|dc.description.abstract||Few scientific fields spark the imagination of the public quite like the study of animal intelligence. For centuries, humanity’s ability to think and to create have represented characteristics that many have cited as being “what separates us from the animals”. However, through the careful studies of a number of key taxa, the basis for this view of human exceptionalism has been heavily eroded. Tool use, for example, was found in chimpanzees, while accounts of intelligent behaviours in dolphins were frequently reported. Humanity’s apparent uniqueness became just one end of a continuum of intelligent behaviours displayed by a wide range of species. Nonetheless, it is only recently that certain groups of birds have been elevated to the realm of intelligent fauna. These investigations have been dominated by two taxa: the parrots (Order: Psittaciformes) and the corvids (Family: Corvidae).
Many facets of high cognition have been described in the corvids, including tool use, social learning and theory of mind. The knowledge garnered from studies of the corvid family has been promising, yet it remains limited by the small number of species. This study attempts to complement and expand current knowledge by undertaking carefully designed experiments on a little-studied but eminently suited Australian species, the Torresian crow (Corvus orru), in a suburban environment.
This thesis is comprised of four sections, detailing the research project conducted between 2012 and 2016 in Brisbane. Within these four sections are nine chapters. The first two chapters provide an overview of the nature of the corvids and of the Superfamily Corvoidea, before detailing the potential consequences of conducting behavioural research entirely in a controlled setting, describing the environment in which the study was conducted as well as the study species and sites.
The third chapter provides an in-depth review of trends in the literature regarding cognition in the corvids. This review found that the research was heavily biased towards studies on animals held in captivity, whether they be hand-reared or wild-caught birds. The study also found that, although the corvids are a near-worldwide group of approximately 120 species, almost all of the scientific literature surrounding the group has been conducted on just 14 species, almost entirely in Europe and North America. Based on the results of this review, it appeared that an intensive study into a species from a different region and conducted on wild animals would be beneficial to the field.
The fourth chapter documents the methods devised and tested in the suburban environment, and discusses their level of success. Conducting experiments originally designed for use in a controlled laboratory setting on free-living animals in the wild required significant modifications to apparatus and methods. The need to attract crows while minimising interference from the many other bird species present in the area was a constant concern, and required the trialling of several kinds of bait. Apparatus were significantly modified several times throughout the study. This chapter documents these modifications to provide an extensive list of what does and does not work when studying learning behaviours of wild birds.
The second section includes four chapters, each of which is a piece of original research written in the form of a publication. The first of these (already been published in the journal Ethology) investigated the extent to which Torresian crows displayed neophobic behaviours in the wild, compared with three other corvoid birds common in suburban Brisbane: the Australian magpie (Cracticus tibicen), grey butcherbird (Cracticus torquatus) and pied currawong (Strepera graculina). It was found that Torresian crows, like overseas corvids, are highly neophobic, being both significantly delayed in attaining food and displaying more neophobic behaviours when a novel object was present. The Torresian crow was affected by the novel object far more than any of the other three species studied.
Associative learning and shape discrimination were tested on five wild Torresian crows, three of which were successful. Numerosity, a form of rule-learning, was then tested in two of the crows that succeeded in the shape discrimination task. No crows were successful in distinguishing between different quantities.
Crows and other urban species were also exposed to Mirror-Image Stimulation (MIS), in light of several other studies on MIS and Mirror Self-Recognition in corvid species. Crow behaviour was significantly affected by the presence of a reflective surface, and this behaviour was relatively consistent between individuals. Crows behaved in a similar fashion to two native Cracticids, the Australian magpie and pied currawong, and two invasive species common in urban areas, the common myna (Acridotheres tristis) and spotted dove (Spilopelia chinensis). Other native species studied acted in a dissimilar way. Both territorial crows and magpies reacted in a very dissimilar way to their usual response to seeing a conspecific in their territory.
Finally, crows and pied butcherbirds (Cracticus nigrogularis) were tested for their ability to solve a horizontal string-pulling problem and to discriminate between two parallel and two crossed strings, only one of which would result in a food reward. Only one crow spontaneously solved the problem, but then went on to fail the string discrimination task. Only two Torresian crows were able to distinguish between the parallel strings and none were capable of solving the crossed string experiment, producing similar results to what has been recorded in New Caledonian crows (Corvus moneduloides). Results suggested that the majority of crows tested were simply pulling a string without considering whether it was attached to the food reward. Conversely, pied butcherbirds excelled in this task, not only solving both tasks to a much greater degree than crows but also doing so without training.
This thesis has attempted to complement the existing literature on corvid cognition by studying an Australian species intensively in the wild environment. In doing so it has established that the behaviour of Australian crows was similar to their overseas cousins in a number of ways, including neophobia, associative learning and string discrimination. However, it is likely that the uncontrolled nature of wild experimentation may have hampered crow performance, and more controlled investigations may yield better results. On the other hand, this thesis has unveiled the butcherbirds (Cracticus spp) as being a group performing particularly well in cognitive tasks. It is quite possible that this genus could provide a third group of highly-intelligent birds, in addition to the corvids and parrots, for experimentation. Further investigations into this group could prove particularly fruitful.||