Read Aloud the Text Content

This audio was created by Woord's Text to Speech service by content creators from all around the world.

Text Content or SSML code:

At the time of recovery, the cachers moved more items out of the observed tray than out of the other tray for recaching when the corresponding observer was present. This recaching behaviour was not a response to the behavioural cues of the observers because the recaching preference did not emerge when they recovered in the presence of control birds who had previously watched different individuals’ cache food in the same tray. These results suggest that scrub-jays can keep track of particular individuals that were present at the time of caching and alter their recaching behaviour to protect the caches from being pilfered. Like ‘‘theory of mind’’, where the knowledge about other’s mind is related to that of one’s own mind (Gordon 1995), animals with an understanding that others gain knowledge by seeing are also likely to understand that act of seeing would lead to its own knowledge. One way to study concurrent metacognition judgement is to use the foraging task designed by Call and Carpenter (2001) that requires the subjects to search for a reward hidden by an experimenter. In a typical trial, the subjects observe the experimenter place a reward item in one of the multiple tubes and then they are allowed to choose the tube that they thought contained the reward. Before making a choice, the subjects can look into one or more tubes to check the location of the reward. Some trials are given under a high-uncertainty condition by, for example, placing a visual barrier between the subjects and the tubes during the baiting so that they could not see which tube was being baited or by inserting a delay period prior to the choice phase. In such experiments, it is predicted that the subjects would look into the tubes more often in the high-uncertainty condition than in the control condition if they can monitor their own knowledge state. Because the foraging task uses searching behaviours that are already in the animals’ behavioural repertoire, there is usually no need to train the subjects to use the uncertain response, i.e. looking into the tubes. This has an advantage over studies that use arbitrary visual cues, such as Hampton (2001), which are sometimes criticised for allowing the possibility of reinforcement learning to occur during the training process (Le Pelley 2012). The current experiment used a modified version of Call and Carpenter’s (2001) foraging task to investigate information-seeking behaviour of western scrub-jays. The birds watched while an experimenter hid a bottle cap containing food in one of the two aluminium tubes. They could, then, approach the tubes and choose one by entering within 10 cm radius of the baited end of the tube. The closer opening of the tubes had peepholes, which enabled the birds to check the presence of the bottle cap before making the choice. The birds’ uncertainty levels were manipulated by varying the visibility of the baiting condition, inserting a delay after baiting, and inducing confusion by moving the bait. Their accuracy in choosing the baited tube and use of looking behaviour were measured and compared across conditions. If the birds can monitor and control their knowledge during memory retrieval, then they should look into the tubes when they do not know the location of the bait and are likely to err. Experiment 1 In Experiment 1, there were two baiting conditions, designed to create different knowledge states in the subjects. During the seen condition, the birds could observe the baiting process and so should be certain of the location of the bait without looking into the tubes. During the unseen condition, the jays could not observe the baiting because an opaque divider was inserted between the birds and the tubes. The probability of choosing the correct tube based on this unseen condition alone would be at chance but could be improved by looking into the tubes and checking the location of the bait before making a choice. Methods Subject and experimental set-up The subjects were five adult western scrub-jays (four males and one female, identified by letters A–E). All of the birds were originally collected in 2007 as wild nestlings from California, USA (Federal Fish and Wildlife permit OMB No. 1018-0022; State of California Collecting permit SC8592). They were hand-raised at the University of California, Los Angeles, and imported to the UK via air transport at 2–3 months of age (US Department of Agriculture Animal and Plant Health Inspection Service certificate S 088337; UK DEFRA import permit SIL/2007/33/ 02). Since then, the birds have been housed at Clayton’s Corvid Cognition Facility in the Sub-department of Animal Behaviour at the University of Cambridge, which consists of a series of indoor (3 9 2 and 2.4 m high) and outdoor (6 9 3 and 3 m high) enclosures specifically designed for corvids. Each aviary contained two to eight individuals with ad libitum access to food and water, and various forms of environmental enrichment. For the duration of the experiment, the birds were brought into experimental rooms where they were kept in pairs in cages measuring at least 2 9 1 9 1 m under 12-h dark/light cycle at 21 ± 2 C. Their maintenance diet consisted of dog biscuits, boiled eggs, cheese, fruits, vegetables, and Mazuri pellets. They had free access to water throughout the experiment. The experiment was conducted between November 2009 and April 2011 under the UK Home Office project licence PPL 80/1975. Three of the subjects, birds A, B, and C, were used in the previous metacognition study (Watanabe et al. 2014). Before the commencement of each session, the bird was individually placed into the observation cage, measuring 1 9 1 9 1 m. This observation cage was adjacent to the experimental cage (see Fig. 1). Both cages were visually occluded from other birds kept in the same room. A Plexiglass board acted as a divider between the observation cage and the experimental cage. This divider was transparent for the training and for the seen trials of the test, enabling the bird in the observation cage to see the activity in the experimental cage, and opaque for the unseen trials, preventing the birds to see inside the experimental cage. The experimental cage contained two aluminium 3 9 3 cm square tubes of which there were two types—short tubes (30 cm) used for the first two phases of training and longer tubes (45 cm) used for the last part of the training and testing. The opening of the long tube, facing the observation cage, was covered with white plastic sheet with a hole of diameter 1, 1.5, or 2 cm. These holes required the birds to bend down to look through them. The openings of the short tubes were uncovered and completely open so that birds could look through them without much effort. The other end of the tube, facing away from the observation cage, was baited with a plastic bottle cap containing either a mealworm or a wax worm, depending on the bird’s preference. To prevent the bird from making more than one choice of the tube, a small board (50 9 50 cm) was placed between the ends of the two tubes. The bird’s behaviour in the experimental cage was monitored by the experimenter in the same room with a video camera so that she could end the trial as soon as the bird made a choice. It was also video-recoded with a Multicam Surveillance Camera System (GeoVision GV-900, London, UK) for later coding of the bird’s behaviours by the experimenter and another coder to check inter-observer reliability. Procedure The birds were first habituated to the experimental cage and the metal tubes by retrieving food from a bottle cap that was placed at various locations around (but not inside) the tubes in the experimental cage. Once the birds were retrieving the food within 1 min on all eight trials within a trial block, the training phases began. All five birds passed this criterion in the first block of trials that was presented to them. Training 1 Two short tubes were placed in the experimental cage at 60 degrees (see Fig. 1a). At the beginning of each trial, the experimenter placed the bottle cap containing a worm at the end of one of the tubes, while the bird watched from the observation cage. The divider was then removed and the bird could enter the experimental cage and look for the food. The trial ended when the bird found and consumed the food or after 3 min passed. The position of the baited tube was pseudo-randomised so that the same tube was not baited on more than two trials consecutively. Two blocks of eight trials were given to each bird per day. When the bird found the food within 1 min on all trials within a block, the next part of the training began. This criterion was reached after 3.0 ± 1.0 (mean ± SD) blocks of trials. Training 2 Training 2 was similar to Training 1 except that the birds could only make one choice. A choice was defined as entering within 10 cm diameter of the further end of the tube. The bird was given time to consume the food if it had chosen the correct tube. If it chose the incorrect tube, however, then the bait was removed by the experimenter and it immediately went back to the observation cage. The birds received three blocks of eight trials per day. They moved onto the final part of the training once they chose the correct tube on seven or more out of eight trials on two consecutive blocks. This criterion was reached after 9.4 ± 3.8 blocks of trials.