QL 78 Ha 3 300 7flQ BEHAVIOR MONOGRAPHS Volume 2, Number 6, 1915 Serial Number 1 1 Edited by JOHN B. WATSON The Johns Hopkins University? The Effect of Age on Habit Formation in the Albino Rat DISSERfAliON Submitted to the Board of University Studies f the Johns Hopkins University , in conformity with the re ;uirements for the Degree of Doctor of Philosophy BY HELEN B. HUBBEHT 1915 Published at Cambridge, Boston, Mass. HENRY HOLT & COMPANY 34 West 3M Street, New York G. E. STECHERT & CO., London, Pari« and Leipzij, Foreiga Agenta EXCHANGE BIOLOGi LIBRARY G THE EFFECT OF AGE ON HABIT FORMATION IN THE ALBINO RAT DISSERTATION Submitted to the Board of University Studies of the Johns Hopkins University, in conformity with the requirements for the Degree of Doctor of Philosophy BY HELEN B. HUBBERT BALTIMORE, MARYLAND JUNE. 1915 H? BIOLOGY LIBRARY G CONTENTS Page •Acknowledgments v Introduction 1 Historical 1 Apparatus and Procedure 4 Experimental Results: Twenty-five day rats 15 Sixty-five day rats 23 Two hundred day rats 29 Three hundred day lats 33 Five hundred day rats 38 Comparison of Results for Different Ages 42 Incidental Tests: Effect of Sex on Rapidity of Learning 46 Day and Night Work ; . 47 Continuation of Work after Problem has been Learned 48 Blood Relationship and Learning 51 Retention 52 Resume of Conclusions . . 54 iii ACKNOWLEDGMENTS The writer's heaviest obligation is to Professor John B. Wat- son, Director of the Psychological Laboratory of The Johns Hopkins University, without whose unsparing suggestions and criticisms, the work could not have been brought to a successful conclusion. Dr. Knight Dunlap, Associate Professor of Psychology at The Johns Hopkins University, offered much needed assistance in the handling of the data. Dr. Gardner C. Basset, of Pittsburgh University, gave valu- * able suggestions regarding technique and method. Dr. Henry H. Donaldson of The Wistar Institute of Anatomy, furnished practically all of the rats used for breeding purposes as well as a few of those employed in the experiment, and was most helpful in suggestions throughout the experiment. Dr. Shinkishi Hatai of The Wistar Institute of Anatomy, pre- pared anatomical data for each rat used in the experiment, which cannot be adequately handled in this discussion, and which will probably be taken up in a separate publication to appear later. THE EFFECT OF AGE ON HABIT FORMATION IN THE ALBINO RAT INTRODUCTION The present investigation is concerned with the problem of the relation of the age of an animal to its learning capacity. Experiments were begun in the Psychological Laboratory of The Johns Hopkins University during the winter of 1912, and continued until the spring of 1915. HISTORICAL So far as the writer has been able to ascertain, practically no prolonged experimental work has been undertaken hitherto on the relation of age to learning ability, although the import- ance of the problem has been generally conceded. In the field of human psychology, Munn1 carried out a series of " substitution tests " on children in the grades, on normal school pupils, and on two elderly persons, to determine the relative rapidity of gain in ability to make the required sub- stitutions. Her records were taken in the terms of time, and showed that although the children gained much more rapidly than the adults, their actual rate of speed at the beginning was lower, and that they did not reach the same level of effi- ciency within the limits of the experiment. Only two elderly subjects were used, hence too much reliability cannot be at- tributed to the results from the last group, but apparently, while their initial rate is intermediate between that of the children and the normal school pupils, they fail to reach the final rate attained by either of them. Munn gives neither the average nor the rate of gain for this last group, but the former was easily obtained, and appears in the table below. Adults — first test 42 seconds — last test 14 seconds Children — first test 184 seconds — last test 32 seconds Old persons — first test 72 seconds — last test 39 seconds 1Munn. Curve of Learning. Archives of PsychoL, no. 12 p. 37. 1 HELEN B. HUBBERT Gain in first 5 tests: Adults 16 seconds Children 34 seconds Gain in second 5 tests: & Adults 5 seconds Children 14 seconds It would appear from these results, that, if the rate of improve- ment is the question considered, children learn about twice as fast as adults. Turning to the field of animal behavior we find a somewhat larger amount of experimental work on the matter under dis- cussion, although practically all of it occurs as a side issue to some other problem. Slonaker2 undertook a study of the nor- mal activity of the white rat at different ages, hoping to " ascer- tain how the age of greatest activity compared with that at which the rats were most capable of education." His con- clusions which relate particularly to the subject of this discus- sion are as follows: 1. " White rats of different ages show a marked difference in their activity. 2. " The very young rat and the very old rat are each notice* ably inactive. 3. " These experiments indicate that the age of greatest activ- ity ranges between 87 and 120 days. 4. " From these preliminary experiments no correlation can be made between the age at which they are most active and the age at which they learn most rapidly." In a later paper3 he places the age of greatest activity for the males at three hundred days, and for the females at three hun- dred and seventy-five days.4 The daily activity increases with the advance in age until a certain age is reached, after which there is a gradual reduction till death occurs.8 " The female is much more active than the male."6 There is seen to be a 2 Slonaker, J. R. The Normal Activity of the White Rat at Different Ages. Journ. Comp. Neur. and Psych., 17 ('07), 342-59. 3 Slonaker, J. R. The Normal Activity of the Albino Rat from Birth to Nat- ural Death, etc. Journ. Animal Behav., II ('12), 20-42. 4 Op. cit., p. 30. 6 Op. cit., p. 26. 6 Op. cit., p. 42. HABIT FORMATION IN THE ALBINO RAT 3 discrepancy in the results of the two papers which Slonaker does not attempt to explain. In the latter paper as in the earlier one no attempt is made to correlate amount of activity with capacity to learn. Yerkes7 raised the question of the relation of age to habit formation in the dancing mouse. He worked first on the ac- quisition of the white-black discrimination habit, and later on the learning of simple labyrinth pathways. The indices of modifiability as given by the number of training tests required to complete the habit for dancers of one and four months re- spectively show that the males learned the white-black dis- crimination habit more quickly at one month (30 days) than at four months (120 days) while the reverse was true of the females.8 The female was superior to the male, however, in the formation of the labyrinth habit.9 In later work10 he finds that male dancers ten months old learn the labyrinth more rapidly than those one to two months old, while there is prac- tically no difference in rapidity of learning of one to two month and ten month females. The old dancers are somewhat superior to the young in their ability to learn the labyrinth paths.11 With regard to the sensory habit he says: " 1. The dancer at one month of age acquires a particular white-black visual discrimination habit more rapidly than do older individuals. From the first until the seventh month there is a steady and marked decrease in rapidity of habit formation; from the seventh to the tenth month the direction of the change is reversed. These statements hold for both sexes. " 2. Young males acquire the habit more quickly than young famales, but between the ages of four and ten months the fe- males acquire the habit the more quickly."12 Haecker,13 in work on the Mexican axolytl, found that the habit of distinguishing between wood and meat when offered to the animals in forceps, was learned with far greater difficulty 7 Yerkes, R. M. The Dancing Mouse. The Macmillan Co., 1907. 8 Op. cit., p. 274. 9 Op. cit., p. 273. 10 Yerkes, R. M. Modificability of Behavior in its Relation to the Age and Sex of the Dancing Mouse. Journ. Comp. Neurol. and Psychol., 19 ('09), 237-271. 11 Op. cit., pp. 266-267. 12 Op. cit., p. 269. 13 Haecker. Arch. f. d. ges. Psych., 25, 1-35. 4 HELEN B. HUBBERT by the young (nine month) individuals than by the old ones, whose age is not given. Watson14 in his Animal Education, discusses work both on habits involving simple motor ability and on those requiring skill in manipulation. He concludes that " a young rat will solve for the first time more quickly than a mature rat any problem conditioned on mere random activity, but that a prob- lem involving associative activity and manipulation is more easily solved by the older animals." He found that with the simple saw-dust box the average time of entrance for the old rats was 85.50 minutes, while that for the young ones was 6.87 minutes and says further, " there is a gradation in the number of useless movements made by rats at different ages. At thirty- five days of age, when physical activity appears to have reached its highest stage, the percentage of useless movements is largest. As the rats grow older, this stiperabundant activity disappears, and in its place comes direction of activity." To summarize the main points in this brief historical survey, we may note: First: — That there is disagreement as to the age of greatest activity, Slonaker putting it first between eighty-seven and one hundred twenty days, and later at ten months for the males and twelve and a half months for the females, while Watson believes it to be at about thirty-five days; Second: — That Yerkes finds the labyrinth habit more easily learned by the old dancers than by the young, while if Watson's interpretation is correct the reverse should be true; Third: — Yerkes concludes that the female is superior to the male in learning the labyrinth. APPARATUS AND PROCEDURE Albino rats were chosen as subjects in this investigation, for several reasons: Slonaker, Watson and Yerkes worked with rodents, and we desired to compare our results with theirs. Nearly two hundred animals were required for actual experi- mental work and many more than that had to be kept on hand to provide for replacing any which might become unfit for work, and to allow for the usual losses through death and sickness. It has been found that white rats are easier to breed, handle, "Watson, J. B. Animal Education. University of Chicago Press, 1903. HABIT FORMATION IN THE ALBINO RAT 5 and care for in large numbers than any other small mammal. For reasons which will appear later, we adopted the circular maze as our problem since it is generally conceded that the rat is pre-eminent among animals in his ability to thread a labyrinth, while his satisfactoriness as a subject for experimental work is attested by the number of experimenters who have employed him in various capacities. The rats were bred in our own laboratory as needed, inbreeding being carefully avoided and all possible care being taken to maintain uniformity of breeding conditions. All of the rats were weaned at from eighteen to twenty-three days,15 and the sexes were separated at thirty-five to forty days and kept sepa- rate thereafter. The living cages were protected from mice and gray rats by screened compartments constructed of pine and one fourth inch wire mesh. Every two weeks the cages were thoroughly cleaned, the shelves washed with a disinfecting solu- tion, and the rats dipped in a one per cent solution of "Kreso" to prevent the rise and spread of vermin. The animals were carefully watched and treated immediately upon the appear- ance of parasites, so that they were kept continually in a healthy condition. The diet consisted of milk-soaked bread given every day, and a mixture of cracked corn and sun-flower seed every other day. They seemed to thrive on this somewhat restricted diet, so that no additions were made to it although both Basset and Ulrich used carrots and fruit occasionally. The rats were handled freely from birth, and consequently were perfectly tame and evinced no fear of the experimenter. Special care was taken to tame any rat seeming a little wild, before beginning work with him; since it was believed that fear and timidity might cause irregularities in behavior, a belief which was substantiated during the course of the experiment. It was desired in this work to obtain not only a record of time but also a distance record of the learning process, since it was felt that this might throw considerably more light on the factors involved in learning than had yet been obtained. The maze problem seemed to offer greater possibilities in this line than either sensory problems requiring a long and tedious course of preliminary training, or problems of manipulation permitting 15 No bad effects were noticed from this early weaning and the rats were found to be extremely active as early as the sixteenth day. See Slonaker, op. cit., p. 350. 6 HELEN B. HUBBERT of movements in two dimensions which would be practically impossible to trace. We therefore selected as our problem the learning of the circular maze. Heretofore, the only data possible on such a problem have been in terms of time and errors, the time being the only reliable record since it is practically impossible to evaluate and stand- ardize errors.18 With regard to this Miss Hicks 17 says: "The prevalent practice of omitting all total and partial returns from the error record, and of making no attempt to evaluate varying degrees of error gives a curve which is not only worthless but false." She says further: "The total distance criterion pre- sents so many difficulties as to render it impracticable for ordi- nary work. One difficulty lies in the matter of taking records accurately. The rats, after a few trials, run so rapidly that it is extremely difficult for one person to observe and record at the same time. To do this, it is necessary to mark off the maze into small segments and commit to memory some scheme of representation so that records can be jotted down in a purely automatic manner. The work of transcribing this record into distance terms and computing the same is very laborious. Eliminating these practical difficulties, the distance criterion is in some ways an ideal one. (italics mine.) There can be no diver- gence of practice as to what shall be omitted or included and results obtained by different experiments upon the same maze will be strictly comparable." " The distance and error criteria are fundamentally alike. The distance curve is the better repre- sentative of the progressive approximation of the act towards automatic accuracy. It portrays all the details of this elimina- tive process and it approximates the ideal of uniformity and regularity of descent. However, it is impracticable from the standpoint of recording and manipulating the data." These practical difficulties in " recording and manipulating the data " have been overcome, at least where small animals are the subjects used in the maze. The total distance can be obtained accurately by means of the camera lucida attachment designed by Professor Watson (see Fig. 1) for use with his "Watson, J. B. Noddy and Sooty Terns. Carnegie Pub., no. 103, p. 249, note 1. 17 Hicks, V. C. The Relative Values of Different Curves in Learning. Journ. Animal Behav., I, 138-156. HABIT FORMATION IN THE ALBINO RAT 7 circular maze. This maze has a wooden base one hundred and fifty centimeters in diameter and six aluminum runways fifteen and five tenths centimeters high and ten centimeters wide. The entrances to the alleys are ten centimeters wide, and are at alternate ends of a quadrant arc. The radial stops in alleys 1 to 5 are also placed at alternate ends of a quadrant arc, the stop in each alley being directly opposite its entrance. Thus, it is possible for a rat to run only one half the circumference of a runway in either direction before being forced to turn. This is not true of alley 6, where no stop is employed. The central circle, or food compartment is twenty centimeters in diameter. A three quarter inch mesh wire top prevents the animals from escaping, without interfering with observations of their movements. The camera lucida attachment consists pri- marily of two mirrors and an achromatic lens. The arrange- ment is as follows: A large plate glass mirror is fastened by supporting framework at an angle of forty-five degrees," with its center directly above and one and eight tenths meters from the center of the maze. A somewhat smaller mirror is placed facing the first and making an angle of ninety degrees with it at such a distance away that the light reflected downward falls outside the maze area. In the path of this reflected light is placed a single achromat six centimeters in diameter and of fifty centimeters focus in a mounting provided with rack and pinion adjustment which is fitted into the center of a wooden disc thirty centimeters in diameter. Below this at the focus of the lens is placed a second wooden disc of the same diameter as the first, which serves as a holder for the paper upon which the image of the maze is reflected. Both of these discs are attached to iron collars which slide independently up and down the rod CR, thus making it possible to vary the size of the image. A small curtain of dull black velvet attached to the upper disc serves to exclude all extraneous light from the re- cording table and as a further aid in sensitizing the eye, a large curtain of dark material encircles the space occupied by lens and recording apparatus as well as the experimenter's chair. This curtain also serves the purpose of completely hiding the experimenter from the animals while they are running in the maze. HELEN B. HUBBERT FIGURE 1 Maze with Watson Camera Lucida Attachment Illumination is obtained by means of six 40-watt tungsten lamps placed symmetrically around the maze and one 150-watt tungsten in the center. These lights are mounted on brass rods and fitted with aluminum shades blackened on the upper HABIT FORMATION IN THE ALBINO RAT 9 surface. The central shade is circular, those for the peripheral lights are half shades. The floor of the maze is covered with white linoleum, which can be thoroughly scrubbed whenever necessary. The entrance to the starting box is supplied with a hinged door which can be securely fastened after the animal has been placed inside. The exit is provided with a sliding door which is raised by means of a cord, and closes of its own weight when the tension on the cord is released, thus making it impossible for a rat to return into the starting box after it has once entered the maze.18 By means of the two mirrors (M and M'), and the lens (L), an exact image (I M) of the maze is thrown on the recording table where the experimenter can follow every movement of the animal during any passage through the maze. Actual records of these trips are made by tracing on the record sheet with a soft pencil the successive movements of the rat. (See Fig. 2). These tracings, measured with a chartometer shown by calibration to be accurate to within one per cent, form the basis for the distance record. Since the maze is six and four tenths times as large as the image, the distance record obtained in centimeters by the chartometer, must be multiplied by six and four tenths to obtain the actual distance traversed in the maze. For example, if the distance indicated by the char- tometer is one hundred and twenty-one centimeters we obtain the actual distance run, thus, 121 centimeters x 6.4 = 774.4 centimeters. The values given in the tables represent the actual distance covered by the rats. Both chart and maze distances were tabulated, and the multiplications made to ob- tain the latter were checked on the adding machine. In addi- tion to the distance record, such charted pathways also furnish 18 The maze was used exactly as described above throughout the work in order to maintain the same experimental conditions for all the groups. However, in the course of the experiments, several possible improvements suggested them- selves as being desirable: First. — The maze should be constructed of such material that it could be frequently flushed out with a hose, and a plug should be fitted into the bottom to facilitate cleaning. Second. — Nitrogen lamps placed at crossfire above the maze would be better than the tungstens surrounding it, and would do away with the shadows caused by the aluminum shades, which, when a very small animal is the subject in the maze, are troublesome. Third. — A change in the lighting arrangement would make it possible to have the mesh top made in two pieces instead of four, and hinged to the sides of the maze so that it could be lifted easily and noiselessly, thus avoiding frightening the animal within the maze. 10 HELEN B. HUBBERT accurate account of the excess effort expended, enabling a com- parison as to the frequency and extent of the several possible errors as well as a record of the exact steps in their elimination. It can be determined whether a certain error is lessened at each trial and finally disappears, or whether it is dropped out all at once. In short we have an accurate method of tracing the several factors involved in the learning of the maze problem, and a basis for the analysis of the learning process which has heretofore been lacking. FIGURE 2 Actual Tracing of Pathway Traversed by a Rat in the Maze The exact method employed in this research concerning the relation of age to the learning ability was as follows: One week preceding the day on which the animal was to begin work, food was removed from the living cage and the rat was fed each day in the center of the maze which was tempor- arily partitioned off from the remainder, making it impossible for him to roam at will through the maze. The first day, he was allowed to eat for forty-five minutes; the second day, for thirty minutes; the third day, for twenty minutes. The feeding time was then diminished five minutes on each succeeding day, so HABIT FORMATION IN THE ALBINO RAT 11 that the day before beginning the problem, the rat had been fed for five minutes in the food box of the maze.19 Two things were accomplished by this procedure. 1st: The rat was ren- dered quite hungry, a necessary step since food was the stim- ulus used, but the shock which would have resulted from entire absence of food was avoided. 2nd: It became accustomed to some extent to experimental conditions. On the day when the problem was actually begun, the tem- porary partition was removed from the maze, a dish of milk- soaked bread placed at the center and the rat put into the starting box (S. B. Fig. 1). The instant it emerged into the maze proper, the door (indicated but not shown in the illustra- tion), was closed behind it, making return into the starting box impossible, the stop watch was started and the tracing begun. Twelve or fourteen minutes might be required to reach the food, and as many as sixteen sheets of paper have been necessary to trace the pathway during a single trial. At the moment of entrance into the food box (F. B.), the watch was stopped, the time noted, and the animal at once removed. This constituted one trip or one trial. The rat was immediately introduced for a second trial, in which the same procedure was followed except that on reaching the food it was allowed to eat for five minutes before being removed. The feeding period was carefully timed with the purpose of keeping the hunger stimulus as uniform as possible. A short ration of grain was thrown into the living cage, and no more food was allowed until the next day's work. Basset20 had given grain only twice a week, and noted in consequence a disturbance in behavior on the day following that on which grain was given. Ulrich21 fed his ani- mals in the cage after work, which may account for their slow- ness in learning the maze as compared with the rats used in this problem. Two trials were given each day until the problem was learned, i. e., until in six trips made on three consecutive days no error was made from start to finish. In both Basset's and Ulrich 's work, a time norm was set, and, although no useless movements were made, unless the act was performed within the limits of 19 Grain was given in the cage each day at the end of the feeding period. 20 Basset, G. C. Habit Formation, etc. Behavior Monograph, 2 ('14), no. 4. 21 Ulrich. Behavior Monographs. Vol. II, No. 5. 12 HELEN B. HUBBERT the time set, it was not considered perfect. For the purposes of this experiment such a norm was not desirable, since one of the points under investigation was the relative final rate of efficiency attainable by rats of different ages. Elimination of all useless movements for three days was therefore considered as sufficient evidence that the problem had been learned. The number of trials required to reach this level of efficiency varied with each rat, the extreme limits being fourteen and one hun- dred twelve trials. In a single trial any distance greater than four and five tenths meters, which is the length of the errorless pathway from the entrance to the food, represents excess effort on the part of the animal. If a rat remained in the maze for fifteen minutes without reaching the food box he was taken out and replaced in the entrance box for a second attempt. Distance and time were recorded in the same way as for a successful run, i. e., if the first effort to reach the food proved unavailing after fifteen minutes, and the second attempt was successful after eight minutes, the total time for the first trial would be twenty-three minutes and the total distance the combined distance of the two attempts. Should the rat fail on the second effort also, it was fed for three minutes in the maze with the food box partitioned off as for preliminary feeding, and tried again the following day. The time and distance records for each trip were carefully tabulated, and form the basis for the conclusions which appear later. In many cases the actual tracings were kept for reference. EXPERIMENTAL RESULTS It was planned to work with five groups of rats, twenty-five, sixty -five, two hundred, three hundred and five hundred days old respectively, since it was thought that these ages represented fairly well the successive stages in the growth and development of the animal; twenty-five days for youth, sixty-five days for sexual maturity, two hundred days for maturity, three hundred days for age, and five hundred days for old age. The attempt was made to have thirty rats in each group, but sickness and unavoidable accidents among the animals have brought the number somewhat lower. It has been found extremely difficult to obtain rats for the last group (500 days). Although Slonaker finds the average length of life of the white rat to be thirty-four HABIT FORMATION IN THE ALBINO RAT 13 months,22 and Donaldson gives it as three years, from three hundred to four hundred days is the maximum longevity for most of the rats used in this laboratory, and up to this time only twelve have lived to work at the five hundred day age, one of these dying apparently of old age before the problem was learned.23 The groups used, with the number of rats in each group were as follows: Age at which work began Number of rats in the group 25 days 27 65 « 27 200 " 28 300 " 28 500 « 12 Throughout the experiment two conditions have been rigidly complied with. 1st: Every animal was started on the problem upon the exact day at which the proper age was reached. This procedure was followed even when it necessitated starting eighteen rats on the same day, in order that experimental con- ditions might be kept strictly comparable. 2nd: Every animal was run twice every day from its first trial until the last, even though at one stage of the work this required having as many as fifty-eight rats under observation at one time. Such strict continuity of trials precluded the introduction of any factors aside from those involved in the learning process proper. Removal from experimental conditions for even one day would not only cause a change in the physiolog- ical tonus of the organism, but would also bring in the matter of retention. So far as was possible the rats were run at the same hour each day, but where large numbers were being used, it was impossible to adhere strictly to this rule, although rats accustomed to run at night did not do so well if used in the mornings and vice versa; this was probably attributable to the acquiring of a certain food rhythm that might not be broken with impunity. Thus it was found that while a difference of an hour in the working time, and hence the feeding time, caused no 22 Op. cit., Journ. Animal Behav., pp. 37-38, tables. 23 Dr. Wats9n has informed the writer that his experience was quite similar, very few of his rats living to be more than 500 to 600 days old. 14 HELEN B. HUBBERT noticeable change in the behavior of the rats, marked disturb- ance resulted from a delay of four or five hours. In general, the behavior of individuals of each group on first entering the maze was the same. The fats showed great hesi- tancy in leaving the starting box, returned to it frequently after finally entering the maze proper and endeavored to push up the sliding door; they were slow to leave a familiar alley for one unexplored, became excited when a stop was encoun- tered, trying repeatedly to push it aside or to gnaw through the mesh top, and made frequent efforts to escape from the maze. Departure from this type of behavior was noticed among the very old rats and the very young ones. Many of the former evidenced no excitement whatever, often sleeping for several minutes between period of activity, while the latter were far more active than the rats of any other group, and showed no hesitancy in entering unfamiliar portions of the maze. The time usually decreased very rapidly, the distance less so, during the first three or four trials. For example, on its first trial, rat 34 of the three hundred day group required eleven minutes and forty seconds to reach the food, and the distance covered was forty-nine and six tenths meters. On its second trial, seven minutes six seconds were required, and the distance run was thirty and nine tenths meters ; at the fourth trial, success was attained after one minute nineteen seconds, the pathway traversed measuring ten and two tenths meters, while for the sixth trial, the time record was only forty-nine seconds, the distance eight and six tenths meters. By the tenth or fifteenth trial, the decrease in both time and distance had become much more gradual, and continued so until the problem was learned. The rat referred to above, required on the fifteenth trial, twenty- four seconds, and ran seven and two tenths meters; on the thirtieth trial the trip occupied fourteen seconds, and covered five and six tenths meters. This particular animal completed the problem at the sixty-sixth trial, when the time record was seven and two tenths seconds, and the distance record four and five tenths meters, which, it will be remembered, constitutes a perfect run. The data set forth above may be conveniently tabulated thus: HABIT FORMATION IN THE ALBINO RAT 15 Time required Distance run in maze before reaching food. Trial to reach food (The true pathway 4.5 m.) 1st 11 min. 40 sec. 49.6 meters 2nd 7 " 6 « 30.9 u 4th 1 " 19 « 10.2 u 6th 49 « 8.6 u 15th 24 « 7.2 • 30th 14 « 5.6 u 66th 7.2 « 4.5 u This rapid decrease in time and distance at the beginning of the problem was characteristic of all groups, and is clearly shown in the initial drop in all the curves. TWENTY-FIVE DAY RATS Work on this group began at twenty-five days when the rats were so small that they could crawl through the mesh top of the maze, and could touch the sides of the alleys only by running from side to side, while other rats could remain in the center of the path and touch both walls of the runways with their vibrissae. These rats were weaned at eighteen days, and were fed in the maze for five days preceding the experiment, the forty-five and thirty minute feeding periods being omitted. For the first day or two after starting the problem, they were allowed to eat for six or seven minutes instead of five minutes at the end of each day's work, since it developed that a shorter ration had a weakening effect on animals so young. The little rats were exceedingly active, and on entering the maze ran so rapidly that it was very difficult, but never impossible, to trace their movements. For the most part they showed great eagerness to escape from the starting box, some even acquiring the habit of lifting the door partway with the nose, and as a rule they had no hesitancy in entering unexplored portions of the maze, in this respect differing from most of the rats in this experiment. The error of circling the food box occurred more often with rats of this group than with those of any other, the explanation being, perhaps, that in their over-eagerness to reach the food they acquired such momentum than they ran past the entrance to the food box. 16 HELEN B. HUBBERT Twenty-seven rats were experimented with at this age, eleven males and sixteen females, eight strains being represented as follows : WM 1/9/14 Y(CF) 2/10/14 GJ 3/12/14 AL 3/14/14 XL 3/15/14 KL 9/20/14 X W 11/1/14 Y W 11/1/14 Total Males . . . Females. All 0 2 2 1 1 1 1 1 1 3 3 5 1 3 2 0 11 16 2 3 2 2 4 8 4 2 27 The first letter indicates the father, the second letter the mother, of the litter. Individual rats were distinguishable from each other by a convenient system of ear marks, and on every cage was a tag showing the experimental number, parentage, date of birth, sex and ear mark for each rat contained therein. Thus, W M 1/9/14 R — 9 4, would be deciphered, rat number four, female, right ear straight, born January ninth, 1914, mother M, father W. The number of trials required by animals of this group in learning the problem varied from fourteen to fifty-one, the absolute time from four and nine-tenths seconds to nine and one-tenth seconds, the total time from sixty-four minutes to six hundred forty-nine minutes; and the total distance from one hundred thirty-nine meters to four hundred eighteen meters. The " absolute time " is the average time for the last six trials, represents the limit of efficiency in speed for a given group, and varies among individual rats within the group as well as for the groups themselves. Thus, the record time for the twenty-five day group was made by a rat which could run from entrance to food box in four seconds, but no other rat attained this speed, and one in particular could not make the run in less than eight seconds. The last six trials were all with- out error and would seem to afford a fair basis for judging the average final efficiency, which for this group was five and seven- tenths seconds. The absolute distance is the same for each group, since the last six trials are errorless, and the true pathway measures approximately four and five-tenths meters. HABIT FORMATION IN THE ALBINO RAT TABLE I RUN IN DAY TIME. RAT 15 — 25 DAYS Distance 17 Day Trial Time Seconds Chart Maze 1248.0 896.0 678.4 448.0 1011.2 768.0 460.8 678.4 448.0 768.0 X L 3/15/14 B- male 4/9/14/ 2 \ 4/10 / 3 \ 4/11 ) 4/12 j 5 \ 4/3 / 1 2 3 4 5 6 7 8 9 10 1'7.8" 67.8 37.2 15 10 25.8 17.4 11.8 18.6 10.4 25.8 195 140 106 70 158 120 72 106 70 120 Elim. 4-5-6 3-4-6-wrong turn-1 « 1-3-4-5 Perfect Elim. l-2-4-5-6-lost in 3 « 2-4-5-6-w. t. 1-3 « 2-3-4-5-6-W. t. 1 " 2-3-4-5-6-W. t. 1 Perfect Elim. 2-4-5-6-w. t. 1-3 6 \ 4/14 J 11 12 6 18 70 132 448 844 .0 .8 Perfect Elim. 3-4-5-6-w. t. 1-2 7 1 4/15 ) 13 14 5.6 6.2 70 70 448 44 .0 .8 Perfect u 8 \ 4/16 j 15 16 11. 32. 8 4 99 129 633 825 .6 .6 Elim. 2-3-4-5-6-ret. 1 « 2-3-4-5-6-ret., w. 1. 1 9 \ 4/17 / 17 18 7.8 5. 7. 7. 4 4 70 83 448 531 .0 .6 Perfect Elim. 2-3-4-5-6-W. t. 1 10 \ 4/18 j 19 20 5. 5. 4 2 70 70 448 448 .0 .0 Perfect u 11 \ 4/19 / 21 22 5. 6 4 70 ' 72 448 460 .0 .8 u Elim. 2-3-4-5-6-too far 1 12 1 4/20 / 23 24 5. 9 2 70 101 448 464 .0 .4 Perfect Elim. 2-3-4-5-6-W. t. 1 13 \ 4/21 / 25 26 5 5. 6 70 70 448 ' 448 .0 .0 Perfect u 14 \ 4/22 / 27 28 7. 5. 6 4 72 70 460.8 448.0 Elim. 1-2-3-4-6-W. t. 5 Perfect 15 \ 4/23 / 16 \ 4/24 / 29 30 31 32 5 5. 4. 4. 2 8 4 70 70 70 70 448 448 448 448 .0 .0 .0 .0 a u u Finished 4/25/14 34 trials 17 \ 4/25 / 33 34 4. 5. 6 6 70 70 448 448 .0 .0 u u 18 HELEN B. HUBBERT Rat 46* 6 8 10 15 16 19 24 27 36 37 TABLE II TWENTY-FIVE DAY RATS 11 males — 16 females Trials 18 45 32 28 34 27 40 24 51 30 30 Time (sees.) 2541.4 1903.4 959.8 699.6 423.2 1592.2 721.2 388.4 1588.0 1603.2 1632.8 Absolute Distance time (cm.) (sees.) 18035.2 38666.8 36732.4 23770.4 19424.0 30016.4 29207.2 16131.6 41816.0 28454.4 23070.4 5.7 5.1 5.3 5.0 4.9 5.5 5.5 6.1 5.6 9.1 5.5 19 2 5 9 11 12 13 14 17 20 21 22 25 28 31 41 Totals Averages Average for Average for 23 26 14 38 18 32 46 24 26 34 36 32 36 28 44 17 822 30 32 29 2623.2 37819.2 562.2 21913.6 3897.2 30348.8 723.6 27129.6 401.4 13900.8 2381.0 32648.8 999.4 30407.0 519.6 18553.6 612.8 18771.4 945.6 25779.2 900.6 26208.0 536.2 23979.2 597.6 30882.8 2643.2 36672.0 1838.2 30332.2 2082.8 23212.8 36317.8 733383.8 1345.1 27162.3 1277.5 27711.3 1391.5 26784.9 7.2 5.5 5.2 5.4 5.7 6.1 4.9 5.6 5.6 5.3 5.1 5.4 5.2 5.2 5.2 7.6 153.5 5.7 5.7 5.6 The enormous number of figures involved, makes the showing of individual records inexpedient. An exact copy of the daily record for rat number fifteen of the twenty-five day group, from the first to the last trial appears as table I. The averages and totals for each rat which appear in table II are obtained by adding and averaging the daily records for the individual rats. The total time, total distance, total number of trials and the absolute time of the twenty-seven rats were added and averaged to give the average total time, the average total dis- HABIT FORMATION IN THE ALBINO RAT 19 tance, the average number of trials, and the average absolute time for the group. The speed is the average number of centimeters traveled per second throughout the learning process, and is obtained by dividing the total time into the total distance. For the twenty- five day group these averages were : Time Trials Absolute Total Distance Speed 30 5.7 sec. 224 min. 271.6 meters 20.1 The curves shown in fig. 3 are based on the figures in col- umns 2, 4 and 6 of table II. Only one rat finished in less than fifteen trials, and this is indicated on the first point of the curve. Three rats finished at between fifteen and twenty trials, one at seventeen and two at eighteen trials each, the average being seventeen as indicated on the curve. Between twenty and twenty-five trials three rats finished, at twenty-three, twenty- four and twenty-four trials respectively; the average is twenty- three, and the third point on the curve indicates that three rats finished at twenty-three trials. The same procedure is followed in drawing the time and distance curves except that they are necessarily more condensed. Three rats required approximately four hundred seconds each in which to learn the problem, (numbers 11, 15 and 24), and the first point on the time curve indicates this fact, The fourth point shows that six rats consumed from fifteen thousand to twenty thousand seconds (average for the six, seventeen thousand seconds), in their total number of trials. The fifth point in the distance curve is interpreted to mean that six rats covered between three hundred thousand and three hundred fifty thousand cen- timeters, (average for the six, one hundred seventy thousand), in learning the maze. It might be well to notice at this point that all of the curves appearing in this paper are constructed on this same plan. 20 HELEN B. HUBBERT FIGURE 3-A Trial Curve for Twenty-five Day Rats. The trial curve, (Fig. 3-A) for this group reaches the apex at about thirty, which is the average number of trials for this age. HABIT FORMATION IN THE ALBINO RAT 21 FIGURE 3-B Time Curve for Twenty-five Day Rats. Two maxima appear in the time curve, (Fig. 3-B) at eight hundred seconds, and at seventeen hundred seconds re- spectively. A point intermediate between the two would give the time average for the group, approximately thirteen hundred seconds. 22 HELEN B. HUBBERT FIGURE 3-c Distance Curve for Twenty-five Day Rats. The apex of the distance curve, (Fig. 3-C) , is at three hundred thousand centimeters, which is not far from the group average of two hundred seventy thousand. No very close relation seems to obtain between the number of trials required for finishing the problem and the total amount of time or distance. Thus, the rat which finished in fourteen trials had the highest total time record in the group, and a high distance record, while the rat which finished in fifty-one trials had the highest distance record, but a mean time record. The lowest time record was made by a rat finishing in twenty-four trials whose distance record was also low; the lowest distance record, by a rat requiring eighteen trials for the problem, whose time was lower than the average. Except where the time values are very high, time and distance bear a fairly constant HABIT FORMATION IN THE ALBINO RAT 23 ratio to each other. The exceptions occur when the time record is increased on account of failures (each of which means a count of 900 seconds) in the first part of the learning process. In a trial which has one or more failures as a component part, the distance run is never proportional to the time spent, since if the rat does not refuse to run altogether after several vain efforts to reach the center, it will make frequent halts at the radial stops and at the entrance to the maze. It appears that we have here additional evidence for considering the distance as a more accurate measure of the learning process than the time. According to the averages which appear below, the females of this group are superior to the males in number of trials re- quired to learn, and in final efficiency (absolute time), but inferior in total time consumed, total distance covered, and speed attained throughout the learning process, so that on the whole, the males may be considered as slightly superior to the females. Time Trials Absolute Total Distance Speed Males 32 5.7 sec. 213 min. 277.1 m. 21.6 cm. per sec. Females.. 29 5.6 " 232 " 267.8 « 19.2 " " " SIXTY-FIVE DAY RATS24 These rats began the problem when sixty-five days old, and were fed in the maze for one week before actual experimentation 24 A group of twenty-seven thirty-five day rats was used early in the experi- ment, but the results obtained from their records were so at variance with those for other groups that it was felt there must have been some error in the experi- mental work. Accordingly a control group consisting of thirteen rats was trained near the close of the experiment, with no better results so far as consistency with the main body of averages of the other groups was concerned, but with exactly opposite results from those of the first thirty-five day group. Therefore neither set of figures is given in the body of this paper, since still further work is neces- sary at that age before any statements can be made regarding it. The averages obtained for the two groups are given below: Time Trials Absolute Total Distance First group 63 6.7 sec. 564 min. 552.2 m. Control group...... 22 7.3 « 160 « 156.7 " 24 HELEN B. HUBBERT began. They were lively, but did not show the superabundant activity of the twenty-five day group, and were not so speedy. Twenty-seven rats were run, sixteen males and eleven females, representing nine strains as follows: EJ 11/1/13 EL 11/19/13 EL 12/3/13 WM 1/8/14 Y(CF) 2/10/14 G J 3/12/14 AL 3/14/14 XL 3/15/14 Total Males.. Females 0 3 3 0 2 0 4 3 3 2 0 1 1 0 1 2 16 11 All 3 3 2 7 5 1 1 3 27 Trials varied from fourteen to sixty-five, absolute time from four and seven tenths to eleven and eight tenths seconds, total time from sixty-four minutes to seven hundred thirty-one min- utes, and total distance from ninety-one and eight tenths meters to seven hundred fifty meters. Here, as in the preceding group, we can trace no close connection between number of trials and time or distance. The rat which finished in the fewest number of trials had a low time record and the lowest distance record for the group, while the one requiring the greatest number of trials had the highest time and distance records. So far the relation seems very close. But the lowest time record was made by a rat finishing in twenty- two trials, whose distance record was high, while two other rats which finished at twenty- two trials had very high time and distance records. The next to the highest distance record was by a rat finishing in fifty- four trials, while the next to the highest time record was made by one which finished in twenty-two trials. In general, where the trials run very high (65, 54) or very low (14, 16) the dis- tance corresponds rather closely, but for the trials lying between these extremes no such correspondence can be traced. The ratio between time and distance in this group is by no means constant. See table III. The group averages are: Time Trials Absolute Total Distance Speed 31 6.8 sec. 219 min. 260.6 m. 19.8 cm. per sec. HABIT FORMATION IN THE ALBINO RAT 25 FIGURE 4-A Trial Curve for Sixty-five Day Rats. There are two maxima in the trial curve for this group, (Fig. 4-A), at twenty- two and thirty-seven respectively, and the group average lies midway between the two, at thirty-one. 26 HELEN B. HUBBERT FIGURE 4-B Time curve for Sixty-five Day Rats. The time curve (Fig. 4-B), shows the highest point at six thousand seconds and one almost as high at twelve hundred seconds, while the average number of seconds for the group was eleven hundred. HABIT FORMATION IN THE ALBINO RAT 27 FIGURE 4-c Distance Curve for Sixty-five Day Rats. For the distance curve, (Fig. 4-C), there is a clearly marked maxima at twenty-six, which corresponds exactly with the group average of twenty-six hundred centimeters. 28 HELEN B. HUBBERT TABLE III SIXTY-FIVE DAY RATS 16 males — 11 females Absolute Time Distance time Rat Trials (sees.) (cm.) (sees.) 26 c? 14 454.0 9184.0 11.8 4 16 496.0 11603.2 9.9 5 22 2378.0 26675.2 6.5 6 24 452.4 18182.4 6.3 7 22 645.2 16454.4 6.6 8 30 385.8 19718.4 6.6 9 38 981.4 25117.6 7.2 10 26 519.2 19481.6 5.6 11 18 1184.8 18899.2 5.8 12 34 1724.8 30498.8 6.4 13 22 721.0 16626.4 5.6 14 22 2700.4 29484.8 5.9 20 36 1051.8 29491.2 6.9 21 46 1436.4 33280.0 5.9 22 21 1344.0 21305.6 6.6 24 38 1982.0 39326.0 10.6 19 30 1327.4 23033.6 8.2 2 54 1651.8 42368.0 7.9 3 21 1542.0 19603.2 7.1 15 28 2028.0 28921.6 5.5 16 36 1139.4 25177.6 5.3 17 36 454.4 20819.2 4.7 18 65 4388.4 75001.6 5.5 19 32 1312.0 25638.4 4.9 23 38 815.8 28409.6 7.6 27 40 1714.6 35584.0 6.4 28 20 597.4 13849.6 6.7 Totals 830 35428 4 703735 2 184 0 Averages 30.7 1312.1 26064.2 6.8 Average lord4 ... 26.8 1153.6 22833.0 7.1 Average for 9 ... 36.5 1542.8 30764.2 6.3 Comparison of the male and female records in the group show that while the absolute time of the females is less than that of the males, in other respects the male showing is the better. Time Trials Absolute Males... Females. 27 36 7.1 sec. 6.3 « Total 192 min. 257 " Distance 228.3 m. 307.6 « HABIT FORMATION IN THE ALBINO RAT 29 Two HUNDRED DAY RATS These rats were put to work when two-hundred days old, having been fed in the maze for one week preceding the begin- ning of the problem. They were more erratic than those of any other group used, being jerky and irregular in their movements. Often after making a perfect run in six or seven seconds a rat would drop back to one, two or three minutes with many errors! This behavior was not noted in any other group except in one or two isolated cases. Twenty-eight rats were used, fifteen males and thirteen fe- males, eight families being represented as follows: BD 7/26/13 CC2 7/30/13 WG 8/24/13 BD 8/28/13 BH 9/18/13 BH 9/25/13 WK 10/2/13 XL 3/15/14 Total Males.. Females All 1 0 1 2 3 2 1 2 2 4 3 2 2 0 2 1 15 13 1 3 5 3 6 5 2 3 28 Trials varied in number from fourteen to one hundred twelve, absolute time varied from five and two tenths seconds to twenty- four and one tenth seconds, total time from eighty-two to nine- hundred forty-nine minutes, and total distance from one-hundred twenty-five and five tenths meters to nine-hundred and twelve meters. As for previous groups, trials bear little relation to time or distance, but the proportion between total time and total distance is fairly constant. (See Table IV). The lowest 30 HELEN B. HUBBERT TABLE IV Two HUNDRED DAY RATS 15 males — 13 females Rat 2C? 5 8 9 10 11 23 24 29 30 31 33 34 36 38 39 4 6 7 15 17 18 19 20 21 25 27 35 Totals... Averages. Average Average for Q Trials 81 28 44 26 20 20 30 30 104 107 64 32 22 32 14 26 112 18 54 32 56 79 49 27 32 35 22 37 1170 41.7 39.4 44.5 Time (sees.) 810.2 1548.2 1767.0 646.3 496.4 1548.6 1191.0 1564.8 2636.1 4735.6 1877.2 1343.8 1084.6 777. 8 732.8 1280.4 3014.0 1825.2 3851.1 2409.0 2995.8 3663.4 2031.4 5694.8 2440.0 2990.6 963.4 3135.8 36294.9 2109.1 1517.3 2791.9 Distance (cm.) 51010.4 21932.4 30275.2 16672.0 12559.2 18387.2 22828.8 28355.2 64632.2 91293.8 47627.2 25446.4 18240.3 18675.2 12569.6 19744.0 69781.6 13475.2 51143.4 31522.8 46335.8 59373.2 44944.4 32960.0 35686.0 35897.6 19043.6 45104.8 949517.5 33911.1 29633.6 38847. 1 Absolute time (sees.) 8.7 8.6 10.7 5.7 6.8 7.4 7.0 7.6 17.3 24.1 6.5 8.5 10.5 8.1 7.9 9.4 7.4 8.0 7.2 5.2 6.5 6.6 5.4 8.7 6.1 7.8 7.8 7.3 238.8 8.5 9.7 7.2 trial record was fourteen, and the time and distance records for the rat making this record were also low. The lowest time record, as well as the lowest distance record was made by a rat finishing in twenty trials. The rat requiring the largest number of trials (112) had a time record but little higher than one which required only thirty-four trials, while its distance record was next to the highest. The highest time record was that of a rat which finished in one hundred seven trials, with the next to the highest distance record, the highest distance record by one finishing in twenty-seven trials, whose time was consid- erably above the average. HABIT FORMATION IN THE ALBINO RAT The averages for this group are: Time Trials Absolute Total Distance Speed 31 42 8.6 sees. 351 min. 339.1 m. 16 cm. per sec. FIGURE 5-A Trial Curve for Two Hundred Day Rats The apex of the trial curve, Fig. 5-A, lies at thirty-three, while the average for the group is forty-one, but the explanation of the apparent discrepancy is to be found in the records of the six rats who required from fifty-six to one hundred twelve trials to learn the problem, thus running the group average up. No well defined apex can be found in the time curve (Fig. 5-B), the group average, twenty-one hundred seconds showing rather as a depression, nor could it be divined by a glance at the dis- tance curve (Fig. 5-C) that the average lay in the neighborhood of thirty-three thousand centimeters. 32 HELEN B. HUBBERT FIGURE SB Time Curve for Two Hundred Day Rats HABIT FORMATION' IN THE ALBINO RAT 33 FIGURE 5-c Distance Curve for Two Hundred Day Rats. If the averages for the males and females be compared, it appears that the former are somewhat superior to the latter except in absolute time. The averages are: Time Trials Absolute Total Males. .. Females. 39 45 9.7 sec. 7.2 « 219 min. 465 « Distance 296.3 m. 388.4 " THREE HUNDRED DAY RATS This group consisted of twenty-eight rats from ten families, thirteen males and fifteen females, who began the problem when three hundred days old. 34 HELEN B. HUBBERT