Abstract: Keas, a species of parrots from New Zealand, are an interesting species for comparative studies of problem solving and cognition because they are known not only for efficient capacities for object manipulation but also for explorative and playful behaviors. To what extent are they efficient or explorative, and what cognitive abilities do they use? We examined how keas would solve several versions of artificial-fruit box problems having multiple locks. After training keas to remove a metal rod from over a Plexiglas lid that had to be opened, we exposed the birds to a variety of tasks having two or more locks. We also introduced a preview phase during which the keas had extended opportunity to look at the tasks before the experimenter allowed the birds to solve them, to examine whether the preview phase would facilitate the birds' performance on the tasks. In a large number of tests, the keas showed a strong trend to solve the tasks with no positive effect of previewing the tasks. When the tasks became complex, however, the keas corrected inappropriate responses more quickly when they had had chance to preview the problems than when they had not. The results suggest that the keas primarily used explorative strategies in solving the lock problems but might have obtained some information about the tasks before starting to solve them. This may reflect a good compromise of keas' trial-and-error tendency and their good cognitive ability that result from a selection pressure they have faced in their natural habitat.

Abstract: Learning about food palatability from watching what conspecifics eat might be one of the advantages of group living. A previous study investigated whether group members' presence or eating activity account for social facilitation of eating of foods never previously tasted. Capuchins encountered novel colored foods when (1) alone (Alone condition) or (2) with group members visible in the nearby cage (Group-present condition) or (3) with group members present and eating a familiar food that had not been colored (Group+food condition). Social facilitation of eating occurred when group members were eating, despite the difference in color between the familiar food eaten by them and the novel food presented to the experimental subject. To clarify what subjects learnt from group members when social facilitation occurred, we further analyze here the data from the previous study. The number of visual exposures to the colored novel food (as a group member) correlated with increased consumption of that novel food when encountered later (as experimental subject). In contrast, the number of times that an individual fed on the familiar food (as a group member) did not decrease its consumption of novel food (as experimental subject). Therefore, capuchins (1) habituated to the colors of the novel foods, and (2) did not take into account that seeing group members eating a food does not provide information about the palatability of a differently colored food. Since social facilitation of eating occurs when foods do not match in color, at least in capuchins, social facilitation of eating should not be considered as a way of learning about a safe diet, but rather as a way of overcoming neophobia.

Abstract: Like other corvids, food-storing ravens protect their caches from being pilfered by conspecifics by means of aggression and by re-caching. In the wild and in captivity, potential pilferers rarely approach caches until the storers have left the cache vicinity. When storers are experimentally prevented from leaving, pilferers first search at places other than the cache sites. These behaviours raise the possibility that ravens are capable of withholding intentions and providing false information to avoid provoking the storers' aggression for cache protection. Alternatively, birds may refrain from pilfering to avoid conflicts with dominants. Here we examined whether ravens adjust their pilfer tactics according to social context and type of competitors. We allowed birds that had witnessed a conspecific making caches to pilfer those caches either in private, together with the storer, or together with a conspecific bystander that had not created the caches (non-storer) but had seen them being made. Compared to in-private trials, ravens delayed approaching the caches only in the presence of storers. Furthermore, they quickly engaged in searching away from the caches when together with dominant storers but directly approached the caches when together with dominant non-storers. These findings demonstrate that ravens selectively alter their pilfer behaviour with those individuals that are likely to defend the caches (storers) and support the interpretation that they are deceptively manipulating the others' behaviour.

Abstract: Little is known about how a prey species' cognitive limitations might shape a predator's prey-capture strategy. A specific hypothesis is investigated: predators take advantage of times when the prey's attention is focussed on its own prey. Portia fimbriata, an araneophagic jumping spider (Salticidae) from Queensland, is shown in a series of 11 experiments to exploit opportunistically a situation in which a web-building spider on which it preys, Zosis genicularis (Uloboridae), is preoccupied with wrapping up its own prey. Experimental evidence supports three conclusions: (1). while relying on optical cues alone, P. fimbriata perceives when Z. genicularis is wrapping up prey; (2). when busy wrapping up prey, the responsiveness of Z. genicularis to cues from potential predators is diminished; and (3). P. fimbriata moves primarily during intervals when Z. genicularis is busy wrapping up prey. P. fimbriata's strategy is effective partly because the wrapping behaviour of Z. genicularis masks the web signals generated by the advancing P. fimbriata's footsteps and also because, while wrapping, Z. genicularis' attention is diverted away from predator-revealing cues.

Abstract: Locating food and refuge is essential for an animal's survival. However, little is known how mammals navigate under natural conditions and cope with given environmental constraints. In a series of six experiments, I investigated landmark-based navigation in free-ranging Columbian ground squirrels (Spermophilus columbianus). Squirrels were trained individually to find a baited platform within an array of nine identical platforms and artificial landmarks set up on their territories. After animals learned the location of the food platform in the array, the position of the latter with respect to local artificial, local natural, and global landmarks was manipulated, and the animal's ability to find the food platform was tested. When only positions of local artificial landmarks were changed, squirrels located food with high accuracy. When the location of the array relative to global landmarks was altered, food-finding accuracy decreased but remained significant. In the absence of known global landmarks, the presence of a familiar route and natural local landmarks resulted in significant but not highly accurate performance. Squirrels likely relied on multiple types of cues when orienting towards a food platform. Local landmarks were used only as a secondary mechanism of navigation, and were not attended to when a familiar route and known global landmarks were present. This study provided insights into landmark use by a wild mammal in a natural situation, and it demonstrated that an array of platforms can be employed to investigate landmark-based navigation under such conditions.