Although today coral reefs need to be managed to ensure their persistence in the challenging conditions imposed by a rapidly changing environment, practitioners worldwide often lack adequate tools to achieve this. Spatial patterns of the abundance of reef resources, but also of the processes that govern reef recovery after disturbance, need to be mapped at relevant scales to identify priority conservation measures. In this thesis I took important steps towards the construction of adequate tools for reef managers: towards creating maps of reef resilience. The first step comprised the accuracy enhancement of thematic maps to discriminate typical forereef habitats that differ in their structural complexity. The second step consisted of the creation of statistical models to predict spatial patterns of the density and biomass of several fish species including grazers. To date, fish species richness can be mapped but not the spatial patterns of abundance or biomass of key species of reef fish. Here, I demonstrate that it is possible to predict spatial patterns of the abundance of key species of grazers across large scales on Caribbean reefs by mapping their acoustic roughness. The third and fourth steps focused on improving our understanding of the process of parrotfish grazing on Pacific reefs. I created models to predict the grazing impact of populations of grazers from their abundance data incorporating sources of spatio-temporal variability in their grazing behaviour. Although grazing of Pacific parrotfish communities is a subject of growing concern and several aspects of its dynamics are well understood on the Great Barrier Reef, this thesis contributes with two major future goals: (1) to rank Micronesian parrotfish species according to their relative contribution to grazing impact and (2) to aid the generation of maps of grazing.

Gaining an understanding of the spatial ecology of marine turtles is essential for elucidating aspects of their life history ecology and for effective conservation management. This thesis presents a collection of chapters seeking to investigate the spatial ecology of this taxon. An array of technologies and methodologies are employed to ask both ecological and spatial management questions. Work focuses on foraging and thermal ecology, spatial appropriateness of Marine Protected Areas, movement models to describe habitat utilisation, analysis of data from sightings and strandings schemes and the use of a large synoptic fisheries dataset to describe fisheries patterns and putative risks to marine megavertebrates.

This thesis aims to investigate causes of variation in the condition of Atlantic cod using long-term, high-resolution liver-weight databases.  The Liver Condition Index (LCI), the weight of the liver as a percentage of total body weight, is used as a measure of condition. The first study presented in this thesis is an analysis of the causes of inter- and intraannual variation in the condition of Northeast Arctic cod.  Temperature is identified as an important factor influencing condition.  Increasing temperature is shown to positively impact condition at both inter- and intraannual timescales.  Greater availability of the cod’s preferred prey, a monthly timescale.  A time lag of one month between the change in capelin consumption and the subsequent change in condition, or ‘latency’, is demonstrated.  Herring (<i>Clupea harengus</i>) are shown to have an indirect impact upon cod condition through competition for energy-rich capelin. The second study presented in this thesis is an investigation of sex-dependent variation in the condition of Northern cod.  Differences in condition are mainly seen in mature fish, with mature female cod in consistently higher condition throughout the year than male.  However, the only season for which a significant difference was identified was spring, suggesting that pre-spawning females are able to acquire more energy than pre-spawning males.  In addition, a greater decline in post-spawning condition was found in females.  Later maturation and slightly faster growth in mature females provide further verification of variation in energy allocation dependent on sex.  The analyses presented in this chapter suggest that individual-level analyses of condition in cod should, in future, takes account of the effect of sex on energy usage.

A study was carried out to investigate the effect of different levels of supplementary food on populations of wood mice Apodemus sylvaticus. The objective was to test the hypothesis that population density of wood mice is associated with habitat productivity. An experiment was set-up whereby three different levels of supplementary food in the form of wheat was given to a population of wood mice living in a sand-dune area in North-East Scotland. Using capture-recapture data, the population dynamics and demography of the supplemented populations were compared with a control population with natural levels of food. Supplemented populations had higher peak densities compared to the control population but not in proportion to the level of supplementation. Extra food however did not bring forward the onset of the autumn increase in numbers and populations in summer remained low and depressed. Recruitment was low in all populations and only showed substantial increase in autumn. Recruitment was associated with the level of supplementary food and with sex ratios increasingly biased towards males. Supplemented population had a lower proportion of reproductive females. Food supplementation also did not improve survival over the breeding season. A second supplementary food experiment was carried out using a replicated treatment design. Control populations were compared with populations that received a single level of supplementary food. The experiment was repeated for another year but with treatments exchanged between grids; previously supplemented grids treated as controls and vice versa. The objective of the experiment was to verify that the observed changes in density and dynamics can be attributed to the additional food provided and not to other environmental factors. Supplementation was confined to the breeding season only during both years. The results of the second experiment confirms the positive relationship between food availability and population density. Supplementary food led to higher recruitment of both males and females. However variation in population abundance was positively related to male recruitment only. Using statistical modelling to model survival and recapture parameters, survival rates of supplemented populations appeared to have improved during early summer but were not different from controls during the later period. The third study looked at the effect of food supplementation on ranging and spacing behaviour of wood mice. The objective was to test the hypothesis that supplementing with extra food early in the breeding season would cause existing reproductive females to reduce their range sizes, in so doing giving rise to gaps between previously adjoining territories. The creations of this gaps would allow new females to become recruited and join the breeding population. Comparison between range sizes indicated that supplementation led to a reduction in female home range and core range sizes but not that of males. Supplemented females also had a smaller mean percent core range overlap, suggesting less interaction between neighbouring territorial females at the range boundaries. The spatial distribution of female territories shows that established females decreased their core areas with supplementary food, allowing new females to become established in the area. Since food addition led to a reduction of territory size of resident females and allowed the establishment of new females in the supplemented area, females therefore play a more crucial role than males in limiting numbers during the breeding season. The low population density during the breeding season can be attributed to the spacing behaviour of reproductive females. Increased food availability may increase the number of resident females. However the territorial nature of females means there is an upper limit to the number of females accommodated for a level of food availability. Therefore the lower recruitment and poor survival at all levels of food over the breeding season was due to the failure of new females to establish a home range. At the end of the breeding season in autumn, when females no longer possess territories, recruitment rose to a peak and survival improved considerably.

The study of sympatric red (<i>Cervus elaphus</i>) and roe deer (<i>Capreolus capreolus</i>) in a mixed-aged, second-rotation conifer forest in Argyll, Scotland, used radio-tracking and dung accumulation techniques to identify differences in the habitat utilization between the two species. The home range and core area size of roe deer was approximately a third of the size of those of red deer. A seasonal difference in range and core size occurred in male roe deer (larger in winter) but not for female red or roe deer. Habitat selection was similar by day and night within each species. The use of habitats with younger trees (upto 15 yrs), was greatest in slightly older tree stages (20-30 yrs). Both species used open habitats for feeding. In all `cover' habitat types, roe deer were found comparatively closer to the cover-edge than were red deer. The distance travelled about the home range varied according to time of day and season. Furthest distances moved occurred between 08.00hrs and 12.00hrs, and 18.00hrs and 24.00hrs. Male roe deer moved greater distances in winter, whereas the distances travelled by female red and roe deer were similar between seasons. Red deer were proportionately more active (59%) than roe deer (54%) through the 24hrs, but roe had more bouts of activity. Botanical and chemical analyses of rumen contents, showed that the diets and nutritional quality of the food eaten, differed between the deer. The diet of red and roe overlapped slightly in summer (35%) as red deer fed predominantly on grasses, and roe deer fed on forbs. In winter however, the overlap was much larger (75%) with both deer species feeding upon heather and conifers, although red deer ate larger amounts of grass than roe. The proportion of N, P, K, Ca and Mg in the winter diet of roe deer was greater than that for red deer. The observed feeding behaviour of roe and red deer highlighted the different feeding strategies adopted. Roe deer were more selective, feeding on patchily distributed plants, so consequently moved more often whilst feeding. Red deer fed predominantly on grasses that were widespread; they moved and raised their heads less often than did roe deer. The differences in habitat selection, preference and diet between red and roe deer were unlikely to be the result of interspecific competition. The observed differences follow the Jarman-Bell principle that body size dictates their nutritional requirements and consequently their behavioural patterns of foraging and habitat selection.

Red grouse digested about 85% of the soluble carbohydrates in heather, their main food, about 40% of the crude fat and amounts of lignin, α-cellulose and hemicellulose varying up to 40%. The metabolisable energy of heather was 1.5-1.9 kcal/g (dry basis) and that of lignin 2.25 ± 0.59 kcal/g. The fatty acid composition of the depot fat of tetraonids reflected that of the food eaten, presumably because the fatty acids were derived directly from the food. The nutrient requirements of poultry were compared, on the basis of metabolisable energy, with the nutrient content of heather. By analogy with poultry, nitrogen and phosphrus were the nutrients most likely to be limiting to breeding hen grouse. Wild grouse in November selected henther containing more nitrogen and phosphorus then the current year's growth of the heather available. Scottish ptarmigan eat the three main plants commonly available. These same species are available to Icelandic ptarmigan, but other available species which contain more nitrogen and/or phosphorus or soluble carbohydrates are preferred, apparently because they are more nutritious. The ecology of red grouse populations at two 'rich' moors overlying basic rooks was contrasted with that at two 'poor' moors overlying acidic rocks. Differences in the nutritive value of the heather, associated with differences in phosphorus content, may have caused the higher grouse numbers and better breeding success recorded at the rich moors. Shoot tips of very young heather appeared to be more nutritions than those of older heather, but no changes in chemical composition with age were noted beyond six years.

The ecology of the plaice Pleuronectes platessa L. and other flatfish in the Clyde Sea Area have been described by M.G.Poxton for the award of the degree of PhD (1976). The dates and duration of spawning have been deduced and the maximum concentrations of eggs and larvae given. Information on the distribution of the eggs and larvae in the plankton is presented. The ecology of O-group plaice on their nursery grounds has been examined in particular detail. Every sandy bay in the Clyde Sea Area has been sampled in order to determine the most important nursery areas. Population estimates have been calculated by making right-angled tows off the beaches with a 2-metre beam trawl. Estimates of mortality and growth rates have been obtained and a length/wet-weight relationship and a wet/dry weight ratio have been calculated. Particular attention has been paid to the food of the plaice, samples of stomachs being taken from the trawl caught fish. Comparisons have been made between plaice taken from different areas (including Firemore Bay, Loch Ewe) and between plaice and the dab Limanda limanda L. which shares the same nursery environment. The main nursery area, Irvine Bay, has been described and details of the bottom fauna given. Offshore sampling for juvenile flatfish has also been carried out using an 18ft. beam trawl in conjunction with F.R.V. Mara. Details concerning the adult plaice stock have been obtained by sampling at Ayr fishmarket where most of the catches are landed. The commercial fishery for plaice has been described and data on the landings have been compiled. The age and length composition of the plaice stocks have been given and assessments of their population dynamics have been made. Research has also been conducted using commercial boats from which samples of adult flatfish have been taken and details of the food eaten have been obtained.

The lizard Anolis oculatus, endemic to the small Lesser Antillean island of Dominica, was selected as a model to study the causes of geographic variation. Patterns of geographic variation were investigated using a variety of univariate, bivariate and multivariate numerical methods. Patterns were found to be predominately clinal, with some incongruence within and between the character systems (body proportions, scalation, and colour pattern), and included altitudinal, east-west and (on the Caribbean coast) north-south clines. Overall four ecotypic groups were recognisable, and the current subspecies were invalidated. Simultaneous Mantel tests pointed strongly to adaptation to various environmental factors (both biotic and physical) as the cause of variation in most of the characters studied. Canonical correlation analysis was used to study the covariance of character constellations and their relationship to environmental variation. A closely related species from a neighbouring island with an independent evolutionary history, A. marmoratus from Basse Terre (Guadeloupe) showed many parallels, both in patterns of geographic variation and associations with environmental variation, which can be attributed solely to the effect of current ecology. A direct test of the action of natural selection was then provided by a translocation experiment in Dominica, where representative populations of the four ecotypes were translocated into a single habitat. In a very short period, evidence was obtained of strong directional selection acting proportionately to the extent of ecological change experienced by the different ecotypes.

The feeding ecology of the short-nosed fruit bat, Cynopterus brachyotis, was investigated in selectively-logged lowland dry evergreen forest in south east Thailand between March 1998-April 2000. C. brachyotis consumes fruit of 14 plant species, pollen/nectar of four and leaves of nine species. Faecal analysis revealed that fruit comprises more than 95% of the diet, and nectar and pollen supplements the diet during the dry season. Fruits of successional plants contribute two-thirds of the diet. The diet of C. brachyotis varies significantly with the age and sex of the bats. Juvenile and immature bats ingest more fruits from plants producing large crops over a relatively short period, while mature bats consume more fruits from plants producing a few fruits nightly over a long period. Females, especially reproductive ones, consume disproportionately more fruits from selectively logged dry evergreen forest species than males. Different study techniques reveal a significant variation in the diet of fruit bats. Leaves are prominently represented in the diet deduced from pellets collected below day roosts than from faeces of captured bats. Radiotracking carried out on 18 individuals for 18-149 days revealed that home range of mature C. brachyotis is between 30-365 ha with the average maximum width of 2,545m. Females have marginally significantly larger home range than males. The roost area contributes an average of five percent of the home range, and there was no incidence of overlap of roost area between any solitary roosting bats. Male C. brachyotis has a high fidelity to its home range and roost area both between seasons and years. Radio-tagged bats visit 1-6 feeding areas each night and spend an average of 1 h 38 minutes in each feeding area. In feeding areas, they undertake 2-10 feeding bouts at fruiting trees each hour. During each night, individual bats commute an average of 7.14 km and up to 14.5 km between day roosts and feeding areas. Mature bats spend 17.9 % of their nightly foraging activity at day roosts. Moonlight inhibits foraging activity of C. brachyotis. During bright nights, it visits fewer feeding areas, reduces its rate of movement between feeding areas and the number of feeding bouts.

A study of the feeding ecology of the flounder in the Ythan estuary was carried out between October 1970 and August 1973. The diet of the fish varied along the length of the estuary. On the mudflats in the upper estuary C orophium vo lu tat or was the main food item with Nereis divers color and Hydrobia ulvae secondarily important, but on the very soft mudflats in the middle estuary small oligocbaetee (mainly Peloscolex benedeni) and polychaetes predominated in the diet. Over mussel beds 3\Terels dlveraicolor and Carcinus maenas were important and the lower and outer part elupeids were taken in large quantities by the older age groups. Intensive study on one mudflat" the Sleek of Tai'ty, showed a seasonal change in the diet; Corophi\:un being more important in the summer, and Nereis, feco. lthica,Hydrobia and P. os. io_&pound;l. ans in the autumn and winter.. Changes in dietbetween day and night were relatively unimportant as were the differences between age classes, though the younger fish (0 & X-group) took many Crangon vulgaria, Gammarus sp. and Meomysis integer, whilst Macoma and Hydrobia wore secondarily important to the older age groups. Trapping throughout 24-hour periods showed that the stomachs of flounders were filled over the high tide period on the mudflatsf irrespective of whether high tide occurred by day or night, whereas at low water most fish had empty stomachs. However, variations in stomach fulness at high tide occurred. 0-group flounders had significantly fuller stomachs during the day than by night, but no differences were found in the >I-group. Also to show conclusively that the >I-group fill their guts to the same extent on different tides an index of gut (stomach plus intestine) fulness was determined and found not to vary significantly between day and night.