~ 1" AN INVESTIGATlON OF SOME FACTORS REVELANT TO THE DEVEiO?MENT OF A SECONDARY HUNGER DRIVE Thesis for #he- Degree of M. A. MiCHiGAN STATE COLLEGE Richard Arflwr Behan 1949‘ THESIS This is to m-rtill] that the lhvsis Pntitlcd AH Investi; AeleVPHt to a Seccngary of Some Vactcrs Developucut of n .6 lunfier Drive lu‘c‘svmm] bl] has been am‘vlm-ql tnwnnlx fullillnn-nt ml llw requirements fur L1y*tgz' (I.‘vFUGpwrin_lllifiiullfléfi’ X§£w~4fl ,' ' lvl‘Hfo'vaI' 3/ > V K4. “mp ".g £m,.‘:l’3 ‘5" /2y2 :- .nu- .A."'b-.‘4 5-.— .‘ULAE-_- g AN INVESTIGATION OF SOME FACTORS RELEVAJT TO THE DEVELOPMENT OF.A SECONDARY HUNGER DRIVE By Richard Arthur Behan A THESIS submitted to the Graduate School of Michigan State College of Agriculture and.Applied Science in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 19h9 ACKNOWLEDGMENT Grateful acknowledgment is made to Dr. M. Ray Denny for his willing guidance and encouragement, and for his tireless assistance in the study here reported. ii TABLE OF CONTENTS Page LIST OE TABLES 111 LIST OF FIGURES iv I. INTRODUCTION To THE PROBLEM I II. STATEMENT OF THE PROBLEM 6 III. EXPERIMENTAL PROCEDURE 7 Apparatus 7 Subjects 7 Preliminary Training 8 Method 8 IT. RESULTS 13 v. INTERPRETATION OF RESULTS 19 v1. SUIIIERY AND CONCLUSIONS 21+ BIBLIOGRAPHY 27 APPENDIX 29 I. II. III. IV. V. VI. VII. LIST OF TABLES Inside dimensions of the various parts of the apparatus Comparison of mean and median running time for control and experimental groups on the first test trial Comparison of the median running times for the control and experimental groups for each of the five test trials Comparison of the median running time per trial for the experimental and control groups for all of the trials Summary of analysis of variance of four groups of six animals each, of the number of correct responses each day for five days Summary of analysis of variance of four groups of six animals each, of the number of correct responses on the first two trials of each day for five days Summary of analysis of variance of four groups of six animals each, of the number of correct responses on the first trial of each day for five days iii Page 15 16 17 18 18 18 LIST OF FIGURES Diagramatic presentation of the straight alley maze Diagramatic presentation of the T-maze Jedian running time per trial for experimental and control groups on the test trials Total number of correct responses for each group for each of the five days on the T—maze trials iv Page 12a 12b 18a 18b I. INTRODUCTION TO THE PROBLEM The problem of secondary or externalized drive of the positive or adient type has received considerable textbook attention in the past few years, but little attention in the way of research. E. E. Anderson (1), who is responsible for the term externalized drive, describes the concept in the following manner. It is ordinarily conceded that when an animal is performing in an experimental situation under some sort of primary drive, that the drive stimulus is within the organism. With a drive like hunger, the drive stimulus is also considered to be internal to the organism. Now as the animal continues to perform in the experimental situation, it is felt that the ex- perimental Situation (maze, etc.) gradually Comes to acquire the property of eliciting at least some of the components of the original drive stimulus. In other words, the conditions for drive become ex- ternal to the animal, and drive is said to be externalized. After the drive has become attached to stimuli external to the animal, and in the absence of the original internal conditions, the mere presence of the proper external conditions will suffice for its arousal. In three later papers Anderson (2, 3, M) gives the results of rat studies which would seem to substantiate certain hypotheses suggested.by his concept. The first of these studies was undertaken to investigate the effect of satiation and removal of reward at various stages in the learning process. Anderson (2) used three groups of ten rats each in a 1h unit T-maze; Group I was hungry and rewarded on each of 38 trials; Group II was hungry and rewarded except on trials 8 to 12 and 29 to 38. Group III was hungry and rewarded on trials 1 to 8 and 1“ to 28, but was satiated and rewarded on trials 9 to 13 and 29 to 38. The results indicated that satiation was a more disturbing factor early in learning, and that removal of reward was more disturbing late in learning. A second part of the same experiment was concerned with the ef- fect of changing food boxes with hungry and satiated, and rewarded and non-rewarded animals. The results, on the whole, were in line with Anderson's theory that changes in external "drive" conditions may be as detrimental to performance as altered internal drive. The second study (3), concerned with the ability of non—rewarded and satiated rats to learn a maze, was a test of the hypothesis that once a drive mechanism Can be aroused.by external stimuli, such an ex- ternalized drive should make possible the establishment of new behavior patterns. Hunger motivated rats were given 73 food-rewarded trials in a 1M unit T—maze. Groups of animals were then run on two other mazes under the following conditions: (1) hungry and rewarded, (2) hungry and nonprewarded, (3) satiated and rewarded, (u) satiated and non-rewarded. Control groups did not receive the 73 trials of rewarded preliminary training. All experimental groups showed marked improvement and were superior to the control groups which ran under the same conditions of motivation. The last of the studies reported by Anderson (M) dealt with the effect of pre-feeding on the maze performance of hungry non-rewarded rats. This group was given H2 non-rewarded trials in a six-unit T-maze. Preceding each trial, and before being placed in the maze, each animal was allowed to eat a small amount of food. After 36 non-rewarded trials the results showed a statistically significant reduction in error scores on trials 37 to M2, although the animals were never rewarded in.the goal box on these trials. Additional investigations of secondary hunger drive have been carried out by N. E. Miller. According to an abstract (10) one study is concerned with the relative efficiencies of Shock and hunger motiva- tion in setting up acquired drives. A group of satiated animals was trained to get out offwhite compartment into a black one to escape shock. A second group of hungry animals was trained to go from a white to a black compartment to secure food. Both groups of animals were then tested on the same task in the absence of the motivation that had been in operation during the training period. The habit set up with the aid of hunger motivation extinguished more quickly than did the habit set up with the aid of shock motivation. .Also those trained under shock motivation were able to learn a new habit on the basis of their acquired motivation, while the group that had been trained originally under hunger motivation did not learn the new habit. The hunger motivated group was then divided into three sub-groups and started on a task which involved running down a short black alley, through a door into a white food box. The groups were run under the following conditions: Group I encountered a variable delay at the door; Group 2 had a longer run to make; Group 3 had neither of the above mention- ed conditions. The subgroup that was frustrated (delayed) ran longer during satiation trials than did either of the other two groups. Miller concludes that electric shock produces a stronger acquired drive than does 22 hour hunger privation, and that the introduction of frustration increases the strength of acquired drive based on hunger privation. More success has been achieved in the attempt to demonstrate an avoidance type of drive. For example, 0. H. Howrer and R. R. Lamoreaux (12), N. E. Miller (11), and M. A. May (9) have clearly demonstrated that a secondary fear drive, i. e., a negative or avoidance drive, may be acquired by rats. When electric shock is associated with a previously neutral stimulus, this neutral stimulus acquires properties originally possessed only by the electric shock. The fact that the positive type of secondary drive is readily ex- tinguished, as is shown in Miller's (10) study and.also by Anderson's (2, 3) satiated animals, raises the question of the relationship of secondary reinforcement to secondary drive, assuming that they are actual- 1y independent concepts. In other words, will secondary drive be more resistive to extinction if there is secondary reinforcement of the re- sponse mediated.by the secondary drive mechanismt Also, since many studies, for instance those of Saltzman (l3) and Grindley (7), show that the secondary reinforcing value of a stimulus readily extinguishes, it might be that if the secondary reinforcing property of the stimulus is frequently reinstated that the secondary reinforcement would be able to reinforce or maintain the associated secondary drive. Since Saltzman (13) has demonstrated that secondary reinforcement with a hunger drive present will mediate learning without the presence ‘ of so called primary reinforcement, it may be possible that secondary reinforcement could operate in a similar fashion without the presence of the primary drive. II. STATEMENT OF THE PROBLEM The present experiment serves as a check on the work of E. E. ,Anderson, but is more specifically designed to investigate whether secondary reinforcement in the absence of the original motivating con- ditions can maintain or prolong a secondary drive. Given animals which have had sixty trials on a straight alley maze to a white food box, under hunger motivation, the Specific hypotheses to be tested are as follows: EEL32_ When run in the same situation in the absence of hunger motiva- 'tion the animals will perform significantly better in terms of lower running times than a control group which has not had the pre-training and which also runs the test trials under conditions of satiation. §L_§L The experimental group will learn, at least at the 5 § level of significance, to go to the white goal box when it is placed in a single unit T-maze situation with only the postulated secondary drive as motivation. H. . If for half of the experimental animals secondary reinforce- ment is reinstated to the white goal box after each day's block of T-maze trails, and for half it is not reinstated, then the former group will show more learning and less extinction effects than the subgroup which receives no reinstatement of secondary reinforcement. III. EXPERIKEHTAL PROCEDURE A. Apparatus The apparatus is presented diagramatically in Figures 1 and 2. It consisted of a single unit T-maze constructed of half inch plywood, with a removable cross-arm so that it could be used as a straight alley maze. The starting box, stem, and the arm of the T were painted light grey inside and out, and the top was covered with plastic screening. The positive and box was painted light grey outside and white inside. The negative end box was painted light grey outside, and black inside. The floor of the negative and box was covered with plastic screening. Both end boxes were covered on top with plastic screening. The inside dimen- sions of the various parts are given in the table below. The starting box was an integral part of the stem, with only a door to mark a separation. TABLE I Inside dimensions of the various parts of the apparatus in inches; Maze Unit Length Width Height Positive goal box 1% 7 9 Negative goal box 1h 5.5 9 Starting box 8 H 6 Stem 3% H 6 Arm of T 21+ 1; 6 B. Subjects The subjects used in this experiment were 2H naive male albino rats from the rat colony maintained by the Department of Psychology of Michigan State College. The animals were approximately one hundred days old when started on the portions of the experiment in which they were to perform. C. Preliminary training Habituation was accomplished by handling, petting, allowing the animals to run about on a table top, and placing them in individual feed- ing cages for a short time each day for a period of five days for M5 minutes per day. On the fifth day all animals were allowed to explore in the starting box and stem of the maze for a period of ten minutes. All of the groups received this training. Setting up the hunger drive required four days and was accomp- lished as follows. On the second of the habituation days the animals were given only eight grams each of their usual diet, in individual feed— ing cages. They were given ample water at all times. This feeding pro- cedure was followed on each of the remaining three days. All of the animals received this experience. D. {ethod Part I of the experiment consisted of giving twelve animals sixty trials on the straight alley maze to the white end box with a hunger drive of 22 hours food privation. The incentive used was a small pellet (approximately 0.3 grams) of Dickinson's Puppy Food placed in a glass coaster in the food box. The trials were given six per day for ten days. Each trial was timed as follows: The animal was placed in the starting box, and when it was facing the door, the door was opened and the stop watch was started. The watch was stopped when all of the animal& except its tail was in the goal box. The watch was operated manually. The measure used was, therefore, a combined latency and running time measure. Each animal received eight grams of food per day during this training and was fed separately in individual feeding cages one half hour after each day's runs were completed. Part II of the experiment consisted of straight alley maze test trials given according to the following procedure. 1. Procedure for satiation. Satiation was accomplished over a period of M8 hours while the animals were in their home cages. Large amounts of dry food, wet mash, and unlimited water were available at all times. This procedure was used with all of the animals. 2. Procedure for test trials. At the end of the MS hour satiation period five unrewarded test trials were given all on one day. At least five minutes elapsed between trials for any rat. Before any animal ran its test trials it was given an opportunity to eat as much of the food as it ordinarily found in the goal box as it desired. This was to insure that the animals would be satiated on the test trials as in Part I. The door to the goal box, three feet distant from the starting box, was open and the starting box door closed. S was placed in the starting box, and when S was facing the door, the door was opened and the stop watch was started. When all of the animal except its tail was in the goal box, the watch was stopped and the door to the goal box closed. If an animal did not leave the starting box it was allowed to remain there until five minute extinction criterion was reached. If S did not enter the goal box at the end of five minutes it was considered to be extinguished on 10 that trial, and was removed from the maze to await its next trial. All groups were tested in this manner. Part III of the experiment consisted of giving each of the groups fifteen trials on a single unit T-maze. 1. Procedure for T—maze trials. T-maze trials were given over a five day period for each group, three trials per day, for a total of 15 trials. Before an animal ran on the T-maze it was given an opportunity to eat as much of the food as it ordinarily found in the goal box as it desired. For this procedure the animals were placed in individual feeding cages supplied with food and water. Trials were given in such an order that each animal had its first trial before any animal had its second, but the order of running the animals was changed for eacn trial. The white goal box was on the side opposite the rat's preference. Preference was determined for each rat according to the direction it chose on its first trial. This choice was automatically counted as a wrong choice, i. e., as each rat made its first choice the black end box was placed at the end of that arm. Trials were given as follows: The door between the starting box and the maze stem, and the ones at the ends of the T were closed. The three doors at the choice point were open. An animal was placed in the starting box and when it was facing the door, the door was opened, and the stop watch was started. The starting box door was left open during the remainder of the trial. As S entered an arm of the T a door was closed behind it to prevent retracing, and the door at the entrance to hhe end box was opened. When all of the animal except its tail was in the end box, the watch was stopped and the door was closed. The time for 11 each trial and the direction turned (towards the white or the black and box) were recorded. If an animal had not left the starting box at the end of one minute it was pushed out into the stem of the maze. If an animal had not made a choice at the choice point by the end of five minutes, it was taken out if the maze to await its next trial. If an animal en- tered one of the arms of the T, but had not entered an end box by thirty seconds from the time it made its choice it was removed from the maze to await its next trial. Each rat remained in an end box for ten seconds before being removed. There was food in the goal box at all times (but no rat offered to eat it). 2. Procedure for reinstatement of secondary reinforcement. After the animals which were to recieve reinforcement had finished their T-maze trials for any day they remained in the feeding cages (without food) for a period of six hours. At the end of the six hour period the animals were placed, one at a time, in the white goal box and allowed to eat un- til they refused more food, or for ten minutes, whichever was shorter. After this the animals were returned to their home cages to await their next day's trials. 3. The subgroups 9: Part III. Group A... Experimental group A consisted of six animals. After satiation the animals were given five test trials, following which they were given fifteen T-maze trials with feeding in the white goal box six hours after each day's run. 'Egggp E, Control graup B consisted of six animals treated exactly like group A, except that they did not receive 60 straight alley trials. 'Egggplg. Experimental group C consisted of six animals which had train- ing similar to that of group A, with the exception that they received no reinstatement of secondary reinforcement after running each day's block of T-maze trials. Group 2. 12 Control group D consisted of six animals which were trained under conditions similar to that of group B, except that group D received no reinstatement of secondary reinforcement after running each day's block of T-maze trials. h. 1. Summary of experimental procedure. Group _A_ Habituation 60 straight alley trials Satiation Test trials T—maze Reinstate secondary reinforcement Group‘g Habituation Satiation Test trials T-maze Reinstate secondary reinforcement Group 9 Habituation 60 straight alley trials. Satiation Test trials T-maze Group‘g Habituation Satiation Test trials T-maze SB r 1- )h L .3 . 12b . \ | N6 F: LA Jcpln RA 0 P8 SE! Figure 2. Diagramatic presentation of the T-maze. SB--starting box, S--stem, CP--choice point, RA--right arm, IA--left arm, PB--positive end box, NB--negative end box, D--door. 15 RESULTS Table II summarizes the differences between the running times of the experimental and control groups for the first trial. This trial is considered crucial because the effects of secondary reinforcement by way of the food in the goal box has not had a chance to operate for the con- trol group. dean differences and median differences are compared. The first test trial differences were significant beyond the one percent level of confidence, as predicted. The experimental group ran significant- ly faster than the control. Table III is a summary of the differences between the control and experimental groups for each of the five test trials with reference only to the median running scores.1 It will be noticed that only the first two trial differences are significant, and that these are significant be- yond the one percent level of confidence. The probabilities of obtain- differences as large as were observed in the last three trials, purely by chance, are very high. These results at first analysis seem to offer a confirmation of hypothesis I, particularly because of the faster running time of the experimental group on the crucial first test trial. A comparison of the median running scores of the experimental and control groups for all trials is found in Table IV. This difference is significant at only the twenty percent confidence level. Using all five test trials hypothesis I is supported with a rather low degree of confidence. 1The median scores were used here because there were in the con— trol group several animals which met the extinction criterion of five minutes without entering the goal box. None of the experimental animals extinguished during the test trials. in With respect to hypothesis II, namely that the experimental animals would learn a new habit on the basis of the postulated secondary drive, and hypothesis III, that those animals which had received rein- statement of secondary reinforcement after each day's block of three T-maze trials would perform in a manner superior to those animals which did not receive reinstatement of secondary reward, refer to Tables V, VI, and VII. Here are summaries of the analysis of variance for the four groups for (l) the number of correct responses each day for five days, (2) the number of correct reSponses on the first two trials of each day, and (3) the number of correct responses for the first trial each day. It will be noticed that there is a greater difference within groups than between groups wherever a significant F score appears in the tables. In no case is there a significant difference between the control and experimental groups, nor is the mean square for the experimental con- dition significant. It is interesting also to note that the analysis for the number of correct responses on the first trial of each day yields as much information as do the other two analyses. These data do not support either hypothesis II or hypothesis III. Consequently these hypotheses must, in the light of these results, be abandoned. The fact that hypotheses II and III are not confirmed casts doubt on the validity of hypothesis I, because a EEEE‘EEEE drive should mediate further learning. TABLE II 15 Comparison of mean and median running times for the control and ex— perimental groups on the first test trial. Mean Data Median Data # Sig. Group N M 6' t j N Mdn P.E. ratio 2 Experimental 12 6.75 3.36 12 3.0 0.31 2. .01 6.3M .01 Control ‘11 32.55 8.16 12 36.5 5.27 16 TABLE III Comparison of the median running times for the control and experimental groups for each of the five test trials Diff _ Trial Mdnl - Mdng PE diff PE diff p 1 3.5 .28 6. h .01 2 i9.0 E.u9 n.33 .01 a 2.0 13.M2 0.15 .98 27.5 h2.h8 0.66 .51 5 u.o 17.57 0.23 .82 17 TABLE IV Comparison of the median running time for the ex- perimental and control groups for all of the trials Trial Experimental Group Control Group 1 3.0 36.5 2 5.5 211.5 a 29.0 27.0 21.5 M9.0 5 no.0 1111.0 x 19.8 36.2 mm. 21.5 36.5 a’mdn. 8.l+7 5.87 sig. ratio l.h5 P 0.20 18 TABLE V Summary of analysis of variance of four groups of six animals each, of the number of correct responses each day for five days. Source of variance d.f. S.S. M.S. sig.F p Total 113 75.99 Rows 3 fi5 0.86 — Columns 23 22.39 0.97 1.79 (.05 Between groups 3 0.88 0.29 - Within groups 20 21.51 1.08 1.97 14.05 Experimental condition 1 0.u1 0.u1 - Error 92 50.15 0.55 TABLE VI Summary of analysis of variance of four groups of six animals each, of the number of correct responses on the first two trials of each day, for five days. Source of variance d.f. 5.8. M.S. sig.F p Total 119 u?.70 Rows h 2.28 0.57 - Columns 23 10.90 0.u7 - Between groups 3 1.17 0.39 - Within groups 20 9.73 0. 9 - Experimental condition 1 0.30 0.30 - Error 92 39.52 0.38 TABLE VII Summary of analysis of variance for four groups of six animals each, of the number of correct responses on the first trial of each day for five days. Source of variance d.f. S.S. M.S. sig.F p Total 95 33-33 0.25 Rows 3 0.75 0.25 - Columns 23 8.33 0.36 1.7M .< .05 Between groups 3 1.00 0.33 - Within groups 20 7.33 0.37 1.79 <_.05 Experimental condition 1 0.16 0 16 - Error 69 1h.25 0.21 PFP TD IAL ‘ I N '1‘ I13 31 DI r'fi tJ IN SE CONDS on C? m 9 18a /_ __ _ _ / \ \ \ \ \ v FY1313. ‘3. II it}? TA L CSETTROL _____ '3 1 2 5 4 5 TRIAL Figure 3. Lledian running time per trial for control and experimental groups on the test trials. TOTAL 1101-28793 0? CORRECT RFSPOI‘JSFTS 18b 1 p 2 /\ I \ 11- ,’.T\<_ A // ~ \ \ ” \ I I l l I 7 - I, \\ // I \ , 6- I / CBC-UP A _ -- _ _- 5’ GROUP C .____.___. _. GROUP B _____ GROUPD I; " O l 2 3 4 5 DAYS Figure 4. Total number of correct responses for each group for each of the five days on the T-maze trials. 19 V. INTERPRETATION OF RESULTS The present finding, that animals which have had 60 rewarded trials on a straight alley maze with a hunger drive will run to the goal box significantly faster, under conditions of satiation, than will animals which have not had this training seems to confirm.Anderson's hypothesis of externalized drive (1). However, it should be emphasized that the differences between the control and experimental groups on the test trials were significant for only the first two test trials. The mechanism mediating faster perfor- mance by the experimental group seems to be very transient or labile. Furthermore, since both hypotheses II and III were not confirmed the mechanism of secondary drive, as Anderson conceives it, at least, receives no support from the present investigation. The present results are more in line with those of Miller (10). For a bona fide drive state to exist there must be the mediation of learning. Otherwise, the results of better performance during the test trials by the experimental group as compared with the control group may be explained on the basis of a running habit multiplied by residual drive.1 Such an explanation is in terms of Hull's theoretical analysis of performance. or effective reaction potential. For Hull, effective reaction potential, or performance is a joint multiplicative function of drive and habit. Therefore, if the one is increased, the resultant is 1Any number of irrelevant and unsatiated drives operative at the time the rat is satiated for food and water. 20 increased. In the present study, habit and presumably residual drive remain relatively constant, and drive is greater for the experimental group. It is also likely that fractional anticipatory goal responses, as differentiated from a secondary drive state, were learned to the maze situation. However, the more permanent habit phenomenon that would be established in the straight alley trials is the running response. This habit is maintained in all of the animals of the experimental group on all five test trials. No animals of the experimental group reached the extinction criterion, of five minutes without entering the goal box, while three animals of the control group refused to enter the goal box at least one time. The animals of the experimental group showed a fairly . steady increase in running time with successive test trials (see Figure 3), but this phenomenon may be readily explained by the rapid extinction of fractional anticipatory goal responses. Incidently, the control group of the present experiment does not show much of an increase in running time with successive test trials (see Figure 5). Koreover, all of Anderson's results are explicable in terms of maze habit times residual drives, or in terms of the extinction or dis- ruption of fractional anticipatory responses. It is not necessary, or desirable to postulate an independent secondary drive. The results of the present study which indicate rather definitely that secondary reinforcement does not operate in the absence of some ap- preciable primary drive state are entirely in accord with the results of two recent studies by Estes (5, 6) which were published after the com- pletion of the present study. In this study thirsty rats learned to operate a modified Skinner 21 type apparatus for water reward. Test trials were given under hunger motivation with only a secondary reinforcing agent as a reward. Re- sistance to extinction with a six hour food privation and the secondary reinforcing click was no greater than for the control group for which the click was omitted. The click plus 23 hour food privation, on the other hand, gave significantly increased resistance to experimental extinction. The fact that Saltzman (13) has demonstrated that rats will learn a maze under a hunger drive when the only reinforcement is of a secondary nature, to indicate that our results of no learning on the T-maze may not be attributed to failure of secondary reinforcement to mediate learning. Saltzman not only demonstrated that rats would learn on the basis of secondary reinforcement, but also that, under the proper conditions second- ary reinforcement is as efficient as primary reinforcement (if there be an actual difference). Results similar to those obtained in the present investigation with the five test trials are found in a recent abstract by Kimble (8). Kimble trained rats to operate a panel pushing apparatus to get a pellet of food, satiated them, and gave them four sets of five trials a day. Using reaction latency as a measure he found that time increased signific- antly within each day's block of five trials. On the first trial of each day latency was very short, and approximately the same on each day. The slope of the performance curves increased, however, with succeeding days. Such results would seem to indicate that the fractional anticipatory re- sponses extinguish rapidly in one day's set of five trials, but recover in the intervening time, while the habit remains fairly constant over the four day period. Thus Kimble's results seem also to be explicable in 22 terms of an instrumental habit, residual drive, and fractional anticipatory goal responses. This author believes that the present results might be better ex- plained in terms of secondary reinforcement mediated by anticipatory goal responses than by a conditioned drive state. With each of the 60 straight alley trials the animals made a goal response. This goal response is com- posed of a number of discrete and independent, but sequential acts. Some of these are: seizing the food, chewing, salivating, swallowing, taking a particular stance, etc. It may be assumed that as training progresses some of these reSponses, which do not interfere with the gross behavior of the organism, come forward in the behavior sequence, until the entrance into the starting box is a stimulus for eliciting these partial goal re- Sponses (for example, chewing, salivating, and swallowing responses). In other words, the maze itself comes to mediate the goal response. Some observational evidence in support of this view is on hand. It was noticed that some of the animals, after they had had a few trials, would, if placed beside the starting box, climb into it. Once inside the start- ing box they would display behavior which one might interpret as intent to get out of the starting box and to run to the goal box. Not only the animals which would climb into the starting box displayed this latter type of behavior, but after a few trials every animal, when placed in the starting'box behaved as though it were very anxious to be OIlitS way to the goal box. Another bit of evidence of fractional goal responses on the part of the experimental animals is to be found in their behavior during the 91-maze trials. The majority of animals, when they entered the positive gpal box, would go immediately to the food dish and sieze the piece of 23 food. However, no animal attempted to eat this piece of food. This expectancy which consists of fractional anticipatory goal responses seems to be very labile when satiated animals are subjected to extinction trials (Figure 3). However, as Fstes (5, 6) and Saltzman (13) have shown, when animals are motivated by a primary drive, the eXpectancy mechanism operates rather effectively. A great deal of additional research is necessary to understand the behavior of animals under conditions of satiation. 24 VI. SUIZ'ARY AND CONCLUSIONS The present study was conducted to check on the validity of the concept of secondary or externalized drive of a positive or adient type, and to investigate whether secondary reinforcement would maintain or pro- long a secondary drive in the absence of the original motivating condi- tions. There were two experimental and two control groups each with an N of six. Both experimental groups were given 60 trials on a straight alley maze with food reward under conditions of hunger motivation. The control group did not receive this training. All four groupsjwere then given five trials on.the straight alley maze under conditions of satiation. All groups were then given fifteen trials, three a day, on a simple T-maze. One of the experimental groups and one of the control groups received secondary reinforcement in the white goal box six hours after each day's block of T-maze trials. The apparatus consisted of a straight alley maze which could be converted into a simple single unit Tdmaze. All animals received five days of handling before the straight alley trials began. The results revealed a significant difference between the control group (N I 12) and the experimental group (N I 12) on the first two of the test trials only, and no significant differences between the four groups (N = 6 each) in their performance on the T-maze. On the basis of the results found in this study the following conclusions may be drawn: 1. Animals which have been conditioned to make a response which leads to food.when they are motivated by hunger tend 25 to make that response when the original primary motivation is absent. The mechanism which is reSponsible for the tendency to make a learned response in the absence of original primary motivation is very labile. Secondary reinforcement is not effective in strengthening a new response in the absence of any primary motivation. The mechanism which is reSponsible for the tendency to make a learned response in the absence of the original primary motivation is not of sufficient strength or duration to mediate the learning of a new response. BIBLIOGRAPHY 10. ll. 12. 13. REFEREE: C ES Anderson, E. E., The externalization of drive. I. Theoretical considerations. Psychol. rev., l9hl, #8, act—22M. Anderson, E. E., The externalization of drive. II. The effect of satiation and removal of reward at different stages in the learning process of the white rat. g, ggnet. Psychol., 19h1. 59. 359-376. Anderson, E. E., The externalization of drive. III. Maze learn— ing by nonprewarded and satiated rats. g, genet. Psychol., 19h1. 59. 397-h26. Anderson, E. E., The externalization of drive. IV. The effect of prefeeding on the maze performance of hungry non-rewarded rats. g. comp. Psychol., 19u1, 31, 3u9-352. Estes, W. K., Generalization of secondary reinforcement from the primary drive. i, exp. Psychol., 19H9, 39, 28b-295. Estes, W. K., A.study of the motivating conditions necessary for secondary reinforcement. g, exp. Psychol., l9u9, 39, 306-310. Grindley, G. C., Experiments on the influence of the amount of re- ward on learning in young chickens. Brit. g. Psyghol., 1929, 20, 173-180. . Kimble, G. A., Behavior potentiality under conditions of satiation. Amer. Psychol., 19kg, u, 218. Abst. May, M. A., Experimentally acquired drives. g, exp. Psycholl, l9h8, 38: b0'770 Miller, N. E., Experiments on the strength of acquired drive based on hunger. Amer. Psychol., 19u7, 2, 303. Abst. Miller, N. E., Studies of fear as an acquired drive. I. Fear as motivation and fear reduction as reinforcement in the learn- ing of new responses. 1. exp. Psychol., l9u8, 38, 89—101. Mowrer, O. H. and Lamoreaux, R. E.,.Avoidance conditioning and signal duration. .A study of secondary motivation and reward. Psychol. MOnogr., 19u2, 5h, No. 5. Saltzman, I. J., Maze learning in the absence of primary reinforcement. A study of secondary reinforcement. g. comp. physiol. Psychol., 19kg, M2, 161-173. l. 5. 28 REFEMNCES NOT CITED Hull, Clark 1..., Principles of behavior. New York: D. Appleton Century Co., 19h3. pp. 231-232, ans-2M9. Kendler, H. H., Drive interaction. I. Learning as a function of simultaneous presence of hunger and thirst drives. 9;. 9352. Psychol., 19u5, 35, 96-109. Siegel, P. 5., Alien drive, habit strength, and resistance to ex- tinction. g. comp. Psychol., 19kb, 39, 307-317. Viebb, W. 28., The motivational aspect of an irrelevant drive in the behavior of the white rat. i. exp. Psychol., 1949, 39, l-lu. Young, P. T., Food seeking drive, affective processes, and learning. Psychol. rev., 19kg, 56, 98-121. APPEICDI X Behavior data of the animals of the experimental groups on the five straight alley maze test trials in terms of seconds of running time. Animal No. H H to UTKJ'I 4r WNKJJKNO‘sKNNNONI-JKN TABLE.A 120 0 rs TABLE B BehaNior data of the animals of the control group on the five straight alley maze test trials in terms of the seconds of running time. .Animal No. Trial 1 2 3 u 5 1 M5 8 5 19 6 2 7 1h 2M 2M 17 a 8 15 31 51 #9 10 2h 9 13 13 5 5 6 7 ll 17 6 M3 39 22 122 62 7 23 lg 35 16 39 8 31 1 u 300* 281 9 300* 300* 300* 300* 300* 10 H7 31+ 79 37 17 11 Me 300* 255 300* 300* 12 97 3M 8 108 100 *Denotes reached the extinction criterion. "L TABLE C IBehavior data of the two experimental groups on the T-maze trials “W" represents a correct response "B" represents an incorrect response "E‘" denotes extinction criterion reached Trials GroupA 123M56789101112131h15 .Animal No. 1 swssswwwswsswww 2 swwwwwwxsssswss a BBWWWWWWBBBBWBB Bswwswswwwwswsw 5 BBWWBWBWWBBWWBW 6 BWBBBBBWWWWWWBB Group C Animal No. 1 BWWBWWWBBWBBBBW 2 wawwwwswwssssw a BBWWBBWBWBWWWB'B swwwsmwwwwwwwww 5 BWBWBWBBBBBWBBB 6 BWBBBBBBBBBWBBB 31 TABLE D "W" represents a correct response "B“ represents an incorrect response “E*" denotes extinguished in that trial IBehavior data of the control groups on the T—maze trials. Trials 1 2 3 u 5 6 7 8 9 10 11 12 13 1M 15 Grouq>B .Animal m. BWPWBW WBWBBW BWPWBB BWWWWB WBWWBB BWWWWW BWPBWW BBWWBW WWWWWB WBPWWW WWVBWB BWWWWW WWWWBB WWWBWW 333333 123456 Gme .Animal No. WWBWWB WWWWBW BWBBBB BWBWBW WBBWBB WBBBBB BWBWBW BBBWWW WWWBWB BWWWWW BWWBBB WWWBBW BWBBBW WWWWWB 333333 12721456 Y LIBR 057 ARIES M'CIIII'IIIIIHIITIIIIII IIIIIIIIIIIII 3 1 19 3 0 3 0 8 3 0