Recently, empirical studies combining both an ethological and comparative psychology approach have revealed that some species of birds, particularly corvids, are remarkably intelligent, both in the social and technical domain. evidence has come from a range of species within this family, leading to claims that individual corvid species have evolved particular adaptive specialisations rather than possessing a more generalised intelligence. This thesis investigates some of the key components of the social and physical cognition of one species, the rook, in order to gain a more complete picture of its intelligence. The experiments in this thesis show that rooks possess a number of core components of social intelligence, such as individual recognition, behaviour reading and coordination, that are not necessarily cognitively complex but that may be used to solve social problems; suggesting caution must be used when speculating as to whether they possess more advanced social intelligence. On the other hand, it appears that rooks often solve physical problems using cognitively advanced and generalised causal reasoning, rivalling the great apes in their physical intelligence. Rooks were able to insightfully solve novel problems by using tools and demonstrated an understanding of object properties and unobservable folk-physical forces. The experiments introduce novel methodologies that may prove useful in further revealing how rooks and other animals ‘see’ the world from a cognitive perspective.

In this thesis I explore assumptions of current models of cooperative breeding and reproductive skew. In the highly social Damaraland mole-rat I examine the questions (1) What benefits do breeders gain from subordinates? (2) Why do helpers help? and (3) Why don't subordinates breed? In answer to (1), benefits to breeders appear to accrue primarily through communal foraging rather than through the provision of alloparental care to young or the reduction of workload. Breeding females contributed most to care of pups, but worked at high rates, even during energetically costly reproductive activities. Subordinate number did not increase litter sizes or pup survivorship. In answer to (2), all cooperative behaviour of subordinates could be explained by simple maximisation of current individual fitness. Mole-rat work levels simply reflected their weight, rather than breeding status, potential kin-selected benefits, or energetic needs of the breeding female. Experimental investigation of communal food storing showed that food storing reflects mole-rats' own weight and hunger state and is insensitive to the hunger state of other colony members. Participation in colony defence was consistent with defence of reproductive status. Attacks on intruders were carried out almost exclusively by breeding mole-rats who attacked primarily same-sex intruders, and subordinates typically did not participate. In answer to (3), there was some evidence that subordinate males pose a reproductive threat to dominant males, but only slight evidence that dominants may control their reproduction. Among females, I found no evidence that dominant control limited the reproduction of subordinate females, and strong evidence that reproductive skew was due to inbreeding avoidance.

This thesis investigates the roles of the amygdala and hippocampus in Pavlovian conditioning. Three questions are addressed. First, what is the involvement of discrete sub-nuclei of the amygdala in Pavlovian conditioning? Second, what molecular processes accompany the retrieval of Pavlovian associations? Third, what changes in gene expression are induced during the acquisition of a Pavlovian association? The basolateral region of the amygdala (BLA) is widely believed to represent the site of CS-US association in Pavlovian conditioning. Recent experiments have however suggested that the BLA may play a more restricted role in Pavlovian conditioning, and that other nuclei and within and outside the amygdala can support the formation of Pavlovian associations. In a first series of experiments the effect of BLA lesions on aversive Pavlovian conditioning were investigated. Lesions of the BLA were found to disrupt conditioning to both discrete and contextual stimuli, as assessed by conditioned freezing, but the effect of BLA lesions on contextual conditioning was ameliorated by additional training. In a second group of experiments the effect of amygdala lesions on the ability of Pavlovian cues to motivate instrumental responding (Pavlovian to instrumental transfer, PIT) was studied. Lesions of the BLA were found to be without effect on PIT, but lesions of the central nucleus of the amygdala (CeN) abolished this effect, as did lesions of the nucleus accumbens (NAcc) core. These results demonstrate that there are dissociable roles of the BLA and CeN within Pavlovian conditioning, and suggest that the CeN and NAcc core interact in PIT.

I studied the reproductive behaviour of the great tit (<i>Parus major)</i>, looking in detail at the factors influencing parental investment decisions made before and after hatching. Using playback experiments and brood-size manipulations, I found that nestling begging behaviour and partner provisioning rates independently influenced the rate at which food was brought to the nest during chick rearing. In general, females has more control over parental provisioning rates than either their partner or their offspring. Males, in turn, had greater control over food delivery rates than offspring. During egg laying, females appeared to optimise egg investment according to their own provisioning capabilities, measured via provisioning levels to a standard brood size. By removing eggs as they were laid, and therefore inducing females to lay more eggs than they intended, I found a trade-off between egg-laying and chick-provisioning levels. The more eggs a female laid, the less able she was to care for the young after hatching. Females may benefit from advertising their reproductive strategies, and this possibility was explored with a cross-fostering experiment. I found that variation within the female breast stripe revealed information regarding alternative female reproductive strategies within the current reproductive attempt. Females with small, narrow breast stripes invested more pre-hatching; they laid heavier eggs and their chicks grew faster when reared by foster parents. By contrast, females with large, wide breast stripes invested more post-hatching; the chicks they reared grew faster. Additionally, females laid heavier eggs if they were paired with a male that had a wide stripe, which may reflect differential allocation for indirect benefits.

Behavioural assays were used to investigate how flies' <I>(Drosophila melanogaster</I>) responses to a chemical were influenced by pre-exposure to the same chemical. A modified trap assay, a T-maze olfactometer and several assays of mating behaviour were used to investigate the response of <I>D. melanogaster</I> to various chemicals. Menthol was used most extensively for pre-exposure and behavioural testing. The trap assay was used to produce dose-response curves to menthol for flies that had been pre-exposed, and flies that had never been exposed to menthol before testing. There was a threshold pre-exposure dose that produced a behavioural change. Pre-exposure to mentholic food resulted in reduced aversion to some concentrations of menthol. In the T-maze, flies normally were averse to menthol, but flies pre-exposed to mentholic food were less averse and sometimes responded positively to the odour of menthol. Exposure to mentholic food in only the larval stage did not change adult behaviour in the T-maze or trap assays. There was no evidence for pre-imaginal conditioning. Evidence from the trap assay suggested that menthol contamination on the puparium was sufficient to change adult behaviour. Exposing adult flies to menthol did result in behavioural changes. Even brief exposure to menthol for just an hour before testing was enough to change behaviour. Fly age influenced behaviour in both the T-maze and trap assay. This factor may help to explain discrepancies in earlier investigations of olfactory conditioning and pre-imaginal learning in <I>D. melanogaster</I>. Menthol exposure appeared to change the dose-response curves to chemicals other than menthol in the T-maze. I discuss what mechanisms might underlie this behavioural change and suggest work to investigate this issue further.

The aim of this study was to understand the function of song and its role in the speciation process in a duetting Neotropical passerine, the gray-breasted wood-wren (<i>Henicorhina leucophrys).</i> This species lives in the dense understory of mid-elevation cloudforests throughout the Andes and up into southern Mexico. Gray-breasted wood-wrens sing sex-specific songs, either as solos or as complex duets, with the male and female of a mated pair alternating their songs with variable degrees of coordination. Males sing more than 50 different song types, while females sing at least 15-20 song types. Duets are initiated by males and females at equal rates. Using playback experiments simulating intrusions by paired and single birds, I show that males and females sing to advertise their partner’s mated status and to repel potential rivals. These results show that duets reflect conflict between the sexes rather than cooperation between the pair. I studied song variation in two subspecies of the gray-breasted wood-wren which occupy different habitats in the Ecuadorian Andes. I found that the subspecies sing significantly different songs and showed that these differences are driven by acoustic differences in their preferred habitats – loud cicada choruses in the habitat occupied by one subspecies have caused that subspecies to sing at a lower frequency in order to be heard over the noise. Where the two subspecies come into secondary contact, their songs are more different from each other than in areas where they do not meet. This suggests that there is a cost to mating with the wrong subspecies and that selection has driven the songs apart in order to prevent such an occurrence. This result provides the first convincing case of character displacement in bird song and strong evidence that song divergence acts as a barrier to gene flow between these two subspecies.

In this work, the social dynamics and the dynamics of coalitionary behaviour observed in a semi-free provisioned group of <i>Cebus apella</i> were used to explore predictions of socio-ecological models regarding the formation and organisation of social groups, and the hypothesis that complexity of social life is pressure to enlarged brain. In accordance to socio-ecological models but in contrast to accepted classification of <i>C. apella, </i>the group was not matrilinally organised, consequence of the prevalence of scramble type of food competition. The group hierarchy had a clear alpha male, an alpha female, and was age-size based for the rest of the group. The rank of infants did not correlate with the rank of their mothers. Again indicating the non-matrilineal organisation, females did not tend to be closer to each other or to their relatives, but rather the alpha male and dependent infants were most the attractive members of the group. Grooming was defined as a multifunctional behaviour, indicating both tolerance and affiliation for different dyads. Grooming occurred down the hierarchy and the alpha female was the most active groomer. More importantly, amount of grooming exchanged between dyads did not influence coalitionary behaviour. Although infrequent, opportunistic conservative and ambivalent coalitions did occur in conflicts involving adults, suggesting its strategic use. In the Main group, most of third party interference in aggressive events was protective, with the alpha male supporting immatures against other adult males. It is suggested that a mild form of male-male sexual competition influences the social organisation of <i>C. apella</i>, with females clustering around an alpha male to obtain benefits from his protective abilities and tolerance. Finally, it is suggested that for a large brained species with wide dietary variability as <i>C. apella</i>, social organisation may be variable and flexible, and to reflect a continuum rather than discrete categories.

In common with most squid fisheries, the Falkland Islands fisheries for <i>Illex argentinus </i>and <i>Loligo gahi </i>are characterised by unpredictable, inter-annual variations in distribution and abundance. In recent years, much research effort has been directed at understanding the influences of physical oceanographic variability, such as sea surface temperature, on the recruitment, distribution and abundance of these species. This study considers the influence of predation and food availability on the recruitment, distribution and abundance of these species. Cannibalism (defined as intraspecific predation) is a feature of cephalopod feeding biology and impacts on natural mortality and population dynamics. The research provides the first comprehensive account of cannibalism in two commercially exploited squid species in the southwest Atlantic. The diets of <i>I.</i><i> argentinus </i>and <i>L. gahi </i>were examined over a wide spatial area and over several years and results showed that this behaviour is related to population density and the availability of food. This study also quantified cannibalism in <i>I.</i><i> argentinus </i>and <i>L. gahi </i>as a major source of natural mortality and has assessed the impact on the population dynamics of these species. This links between food availability at the base of the food chain and the distribution and abundance of <i>I.</i><i> argentinus </i>and <i>L. gahi </i>were examined using satellite-derived chlorophyll-<i>a</i> data imaged by Sea-viewing Wide Field-of-view Sensor (SeaWiFS). Results showed that phytoplankton biomass appeared to be related to <i>I.</i><i> argentinus </i>biomass during the period of the fishery. The addition of information on food availability using SeaWiFS technology may improve current forecasting models for <i>I.</i><i> argentinus </i>and <i>L. gahi. </i>Within season estimates of the natural mortality due to cannibalism may be used to make real-time modifications to the target escapement of the <i>I.</i><i> argentinus </i>and <i>L. gahi </i>fishery populations.

In this thesis, I use nestling provisioning in cooperatively breeding chestnut-crowned babblers as a measure of cooperative behaviour, and, using this measure, I investigate how and why members of the same group vary in their investment in young. Previous studies of investment have routinely used provisioning rates, rather than biomass of prey fed to young, as a proxy for investment. I show that provisioning rate closely reflected biomass delivered to the nest, making it a good measure of investment in provisioning behaviour in chestnut-crowned babblers. I explore how different components of provisioning effort (rate, prey size and type) traded off against each other in response to changes in brood demand. I show that contributions to care were strongly influenced by the interaction between the sex and age of helpers, and that observed patterns of care could only be explained by considering the costs as well as the benefits of care. Using brood size manipulations, I highlight that the effects of group size on helper contributions may well extend beyond changes in brood demand. I show that breeders and helpers did not follow the same investment rules. In particular, breeding females invested least in provisioning young. I discuss whether this strategy arose because resources were allocated to other components of reproduction. Finally, I use manipulations of provisioner:nestling ratios to demonstrate that helpers did enhance the fitness of the current brood and could therefore gain direct and/or indirect fitness benefits from provisioning young.

I use simulated pedigree and fitness data to test the statistical power of a regression model proposed by Ballou (J. Heredity, 88, 169-178, 1997) to detect inbreeding depression and purging. Finding the model to be lacking in power when used to analyse typical zoo pedigrees, I develop an alternative, more powerful model. I use both of these models to investigate the effects of inbreeding in 136 zoo populations, encompassing 109 species of mammals, birds, reptiles and amphibians. A significant cross-population trend in inbreeding depression is detected, as is a cross-population trend of purging in those populations which showed negative effects of inbreeding. The average change in inbreeding depression due to purging is < 2%, however, suggesting that fitness benefits are rarely appreciable. The study re-emphasises the necessity to avoid inbreeding in captive breeding programs and shows that purging cannot be relied upon to remove deleterious alleles from zoo populations. The severity of inbreeding depression appears to vary among taxa but few predictors of a population’s response to inbreeding are found. The models are most likely to detect inbreeding depression in large populations, i.e. in populations in which their statistical power is maximised. By analysing founders’ contributions to inbreeding, I find that random founder effects play a part in determining whether a population suffers from inbreeding depression. I also show that inbreeding depression may have differing effects on sexes but find no evidence of a consistent sex-bias. Susceptibility to inbreeding depression may therefore depend on a complex interaction of genetic, environmental and stochastic factors.