ABSTRACT. Spatial memory and foraging competition were investigated in three mother/offspring pairs of western lowland gorillas, Gorilla gorilla gorilla, using a ...
PRIMATES,41 (2): 147-160, April 2000
147
Spatial Memory and Foraging Competition in Captive Western Lowland Gorillas (Gorilla gorilla gorilla) STEPHANIE GIBEAULTand SUZANNEE. MACDONALD
York University
ABSTRACT. Spatial memory and foraging competition were investigated in three mother/offspring pairs of western lowland gorillas, Gorilla gorilla gorilla, using a naturalistic foraging task at the Toronto Zoo. Sixteen permanent food sites were placed throughout the animals' enclosures. All of the sites were baited and a pair of animals was free to visit the sites and collect the food. Five of the subjects collected the food with accuracy better than chance. Most of the subjects visited the sites using a pattern, and for half the subjects this was one of adjacency. The high accuracy of five of the subjects and the lack of a consistent adjacency pattern suggest that the animals did in fact use spatial memory. Furthermore, the gorillas tended to avoid visiting food sites that had been previously depleted by their partner. They also appeared to split their search of the enclosures, each visiting only a proportion of the food sites. This indicated that the animals were competing and altering their foraging behaviour based on the behaviour of their partner. Therefore, accuracy was recalculated to take this into account. When the sites depleted by either animal in a pair during a given trial were worked into the accuracy calculations for individual animals, three of the animals still maintained accuracy above chance. This suggests that the animals were not only able to remember which sites they had depleted, but those sites depleted by their foraging partner as well. Key Words: Western lowland gorilla; Spatial memory; Competition; Foraging behaviour.
INTRODUCTION The ability to remember and use information about the distribution of resources can be critical to an animal's foraging success (GARBER, 1989). For many species of rainforest primates, such as the western lowland gorilla, Gorilla gorilla gorilla, food resources, such as fruit or insects, are patchy and vary in both space and time (HLADIK, 1988). Those animals that gather more or better quality food in less time, thus using less energy, will have an evolutionary advantage (KAMIL & YOERG, 1982). The ability to remember the location of food sources, then, may provide a distinct advantage (MmToN, 1981). This ability to remember locations in the physical environment is termed spatial memory. The examination of spatial memory in nonhuman primates is a growing field of interest. Most of these studies used a modified free-range version of the radial arm maze initially introduced with rats (OLTON & SAMUELSON, 1976). Monkeys (GARBER, 1989; MACDONALD & WmKm, 1990; MACDONALD et al., 1994) and apes (TINKLEPAUGH, 1932; MENZEL, 1973; MACDONALD, 1994; MACDONALD & AGNES, in press) have both exhibited accurate memory for spatial location. In addition, many primates also exhibit efficient strategies when retrieving hidden food. For example, MENZEL (1973) found that captive chimpanzees, Pan troglodytes, used a 'least distance' strategy when collecting food from locations they had previously seen baited by the experimenter. In similar foraging experiments, yellow-nosed monkeys, Cercopithecus ascanius whitesidei (MACDONALD & WILKIE, 1990), marmosets (MACDONALD et al., 1994), gorillas (MACDONALD, 1994), and orangutans, Pongo pygmaeus abelii (MACDONALD & AGNES, 1999)
148
S. G1BEAULT& S. E. MACDONALD
also minimized the total distance travelled when visiting baited locations by moving from each site to the next adjacent site, thus increasing their overall foraging efficiency. Although field studies have suggested that western lowland gorillas use spatial memory while foraging, only one study has specifically examined this issue. MACDONALD(1994) tested an adult and a juvenile male gorilla, both zoo residents, using a foraging task at the Toronto Zoo. Eight containers were baited with food and placed in a circular pattern around the animals' enclosure. Individual subjects were then released into the enclosure and were free to collect the food. The adult male showed excellent accuracy and the juvenile, although not as accurate as the adult, also showed accuracy levels above chance. Both gorillas used an adjacency strategy in site choice almost exclusively in the foraging task, moving from one container to the next in a circular pattern. The results of MACDONALD'S (1994) experiment raises the question of how universal this cognitive ability is in western lowland gorillas. The diet and social system of western lowland gorillas in the wild would lead one to expect a sophisticated use of spatial memory. To exploit their varied diet, which contains seasonal fruits, stems, leaves, roots, and insects, gorillas face the challenge both of finding vegetative plant material and of spatially mapping and remembering the fruiting state of widely dispersed fruit plants (TUTINet al., 1991; BYRNE, 1995). Gorillas also live in social groups of usually six to eight animals. These family groups contain an average of three adult females, their immature offspring, and one, or sometimes more, mature males (HARCOURT & STEWART, 1989; TUTINet al., 1992). The adult females in a given group are probably unrelated and have not known each other from immaturity (HARCOURT, 1979). This form of group living may have led to greater competition among individuals for resources, providing an even stronger evolutionary advantage to those animals with more advanced spatial skills. The purpose of our study was to examine spatial memory in the western lowland gorilla further. We employed a foraging task at the Toronto Zoo using a captive gorilla group as subjects (based on MACDONALD, 1994). The animals were tested in mother/offspring pairs, because zoo staff felt that individual testing of the animals would cause excessive stress. Based on western lowland gorilla feeding habits and diet, and the seasonality in their natural habitat, we expected all the gorillas to use spatial memory in their foraging and to use an adjacency strategy to reduce foraging effort. In addition, western lowland gorillas in the wild tend to feed alone or in small clusters within food patches, with different animals feeding on separate foods and in different sized patches (REMIS, 1995). Therefore, we expected individuals to forage independently of each other throughout the study.
METHODS SUBJECTS
Six experimentally naive western lowland gorillas served as subjects. All six subjects were housed as a family group at the Toronto Zoo in Toronto, Ontario, Canada. Pair 1 consisted of a wild born female, Josephine (age 23), and her juvenile male offspring, Jomo (age 4); pair 2 consisted of a wild born female, Caroline (age 23), and her adult captive born female offspring, Catherine (age 12); pair 3 consisted of a wild born female, Samantha (age 23), and her juvenile female offspring, Sekani (age 5). The females each weighed approximately 100 kg, and stood 1.25 m tall. The two captive born juveniles each weighed approximately 35 kg, and stood approximately 1 m tall. The adult female in pair 3 was pregnant for the first five months of the study and after giving birth no longer participated in testing. The only member of the group that
Spatial Memory in Gorillas
149
did not participate in the experiment was the dominant adult male, who had previously participated in a similar study (MACDONALD, 1994). TESTING AREA
The experiment was conducted at the Toronto Zoo in the large gorilla exhibit shown in Figure 1. The exhibit consisted of two areas, the indoor enclosure (which measured 10 m wide by 30 m long by 4 m high) and the outdoor enclosure (which measured 10 m wide by 20 m long by 10 m high). The floor of the indoor area was concrete with a thick bed of wooden shavings and alfalfa hay. There were three wooden climbing structures, several metal and Plexiglas sleeping platforms, three large hammocks, a tire swing, and numerous climbing handles scattered across the walls. The public viewing side of the enclosure was constructed of thick Plexiglas. The other three walls were concrete. Two electrically-controlled shift doors led into the area, as well as a steel mesh door for keeper access. The outdoor enclosure had a natural soil and grass floor. There were four wooden climbing walls, a tire swing, and a 1 m by 2 m pool that was usually empty. All of the walls were steel mesh, except for two Plexiglas viewing windows on either side of the enclosure. The indoor enclosure was lit by natural light as well as overhead fluorescent lights. The outdoor enclosure was illuminated by natural light. As shown in Figure 1, the two enclosures were joined by a connecting hall, with one shift door into the
/
HALL
HOLDING AREA
-t
I
10X IX
4x 3X
INDOOR
O
0
0
EXHIBIT
(lID
,x
Fig. 1. Schematic drawing of the test area and the location of the food sites. X: The location and the number of each food site.
150
S. GIBEAULT& S. E. MACDONALD
indoor enclosure at one end, and two shift doors leading to the outdoor enclosure at the other end. Several internal doors led from the connecting corridor to the animals' holding area, which was not visible from the public viewing areas. The holding area was divided into eight rooms, in which the animals could be separated and fed their daily rations. All the animals were allowed free access to the indoor and outdoor enclosures immediately after the termination of a trial. After testing, a variety of vegetables and hay was scattered throughout the exhibit so the gorillas had access to food throughout the day. Both the indoor and outdoor viewing enclosures contained permanent food sites throughout the study. Each of the two enclosures contained 8 sites, for a total of 16 sites. As shown in Figure 1, the sites were scattered throughout the enclosures. Several sites were placed behind poles and around comers so that the animal could not see all the sites at once. Each enclosure was not visible from the other. Each site consisted of a 15 cm long by 15 cm wide section of metal tubing, painted green. The tubes were positioned upright on the ground against a wall or pole. We used one 2.5 g piece of dry cereal (Kellogg's Frosted Mini Wheats) to bait each site. The inside of every tube was cleared of all debris before baiting, and the food item placed such that it was not visible unless the animal was looking directly down into the tube. PROCEDURE
Trials took place in the morning, between 08:30 and 10:30, after the exhibits had been cleaned by the keepers, and they were usually conducted four days per week. All of the trials were monitored by at least one observer positioned in the public areas and most were also videotaped with an 8-mm Sony video camera (model number CCD-TR4). The gorillas ate their morning meal, which consisted of a variety of fruits and vegetables, in the holding area almost immediately before testing. The gorillas were kept in the holding area every morning before the experiment began, while the food sites were baited. The animals that did not participate on a given day remained in the holding area during the trial. Only one pair of animals was tested each day. After the sites were baited, a door to the indoor enclosure and one to the outdoor enclosures were opened, making access to both enclosures available. A pair of subjects was released through the internal holding doors into the hallway in the holding area, and was free to choose which enclosure to visit first. The exception to this was pair 1. Although the doors to the passageway were opened before each trial, this pair was released into the indoor enclosure through the left-hand electric shift door (see Fig. 1) because of the position of the rooms they occupied in the holding area. The subjects were then free to explore both enclosures and collect the food. The small size of the food items and the depth of the tubes made it impossible for the gorillas to see at a glance whether the site was baited, so it was quite easy to judge when a visit to a site occurred. We considered a site "visited" if a gorilla came within 30 cm of the site, with its eyes oriented to the tube, or if a gorilla put his or her hand or foot into the tube. We recorded the order in which each animal visited the sites and behavioral data, such as aggressive displays and fighting. Interobserver reliability was 100%. After the animals had visited all 16 food sites the trial was terminated. No time limit was placed upon the trial, although most ended within 30 min. When the gorillas had finished the task, the keeper called them back into the holding area and the shift doors were closed. With this basic design established, the experiment consisted of two phases. In the first phase all 16 sites were baited equally with one piece of cereal. In the second phase only the inside enclosure was used (due to adverse weather conditions). Only pairs i and 2 continued to participate in phase 2. During phase 2, the behaviour of each animal was more closely recorded. We
Spatial Memory in Gorillas
151
now noted all the visits made by one animal that seemed to be observed by the other animal in the pair. Although this was a subjective judgement, based mainly on the position of an animal's head and body, it would usually be obvious, such as when both animals stood together or one animal physically hit the other or vocalized during or immediately after a visit. Each visit to a site was classified as either a new visit (made to a baited site) or a revisit (made to a site that the subject had already visited). The total number of visits made (new visits plus revisits) was calculated. To analyze accuracy, we compared the number of new visits made by a subject with that would be made by a "random forager" (ROBERTS et al., 1993). A random forager makes each visit independently of any previous visits, with accuracy levels based on chance probabilities alone. For example, if a random forager has 16 food sites available, in its first 16 choices it will retrieve food from 10.30 sites by chance alone, leaving 5.70 of the food sites unvisited. To be "perfectly" accurate on this task, an animal should make only 16 visits, going to all 16 sites once, without revisiting previously depleted sites. The order of visits to sites was also recorded so that foraging patterns could be determined. A total of 36 trials was run with two of the three pairs (Josephine/Jomo and Caroline~Catherine). Samantha, the adult female in the remaining pair, was pregnant during the study, so only eight trials were run with the Samantha/Sekani pair.
RESULTS The results are divided into three sections -accuracy, foraging patterns, and competition- for ease of analysis and interpretation. The accuracy and foraging pattern calculations were carried out with the assumption that the subjects behaved independently of their foraging partner. This allowed us to examine and compare the performance of individual animals. However, in the competition section, we reexamined the data with the assumption that the two animals in a foraging pair interacted. This allowed us to examine the impact of competition on foraging efficiency. ACCURACY
Phase 1 Typically, in tests of spatial memory, the standard measure of accuracy involves examining the number of food items an animal collects (number of new visits) out of the total number of food items available (number of sites), and then comparing this value to chance probabilities (T1LLE et al., 1996). However, we noted that our subjects often did not make 16 visits and therefore could not have visited all of the available sites. We used a two-tailed t-test to determine if the mean number of visits made by each animal per trial, in those trials it participated in, was equal to 16. Only the juvenile animals made a mean number of visits that was not statistically different from 16 (see Table 1). Therefore, four of the animals did not make enough visits to deplete all of the sites. Next, accuracy was calculated by comparing the number of new visits made by each subject with the number that would be made by a random forager (ROBERTS et al., 1993). Because four of the animals rarely made enough visits to deplete all 16 sites, each animal's accuracy had to be evaluated based on its own performance. A different number of random visits was used for comparison with each animal's actual performance (Table l, "Random forager criterion number"). This random forager criterion number was chosen to be as high as possible while still
152
S. GIBEAULT& S. E. MACDONALD
Table 1. Phase 1 accuracy data. Mean No. new Mean No. of MeanNo. visits Randomforager visitsmade in Accuracycompared Subject visits ( s d ) comparedto 16 criterionnumber criterionNo. (sd) to random Josephine 4.71 (2.70) to.o5(2).)3=15.6** 6 5.40 ( 0 . 8 9 ) to.o50),4=3.76 ** Jomo 14.45 (6.76) to.o5(2),1o=0.76ns 16 14.43 ( 1 . 9 0 ) to.05o),6=5.74 ** Caroline 12.00 (4.60) to.o5(2),14=3.45** 14 13.13 ( 1 . 1 2 ) t0.050),7=9.06 ** Catherine 8.57 (4.55) t0.05(2),13=6.10** 10 9.86 ( 0 . 3 8 ) t0.050),7=15.74 ** Samantha 3.50 (1.52) t0.osc2Ls=20.19** 3 2.67 ( 0 . 5 0 ) to.o50),2=0.08 ns Sekani 12.25 (6.41) to.05(2).7=1.66ns 9 8.28 ( 0 . 9 5 ) t0.05(1),6=3.44 ** ** p
