VAEYENG HEMULATEON AND EMPREE‘VTENG {N DQG$ ThesEs {‘or fin Degree of DH. D. MECHECAN STATE UN ‘JEKSETY John N. Mart EQBL This is to certify that the thesis entitled VARYING STDTULATION AND IMPRINTING IN DOGSV' presented by John N. Marr has been accepted towards fulfillment of the requirements for m degree in imbalagy Maior professor Date November 16, 1961 0-169 LIBRARY Michigan State University s" : A -. h. :1“ blah... - . ‘ V University Ii ‘. . u‘V Ll ABSTRACT VARYING STIMULATION AND IMPRINTING IN DOGS by John N.. Marr The process by means of which certain visual stimuli acquire with extreme rapidity the capacity to elicit responses that an animal ordinarily directs only towards members of its own species has been called imprinting. A review of the studies of imprinting and observa- tions of the care-giving activities of the mammal mother suggests that varying stimulation is an important variable in the development of the preference for the species in the young animal. As a result of the mother consistently supplying varying exteroceptive and proprioceptive stimulation to the infant animal, the infant associates the varying stimulation with the stimulus-characteristics of the mother, and the mother comes to elicit approach responses from the young animal. It was the purpose of this study to demonstrate that a visual pattern which has been presented to young animals in conjunction with varying exteroceptive or proprioceptive stimulation will come to elicit approach responses from the young animal. Thirty puppies from four different breeds and nine litters were assigned to three experimental and two control groups. When the 58 were three weeks of age, each was placed in a training box for three minutes three times a day for seven days. The experimental group 53 were given varying stimulation while facing a white circle in the box. Ss in Group I were rubbed (tactile stimulation), .83 in Group II were rocked (proprioceptive stimulation), and a light was intermittently John N. Marr flashed on the circle while 55 in Group III were in the box (visual stimu- lation). Ss in the two control groups were not stimulated in the box. The circle was in the box for one of the control groups (Group IV) but not for the other (Group V). The emotional disturbance of each S was rated during training. When 53 were 28 days old, each was tested in an open-field box in the presence of the circle and alone. On day 29, they were retested once alone and once with the circle present. The length of time 53 spent near the circle was automatically recorded by a platform wired to a clock, and was considered a measure of the puppy's attachment to the circle. Measures of the emotional disturbance of the Ss in the tests were also taken. The groups differed in emotional disturbance during training. The visually stimulated group and the control groups were rated high in emotionality and the 83 that were rocked and rubbed were rated low in emotionality. The comparisons of the breeds during training indi- cated that the German Shepards were rated highest and the Poodles lowest in emotional disturbance under all training conditions. An analysis of variance of the platform time during testing indi- cated that the groups that were stimulated during training Spent more time on the platform during circle tests than during tests alone while there was no difference in platform time between tests for the control groups. These results supported the hypothesis that a visual pattern which has been presented to puppies in conjunction with varying stimu- lation will come to elicit approach responses from the puppies. All groups were less emotionally disturbed in tests with circle than in tests alone. This finding was believed to be due to the decrease in crying when 83 approached the circle and the resumption of crying when 53 reached the circle. None of the animals had been allowed to approach the circle during training. Date \k 114 (76/ Approved -jfifl’l C' @M Major flofessor VARYING STIMULATION AND IMPRINTING IN DOGS BY \ If ‘ .\/ John NE" Marr A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1961 I W' v 4‘ ’C' ’p_ .' I "a. .l 2’" ~ .._.. . gas I 4 ACKNOWLEDGMENTS The author wishes to express his appreciation to Dr. Ratner whose excellent guidance helped make this study possible. I am in- debted to him for the contribution of his time and experience. .I am especially grateful for the interest and encouragement which he gave me during the formulation of the problem and which resulted in the purpose of the study being changed from an attempt to describe the development of behavior in the puppy to an attempt to explain this development. I am obligated to Dr. Terry Allen, Dr. Charles Hanley, and Dr. Joseph Reyher for their time and their assistance. Dr. Allen's advise and suggestions were invaluable in the analysis of the results. The original work leading up to the present study was done under the guidance of Dr. Hanley, and his criticisms have had considerable in- fluence on my thinking during the formulation of this problem. I am very grateful to Dr. Reyher for his friendship, his interest in my work, and for his willingness to serve on my committee under such short notice. I would also like to thank the members of the Ingham County Kennel Club who have kindly helped me locate the animals used in this study. I am especially indebted to my friends, Mr. and Mrs. T. W. Bradbury Jr. , Mr. Harry Bridge of Harbridge Kennel, Mr. and Mrs. Frank Mainville of Val-Valle Kennel, and Mr. and Mrs. D. Ruth of Pine Tree Kennel for the use of their puppies, their homes, and their kennels. This paper would not have been possible without the patience, understanding and assistance of my wife, Jo. . She has given me constant encouragement and support throughout my graduate education. >k >:< 3:: >:< 3;: 3:: >{z :1: a}: >:< >:< >:< >3 3:: >}< a}: ii TABLE OF CONT ENTS Page INTRODUCTION . ................. . ...... l Imprinting . . . ...................... 1 Development of the Species-Preference in Mammals Varying Exteroceptive and Proprioceptive Stimulation in Infancy . ............ . ........ 5 THE PROBLEM ......................... 8 METHOD ............................. 10 General . . .......... . ............. 10 Subjects .......................... 10 Apparatus ........... . ............. 12 Procedure ................. . ....... 12 RESULTS . . . .................. . ...... l7 OBSERVATIONS ......................... 26 DISCUSSION ........................... 30 Test Results ........................ 31 SUMMARY . ........................... 40 REFERENCES .......................... 44 APPENDICES .......................... 47 iii TABLE II. III . IV. VI. VII. VIII . - LIST OF TABLES Analysis of Variance 'Rating of Type of Sound in Tests iv Page Distribution of Breeds Within Groups ......... 11 Mean Rating of Type of Sound During Training . . . . l8 . Mean Rating of Type of Sound During Training of Breeds ...... . .............. . . . . 18 Mean Time on Platform During Tests ....... . . 19 Analysis of Variance Mean Time on Platform During Tests .......................... 21 Mean Sound Duration in Seconds During Tests 22 Analysis of Variance Mean Duration of Sound in SecondsinTests............... ..... 23 Mean Rating of Type of Sound During Tests ...... 24 25 LIST OF FIGURES FIGURE Page 1. The Training Box . . . . ............... 13 2. The Testing Box . t ................. 13 3. Test behavior of a puppy that was rocked while facing the circle during training .......... 33 4. Test behavior of a puppy that was rubbed while facing the circle during training .......... 33 'LIST OF APPENDICES APPENDIX Page A. Sex Distribution within Breeds and Groups ..... 48 B. Correlations Between Sound Ratings and Sound Duration in Tests . . . . . . ......... . . . . 49 vi INTRODUCTION In recent years considerable attention has been givento the problem of how the organism develops its preference for its own species. . Investigative techniques have changed from observations of the mother-infant relationship to studies of the specific variables which are involved- The results of these studies are examinedlin this paper and compared to the results found in those studies in which infant organisms have been subjected to stimulation and stimulus-depriv- ation. (Specifically, theories on imprinting, stimulus deprivation, and varying stimulation are discussed. Imprinting Lorenz has described imprinting as a process or mechanism by means of which certain visual stimuli acquire with extreme rapidity the capacity to elicit responses that an animal ordinarily directs only towards members of its own species. Thorpe (1956) has stated that imprinting seems to have four unique characteristics, (1) the confine- ment of the process to a very definite and brief period of the individual's life, (2) the stability or perhaps total irreversibility of the capacity once it is accomplished, (3) the completion of the process long before the various specific reactions to which the imprinted pattern will become linked are established, and (4) the learning of the broad characteristics of the species. The characteristics of imprinting are important because imprint- ing has been suggested as a proceSS whichleads to the preference for the species. . In a review of recent studies of imprinting, Moltz (1960) has shown that the characteristics suggested by Lorenz and his co-workers arenot as typical of imprinting as previously thought. Although precocial birds, those which move about freely when hatched, show a brief stage of development during whichimprinting is most likely to take place, the period is not rigid or absolute. That is, some » birds do not follow an object presented to them after this period. Furthermore, the process is not irreversible as formerly believed but has been shown to decrease with time if the imprinted object is with- held from contact with the subject. To account for the apparent irre- versibility of the process, Beach and Jaynes (1954) have suggested that imprinting can be explained in terms of the learning of a response to a stimulus, (such that the occurrence of that re8ponse each time the stimulus presents itself prevents the acquisition of other types of behavior which could compete with the original response. Moltz (1960) points out that some generalization of the response has been shown in all imprinting studies which would account for the supra-individual characteristics of the response. Since imprinting has also been shown to increase when the object is accompanied by aversive stimulation (shock), Moltz postulates that anxiety reductionis the rein- forcement for the behavior observed in imprinting studies. . In the natural situation, low anxiety is created by the unfamiliar and results in anxiety reactions. These reactions are conditioned to an attention- evoking object. Specifically, Moltz says, . As a result, the object acquires the capacity to function as a reinforcer, henceforth mediating new learning. When anxiety is aroused, any response instrumental to bringing the animal in contact with the familiar object will be closely followed in time by anxiety reduction due to the previously acquired capacity of the object to elicit responses incompatible with anxiety. Vigorous pursuit of the object should then occur until either the stimulus situation ceases to arouse anxietyor, the reward value of the object is extinguished. (Moltz, 1960) Development of the Species-Preference in Mammals Although a learning model is proposed to account for imprinting, the conclusions drawn from the studies on birds can not necessarily be generalized to mammals. Both birds and mammals show preference as adults for their own species, but the Species preference may have developed in entirely different manners. . In a comparative study of the different phyla, Thorpe (1956, p. 8) has pointed this out, "living beings do the same thing by many different routes and in many different ways, and they may well all be different from the way in which the calculating machine achieves its results. " Hess (1959) and Thorpe (1956) cite reports of imprinting in guinea pigs, sheep, deer, and buffalo but most of these reports are of an observational nature. The relationship between young and adult animals varies consider- ably from species to Species, as it does, with birds. . Some animals are so mature at birth that they are only dependent upon the adult for food, while others are completely dependent upon the adult for food, warmth, cleaning, protection and in some Species, tuition. (Cruikshank, 1954, Thorpe, 1956, and Scott, 1958) The association of the mother with these need-reducing activities might account for the preference for the species whichis shown by the young in later life. The mother is associated with need-reduction by the young animal because She feeds, cleans, and warms the young. As a result of this association, the stimulus-properties of the mother acquire the capacity to function as a reinforcer. . It can be further postulated that the responses elicited by the mother are later generalized by the young animal to the Species. However, recent studies have indicated that a need-reduction model may not be necessary to account for the development of the attachment between the young animal and its parent. . In a study by Harlow and Zimmerman, (1959), infant monkeys separated from their natural mothers at birth preferred non-food-giving surrogate mothers covered with cloth to food-giving mothers not Covered with cloth. Harlow (1959) has called this contact comfort. Studies on dogs have also indicated that contact comfort may be an important variable in the development of the attachment of the young to the species (Igel and Calvin, 1960, and Marr, 1960). Bowlby (1958) has suggested that a theory of component instinctual responses, species Specific, can best account for the human infant's tie to its mother and that the instinctual responses of humans include con- tact, clinging, sucking, crying, smiling, and following. Such responses as following a moving object for precocial birds, moving to maintain contact with a soft object for dogs, and clinging to a soft object for primates can be thought of as examples of such instinctual responses of the respective Species. Both Bowlby (1958) and Harlow (1959) believe that these responses are independent of drive reduction but serve to keep the young organism in contact with the adult. Although the component instinctual responses differ from one species to another, they are similar in that they supply the young organ- ism with varying degrees and kinds of stimulation. . The movements of the young animal which keep it in contact with the adult result in varying exteroceptive and proprioceptive stimulation for the young animal. . In this paper, exteroceptive stimulationwill be considered to be physical energy which is converted into receptor energy by the senses of vision, hearing, taste, smell, touch, pain, and temperature- Proprioceptive stimulation will refer to that mediated by the vestibular and kinesthetic senses. The distinction‘between various kinds of stimulation such as exteroceptive and proprioceptive, is still a serious theoretical prob- lem but for the purposes of the present investigation conventional names are being used. Hebb (1958) has pointed out that stimulus deprivation studies with mammals and "brain washing" experiments .on humans indicate that "varying exteroceptive stimulation is a universal reinforcing agent. " He also suggests that these studies indicate that varying exteroceptive stimulation serves to arouse the organism and prepare it for inter- action with itS environment. Hebb (1955) and others have also suggested that there seems to be an optimal level of arousal for effective behavior. That is, a mild degree of stimulation is pleasing but a high degree of the same kind of stimulation brings about emotional disturbance. Varying Exteroceptive and Proprioceptive Stimulationin Infancy The question that is raised in this paper is whether or not varying exteroceptive and proprioceptive stimulation (VEPS) is important in the development of the attachment of the young animal to the adult of its own Species. . If it is a crucial and general variable, it Should have been present in those experiments in which imprinting has been demonstrated. The imprinted object, in all of the observations reviewed by Hess (1959) and Moltz (1960) moved past the young birds and they followed it. That is, varying exteroceptive and proprioceptive stimulation can be con- sidered to have been present. Specifically, the moving object supplied visual stimulation and the movements of the young bird resulted in kinesthetic stimulation. In those studies in which shock and flickering lights were shown to enhance imprinting, VEPS could also be postulated. The adult mammal characteristically supplies VEPS to the young animals as She "mothers" them. The mother rat, dog, or cat upon returning to the nest often licks the bodies of the young before feeding them. This licking appears to be very vigorous to the point of being rough. A The young are rolled over, pushed, and even held with a paw during the "mothering. " The primate mother is also very rough with the young as can be seen in grooming, playing, and exercizing activities (Yerkes and Tomilin, 1935). Varying stimulation is also characteristic of the mother-child relationship of humans. , The mother rocks, pats, grooms, and talks to the baby. Sensory deprivation studies of humans and non-humans suggest the importance of VEPS during infancy. Boulby (1953) and Brody (1956) conclude from their reviews of studies of institutionalized children that the lack of "mothering" results in. emotionally withdrawn and isolated children. Boulby, in describing these children, says, Prolonged breaks (in the mother-child relationship) during the first three years of life leave a characteristic impression on the childs personality. . . . They fail to develop loving ties with other children or with adults and consequently have no friend- ships worth the name. While Bowlby attributes the resultant personality of the children to the prolonged breaks in the mother-child relationship, it is also possible to attribute the effect to the lack of varying stimulation that so often characterizes the childrens' institution. . This lack of stimulation may have resulted in the failure of the infant to develop a preference for his Species. . Sensory deprivation studies with animals indicate thatthe lack of VEPS during infancy affects other classes of behavior later in the animals? lives. Bovard (1958), reviewing the studies on the effects of early handling on the viability of the rat, concludes that early handling increases viability under stress. Bovard reports additional research on stimulus deprivation that tends to support a hypothesis by Thompson that the total quantity of sensory interaction with the environment rather than the specific nature of the sensory input is the essential factor involved in the effects of early experience on viability. Early sensory deprivation also results in a profound change in the intellectual, perceptual, and emotional behavior of adult dogs (Beach and James, 1954; Melzack, 1954; Melzack and Thompson, 1956; Thompson and Heron, 1954). Specifically, early stimulus-deprivation has been shown to lead to diffuse emotional activity towards strange objects, poorer performance on intelligence tests, permanentwith- drawal from the experimenters, and an inability to cope with other dogs. . It is important to note that experiments on sensory deprivation and stimulation have used a wide variety of procedures, species, ages, and measures. The findings, which are summarized above, may be specific to the particular Species, age, etc. used in the studies. However, as a body, the findings support the notion that the presence or absence of VEPS has an effect on later behavior. THE PROBLEM The major hypothesis of this paper is that as a result of the mother’s supplying varying exteroceptive and proprioceptive stimu- lation to the infant animal, the infant associates the varying stimulation with the stimulus-characteristics of the mother. As a result of this association, the mother comes to elicit approach responses from the young animal. These approach responses are later generalized to other animals having similar stimulus-characteristics, the species. It is the purpose of this experiment to attempt to demonstrate only the first step of the development of the preference for the Species. That is, the experiment will be designed to demonstrate that a visual pattern which has been presented to young animals in conjunction with VEPS will come to elicit approach responses from the animals. 9 Dogs were used in this study because of the wealth of infor- mation that is available on the maturation of their sensory modalities and because the experimenter was more familiar with this species than with any other animal. Puppies respond to tactile, thermal, gustatory, and olfactory stimuli at birth, but do not, respond to visual and auditory stimuli until approximately 21 days of age and do not walk until approxi- mately 18 days of age (James, 1952; Scott and Marston, 1950). Puppies were trained and tested at the third and fourth week of age since {Scott (1958) has suggested that this is a critical stage in the development of puppies. The socialization stage is considered to begin at three weeks of age when puppies begin to respond to their surroundings as vision and hearing mature. Specifically, it is hypothesized that puppies which receive VEPS daily from three to four weeks of age while facing a visual pattern (solid white circle) will approach the patternmore and Show less emotional disturbance in an open-field test at four weeks of age than puppies which do not receive VEPS while facing the pattern during the training period. It is further hypothesized that, although the stimulated animals will differ from the non- stimulated animals, any kind of varying stimulation will lead to this kind of differentiation. The open fieldtest is used in this study because the attachment of the young to the mother is clearly seen when they are in surrounds that are unfamiliar to the young. The young animal characteristically remains very close to the mother in a strange environment. . In addition, studies. of monkeys (Harlow and Zimmerman, 1959) and of humans (Arsenian, 1943) have indicated that the young were much less emo- tionally disturbed by the strange room when the mother was present than when she was not present. Puppies raised in the presence of a toy cloth animal also Showed less emotional disturbance in the open field test when the cloth animal was present than when it was absent (Marr, 1960). METHOD General Puppies which were three weeks of age were‘assigned to one of three experimental groups or to two control groups. Three times a day for seven days each experimental animal received stimulation while facing a white circle which was fixed in a box... Experimental Group I received tactile stimulation, Experimental Group 11 received visual stimulation, and Experimental Group 111 received proprioceptive stimulation. Animals in the two control groups were placed in the box for the same period of time as the experimental groups but were not stimulated. The circle was in the box for one of the control groups but not for the other. The emotionality of each S was rated during each training period. After a week of training when _S_s were 28 days old, each was tested in an open field test one time with circle present and one time alone. Each test was repeated on the following day. The length of time §S spent near the circle or near the area in which it had been located, and the degree of emotional disturbance .were measured on each test trial. Subjects SS were trained and tested during July and. August 1961. The animals used in the experiment were from nine litters of pure bred dogs. The total group of SS consisted of eight Minature Poodles, ten English Cockers, seven German Shepards, and five Shetland Sheepdogs. The distribution of SS by breed, litter, and experimental group is Shown in Table I. This table shows that the 30 _S_s were divided among 5 groups. 10 11 The assignment of _S_s to the groups was random with two restrictions-- every breed was represented in every group and each group contained the same number of _S_s. . Table 1. Distribution of Breeds Within Groups . Stimulation'W/circle‘ Mulation G r oup: ‘I‘ III 117: IV V Rubbing Rocking Flashing Circle No Circle Litter Breed Light 1 German Shepard 2 1 l 2 l 2 Shetland Sheepdog 1 l 1 1 3 Shetland Sheepdog \ l 4 Poodle 1 5 Poodle 1 6 Poodle 1 1 l 7 Poodle 1 1 1 8 English Cocker 1* l l 1* 1 9 English Cocker 1 1 1 1 1 Total 6 6 6 6 6 * Reared by Poodle Mother of Litter 7. Litters 1, 2, and 3 which included the German Shepards and Shetland Sheepdogs, were reared in private homes; the other litters were reared in commercial kennels. All puppies were reared by their natural mothers with the exception of two of the English Cockers from Litter 8 which were fed and cared for by the poodle mother of litter 7. All puppies ' 12 had been handled by their owners or by the kennel managers prior to training. The handling included almost daily petting and weekly examinations. The German Shepherds and» Shetland Sheepdogs were also handled by their owners on those days that they were trained and tested. Apparatus The training box, shown in Fig. 1, was 25 in. long, 13 in. wide, and 17 in. tall. i It was constructed from brown peg board, and had no bottom or top. The wooden circle was 5%- in. in diameter and painted flat white. On training days, the box was placed on a cement floor, and the circle was fastened inside and at one end of the box as Shown in Fig. 1. The circle was located with its center 7" above the floor and equidistant from the Sides. The box shown in Fig. 2 was used for the open field tests. The box was constructed on i— in. plywood and each side was 6 ft. long and 4 ft. high. Inside the box and attached to one side, was one end of the pegboard box which had been used during training. This is also shown in Fig. 2. During the tests with circle, the white circle was fastened to the end of the peg board box, 8 in. from the floor. A 2 ft. by 2 ft. plywood platform lay on the floor, immediately in front of the pegboard. Attached to the sides of the platform were five micro- switches which were wired to an electric clock. . The switches were activated by _S_S placing two or more paws on one exposed portion. of the platform, so that the clock was recording when S was on the platform. The clock which was calibrated in tenths of a second could record up to 600 seconds. - It was located outside of the test box. Procedure Each _S_ was given three training trials a day beginning when it was 21 days of age and continuing through the 27th day of age. For each 13 NH. ,4 ~flw‘ v —\.~VV ~\ ~\¥ ~ ~ ~ —\._. A A x . a. . a v \ _ \. .__. ,\ ,\. FIGURE I The Training Box ‘Vm. \VM K.“WWM\ \ .-W\.. FIGi'RE 2.. The Testing Box 14 trial S was removed from its whelping box and placedin the training box for three minutes. During this trial, S was held in the box, facing one end, .with E's hand under its stomach and chest, so that _S_‘S rear paws touched the floor. The _S_S in Groups I-IV faced the white circle. The circle was removed from the box when _S_S in Group V (no stimulation, no circle) were in the box. . The groups were differ- entiated in the following ways. Groupl: The _S_S were stroked 1-3 times per sec. on the back of the head, neck, and back with a moderate pressure of E's hand in a manner described by Bovard as Type I handling (1958). One stroke consisted of a cephal-caudual movement of the hand followed by a caudual-cephal movement of the hand. (Rubbing w/circle) Group II: A light from a flashlight was flashed on the circle 1-3 times per sec. . One flash of the light consisted of turning the light on by depressing a button on the flashlight and turning it off by releasing the button. (Flashing light w/circle) . Group III: The _S_S were rocked back and forth 1-3 times per sec. with their rear paws touching the floor. As _S_ was rocked, its head moved forward approximately 1 in. and then back 1 in. (Rocking w/ circle) Group IV: The _S_S were held in the box in the same manner as animals in Groups I-III but the stimulation was not varied. (No stimu- lation with circle) Group V: The _S_S were treated in the same manner as Group IV except that the circle was not present. (No stimulation and no circle) Upon the completion of a trial, _S_ was returned to the whelping box. . Each S received two trials in the morning and one in the late afternoon. The morning trials were separated by at least 30 min. The sound emitted by each _S_ was rated during each minute of a training trial. The ratings were made as follows, O-no sound, 15 l-whining, 2-.yelping, and 3-continuous yelping and howling. The ratings for the three minutes were summed which permitted at maximum rating of 9'for each training trial. An average rating of the 21 training trials was obtained for each animal. The 12 had made similar ratings in a previous study. (Marr, 1960) All animals were tested on days 27 and 28. On each test day, each S was placed in the opening field box for 3%- min. once with the circle present and once with it absent. In all tests the SS were placed 6 in. in front of the platform facing the pegboard. Half of the animals in each group were first tested with the circle present and then retested alone. The other half of each group received the opposite order. . On day 29 the orders were reversed. On each day, the interval between tests for each _S_ was approximately 30 min. Three criterion measures were obtained from each test trial. The length of time each S was on the platform was automatically recorded by the clock connected to the microswitches attached to the platform. The time on platform when the circle was present was assumed to be an estimate of the attachment of the puppies to the circle when compared to the time on platform when the circle was not present. Two measures of emotional disturbance were also obtained from each test. The E recorded the length of time the animal made sounds (whining, helping, or howling) with a stop watch. Ratings of the type of sound were also made in the same manner as that described for training trials. It was assumed that the longer the animal vocalized and the higher the rating, the more emotionally disturbed it was. Following the last test trial on the second day of testing, an attempt was made to increase the fear of the SS while in the open field box. . Each S was placed in the box for one min. and a small horn from a child's tricycle was honked outside the box while a child's pull-toy 16 (Donald Duck on wheels) was pulled back and forth inside the box by E who stood outside the box. The toy was at the far end of the box from the circle. . Platform time was recorded and observations of _S_S behavior were noted. Only nine animals were given this fear test because the owners of the dogs expressed dissatisfaction with the test. RESULTS The number of animals from the various breeds differed among the training groups. . In order to investigate breed differences in emo- tionality and to control these differences in comparing training groups, one animal from each breed was randomly drawn from each group to obtain equal breed representation across groups. . In. those groups where only one of a breed was present, that animal was used as the representa- tive from the breed. This procedure (resulted in 4 animals in each of the five groups for training comparisons. However, all six animals within each group were used in the analysis of the 1_:_e_s£ results. Table II Shows the means and standard deviations of the sounds _S_S made during training trials. . The ratings from the 21 training trials were averaged for each animal, and a mean of these mean ratings was obtained for each group. The ratings of the type of sounds made during training varied considerably between groups. . In the visually stimulated group, Group II, and in the no-stimulated groups, Groups IV and V, there was considerable whining, yelping, and howling while the animals were being held in the box. The _S_S in Groups I and II (rubbing and rocking, respectively) made very little noise during training trials. To test the differences among training groups, groups were ranked within each breed, and the degree of agreement among the breeds was estimated using Kendall's coefficient of concordance (Edwards, 1954, pp. 402-412). The W of . 734 with four sets of five ranks is significant at the . 01 level. This shows that the differences amongtraining groups were consistent for all breeds. 17 18 Table II. Mean Rating of Type of Sound During Training Type of Training Stimulation W/Circle . No Stimulation Rubbing Rocking Flashing Light W/Circle No Circle Mean .11 .68 2.44 2.71 ' 1.74 S.D. .14 1.36 1.61 1.22 1.31 N 4 4 4 4 4 Table III shows the means and standarddeviations of the sound ratings for each breed. The breeds were ranked within each training group so that the degree of agreement among groups could be estimated using Kendall's coefficient of concordance. The W of .633 with.5 sets of 4 ranks is significant at the . 01 level. Table III. Mean Rating of Type of Sound During Training for Breeds w German English Cocker Shetland Minature Shepard Spaniel Sheepdog Poodle Mean 2.52 ' 1.12 1.84 .66 S.D. 1.52 1.08 2.00 .64 N 5 5 5 5 The three dependent variables measured in the open-field tests were time on the platform, duration of sound, and rating of type of sound. The time on the platform was used as a measure of the degree of attach- ment to the circle. The duration of sound and sound rating were used as estimates of the emotional disturbance of the animals while they were in the open-field box. 19 The three variables were analyzed separately. The measures used in each. analysis were the mean of the measures obtained from the two tests with circle and the mean of the measures obtained from the two tests with no circle. 3 The analysis of variance performed on each variable was a three factor designwith a fourth, factor nested within one of the other factors in a manner suggested by Allen (1961). This alloweda comparison of the stimulated groups and the non-stimulated groups, a. comparison of the groups nested within each of these two levels of stimulation, .a comparison of the tests with circle and tests without circle, and a comparison of the two different orders in which the tests were presented. The orders of presentation were compared because having been in the open-field box. first with the circle might have affected the animal's subsequent behavior when the circle was not present. Table IV shows the mean platform times for each training group in each test. The animals stimulated during training trials Spent more time on the platform when the circle was present than when the circle was absent. Table IV. . Mean Time on Platform During Tests Type of Training Stimulation W/Circle No Stimulation Test Rocking Rubbing Flashing Light W/Circle No Circle Alone Mean 42. 93 29.17 40. 03 29.18 48. 85 .S.D. 10.54 29.15 25.28 21.66 49.63 W/Circle Mean 56.65 57.17 69.17 29.42 47. 97 S.D. 28.97 8.31 33.18 23.43 46.65 20 Since the variantes of the distributions of platform times were found to be significantly different, a more conservative level of signifi- cance'was chosen for the analysis of variance test as suggested by Lindquist (Lindquist, 1953, pp. 78-90). 4 For this analysis 3 of less than . 03 was selected. .Table V shows the results of the analysis of variance of platform time. It can be seen from Table V that the stimulation groups did not differ significantly from the no- stimulation groups when the platform times in both kinds of test conditions were considered together. However, the interaction between level of stimulation and type of test was signifi- cant (B' < . 03). This interaction indicates that there was a significant difference in time on platform between tests with circle and tests with no circle for the stimulated animals but not for the non- stimulated animals. The difference between the two types of tests for the stimu- lated animals was large enough to make the difference between tests significant for all groups. It is apparent from an examination of the means in Table IV that the results from the two tests are almost identical for each of the non- stimulated groups. However, time on platform is substantially larger in tests with circle than in tests alone for each of the three stimulated groups. Individual comparisons by t tests between circle and no circle tests for each stimulated group yielded t's that were Significant at the . 05 level. The mean durations of sound for each group on the tests are shown in Table VI. All groups spent less time. crying during tests with circle than in tests alone. Table VII presents the summary of the analysis of the measure of duration of sound. From Table VII it can be seen that orders of testing were Signifi- cantly different (p. < . 05) and the interaction of order withlevel of stimulation approached Significance (p < . 10). Animals in the 21 Table V. -Analysis of Variance Mean Time on Platform During Tests w m Source df MS FM! Training (Stimulation vs- No stimulation 1 1537. 19 Grps. wn Stimulation 2 395. 46 Grps. wnNo Stimulation 1 2190. 77 Order of Tests 1 76.16 Order by Levels of Stimulation 1 2983. 10 Order by Grps. wn Stimulation 2 159. 10 Order by Grps. wn No Stimulation 1 2202. 25 Error (Subjects in Groups by Order) 20 1581. 54 Tests (Circle vs. No Circle Present) 1 2956. 88 7. 92* Tests by Levels of Stimulation 1 2063. 57 5. 53* Tests by Grps. wn Stimulation 2 221.49 Tests by Grps. wn No Stimulation 1 1. 84 Tests by Order 1 29.44 Tests by Order by Grps. 4 397. 67 Error (Tests by Subjects in Groups by Order) 20 373. 33 Total 59 *p < . 03 >' 9; s F ratios of less than 2. 00 are not shown. 22 Table VI. Mean Sound Duration in Seconds During Tests Type of T raining Stimulation W/Circle - No Stimulation 7 Test Rocking Rubbing Flashing Light W/Circle No Circle Alone Mean 110.67 94.17 104.17 . 126.33 96.17 S.D. 47.44 32.16 38.15 46.12 48.73 W/Circle Mean 83.17 86.50 77.00 103.83 86.83 S.D. 34.77 49.55 36. 74 51.99 47.66 non-stimulated groups and the animals that had been rocked during training cried longer in all open-field tests if they were tested alone first than animals which had been tested with circle present first. The differences in duration of sound for all groups between the two types of tests resulted in an F_ of 7. 57 (p.< . 05). The mean ratings of sounds for the groups during the tests are presented in Table VIII. This measure of emotional disturbance was also lower in tests with circle than in tests alone. The summary of the analysis of variance of this measure is shown in Table IX. The 19; test of the difference between types of tests for all groups was highly significant (p > . 01). The interaction between Tests and Levels of Stimulation was not Significant (p < . 20). 23 Table VII. Analysis of Variance Mean Duration of Sound in Seconds In Tests Source df MS F** Training (Stimulation vs. No Stimulation) l 1642 Grps. wn Stimulation 2 167 Grps. wn No Stimulation 1 3337 Order of Tests 1 17579 7.123%< Order by Levels of Stimulation 1 10573 4. 28 Order by Grps. wn Stimulation 2 1720 Order by Grps. wn No Stimulation 1 196 Error (Subjects in Groups by Order) 20 2470 Tests (Circle vs. No Circle Present) 1 5320 7. 57* Tests by Levels of Stimulation 1 86 Tests by Grps. wnStimulation 2 387 Tests by Grps. wn No Stimulation 260 Tests by Order 1 874 Tests by Order by Grps. 4 194 Error (Tests by Subjects in Groups by Order) 20 702 Total 59 * p < . 05 >3 :4: F ratios of less than 2. 00 are not shown. 24 Table VIII. 7 Mean Rating of Type of Sound During Tests Type of T raining Stimulation W/Circle No Stimulation Test Rocking Rubbing Flashing Light W/Circle No Circle Alone Mean 3.58 3.21 2.94 3.79 3.29 S.D. 2.15 1.25 1.15 1.84 1.95 W/Circle Mean 2.17 3.12 2.33 3.58 3.08 S.D. 1.66 1.86 1.62 2.28 1.91 25 Table IX. Analysis of Variance Rating onype of Sound in Tests m Source df MS F* Training (Stimulation vs. No Stimulation) 1 4. 3319 Grps. wn Stimulation 2 . 8454 Grps. wn No Stimulation 1 1. 6276 Order of Tests 1 13. 8913 2. 28 Order by Levels of Stimulation 1 6. 5367 Order by Grps. wn Stimulation 2 4. 2286 Order by Grps. wn No Stimulation 1 1. 6302 Error (Subjects in Groups by Order) 20 6. 1022 Tests (Circle vs. No Circle Present) 1 3. 5673 8. 66** Tests by Levels of Stimulation 1 . 9120 2. 21 Tests by Grps. wn Stimulation 2 l. 2778 3. 10 Tests by Grps. wn No Stimulation 1 . 0026 Tests by Order 1 . 9703 2. 35 Tests by Order by Grps. 4 . 0066 Error (Tests by Subjects in Groups by Order) 20 .4120 Total 59 * F ratios less than 2. 00 are not shown. ::< * p>.01 OBSERVATIONS A number of other observations were made of the behavior of the young dogs during the training and testing periods of this study and during the pilot study. This data is relevant to the topic of this study andalso suggests hypotheses for further studies. . Although animals in all treatment groups remained very passive while being held-in the training box, the _S_S in the different groups appeared to differ in terms of their head movements. Whereas_S_s that were being rockedand rubbed held their heads in one position facing the circle, .Ss in both control groups occassionally moved their heads. A control animal would first face the circle or the area where circle was located (Group V), change its head position so that it was facing the adjacentcorner or wall of the box, change again so that it was facing the circle, and then change again so that it was facing the other adjacent corner or wall of the box. . These head movements occurred as frequently as five times a minute. Animals undergoing the flashing light treatment, however, would turn their heads back and forth as much as 15 times a minute as if they were attempting to avoid looking at the light on the circle. . Activity in the training box increased onthe 27th day to the degree that it was difficult to hold a puppy in‘position with one hand. . If it had been intended to train the animals for‘more than a week, special appara- tus would have been necessary to hold the puppies in one position in thetraining box. When the animals were 28 days of age, avoidance responses were first noted. The mongrel puppies used in the pilot study were first seen to avoid the E while housed in a nesting box that was located outdoors. These puppies had never showed any avoidance of _E_2 in days 21 to 27, and sometimes would approach E's arm when he reached into the box. 26 27 On days 28 and 29, all puppies in this litter ran to the nesting box and/or crouched down in the box when._E_2 approached. At the conclusion of the test trials on days 28 and 29, as _E_Z approached S in the open-field test to return it to the nesting box, the animal would run to the far end of the room, crouch,andlyelp until picked up. It was at first believed that the‘pups had had some experience between the last training trial on day 27 and the first. test trial on day 28 which had resulted in the pups' learning to avoid an approaching human figure. . However, the same type of response was observed in this experiment. On days 28 and 29, many of the _S_S ran away from E, crouched, and yelped as 12 entered the open field box at the conclusion of a test trial. . Neither the puppies .used- in the pilot study or those used in this experiment made any of these responses to E once they had been picked up and set down again. None of the SS used in this experiment were observed to make an avoidance response to _E_3 on days 28 and 29 when they were in their whelping cages. It appeared that the three and a half minutes in the unfamiliar surrounds of the open field test had lowered the "fear" threshold of the pups. As a result, the large moving figure approaching _S_ elicited fear responses in the open field test but not in the vicinity of the whelping nest. The avoidance responses were not observed in all breeds. The pilot study animals were part GermanShepard and part Collie. . The avoidance responses were noted in this study to be present in the German Shepards, English Cocker Spaniels, and Shetland Sheepdogs. None of the eight Minature Poodle puppies were observed to make these responses. They either made no response to the approach of _E_ or approachedhim, wagging their tails. It is doubtful that these breed differences were due to differences in the type of rearing of the puppies. . Five of the English Cocker Spaniels were reared in the same kennel as the poodles. 28 In other species, similar observations have been made of the sudden onset of fear responses. This has been noted by Hess (1959) with ducks and Ratner and Thompson (1960) with domestic fowl. The behavioral differences observed among the breeds suggest different rates of maturation among the different breeds. The Poodles that were used in this study opened their eyes later and walked later than the other breeds. The observation that avoidance responses were not present in Poodle puppies by the 29th day but were present in the other breeds suggests that fear also appears later in Poodles. Other differences were noted among the breeds during test trials. Although all puppies smelled the cement floor when first placed in the test box, the German Shepards, Shetland Sheepdogs, and Poodles ap- peared to orient themselves primarily in terms of vision. They would explore the box visually, take a few steps, and again visually explore the box. The English Cockers appeared to depend more upon smell than the other breeds. At the beginning of a test trial, an English . Cocker puppy would characteristically turn around and around with its nose pointing towards the floor, sniffing continuously. It would walk a few steps and then turn around again as it sniffed the floor. This type of behavior would continue for approximately 30 sec. and then it would begin to visually explore its surrounds. Visual exploration appeared to be accompanied by olfactory exploration throughout the test trials. After both open-field tests were completed on the 29th day, nine of the animals were again placed in the test box, one at a time, with the circle present. This group consisted of seven German Shepards and two Shetland Sheepdogs. An attempt was made to increase the SS fear by honking a bicycle horn and moving a mobile, child's toy back and forth at one end of thebox (as described in the procedure section of 29 this paper). . Eight of the nine animals began to tremble at the onset of the noise but then approached the toy, trembling all the way. One GermanShepard puppy, Group V, remained in one spot, crouching and watching the toy. . None of the animals cried during this test. DISCUSSION The results of the analyses of the-ratings .made during training indicated that the type of treatment during training significantly affected the kinds of sounds the dogs made during training trials. Rubbing and rocking in the presence of the circle resulted in less intense sounds than exposure to the flashing light or either of the control conditions. The sound ratings are considered to be an index of emotionality (Scott and m): 1950; Fuller, 1955; Marr, 1960). This. finding suggests that the emotional disturbance of three to four week old puppies is dependent upon the kind and intensity of stimulation. Hebb (1955) has suggested that a mild degree of stimulation is pleasing but a high degree of the same kind of stimulation can bring about emotional disturbance. The stimulation supplied to the puppies by the flashing light may have been more intense subjectively than the rubbing and rocking. However, this interpretation does not explain the behavior of the control groups whose sound ratings were similar to those with the flashing light. The results of theratings of sound during trainingalso indicate that there are differences in vocalization as a function of breed. Such differences have been previously reported by Fuller (1955).. . The German Shepards were rated higher than the other breeds and the Minature Poodles were rated lower than the other breeds in every training group. Although the difference in ranking of emotionality among the breeds suggests inherent differences, other factors that were confounded with breeds may have contributed to the differences. For example, the breeds used in this study were reared under different kennel conditions by different bitches at different times during the summer. . In addition, 30 31 the German Shepherds and English Cockers weighed 2%- to 4 lbs while the Shelties and. Poodles weighed-i— to 1%- lbs. If intensity of a type of sound (e. g. , yelping) is roughly proportional to the Size of the animal, the different ratings could be partially a function of these differences. , Although it is relatively easy to discriminate between the whining and yelping in the same breed, it is much more difficult to discriminate between the whining of a German Shepard and the yelping of a Minature Poodle because of the similarity in intensity of sound. Test Results The different treatments in training also resulted in differences among the groups in test measurements. The interaction between types of tests and levels of stimulation was significant as measured by time on platform. . That is, the groups that received varying stimulation (VEPS) in the presence of the circle during training spent more time on the platform during tests with circle than during tests without the circle. However, the groups that were not given varying stimulation on training trials, Groups IV and V, did not spend more time on the platform when tested with circle than when tested alone. ' . Furthermore, the interaction of tests with stimulated groups and the interaction of tests with non- stimulated groups were not signifi- cant in the analysis of variance of platform time. This means that the groups which had received different kinds of varying stimulation in . training did not differ on platform times in either test condition. . In addition the group which had faced the circle during the training trials but which had not received varying stimulation during these periods, Group IV, did not differ in platform time from the group that had neither been stimulated nor had the circle in the. training box, Group V. The time on platform was considered to be a measure of the degree of attachment of a puppy to the circle. It was hypothesized that a visual 32 pattern which has been presented to puppies in conjunction withvarying exteroceptive or proprioceptive stimulationwill come to elicit approach responses from the animals. . The SS that were given varying tactile stimulation.(rubbing), varying kinesthetic stimulation (rocking), and varying visual stimulation (flashing light) from three to four weeks of age .while facing a visual pattern. showed a stronger attachment to the pattern at four weeks of age than _S_S which were not given varying stimu- lation but which had faced the pattern during training. . Also the latter group of _S_S did not show any more attachment to the visual pattern than puppies which had never seen the pattern prior to testing. Figures 3 and 4 Show puppies that had been stimulated during trainingwhich are approaching the visual pattern during tests. Other interpretations are possible as explanations of the difference in platform time between tests. with circle and tests alone for the stimu- lated groups. Generalization of the approach response from mother to white circle is one interpretation. However, the dissimilarity in size, shape, and odor between the circle and the mothers makes this inter- pretation doubtful. Although the circle was white, only two of the eight bitches were white. These two bitches had four'pups in the control groups and four in the experimental groups. Another explanation would involve the postulation of anxiety re- duction as the reinforcement for the behavior observed in this study. . Moltz (1960), has proposed such, a theory of anxiety reduction to explain imprinting phenomena. The application of this theory to the results of this study would require that the circle be associated with the reduction of anxiety for the groups that received varying stimulation but‘not for those groups which. had not received varying stimulation during training. The sound ratings taken during training were considered to be a-measure of the emotional disturbance of the animals and could also be considered a v» ~¥_ “Wm-x «fix‘mm WM.X~ 33 Ml. ix 19..- \‘~,.\ \\‘\ \.\\\\\._\\\\‘\h_, FIGURE 3 Test behavior of a puppy that was 'rocked while facing the circle during training. w ‘FIGURE 4 Test behavior of a puppy that was rubbed while facinq the circle during training. 34 a rough measure of the anxiety of the animals during the training sessions. . Since there were significant differences in emotional disturbance among the stimulation groups and since these groups did not differ in the measure of attachment to the circle during testing, it would be difficult for an anxiety reduction theory to explainthese findings. Animals that were rubbed or rocked made very little sound during training trials. . Animals that were subjected to the flashing light were rated high in emotional disturbance during training trials. The results of this study indicate that the association of VEPS with a visual pattern results in the visual pattern eliciting approach responses in the dog. In order to Show that this association is part of the process by which the dog learns its species characteristics, it is necessary to demonstrate that the approach responses elicited by the visual pattern would also be elicited by visual patterns with Similar stimulus character- istics. . If generalization from one pattern to a Similar pattern could be demonstrated, the process by which the dog develops its preference for itsown Species could be better understood. Imprinting has been described as a process or mechanism by means of which certain visual stimuli acquire with extreme rapidity the capacity to elicit responses that an animal ordinarily directs only towards members of its own species (Thorpe, 1956). This study suggests that the associ- ation of VEPS with the visual stimuli is part of this process in dogs. The puppy ordinarily does not learn to make these responses to other animals because the only objects associatedwith varying stimulation in the first few weeks of life is the bitch and the litter mates; In the natural state, the bitch prevents any other mammal from entering the den and the pups do not normally venture outside the den until four weeks of age. Scott (1958) has reported a critical stage in puppies beginning at three weeks of age and lasting until 10 weeks of age. The puppies are beginning to respond to their surrounds at three weeks as vision and 35 hearing mature, and are weaned at approximately ten weeks by the bitch. This socialization stage may be critical for the learning of the Species characteristics. Since this study gives evidence that dogs can learn to make approach responses to an object that has been associated with varying stimulation from 3 to 4 weeks, an investigation at other ages is warranted. Some of the work on imprinting which has been reported by Thorpe (1956), Hess (1959), and Moltz (1960) suggests that there is a brief stage of development during which imprinting is most likely to take place, although the period is not rigid nor absolute as Lorenz and his co-workers had suggested. Hebb (1958) has suggested that varying stimulation is a- "universal reinforcing agent, " and this study gives some evidence that an object associated with varying stimulation acquires special properties. While this experiment was being carried out, McV. Hunt and Quay (1961) also reported evidence that varying stimulation has reinforcement value, and that any receptor input is innately reinforcing. Receptor input does seem to have reinforcement value but only if the input changes. The physical characteristics of the training box, the circle, and E's hand in this study were probably impinging upon the receptors of the puppies who were in the box. with the circle but not given varying stimu~ lation. Yet, they showed no evidencerof having learned the character- istics of the circle as measured by time on platform during tests. The animals that were subjected to different kinds of stimulation that varied in intensity and rate did Show evidence of learning. . A constant input may satiate receptors, neurons, or synapses while varying input does not. Thus, it is suggested that as a result of the neurological character- istics of the organism, varying stimulation has more reinforcement value than unchanging stimulation. _ It was also hypothesized in this study that _S_S which had received varying stimulation during training while facing the circle would be 36 less emotionally disturbed in those tests .with circle present than in the tests in which the circle was not present, while there would be no difference in emotional disturbance for those animals which were not given varying stimulation during training. The two measures of emotionaldisturbance were duration and ratings of sound during test trials. The results of analyses of both measures Show that all groups were less emotionally disturbed in those tests with circle than in tests alone (Tables VI and VIII). Furthermore, the interaction of types of tests and levels of stimulation was not significant for either measure of emotional disturbance (Tables VII and IX). This means that the decrease in emotional disturbance from tests alone to tests with circle for stimulated animals was not greater than, the decrease for control animals. . Observations of the puppies in the open-field box with circle showed that a puppy would characteristically stop crying while approaching the circle and then resume crying after it reached the circle regardless of whether it left the platform or sat down facing the circle. These short lapses in sound when approaching the circle could account for the general drop in rating and duration of sound for animals in all groups from tests alone to tests with. circle. One of the reasons for the resumption of crying by the stimulated animals after they reached the circle may have been their lack of familiarity with the odor of the circle. That is, an object that was familiar in appearance was not familiar in odor. The encounter of the animalswith the circle during the open-field test were the first time that they had been allowed to approach and smell the circle. . In the traine- ing box they were held about 1%- ft. from the circle during stimulation. In a previous study by Marr (1960), puppies which had been reared by their natural mother in the presence of a toy cloth animal Showed less emotional disturbance in an open-field test if the toy cloth animal was 37 present than they did in the test alone. Those pups had slept next to the'toy, climbed on it, chewed on it, etc. for the first four week of life. They had also been-stimulated by the mother in the presence of the toy. The pups in this experiment were stimulated in the presence of the circle but not allowed any other associationwith it. The order of tests was significant (p < . 05) in the analysis of variance in terms of the variable of durationof sound (Table VII). The animals that were tested first with circle cried for shorter periods of time in both tests than the animals that were tested alone first. This difference is difficult to interpret because the order of tests was con- founded with breeds, sex, and the number of animals that had preceded any other animal into the test box. . Since there were emotional disturb- ance differences among the breeds during training, the disproportionate representation of a quiet breed in one order could cause the order of tests to appear important. A disproportionate number of animals of one sex in an order could also cause the difference in order of tests if there is an innate difference in vocalization between sexes. Although the total sample included 15 males and 15 females, nine of the females were given the test with circle first. The number of animals that had preceded the puppy being tested into the test box could also affect the animal's behavior during a test. . Although the test box floor was cleaned after each test, the odor of feces and urine was probably still present. It was noted that the animals would smell the areas where other pups had urinated or defecated. When a puppy was smelling the floor, it usually was not crying. This lapse. in sound could also have contributed to the difference attributed to order. . In general, the platform time was a' more reliable measure than sound rating or sound duration. The platform time was automatically registered and independent of 133 while the other two measures were recorded by E and subject to experimental bias. 38 The experimental procedure made it more difficult to. demonstrate the predicted behavior in this experiment than was expected. The circle was attached to one end of the training box during training and that end of the training box was later attached to the inside of the test box for both circle tests and tests without circle. ~ If any of the animals learned the stimulus characteristics of the pegboard box during training, the end of the box that was attached to the inside of the open field box may have elicited approach responses in tests with no circle present. .As a result, the strength of the attachment of the stimulated animals to the circle may have been underestimated because of the presence of the pegboard in both tests conditions. Observations made during the tests also suggest that the duration of the tests should have been longer. It was noted that the experimental animals were usually on the platform at the conclusion of the tests with circle, whereas the control animals were usually in one of the corners of the open-field box at the conclusion of both tests. Longer test periods might have resulted in alarger difference in platform time betweenthe experimental and control groups. It also would have been desirable to have the animals reared under more similar kennel conditions. For example, differences in whelping nests or feeding times of the bitches could have created error variance that would not be present if all animals had been reared in the laboratory. The effects of the association of varying stimulation with the circle on dogs might also have been better demonstrated if the pups could» have been reared without their mothers. During the week that they learned the stimulus characteristics of the circle they were also learning the stimulus characteristics of their mother who was also stimulating them. As a result, the puppies may have left the platform after they had A encountered the circle in the open-field test in search of the mother. 39 Two other variables that should be investigated are rate and intensity of stimulation. . The stimulation given to the animals in this experiment varied at a rate between 1 to 3 per second. VA repetition of this experiment using other rates of stimulation might reveal maximum and minimum rates at which stimulation is effective. The intensity of stimulation should also be further investigated. Hebb (1955) has suggested that there may be an optimal level of arousal for effective behavior. The variation in intensity of stimulation that was given to the experimental animals in this study may have been restricted to a range of intensity that results in an optimal level of arousal in the three week old dog. A larger range of intensity might have had different effects on the behavior of the animals. Since the animals that were subjected to the flashing light cried much more in training than the animals that were rocked or rubbed, the light may have been muchlmore intense subjectively than the tactile pressure. There may be a particular range of intensity for each modality which is optimal for learning . SUMMARY It was the purpose of this paper to examine the effect of varying exteroceptive and proprioceptive stimulation in the development of the dog's attachment to its own species. A review of the studies on im- printing and stimulus deprivation, and observations of the mother's activities when she is caring for the young suggested that varying extero- ceptive and proprioceptive stimulation is important in the development of the preference by the young animals for the species. It was hypothesized that as a result of the varying exteroceptive and proprioceptive stimulation applied by the mother to the infant animal, the infant associated the varying stimulation with the stimulus-character- istics of the mother. As a result of this association, the mother comes to elicit approach responses from the young animal. These approach responses are later generalized to other animals having similar stimulus- characteristics, the Species. This experiment attempted to demonstrate only thefirst step of the development of the preference for the species in the dog. That is, a visual pattern which had been presented to puppies in conjunction with varying exteroceptive or proprioceptive stimulation was tested for its properties to elicit approach responses from the ani- -malS. ~ Itwas also expected that the puppies would be less emotionally disturbed in. an open-field test when the visual pattern was present than whenit was not present. Thirty dogs from four breeds and eight litters were assigned to one of three experimental groups or two control groups. . From 21 to 27 days of age, each animal was placed in a training box for three minutes three times a day. . The SS in the experimental groups were given varying stimulation while facing a white circle in the training box. The _S_S in Group I were rubbed (tactile stimulation), _S_S in Group II were rocked 40 41 (proprioceptive stimulation), and _S_S in Group III were subjected to a light flashing on the circle (visual stimulation). The _S_S in the control groups (Groups IV and V) were not given varying stimulation in the training box. Group IV animals were held while facing the circle; Group V was held but the circle was not present in the training box. The emotional disturbance of the puppies was estimated by rating the sounds made during the training trials. When the dogs were 28 days of age, each was tested in an open- field box in the presence of the circle and alone. . Half of the animals in each group were first tested with the circle present and then retested alone. The rest of the animals first received the test when alone. On day 29, the orders of testing were reversed. Three criterion measures were obtained from each test. The length of time each _S_ was on a 2 ft. by 2 ft. platform, located in front of the circle, was automatically recorded. The platform time obtained from tests with circle when compared to the platform time obtained from tests when the circle was not present was considered a measure of the attachment of the puppies to the circle. Two measures of emotional disturbance were also obtained from each test. The duration of sound made by S was recorded with a stop watch and the type of sound was rated by E. It was assumed that the longer the animal vocalized and the more intense the vocalization, the more emotionally disturbed it was. The groups were found to be Significantly different in emotional disturbance during training. The control animals and those whichrhad received varying visual stimulation were rated high in emotional dis- turbance while the animals that had been rocked or rubbed were rated low in emotional disturbance during training. The breeds were also found to differ in emotional disturbance in training. The German Shepards were rated high and the Minature Poodles 42 were rated low in emotional disturbance. . Although Similar, breed dif- ferences have been reported previously, these findings might be attributed to the intensity of sound which may be a. function of the size of the animal. The three dependent variables which were measured in the open- field tests were analyzed separately. An analysis of variance design was used which allowed comparisons of the training groups, tests, and order of tests. The results of the platform time comparisons indicated that the experimental animals were more attached to the circle than the control animals. The hypothesis that a visual pattern which has been presented to young dogs in conjunction with varying exteroceptive or proprioceptive stimulation will come to elicit approach responses in the puppies was supported. The results of the analysis of variance of the emotional disturbance measures did not support the hypothesis that the association of varying stimulation with the visual pattern would result in the pups being less emotionally disturbed in the test with circle present than in the test with circle not present. It was noted that the experimental animals cried during the tests even while standing. on the platform and facing the circle. . The suggestion was made that the puppies may have been dis- turbed by the fact that an object which was familiar in appearance did not have a familiar odor. Their contact with the circle in the test box was the first time that they had been allowed to approach the circle. The duration of sound in both tests was significantly less for all groups when the animals were given the testwith. circle first than when they were given the tests without circle first. . It was difficult to inter- pret this finding because breeds, sexes, and the number of animals that had preceded any animal into the test box was confoundedewith. the order of tests. Other variables that need further study are age of the pup when trained, and rate and intensity of stimulation. Specific rates and 43 ranges of intensity of stimulation at a specific age of the animals may be most effective for the learning of the stimulus-characteristics of the species. REFERENCES Allen, T. -Personal Communication, 1961. Arsenian, J- M. Young children in an insecure situation.. J. Abnorm. Soc- Psychol., 1943, 3___8, 225-249. Beach, F. A. and Jaynes, J. Effects of early experience-upon the behavior of animals. PsycholL Bull., 1954,31, 239-263. Blauvelt, H. Dynamics of the Mother-Newborn Relationship. in Goats. InB. Schaffner (Ed). Group Processes. New York: Iosech Macy Jr. Found. 1955. Bovard, E. W. The effects of early handling on viability of the albino rat. -Psychol. Rev., 1958, 62, 257-271. Bowlby, J. Child care and the growth of love. . In Haimowitz and Haimowitz (Eds.) Human Development. New York: Thomas Crowell, 1960. Bowlby, J. The nature of the child's tie to his mother. . Intern J. Psycho-Analysis, 1958, _3_9, 350-373. Brody, Sylvia, Patterns of Motherin& New York: Hallmark, 1956. Cruikshank, Ruth M. . Animal Infancy. Manual of Child Psychology. In Carmichael (Ed.),. New York: Wiley, 1954. Edwards, A- L. Statistical Methods for the Behavioral Sciences, New York: Rinehart, 1954. Gibson and Walk, R- O. The effect of prolonged exposures to visually presented patterns of learning to discriminate them. .. J- Comp. Physiol. Psychol., 1956, :l_9, 239-242. Goy, R. W. and Young, W. C. Somatic basis of sexual behavior patterns in guinea pigs. PsychosomaticMed., 1957, 52, 144-151. Gray, P. H. . Theory and evidence of imprinting in human infants. . J. Psychol., 1958, 46, 155-166. 44 45 Harlow, H.. F. and Zimme rmann, R. R. . Orphaned baby monkeys develop a strong and persistent attachment to inanimate surrogate mothers. . Science. 1959, 130, 421-432. Hebb, D. O. The motivating effects of exteroceptive stimulation. .Amer- Psychologist, 1958, L3, 109-113. 7 Hess, E. H. Imprinting, Science, 1959, 130, 133-141. Hess, E. H. . Two conditions limiting critical age for imprinting. J.. Comp. Physiol. Psychol., 1959, _S_, 515-518. Igel, G. J. and Calvin, A. D. The development of affectional responses in infant dogs. .J. Comp. Physiol. Psychol. 1960, 23, 302-305. James, W. T. Observations on behavior of new-born puppies: I method of measurement and types of behavior involved. J. Genet- Psychol. , 1952, _82, 65-90. Lindquist, E- F. Design and Analysis of Experiments, Boston: Houghton-Mifflin, 1953. . Marr, J. N. The influence of soft body-contact on the development of canine affection. Paper read at Michigan Academy of Science, Arts, and Letters, 1960. . McV. Hunt, J. and Quay, H. C. - Early vibratory experience and the question of innate reinforcement value of vibration and other stimuli: a limitation on the discripancy (Burnt Soup) principle in motivation, Psych. Review., 1961, 6_8_, 149-156. Melzack, R. The genesis of emotional behavior: an experimental study of the dog. J. Comp. Physiol. Psychol. 1954, :1_7, 166-172. Melzack, R. and Thompson, W. R. . Effects of early experience on social behavior. Canad. J. Psychol. 1956, 12, 82-90. Moltz, H. Imprinting, empirical basis and theoretical Significance. chhol. Bull., 1960, E, 291-314. Scott, J. P. Animal Behavior, Chicago: University of Chicago, 1958. 46 Scott, J.. P. and Marston,. Mary-Vesta. Critical periods affecting the development of normal and mal-adjustive social behavior of puppies. J. Genet. Psychol., 1950, 11, 25-60. Thompson, W. andHeron, W. The effects of restricting early experi- ence on the problem-solving capacity of dogs. Canad. J. Psychol. , 1954, _8, 17-31. Thorpe, W. H. . Learning and Instinct in Animals, Cambridge: Harvard, 1956. Valenstein, E- 5., Ross, W. and Young, W. C. . Experimental and genetic factors in the organization of sexual behavior in male guinea pigs.. J. Comp. Physiol. Psychol., 1955, g, 397-403. Yerkes, R- M. and Tomilin, M. J. .Mother-Infant relationship in chimpanzees. J. comp. Physiol. Psychol. 1935, 2__0, 321-359. APPENDICES 47 48 L0 H M s N m a. use-um 36.- OM .12 m m a. m N 32a Hebe.- m 2 m 2 2 2. Sauce annmem _ a m a 2 2 a 2. e380 germ-am m m 2 2 2 3869 a m 2 a m baboon e a m baboon. m a a . sauce“ I. n m moenueam senseam m a m a E a woenuuam censusm N e a m .2. 2 a ,. m a eunnenm 8880 a z - £16.30 62 126:0 1 messes Ewan manners museum 1 - sebum nun-3 . > 9.80 3 a = macho ammo. . Gowudfidfiflm oz EQQW sir- 26.30?» 832583 $5.8wa x5153 _. 4 XHQZManH< eon-333.99 Jam...» 49 APPENDIX B Tests with Circle Not Present . 706 Tests with Circle Present .. 855 Correlations between Sound Rating and Sound Duration in Tests