International Journal of Primatology, VoL 15, No. 4, 1994
Effects of Modeling and Lineage on Fishing Behavior in the Small-Eared Bushbaby
( Otolemur garnettii) Sheree L. Watson, 1 Martha Schiff, 1 and Jeannette P. Ward 1,2 Received January 12, 1993; accepted April 5, 1993
Thirty-eight bushbabies (Otolemur garnettii) were subjects in an observational learning study. We exposed them to one o f three modeling conditions: (1) fishing model--one that actually performed fishing behavior; (2) nonfishing model-----one that performed as a model in every way except performance of fishing behavior; and (3) no model. We assessed them with regard to latency to approach the fishbowL latency to make an initial fishing attempt, duration of time spent in the vicinity of the fishbowls, and number of actual fishing attempts. Results indicate that subjects that were exposed to either fishing or nonfishing models were faster to approach the fishbowls and spent more time in the vicinity of the fishbowls than animals in the no-model condition Lineage, i.e., whether or not the animals' parents fished, rather than modeling condition, was the best predictor of the latency to initial fishing attempt and the number of attempts made. KEY WORDS: small-eared bushbaby; observational learning; heredity; socially enhanced learning.
INTRODUCTION Although socially enhanced learning (Galef, 1988) has been demonstrated in several anthropoid primate species (Cambefort, 1981; Myers, 1970; Huffman, 1984; Huffman and Quaitt, 1986; Presley and Riopelle, 1959; Riopetle, 1960; Strayer, 1976; Tomasello et al., 1987; Watanabe, 1989; Wec'hkin, 1970; contra Chamove, 1974; Fragaszy and Visalberghi, IDepartment of Psychology, The University of Memphis, Memphis, Tennessee 38152. 2To whom correspondence should be addressed. 507 0164-0291/94/0800-0507507.00/0 O 1994 PlenumPublishingCorporation
508
Watson, Schlff, and Ward
1989; Visalberghi and Fragaszy, 1990), only two prior studies assessed the extent to which prosimian primates may benefit from observing the behavior of conspecifics. Feldman and Klopfer (1972) provided evidence that juvenile lemurs (Lemurfulvus) were able to learn a stimulus-discrimination task faster if they first observed their mothers perform the task, though a similar saving in learning was not apparent if the subject observed another juvenile. However, the small number of subjects and the high individual variation in scores rendered the results of this study somewhat inconclusive. Welker (1976) reported the first observation of fishing behavior in a prosimian species---4he small-eared bushbaby. H e also found evidence that acquisition of fishing was enhanced by the opportunity for juveniles to observe their more experienced parents perform the task. However, all bushbabies were free to move about together in the test area and no control group was employed. Therefore, the possibility that some direct learning occurred cannot be ruled out. It is possible, for example, that bits of dropped fish may have served as a reward for incomplete responses. Our research is a systematic replication of Welker's (1976) original work, with two modifications. We assessed fishing behavior in three groups of bushbabies for which all test conditions were identical except that, before evaluation, one group observed models that fished, one group observed models that did not fish, and the third group had no model. Moreover, the observers and models were physically separated throughout the course of the experiment. W e employed the dual control design to determine whether social enhancement, if it occurred, was most consistent with local enhancement or direct imitation. Local enhancement theory suggests that socially enhanced learning results from an animal's attention being directed to a particular portion of the environment by a conspecific's activity there (Fragaszy and Visalberghi, 1989; Galef, 1988; Pallaud, 1984; Thorpe, 1956). Imitation implies purposeful, goal-directed copying of an action performed by a conspecific (Galef, 1988; Pallaud, 1984; Thorpe, 1956; Tomasello, 1987). If the observed phenomenon is that of local enhancement, no difference in learning should be evident between the group that observed the fishing models and the group that observed the nonfishing models. However, these two groups should be distinguished from the group that observed no model. If imitation has occurred, the experimental group should differ in performance from both the nonfishing-model and the nomodel groups.
Effects of Modeling
509 METHOD Subjects
We used small-eared bushbabies [Otolemur garnettii (Nash et al., 1989)] from The University of Memphis breeding colony as subjects. In the colony, all 61 bushbabies were housed socially and were maintained on a 12:12 light/dark cycle, with dark onset at 1300 hr. They were watered and fed an ad libitum diet of vegetables, fruits, and Mazuri High-Protein Primate Chow (PMI Feeds, Inc., St. Louis, MO). Test E n v i r o n m e n t and Testing Materials
We conducted the experiment in a 267 x 267 x 288-cm observation room, which was bedded with cedar shavings (Nature's Best, Smithton, MO) and contained two climbing logs, approximately 175 cm long and 20 cm in diameter. We placed a 42 x 61 x 122-cm cage in the room. The cage consists of 1-cm2 plastic-covered wire mesh that was covered with opaque brown pegboard on the top and three sides. A table, 112 cm tall, was 30 cm directly in front of the unobstructed side of the cage. It had a surface area of 3400 cm 2 and supported two fishbowls, separated by approximately 3 cm. We used two fishbowls to discourage competitive inhibition of behavior. The fishbowls were clear glass, 14 cm high, with a 15.5-cm opening at top. They contained 5.5 cm of uncolored gravel and approximately 3 cm of water. When the bowls were filled, the surface of the water was approximately 5.5 cm below the rim of the bowl. The commercially purchased goldfish are commonly referred to as feeder fish. They measured approximately 2 cm in length. Procedure
To establish the acceptability of goldfish as a dietary item, and thus a suitable incentive for fishing behavior, every colony animal was offered a freshly killed goldfish. After it was determined that all bushbabies consumed goldfish readily, we selected models and subjects. Previous research in our laboratory demonstrated that bushbabies tend to become lethargic and anorexic when isolated from conspecifics. Therefore, throughout the experiment, each subject was paired with its colony cagemate.
510
Watson, Schiff, and Ward
Selection and Training of Models. Models were randomly selected from among the seven wild-born bushbabies available in the colony. They were the only animals whose life histories were not entirely known and that may have had experience with the fishing task. We trained one pair to fish as follows: on days 1 and 2, we placed freshly killed goldfish in the subjects' food cups along with the standard diet. On days 3 and 4, freshly killed fish were floated in a fishbowl conraining approximately 3 cm of water. On day 5, we presented the subjects with fishbowls of water containing live fish. The models fished robustly. Beginning on day 8, they caught and ate every fish presented to them. We designated them fishing model. The other pair was not trained. We designated them nonfishing model. Subject Selection and Assignment to Conditions. From the remaining 57 colony members, we selected 38 bushbabies randomly. The only colony members excluded from the selection process were those that were <6 months old or that had physical limitations that could interfere with testing (e.g., missing digits). Twelve subjects observed the fishing models (Group F), 12 observed the nonfishing models (Group NF), and 14 were not exposed to any model (Group N). Within the constraints of the colony composition, the age ranges represented in each group are roughly equivalent. Table I lists descriptive data about each subject. Training and Testing Procedures. The procedure was identical for all three groups with the exception of the models' presence or behavior or both. On day 1, we placed the observing pair in the holding cage and released the models for Groups F and N F into the room approximately 2 hr before the scheduled dark onset, i.e., at about 1100 hr. Between 2 and 4 hr after dark onset, i.e., 1500-1700 hr, we placed two fishbowls, each containing five live fish, on the table. In Group NF, the fishbowls were covered with clear plastic---Saran Wrap---as a precaution against possible spontaneous fishing by the nonfishing models. We allowed the observers to observe the models (or fishbowls for Group N) for 30 min. To ensure that the behavior of the two model types was comparable, we videotaped representative samples of the models' behavior throughout the experiment. Then we removed the fishbowls the from the observation room and placed the models' food cups on the table where the fishbowls had been. We fed the observing animals in the holding cage. On day 2, we repeated the procedure with the following modifications. When the 30-min observation period had ended and the fishbowls were removed, we returned the models to the colony. Then we released them from the holding cage and placed their food cups on the table. Testing began on day 3 and continued through day 5. On these 3 days, we placed the two open fishbowls, each containing five live fish, on the table for 1 hr daily. We videotaped sessions via an
Effects of Modeling
$11 Table I. Descriptive Information About Subjects
Name
Observer pair
Modela
Sex
Bilbo Frodo Giles Justin Darryl Maude Samantha Nickolas Barney Abbey Noel Skinner
1 1 2 2 3 3 4 4 5 5 6 6
F F F F F F F F F F F F
M M M M M F F M M F F M
7 7 17 17 30 25 66 31 68 > 115 31 44
Daisy/Bogart Daisy/Bogart Lombard/Gahle Daisy/Bogart Hepburn/TracT Daisy/Bogart Livy/Twain Abbey/Barney Hepburn/TracT Unknown Abbey/Barney Lombard/Gable
Burice Gadriviei Sylvia Buffie Mickey Daphne Michael Garrett Daisy Bogart Sadie Lashley
7 7 8 8 9 9 10 10 11 11 12 12
NF NF NF NF NF NF NF NF NF NF NF NF
M M F F M F M M F M F M
8 7 12 12 25 24 24 24 > 104. > 120 67 52
Livy/Twain Lombard/Gable Abbey/Barney Sadie/Lashley Dalsy/Bogart Livy/Twain Lombard/Gable Sadie/Lashley Unknown Unknown Lombard/Gable Hepburn/TracT
Hope Niko Niles Trina Eden Uriah Joey Walt Peter Odie Clara Robert Bib Bub
13 13 14 14 15 15 16 16 17 17 18 18 19 19
N N N N N N N N N N N N N N
F M M F F M M M M M F M F M
41 52 52 25 42 22 20 19 51 50 30 8 20 20
Lombard/Gable Hepburn/TracT Hepburn/TracT Livy/Twain Daisy/Bogart Sadie/Lashley Sadie/l.ashley Abbey/Barney Abbey/Barney Daisy/Bogart Abbey/Barney Livy/Twain Hepburn/TracT Hepburn/TracT
~
Age (mo)
Parents
fishing model; NF, nonfishing model; N, no model.
R C A Camcorder (Model CPS04). Upon removal of the fishbowls each day, we placed the animals' food cups on the table. After the day 5 testing session, we returned the observers to the colony. We cleaned the table and
Watson, Schiff, and Ward
S12
bedding and aired the room for 2 days before the next group of observers and models were introduced into it. Data Analysis. We scored the subjects on five variables: (1) latency to approach the fishbowl, defined as the amount of time (minutes) that each subject took to approach the fishbowl and display interest in it by touching, rubbing, or sniffing it; (2) duration of interest-----the amount of time (minutes) that a subject spent on the table that contained the fishbowls; (3) latency to make an initial fishing attempt---the amount of time (minutes) that a subject took to make the first fishing attempt; (4) an attempt, defined as a manual thrust into the water during which any part of the hand went below the water level; and (5) a success, defined as actually catching a fish and consuming a portion of it. Thus, if a subject caught a fish but dropped it back into the bowl, we counted it as an attempt, but not as a success. If the subject ate all or part of the fish, we counted it as an attempt and a success. Two of the five behavioral measures latency to approach the fishbowl and duration of time spent in the vicinity of the fishbowls reflect interest, curiosity, and/or absence of neophobia. Two latency to initial fishing attempt and number of fishing attemptz more specifically reflect a proclivity to fish. The fifth measure----number of successes is a measure of skill in fishing. RESULTS Fish counts following each modeling session verified that the fishing models successfully caught and ate all fish during the allotted 30 rain. They also verified that the nonfishing models were unsuccessful in breaking the plastic wrap to remove fish from the bowls. An analysis of representative samples of videotaped modeling sessions indicated that, except for fishing activity, the fishing and nonfishing models behaved in a similar manner during the observation sessions. That is to say, both model types engaged in a comparable amount of activity in the vicinity of the fishbowls, as defined by amount of time spent on the table. The fishing models spent an average of 11.7 min per 30-min session on the table, while the nonfishing models spent an average of 12 rain per 30-min session on the table. This difference was not statistically significant It(4) = --0.05, n.s.]. Effects of Model
Examination of the five measures for differential effects of the three modeling conditions demonstrated that only the two measures that indicate
Effects of Modeling
$13
interest or absence of neophobia, i.e., latency to approach the bowls and duration of time spent on the table, were systematically influenced by the presence or absence of a model. Latency to approach the bowl was influenced by modeling condition as indicated by a Kruskal-Wallis one-way A N O V A (Z2 = 19.74, p < .0001). Kruskal-Wallis follow-up tests revealed that Group F differed from Group N (z = 6.54, p < .05) and that Group NF differed from Group N (z = 16.58, p < .05), but Groups F and NF did not differ from each other. These results are depicted graphically in Fig. la. The analyses also indicated that type of model influenced the amount of time that observers spent in the vicinity of the fishbowls (Z2 = 21.89, p < .0001). Kruskal-Wallis follow-up tests demonstrated that, whereas group F differed from Group N (z = 14.50, p < .05) and Group N F differed from Group N (z = 19.42, p < .05), Groups F and NF did not differ from each other. Figure lc depicts the performance of the three groups on this measure. A similar analysis detected no difference in number of attempts (Fig. ld) or latency to initial attempt (Fig. lb) among the three groups. There were very few successful fishing attempts. Only six animals caught and consumed a combined total of 27 fish (Table II). Thus, this variable was excluded from the analyses.
Effects of Lineage Because we noted that only about half the bushbabies attempted to fish, we further examined the data to discern the possible source of these individual differences in behavior. This revealed an unexpected, but interesting trend. It appears that there was a clustering of fishing behavior among first degree relatives. Because all of the animals in the colony are descended from the original colony founders, the lineage of all but three animals, which were among the original pairs, in the study are known. Depending on whether their parents displayed fishing behavior, we classified the animals of known lineage (n = 35) as being descended either from demonstrated fishers (n = 17) or demonstrated nonfishers (n = 18). Two lines (Gable/Lombard, Bogart/Daisy) tended to produce animals with a proclivity to fish; three lines (Twain/Livy, Barney/Abbey, Tracy/Hepburn) tended to produce offspring that were reluctant to fish. Eleven of the 17 descendants of fishing parents attempted to fish, while only 6 of the 18 descendants of nonfishers did. Although this frequency distribution did not achieve statistical significance (Z2 = 3.445, p < .10), the directional trend prompted further analysis of behavioral measures based on lineage. The descendants of fishing bushbabies made a combined total of 661 attempts, while the descendants of animals that did not fish made a corn-
Watson, Schiff, and Ward
514
7O
7O
6O
6O
50
50
== 40
4o
.E
30
.c_
30
20
20
lo
10
o
0
d 7O
6O
60
5O
50
40
40 3O 20 10 0
NI T
E
3o 2o
lo o
Fig. 1. Results on four variables for animals that observed fishing, nonfishing, or no models. (a) Mean latency to approach the fishbowls; (b) mean latency to make an initial fishing attempt; (c) mean amount of time spent in vicinity of fishbowls; (d) mean number of fishing attempts.
bined total o f only 63 attempts. B e c a u s e o f this striking difference w e anal y z e d t h e v a r i a b l e o f l i n e a g e by K r u s k a l - W a l l i s o n e - w a y A N O V A . D e s c e n d a n t s o f fishing animals m a d e significantly m o r e fishing attempts than descendants o f nonfishers (Z 2 = 4.39, p < .036). M o r e o v e r , the descendants o f fishing animals were faster to m a k e an initial fishing attempt than the descendants o f animals that did not fish (Z 2 = 5.41, p < .02). T h e s e data are depicted graphically in Figs. 2d and b, respectively. A similar analysis o f the variables o f latency to approach the fishbowls (Fig. 2a) and duration o f time spent in the vicinity o f the bowls (Fig. 2c) did not reveal differences b e t w e e n the two groups. All six o f the bushbabies that successfully caught fish are d e s c e n d e d from fishing parents.
Effects of Modeling
515 Table I1. R a w Scores on Six Variables
Name
LAB a
Duration b
LFAc
Attd
Suc e
Lineage f
Fishing m o d e l Bilbo Frodo Giles Justin Darryl Maude Samantha Nickolas Barney Abbey Noel Skinner Nonfishing m o d e l Burice Gadriviel Sylvia Buffie Mickey Daphne Michael Garrett Daisy Bogart Sadie Lashley
1.3 3.3 27.0 61.0 8.6 43.7 5.0 10.0 6.3 6.3 11.7 26.3
70 98 10 0 20 18 25 22 76 34 28 28
1.3 4.0 32.3 61.0 43.3 56.3 61.0 61.0 61.0 61.0 61.0 46.3
95 218 8 0 12 1 0 0 0 0 0 3
5 8 0 0 0 0 0 0 0 0 0 0
Yes Yes Yes Yes No Yes No No No Unknown No Yes
20.3 25.0 11.7 27.3 10.0 8.3 5.0 9.3 1.0 16.7 13.7 25.0
30 26 35 37 34 31 51 34 49 20 36 25
61.0 61.0 61.0 29.3 61.0 61.0 8.3 41.7 20.0 45.3 15.0 42.0
0 0 0 65 0 0 148 4 8 4 77 7
0 0 0 2 0 0 8 0 0 0 2 0
No Yes No Yes Yes No Yes Yes Unknown Unknown Yes No
No m o d e l Hope Niko Niles Trina Eden~ Uriah Joey Walt Peter Odie Clara Robert Bib Bub
14.7 45.7 61.0 61.0 53.0 54.0 54.0 60.3 61.0 61.0 52.6 37.7 37.0 61.0
17 10 0 0 5 2 2 1 0 0 11 14 20 0
18.7 45.7 61.0 61.0 53.0 61.0 61.0 61.0 61.0 61.0 53.3 42.3 47.0 61.0
36 29 0 0 6 0 0 0 0 0 3 3 9 0
0 0 0 0 2 0 0 0 0 0 0 0 0 0
Yes No No No Yes Yes Yes No No Yes No No No No
a M e a n latency to a p p r o a c h t h e fishbowl. bTotal time in vicinity o f fishbowls over 3 days. CMean latency to initial fishing attempts. a'I'otal n u m b e r of fishing a t t e m p t s over 3 days. eTotal n u m b e r of successful catches over 3 days. f D e s c e n d e d from d e m o n s t r a t e d fishers. 9 to e q u i p m e n t failure, we filmed only a portion o f E d e n ' s trial. H e r two successes required m o r e t h a n six attempts, b u t the total n u m b e r of a t t e m p t s is unknown. Only t h o s e recorded are reported. T h e fish c o u n t verified t h e two successes.
Watson, Schiff, and Ward
516
b
(]
.E
70-
70
60-
60
50-
50
4o-
40
30 -
9--q 30
20-
2O
10-
I0
0
0
d
c 70-
60
60-
50
50-
"5 .c -5
-&
4030
~
2O
2O
I0
10
0
0
I
[~
= Fishing parents
mlm = NonFishing parent=
I
Fig. 2. Results on four variables for animals descended from fishing and nonfishing parents. ( a ) Mean latency to approach the fishbowls; (b) mean latency to make an initial fishing attempt; (c) mean amount of time spent in vicinity of fishbowls; (d) mean number of fishing attempts.
DISCUSSION There is little doubt that learning among nonhuman animals is influenced by social factors. However, the nature of the information being transmitted has not been clearly established. Our results indicate that the presence or absence of a model strongly influenced two aspects of behavior. Bushbabies that observed either model type approached the fishbowls sooner and spent more time in the vicinity of the fishbowls. This result, along with the finding that fishing behavior was not affected by type of model, is consistent with local enhancement theory. The model's behavior seems to have served to direct the observer's attention to the relevant area of the experimental environment-the tabletop with the two fishbowls. These results further suggest that, in this experiment, an unrewarded model
Effects of Modeling
517
was as effective as a rewarded model in providing local enhancement (Giraldeau and Templeton, 1991; Riopelle, 1960; Sherry and Galef, 1990). That actual fishing behavior was not influenced in a similar manner suggests that imitation---a deliberate copying of behavior topography is an unlikely possibility. In this instance, the local enhancement may have occurred through a reduction of neophobia. The observer may have been less fearful as a result of watching the model interact with the novel object without harmful consequences (Clayton, 1978). This interpretation is in accordance with Feldman and Klopfer's (1972) finding that observers subsequently approached the novel stimuli more readily than the models did. In our experiment, all subjects were fed on the tabletop. Thus, the observers had experience with the table before their first fishing opportunity. Therefore, any neophobia must have been directed toward the fishbowls. An unanticipated but important finding is that fishing behavior tended to occur in family lines. Lineage is the best predictor of both rapidity of initial fishing behavior and number of fishing attempts. In fact, all 27 of the successful fishing attempts were made by descendants of demonstrated fishers. This findhag raises interesting questions about the source of the fishing behavior. One possibility is that the mothers of the fishing lineage interacted with offspring such that long-term behavioral effects resulted. Rosenson (1972) has reported conspicuous differences in the ways that O. garnettii mothers interacted with their respective infants. Possibly, some of the mothers in our study had interacted with their infants in a manner that resulted in bolder or more outgoing offspring. However, if this were the case, one would expect to see approach latencies and interaction durations influenced by lineage as well as modeling condition. This was not the case in our study. However, it is possible that these mother-infant interactions influenced later problem-solving abilities in offspring, i.e., some mothers produce smarter babies, and that this was a factor in the earlier and more frequent fishing attempts of descendants of fishers. Another possible explanation is that fishing behavior is a manifestation of adaptive behavioral changes that are transmitted genetically to offspring. The results of this investigation, along with those of Welker (1976), provide compelling evidence that at least one species of bushbaby----Otolemur garnett// will readily accept fish as a food item and, indeed, that some are skillful fishers. Whether they consume fish in their natural habitat is unknown. Because many natural colonies live in coastal areas and near rivers (Olson, 1979), fish may well be a natural food item. Possibly, during times of food scarcity, some of the wild-born bushbabies exhibited behavioral tendencies, e.g., fishing behavior and affinity for water, that enhanced survival. Offspring that shared these tendencies may have proliferated in certain areas. The off-
Watson, Schiff, and Ward
518
gin of the wild-born bushbabies in our colony is unknown. Thus, it is possible that some of them originated in coastal regions, while others came from more inland areas. Despite the fact that bushbabies are arboreal, at least one spedee Otolemur crassicw_,dams--foraging has been observed on the ground. Further, by foraging on the ground, they are able to take advantage of food sources that are available during only a few days of the year (Happold and Happold, 1992). Given the availability of small fish during times of food scarcity, some of the coastal inhabitants may have developed a behavioral adaptation not available to their more inland relatives. Lineage did not predict fishing behavior in an all-or-none fashion. Of the 17 descendants of fishers, 11 made fishing attempts; of the 18 descendants of nonfishers, only 6 attempted to fish. This lends support to the notion that heritable factors interact with other factors to produce fishing behavior. It is not known whether the nonfishing animals could be induced to fish through shaping procedures or intensified motivation, i.e., extended food deprivation. Certainly, we observed that fish seem to be a highly preferred food item because in our pretests, every bushbaby, though not food deprived, immediately seized and consumed all fish that we offered. The proclivity to fish may exist in many, if not all, bushbabies, but may be evident only if environmental factors potentiate the behavior. Because bushbabies are social animals in the wild (Clark, 1978), testing the animals in pairs added a dimension of ecological relevance to our experiment. However, this manner of testing may have allowed for social influences during the testing phase. Whether testing-based social enhancement had any effect on fishing behavior in our study cannot be determined because to do so would require a control group of bushbabies tested individually. However, of 12 pairs in which at least one subject fished, there were 5 pairs in which only one animal fished. Thus, there is no compelling evidence that testing-based social enhancement occurred. Consistent with local enhancement theory, our results suggest that, for bushbabies, the likelihood of interaction with a novel stimulus object is increased if the subject has previously observed a conspecific interacting with it. However, fishing behavior appears to include a familial component. Whether these filial differences indicate regional foraging differences or different responses to novel stimuli is yet to be determined.
ACKNOWLEDGMENT This research was supported in part by a Centers of Excellence grant awarded to the Department of Psychology, the University of Memphis, by the State of Tennessee.
Effects of Modeling
519
REFERENCES Cambefort, J. P. (1981). A comparative study of culturally transmitted patterns of feeding habits in the chacma baboon (Papio ursinus) and the vervet monkey (Cercopithecus aethiops). Folia Primatol. 36: 243-263. Chamove, A. S. (1974). Failure to find Rhesus observational learning. Z Behav. ScL 2: 39-41. Clark, A. B. (1978). Sex ratio and local resource competition in a prosimian primate. Science 201: 163-165. Clayton, D. A. (1978). Socially facilitated behavior. Q. Rev. BIOL 53: 373-392. Feldman, D. W., and Klopfer, P. H. (1972). A study of observational learning in lemurs. Z. TierpsychoL 30: 297-304. Fragaszy, D. M., and Visalberghi, E. (1989). Social influences on the acquisition of tool-using behaviors in tufted capuchin monkeys (Cebus apella). Z Comp. Psyc. 103: 159-170. Galef, B. (1988). Imitation in animals: History, definition, and interpretation of data from the psychological laboratory. In Zentall, T. R., and Galef, B. G. (eds.), Social Learning: Psychological and Biological Perspectives, Lawrence Erlbaum Associates, HiUsdale, NJ, pp. 15-23. Giraldeau, L. A., and Templeton, J. J. (1991). Food scrounging and diffusion of foraging skills in pigeons, Columba livia: The importance of tutor and observer rewards, Ethology 89: 63-72. Happold, D. C. D., and Happold, M. (1992). Termites as food for the thick-tailed bushbaby (Otolemur crassicaudatus) in Malawi. Folia Primatol, 58." 118-120. Huffman, M. A. (1984). Stone play of Macaca fuscata in Arishiyama B troop: Transmission of a non-adaptive behavior. Z Hum. EvoL 13: 725-735. Huffman, M. A., and Quaitt, D. (1986). Stone handling by Japanese macaques (Macaca fuscata): Implications for tool use of stone. Primates 27: 413-423. Myers, W. A. (1970). Observational learning in monkeys. J. Exp. AnaL Behav. 14: 225-235. Nash, L. T., Bearder, S. K., and Olson T. R. (1989). Synopsis of Galago species characteristics. Int. J. PrimatoL 10: 57-79. Olson, T. (1979). Studies on Aspects of the Morphology of the Genus Otolemur coquerel, 1859, Unpublished doctoral dissertation, University of London, London. Pallaud, B. (1984). Hypotheses on mechanisms underlying observational learning in animals. Behav. Proc. 9: 381-394. Presley, W. J., and Riopelle, A. J. (1959). Observational learning of an avoidance response. Z Genet. Psychol. 95: 251-254. Riopelle, A. J. (1960). Observational learning of a position habit by monkeys. J. Comp. PhysioL PsychoL 53: 426-428. Rosenson, L. M. (1972). Observations of the maternal behavior of two captive greater bushbabies (Galago crassicaudatus argentatus). Anita. Behav. 20: 677-688. Sherry, D. F., and Galef, B. (1990). Social learning without imitation: More about milk bottle opening by birds. Anita. Behav. 40: 987-989. Strayer, F. F. (1976). Learning and imitation as a function of social status in Macaque monkeys (Macaca nemestrina). Anirn. Behav. 24: 835-848. Thorpe, W. H. (1956). Learning and Instinct in Animals, Methuen, London. Tomasello, M., Davis-Dasilva, M., Camak, L., and Bard, K. (1987). Observational learning of tool-use by young chimpanzees. Hum. EvoL 2: 175-183. Visalberghi, E. and Fragaszy, D. M. (1990). Do monkeys ape? In Parker, S. T., and Gibson, K. R. (eds.), "Language" and Intelligence in Monkeys and Apes: Comparative Developmental Perspectives, Cambridge University Press, Cambridge, pp. 247-273. Watanabe, K. (1989). Fish: A new addition to the diet of Japanese macaques on Koshima Island. Folia PrimatoL 52: 124-131. Weehkin, S. (1970). Social relationships and social facilitation of object manipulation in Macaca mulatta. J. Comp. PhysioL Psycho/ 73: 456-460. Welker, C. (1976). Fishing behavior in Galago crassicaudatus. Folia PrimatoL 26: 284-291.