In the UK, translocation is increasingly being used to resolve conflict between great crested newt (Triturus cristatus) conservation and land development. Due to a lack of objective study on the translocation procedure, there remains little evidence of the success of employing this strategy despite widespread implementation. Reviews of translocations highlight the need for case studies that include longer term pre and post translocation monitoring. To allow redevelopment of the Gartcosh Industrial Site, the decision was taken to translocate the resident great crested newt population to the purpose built Gartcosh Nature Reserve around the periphery. This provided an opportunity for in-depth analysis of the largest project of its kind in Scotland. This project was designed to test the effectiveness of translocation in producing a self-sustaining, viable population. The key aims were: to ascertain if the population was successfully re-established in the receptor site at a level comparable with the donor site; to assess whether the newly created habitat was suitable for supporting a population of great crested newts; to determine what constitutes a successful translocation and how best to achieve this within the Scottish context. The following points summarise the projects findings: • Simple counts of adults are being maintained at a level comparable to or greater than pre translocation counts. • The favourable status of the adult population is supported by a capture-mark-recapture study. Population estimates are on a par with numbers of adults translocated to the Gartcosh Nature Reserve. • Juvenile lifestages indicate declines. Further monitoring is required to determine if this is an effect of the translocation or a natural fluctuation. • Survival rate of adults is measured at 43%. • There is significant recruitment of ‘new’ adults. • Good quality terrestrial and aquatic habitat has been produced, with an overall loss of land and pond surface area but increased number of ponds. • Increased individual growth rates of adults are indicative of a habitat capable of meeting adult resource requirements. • The nature reserve is internally fragmented into zones preventing movement through the site and is isolated within the wider landscape. For the population to be viable, connectivity requires improvement. • To ascertain long-term success of the Gartcosh translocation it is recommended that post monitoring extends beyond simple adult counts and continues capture-mark-recapture study specifically within the Railway Junction area. • Guidelines have been produced detailing best practice in translocation, monitoring and habitat creation.

Animals often respond differently to the same environmental cues. Where behavioural responses differ consistently between individuals over time or contexts, this is “personality”. In wild animals, personality is linked to variation in fitness and survival. Predictions on the behavioural mechanisms underlying this variation come from captive studies, on the often untested assumption that captive behaviour reveals how animals would behave in the wild. In chapter 2, using blue tits (Cyanistes caeruleus) I tested first whether behaviour in captivity predicted foraging behaviour in the wild. I measured the personality traits neophobia (latency to feed in novel scenarios) and exploratory tendency, first by relatively standard captive protocols and second, using an electronic monitoring system at feeding stations, by novel wild methods. As predicted, analogous traits correlated across contexts. Moreover, neophobia and exploratory tendency were uncorrelated within individuals in both contexts, in contrast to many other species. In captive studies, personality types also respond differently to changing environmental cues, or “environmental sensitivity”: neophobic and non-exploratory types adjust behaviour whilst neophilic and exploratory types maintain foraging routines. In chapter 3, I tested this second captive prediction in the wild, defining environmental sensitivity in the wild by changes in feeder use with varying air temperature or food supply. Neophobic and, contrary to expectation, exploratory blue tits were most environmentally sensitive. By contrast, neophilic and nonexploratory birds visited feeders at a fixed level independent of temperature and continued to visit feeders for a prolonged period even after they were emptied. Age and body size also influenced environmental sensitivity, suggesting learning and dominance interactions modify the expression of personality in the wild. From potential behavioural costs, in chapter 4 I turned to the physiological costs of personality. Variation in metabolic rate and stress metabolism may be proximate mechanisms for personality. Whilst these physiological traits are linked to oxidative stress directly, with pro-oxidants that damage body tissue a by-product of metabolism, few studies link personality to oxidative stress. I found that oxidative profile (pro-oxidants, antioxidants, oxidative stress and oxidative damage) and hence physiological costs differed notonly within traits but also related differently to neophobia and object exploration in captive-bred greenfinches (Carduelis chloris). Finally, variation in response to environmental cues may reflect differences in learning between individuals, as perhaps illustrated by age differences in environmental sensitivity (Chapter 3). In chapters 5 and 6, I investigated whether learning that a feeding site is temporally stable could cause changes in response to food appearance (“local cues”) when foraging. I predicted that birds would re-find food by spatial rather than local cues in these scenarios, as appearance can change hence local cues become unreliable over time. In chapter 5, I carried out an associative learning test to test this prediction in captive-bred greenfinches. Within a simple foraging scenario, the prediction was upheld: greenfinches favoured local cues in situations where the temporal stability of food was unknown, but switched to spatial cues when temporal stability was learnt through repeated encounters. In chapter 6 though, four of five wild bird species foraging at temporally stable bird feeders continued to respond to local cues, selecting feeders on the basis of colour. Most species were biased toward red feeders, and also responded to social cues when finding feeders: foraging strategies better suited to finding ephemeral food than re-finding temporally stable feeding sites. I suggest that wild birds use information on temporal stability from the broader environment (i.e. natural ephemeral food beyond temporally stable artificial feeders). This illustrates how animals may not necessarily forage in the wild as we would expect within specific contexts. Throughout this thesis therefore, my findings illustrate the importance of testing predictions generated from captive behaviour in the wild. Moreover, identifying variation in both the foraging strategies and physiological costs to individual variation in behaviour, this thesis provides new insight into the adaptive significance of animal personality.

Chapter 2 Natural concurrent infections of rats with H. diminuta and M. moniliformis may increase the chances of individual insects being infected by both parasites simultaneously. Although previous work has shown H. diminuta and M. moniliformis can co-exist in the small intestine of the rat, it was considered important to determine the length of time the two species would be simultaneously patent. This time period would occur during maximum rate of egg production by both parasite species, determined from numbers of eggs found the rat's faeces. It was concluded that although eggs of both species were found in the faeces for proximity 11 weeks post-infection, the chances of an insect acquiring an infection would be highest at the time of peak egg production. This was approximately 6 weeks in the mid-patent period of M. moniliformis. Chapter 3 The initial process leading to an infection of the intermediate host by H. diminuta and M. moniliformis were examined; i.e. egg ingestion, hatching, passage along the insect gut and penetration of the midgut wall. Adaptations by the larvae of both species of parasite to infect the intermediate host were shown to influence host-specificity. H. diminuta and M. moniliformis hatched in the guts of a wide variety of insect species, but only M. moniliformis acanthors penetrated the gut wall of P. americana. H. diminuta oncospheres only penetrated the gut walls of the locusts, Schistocerca gregaria and their natural hosts, the flour beetles Tribolium confusum and Tenebrio molitor. Transit time for food material passing along the gut was found to be important in the synchronisation of parasite hatching with arrival at the site of gut penetration. P. americana holds food items and parasites in the foregut (crop) until partially digested before allowing them into the midgut. Thus any H. diminuta oncospheres stimulated to hatch by the insect's mouthparts only progressed as far as the crop until initial digestion was completed; the time taken for crop contents to be passed in to the midgut tended to exceed the time larvae remained active and capable of gut penetration. In contrast, M. moniliformis acanthors hatch over a longer time period and therefore enter the midgut in a state capable of gut penetration. The abilities of the two species of parasite to tunnel through the midguts of different insect species was compared in vitro using a qualitative assay technique. H. diminuta oncospheres were unable to penetrate the tissues of P. americana midgut. Chapter 4 Eggs from both species of parasite were fed simultaneously to insects. It was initially proposed that oral infections of cockroaches with M.moniliformis might facilitate penetration of the cockroach gut by H. diminuta oncospheres, if the gut tissues were sufficiently disrupted by the former parasite. However, only locusts could be simultaneously infected with both species orally, infected, and viable H. diminuta oncospheres fed to cockroaches were found to adversely affect the success of a simultaneously offered dose of M. moniliformis eggs (acanthors). A hypothesis was put forward to explain this result; that H. diminuta oncospheres perturbed the midgut tissues in their unsuccessful attempt to burrow through the gut wall, thus initiating a wound-healing response by the hosts immune system. This resulted in gut-penetrant M. moniliformis larvae being killed by a melanotic encapsulation reaction. Unfortunately, light and electron microscopy has revealed little evidence of such a wounding in the gut which might have initiated such an immune response. Chapter 5 H. diminuta oncospheres were injected directly into the cockroach haemocoele, as it has been previously shown that a small number of parasites survive. By repeatedly passaging the few surviving cysticercoids from each infection through the rat/cockroach system it was hoped to raise a cockroach-infective strain of H. diminuta. However, their infectivity to cockroaches did not increase in successive generations; several explanations for the possible failure of this selection programme have been put forward. Intrahaemocoelic injections of pre-hatched H. diminuta oncospheres or M. moniliformis acanthellae into the host made it possible to by-pass the gut and thus investigate concurrent haemocoelic infections of cockroaches with both species of parasite. When H. diminuta was injected into M. moniliformis-infected cockroaches, prevalence and intensity of the former were significantly elevated compared to naive controls, indicating that a putative immunosuppressive action from M. moniliformis facilitated H. diminuta development. In some instances, H. diminuta was found to have burrowed through the envelope surrounding M. moniliformis and continued normal development within, unmolested by the host's haemocytes. This was considered as further evidence for the protective nature of the acanthocephalan envelope. Chapter 6 In Chapter 6, assays for aspects of haemocyte behaviour was performed on insects (in particular, P. americana) experimentally infected with either M. moniliformis or H. diminuta. It was found that the phenomenon seen in Chapter 4, whereby H. diminuta adversely affected the success of M. moniliformis when fed simultaneously to cockroaches, appears to be a direct consequence of the stimulatory effect of H. diminuta on the immune system. Conversely, the developing larvae of M. moniliformis were shown to depress haemocytic activity; possibly explaining why elevated numbers of injected H. diminuta survive in M. moniliformis-infected cockroaches. To investigate the affects of immune stimulation on the survival of parasites, locusts were injected with Zymosan, a derivative of yeast cell walls containing 1,3-glucans. H. diminuta oncospheres, injected into Zymosan-stimulated locusts appeared to be partially encapsulated, resulting in a temporary arrest in their development when compared to controls. An assay was devised to observe the encapsulation of materials in vitro by the haemocytes of P. americana. This method was used to show differences in the haemocytic encapsulation reaction to the different larval stages of M. moniliformis and H. diminuta. Both the gut-penetrant stages of each parasite (i.e. oncospheres and acanthors) were encapsulated, whereas the haemocoelic stages (i.e. cysticercoids and acanthellae to cystacanths) remained free of haemocytes. Chapter 7 Finally, in Chapter 7, a model for the alternative pathways leading to success or failure of parasitism by H. diminuta and M. moniliformis in the insect host has been discussed. The results presented here contribute to the fuller understanding of how immune stimulation and immunosuppression affect the survival of helminth parasites, particularly in the cockroach host. These two phenomena have also been shown to be effected by the parasites themselves.

Kin recognition allows individuals to assess their relatedness to conspecifics, thus they may then show kin discrimination and make informed choices as to with whom to associate and/or breed. Cooperatively breeding species, such as the cichlid Neolamprologus pulcher, are an excellent model system for investigating kin recognition, as they live in complex social groups, containing both kin and non-kin group members. Cooperation involves individuals helping to rear the offspring of the dominant pair. Helping is costly, but helpers may gain direct fitness benefits through living in a group, and if they aid relatives, they can also gain indirect fitness benefits through kin selection. Furthermore, by being able to recognise kin, individuals can also avoid inbreeding and the potentially deleterious consequences of it. Thus, N. pulcher are predicted to have good kin recognition abilities. In this thesis, I investigate kin recognition and its consequences for helping and mate choice in a captive population of N. pulcher. In chapter 2, I investigated the kin recognition capabilities of juvenile N. pulcher whilst controlling for familiarity. I found that N. pulcher preferred to associate with unfamiliar kin over unfamiliar non-kin. Kin recognition was via some form of phenotype matching, with chemical cues being more important than visual cues. Additionally, I found no discrimination between familiar and unfamiliar kin; thus, relatedness rather than familiarity was important in the association preferences of juvenile N. pulcher. Chapter 3 explored whether relatedness to the breeding pair, or differences in individual behavioural types affected the amount or type of helping shown by N. pulcher. Controlling for group size and helper relatedness, I found that the relatedness of the helpers to the breeders had no influence on the amount or type of help carried out. Thus, kin selected benefits alone cannot explain variation in helping behaviour in N. pulcher. The amount of territory maintenance carried out correlated with the amount of territory defence, thus, some individuals were consistently helpful. Individuals varied consistently in their aggressiveness, risk-responsiveness and activity levels, but these traits did not correlate with one another. More aggressive, risk-prone or more active helpers carried out more territory defence than submissive, risk-averse or inactive helpers. In contrast, the amount of territory maintenance carried out by helpers, was not correlated with the behavioural types. Thus, differences in behavioural types explained more variation in helping behaviour in N. pulcher than relatedness. Since motivation to associate with kin might vary with age and individual state, in chapter 4, I investigated whether N. pulcher avoided kin when sexually mature, and examined the fitness consequences of inbreeding. In standard two-way choice tests, I found that whilst male N. pulcher showed no preferences for associating with sisters over female non-kin, female N. pulcher preferred to associate with brothers over male non-kin. However, when given the opportunity to breed, latency to breed and hatching success did not differ between brother-sister pairs and unrelated pairs. Thus, in N. pulcher inbreeding is not actively avoided and does not appear to be detrimental to fitness. I suggest that sex-biased dispersal and regular breeder replacement on territories may minimise the occurrences of inbreeding in the wild and that inbreeding may be opportunistic, rather than a strategic decision. The final theme of my thesis investigated the effect of phenotypic traits on mate choice. In N. pulcher (chapter 4) I found that the size of an individual’s facial stripe, which varies between individuals, played no role in mate association preferences. I then investigated male mate choice for female body size in the non-cooperatively breeding green swordtail, Xiphophorus hellerii. In chapter 5, I found that males showed preferences for large over small females when presented only with visual cues, but not with only chemical cues. However, as the size differential between the large and small female increased, males showed preferences for the larger female based on chemical cues. So, male X. hellerii prefer larger females, which are predicted to be more fecund and hence, bring them greater fitness returns. In conclusion, my study has shown that N. pulcher can recognise kin, but the ability to do this does not compel individuals to show kin directed cooperation, or inbreeding avoidance. Instead, factors such as an individual’s behavioural type have more influence on decisions to help, and inbreeding does not appear to be detrimental to fitness. Overall, this project shows that under the conditions we tested, kin selection alone does not drive the social interactions in N. pulcher groups. Further, it highlights the need to consider multiple factors affecting an individual’s fitness, in order to fully understand why different species show a propensity to recognise and discriminate between kin and non-kin.

1. The ecology of Neoechinorhynchus rutili (Acanthocephala) has been investigated in a Scottish population of brown trout (Salmo trutta) inhabiting a small highland loch, in Central region. The results have been compared with data from other hosts and localities. In addition, a complementary study on the ecology of Echinorhynchus truttae in brown trout was carried out at a reservoir in Lothian Region. 2. A review of the knowledge of the classification, morphology of developmental stages, life cycle, ecology and distribution and biochemistry and physiology of Neoechinorhynchus rutili was undertaken. 3. Examination of 1189 fish (11 species) from 8 Scottish regions and collation of available records indicated that Neoechinorhynchus rutili infects 8 species of freshwater fish (Esox lucius, Gasterosteus aculeatus, Perca fluviatilis, Phoxinus phoxinus, Salmo gairdneri, Salmo salar, Salmo trutta and Salvelinus alpinus) at 41 sites in 6 Scottish regions. These sites encompass a wide spectrum of aquatic environments in terms of size, water quality and faunal community structure. This wide distribution of Neoechinorhynchus rutili is explained through the multifarious habits of the definitive host species and possibly interactions with human, avian or insect factors. There is evidence for temporal stability of the infection at some of these sites. 4. The main field site for the examination of the ecology of Neoechinorhynchus rutili was Loch Maragan (Grid ref. NN 402278), Central Region. This small loch (surface area 7.3 ha, maximum depth 10.2 m, volume 153943 m3) lies at 472 m above sea level and had slightly acid water conditions (p.H. 6.44, October 1986.) Three species of fish (Anguilla anguilla, Phoxinus phoxinus and Salmo trutta) were found to inhabit the loch. In August 1987 the brown trout population size was estimated, by a simple mark and recapture technique, as 2641 (1 to 3 years olds) (Maximum value). 5. Two benthic faunal surveys, carried out in November 1986 and May 1987 respectively, did not reveal the species of invertebrate which was acting as intermediate host for Neoechinorhynchus rutili at Loch Maragan. 6. A total of 226, between 1 and 5 years old were caught at Loch Maragan over a 25 month period between July 1986 and August 1988 and examined for visceral and gut macroparasites. Minnows (n = 207) and alder flies (Sialis lutaria) (larvae and adults) were also examined. Sialis lutaria larvae were infected with Neoechinorhynchus rutili and unidentified trematode metacercaria. The minnows harboured Neoechinorhynchus rutili, Ligula intestinalis and Crepidostomum spp. The brown trout harboured Capillaria salvelini, Crepidostomum spp., Diphyllobothrium ditremum and D.dendriticum and Neoechinorhynchus rutili. The distributions of these parasitic infections amongst their hosts are described. 7. The collection of 4992 Neoechinorhynchus rutili from 226 brown trout, revealed the overdispersed nature of the infection (k = 0.7893), individual fish harbouring between 1 and 324 worms. The overall prevalence and mean intensity of infection were 87.6% and 21.5 respectively. Monthly prevalence values (of samples) never fell below 50%. Trout age, but not sex, influenced the infection parameters. 8. Neoechinorhynchus rutili exhibited a definite seasonal cycle of intensity and maturation in brown trout at Loch Maragan. Although recruitment apparently occurred throughout the year, worm intensitites, particularly of gravid females, peaked in summer when water temperatures and host feeding rates were maximal. Acanthor production occurred between March and November. In summer 1987 the Neoechinorhynchus rutili metapopulation (in trout aged 1 to 3 years) was estimated to be 36154, of which 5566 were gravid females. These females were estimated to have produced 1.6 x 108 shelled acanthors which represents a reproductive success rate of 3.5 x 10-3 % i.e. 1 in 38030 shelled acanthors becoming a reproductively active female in the next parasite generation. No evidence for density dependent effects upon worm fecundity were found. Similar seasonal patterns were found at other Scottish locations. 9. Neoechinorhynchus rutili was typically found in the ileas and rectal regions of the trout gut, but the distribution was dynamic with respect to season, worm sex, state of maturity and infection intensity. 10. No adverse effect upon the condition factor of brown trout in Loch Maragan could be attributed to the presence of Neoechinorhynchus rutili. 11. The diet of brown trout at Loch Maragan was analysed and Sialis lutaria larvae formed an important element, especially in spring. Larger trout were found to be piscivorous. 12. Acanthocephalan specimens found in the haemocoel of Sialis lutaria larvae collected from 2 Scottish sites (Loch Maragan and Loch Monzievaird) were identified, on the basis of their morphology, as the eoacanthocephalan Neoechinorhynchus rutili. 13. Sialis lutaria larvae were found to be infected with Neoechinorhynchus rutili at Loch Maragan and 3 other Scottish sites where fish were also infected (Bridge of Weir fish farm, Drumore Loch and Loch Monzievaird). The distribution of infection was typically overdispersed (k values ranged from 0.27 to 1.25). Overall prevalence values ranged from 4.2 to 40.7%. Infection parameters varied with insect size and season. 14. Establishment of Neoechinorhynchus rutili in experimental infections of rainbow trout (Salmo gairdneri) via feeding upon infected Sialis lutaria larvae was up to 33% successful. Experiments to infect ostracods (Herpetocypris reptans) and Sialis lutaria larvae via feeding with shelled acanthors were unsuccessful. 15. Postcyclic transmission of Neoechinorhynchus rutili occurred in the laboratory when rainbow trout were exposed to worms established in the three-spined stickleback (Gasterosteus aculeatus) hosts. The re-establishment rate was estimated as 92.8%. There is circumstantial evidence to suggest that this form of transmission occurs in natural populations in Scotland. 16. In the complementary study at Gladhouse Reservoir, 4 visceral and gut parasites were identified in the brown trout: Cystidicola farionis, Eustrongylides sp. (Nematoda), Eubothrium crassum (Cestoda) and Echinorhynchus truttae (Acanthocephala). Both E.crassum and Echinorhynchus truttae were overdispersed in their hosts and the overall prevalences and intensities of infection were 61.9% and 92.9% and 1.48 and 156.4 respectively. Echinorhynchus truttae exhibited a seasonal pattern of maturation, females releasing shelled acanthors in the summer months. This study also considered the logistics of carrying out scientific research in collaboration with members of the public.

The aim of this study was to investigate factors affecting the uptake of dissolved inorganic nitrogen (DIN) by the temperate symbiotic sea anemone, Anemonia viridis. Laboratory experiments were used to test DIN uptake rates of symbiotic and aposymbiotic (alga-free) individuals, to enable the prediction of the magnitude of DIN flux rates in anemones in their natural environment, and to allow the construction of a model of the DIN fluxes in the association. It was also hoped to assess the potential for fully autotrophic growth in this species. Uptake rates were measured by following depletion of DIN from small-volume incubation chambers that had been enriched to produce DIN concentrations across the reported range for temporate coastal waters. At light levels of 190 E.m^-2, ammonia was taken up by symbiotic anemones from all concentrations tested, whilst aposymbionts lost ammonia to the surrounding sea water. Nitrate was not taken up by any of the anemones. The relationship between the weight-specific ammonia flux rate and ambient ammonia concentration was linear for both symbionts and aposymbionts. In all cases the slope of the relationship was positive, symbionts showing increasing rates of uptake and apsoymbionts showing decreasing rates of efflux with increasing ammonia concentration. There was no evidence of Michaelis-Menten uptake kinetics across the range of concentrations used, despite extending this to 30 g-at NH3 - N.1-1 in an attempt to show saturation of uptake. In darkness, the elevation of the rate/concentration relationship was depressed for symbionts, but not for aposymbionts, so that symbionts showed efflux of ammonia at concentrations below 4 g-at NH_3-N.1^-1. The elevation increased with light intensity up to 50 E.m-2.s-1, whilst the slope of the relationship remained constant. Above 50 E.m^-2 the elevation could not increase, but the slope became greater. Feeding the anemones with squid mantle tissue three days prior to experiments had variable effects on ammonia flux rates, but there was a tendency for the elevation of the rate/concentration relationship to drop in recently fed anemones, compared to starved anemones at the same light intensity. Effects are likely to be more marked over the first 24 hours after feeding, and ammonia flux during this period may have significant consequences for the nitrogen balance of the association. Freshly isolated zooxanthellae took up ammonia, and showed Michaelis-Menten type uptake kinetics, with k_s between 3 and 7 g-at NH3 -N.1-1, and max between 0.025 and 0.043 g-at NH_3-N.(g chla)-1.h-1. Uptake rates were not affected by light level, and were an order of magnitude higher than those of the intact association over the range of experimental concentrations used. Incubation of the intact association with 15 NH3-enriched sea water resulted in the 15 N being incorporated into the zooxanthellae, and not the host tissue, over 30 minutes exposure to the labelled ammonia. This showed directly that the algae are responsible for the ammonia uptake of the association. The results obtained are consistent with a two-component uptake system, diffusion across the animal cell membrane being driven by a concentration gradient produced by algal ammonia assimilation. The observed kinetics are the result of two antagonistic processes; the catabolic production of ammonia by the animal cell, and the active uptake of ammonia by the algae. It appears that algal ammonia uptake is saturated at irradiances of 50 E.m^-2.s^-1 and below, and unsaturated at higher irradiances. Light levels and ammonia concentrations were measured in the vicinity of a natural population of Anemonia viridis in Loch Sween, Argyll. The laboratory experiments were used to predict the net metabolic gain or loss of DIN experienced by anemones under the measured field conditions. Ambient ammonia concentrations were negligible around the anemones in Loch Sween throughout the year, and under such conditions a net loss of nitrogen would have occurred, making a heterotrophic input of nitrogen necessary for growth. Net nitrogen gain may, however, occur at ambient ammonia concentrations that could exist in rockpools further south in Britain, where A. viridis is frequently the dominant animal species. A laboratory experiment was conducted to test for the possibility of autotrophic growth in the presence of adequate levels of DIN. Unfortunately, experimental conditions did not consistently provide high enough concentrations of ammonia for net nitrogen gain to be predicted, but anemones exposed to elevated levels of DIN for 8 weeks lost significantly less weight than anemones kept in sea water with negligible ammonia concentrations. The weight loss of the DIN-'fed' anemones corresponded closely to the weight of lost nitrogen predicted from the flux experiments, after conversion to equivalent weight of protein. It remains to be seen if Anemonia viridis can survive and grow autotrophically, but even in the nitrogen-poor environment of a Scottish sea loch, the algae can potentially supply 86-90% of the associations basal nitrogen requirements. The capacity for DIN uptake is clearly an important factor in the success of this species in temperate waters.

The aim of this project was to investigate the ecology of Manx shearwaters on Rhum, with particular emphasis on the influence of predation and other extrinsic factors upon breeding success and population dynamics. Brown rats were not found to be active predators of shearwaters on Rhum. Few rats were present in the shearwater colony during the birds' breeding seasons, but larger numbers scavenged on the remains of shearwater eggs and chicks during the autumn and early winter. Rats are apparently prevented from becoming established in the shearwater colony by food shortage in late winter and by the colony's location within Rhum. Golden eagles, peregrine falcons and great black-backed guils do prey upon shearwaters on Rhum but have an insignificant effect upon the shearwater population. Hatching success in the colony varies markedly from year to year and is significantly negatively correlated with the incidence of heavy rain during the incubation period. This arises from the variation in susceptibility to flooding among nesting burrows. The tendency for individual burrows to flood was quantified and was found to have a significant influence upon the probability of successful breeding. Weights at fledging showed a significant seasonal decline, apparently due to increased difficulty for adults in provisioning chicks late in the season. Heavy fledglings are apparently more likely to survive than light ones so that there is selective advantage in breeding as early as possible. The ultimate and proximate control of timing of laying is discussed. There was evidence of year to year variation in feeding conditions around the colony. Chicks grew more slowly in 1985 than in 1984 and were fed less frequently. The implications of these findings to the population dynamics and conservation of the colony are discussed.

This study investigated the impact of kleptoparasitism and predation by large gulls (Larus spp.) on the recruitment rate, foraging economics and reproductive performance of Atlantic puffins Fratercula arctica breeding on the Isle of May, Firth of Forth, southeast Scotland. The implications of the findings for nature reserve management strategies are discussed. During the period from 1972 to 1989 the population of herring L. argentatus and lesser black-backed gulls L. fuscus nesting on the island was reduced as part of a gull control programme. Since 1989, gull management has involved maintaining gull-free areas by repeated removal of nests. Analysis of data collected over a 23 year period showed that the recruitment rate of puffins to the colony was significantly higher in the areas of the island where gull nest density was low. This suggests that, by reducing the density and spatial distribution of nesting gulls, the control programme successfully increased the attractiveness of the colony as a potential breeding site for puffins. Maintaining gull-free areas was also effective in reducing the frequency with which puffins were attacked by gulls. Kleptoparasitism risk during the morning and evening peaks of puffin feeding activity was 37% in gull-occupied habitat compared to only 5% in gull-free areas. These results indicate that gulls predominantly attacked puffins that bred in close proximity to them and did not move outwith the main gull breeding areas to attack puffins elsewhere on the island. The presence of gulls also significantly reduced the actual rate at which adult puffins attempted to deliver food to their chicks. The lower provisioning rate coupled with the higher kleptoparasitism risk resulted in puffins breeding in gull-occupied habitat delivering 44% fewer food loads to their chicks in the morning and 25% fewer in the evening compared to puffins breeding in gull-free areas. However, despite this difference, there was no evidence of a significant reduction in the growth of puffin chicks or their survival to fledging.

The elaborate courtship ritual performed by Drosophila melanogaster males to interest females in copulation is used as a model to investigate the genetic and neurobiological basis for the specification of complex behaviours. One component of this courtship ritual is the ‘lovesong’, which both promotes mating and carries vital species-specific information. Sex mosaic studies have shown that specific regions of the central nervous system (CNS) must be male in order to produce this courtship song. Indeed, two genes in the sex determination hierarchy, fruitless (fru) and doublesex (dsx), which are responsible for sexual differentiation in both the CNS and soma, are required for wild-type song production. However, the critical differences underlying the sexually dimorphic production of courtship song are unknown; moreover, the relative contributions of fru and dsx to the generation of these differences are not known. The central aim of this thesis was to investigate the neurobiological basis for the sexually dimorphic production of courtship song; and in addition, to determine the individual and combined contributions of fru and dsx in creating these essential differences. The long-term goal of this study was to determine the mechanism by which genetic factors such as fru and dsx can create the potential for courtship song by specifying aspects of CNS development and function. As a first step towards this long-term goal, the work presented in this thesis has identified the individual and combined contributions of fru and dsx to the production of courtship song. Moreover, a requirement for both fru and dsx in the specification of a neural substrate underlying courtship song was shown. Using females constitutively expressing the male-specific fru isoforms, the individual contribution of fru to the production of courtship song was determined; and the amount and quality of song generated by these females was found to be highly aberrant. Thus fru alone does not specify courtship song production. In fact, wild-type courtship song was only achieved when females expressed the male-specific isoforms of both fru and dsx, demonstrating that both genes are required for the specification of courtship song. Next, the co-expression of fru and dsx was examined, where fru and dsx were found to be co-expressed in three distinct regions of the CNS. One of these regions, the mesothoracic ganglion (Msg), is strongly implicated in the production of courtship song, suggesting that fru and dsx may act in concert to specify courtship song production. A closer examination of this region revealed a male-specific population of fru-expressing neurons in the Msg. This population of neurons was not present in females constitutively expressing fru’s male-specific isoforms, demonstrating that fru alone does not specify the development of a sexually dimorphic Msg. Instead, a critical requirement for dsx, alongside fru, in the specification of this sexually dimorphic population of neurons was shown. Thus a critical requirement for both fru and dsx in the specification of courtship song, and in the creation of a sexually dimorphic population of neurons in the Msg has been revealed. Finally, to determine how this sexually dimorphic population of neurons is linked to courtship song production, this thesis described the generation of genetic tools to allow the visualization and manipulation of these sex-specific neurons. Using the UAS/GAL4 system of targeted gene expression, another sexually dimorphic population of neurons was identified in the posterior brain, which is, like the Msg, a region where fru and dsx are co-expressed. Together, these results suggest that fru and dsx act to generate sexually dimorphic populations of neurons in regions of co-expression, where these neurons may form part of the sex-specific neural circuits underlying the performance of sexually dimorphic behaviours.

Field observations were made at Loch Sween in an attempt to describe the behaviour and ecology of L.depurator. Abundance varied over time with peak abundance occurring during periods of warm water. Peak densities of 1 crab/7m [sq] were observed but average density was much lower at 1 crab 25m [sq]. Some evidence was obtained for migration from the study site similar to that reported for other species. Tagging studies revealed no long term site attachment. Populations were found to be sex biased in favour of male during summer months and individuals were most commonly observed within 1m of the nearest conspecific. Ovigerous crabs were observed in two distinct phases between April and September and recruitment of juvenile crabs; occured in early spring. Observations on food and feeding behaviour showed L depurator to be a predator of sessile or slow moving benthic macroinvertebrates. The main types of escape response were observed and discussed. The incidence of physical damage and naturally occurring aggression in the field were low. Investigation into the locomotor activity of individual swimming crabs in the laboratory showed that shallow water (<9m) crabs displayed significant cyclic locomotor activity while deep water (>30m) crabs displayed arhythmic activity.