after being treated by a physician. This continued throughout the remainder of the adjectives; it could hardly be said that his reactions were perfectly normal until this time, which was about a month later. From then on during the noun series there was nothing con- spicuous in his behavior or reactions. Notes taken during his period of disturbance may prove of interest in emphasizing the dependence of the reverse order of reaction times upon his health at the time. Subject I. November 18. A tendency today to respond to an ad- jective with a noun cognate to it, e.g., awkward-awkwardness. When questioned in regard to this, says he has a feeling of going along the line of least resistance, which he can not control. Went to see Dr. X. last Friday about his nervousness. Dr. X. gave him some bitter medicine and it made him better the next day, and he has felt better ever since. Subject I looks better and is much less nervous. Face not so scratched — scratches it when nervous. Has, he says, especially hard attacks twice a year. Has been playing heretofore during experiment with a collec- tion of clamps, bars, wire, etc., collected from what is within his reach. No such behavior today. November 19. Some tendency to react as on yesterday, e.g., peevish, peevishness. Some tendency to repeat the response of the previous stimulus word. Much repetition of "being" as reaction word. Nerv- ousness present again today. Medicine taken last Friday was nux vomica — has been taking it every day, but not today, because "effect 384 MILDRED WEST LORING wearing off." Now attributes his well-being of yesterday to the good weather, and vice versa for today which is cloudy. November 26. Reactions of this type no longer occur: awkward- awkwardness, but new type has appeared — the same reaction word is given many times in the same list. "Spirit" was given 15 tunes in today's work (160 words). Other responses of this type were "mean" money," "mood." December 2. Reaction types occurred like those of November 26. The repeated responses were, List I — Occasion 3 times Gift 3 times Condition 5 times List II — Occasion 2 times Gift 2 times Condition 1 time List III — Occasion 1 time Gift 5 times Condition 0 time Spirit 6 times It may be that "spirit" in the last list was substitution, voluntary or involuntary, for the word "condition" In the middle of list III, sub- ject said that when he gave a reaction that was being duplicated so often, it was not the word that first came to mind, but to his lips, i.e., Maternal — gift (spoken reaction) Maternal — care (thought reaction) After being told that hereafter on such reactions the thought word as well as the spoken word would be called for, these reactions began immediately to fall off, i.e., List IV — Occasion 1 time Gift 0 time Condition 0 time Spirit 2 times but Mind 5 times as if a new word were being introduced to avoid the anticipated question- ing. There was still some evidence of the first type of odd reaction, i.e., awkward— ^awkwardness. December 3. Duplicate responses continue. STUDYING CONTROLLED WORD ASSOCIATIONS 385 List II — Conduct 8 times List III — Conduct 1 time Spirit 5 times Action 5 times List IV — Conduct 2 times Spirit 4 times Action 2 times Condition 1 time Man 7 times Mind 5 times Nouns — first day. Likes adjectives better than nouns ; nouns call up an object with no particular emphasis on its qualities, while the adjective can not appear without an object. Subject feels that the nouns are going faster however. Tables 2 and 3. These tables show the total distribution of reaction times for each of the four subjects who acted in the adjective noun experiment. The reaction time has been divided into steps of 100 sigma, and opposite each reaction time is re- corded the number of cases for a particular group of words in which the reaction time had a value lying between this particular number of sigma and 100 sigma more. For instance, in table 2, subject I, there are 16 reaction times of values between 1100 and 1200 sigma, for one syllable adjectives. All reaction times greater than twelve seconds were arbitrarily discarded as failures, and since the number of cases of very long reaction times was few for most subjects, it was decided to bunch all cases of reaction times lying between 4000 and 12,000 sigma. This accounts for the number of cases listed opposite 4000 sigma sometimes being much greater than the adjacent number of cases. These distribution tables bear out the results in table 1. On the whole for any given subject, the maximum number of cases for any group of nouns is at a higher reaction time than for ad- jectives of the same number of syllables. The syllab^ variation is not so perfect, but with greater uniformity among the number of cases, the modes would probably follow the averages. These results have also been plotted into distribution curves, which are shown in figures 1, 2, 3, 4. In each figure are given P8YCHOBIOLOQY, VOL. I, NO. 6 TABLE 2 Distribution of reaction times, adjectives and nouns SIGMA SUBJECT I SUBJECT II ONE SYLLABLE TWO SYLLABLE THREE SYLLABLE ONE SYLLABLE TWO SYLLABLE THREE SYLLABLE 1 ! % S I 1 1 JL 0 | 1 0 1 '•3 m a £ 1 ••3 3 1 1 !3 0 • c 1 0 •< 00 a 3 O £ 0 3 | 400 0 0 I 1 i 0 500 0 1 i 0 0 0 2 0 0 i 1 600 0 0 0 0 2 0 0 2 0 0 1 700 1 3 6 1 1 0 0 5 0 0 1 800 2 18 24 11 15 3 1 13 1 5 1 900 3 50 70 48 59 13 3 44 2 15 0 1000 7 91 108 113 108 24 19 87 16 37 0 1100 16 128 136 185 157 29 45 123 31 75 2 1200 21 159 152 250 126 31 76 119 85 88 7 1300 16 131 108 262 100 27 117 97 65 99 19 1400 19 112 102 251 71 19 120 94 98 92 37 1500 21 84 59 206 49 18 113 74 112 70 45 1600 17 80 54 192 21 16 104 64 95 64 64 1700 11 48 33 150 18 10 92 55 110 39 98 1800 11 36 20 132 6 7 95 30 91 36 92 1900 14 43 15 96 4 6 90 23 95 34 106 2000 38 C t- 99 1 0 & 54 28 93 13 77 2100 22 10 80 f^ 1 71 17 93 22 76 2200 26 r 66 t r O 61 8 82 16 87 2300 24 n A 53 f\ 1 50 9 70 14 72 2400 10 1 46 0 1 37 r 0 49 L 79 2500 6 1 41 0 0 38 1 48 5 55 2600 11 2 31 0 0 37 0 61 r t 75 2700 * 1 34 0 0 31 J 44 t! 54 2800 14 0 21 0 0 27 0 32 2 36 2900 r 1 25 0 0 25 0 36 1 51 3000 0 12 i 16 0 0 19 1 28 i 56 3100 0 7 0 18 0 0 10 0 17 • 37 3200 0 4 T 13 0 0 16 1 21 0 35 3300 0 j 0 12 0 0 17 0 26 \ 39 3400 1 2 0 Q 0 0 12 0 16 - 26 3500 - j 0 7 0 0 10 o 14 - 23 3600 0 1 o 11 0 0 10 0 16 ' 18 3700 0 - • 0 6 0 8 0 25 3800 0 0 o 0 6 o 17 0 12 3900 - o 0 o 2 0 11 o 18 4000 to 12,000 2 o 30 0 0 45 0 66 J 156 Total cases 164 1912 282 1190 1580 404 918 1324 240 2522 1744 470 749 1268 192 211 1378 248 1461 2014 598 901 1396 284 1649 2210 630 755 1592 302 1581 2542 828 Average R. T M V 386 TABLE 3 Distribution of reaction times, adjectives and nouns SIGMA SUBJECT III SUBJECT IV ONE SYLLABLE TWO SYLLABLE THREE SYLLABLE ONE SYLLABLE TWO SYLLABLE THREE SYLLABLE 8 I I 1 8 5 1 03 Q '•3 I 8 § q 09 o 1 00 c 9 s 3 3 .® 3 •o £ 3 1 J_ 1 3 I •9 1 -c < I 400 0 2 i 0 0 0 0 1 0 500 0 1 0 0 0 0 0 0 0 600 4 0 6 0 3 1 1 1 1 700 12 2 36 2 5 3 1 3 0 800 27 9 82 9 23 4 4 14 5 900 20 31 131 9 48 16 7 38 24 1000 23 77 123 39 89 25 18 97 66 1100 23 120 131 57 83 74 54 169 109 1200 19 137 97 72 84 92 60 229 134 1300 8 160 80 72 100 90 71 231 203 1400 8 149 66 78 81 132 54 233 195 1500 9 139 41 70 60 118 45 237 150 1600 5 122 39 55 40 91 41 229 143 1700 2 94 23 53 39 92 40 175 112 1800 4 91 16 42 20 106 25 148 96 1900 2 82 13 49 14 101 32 119 73 2000 0 63 11 35 8 99 27 113 75 2100 2 51 4 32 6 64 16 80 63 2200 0 52 3 28 7 55 16 77 42 2300 0 30 0 15 6 46 20 63 46 2400 0 31 2 15 3 62 11 41 22 2500 0 40 1 22 2 46 7 41 21 2600 0 19 1 7 3 28 9 41 18 2700 0 21 0 12 1 29 7 28 15 2800 0 17 0 6 0 25 2 21 14 2900 0 16 0 8 3 31 4 25 5 3000 0 17 0 6 0 26 4 16 9 3100 0 10 0 3 0 23 2 7 6 3200 0 7 0 8 0 26 3 5 8 3300 0 4 0 7 0 17 1 4 4 3400 0 5 0 2 0 12 0 8 5 3500 0 2 0 0 0 12 0 8 6 3600 0 4 0 0 0 14 1 4 1 3700 0 1 0 0 0 7 0 5 2 3800 0 1 0 2 0 3 1 0 0 3900 0 1 0 0 0 9 2 5 0 4000 to 12,000 0 9 0 2 0 29 5 11 1 Total cases 168 1617 907 817 728 1608 591 2527 1674 Average R. T 1138 1730 X328 1750 1380 2018 1708 1718 1692 M. V.. . 218 442 176 414 268 554 414 412 370 387 388 MILDRED WEST LORING six curves, a one, two, and three syllable curve each for adjectives and nouns. Full lines represent the adjectives and dotted lines the nouns. The number of syllables is indicated on the curve itself. Smoother curves might have been obtained if larger steps in the reaction time had been chosen, and if the number of cases for the various groups of words had been more uniform. It is worth while emphasizing the fact that reaction times do not follow the error curve of mathematics. Whereas the error curve is perfectly symmetrical, the reaction time curve is very much skewed, with the maximum of the curve lying at a much lower value than the mean of the extreme reaction times. The reason for this is obvious. There is a physiological limit for the lowest value of a series of reaction times, below which no reaction time can fall, while on the other hand there is no such limit put upon high reaction times. Reaction times may be expected to be of any increasingly greater value beyond the minimum, up to infinity (which we call failure) depending on the ability of the subject. The curve then takes the form indicated here. It should be noticed therefore that only such mathematical formulae should be applied to the data of such curves as have been developed for this type of measurements. For this reason only the number of cases, averages and mean variations are given in the data presented here. There has been an unfortunate tendency to apply formulae that refer exclusively to the mathe- matical curve of error to such curves as are given here. The reaction times may at first sight appear abnormally long. Likewise the mean variations may seem high. But this is not so however. In the first place no conclusion is valid unless based on a sufficiently large number of cases. In a great many investigations only a short list of words has been used and sweeping conclusions are made on the basis of their results. In this experiment over 8000 adjectives and nouns have been used for stimulus words, so that a fairer indication is given of the true nature of the reaction tune under the conditions laid down in the experiment. In the second place no attempt has been made in this work to choose "easy" words for stimulus words. In much of the association work easy words have been chosen STUDYING CONTROLLED WORD ASSOCIATIONS 389 because they are more suitable for educational and pathological tests. It is to be expected that the average reaction time on such a list of words given to normal subjects would be consider- ably lower than when the list of stimulus words is maximally inclusive as is the case here. The aim here was to retain as many words as possible both difficult and easy, rather than to choose a select, homogeneous list. But aside from these two reasons which might justify unduly long reaction times, it must still be concluded here that these reaction times are not long at all. A careful gleaning of the literature shows that where any considerable number of stimulus words has been used, reaction times running up to four and five seconds are not uncommon even for normal subjects. To what extent these high reaction times have been found is difficult to determine because in most cases detailed results are not given, and mean variations are invariably omitted which might afford a clue to this point. The chief interest has been concerning the nature of the associations themselves and this has helped to minimize a dis- cussion of the length of the reaction time. For this reason detailed results are given here showing the number of cases, averages, mean variations, and distributions. All computations were made on the Burroughs electric adding machine. Mean variations were obtained with the Dunlap formula, which is especially adapted for use on a calculating machine. Table 4- This table shows the variation in reaction tune according to the position of the accent in three syllable adjectives and nouns. It was not considered worth while or valid to make comparisons on accent for two syllable words inasmuch as the percentage of these with the accent on the second syllable is extremely small in comparison with those having the accent on the first syllable. And likewise for three syllable adjectives and nouns, comparison is really valid only between words having the accent on the first and on the second syllable, inasmuch as the number of words with the accent on the third syllable is very small. No conclusion can be drawn at all in regard to the effect of accent on reaction time. The variations in reaction time with change of accent for either adjective or nouns follow 390 MILDRED WEST LORING no consistent sequence for the various subjects. It is interest- ing to note though that with both groups of words here sub- divided into three classes each, the conclusion drawn from table 1 still holds, that is, that the reaction time is greater for stim- ulus nouns than for stimulus adjectives. Table 5. Here are shown the results of separating the stim- ulus nouns into seven logical categories and one unclassified group. Considerable difficulty was experienced in rinding a scheme for classification. It was intended at first to classify them into two groups only, abstract and concrete, but when an actual classifica- tion under these simple headings was tried it was found utterly TABLE 4 Variation of reaction time with accent; three syllable unselected adjectives and nouns SUBJECT ACCENT ON F1E8T SYLLABLE ACCENT ON SECOND SYLLABLE ACCENT ON THIRD SYLLABLE Adjectives Nouns Adjectives Nouns Adjectives Nouns 1 > a- 1 > g- J 4 * J > g- 0 *5 5 1 > 5 I II III IV 363 368 352 1232 1604 1358 272 356 158 950 981 1014 2508 2034 1700 768 558 382 374 371 365 1294 1540 1388 198 322 270 579 582 611 2594 1992 1686 700 286 368 12 16 11 1458 1576 1624 144 204 248 42 45 49 2674 1948 1568 586 566 282 impossible. There is no hard and fast line between these two classes. There are, of course, nouns which are obviously con- crete, and othere undeniably abstract, from a certain point of view, but between these and including a very large percentage of all nouns are a great horde which are really of widely varying degrees of concreteness and abstractness. The terms are only relative. From one point of view everything is concrete and from another everything is equally abstract. Even the stock illustrations of abstract nouns, such as triangularity or virtue, may be thought of as being just as concrete as furniture or walking. The words that gave the greatest difficulty were those of a lower level of abstractness than those which are usually used as illustrations of abstracts (nouns ending in ity, hood, ness), STUDYING CONTROLLED WORD ASSOCIATIONS 391 TABLE 5 Reaction times for classified nouns (unselected) SUBJECT I SUBJECT II SUBJECT III SUBJECT IV 1 >•' •< > S I o > ' < > a o > < > a One syllable group* In An 554 144 152 34 46 17 68 175 1518 1720 1530 1554 1622 1630 1654 1664 364 436 452 372 390 268 380 492 404 673 188 198 54 63 18 78 189 2022 2198 1964 2078 1782 2244 2142 2238 2014 544 804 526 648 410 888 618 574 598 725 212 216 63 63 23 103 212 1617 1700 1774 1674 1604 1782 1774 1900 1702 1730 422 472 390 352 322 496 622 544 442 266 73 74 28 28 6 52 64 591 1708 1656 1730 1594 1610 2068 1730 1776 452 400 408 372 410 368 390 364 414 Ac Vg Pb Em Ab Un Total 1190 1580 1461 1708 Two syllable group* In 1051 541 248 78 37 24 213 312 1644 1822 1892 1600 1654 1990 1836 1794 396 382 530 424 486 558 552 486 659 354 161 49 30 16 .186 194 2148 2282 2264 2080 2290 2334 2302 2198 588 612 624 376 558 472 732 678 348 169 88 28 8 11 54 111 1702 1998 1732 1546 1480 1758 1664 16H4 414 488 392 356 450 264 396 366 1024 539 263 76 46 23 246 310 1666 1698 1826 1606 1696 1802 1840 1768 418 384 436 440 320 630 426 440 An Ac Vg Pb Em Ab Un.. Three syllable group1 In 434 322 222 38 12 23 275 255 2538 2716 2698 2338 2334 2774 2864 2656 680 682 722 704 596 710 614 694 440 325 234 42 14 25 279 249 1956 2092 2060 1600 2540 1750 2038 2040 524 514 648 410 536 518 606 554 446 343 236 39 20 26 293 271 1670 1674 1764 1404 1708 1758 1702 1694 1692 334 356 390 268 402 426 420 414 An Ac Vg .. Pb Em Ab Un Total 1581 2542 828 1608 2018 554 1674 370 *In, inanimate objects; An, animate objects; Ac, nouns of action; Vg, vege- table kingdom; Pb, parts of body; Em, feelings and emotions; Ab, abstract nouns; Un, unclassified. 392 MILDRED WEST LORING but which are not of the level of concreteness of apple or book. Such words are cost, fate, style, skill. Do we mean cosiness, fatehood, styleness, skillness, or the actual money, the actual happening, short skirts and high boots, and manipulation of figures? Or can we say that we are ever precise at all in just what we do mean, is now one meant, now the other? For this reason a new classification was sought. Seven categories with an additional unclassified group was the smallest number found possible to use. They were chosen as follows: 1. Inanimate objects. This class is self-explanatory. 2. Animate objects. Here are included all nouns denoting living objects in the animal kingdom. A corpse or a salt herring was classified as inanimate. The noun must also denote the whole organism, not just one of its parts, e.g arm and knuckle come under a later category. 8. Actions. These nouns express pure action, such as leap, riding, for their existence lasts only during the leaping or riding. After the leap has been leaped there is no leap left. This was taken as the test for this class. There are a good many nouns which may express either pure action or the result of the action. The adjective given by the subject was taken as criterion of which interpretation has been made by the subject, and it was so classified. If ambiguous, it was relegated to the unclassi- fied group. Such an example of possible double meaning is the noun crack. This may mean the actual space in the side of the broken object, as in the reaction jagged-crack, or it may mean the action of cracking itself as in the reaction sudden-crack. The former and others of its kind were classified as inanimate objects. 4. Vegetable kingdom. This class is self-explanatory. 5. Parts of the body. This class is self-explanatory. It was of course not possible to list them under animate objects with words denoting a complete organism. Besides it was thought that a separate classification of these might through a possible increased reaction time throw some light on the emot'onal reaction, inasmuch as many parts of the body are of especial -erotic significance. STUDYING CONTROLLED WORD ASSOCIATIONS 393 6. Feeling and emotions. These are nouns actually denoting the feelings and emotion themselves, such as love, anger, distress. In this list were not included words having an emotional con- notation, such as mistletoe. 7. Abstracts. These have been partly discussed above. It was decided to include not only the orthodox abstracts which usually end in ity, hood, ness, but also that lower hierarchy of abstracts including words like vogue, needs, and loss. 8. Unclassified. Here were put the days, months, seasons, sounds, diseases, weights and measures, collective nouns, direc- tions of the compass, times of day, parts of speech, the sciences and arts, and others. The result of this classification as indicated in table 5 shows nothing. There is no greater variation in the average reaction time for the eight classes of nouns for any given subject than would be expected from the large variation in the number of words per group. This of course was not controllable. This uniformity may indicate that emotional disturbances, from what ever cause — the meaning of the word, its relative unfamiliarity, etc. — do not necessarily manifest themselves in an increased reaction time but through some physiological mechanism other than the vocal apparatus, such as respiration, heart rate, blood pressure. This problem is already under investigation in this laboratory. Certainly nothing was indicated by this laborious classification to give any clue to a possible difference in types of response according to the intrinsic nature of the stimulus noun. Because of these negative findings it was decided not to compile reaction tune results on the classification of adjectives. These adjectives had already been classified under the same heads as the nouns, the criterion of classification being the noun cognate with the adjective. Results for the verb-object and verb-subject association The results for verbs in experiment I, given in table 6 are an- alogous to those for adjectives and nouns. The reaction tune for the verb-subject association is longer than for the verb-object 394 MILDRED WEST LORING association, and the reaction time for both types of association increases directly with the number of syllables in the stimulus word. On the basis of these verb associations further elimina- tions in the verb stimulus words were made. It was only lack of time that prevented a more detailed investigation into these types of controlled associations by giving the selected words to a new and larger group of subject in a more systematic manner. This was done only with the selected adjectives and nouns. TABLE 6 Reaction times for verb-object and verb-subject associations SUBJECT ONE SYLLABLE TWO SYLLABLE THREE SYLLABLE Verb-object Verb-subject Verb-object Verb-subject Verb-object Verb-subject 1 > < > a 8 J > <3 > a • > 3 > a 1 4j > a 1 6 > *4 > a 1 > «< > a V 435 1308 432 213 1536 506 985 1420 444 220 1898 598 369 1516 388 50 2124 438 VI 281 1708 606 194 1760 410 941 1756 544 219 1742 432 41 2118 542 VII 718 2026 808 960 2250 902 366 2490 926 VIII 437 1614 520 195 2284 806 215 2690 906 351 2140 782 47 2990 1154 Conclusions for the adjective-noun and noun-adjective association 1. The reaction time for the noun-adjective association is very definitely longer than for the adjective-noun association, for stimulus words having the same number of syllables. The amount of this increase varies with the number of syllables in the groups compared and with the subject. The difference varied from 420 sigma to 950 sigma, but not in any fixed way in passing from words of one syllable to those of two and three syllables. The reason for this definitely longer reaction time for nouns than for adjectives is probably to be sought in the normal order of nouns and adjectives in the English language. With the exception of a few set phrases such as "durance vile" and "choir invisible" the universal order in the English language is adjective-noun. Because of this the habit of reaction in this direction is very stable and as mechanised as is possible with the permutation of adjectives and nouns occurring in language. STUDYING CONTROLLED WORD ASSOCIATIONS 395 The reverse reaction therefore always requires greater effort and gives a longer reaction time. In connection with this pos- sible explanation an experiment is under way in this laboratory to test out the same reactions on French, Italian, and Spanish subjects where the order of adjectives and nouns is on the whole the reverse from the English order. If this language explana- tion has validity, we will expect to find that the adjective-noun reaction is longer than the noun-adjective reaction. A contributory reason for the reaction time being longer for nouns than for adjectives lies in the fact that many nouns are commonly never used with a modifying adjective, e.g., rote and ounce, and it is relatively difficult to find adjectives to modify them. This lengthens the reaction time. This was no factor however in Experiment II which follows, where all such words were eliminated. Subjects invariably stated that it is more difficult to respond with adjectives to noun stimulus words than with nouns to adjective stimulus words. They were questioned in regard to this and their naive explanations all fell into this general scheme; an adjective always suggests some object having that quality, but a noun suggests no particular aspect of itself, making it necessary to "feel" for an attribute. Some notes taken of comments by different subjects may be of interest here. Subject II. November 17 (first day on nouns) . Nouns much harder than adjectives. Have to think backwards then forwards to see if adjective fits. Have to think harder but time seems to go much faster for a list. (This is not so — time really much longer.) Subject I. December 10 (first day on nouns). Likes adjectives better than nouns, because nouns call up an object with no particular emphasis on its qualities, while the adjective can not appear without an object. It feels as if the nouns were going faster. Subject IV. November 22 (first day on nouns). Felt as if doing poorly. Ought to be able to think of better adjectives — less common- place ones. 2. The reaction time for these types of controlled associations increases directly with the number of syllables in the stimulus word. This holds for both adjectives and nouns. The one 396 MILDRED WEST LORING exception found for adjectives was in the results for subject I and for nouns in the results for subject IV. These have already been explained. This increase varies for adjectives from 18 sigma to 196 sigma and for nouns from 10 sigma to 332 sigma. 3. The position of the accent in stimulus nouns and stimulus adjectives has no systematic effect on the reaction tune. 4. There is no interpretable variation in reaction time accord- ing to a logical classification of the stimulus nouns for the noun- adjective association. The reaction time remains relatively constant within the limits of word length. The variations can be attributed to a difference in the number of cases in the various groups. 5. From observations of the subjects during the course of the experiments is to be concluded that two or three separate hours of work are sufficient for the subject to become adapted to the experiment, to lose any emotional disturbance due to sex difference between experimenter and subject, as far as this last can ever be reached. Inhibition of associations from such a cause we are persuaded was at its minimum during the whole experiment for all subjects. 6. It is impossible for a subject to keep at the top notch of tension throughout the whole of one session, and necessarily not for the whole experiment, to the same extent that this is possible in getting reaction tunes on a few words. That which we call "tension," whatever its physiological mechanism, was evident for the first two or three words of each list. It was in- dicated by a distinctly shorter reaction time and an observable muscular rigidity in the subject. The reaction time then sud- denly became longer and remained more or less uniform through- out the list. It is probable that this later uniformity is due to an adjustment to a certain comfortable physiological tension that can be maintained through the course of one hour's work at a time, for three days a week, throughout several months. Conclusions for the verb-object and verb-subject associations 1. The reaction time for the verb-subject association is longer on the average than for the verb-object association for stimulus words having the same number of syllables. This is probably STUDYING CONTROLLED WORD ASSOCIATIONS 397 due to the word order in the English language. Since the normal sequence for subject, verb, object is seldom varied, it is natural that associations between them should be in the order of subject- verb and verb-object. If the reaction time should prove to be equal for these associations in their normal order, it would follow that by reversing one association, for instance the subject- verb to the verb-subject, this reaction time would be longer than for the verb-object association. This is what the results here show. It is probable therefore that the English word order accounts for the relative value of the reaction times for the verb- subject and verb-object associations. 2. For both types of verb associations the reaction time on the average increases directly with the number of syllables in the stimulus word. The explanation for this is probably the same as for the same finding for adjective-noun and the noun- adjective associations, that is, it is due to the increasing un- familiarity and complexity of meaning of a word with the in- crease in the number of its syllables, rather than to the actual increase in the time taken to speak the longer words. EXPERIMENT II A more detailed study of the adjective-noun and noun-adjective association From the results of the previous experiment it was possible to secure a revised list of adjectives, nouns, and verbs, forming a lexicon of words suitable to the types of associations outlined above; to take an equal number of each group for further trial on a large number of subjects and investigate in greater detail the same problems studied in experiment I. The lexicon was compiled for each of the four groups of words and will be later published separately but lack of time prevented any further investigation into the verb-subject and verb-object associations. The first task was to eliminate unsuitable words. These were found to fall into the following classes: 1. Unfamiliar words beyond those already eliminated in mak- ing up the original lists. Throughout experiment I there was a 398 MILDRED WEST LORING growing conviction that it is necessary to distinguish between the reading and the conversational vocabulary. In making the original lists the experimenter had unwittingly but naturally chosen words which are familiar as read. Many of these were pronounced unfamiliar when heard by the subject, though his judgment changed when the word was spelled for him. 2. Words difficult to pronounce intelligibly to the subject, such as leak, which was taken variously as leap, link, etc. 3. Words having an emotional value either obviously or subtly; in the first case words like pregnant, corset, etc., and in the latter case that rather large group of words almost entirely confined to the masculine vocabulary whose dictionary meaning is perfectly unemotional, but in everyday use have also a subtle sexual meaning. The elimination of these words was made by several men in the department. Since however, many of these words seem to be purely colloquial, there are doubtless many still in the revised lists. 4. Homonyms in the narrow sense. A difference of spelling and identity of pronunciation in words of the same part of speech was made the criterion. Fate and fete are homonyms for this experiment, but not great and grate. It was considered sufficient control that the subject knew what part of speech he was being given. Furthermore most homonyms actually do exist in the narrow sense taken. Two of these homonyms escaped detection and got into the revised lists. These are chaste, chased, and dessert, desert/ 5. Words that are intrinsically difficult to respond to with the required type of response, such as the nouns nothing and ounce. Adjectives are not commonly used with these nouns. 6. A small group of words eliminated for various unclassi- fied reasons, long reaction times or absurd responses attributable to none of the above reasons. After these rejections were completed, special lists were made up from the selected words for ten days' work. Each list con- tained 30 words and six lists were given in an hour making 180 words per session. Whereas the two types of stimulus words were run through serially in experiment I, in this experiment one, STUDYING CONTROLLED WORD ASSOCIATIONS 399 two, and three syllable nouns and adjectives were run in parallel so that comparison between the groups would be entirely valid. Therefore each day's series was made up of one list each of the following words and give in this order : (1) One syllable adjectives (2) One syllable nouns (3) Two syllables adjectives (4) Two syllable nouns (5) Three syllable adjectives (6) Three syllable nouns The adjectives and nouns alternated on successive days in oc- cupying the first position in the series. Nouns occurred first on odd days and adjectives on even days. It was discovered unfortunately that there were not enough selected one syllable adjectives to cover ten days' work, in fact there were only 194 of them, enough for six days' work (180) and a few over. For this reason it was possible to carry out the above procedure for only six of the ten days. The remaining days' work was made up only of two and three syllable adjectives and nouns. It was necessary to drop one syllable nouns for the remainder of the series in order that there might be an equal number of one syllable adjectives and nouns for the comparison of their reaction times. To make the number of two and three syllable adjectives and nouns come out even for the remaining four days' work the fol- lowing schedule was adopted: /Two lists each of 2 syllable nouns and adjectives Daws 7 and 9 < ^ ,. A * . (One list each of 3 syllable nouns and adjectives /Two lists each of 3 syllable nouns and adjectives Days 8 and 10 < _ (One list each ot 2 syllable nouns and adjectives This gave the following distribution of the 1800 stimulus words: 1 syllable adjectives and nouns, each 180 360 2 and 3 syllable adjectives and nouns, each 360 1440 In addition to the regular ten days' work, three days' practise work was given beforehand on the basis of the conclusion in experiment I that the length of time required to become adapted 400 MILDRED WEST LORING to the experimental situation is two or three sessions. For stimulus words, were taken whatever material was available from the rejected words of various kinds, and as nearly as pos- sible the scheme of parallelism and alternation was carried out. No three syllable nouns were included in these practise words. Rej ec- tions had not been quite completed on them at the time experiment II began, inasmuch as the three syllable nouns formed the last material for experiment I. The actual three days7 practise words, totalling 540 nouns and adjectives, were distributed as follows: ONE SYLLABLE TWO SYLLABLE THREE SYLLABLE Adjectives 30 180 60 Nouns 120 150 0 It was also intended that the results for the practise period and the regular ten days' work should be compared to see directly the effect of excluding unsuitable words of different kinds. Eight subjects were used in this experiment, four of whom had acted as subjects in the experiment on verbs, subjects V, VI, VII, VIII and four who were entirely new to the whole procedure. All eight subjects took the practise work, for there was no reason for thinking necessarily that the first four might be adapted to adjective and noun associations merely because they had been working on verb associations. Of course practically all emo- tional disturbance in the first four subjects had already been eliminated. At the beginning of the regular ten days' work it may be said that all subjects had lost as much of any emotional disturbance as they would ever lose, and for all except one subject perhaps we would say that any remaining emotional upset was practically nil. This one exception, subject XII, was of a natu- rally shy disposition. Subjects VII, VI, VIII, IX, X and XII were all university men, the first of these being a Ph.D., the second a graduate student, the next two sophomores and the last three freshmen. Subject XI was a junior at a woman's college. In all there were seven men and one woman. It had been hoped to have an equal number of men and women but the difficulty of getting women subjects prevented this. STUDYING CONTROLLED WORD ASSOCIATIONS 401 Results for experiment II Table 7. This is similar to table 1, experiment I. It shows the number of cases, averages, and mean variations for one, two, and three syllable selected adjectives and nouns for the total regular ten days' work. The results are consistent through- out for each observer in corroborating the conclusion in experiment I that the noun-adjective reaction is longer than the adjective-noun reaction. These results also substantiate the earlier conclusion that the reaction time increases directly with the number of syllables in the stimulus word for these two types of associations. TABLE 7 Reaction times for selected adjectives and nouns, regular ten days' work (totals) ONE SYLLABLE TWO SYLLABLE THREE SYLLABLE Adjectives Nouns Adjectives Nouns Adjectives Nouns k 1 ^ | ^ § >- to ^ 8 ^ 8 ^ i 6 4j § 6 $ * 1 4 * 6 j a 6 j; S 0 { S V 179 1196 236 176 1584 418 347 1382 278 352 1666 408 350 1616 450 348 1852 424 VI 176 1196 356 176 1602 386 351 1290 358 344 1836 598 352 1466 414 350 1910 580 VII 177 1640 478 177 2380 928 350 1856 636 346 2660 1014 351 2130 758 344 2902 1046 VIII 168 1766 548 169 2434 730 341 1962 652 334 2752 624 344 2318 920 347 2978 1042 IX 174 1738 634 172 1918 586 347 1830 604 337 2338 804 340 2022 612 349 2568 814 X 173 1338 300 174 1532 294 349 1508 240 334 1714 346 344 1790 546 340 1958 358 XI 174 1694 676 177 2300 886 350 1998 812 348 2860 1286 332 2396 1078 339 3002 1334 XII 172 1736 622 166 1920 562 343 2038 710 344 2336 798 332 2460 810 344 2518 840 Table 8. This table shows the reaction times for the regular ten days, giving the numbers of cases, average reaction times, and mean variations for each of the ten days separately. The con- clusions drawn from the previous table, where, the same reaction times of each class were treated en masse for the ten days, are on the whole borne out here. An examination into the relative length of the average reaction time for adjective and noun stim- ulus words for each subject for each of the ten days shows the following number of cases where the noun stimulus word gives a longer average reaction time than the adjective stimulus word. Of course, only the average reaction times for the same PSYCHOBIOLOGY, VOL. I, NO. 6 402 MILDRED WEST LORING 10 1 'A'K •AV sas«o 1 I 'A'K •AV 888*0 (H Q 1 'A'K •AV 898*0 Adjectives 'A'K •AV S8S*0 CO H i 'A'K •AV 888*0 i 'A'K •AV 888*0 CM a CO 1 'A'K •AV 888*0 1 'A'K •AV 88SBQ ° 09 1 •A-K •AV 888BQ I 'A'K •AV 88SBO «™, 8 8 3 S 8 §8 §2 2 2 £ g § ill il el n SS 00 § g 00 00 O I in g § IM 00 CO S S S C5 rfi ^J 1 1 i g§ §11 0 CM 0 S CO S lii o o CO CO CM CM gss 10 t- «0 STUDYING CONTROLLED WORD ASSOCIATIONS 403 gj§5 8 8 8 (M OO CO 3 115 CO CO 00 2 I i (M O ss s M 00 CO C^ (>^ S 2 §§ — I CSI cJ 111 §8 a III 404 MILDRED WEST LORING 2 S 1 'A'W > ID 3 03 •AV 888*3 1 1 ^d 'A'W •AV 888*3 Oi S 1 'A'W •AV 888*3 Adjectives •A'W •AV 888*3 00 Q § 1 'A'W •AV 888*3 Adjectives 'A'W •AV 888*3 t-. |H § 1 'A'W •AV 888*3 •I 'A'W *AV 888*3 CO 1 'A'W •AV 888*3 i C<) S 1 4 £ J 4 s Subject V 1 29 1534 432 28 1878 680 29 2052 664 58 1120 575 2 58 1772 348 56 2164 620 57 1650 410 54 2012 504 59 1502 344 28 2052 40( 3 25 1662 350 30 1920 370 Subject VI 1 29 1316 336 25 1920 414 25 1830 402 53 2322 574 2 50 1720 550 48 2474 780 56 1996 772 56 2488 728 53 1592 424 27 2846 992 3 26 1976 542 23 2372 816 Subject VII 1 28 1866 416 27 3686 1622 25 2954 702 51 4160 1308 2 52 2784 1194 50 3818 1492 53 2420 906 52 3702 1248 51 rw;^J£J ^DOD 776 26 3762 1198 3 24 3034 1460 26 2080 540 Subject VIII 1 26 1938 532 25 3534 1178 26 2962 874 46 3374 950 2 53 2612 1136 45 4352 1276 58 2742 1128 52 3200 1276 46 2520 780 27 3752 1386 3 23 3186 1154 22 2996 804 Subject IX 1 27 1834 848 24 2524 974 25 2398 794 56 3824 1550 2 49 2036 636 38 2658 1038 49 2866 1318 55 3664 1472 49 2958 1216 21 4690 1442 3 23 2804 820 24 3198 1278 Subject X 1 23 1726 494 21 1708 476 24 1512 394 51 1878 410 2 48 1760 420 42 2100 694 51 1608 338 49 1870 422 50 1624 358 27 1938 332 3 25 1896 516 24 1784 410 Subject XI 1 28 1636 522 20 2864 980 25 2704 1226 49 3240 1230 2 55 2132 840 50 3004 1296 52 2096 652 56 2954 1308 47 2176 808 27 3840 1384 3 24 2644 1482 26 2282 982 Subject XII 1 25 1704 556 24 2584 1104 26 3050 1308 48 3128 1174 2 57 2410 872 51 4174 1398 53 2688 1078 56 3514 1190 45 2640 914 24 3320 1084 3 18 3176 976 19 3080 938 409 TABLE 11 Distribution of reaction times, adjectives and nouns SIGMA SUBJECT V SUBJECT VI One syllable Two syllables Three syllables One syllable Two syllables Three syllables 1 > S 8 $ 1 1 c i 00 1 i 1 1 1 1 1 *o o T? 0 *? 0 *Q o a o *C? o ^ : 8 > 1 >: 1 >: I > 1 ^ i 0 <3 a 6 •5 a 0 < a 0 < a 6 •S a 0 •5 a V 29 1400 266 29 1630 414 30 1458 220 60 1674 432 58 1932 380 59 1638 208 VI 30 1066 184 29 954 122 30 994 170 60 1436 362 59 1200 292 60 1214 248 VII 29 1884 520 29 1318 290 29 1166 266 58 2154 754 56 1666 504 59 1356 432 VIII 30 1350 318 29 1604 402 30 1172 312 59 2044 658 59 1848 652 59 1488 420 IX 30 1562 504 30 1570 408 30 1372 314 56 1776 398 59 1566 416 60 1678 476 X 30 1644 368 26 1398 174 30 1120 156 55 1956 516 59 1376 264 60 1228 230 XI 29 1930 910 30 1758 446 29 1310 400 56 2732 1356 58 2260 806 57 1758 716 XII 30 1760 676 30 1348 546 30 1040 346 59 2086 708 60 1680 692 60 1328 462 TWO SYLLABLE NOUNS THREE SYLLABLE NOUNS V 30 1758 346 30 2210 502 30 1878 426 59 2046 484 59 2200 380 60 2072 364 VI 30 1708 482 30 1198 216 30 1108 140 59 1916 592 57 1500 408 58 1438 358 VII 29 2688 996 29 1922 746 28 1366 332 55 2902 1226 55 2024 688 57 1610 488 VIII 30 2374 714 30 2262 802 29 1808 556 57 3106 1032 60 2582 810 57 2124 702 IX 30 2210 840 29 1848 484 30 1536 360 58 2518 736 57 2162 824 60 2088 688 X 28 1946 360 27 1512 298 30 1198 218 55 2146 542 59 1470 252 58 1204 174 XI 30 3698 1842 30 2806 1226 30 1858 612 53 3214 1320 54 3224 1364 58 2120 850 XII 30 1998 738 30 1628 550 30 1154 360 58 2098 566 53 1784 612 60 1540 418 Conclusion for experiment III 1. Successive repetitions of a list of stimulus words cause successive decreases in the reaction times when the subject is not informed of the fact of repetition. These reaction times would ultimately reach a physiological level. EXPERIMENT IV. DOUBLE ASSOCIATIONS This experiment like experiment III was made only on a small scale in order to get an indication of the probable course of the results on a larger scale. It comprised one hour's work only for each of four subjects, V, VI, VII, VIII. Six lists of stimulus words were given, 20 each of one, two and three syllable adjec- tives and intransitive verbs, run in parallel as in experiment II. STUDYING CONTROLLED WORD ASSOCIATIONS 417 Double associations were required. For an adjective stimulus word, the subject must think of a noun applicable to the ad- jective and respond orally with a verb having this noun as sub- ject. In this way the first association was silent, the second spoken. The verb given might be either transitive or intransi- tive, but auxiliaries were prohibited. With the transitive verbs, the association was made in the reverse order, back to a noun as its subject (silent) and then to an adjective (spoken) modifying this noun. Intransitive verbs were chosen for this type of re- action as a check on the backward direction of the association, otherwise the association might be made forward to an object of the verb and its modifier. Results for experiment IV The results are shown in table 16. With the two exceptions in the three syllable words, there is a uniformly large reaction time TABLE 16 Reaction times for double associations SUBJECT ONE SYLLABLE TWO SYLLABLE THREE SYLLABLE Adjectives Verbs Adjectives Verbs Adjectives Verbs I > «5 >" S 1 > •< >•' a 1 jj >>' 3 I >' a 1 > < > a 0 >' ^ > a V 20 2236 696 20 2466 618 20 2466 466 20 2498 546 19 3054 496 20 2892 680 VI 19 1740 400 20 2576 556 19 2458 762 20 2958 852 20 2428 668 20 2658 516 VII 20 2844 764 18 3954 1496 19 4112 1008 20 4690 1674 20 4378 1720 20 5202 1492 VIII 20 2894 626 20 4206 1230 20 4546 1924 19 4606 1566 20 4382 1290 19 4256 1116 for the backward association from verbs to adjectives. This results agrees with the comments of the subjects when asked which was the easier. The backward association with some ex- ceptions was felt to require the greater effort. The increase in reaction tune varies, to be sure, between wide limits among the different subjects and for stimulus words of different lengths, from 32 sigma to 1312 sigma, with an average increase of 571 sigma. For the two exceptions for subjects V and VIII, PSYCHOBIOLOQY, VOL. I, NO. 6 418 MILDRED WEST LORING respectively, the decrease in reaction time for the backward association is small compared to the average increase for other cases. It amounts only to 162 sigma and 126 sigma. A larger number of cases would probably throw these results in the same direction as the others. Conclusion for experiment IV 1. The reaction tune is longer for "backward" than for "for- ward" double associations, using these terms to refer to the normal language order. The reaction tune is longer in associat- ing from intransitive verbs back through noun subjects to mod- ifying adjectives than from adjectives forward through noun subjects to verbs. SUMMARY One of the most important results of this investigation has been the preparation of lists of stimulus words suitable for use in the word association reaction. These lists total 10,888 words and are classified into groups of adjectives, nouns, transitive verbs, and intransitive verbs, each list arranged in chance order alphabetically and further subdivided into groups of one, two, and three syllables. Furthermore each of these groups has been once tested for the suitability of its words to the word associa- tion reaction and separated into selected words, rejected words, homonyms. It is insisted that this evaluation of the stimulus words can by no means be considered final or of universal fitness, but will surely serve greatly in a proper selection of words for a particular purpose in view. The complete lists without eliminations were used once with four subjects to study the adjective-noun, noun-adjective, verb- object and verb-subject controlled associations, and to obtain data for the selection and rejection of words. The adjective-noun and noun-adjective controlled association was more intensively studied on eight new subjects using 1800 words from the revised lists, keeping the number of stimulus nouns and adjectives equal, and constant for all subjects at each STUDYING CONTROLLED WORD ASSOCIATIONS 419 of the ten sessions, and preserving an exact balance in the order of presentation of the various groups. In both parts of the experiment the associations were recorded, also the reaction tunes, which were measured with the Johns Hopkins chronoscope. Averages and mean variations were calculated on the Burroughs adding machine, using the Dunlap adding machine formula for the mean variations. From this data it was purposed to observe whether there is any definite tune relation between the adjective-noun and the noun-adjective associations, or between the verb-object and verb-subject association and whether the reaction time has any fixed relation to the number of syllables in the stimulus word, the position of the accent or the logical category to which the stimulus word belongs. Some data was secured on double associations and associations to repeated stimulus words, and the emotional adaptation of the subject to the experiment is discussed. The following conclusions were drawn from the results : 1. The normal order for adjectives and nouns in the English language gives on the average a shorter association time than the inverse order, when the number of syllables in the compared groups is the same. That is, the reaction time for the adjective- noun association is shorter than for the the noun-adjective as- sociation. This holds when the groups are made up of carefully selected words, or when the stimulus words are all difficult, i.e., the words rejected from the revised lists, or when the stimulus words are still unevaluated and therefore include both difficult and easy words. 2. The reaction time for both the adjective-noun and noun- adjective associations increases on the average directly with the number of syllables in the stimulus word, and as in (1) holds for carefully selected stimulus words, or difficult words, or mixed groups of words containing both selected and difficult words. 3. The position of the accent in the stimulus word has no systematic effect on the reaction time for either the adjective- noun or noun-adjective association. 4. There is no interpretable variation in reaction time on the average according to the grouping of stimulus nouns into 420 MILDRED WEST LORING logical categories. The reaction time remains relatively con- stant within the limits of word length when large groups of stimulus words are used. 5. The reaction tune for the verb-subject association is longer on the average than for the verb-object association for stimulus words having the same number of syllables. As with the nouns and adjectives the normal order for the English language gives the shorter reaction tune. 6. For both the verb-subject and verb-object associations the reaction tune on the average increases directly with the number of syllables in the stimulus word. 7. Two or three separate hours of work are sufficient for the subject to become adapted to the procedure of the word asso- ciation experiment, and to lose as much as is possible of any emotional disturbance resulting from the novelty of his environ- ment or from sex difference between subject and experimenter. 8. The begining of each list of words is usually marked by three or four reaction tunes faster than the average. There is then a sudden increase which persists throughout the list, and which may be due to the adjustment of the subject to a comfortable steady muscular "tension" adapted to a long period of work. 9. Successive repetitions of a list of stimulus words for both the adjective-noun and the noun-adjective associations, cause successive decreases in the reaction tune when the subject is not informed that the stimulus words are words that have been given before. These reaction times would no doubt ultimately reach a physiological level. 10. The reaction tune for double associations is longer on the average for the inverse order than for the normal order in the English language. That is, the reaction time is longer in asso- ciating from intransitive verbs backward through noun subjects to adjectives than forward from adjectives through noun sub- jects to verbs. 11. In all controlled word association experiments, any inter- pretation of reaction times must surely take into consider- ation the type of control put upon the associations, and the STUDYING CONTROLLED WORD ASSOCIATIONS 421 number of syllables in the stimulus word. It is further insisted that conclusions are valid only when drawn from extended data using large groups of words as have been used in this investigation. FURTHER TECHNICAL PROBLEMS IN THE WORD ASSOCIATION METHOD In connection with the application of the Association Method to practical problems, and for the further understanding of the associative processes and conditions themselves, a considerable number of special points remain to be investigated. Obviously, the accurate use of the association-reaction as a tool for the investigation of mental conditions demands the fullest possible knowledge of the laws of the reaction itself. The problems included in the following list have been formulated in this labora- tory for experimental attack, and it is hoped that work on some of them may be under way before long. 1. A more detailed study of the course of the reaction time for repeti- tions of a group of stimulus words of different types, as sug- gested in Experiment III. 2. A more detailed study of double and triple associations along the line suggested in Experiment IV. 3. The effect of suggestion on the reaction time variation for dif- ferent types of forward and backward controlled associations. 4. Comparison of visual and auditory methods of presentation of stimulus words in the word association method. 5. A comparison of spoken and written types of reaction in the word association method. 6. The variation in reaction time for the adjective-noun and noun- adjective reaction in French, Italian, and Spanish subjects. 7. A study of the emotions in the word association method, when supplemented by plethysmographic, cardiographic, sphygmo- graphic, pneumographic, and galvanometric controls. 8. Sex differences in various types of controlled word associations. 9. Investigation into the comparative length of reaction times for the following types of controlled word associations; (1) noun subject — intransitive verb noun subject — transitive verb noun subject — verb (either v.i. or v.t.) 422 MILDRED WEST LORING (2) intransitive verb — noun subject transitive verb— noun subject verb (either v.i or v.t) — noun subject (3) transitive verb — noun object nouns object — transitive verb (4) noun subject — verb verb — noun subject (5) Comparison of English and German subjects for (3) and (4), (6) noun subject — (verb) — noun object noun object — (verb) — noun subject (7) noun (cause) — verb (effect) verb (effect) — noun (cause) (8) verb — adverb adverb — verb (9) class (genus) — member (species) member (species) — class (genus) (10) opposites for verbs (11) opposites for adjectives (12) opposites for adverbs (13) coordinate members. 10. A study of preferential associations. (1) to observe whether transitive or intransitive verbs occur more often in the noun subject— verb reaction. (2) to observe whether noun subjects or noun objects occur more often in the verb — related noun reaction (3) to observe whether noun subjects or noun objects occur more often in the noun — noun (related through action) reaction (4) to observe which type of reaction occurs more often in the noun — logically related word (other than verb) reaction; adjective, subordinate, supraordinate, or co-ordinate. (5) to observe which type of reaction occurs more often in the adjective — logically related word reaction; substantive or opposite. REFERENCES (1) GALTON, F.: Psychometric experiments. Brain, ii, 1879. GALTON, F.: Psychometric experiments. Inquiries into human faculty. 1883. (2) WUNDT, W. : Ueber psychologische Methoden. Phil. Stud., i, 1883. STUDYING CONTROLLED WORD ASSOCIATIONS 423 (3) TRAUTSCHOLDT, M. : Experimentelle Untersuchung ueber die Association der Vorstellungen. Phil. Stud., i, 1883. (4) WAHLE, R. : Bemerkung zur Beschreibung und Eintheilung der Ideen-asso- ciation. V.-schr. f. wiss. Phil., ix, 1885. (5) CATTELL, J. M. : The time it takes to see and name objects. Mind, xi, 1886. (6) CATTELL, J. M. : Psychometrische Abtheilungen, I und II. Phil. Stud., iii, 1886. (7) CATTELL, J. M. : Experiments on the association of ideas. Mind, xii, 1887. (8) CATTELL, J. M., AND BRYANT, S. : Mental association investigated by experi- ment. Mind, xiv, 1889. (9) SCRIPTURE, E. W. : Ueber den associativen Verlauf der Vorstellungen. Phil. Stud., vii, 1892. (10) MUNSTERBERG, H. i Studien zur Associationslehre. Beitrage zur experi- mentellen Psychologie. Heft. iv. (11) TITCHENER, E. B. : The Leipsic .School of Experimental Psychology. Mind, N. S. I, 1892. (12) WUNDT, W. : Bemerkungen zur Associationslehre. Phil. Stud., vii, 1892. (13) BOURDON, B. : Recherches sur la succession des phenomenes psychologiques. Rev. Phil, xxxv, 1893. (14) HOWE, H. H. : Mediate association. Am. J. Psych., vi, 1893. (15) BERGSTROM, J. A., Relation of the interference to the practise effect of an association. Am. J. Psych., vi, 1894. (16) CALKINS, M. W.: Association. Studies from the Harvard Psychological Laboratory. Psych. Rev. i, 1894. (17) JASTROW, J. : Community and association of ideas. Pysch. Rev. i, 1894. (18) JERUSALEM, W. : Ein Beispeil von Association durch unbewusste MitteK glieder. Phil. Stud, x, 1894. (19) SMITH, W. G. : Mediate association. Mind, N. S., iii, 1894. (20) WARD, J. : Assimilation and association. Mind, N. S., iii, 1894. (21) WUNDT, W.: Sind die Mittelglieder einer mittelbaren Association be- wasst oder unbewusst? Phil. Stud., x, 1894. (22) ALLIN, A. : Ueber die Grundprincip der Association. Diss. Berlin, Mayer und Mueller, 1895. (23) DUMONT, C.: Del la ressemblance et de la contiguite dans Tassociation. Rev. de Metaph. et de Mor., iii, 1895. (24) ASCHAFFENBURG, G. : Experimentelle Studien ueber Associationen. Psych. Arbeiten, i, 1896, ii, 1899, iv, 1904. (25) CALKINS, M. W. : Association. Studies from the Harvard Psych. Labora- tory. Psych. Rev., iii, 1896. (26) CALKINS, M. W. : Association: An essay analytic and experimental. Psych. Rev. Monog. Suppl., i, 1896. (27) KRAEPELIN, E. : Der psychologische Versuch in der Psychiatric. Psych. Arbeiten, i, 1896. (28) ZIEHEN, T. : Die Ideenassoziation des Kindes. 1 u. 2 Abh. Sammlung v. Abh. aus dem Geb. der. pad. Psych, u. Phys., i, 1896, iii, 1900. (29) JOST, A. : Die Assoziationsfestigkeit in ihrer Abhangigkeit der Verteihmg der Wiederholungen. Ztschr. f. Psych., xiv, 1897. 424 MILDRED WEST LORING (30) REGNAULT, F. : Des associations d'ictees dans le ge"nie. Medicine Mod., viii, 1897. (31) CALKINS, M. W. : Short studies in memory and in association from the Wellesley College Laboratory. Psych. Rev., v, 1898. (32) DEFFNER, K. : Die Aehnlichkeitsassociation (Diss. Munich., 1898). Ztschr. f. Psych., xviii, 1898. (33) GOBLOT, E. : Sur la theorie physiologique de Passociation. Rev. Phil., xlvi, 1898. (34) VAN DER PLAATS, N. ; Vrije Woord-Associatie. Akademische Proefschrift., Amsterdam, 1898. (35) SOMMER, R. : Lehrbuch der psychopathologischen Untersuchungsmethoden, 1899. (36) VAN ERP TAALMAN KIP, M. J. : Psychiatrische en neurologische Bladen, vi, 1899. (37) BOURDON, B. : Le type grammatical dans les associations verbales. C. R. IVe Congres Int. de Psychol., 1900 (1901). (38) WRESCHNER, A. : Eine experimentelle Studie ueber die Assoziation in einem Fall von Idiotic. Allgem. Ztschr. f. Psychiatric, 1900. (39) CORDES, G. : Experimentelle Untersuchungen ueber Associationen. Phil. Stud., xvii, 1901. (40) KELLOR, F. A. : The association of ideas. Pedag. Sem., viii, 1901. (41) THUMB A., UND MARBE, K.: Experimentelle Untersuchungen ueber die psychologischen Grundlagen, der sprachlichen Analogiebildung. Leipzig, 1901. (42) MAYER, A., UND ORTH, J. : Zur qualitativen Untersuchung der Association. Ztschr. f. Psych., xxvi, 1901. (43) BOURDON, B. : Contribution a T4tude de I7 individuality dans les associations verbales. Phil. Stud, xix, 1902 (Festsch.). (44) SCHMIDT, F. : Experimentelle Untersuchungen zur Associationslehre. Ztschr. f. Psych., xxviii, 1902. (45) CLAPAREDE, E. : L'association des Ide*es. Biblotheque Internationale de Psychologic Experimental. Paris. O. Doin. 1903. (46) FUHRMANN : Diagnostik und Prognostik der Geisteskrankheiten. Leipzig, 1903 (47) PIERON, H.: L'association mediate. Rev. Phil., Ivi, 1903. (48) PIERON, H. : Les experiences sur 1'association des ide*es. Rev. de Psychiat., 3e S., vii, 1903. (49) SMITH, W. G. : The range of immediate association and memory in nor- mal and pathological individuals. Arch. Neur., ii, 1903. (50) WUNDT, W. : Grundziige. 5te Auflage, S. 526, Leipzig, 1903. (51) BLEULER, E.: Diagnostischestudien Vorwart ueber die Bedeutung von Assoziationsversuchen. J. f. Psych, u. Neur., iii, 1904. (52) FERE, C.: Note sur le role des conditions somatiques dans 1'association des ide"es. Anne"e Psych, xi, 1904 (1905). {53) JUNG, C. G., UND RIKLIN, F.: Diagnostische Assoziationen. J. f. Psych. Neur., iii, 1904. .(54) JUNG, C. G. : Analyse der Assoziationen eines Epileptikers. J. f. Psych. und Neur., v, 1904. STUDYING CONTROLLED WORD ASSOCIATIONS 425 (55) GROSS, A. : Zur psychologische Tatbestandsdiagnostik. Mon. schr. f. Krim- inalpsychologie und Strafrechtsreform, 1904. (56) LIPMANN, O. : Die Wirkung der einzelnen Wiederholungen auf verschieden starke und verschieden alte Assoziationen. (57) RIKLIN, F. : Ueber die diagnostischen Bedeutung von Assoziationsversuch- ungen bei Hysterischen. Centbl. f. Nervenhkl. u. Psychiat., 1904. (58) WATT, H. J. : Ueber Assoziationsreaktionen, die auf optische Reizworte erfolgen. Ztschr. f. Psych, u. Phys. d. Sinnesorgane., xxxvi, 1904. (59) WEHRLIN, K. : Ueber die Assoziationen von Imbecilen und Idioten. J. f. Psych, u. Neur., iv, 1904-1905. (60) WERTHEIMER, M., UND KLEIN, J.: Psychologische Tatsbestandsdiagnostik. Arch. f. Kriminalanthropologie, xv, 1904. (61) EDGELL, B.: Experiments on association. J. of Phys., xxxii, 1905. (62) GROSS, A. : Die Associationsmethode im Strafprocess. Ztshr. f . d. gesamte Strafrechtwissenschaft, xxvi, 1905-1906. (63) GROSS, A.: Zur psychologischen Tatbestandsdiagnostik. Beitrage zur Psych, der Aussage, ii, 1905-1906. (64) GRABOWSKY, A. : Psychologische Tatbestandsdiagnostik. Sonderabdruck, 1905. (65) GROSS, H. : Zur psychologischen Tatbestandsdiagnostik. Arch. f. Krim- inalanthropologie u. Kriminalistik., xix, 1905. (66) ISSERLIN, M. : Zur psychologischen Tatbestandsdiagnostik. Monatsschr. f. Psychiatric u. Neur., xviii, 1905. (67) JUNG, C. G.: Uber das Verhalten der Reaktionszeit beim Assoziations- experimente. J. f . Psychiatric u. Neur., vi, 1905 (68) JUNG, C. G. : Experimentelle Beobachtungen ueber das Erinnerungsvermo- gen. Centbl. f. Nervenhkl. u. Psychiat., 1905. (69) JUNG, C. G.: Zur psychologischen Tatbestandsdiagnostik. Centbl. f. Nervenhkl. u. Psychiat., 1905. (70) KRAMER, F., UND STERN, W. : Selbstverrath durch Association. Beitrage zur Psychologic der Aussage., ii, 1905-1906. (71) PETERS, W.: Ueber Erinnerungsassoziationen. Ber. d. Ill Kongress f. Exper. Psych., 1905. (72) RIKLIN, F. : Analytische Untersuchungen der Symptome und Assoziationen eines Fall von Hysteria. Psychiatrisch-neurol. Woch., 46-52, 1905. <72a) WATT, H. J. : Experimentelle Beitrage zur einer theorie des Denkens. Arch, f. d. ges. Psych., iv, 1905. (73) ARNOLD, F. : The psychology of association. Columbia Univ. Contribu- tions to Philos. and Psych., 1906. (74) BALDWIN, B. T. : Association under the influence of difference ideas. Har- vard Psych. Studies, ii, 1906. (75) HAHN, R. : Ueber die Beziehung zwischen Fehlreaktionen und Klangas- soziationen. (Dis.) Zurich, 1906. (76) JUNG, C. G.: Diagnostische Associationsstudien. Bd. I. Leipsig, Barth, 1906. (77) JUNG, C. G. : Die psychologische Diagnose des Tatbestandes. Juristisch- Psychiatrische Grenzfragen. Halle, 1906. 426 MILDRED WEST LORING (78) JUNG, G. C.: Ueber die Bedeutung des Associations experimentes f. d. Psy- chepathologie. Gross Arch. f. Kriminalanthr. u. Kriminalistik, 1906. (79) KIESOW, F. : Ueber sogenannte frei steigende Vorstelhmgen usw. Arch, f . d. gesamte Psych., vi, 1906. (80) LEDERER, M. : Die Verwendung der psychologischen Tatbestandsdiagno- stik in der Strafrechtspraxis. Mon. Schr. f. Kriminalpsych, 1906. (81) LEVY, M. : Studien iiber die experimentelle Beeinflussung des Vorstellungs- verlaufs. I u. II, Ztschr. f. Psych., xlii-xliv, 1906-1907. (82) MESSER, A.: Experimentell-psychologische Untersuchungen liber das Denken. Arch. f. d. ges. Psych., viii, 1906. (83) PICK, A.: Riickwirkung sprachlicher Perseveration auf den Associations- vorgang. Ztschr. f. Psych., xlii, 1906. (84) WERTHEIMER, M. : Experimentelle Untersuchungen zur Tatbestandsdiag- nostik. Arch. f. d. ges Psych., vi, 1906. (85) DROMARD. : De la "Plasticite"" dans 1'association des ide"es. Rev. Phil., Ixiv, 1907. (86) FtiRST, E.: Statistische Untersuchungen iiber Wortassoziationen und tiber familiare Uebereinstimmung. usw. J. f. Psych u. Neur., ix, 1907. (87) JUNG, C. G. : On psychophysical relations of the associative experiment. J. of Abnorm. Psych., i, 1907. (88) NEUMANN, E. : Uber Assoziationsexperimente mit Beeinflussung der Re- produktionszeit. Arch. f. d. ges. Psych., ix, 1907. (89) SEVERANCE, E., AND WASHBURN, M. F. : Loss of associative power in words after long fixation. Am. J. Psych., 1907. (90) WATT, H. J. : Ueber den Einfluss der Geschwindigkeit der Aufeinander- folge von Reizen auf Wortreaktionen. Arch f . d. ges. Psych, ix, 1907. (91) WRESCHNER, A. : Die Produktion und Assoziation von Vorstellungen. Zts. f. Psych, u. Phys. Erganzungsband, iii, Theil I u. II, 1907-1909. (92) SALING, G. : Associative Massenversuche. Ztschr. f. Psych., xlix, 1908. (93) SCROLL, K. : Versuche iiber die Einfiihrung von Komplexen in Assoziationen. von gesunden und Geisteskranken. Klin. f. psych, u. Nerv. Krankh., iii, 1908. (94) ACH, N. : Ueber eine Methode zur Untersuchung der simultanen Assozia- tionen. Bericht d. Ill Kongress f. Exper. Psych., Leipzig. 1909. (95) DUNLAP, K. : A new key for reaction time work. Psych. Monog, x, 1909. (96) SCHULTZE, F. E. O. : Bericht iiber Assoziationsversuche des psych. Inst. der Frankfurter Akad. Ber. d. Ill Kongress f. Exper. Psych. 1909. (97) THUMB, A.: Assoziationsversuche im Dienste der Sprachwissenschaft. Ber. d. Ill Kongress f. Exper. Psych., 1909. (98) STEIN, P. : Tatbestandsdiagnostische Versuche bei Untersuchunggefangen. Ztschr. f. Psych., iii, 1909. (99) YERKES, R. M., AND BERRY, C. S. : Association reaction method of mental diagnosis. Am. J. Psych., xx, 1909. (100) BOVET, P.: L'originalite' et la banalite" dans les experiences collectives dissociation. Arch, de Psych., x, 1909. (101) EASTMANN, M.: To reconsider the association of ideas. J. of Psych., etc., vii, 1910. STUDYING CONTROLLED WORD ASSOCIATIONS 427 (102) FRANZ, S. I. : Some considerations of the association word experiment. Bull. Gov. Hops, for Insane., ii, 1910. (103) FREUD, S. : Origin and development of psycho-analysis. Am. J. Psych,. xxi, 1910. (104) GEISSLER, L. R. : A preliminary introspective study of the association- reaction consciousness. Am. J. Psych., xxi, 1910. (105) JOUSSAIN, A. : Le cours de nous idees. Rev. Philos., Ixx, 1910. (106) JUNG, C. G. : The association method. Am. J. Psych., xxi, 1910. (107) KENT, G. H. AND ROSANOFF, A. J. : A study of association in insanity. Am. J. of Insanity., Ixvii, 1910. (108) LANGFELD, H. S. : Suppression with negative instruction. Psych. Bull., vii, 1910. (109) LEACH, H. M., AND WASHBURN, M. F. : Some tests by the association re- action method of mental diagnosis. Am. J. Psych., xxi, 1910. (110) NUNBERG, H. : Diagnostische Assoziationsstudien. J. f. Psych, u. Neur., xvi, 1910. (111) OHMS, H. : Untersuchung unterwertiger Assoziationen mittels des Wort- erkennungsvorgangs. Ztschr. f. Psych., Ivi, 1910. (112) REINHOLD, F. : Beitrage sur Assoziationslehre auf Grund von Massen- versuchen. Ztschr. f. Psych., liv, 1910. (113) RITTERHAUS, R. : Die Komplexforschung. J. f. Psych, u. Neur., xv, 1910. (114) RUSK, R. R. : Experiments on mental association in children. Brit. J. Psych., iii, 1910. (115) WHIPPLE, G. M. : Manual of mental and physical tests. Baltimore, 1910. (116) DAUBER, J. : Tiber bevorzugte Assoziationen und vertwandte Phanomene. Ztschr. f. Psych., lix, 1911. (117) DUNLAP, K. : The fall-hammer, chronoscope and chronograph. Brit. J. of Psych., iv, 1911. (118) FOUCAULT, M. : Etude experimentale sur Passociation de ressemblance. Arch, de Psych., x, 1911. (119) GOETT, T. : Associationsversuche an Kindern. Ztschr. f. Kinderhkl., cv, 1911. (120) HUBER, E. : Assoziationsversuche an Soldaten. Ztschr. f. Psych., lix, 1911. (121) LANGFELD, H. S. : Suppression with negative instruction. Psych. Rev., xviii, 1911. (122) LEVY-SUHL, M. : Studien iiber die experimentelle Beeinflussung des Vorstel- lungsverlaufes. Ztschr. f. Psych., lix, 1911. (123) LEY AND MENZERATH: L'e"tude experimentale de Fassociation des id£es dans les maladies mentales. Gand: Van der Haeghen, 1911. (124) WELLS, F. L. : Some properties of the free association time. Pysch. Rev., xviii, 1911. (125) WELLS, F. L. : A preliminary note on the categories of association reactions. Psych. Rev., xviii, 1911. (126) WELLS, F. L. : Practice effect in free associations. Am. J. Psych., xxii, 1911. (127) WELLS, F. L., AND FORBES, A. : On certain electrical processes in the human body and their relation to emotional reactions. Arch, of Psych., xvi, 1911. 428 MILDRED WEST LOEING (128) WHITLEY, M. T. : An empirical study of certain tests for individual dif- ferences. Arch, of Psych., xix, 1911. (129) WOODWORTH, R. S., AND WELLS, F. L. : Association tests. Psych. Rev. Monog. Supp., xiii, 1911. (130) APTEKMANN, E.: Experimented Beitrage zur Psychologic des psycho- galvanischen Phanomens. Jahr. f. Psychonanal. u. psychopath. Forsch., 1912. (131) ATHEKTON, M. V., AND WASHBUBN, M. F. : Mediate associations studies by the method of inhibiting associations. Am. J. Psych., xxiii, 1912. (132) CULLER, A. J. : Interference and adaptability : an experimental study of their relation, etc. Arch, of Psych., xxiv, 1912. (133) DUNLAP, K. : Hipp chronoscope without armature springs. Brit. J. Psych., v, 1912. (134) EASTMANN, F. C., AND ROSANOFF, A. J. : Association in feeble-minded and delinquent children. Am. J. of Insanity, Ixix, 1912. (135) NAGLE, F. : Experimentelle Untersuchungen iiber Grundfragen der As- sociationslehre. Arch. f. d. ges. Psych., xxiii, 1912. (136) PFENNINGER, W. : Untersuchungen uber die Konstanz und den Wechsel der psych. Konstellation bei normalen und Friihdementen. Jahr. f. Psychoanal. u. psychopath. Forsch., 1912. (137) PFERSDORFF: Monat. f. Psychiat. u. Neur., xxxi, 1912. (138) WELLS, F. L. : The question of association types. Psych. Rev., xix, 1912. (139) WOHLGEMUTH, A. i Ueber die Richtung der Assoziationen. Ber. iiber d. V Kongress f. Exper. Psych., Leipzig, 1912. (140) DUNLAP, K. : Apparatus for association timing. Psych. Rev., xx, 1913. (141) KELLEY. T. L. : The association experiment: Individual differences and correlations. Psych. Rev., xx, 1913. (142) ROSANOFF, I. R., AND A. J. : A study of associations in children. Psych. Rev., xx, 1913. (143) SUTHERLAND, A. H. : Critique of word association reactions. Banta Pub. Co., Menasha, Wis., 1913. (144) WELLS, F. L. : The personal factor in Association reactions. Am. J. of In- sanity, Ixix, 5, 1913. (145) WELLS, F. L. : Professor CattelFs relation to the association method. Columbia Contribs. to Phil and Psych., xxii, 1914. (146) CLAPAREDE, E. : Experiences sur la me"morie des associations spontan^es. Arch, de Psych., Ixix, 1915. (147) CRANE, H. W. : A study in association reaction and reaction time. Psych. Rev., Monog. Supp., xviii, 1915. (148) DOOLEY, L. : A study in correlation of normal complexes by means of the association method. Am. J. Psych., xxvii, 1916. WORD-LISTS FOR ADJECTIVE AND NOUN REACTIONS MILDRED WEST LORING University of Washington, Seattle The following word lists, were employed by Dr. Loring in experiments II, III and IV of "Methods of Studying Controlled Word Associations (this journal, current volume, pages 369^428). Each day's work is given in the order in which it was presented, and the order of days is preserved. The method of selecting the words is described on pages 374-376, and 397-400. The consideration of these lists in connection with the range of distri- bution of reaction times is important, as it is impossible to obtain the necessarily long lists of words of equal familiarity. These lists being the results of two successive elimination processes, furnish a good basis for further reduction for work where shorter lists may be used, and greater uniformity, with a certain grade of average familiarity, may be desirable. The complete lexicon of verbs, nouns and adjectives which remained after the two eliminations will be issued later : but in the mean time these lists may be useful as sources. The necessity of making up lists of stimulus words by methodical selection from longer, mechanically prepared, lists, and so avoiding the associative processes of the one who makes up the lists, is obvious. 429 430 MILDRED WEST LORING BB H S ^ "5 rt o> o> S -2 += o .2 >» t* 3 ft -* « «oi?i^oa>>-ca> 02.2.^^ Ot2 03 Oi -^p'osos'Sacj fi > «£3 M « a WORD LISTS 431 itiiiiiiiiiiiiyiiiiiiiiiiiiii i S3 filial I"118 c T; -^ •— •> "i >-i o •""ft 2 § -g d a>&illiii>Ifl •r-^,S^2FMa>C3'9o til ii itii in i iiii^iiiti ill § n i ii ~ s «»»lf I s -a .-a I ^ 5 ^ M r^ w in 432 MILDRED WEST LORING l i o ^4 J® tfl •£ > O •** QJ +0 ^ • -J ^ Si—* C bfiCL"2 -»'r? O"titT^jj o-r^ a) r~^ 5 xj -s ft^ S 2 5 •- 2 S *S "55 ® '~ ® " .2*«2 OrjCn^rjO? Q}«4-i'"oSrt3C! >.S o3Cfl.5T3^3 ScC 0^3 c3 S II bC o.S s s^ •= g ft 03 02 > |i B-'i i§-g g g. i glililid1 liUIIII s«d v ^W > >> 8 S ts fe ^^"§1 III .•S .S .£T£ S WORD LISTS 433 °°H K S » O i H If II =3. 3 o S o '£ o •*3 nj -t-= -*f •*» O o « -^ajCJ^Cgc,^ a ocjoS-gcj^osg.S^H S .S 3 .K S^ 5-g g-^ta'o fl .•** S 2 *| 3 1 H P CQ P .laffllf.*?!. 03 •" "S -S ^3 bc«2 'a >> CQ 8 60 s d 'C .a M g| g^l^^^o^ 1^1 II I ll II II I -II ^ d g fill •*» ,£3 ^! bC § ^.SP fl S 'S 4n W 111 ^a 5o^ §^5 agagcn.g ^-S 3§Sbc5ag£HCDrSg I &£ ^S^Jo^l d ^ -s •a •S-s-Sai3^o.§ PSTCHOBIOLOGT, VOL. I, NO. 6 434 MILDRED WEST LORING .g 1 1 1-1 III Jgl •£•25 o p. j ^ "I § g bl^-2^-^ ^oo§.S>:3la?2'i § -a ^ .« ^ ^ 4 I I 8 1 § 11 111 * j* *J FJ o o ^ < I S ' t->>~^>» caS-3 ^ • - >» ^G-S4d OlJ^Os^^^-^C; llitilfllfrlll&llilllililllll §• 1 e I -S -s a s s I -s o s a ! all -S a-s & § -s £ I § -s j I JS, SKp-S fllliliililiifi Illllllfllllill f.i.UlilliJliiillif klififfji^ S a s a B J P B •» *•« f » a « l f 1 1 8 •» » I * i i -e j j WORD LISTS 435 l^iHliijflf JsJ *> S ^ * a g.s'i.d (U llliliil ^^ra^^gbC^ •a .a V ^^^ i s s § a-a llfllillilsiillflltiilifiilll! 436 MILDRED WEST LORING § -g S g .§ §"8 ri | n ^ c iilir%ll!|!lliflil| 3 g 6 25 a = s 5 S .1 II I | -I S I £ 2 1 § | D ^ A -49 6 fl8 V A «O O v*A M >•' a & a tf ft O Cr fl 111! ill I ^oajH^flSat; a>p4^(uo3op >,^ •SJ -S^.2^^ ^3 . -S^^-d "rft-r! H '55 TH CL -P CT'o a; _o « islfftlill ^°3o5o3^^o3'"a2c bDfli-HpQ UQQiJS £ o3 03 -i3 sgsll&lilflllll ^ bO^ d _rt H3 -fl'S'O o SoQ-d «*S ca 5.S,^-| JS"! g > WORD LISTS 437 § S a g 0 £ .g «> 8 s S o -a g g = _ & S3 a llllliiolilllililliilil.siiisl lllliillllllf iiilililiiilllli^ 5 T* •_£; x o .t; ^^ 2 12 -J3 3 o a 13 ~« 3 » ° ^ d^ fc b ^ fe S^5 ® -^ CM o3 ""O OJ 'O ^2 -M Q P< !"* oo 5E C1 y T^ ^ -rl f O O R **_'' r° "? •j-3 Pi 1* ?1 H ?j is d£ SS |3 H ^ * -a g I . •§ g _ I ?•! ? g I ^ 1 1 1 .1 i-ll •§ I s 1 1 1 |||1 llsliisilli ta.;3mo,fl;i-»n*tj?:^>o«:>®^ 43 ^2 o S ^ 1.1 5 X5 ^_ d 3 c8 QQ -M o 43 pe de TWO SYLLABLE ADJECTIVES uiiliilliititllH afio, ~g«-n3 o Sbc Ililflllillllll Jl !? iilillliililiiilJi^llil cnv-3 OQ ^pQ-*S^25 02 c? T3 «<-!j3 P £ =£ +-> _cj o r-i< ^5 unzJjQ 438 MILDEED WEST LORING T3 c3 *llfj13?l &ss!!lji .SP 5 I & £ . tv-hJ.tlJIli* llflillljl filial!8'-2 S.S So a S q^ o> t» 1 11 |1|| P O O ^J lit £2^0 1 O O «3 02 d 03 O E > 5 3 b 9 Jj 03 ijtfi]]ij]iiiM 0> q bD III *^it-§l-^i1lfl1 bC^D^ KJW^J £ 0 Sc,O e3 WORD LISTS 439 il LLAB IVES REE ADJ u.2 d a .S o s o s S Sob "§>>>> c || g"-*-' « gjB Xi 08 pfi 'In OnCQ C* O* 02 03 PnCQ^ S 2 Pi FH Oli i-2 03 ?- f-i 440 MILDRED WEST LORING •e B & iimililiililiii-i gS* O H3 ^3 Si 2! H 03 O ^ S TO .2 ^3 o ja 5a &a P «G T3 -O S | § S OT3 III!,, -^ ^ >» '•g'S Oj O 03 ^ ^3 02 > ft 0 «E O &R. | 111! -o a « S WORD LISTS 441 >> a & & 51 S S -S I .^f . —j i\ O . /^ .«< Tt 2 i-T C ^* ?r rv — u^ 1 s fl a c a 5Pn s€ $ fl *5 ^3 bo55ca o^ S » .*3 ^ jg So p,o Z S « ^43 442 MILDRED WEST LORING -s S 1 a .5 :§ -3 •§ I § a I § a« OJ 111 SB-si 43 e d a .2 »•>•>-( ai ~i *rS .2^ a S-Q.S «" s "S ^^"2 " *M O en o S -S "ilifliiiiii- .^p3PPrao7343jH4i Illllpll llillltl § a r-H O! s'-a Q) 43 « d .2 | 43 r^3 •a -g an o 48 ^^ ft o £ II 3 a a.s Oi I ¥ I 2 T1 -^ i ._. a „ .g -g I! In I ii 111! til WORD LISTS 443 II 1111 111 I JH o3 -M ^Q CL, o 18 «M "« 21 £ S 43 . 2 ^ S _, 12 £ t, S o -g ^ c? e I Ji 5 "3 !3 2 •& •! P3^ -^^ CQ-l^aj lllitlilJli I' •§£ 83 ^> ^ »••* 'a »—5 T* *-v V ^T L^ r> p, ,™ ^ S '-§ w -fa £ s 2 S 2 03 a **H 42 METHODS OF USING BALANCED-MAGNET CHRONOSCOPES KNIGHT DUNLAP The Johns Hopkins University The following schemes may be used equally well for installing the Johns Hopkins chronoscope1 or the Hipp as used by my method2 without armature springs. The diagrams (figs. 1 and 2) are given for the former instrument: in setting up the latter, the circuits which operate the synchronous motor of the Johns Hopkins chronoscope are of course to be omitted. The Johns Hopkins chronoscope operates either on (1) direct current, with tuning fork interruption for the motor circuit, or (2) alternating current for the motor circuit, without fork, and direct current for the clutch circuits. The alternating current of the usual house lighting circuit is 60 cycle, and power circuits are sometimes 25 cycle : it will be assumed that one or the other of these frequencies is to be employed. The first system is represented in figure 1 : the second system in figure 2. THE MOTOR CIRCUIT The motor circuit and the clutch circuit may be drawn from the same direct current source, which may be the 110 to 120 volt power circuit as indicated in figure 1, or may be storage battery or rotary transformer circuit. Otherwise, the motor circuit and the clutch circuit are entirely distinct, and may be usefully described separately. In using alternating current 110 to 120 volts (the potential 1 Dunlap: The Johns Hopkins chronoscope. 1917, Jour, of Exper. Psychol., ii, 299-252. 2 Dunlap : The Hipp chronoscope without springs. 1912, Brit. Jour, of Psychol., v, 1-7. Poffenberger and Morgan: The Hipp chronoscope: its use and adjustments. 1916, Jour, of Exper. Psychol., i, 185-199. 445 446 KNIGHT DUNLAP of the usual house lighting circuits) , lamp resistance or a rheostat may be used to control the current ; or a rectifier of the proper type may be used without resistance. The scheme with resist- ance is shown in figure 2. The standard four-receptacle voltage- reducing lamp board (lamp board no. 1) is here represented, but LAMP BOARD No.l STIMULUS LAMPS ELELCTRIC FORK Z5 or 50 Vibrations FIG. 1. CHRONOSCOPE OPERATED ON DIRECT CURRENT; THROUGH TUNING-FORK INTERRUPTER Set up for visual discrimination reaction utilizing same current source. BALANCED MAGNET CHRONOSCOPES 447 lamps are to be used in the two primary receptacles (the left- hand ones in the cut) only. Obviously a two-receptacle lamp board is all that is necessary, since the motor circuit is closed. One receptacle is sufficient if a single lamp of appropriate wattage fits the requirements, but two receptacles connected in parallel is better, since it makes possible lamp combinations to fit all cases. LAMP BOARD No. I CHRONOSCOPY A.C o— I IKHZOVHto- 1 FIG. 2. CHEONOSCOPE OPERATED ON ALTERNATING CURRENT Set up with voice keys for association reaction, with battery for clutch con- nections. Two 100 to 120 watt lamps (of the voltage of the circuit), or one 100 to 120 watt lamp and one 50 to 60 watt lamp, will be found satisfactory. There is considerable latitude in amount of current permissible in any case, since the speed of the motor is not dependent on the current, but on the frequency. Enough 448 KNIGHT DUNLAP current to give sure operation, and not enough to heat the motoi seriously, or make starting difficult, is the easy requirement. The motor will run on a single 60 watt lamp, although hard to start on so small current, and easily put out of phase and stopped by the slight jar of the clutch action. The current practically needed depends on the adjustment of the clutch, and the amount of current in the clutch circuit. On the average, lamps totaling 150 watts will do very well; 100 watts may be satisfactory; more than 240 should never be used; and more than 180 will heat the motor if used continuously. Lamp resistance is advised, because of its safety. The lamps act as tell-tales, showing clearly whether the current is on or off, and there is no danger of damage from accidentally putting on too much current, as is apt to happen in moving a rheostat handle in the wrong direction. By having the connections at the lamp board covered, all danger from shock is eliminated. If a rheostat be used instead of lamp resistance, one of 120 ohms total resistance, and 2 amperes continuous capacity is advised. The Jagabi 1506, James G. Biddle, fills these specifi- cations. The resistance should never be allowed to fall below 50 ohms; hence a stop should be put on the slide, a little less than half way from the zero end. The motor has been run experimentally on the whole range of frequencies between 15 and 120 per second: the conditions for this test were arranged by M. W. Pullen of the Department of Electrical Engineering, and I have not standardized the resis- tances to be used with the 25 cycle circuit. The determination of the proper resistance is very simple, but there are probably few places where this frequency would be available. The speed of the motor, when operated on the unrectified 60 cycle current may be too great. Since there are ten poles to the motor, and the 60 cycle current has 120 impulses per second, the armature will make 12 revolutions per second and the measur- ing unit of the dial will be one twelve-hundredth of a second. Not only is this unit needlessly small, even for the measurement of simple reaction times, but the difficulty in starting the motor is high (see below). With 25 cycle unrectified current the arma- BALANCED MAGNET CHRONOSCOPES 449 ture speed is 5 rotations per second, the measuring unit of the dial is 2 sigma — which is small enough for all purposes — and the motor is easily started. If a rectifier is used with the 60 cycle current the impulses are reduced to 60 per second, the measuring unit is one six-hundredth of a second, the motor starts easily and runs well without heating. I have found the small size "Tungar" rectifier, General Electric Company, very satisfactory when connected directly in the circuit without resistance. A small transformer may be used directly in the circuit, reduc- ing the voltage to the point required to force the proper amount of current through the field coils. This has the advantage of economizing current, a large proportion of which is wasted in the resistance of the above described arrangements. The trans- former should step the voltage down one-half; i.e., should reduce the 120 volt potential to 60, or the 110 to 55. It is necessary to bear in mind that the nominal cycles per second of a lighting or power circuit are only approximate, and that there is a slight variation from the standard from time to time. The alternating current can be used for demonstration purposes, and for practise work where accuracy is not essential, without check. But, for research purposes it is necessary to use a frequency-meter from time to time to determine the actual frequency; or else to obtain from the power house data on the diurnal variations in frequency. In many situations it will be found that the variations at the hours at which the apparatus is used are small enough to be neglected in the calculations. Since the frequency determines the value of the units of measure- ment of the chronoscope, reaction times obtained while running the motor on a current of uncertified frequency are not worth much. For the above reasons, it is recommended that the motor be operated on direct current, with tuning fork interruption where- ever possible. In using direct current in the motor circuit, an electric fork is necessary to interrupt the circuit periodically. If the 120 or 110 volt current is used, a voltage-reducing lamp board should PSYCHOBIOLOGY, VOL. I, NO. 6 450 KNIGHT DUNLAP be introduced. This set-up is shown in figure 1. In this lamp- board four 100 or 120 watt lamps are satisfactory, and all are left permanently screwed down. Four 120 watt lamps reduce the voltage to about 50, when the circuit is broken. Storage cells may be used as a source of direct current, the voltage of the battery depending on the fork somewhat. In some cases a standard 12 volt battery will serve, but a twenty to twenty-five volt set, with adjustable resistance capable of carrying 1 to 1.5 amperes and a maximum of 12 to 18 ohms is preferable. Edison storage cells are recommended on account of their requiring less care; and they may be used in the same room with the apparatus, as they have no fumes deleterious to steel as do the acid cells. The variations in voltage of the Edison cells do not affect the measurements. If the fork-magnet is high-wound, it may be necessary to arrange a shunt around it, to provide sufficient current for the motor field without forcing too much through the fork. I have not found this necessary with any of the standard wound forks. The fork contacts should be of dental gold alloy (gold and platinum), which is cheaper than platinum and far more satis- factory. Tungsten contacts might possibly be employed, but I am unable to say definitely that tungsten is satisfactory. Much trouble will be found if platinum be used and I do not find that the motor operates satisfactorily with a platinum contact fork. I find dental wire of 18 gauge (Brown and Sharpe) works well: 19 or 20 might perhaps be used. The table on which the wire strikes should be surfaced with a section (disc) of large gauge wire. With any fork interrupting a large current (more than a small fraction of. an ampere) a condenser ought to be used. The small size 2 microfarad condenser made by the Western Electric Company is cheap in price and quite satisfactory. The con- denser must be connected across the gap of the fork contact, and must not include the electro-magnet (of the fork) between its terminals. In other words, when the fork-contact is broken, the circuit must be traceable unbroken from one contact of the condenser through the magnet windings. For convenience, an BALANCED MAGNET CHRONOSCOPES 451 extra pair of binding posts should be mounted, one on the shank of the fork itself, or, as in the figure, on the metal mounting (base) of the fork (fig. 1, electric fork, 3) ; the other on the metal support of the fixed part of the contact. If the fork be connected as described, and then, while it is in operation, one of the condenser connections be momentarily broken, the large increase in the spark at the break of the fork-contact will demonstrate the value of the condenser in preventing burning up of the contact and consequent loss of adjustment. The adjustment of the fork-contact is a matter of trial and error. I find it best to adjust to the smallest amplitude of vibra- tion which gives uniform vibration. The contact should be broken when the fork is at rest. It should be hardly necessary to warn against touching the steel of the fork with the hand in starting the fork, but with students this is always necessary. Once in operation, the motor should run for hours with no adjustments. If the motor stops, it may be due to one of several causes. (1) Insufficient current. (2) Bad adjustment of the fork-contact, so that the break of the circuit is not sufficiently long. (3) Too tight adjustment of the bearing of the motor shaft. The adjustment is a delicate matter, but once made should be good for a long time. The shaft should not be loose, but should yet spin freely. (4) Too much current in the clutch circuit. Under this condition the jar of the clutch sets the armature "hunting" and it kills itself. The fact is to be empha- sized that the variations in current and contact do not affect the speed of the motor unless they stop it completely. STARTING THE MOTOR A simple synchronous motor will not start itself: the armature must first be put in rotation at the speed required by the fre- quency of the current, or the frequency of the interruption, and must "get in step" with the current phases. Given this start, the motor will continue to run as long as the current is adequate, until it is overloaded, or until a jar starts it "hunting/1 On the 25 or 50 vibration fork, on 25 cycle alternating current 452 KNIGHT DUNLAP with resistance, or 60 cycle alternating current with rectifier, starting the motor is simple. With the thumb and forefinger the armature shaft is given a spin in the proper direction. Wrap- ping a string around the shaft, to get a pull, is not satisfactory. On first trial with the thumb and finger method, a great many spins will perhaps be necessary before the proper knack is acquired. The commonest error is spinning too fast. Another common error is taking hold of the shaft for a fresh spin, before noting whether the preceding one was effective; thus spoiling a good start. After a little practise facility is acquired, so that the motor may be started with little difficulty; usually on the first, second or third spin. On the 60 cycle unrectified current, since the speed of rotation is relatively high, the finger and thumb method may be difficult. In that case, wrap a strip of adhesive tape — electrician's or surgeon's — around the shaft spirally to make a smooth single thickness, and start by drawing the fingers across the wrapped part of the shaft. It may be easier to spin the armature with current off, closing the main switch as the fingers leave the shaft. THE CLUTCH CIRCUIT The clutch circuit must always be direct current, and should have an amperage just sufficient to give sharp, not violent, action to the clutch. The exact voltage will necessarily vary with the accessory instruments used, and it is well to have a rheostat in the circuit, as in figure 2, so that the current may be properly controlled. The essential feature of the Johns Hopkins chronoscope, and the Hipp without armature springs is, that within certain limits, variations in the current strength do not alter the measurements, since the clutch is moved in by one electromagnet, and moved out by another which is the exact mate of the first in core and windings. The two magnets are connected in parallel so that variations in the current affect both equally. In consequence of this feature, it is necessary to avoid differences in resistance and in inductance between the two branches of the clutch circuit. BALANCED MAGNET CHRONOSCOPES 453 This requirement is to be borne in mind in connection with the details which follow. For the set-up in figure 2, with voice keys3 for the association reaction, storage cells connected in series to give over four volts will operate the clutch. The higher voltage with resistance, as in figure 1, is more desirable. In the arrangement shown in figure 1, a voltage-reducing lamp-board is used, with one 100 watt lamp and one 60 watt lamp in the primary (left-hand) pair of receptacles, and two 100 watt lamps in the secondary (right-hand) pair. With this arrangement the total current in the two branches of the clutch circuit, the master key being completely closed, is approximately 0.6 ampere. The current is divided through the two branches of the circuit through the master key (refer to figure 2 : figure 1 is more compli- cated), which, as will be noticed, completes the circuit through the anterior magnet (the non-rotating magnet) a moment before completing the circuit through the posterior (rotating) magnet in the other branch. This is necessary in order that the clutch shall always be locked in the non-rotating position when the key is pressed. The master key is a two-contact key of new design, involving three strips of spring brass, or better, or phos- phor bronze, so arranged as to be self-scouring, and therefore needing no gold contacts. When the master key is released, current is "off" both branches of the clutch circuit, thus permit- ting resetting of the register hand, and making the clutch circuit perfectly safe. OPERATION IN THE ASSOCIATION REACTION The set-up of figure 2 has now been completely described, with indication of alternatives in both motor circuit and clutch circuit. The method of operation will now be described before proceeding to the more complicated set up of figure 1. Operator and reactor being in position before their respective 8 For description of the voice key, see Psychological Review, 1913, xx, 250-253. The Ewald chronoscope described there as satisfactory proved later very unsatis- factory, it got out of adjustment and could not be put in repair again. 454 KNIGHT DUNLAP voice keys, the operator commands "ready:" the master key being open, his voice has no effect on the mechanism. Next he closes the master key : current now flows through both branches of the clutch circuit, the armature automatically taking its position against the non-rotating magnet. Holding the key closed, the operator speaks the stimulus word against the dia- phragm of the voice-key at the right of the cut: the vibrations of which cause the current through the anterior magnet to be interrupted momentarily, allowing the armature disc to be attracted to the posterior magnet. The armature shaft and the register hand thereupon commence to rotate, and continue until the reactor, by speaking the response-word against the diaphragm of the other (left-hand in cut), voice key, interrupts the current through the posterior magnet, thus allowing the anterior magnet to attract the armature disc and stop its rotation. With good adjustment, there is nearly perfect balance both of the latent periods of the two magnets, and of the slip of the armature on the pole of each, so that the time between the beginning of the interruption of the first branch and the beginning of the inter- ruption of the second branch, is accurately registered on the dial. The operator releases the master key as soon as the register hand stops, takes the reading, and sets the hand for the next reaction. SET-UP AND OPERATION FOR MEASUREMENT OF SIMPLE AND DISCRIMINATION REACTIONS It is essential that the circuit through the anterior magnet, after being opened by the stimulus, shall be reclosed before the circuit through the posterior magnet is opened, or else the registration will fail. In the association set-up, this reclosure is automatically effected by the voice key, and the time-relations of the reaction are such that this is always effective. For reactions employing a different sort of stimulation, such as light or faint sound, pro- vision must be made for reclosure. A variety of devices are possible for this purpose. Where a pendulum or shutter is employed to control a light stimulus, the breaking and remaking BALANCED MAGNET CHRONOSCOPES 455 of the stimulus branch of the clutch circuit may be arranged with a precise interval between. In other cases, a simple relay of low resistance and inductance has been employed with no appreciable error resulting, the unbalancing of the two branches being slight. In general, the use of a simple relay is not safe, and hence the balanced relay shown in figure 1 has been developed. As shown in the diagram, the closure of the master key will automatically set the armature lever of the relay against the stop on the side of magnet 3-4, and when the master key is completely closed, current will flow equally through this magnet and through magnet 1-2 : and since these magnets are duplicates, and one is in the circuit of each of the clutch magnets, the two branches of the circuit are perfectly balanced. On depressing the stimulus key, contact is made between 1 and 2 (of the stimulus key), thus giving the stimulus, at the same moment that contact is broken between 2 and 3, interrupting the circuit through the anterior chronoscope magnet. The armature lever of the relay being released from magnet 3-4, moves over to the opposite stop, and recompletes the circuit through the anterior clutch magnet, through the third relay magnet, thus again balancing the circuits, and assuring also that the armature lever of the relay will be held in the new position even after the current through magnet 1-2 is interrupted by the movement of the reactor. In the diagram, the pneumatic reaction key which has been found most reliable, is shown, without the rubber bulb which the reactor squeezes. The stimuli are schematically shown as flashes of two low- wattage mazda lamps, either one being switched in before the measurement. For discrimination, lamps of two colors might be used, enclosed both in the same box with a milk glass window, so that the color would always appear in the same place. In the set-up of figure 1 both master key and stimulus key must be held down from the time each is pressed until the end of the registration. The balanced relay is self-setting, so that the actual procedure is simple. The tune of action of the relay must be well inside of the reaction limit, and must be over 15 sigma, in order to allow leeway for the chronoscope clutch. 456 KNIGHT DUNLAP Action time between 20 sigma and 90 sigma may readily be secured, or if pathological or anticipatory reactions are to be measured, the action may be made within the limits of 20 and 50 sigma. For other types of stimulus, variations in the above scheme are required, but the balanced relay will in almost all cases be the main accessory. For the reaction circuit, no special appa- ratus or wiring are needed on account of the balanced circuit, since the reaction circuit does not have to be reclosed. GENERAL CAUTIONS The Johns Hopkins chronoscope, like the Hipp, requires careful testing to assure balance of the magnet latencies and of the armature slips, but once set up and thoroughly tested, no further checks are needed unless the instrument is allowed to get badly out of condition. We have found that chronoscopes which were properly balanced, have become unbalanced in shipment, either through bending of the armature disc or other injury. The anterior clutch magnet (or in later designs, the magnet-core) has so far been made so that it can be rotated by hand ; an adjust- ment which is not needed in most work, and can be dispensed with. This magnet bearing must be tight enough so that the magnet will not creep under the torsion of the armature disc at stopping. Oil on the magnet face or armature disc will cause clinging. The clutch must be kept clean. The chronoscope should be set on a square of thick harness felt. If the tuning fork is in the reaction room, it also should be set on harness felt. The hum of the apparatus is then reduced to a negligible intensity. In using voice keys, they should be supported on a heavy, solid table, to avoid their being susceptible to the jar of the master key and movements of the operator's hand. To remove the most probable source of error in the Johns Hopkins chronoscope, I propose to replace the friction clutch by a modification of the Hipp toothed clutch. It seems advis- .able also in the next model to gear back the register hand so that BALANCED MAGNET CHRONOSCOPES 457 the units of measurement will be larger. Five sigma units are small enough for any work, and 10 sigma units are most satis- factory for association reactions. A chronoscope arranged to read five sigma units on a 50 vibration fork will read ten sigma units on a 25 vibration fork. THE INFLUENCE OF THE DISTRIBUTION OF BRIGHT- NESSES OVER THE VISUAL FIELD ON THE TIME REQUIRED FOR DISCRIMINATIVE RESPONSES TO VISUAL STIMULI1 H. M. JOHNSON Captain, Sanitary Corps, U. S. A., Air Service Medical Research Laboratory, Hazelhurst Field, Mineola, Long Island, New York INTRODUCTION V The following pages contain a report of a study which was undertaken as a methodological preliminary to an extensive study of the influence on visual performance of the distribution of brightnesses over the visual field. The primary purpose was two-fold: (1) To ascertain whether the method employed is sufficiently sensitive to demonstrate the effects of such moderate differences in distribution of lighting as might be encountered under installations which are volun- tarily used; and (2) to train the subjects so thoroughly that their diurnal performance under a given condition would be free of important fluctuations due to the variable effects of fatigue and practice. This was to permit of their economical use in the more extensive work, which was designed to include a wide range of distributions of brightnesses and variable sizes of the test-field, under such controls as would justify direct intercomparison of the data obtained under the various external conditions. As the author has been prevented from carrying out the complete study, the results of the preliminary investigation are presented without apology, as indicating definitely the sensitivity and reliability of the method employed, and the differential effects of the particular environmental conditions under which 1 This paper from the Nela Research Laboratory, National Lamp Works of General Electric Company, Nela Park, Cleveland, Ohio. 459 460 H. M. JOHNSON the work was done. Reference to the primary purpose of the experiment will make it clear why the range of stimulus- variables was narrow, and why certain refinements of control were not employed. PRELIMINARY CONSIDERATIONS The problem of ascertaining suitable methods of investigating such questions was referred to the writer in the winter of 1915-16 by the director of the Nela Research Laboratory. Except for the excellent work of Cobb2 on the thresholds for detail and for differences in brightness, the field had been scarcely touched. Cobb's results, while interesting in themselves, are perhaps more important in that they demonstrate to the illumi- nating engineer (1) that sensory thresholds are exceedingly un- stable and uncertain things, even when the external situation and the physiological factors are optimal; so that large numbers of observations are necessary to satisfactory evaluation of the results; and (2) that an objective check on the reports of the observer is absolutely essential to intelligent interpretation. These facts have long been known to the psychologists, and they can be readily appreciated by any other scientists who may be brought to consider them. EXPLORATORY EXPERIMENTATION For ad hominem reasons it was inexpedient to adopt as highly "artificial" a method as that of determining discrimination- times without first testing the adequacy of some " simpler, easier and more natural" method. Accordingly the first exploratory work was in the attempt to locate such a method. The most obvious of the possible types of test was first selected for study. It was proposed to assign the subject some definite task, to the performance of which vision is indispensable; and, 2 Cobb, P. W. and Geissler, L. R. : The effect on foveal vision of bright sur- roundings. Psychological Review, xx, 1913, pp. 425-447; also Cobb, P. W. : same title, ibid., xxi, 1914, pp. 23-32; and Journ. Exper. Psychol., i, 1916, pp. 419-425, and pp. 540-566. DISCRIMINATIVE RESPONSES TO VISUAL STIMULI 461 after establishing control of other variables than vision, to use the speed and accuracy of the work done under a given lighting condition as an indicator of the appropriateness of the condition to visual performance. Such a method, while simple in principle, demands the fulfil- ment of certain requirements, which I may be pardoned for enumerating for the benefit of non-technical readers : 1. The work done under the different lighting conditions must be of equal intrinsic difficulty. Otherwise, differences in speed and accuracy might wrongfully be attributed to the influences of differences in lighting. 2. The work done under the different lighting conditions must be distributed among them so that the effects of fatigue and practice are evenly distributed. Otherwise, differential effects of these variables may be erroneously interpreted as effects of the differences in lighting; or they may serve to distort or to obscure genuine effects of differences in lighting. 3. The measure of the subject's performance must be obtain- able independently of his own reports or description. This point was mentioned in the comment on Cobb's work, but it cannot be too strongly emphasized. 4. A sufficiently large number of observations must be obtained from each subject to insure that such effects as may be obtained are not accidental. In other words, enough results must be accumulated to justify statistical treatment, and the application of standard measures of reliability to them. In the past ten years, the pages of American lighting journals have been burdened by reports of tests which fail to satisfy certain of the above requirements, especially the third and fourth. In the judgment of the writer — which at this point accords with that of the best psychological authorities — a proper regard for economy of time would not justify one in reading such reports farther than to verify the description of the defective methods employed. In the beginning of the present work, several types of work- material were examined and found unsatisfactory. The most promising material was arithmetical, compiled by Prof. Knight 462 H. M. JOHNSON Dunlap, for use with the Burroughs adding machine in studies of habit-formation. This material had been derived by subject- ing a constant to a certain systematic treatment which insured an even distribution of difficulties among the several sets of material, and minimized and regulated the occurrence of repeti- tions. The amount of this material available was too small for our purposes. Several weeks were therefore spent in the com- pilation of additional material according to a key furnished by Professor Dunlap, and a fair amount of satisfactory material was thus obtained. A lighting booth was then constructed; various types of distri- bution of illumination were procured, and an adding machine installed. As the machine is large, and under the conditions of operation fills a large part of the visual field with surfaces of varying types and degrees of reflectivity — some being black, others nickelled, and most of them highly polished — it was neces- sary to screen the machine from view of the subject, and require him to learn to operate it by the "touch*1 method. Six subjects were employed, five being undergraduate students in the college for women of Western Reserve University. The college is located several miles from the laboratory, and it proved impossible to maintain a regular or satisfactory schedule. It became evident after some weeks that the subjects would not master the "touch" system of operation in time to complete the preliminary experiment according to plan within the academic year. The diurnal variations in performance under a single lighting condition were too great to justify the expectation that the various lighting conditions would show differential effects, unless the lighting conditions under comparison were extreme, unless an enormous amount of work material was provided and used, or unless the preliminary training period was to be indefi- nitely prolonged. It was therefore decided to abandon the experiment in favor of a method which involved a less intricate muscular coordination, until more favorable conditions for the use of the "simpler" method could be obtained. The work method under fair conditions of operation may be valuable. DISCRIMINATIVE RESPONSES TO VISUAL STIMULI 463 DESCRIPTION OF THE METHOD ADOPTED The use of the time required for visual discrimination as an index of the appropriateness of a lighting condition implies that the time required for this form of reaction will be shorter and the dispersion of the individual reactions will be smaller under the more favorable lighting conditions than under the less favor- able ones. As some students of illuminating engineering may not immediately see the justification for this assumption, certain elementary physiological doctrines will be recalled. When a stimulus adequate to excite a sensory-motor response is presented to an organism, certain physiological changes occur, which for simplicity's sake are described as if they occurred within the fictitious "reflex arc.7' These changes are: 1. An electrical or chemical process is excited in the end-organ, or receptor. 2. An afferent neural impulse is transmitted to some selective ganglion. This may lie in the brain, the spinal cord, or some plexus. 3. An efferent neural impulse is transmitted to an effector, i.e., a muscle or gland. 4. The effector responds by secretion, contraction or change in tonicity. 5. An afferent neural impulse, originating in receptors con- tiguous to the effector, is transmitted to some center, and thence, as an efferent impulse, to the same or some other effector. All except possibly the last of the above series of physiological changes are regarded as a necessary condition of a change in "consciousness." The most satisfactory hypothesis of mind- body relationship, namely, the hypothesis of "functional corre- lation," assumes that these physiological changes occur simul- taneously with the changes in "consciousness." In the special case of responses to visual stimuli, the excita- bility of the receptors — i.e., the rods and cones of the retina- is greatly modified by the character of fixation, of accommodation and of adaptation. All these factors may be influenced by the distribution of brightnesses over the visual field. If the distri- 464 H. M. JOHNSON bution is unfavorable, the "receptiveness" of the sense-organ is diminished, and its latent period is increased. The rate of transmission of a neural impulse along a single axone is fairly constant under constant conditions and is not affected by continued repetition of the stimulus, provided the intervening period is not excessively short.3 A relatively large, and variable, portion of the time required for completing a neural arc is consumed in the synapses (i.e., the connections between single neurones) and especially in the central synapses, where the efferent pathways for a given afferent impulse are selected or blocked. If the stimulation of the sense- organ is not sufficiently intense — and inadequacy may be due either to insufficient intensity of the stimulus or to an unfavorable condition of receptiveness of the sense-organ — the first afferent impulse reaching a given center from the sense-organ may be too weak to be transmitted across the surfaces of separation of the connected neurones until it is reinforced by succeeding impulses.4 The latter may be transmitted over the same afferent pathway, or over a different one. The concept of "intensity" of an impulse may be formulated so as to accord with the "all or nothing" principle, on the assumption that the intensity of stimulation of a set of ganglionic synapses is determined, other factors being constant, by the number of afferent neurones excited; that the receptors belonging to the several afferent neurones may have different thresholds or different latent periods; that their activity is intermittent; and that the more favorable the receptiveness of the sense-organ the greater will be the number of receptors excited at the same instant by a given stimulus. If a neural impulse is conveyed to a synapse and not discharged across it, the energy may be stored for a time in the synapse instead of being immediately dissipated. Thus, the afferent impulse, though inadequate at first, may become adequate by reinforcement. Any factor, therefore, which may tend to reduce 'Howell, W. H.: A textbook of physiology (6th ed.). Philadelphia, W. B. Saunders Company, 1915. 4 Sherrington, C. S. : Integrative action of the nervous system. New Haven, Yale University Press, 1906. DISCRIMINATIVE RESPONSES TO VISUAL STIMULI 465 the sensitivity of the sense-organ would also operate to increase the latency of the synapses, and thus to increase the time required for reaction. In the special case of vision it has been shown that the retina is anaesthetic during rotational eye-movement,5 so that if fix- ation is disturbed, a longer time of exposure to the stimulus may be required for the accumulation of enough energy in the various synapses involved to overcome the resistance of the surfaces of separation. Increased difficulty of maintaining fixation under a lighting condition which does not permit of stability of adaptation, has been hypostasized by Cobb,6 and seems to be supported by the self-observation of a number of subjects, although objective evidence, such as photographic records, is lacking. There is a well-established tendency to frequent and excessive changes in the size of the pupil, as the various parts of the retina are succes- sively exposed to brightnesses to which they are not adapted. Great relative instability of the pupillary size may be a source of much discomfort, and it may certainly operate to reduce the sensitivity of the retina. Disturbances of accommodation operate in the same way, by tending to equalize the differences in brightness of the various portions of the stimulus as imaged on the retina. The latent period of muscular tissue is subject to important modifications from temperature-changes, toxins, etc., especially the toxins resulting from "fatigue." In work such as the present one, these variables can be minimized, as the muscular responses involved are relatively simple, easy and not sustained. Inasmuch as discriminative reactions involve more numerous synaptic connections than simple ones, it was assumed that the time required for executing them would be the more sensitive indicator of the influence of external variables. The fact that 5 Holt, E. B. : Psychol. Rev. Monogr. Supp., iv, 1903; pp. 8 Cobb, P. W. : Physiological points bearing on glare. Tr. Ilium. Eng. Soc., vi, 1911, pp. 153 ff. ; also, Physiological aspects of illuminating engineering. (Two lectures.) In: Lectures on illuminating engineering, vol. 2. Baltimore, The Johns Hopkins Press, 1911, pp. 559 ff. PSTCHOBIOLOQT, VOL. I, NO. 6 466 H. M. JOHNSON the correctness or incorrectness of the reaction furnished an objective check on the subject's mode of response, also influenced the choice of this type of determination. APPARATUS AND PROCEDURE In the present work the conditions of Cobb's studies were imitated as far as was practicable. A photometric field, in sensibly perfect balance, is viewed monocularly7 by the subject from E (fig. 1) through the limiting rectangular window W of a hollow polyhedron (7, the interior surfaces of which fill the remainder of the visual field. The brightnesses of the latter are practically unity with respect to each other, but may be varied at will with respect to the brightness of the photometric field, the two illuminations being almost completely independent. The photometric field is formed as follows : A milk-glasssurface MG3, illuminated by a tungsten incandescent lamp L3, is viewed directly by the subject through a 10 degree double prism P. Two similar surfaces, MGi and MG2, illuminated by lamps Li and Z/2, respectively, are viewed by reflection from the silvered mirrors Ml and Mz respectively, and by partial reflection from the front surfaces of P. The silvered mirrors are used on account of the limited area of P and the limited distance WP. If Li and Z/2 are properly placed, the result is a balanced photometric field projected in the plane of P, 150 cm. from the eye, and limited by the dimensions of W. These, in terms of visual angle subtended by them, are approximately 1.96 degrees vertically and 2.65 degrees horizontally; so that when properly fixated, the image of the test field covers the fovea and extends but little beyond it. The distribution of intensities among the wave- lengths of the visible spectrum is approximately that of a black body at 2400°K. The stimulus to reaction is the darkening of one-half of the photometric field to the extent of the brightness added by Li 7 The unused eye is covered by a sheet of ground glass, through which light from the interior of C is diffusely transmitted with little reduction by absorption. This tends to stabilize the relative adaptation of the two retinas, and reduces the pupillary fluctuation which might otherwise result. DISCRIMINATIVE RESPONSES TO VISUAL STIMULI 467 or L2 as the case may be. The subject holds in either hand a rubber bulb, which when quickly pressed, actuates a Dunlap pneumatic reaction key, to which it is attached by a rubber tube.8 The reaction consists in pressing the bulb corresponding to the side of the field which is darkened, as soon as the darkening is perceived. A warning signal, made by an electrical buzzer, is given by hand from 1 to 1.5 seconds before the stimulus is M, "% °L FIG. 1 r, integrating polyhedron; E, position of eye of observer; W, limiting window P, 10 degree double prism; MG$, milk glass surface forming standard portion of photometric field; LS) tungsten lamp illuminating MG3; Mi and M2, silvered mirrors; MG\ and MG2, milk-glass surfaces viewed from E by reflection from Mi and M* and by partial reflection from P; Lt and L2 auxiliary tungsten lamps, illuminating MGi and MG2 respectively, the extinction of which is the stimulus to reaction. Sundry protective screens used to arrest scattered light, are not shown. presented, the time depending on the desire of the subject, and being kept as nearly constant as possible. In ten to fifteen cases, irregularly distributed within each series of 100 reactions, the warning signal is not followed by the darkening of either side of the field, and the subject is instructed 8 The keys used in this work were modified so as to constitute double-throw switches. Thus, if the wrong key was actuated, it momentarily closed a signal circuit (shown in figure 2). The range of movement of the common connector in the reaction switch was made less than 1 mm., in order to insure closing of the signal circuit. 468 H. M. JOHNSON to inhibit reaction in such cases. If reaction is not inhibited, the fact is recorded; but false reactions are not included in the averaged results. A Johns Hopkins chronoscope,9 operated in an adjacent room, records the time elapsing between the opening of the circuit through LI or L2, and the subject's reaction. A part of the latent period of the lamp is thus added to the subject's reaction- time. Correction for this constant was not made in the results as presented. From data furnished by my colleague in physics, Dr. A. G. Worthing, it appears that with the lamps used, and under the conditions of the experiment, the added brightness would be reduced to 0.5 per cent of the brightness of the test- field and thus become completely effective, in 0.022 second after the current is interrupted; but it would be reduced to 90 per cent of its own original value, and begin to be effective, in 0.0006 second. I am uncertain what correction should be applied. It is clear, however, that the correction is constant and relatively small with respect to the variable to which it belongs. The method of presenting the stimuli is simple, and is readily ascertained from inspection of the wiring diagram in figure 2. The wiring of the pendulum magnet, and of the warning buzzer and the buzzer with which the subject, when necessary, signalled the experimenter, are not shown, but are independent of the rest of the system. The chosen supplementary lamp LI or Z/2 is cut out, and the chronoscope started by the pendulum breaking the lamp circuit and the circuit through the stationary magnet of the chronoscope at PSi. The lamp circuit is broken a second time at P& so that the lamp is not re-lighted when P& is closed an instant later to restore the circuit through CCi before the subject reacts. On its return swing, the pendulum closes PSZ and P$i in the order given, but does not re-open PS3, which is opened by hand after the chronoscope is read. Inspection of figure 2 will make it apparent that if the switch ESi is open and if the bank of double-throw switches ES2, ES*, 9 Dunlap, Knight: The Johns Hopkins Chronoscope. Journ. Exper. Psycho!, ii, 1917; pp. 249-252. DISCRIMINATIVE RESPONSES TO VISUAL STIMULI 469 ES* and ES5 are thrown to the left, the left side of the field, partially illuminated by Li, will be darkened when PS, is opened^ and that the subject will stop the chronoscope if he actuates the reaction key SSi, and flash the signal lamp RL4 and sound FIG. 2 B, 50-ampere storage battery (120 volts) ; M , coils of field of chronoscope motor; TF, 100 d.v. tuning fork; CCi, coil of stationary magnet of chronoscope; CCt, coil of rotating magnet of chronoscope; PSi, PS2, PSt, switches actuated by pendulum; ESi, ESt, ESS, ES<, ES6, switches operated by experimenter; SSi, SSi, reaction-keys, pneumatic type, modified to form two-way switches, operated by subject; L3, principal source of photometric field; LI, L2, auxiliary lamps, whose extinction is the stimulus to reaction; Bz, buzzer actuated by subject in event of incorrect reaction; RL\, RLt, RL3, RL4, lamps operated as resistances. The lamps illuminating the interior of C are not shown. They were wired in multiple with the rest of the system shown in figure 2, but were otherwise inde- pendently controlled. the signal buzzer B if he actuates the right hand reaction switch SSZ. If the bank of switches is thrown to the right, the functions of Li and L2 and of SSi and SS2 are reversed. If ESi is closed, the chronoscope is started by the breaking of PSi but the circuit is not opened through either LI or L2. If the subject, from over- 470 H. M. JOHNSON expectancy, should react in such case, the fact will be registered by the chronoscope, or by the lamp-buzzer combination, or both, according as SSi, SS2, or both are actuated. Each daily sitting consists of three series of 100 reactions each, one series being given at each sitting under each of the three experimental conditions under study. The order of arrange- ment of the three series is varied from day to day so as to distrib- ute the effects of fatigue and practice uniformly among them. Each series of 100 presentations of the stimulus contains an equal number of right-hand and left-hand presentations. The order is predetermined by shuffling a pack of 100 cards. The stimuli to which reaction is to be inhibited are interpolated irregularly. Each series of 100 presentations is subdivided into four groups of 25 presentations each. Each group requires from 4.5 to 6 minutes, and a rest of about 2 minutes is allowed between groups. For adaptation to the condition in which the interior of C is not illuminated, 15 minutes is allowed; for the other conditions, 5 minutes. In the latter cases, the subject keeps the head at the opening in C and looks at any part of the interior, or at the photometric field, as he may elect. Intent fixation during adap- tation is not encouraged or practiced. For adaptation to dark- ness the subject turns away his body so that the photometric field is not in view. The rest of the room is not illuminated. During a series of presentations the subject keeps his head in place. This is rendered fairly easy by the use of a mouthpiece containing a wax impression of the teeth. He is instructed to fixate the center of the photometric field as soon as the warning signal is given. The three conditions the reactions under which were compared are defined in table 1 . All the measurements given are in terms of candles per square meter. The brightness of the interior of C under condition Z>i is too low to be measurable with a Beckstein portable photometer, which is of the Lummer Brodhun type. The measurements given below are those made near the end DISCRIMINATIVE RESPONSES TO VISUAL STIMULI 471 of the work. The absolute brightnesses of the various surfaces can not be perfectly maintained without frequent photometry, as the lamps deteriorate gradually with use. The variation in this case was slight, as the lamps were not actually operated over 100 hours after being seasoned, and those illuminating the interior of C were operated at subnormal voltage. The relative brightnesses were easily kept very constant, as the lamps were connected in multiple on a steady circuit from large accumulator cells, and the variations in potential after the resistances had reached a stable temperature were not detectable with the ordi- nary voltmeter. It will be noted that the absolute brightness of the test field (measured with LI and Lz in operation) varied between 23 and TABLE 1 NAME OP CONDITION BRIGHTNESS OF TEST FIELD BRIGHTNESS OF SURROUNDINGS RATIO RATIO B, 24.37 18.19 1.34 0.746 Di 23.09 27.42 >0.00 61.60 <00 0.45 >0.00 2.25 27 candles per square meter with the different brightnesses of surroundings. This is due to light from the interior of C being reflected into the eye of the observer from MGi, MGZ and MG$. This could have been avoided by varying the distance between L3 and MGz to compensate, and it would have been done had the purpose of the experiment been a test of the conditions rather than a test of the method itself. The mounting of L3 which was used did not permit of convenient and accurate adjustment. The absolute brightness contributed by LI and L2 to their respective halves of the photometric field was 1J4 candles per square meter. This, with respect to the total brightness of the field was as follows : For condition BI, 4.68 per cent ; for condition DI, 4.94 per cent; for condition B2, 4.16 per cent. This consti- tutes a variability in the intensity of the stimulus to reaction, which in the most precise work might cct be desired, but which 472 H. M. JOHNSON is not as important as it might seem to the novice. The bright- ness-difference created by cutting out one of the supplementary lamps is several times as large as a threshold value, and it cer- tainly appeared to be no smaller under one condition than under another. It had been pretty well established that when the absolute intensities of the stimuli are well above the threshold, differences of this magnitude do not appreciably affect the time required for reaction; and none of the subjects used in the present work gave differences in reaction-time under the several con- ditions corresponding to the direction of the differences given above. THE SUBJECTS USED Certain information regarding the subjects is of interest in connection with the results which they yielded. Subject C was an unmarried woman, twenty years old, a junior in the college for women of Western Reserve University. She is exceptionally intelligent, and prior to this work had served as a laboratory subject for several months. Her reactions which are presented below were preceded by only 400, distributed over four days, but she made few errors in the regular work. She showed considerable improvement with practice, however, and this tends to increase the dispersion of the averaged reaction- times, and thus to reduce the apparent reliability of the differences among the several sets. The differences, however, are for the most part quite definite. Her academic schedule and vacation- plans prevented the allowance of more training and the accumu- lation of as many reactions as were desired. A fact of consider- able interest is that in some earlier work she showed a definitely higher tactile sensitivity in "total" darkness than in an illumi- nated room, most of the other subjects used in that experiment showing a contrary effect. In the present experiment she "pre- ferred to work in the dark," asserting that she found it "easier to pay attention" in the dark than in the light. She maintained this preference throughout the work, and expressed herself as being certain that her reactions under condition DI were shorter DISCRIMINATIVE RESPONSES TO VISUAL STIMULI 473 than those given under either of the other conditions. The results show that this estimate was unreliable. Subject E was twenty-three years old, unmarried, a senior in the college for women, Western Reserve University. She had had no previous training in photometric observation or in other laboratory work, and required 4500 reactions, equally distributed among the three conditions, to reduce the percentage of incorrect responses to an acceptable minimum. The reactions accumu- lated in this period of training are not included in the averages. Subject M was an unmarried woman, twenty-one years old, and a high school graduate. Her principal occupation was the manufacture of a trousseau. This with activities incidental thereto occupied most of her interest as well. She was trained with some little difficulty but after 2400 reactions, equally dis- tributed among the three conditions, she gave the results which are presented below. In the 5400 reactions which are presented, a considerable improvement with practice is noticeable. Subject A was a left-handed, unmarried man, twenty-two years old, with high school training. He was employed as an assistant in the laboratory. He exhibited definite psychotic traits which would be regarded by some psychiatrists as definite symptoms of an overcompensated inferiority-complex, having a sexual basis. He collected tickets at evening performances in a cinematographic house in order to supplement his salary, and also to study the actors, some of whom he imitated quite credit- ably as to mannerisms of pose and dress. Despite repeated explanations of the uses of reaction-time determinations, he persisted in regarding the experiment as a test of his "mind" and therefore exerted himself to the utmost in order to give the shortest possible reaction. The result was 289 incorrect or anticipatory reactions given with the 5400 correct ones which are presented. The averaged series were preceded by 1000 reac- tions obtained for purposes of training. All except the first 100 were equally distributed among the three external conditions. None of the subjects required refractive correction in the eye used in observation. 474 H. M. JOHNSON RESULTS: AVERAGES The distributions and averages are presented in tables 2 to 5 inclusive. They are given for the benefit of any reader who may wish to subject them to different treatment than the one employed. It will be noted that the reaction-times of subject G are dis- tributed in classes of iihr second, while the other subjects' reac- tions are in classes of 0.01 second. The records of subject C were made first of all, with the chronoscope motor operated at 720 revolutions per minute. Later, the speed was reduced for the other subjects, to 600 revolutions per minute to give a decimal unit of measurement. In all these tables the time values are given in classes of ten units each, the designation of the class being the numerical value of the median of the class. For example, the class of 0.31 includes all the reactions between 0.305 and 0.314 inclusive. The unit of 0.001 second is smaller than is useful. The chrono- scope is accurate to only 1 per cent, in the average, of the magni- tude of the times measured in this work, as was shown by the daily checks with the pendulum. Table 6 gives a summary of the averages and the probable errors of the averages for each subject under each of the three conditions; also, the differences between the compared averages, and the probable error of the differences. The latter is computed from the formula PE^-m = 0.6745 in which a represents the root mean square deviation from the mean, and N the number of terms included in the average M . From inspection it will be noted that this formula is equiv- alent to which is given by Davenport,10 after Pearson, as the formula 10 Davenport, C. B. : Statistical methods, with special reference to biological variation. 3d ed. New York, J. Wiley & Sons, 1904, p. 15. DISCRIMINATIVE RESPONSES TO VISUAL STIMULI 475 TABLE 2 Frequency-distribution of reactions according to time-classes. Subject C LEFT HAND. CONDITION: RIGHT HAND . CONDITION : CLASS (UNIT = Ti5 SECOND) B, Di Bt Bl Dl B2 57 1* 1* 56 55 t 54 1* 53 52 3* 51 1* 1* 50 1* 49 1 1* 48 2* 2* 47 1* 2 2 1* 2 46 5 4 3 3 45 1 3 4 3 1 44 2 6 2 1* 2 6 43 4 10 8 1 1 8 42 4 19 17 2 10 6 41 6 14 15 9 9 15 40 20 27 25 8 9 21 39 18 36 33 7 20 27 38 21 33 32 15 30 25 37 41 38 51 31 31 35 36 50 72 59 37 43 45 35 53 50 64 55 48 49 34 74 39 50 53 41 41 33 49 50 44 46 43 43 32 41 29 40 60 43 37 31 38 36 29 52 52 36 30 54 33 30 60 56 53 29 37 19 15 39 37 42 28 22 18 11 36 31 22 27 13 4 9 21 19 16 26 3 2 4 7 8 6 25 1 1 3 3 3 24 1 3 1 N 553 546 549 542 545 543 M 33.64 35.32 35.18 32.62 33.28 33.98 M (units of 0.01 sec- onds) 28.03 29.43 29.32 27.18 27.73 28.32 0.087 0.100 0.097 0.084 0.099 0.106 * Excluded from average by Chauvenet's criterion. 476 H. M. JOHNSON TABLES Frequency-distribution of reactions according to time-classes. Subject E CLASS LEFT E [AND. COND ITION: BIGHT BAND. COND CTION: (uinr •» 0.01 SECOND) Bi Di Bt Bi Dl *«, 49 1* 2 2 1* 2* 48 1* 2 1 2* 2* 47 1* 8 2 1* 3* 2* 46 2* 4 1 1* 1 4* 45 2* 5 12 1* 3 4 44 9 6 2 3 43 1 18 12 2 42 2 17 11 1* 1 4 41 4 31 26 1* 8 1 40 7 28 24 1* 10 3 39 11 36 34 3 18 7 38 9 44 47 2 20 18 37 12 46 44 5 14 17 36 28 P 70 56 4 14 18 35 55 68 70 20 37 29 34 54 53 60 13 30 26 33 76 79 81 23 43 47 32 75 50 59 37 71 57 31 101 38 40 45 63 64 30 60 31 26 45 64 54 29 52 24 25 65 55 55 28 63 13 18 68 48 82 27 26 11 11 72 41 59 26 24 4 14 68 44 48 25 15 5 6 81 43 36 24 10 2 4 40 23 22 23 6 2 3 40 22 20 22 1 34 10 14 21 1 23 6 6 20 1 6 1 19 A" 693 700 696 694 694 694 M 31.44 35.18 34.66 27.53 30.76 29.98 PEU 0.091 0.115 0.116 0.096 0.117 0.115 * Excluded from average by Chauvenet's criterion. TABLE 4 Frequency-distribution of reaction according to time-classes. Subject M CLAB8 (UNIT — 0.01 SECOND) LEFT HAND. CONDITION: RIGHT HAND. CONDITION: Bi Di * Bi ft * 59 58 2* 57 56 2* 2* 55 2* 1* 54 53 1* 52 2 1 51 50 6 2 49 2 1* 1* 1* 48 5 1* 1* 4 1* 47 2* 3 2* 4 46 7 2 5 3* 45 1* 12 3 6 6 44 13 4 16 2 43 2* 11 3 19 3 42 22 5 1* 13 3 41 2* 17 6 2* 19 5 40 3 21 9 1* 27 8 39 3 37 18 1 40 5 38 5 44 26 5 47 15 37 14 60 25 7 61 23 36 17 70 45 9 61 21 35 19 62 52 16 65 37 34 26 64 43 21 55 54 33 34 78 69 28 62 49 32 47 73 73 46 70 79 31 57 67 91 55 70 82 30 78 72 83 82 58 89 29 88 42 92 92 49 84 28 97 32 80 100 35 77 27 114 22 65 101 44 76 26 99 18 29 92 25 58 25 75 14 37 86 19 48 24 55 7 16 56 10 31 23 28 4 12 36 4 19 22 21 4 2 38 2 11 21 11 2 6 18 2 10 20 1 4 2 19 1 1 N 893 28.37 0.076 893 34.17 0.118 896 31.21 0.097 894 27.82 0.079 897 33.76 0.117 895 30.15 0.096 U. .. PE* * Excluded from average by Chauvenet's criterion. 477 TABLE 5 Frequency-distribution of reactions according to time-classes. Subject A CLASS LEFT HAND. CONDITION : RIGHT HAND. CONDITION: (UNIT = 0.01 SECOND) Bl Di B2 Bt D! Bi 57 1* 56 55 54 53 3* 1* 52 1* 51 1* 50 1* 49 1* 1* 1* 1* 48 1* 2 2 47 2* 1* 2 2 46 5 4 45 6 2* 4 3 44 2 1 10 4 43 5 1 1* 3 6 42 1* 6 3 1* 9 7 41 1* 7 6 1 12 7 40 12 9 1 15 16 39 1 15 4 3 17 19 38 2 23 12 7 29 23 37 5 37 19 11 36 37 36 7 17 37 5 46 39 35 7 50 36 14 44 47 34 14 51 37 33 54 39 33 28 83 56 49 52 57 32 40 70 62 73 66 60 31 62 73 66 86 86 60 30 88 99 102 117 102 98 29 121 106 102 100 84 78 28 100 77 100 97 66 77 27 118 63 76 M 57 62 26 93 42 70 56 37 50 25 82 31 40 56 25 33 24 52 10 24 51 14 33 23 38 6 17 15 8 13 22 32 7 11 14 4 6 21 6 2 3 6 3 6 20 3 2 6 1 1 19 1 1 18 1 17 1 N 901 900 896 896 893 890 M 27.80 31.22 29.96 28.96 31.83 31.15 PEM 0.072 0.095 0.094 0.077 0.107 0.111 *Excluded from average by Chauvenet's criterion. 478 DISCRIMINATIVE RESPONSES TO VISUAL STIMULI 479 £= 3s 41? SIQ CO rH Y— 1 Oi rH I— I l-H O T— ( dodo (N QO »O iO CO 05 rH rH d I-H CO CO 06 os I-H (N §t>- i^ r* I— ( i-H O 1-H T— ( T— 1 d d o d CO CO O CO t>. i> i> oo i^ d co I-H - d d d d t^ 00 T? r-i ci 1C rH O l^ C35 Tf< Tji TjJ CO T^ ^ CO »O O i-H rH rH rH dodo Oi 00 i-H QO iO t^> CO O d d co d 1 1 1 (N oo O (N S 00 co O5 oo rH CO (N *0 co O CO OO °* oo c rH 1 1 1 1 oo - 00 CO rH CO «3 CO O W 480 H. M. JOHNSON of the "probable difference" between Mi and Af2. This desig- nation may be due to a typographical elision, although its use might be justified. The ratio of the magnitude of the difference to the probable error of the difference is given as a means of ascertaining from the probability integral table of the function x + PEX what is the probability for and against the difference being due to chance. Certain extreme reactions, indicated by asterisks in tables 2 to 5 inclusive, were not included in the averages. Their exclusion was accomplished by the use of Chauvenet's criterion.11 Its application to an asymmetrical distribution is not wholly justifiable, but the calculation of a theoretically perfect one is rather tedious and the gain in accuracy is not sufficient to justify the labor. The arbitrary use of a standard criterion of exclusion is evidently better than rendering judgment from simple inspec- tion. In any set of data of this type a number of stray deter- minations will be found which clearly do not belong to the same distributions as the others and probably were not obtained under comparable conditions. To admit them to a distribution containing a limited number of cases may tend to distort com- parisons with other distributions in which, owing to the limited number of observations, similar strays had not yet appeared. And yet, it is evidently unfair to apply an arbitrary criterion of exclusion which may not apply uniformly to all the compared distributions. The distributions, reduced to a percentage basis, are also shown graphically in figures 3 to 6 inclusive. From inspection of the graphs it is evident that all the distributions are multi- modal. This is characteristic of all the reaction-time distribu- tions which the author has seen. His attention was called to the fact some five years ago by Professor Dunlap, who asserted at the time that this form of distribution is typical, and does not disappear with greater accumulation of data. This was corroborated in an interesting manner in tallying the daily records of the present work, according to classes of tune-values. 11 Cf. Davenport, C. B. : op. cit., pp. 12 f. DISCRIMINATIVE RESPONSES TO VISUAL STIMULI 481 All the prominent modes were prominent throughout the process of tallying, although their relative values might shift as a result of practice. This fact indicates that a given curve of this type is really compounded from several separate distributions, more or less widely separated; each of which, if isolated, might be a normal IS 25 30 35 19 25 30 JS 40 41 35 40 4S FIG. 3. PERCENTAGE DISTRIBUTION OF REACTION-TIMES OF SUBJECT C Ordinates, percentages of reactions. Abscissae, time in units of 0.01 second. Graph 1. Light line, 553 left hand reactions under condition Bi. Graph 1. Heavy line, 546 left hand reactions under condition Di. Graph 2. Light line, 553 left hand reactions under condition B\. Graph 2. Heavy line, 549 left hand reactions under condition B*. Graph 3. Light line, 542 right hand reactions under condition B\. Graph 3. Heavy line, 545 right hand reactions under condition D\. Graph 4. Light line, 542 right hand reactions under condition B\. Graph 4. Heavy line, 543 right hand reactions under condition J52. Vertical lines indicate averages. PSTCHOBIOLOGT, VOL. I, NO. 6 482 H. M. JOHNSON curve. The analysis of such a curve is said by Davenport to be very laborious, and I have not yet found time to attempt it. Meantime, the constants of dispersion and of reliability which are presented are not to be regarded as the closest possible approximations to the true values. 13 25 30 35 40 45 50 tt 25 30 35 40 45 50 w » 30 35 40 45 ' SO 25 30 35 40 45 50 FIG. 4. PERCENTAGE DISTRIBUTION OP REACTION-TIMES OF SUBJECT Ordinates, percentages of reactions. Abscissae, time in units of 0.01 second. Graph 5. Light line, 693 left hand reactions under condition Bi. Graph 5. Heavy line, 700 left hand reactions under condition D\. Graph 5. Light line, 693 left hand reactions under condition BI. Graph 5. Heavy line, 696 left hand reactions under condition B2. Graph 7. Light line, 694 right hand reactions under condition BI. Graph 7. Heavy line, 694 right hand reactions under condition Di. Graph 8. Light line, 694 right hand reactions under condition B\. Graph 8. Heavy line, 694 right hand reactions under condition BZ. Vertical lines indicate averages. DISCRIMINATIVE RESPONSES TO VISUAL STIMULI 483 This type of distribution may be explained on the hypothesis that between the retina and the muscles of the hand are several distinct systems of neural arcs of varying "length," and that under favorable conditions of receptiveness of the organism, the "shorter" ones may be brought into operation; while under less favorable conditions the impulse must pass over certain of 25 4547 (620 25 SO S3 25 FIG. 5. PERCENTAGE DISTRIBUTION OF REACTION-TIMES OP SUBJECT M Ordinates, percentages of reactions. Abscissae, time in units of 0.01 second. Graph 9. Light line, 893 left hand reactions under condition B\. Graph 9. Heavy line, 893 left hand reactions under condition D\. Graph 10. Light line, 893 left hand reactions under condition B\. Graph 10. Heavy line, 896 left hand reactions under condition -B2. Graph 11. Light line, 894 right hand reactions under condition Bi. Graph 11. Heavy line, 897 right hand reactions under condition D\.. Graph 12. Light line, 894 right hand reactions under condition Bi. Graph 12. Heavy line 895 right hand reactions under condition Bz. Vertical lines indicate averages. 484 H. M. JOHNSON the "longer" pathways. The expressions "short" and "long" do not refer to geometric distances, but to the relative degree to which the paths retard the propagation of the impulse over them. (Cf. the roughly analogous expression "optical distance.")12 17 tS 30 35 40 45 25 30 3S 40 'is' 45 50 40 FIG. 6. PERCENTAGE DISTRIBUTION OF REACTION-TIMES OF SUBJECT A Ordinates, percentages of reactions. Abscissae, time in units of 0.01 second. Graph 13. Light line, 901 left hand reactions under condition B\. Graph 13. Heavy line, 900 left hand reactions under condition D\. Graph 14. Light line, 901 left hand reactions under condition B\. Graph 14. Heavy line, 896 left hand reactions under condition B*. Graph 15. Light line, 896 right hand reactions under condition B\. Graph 15. Heavy line, 893 right hand reactions under condition D\. Graph 16. Light line, 896 right hand reactions under condition B\. Graph 16. Heavy line, 890 right hand reactions under condition #2. Vertical lines indicate averages. 12 This paragraph was inspired by oral discussion of the question with Professor Dunlap, but it does not purport to be an accurate presentation of his views. DISCRIMINATIVE RESPONSES TO VISUAL STIMULI 485 A similar concept, formulated in accordance with the "all or nothing" principle might be substituted for the above. Under favorable conditions the impulse conveyed to the effector over the first system of neural arcs excited might be adequate to arouse reaction. Under less favorable conditions the " intensity" of the first impulse alone might be inadequate, but might have to await reinforcement by succeeding impulses conveyed over other pathways which were excited later. Or, a part of the delay may be assumed to have taken place in the central synapses, rather than in the muscle, as was suggested above (p. 464). A study of the distribution-curves of the reaction-tunes, not only of one subject, but of several subjects compared with each other, suggests strongly that there are several definite time- values, more or less widely separated, about which the reactions tend to group rather closely. Under the more favorable con- ditions, the preponderance is among the smaller values. The foregoing discussion should make it clear that the practice of smoothing curves of distribution of reaction-times by averag- ing the ordinates in overlapping groups of threes is illegitimate, and may also serve to obscure some very instructive features. DISPERSION The averages of reaction-times obtained under different con- ditions are much more significant when the dispersion of the reaction-times is considered. This measured by the "standard deviation," or quadratic mean of the deviations from the mean. In table 7 is given the standard deviation * 00 oo CO rj< t>- oo co 5^ f^ f*^ f^ ^2 dodo ^ iO ?5 l> CO Tti 10 TjJ oo C< CO CO CO oo CO S CO O5 CO ^ TjJ CO t> C^J >O 1-1 O d d d d rf i— 1 i-H O5 O 1O 10 r-l CO CO CO CO si P ^ OS t^ O CO o CO oo 0 0 TH TH o os o> (N CO OS GO TH rH 00 0 dodo rj< CO O O CO 0 CO 00 O il rH OS CO rH OS O rH CO d d 06 • •* r^ O CO O O ^H rH rH DISCRIMINATIVE RESPONSES TO VISUAL STIMULI 487 The results show in general that the condition which yielded the shortest average reaction-time for all the subjects, gave also the most narrow distribution of the individual reaction-times about the mean. This fact is inferable from a comparison of the probable errors of the compared averages. DISTRIBUTION OF ERRORS In the course of accumulating the averaged reaction-times, a number of wrong or anticipatory reactions were also obtained on each subject. The distribution of these errors according to the lighting condition under which they were given, is of consider- able interest. The data are summarized in table 8. TABLE 8 SUBJECT CONDITION Bl, NUMBER ERRORS CONDITION Dl. NUMBER ERRORS CONDITION Bj, NUMBER ERRORS Absolute Relative Absolute Relative Absolute Relative c 22 9 47 57 1.00 1.00 1.00 1.00 23 24 70 129 1.04 2.66 1.44 2.26 23 23 54 103 1.04 2.55 1.15 1.89 E M A The results follow the same general tendencies as the reaction- times. Those of subjects M and A suggest that discrimination was more difficult under conditions DI and B2 than under B1 and that the subjects sometimes attempted to compensate by an excessive strain of "attention;" i.e., that they were sometimes set for a "motor" type of reaction, in the expectation that a given side would be darkened; and that this expectancy was so strong that when the stimulus finally appeared they reacted without regard to its direction. The same tendency is very clear in the results of subject E, but errors appeared much less frequently than in the results of subjects M and A. I do not regard the differences in the distribution of the errors of subject C as significant. 488 H. M. JOHNSON SUMMARY In all 18,300 discriminative reactions are presented as obtained on four subjects under three lighting conditions. For all the subjects, the time required for making a discrimina- tive reaction is longer for the condition under which the surround- ings are dark than for the condition under which the surroundings are three-fourths as bright as the stimulus. The differences between the averages vary between 2 per cent and 20 per cent for the several reactors; and in seven sets out of eight the prob- ability is many billion to one against the differences being due to chance. The one exception is one series given by subject C, whose preference is for. the dark. In this case there was no reversal of the effect, but merely a reduction of the difference to a value which is four times its probable error. All the subjects required a longer time for discriminative reaction when the surroundings were 2.25 times as bright as the stimulus than when they were 0.75 times as bright as the stimulus. The magnitude of the difference varies between 4 per cent and 10 per cent for the different subjects and in every case its value with respect to its probable error indicates a probability of many billion to one against the difference being due to chance. For three of the four subjects the retardation due to excessively bright surroundings is less than that due to dark surroundings. The difference for subject E is between 3 and 5 times its probable error, and therefore reasonably probable; for subject M it is from 19 to 24 times its probable error, and therefore almost absolutely certain; while for subject A it varies between 5 and 9 times its probable error. One set of subject C shows a reversal of this effect and the other set is ambiguous. For all the subjects, the dispersion of the results is greater when the surroundings are dark than when they are 75 per cent as bright as the stimulus. The magnitude of the difference varies between 15 per cent and 54 per cent for the different sub- jects. For subject C the difference is approximately 5 to 6 times its probable error; and for the other subjects, from 8 to 17 times its probable error. DISCRIMINATIVE RESPONSES TO VISUAL STIMULI 489 For all the subjects the dispersion of the results is greater when the surroundings are 2.25 times as bright as the stimulus, than when they are 0.75 times as bright as the stimulus. The magnitude of the difference varies between 11 per cent and 44 per cent for the different subjects, and in seven sets out of eight it is 8 to 15 times its probable error. For one subject, M, the dispersion of the results is about twice as great when the surroundings are dark as when they are excessively bright; and the differences are large with respect to their probable errors. For the other three subjects the differ- ences are small and their significance is doubtful. Three of the four subjects made a much smaller number of incorrect or anticipatory reactions when the surroundings were 75 per cent of the brightness of the stimulus, than when they were dark or excessively bright. Subject C is an exception, exhibiting no significant difference. She made few errors under any condition after the first few days of practice. DISCUSSION The results show that while individual differences exist among the subjects used, visual performance was more accurate and certain when the surroundings were of the same order of bright- ness as the stimulus than when they were dark or when they were considerably though not excessively brighter. If the latent period of the muscular tissue involved, and the latent period of the retina were added and the sum taken from the time re- quired for the type of reaction given by the subjects, the small absolute differences obtained would appear relatively much larger than they do when added to a larger quantity. If moder- ate differences in the distribution of brightnesses have in general the effect of speeding and retarding visual performance to the extent indicated by the results obtained under these special conditions, the matter of appropriate distribution of brightnesses would seem to be of greater importance than has been attributed to it by illuminating engineers. So far, they have concerned 490 H. M. JOHNSON themselves chiefly with eliminating extreme conditions of "glare."" While the preliminary experiment was not planned to ascertain what is the optimal distribution of brightnesses, certain definite results were incidentally obtained. The author believes these results to be susceptible of misinterpretation unless certain facts are especially considered. It will be recalled that the area of the stimulus was small, its image barely covering the fovea. Under this condition the brightness of the surroundings deter- mined the state of adaptation of the retina. In such case, therefore, one may suspect that the brightness of the surround- ings which best adapts the retina to the brightness of the stimulus will always be the optimal condition. This has been borne out by the work of Cobb and of Cobb and Geissler on the thresholds for pattern and for difference of brightness, and has been cor- roborated by the present work to the extent of the limited range of distributions of brightnesses employed. Under ordinary lighting conditions, however, as in the home or office, the page which one may be reading, or the sewing on one's lap, may cover a large part of the visual field. The paper on which this report was written subtended a visual angle of 40 degrees in one dimension and 60 degrees in the other, at the distance from the eye at which it lay during the work. The writer has tried several "uniform" distributions of illumination in the home and office, and has discarded all of them in favor 13 For the benefit of any student of illuminating engineering who may still be puzzled by the significance of differences of a few thousandths of a second in the time required for recognition, I have the kind permission of Capt. P. W. Cobb to refer to a suggestion of his, contained in a paper which, I trust, will soon be published. The suggestion is that the normal resting period of the eye is but momentary; and that any condition which increases the time required for perception, even by a small amount, may tend to retard the normal rate of eye- movement and hence to induce excessive fatigue of the extrinsic muscles. The case is somewhat analogous to the movements in walking. If one is compelled to walk more slowly than is one's wont, or to adopt an irregular gait, the added exertion is at once noticeable. It is often noted that the reading of unfamiliar subject-matter is much more fatiguing than the reading of familiar matter — for the reason that the eye must rest for a longer period on the words or group of words comprehended in a single observation. DISCRIMINATIVE RESPONSES TO VISUAL STIMULI 491 of a floor-lamp or desk-lamp, placed so as to give a comfortable brightness over a large central area, leaving the extreme periph- eral field fairly dark. In the present experiment, if the area of the test-object had been sufficiently increased, a value would have been found at which the stimulus-surface, instead of the surroundings, deter- mined the state of adaptation. In such a situation, the influence of the surroundings might have been nullified or even reversed. The results as reported are therefore to be regarded as holding under the special conditions of the experiment, and not neces- sarily holding otherwise. The plan of experimentation required next an application of the method to a wide range of relative brightnesses of field and surroundings, to find the optimal relation between them with the original area of the stimulus; and next, a variation of the area of the stimulus to ascertain whether the effects already found are general or not. The work, however, was interrupted by the national emergency of 1917; and not long afterwards, the experimenter's relations with the laboratory were terminated upon his entering the military service. The problem is therefore abandoned, with the following comment on method : Threshold-determinations as a measure of the influence of external conditions are meaningless unless a time-limit is placed on the period of observation, or unless the time required for observation is considered. If the observer maintains — as in an ideal case — a constant criterion, the difference between the threshold readings made under different external conditions need give no indication of the relative facility of observation. Under a relatively difficult condition the better observer tends to com- pensate by making a greater effort, and extending the period of observation; the less reliable observer, if he fails, may fail by shifting his criterion or by declining to make the necessary increase of effort. A special form of threshold-determination consists in requiring the observer to keep an instrument continuously in threshold adjustment for a considerable period of time. This method is unsatisfactory, as the subject's attention is variable, and the 492 H. M. JOHNSON settings made during the lapses of attention are not necessarily related to the subj ect's instructions or intention. In other words, the setting of the instrument at a given instant does not neces- sarily measure the subject's threshold. The method was thor- oughly tried recently in the work on low oxygen at the Air Service Medical Research Laboratory, and had to be discarded. Work-tests are unimpeachable in principle, and the difficulties in the way of applying them are not insuperable. But the preparation of an adequate amount of uniformly difficult material is a task which many students would find appalling; and the use of such material is an absolutely indispensable condition of getting interpretable results. If the type of work chosen for the test involves a high degree of muscular coordination, a very long period of preliminary training is necessary. And, since the time required to execute a coordinated muscular response may be highly variable in comparison with the short time required for simpler registration of the act of discrimination, such a work- test will require the accumulation of a larger quantity of data than most students would consider to be worth the expense. The reaction-time method requires a simple muscular response, which is usually stabilized within a few weeks of training so that further improvement is slow and gradual. The determination of discrimination-time is probably sufficiently sensitive to show the effects of any condition which tends to increase the difficulty of sensory response. The price of definite results, especially if the effect is small, is threefold : (1) adequate control of the external conditions so that secondary stimulus-variables are not effective; (2) adequate training of the subjects, to stabilize effects of fatigue and practice, and insure an approximation to uniformity in methods of observing and reacting; and (3) the accumulation of a sufficient number of reactions to make the results reliable. The number of reactions necessary to demonstrate a differ- ential effect depends, obviously on the magnitude of the effect and the dispersion of the results. For example, the last 600 left-hand reactions of subject A, equally distributed among the three lighting conditions described above, show a difference in favor of condition Bi over DI which is 14 times the probable DISCRIMINATIVE RESPONSES TO VISUAL STIMULI 493 error of the difference; and a difference in favor of condition BI over B2 which is 13 times its probable error. This number of reactions is evidently enough to demonstrate the direction and the relative certainty of these differences. The same data show a difference in favor of condition B% over condition D1 of 3 times the probable error of the difference. This means that in a large number of repetitions of the experiment a reversal of the effect would be expected once in each 22 times. Whether this difference is sufficiently reliable is of course a matter of opinion. I should consider it desirable to quadruple the number of data, since if the effect is persistent this procedure would halve the probable error and would increase nine hundred fold the probability against the difference being due to chance. Obviously the most economical procedure is to tally the results daily as they are accumulated; summate and average them from time to time, and obtain the constants of dispersion and reli- ability; and stop work under any compared conditions when the results show a sufficiently definite effect, as judged by the degree of precision which may be demanded. The number of reactions necessary for definite results from a trained subject is of course far smaller than from an untrained subject, as the effect of training is to enable the subject to reduce the dispersion of his results very greatly. The most economical method therefore requires the subjects to be thoroughly trained before the external variables whose effects are to be compared, are introduced. CONCLUSION I wish to acknowledge the technical assistance of Mr. George Hathaway in the experimental work; and the aid and counsel of a number of colleagues, especially of the following: Dr. W. Weniger, in planning the electrical system; Captain P. W. Cobb, in the photometry of the various surfaces and in criticism of the paper; Dr. C. F. Lorenz and Dr. A. G. Worthing, in the selection of flashing lamps and in valuable criticism of technical methods; and Captain J. E. Coover, who acquainted me with the valuable abbreviations of statistical procedure which he has 494 H. M. JOHNSON used so fruitfully in his work in psychical research at Stanford. Major William MacLake, of the department of psychiatry in this laboratory, favored me by a careful and critical perusal of the paper, with special reference to the neurological questions in- volved. My general indebtedness to the work of Professor Dunlap will be readily recognized by psychological readers, al- though I have seen fit to deviate rather widely from his pub- lished opinions on the treatment of data,14 which I understand he is now willing to modify somewhat. 14 Dunlap, Knight: Some experiments with reactions to visual and auditory stimuli. Psychol. Rev., xvii, 1910; pp. 319-335. Of. p. 323. AUTHOR AND SUBJECT INDEX ANALGESICS, Action of some anti- pyretic, on psychological reaction time. D. I. Macht, S. Isaacs and J. Greenburg, 327-338 Announcement, 1 Associations, Method of studying con- trolled word. M. W.Loring, 369-428 Association test, The stop-watch and the. Knight Dunlap, 339-352 BOCK, CARL W. A classification of groups, 277-318 CEREBRAL cortex of the cat, On the motor functions of the. J. D. Stout, 177-230 Cerebral destruction, The effects of, upon habit formation and retention in the albino rat. K. S. Lashley and S. I. Franz, 71-140 Cerebral motor control : The recovery from experimentally produced hemi- plegia. R. Oden and S. I. Franz, 33-50 Chronoscopes, Methods of using the balanced-magnet. Knight Dunlap, 445-458 Continuous stimulations versus transi- tional shock, in the phototactic response. S. J. Holmes, 65-70 DESTRUCTION of the frontal por- tion of the cerebrum, The retention of habits by the rat after. S. I. Franz and K. S. Lashley, 3-18 Discriminative responses to visual stimuli. H. M. Johnson, 459-494 Distribution of practice to the rate of learning, A simple maze: with data on the relation of, 353-368 DODSON, J. D. Relative values of reward and punishment in habit formation, 231-276 DUNLAP, KNIGHT. Internal secretion in learning. (Discussion), 61-64 DUNLAP, KNIGHT. The stop-watch and the association test, 171-176 DUNLAP, KNIGHT. A synchronous motor kymograph, 319-324 DUNLAP, KNIGHT. Methods of using balanced-magnet chronoscopes, 445- 458 Dunlap's method for the mean varia- tion. Buford Johnson, 325-326 JTRANZ, S. I. AND LASHLEY, K. S. The retention of habits by the rat after destruction of the frontal portion of the cerebrum, 3-18 FRANZ, S. I. AND ODEN, R. On cere- bral motor control; the recovery from experimentally produced hemiplegia, 33-50 FRANZ, S. I. AND LASHLEY, K. S. The effects of cerebral destruction upon habit formation and retention in the albino rat, 71-140 Feeding, Effects of delayed upon learning. J. B. Watson, 51-60 Q.ROUPS, A classification of. Carl W. Bock, 277-318 GREENBURG, J., MACHT, D. I. and ISAACS, S. Action of some anti- pyretic analgesics on psychological reaction time, 327-338 J-jABIT formation and retention in the albino rat, Effects of cerebral destruction upon. K. S. Lashley and S. I. Franz, 71-140 495 496 AUTHOR AND SUBJECT INDEX Habit formation, Relative values of reward and punishment in. J. D. Dodson, 231-276 Hemiplegia, the recovery from experi- mentally produced: On cerebral motor control. R. Oden and S. I. Franz, 33-50 HOLMES, S. J. Continuous stimula- tions versus transitional shock in the phototactic response, 65-70 HUNTER, WALTER S. Some notes on the auditory sensitivity of the white rat, 339-352 JSAACS, SCHACHNE AND MACHT, D. I. The action of some opium alkaloids on the psychological reaction time, 19-32 ISAACS, SCHACHNE, MACHT, D. I. AND GREENBURG, J. Action of some antipyretic analgesics on psycho- ogical reaction time, 327-338 Internal secretion in learning. (Dis- cussion.) Knight Dunlap, 61-64 JOHNSON, BUPORD. Dunlap's method for the mean variation, 325-326 JOHNSON, H. M. Discriminative re- sponses to visual stimuli, 459-494 CYMOGRAPH, A synchronous motor. Knight Dunlap, 319-324 LASHLEY, K. S. AND FRANZ, S. I. The retention of habits by the rat after destruction of the frontal por- tion of the cerebrum, 3-18 LASHLEY, K. S. AND FRANZ, S. I. The effects of cerebral destruction upon habit-formation and retention in the albino rat, 71-140 LASHLEY, K. S. The effects of strych- nine and caffeine upon the rate of learning, 141-170 Learning, Internal secretion in. Knight Dunlap, 61-64 Learning, The effects of strychnine and caffeine upon the rate of. K. S. Lashley, 141-170 Learning, The effects of delayed feed- ing upon. J. B. Watson, 51-60 LORING, MILDRED WEST. Methods of studying controlled word associa- tions, 369-428 LORING, MILDRED WEST. Word-lists for adjectives and noun reactions, 429-444 ]\|ACHT, D. I., ISAACS, S. AND GREENBURG, J. Action of some antipyretic analgesics on psycho- logical reaction time, 327-338 MACHT, D. I. AND ISAACS, S. Action of some opium alkaloids on the psychological reaction time, 19-32 Mean variation, Dunlap's method for the. Buford Johnson, 325-326 Methods of using balanced-magnet chronoscopes. Knight Dunlap, 445- 458 Methods of studying controlled word associations. M. W. Loring, 369- 428 Motor functions of the cerebral cortex of the cat, On the. J. D. Stout, 177-230 ()DEN, ROBERT AND FRANZ, S. I. On cerebral motor control: the recov- ery from experimentally produced hemiplegia, 33-50 Opium alkaloids on the psychological reaction time, Action of some. D. I. Macht and S. Isaacs, 19-32 pHOTOTACTIC response, Continu- ous stimulations versus transi- tional shock in the . S. J. Holmes, 65- 70. Psychological reaction time, Action of some antipyretic analgesics on. D. I. Macht, S. Isaacs and J. Greenburg, 327-338 AUTHOK AND SUBJECT INDEX 497 Psychological reaction time, Action of some opium alkaloids on. D. I. Macht and S. Isaacs, 19-32 J^EACTION time, Action of some antipyretic analgesics on psycho- logical. D. I. Macht, S. Isaacs and J. Greenburg, 327-338 Reaction time, Action of some opium alkaloids on psychological. D. I. Macht and S. Isaacs, 19-32 Reactions, Word-lists for adjective and noun. M. W. Loring, 429-444 Retention of habits by the rat after destruction of the frontal portion of the cerebrum, The. S. I. Franz and K. S. Lashley, 3-18 Retention, The effects of cerebral de- struction upon habit-formation and. K. S. Lashley and S. I. Franz, 71-140 Reward and punishment in habit for- mation, Relative value of. J. D. Dodson, 231-276 gENSITIVITY, of the white rat, Some notes on the auditory. W. S. Hunter, 339-352 Stop-watch and the association test, The. Knight Dunlap, 171-176 STOUT, JOSEPH DUERSON. On the motor functions of the cerebral cortex of the cat, 177-230 Strychnine and caffeine, The effects of upon the rate of learning. K. S. Lashley, 141-170 \7lSUAL stimuli, Discriminative re- sponses to. H. M. Johnson, 459- 494 \YATSON, JOHN B. The effect of delayed feeding upon learning, 51-60 Word-lists for adjective and noun reactions. Mildred W. Loring, 429- 444 BF 1 P6 v.l Psychobiology PLEASE DO NOT REMOVE SLIPS FROM THIS POCKET UNIVERSITY OF TORONTO LIBRARY