This study investigated the ecology of the chequered skipper butterfly Carterocephalus palaemon (Pallas) and was designed to reveal aspects particularly relevant to its conservation. C.palaemon is a rare species confined to a small area on the west coast of Scotland and was once common in central England. It was one of only four butterflies given full legal protection in the Wildlife and Countryside Act of 1981 and thus has a high conservation priority. The study was funded by the nature Conservancy Council. C.palaemon occurs at specific sites throughout its range but, in contrast with other butterflies that occupy precise habitats, does not exhibit colonial characteristics i.e. it exists at low density and has an open population structure with a high degree of dispersal of adults from the flight area. This may be attributable to the abundant semi-natural resources in the region, posing few barriers to movement. It is a butterfly of woodland edge and scrub habitats but adults make use of other areas, such as wet meadows. Features important to adults are shelter, abundant nectar sources for females and areas that provide favourable conditions for mate location by males. Larvae also occur at low density but in different situations to adults. Although some are found within the obvious flight areas, many more probably occur elsewhere, indicating the large area of habitat C.palaemon requires. They have a long development period, in excess of four months, and the availability of the foodplant, Molinia caerulea Moench, in good condition at the end of the year is critical to survival. M.caerulea is common and widespread in Scotland but the majority of this becomes unsuitable for feeding before larvae have completed development. Only where it grows in areas with a favourable water supply and soil relations, usually with soil-enriching plants, that enhance the quality of M.caerulea and prolong its season, can larvae survive. Hence, it is proposed that the length of season available to larvae is important and success may also be determined by the timing of the adult flight period, which is already early in the year.

The main objective of this thesis was to investigate the role of biotic interactions, biogenic structures and successional patterns in several invertebrate soft bottom assemblages in the Ythan estuary, Aberdeenshire, Scotland. The seasonal dynamics of the fauna found within a dense tube-field of <I>Lanice conchilega</I> were investigated. The abundance and composition of this fauna differed from that in the adjacent sediments and from that in the nearby sandflat at Red Inches. No consistent differences in abundance could be detected for any species throughout the tube-field, and the mussel (<I>Mytilus edulis</I>) was the only species significantly positively correlated with <I>Lanice</I> tube density. <I>Pygospio elegans </I>was found in significantly reduced numbers in tube-field areas affected by weed. Accumulation of weed and subsequent settlement of mussels in summer and periodical intense sediment movement during winter may have partly accounted for the large seasonal fluctuations seen in the abundance and composition of the tube-field fauna. The size of distributed patches is thought to influence the rate of colonization, but there have been few experimental tests of this proposition. The hypothesis was therefore tested by a field experiment where the rates and modes of colonization of azoic patches of different size (1600cm<SUP>2</SUP> vs. 55cm<SUP>2</SUP>) were compared. The results of this experiment were also considered when designing subsequent experiments (see below). None of the invertebrates recorded differed significantly in abundance between patch size on any sampling occasion. The total number of species and individuals rapidly attained control (natural sediment) values. However, the patterns of colonization of individual species could be categorised into three types. The abundances of Type I species mirrored closely those in control sediments. Type II species were opportunists, such as the polychaete <I>Capitella</I> spp., which exceeded the control density rapidly but then showed a decline. Type III species, such as the polychaete <I>Pygospio elegans</I>, showed a very slow increase in azoic areas and never attained control densities, although size distributions were markedly different between the two patch sizes. This experiment showed that over the range studied, size of patch was not important in determining overall colonization patterns on sandflats.

The studies were set against a background of concern over the effects of high levels of recreation in the Gairn Gorm area on vegetation and fauna. The aims were to compare the effects of disturbance by trampling on different mountain vegetation types, to examine these effects in detail on one locally widespread vegetation type, Open Juncus trifidus heath, and to make detailed studies of the structure and dynamics of this vegetation type. Comparison of the effects of experimental trampling on six vegetation types showed Alpine Nardus grassland to be the least damaged, followed by Open Juncus trifidus heath. Dwarf-shrub heaths were more heavily damaged. Open Juncus trifidus heath also showed moderate recovery. Vegetation responses were largely determined by the responses of the most abundant species. Such responses supported groupings of species as showing high, moderate or low susceptibility to trampling. Density (4- 4-1 m-** excluding seedlings) and size-class structure were determined for a J.trifidus population at Cairn Lochan, and J.trifidus tussocks were found to be randomly dispersed. Numbers of seedlings seemed to vary from year to year and their spatial distribution was related to the type of surface. Most seedlings occurred on bare soil and on surfaces of fine and medium gravel. A reduction in seedling density from in June, 1980 to 1 . 5m22 in June, 1981 was recorded. The quantitative morphology of J.trifidus tussocks was documented. Tussock size was related to age, based on node counts and the annual production of tillers. The annual cycle of growth and the changes in tussock morphology with age were described. The annual rate of radial growth of tussocks was indicated to be 2 - 3mm per year. An age-state classification for J.trifidus was described. The effects of different surfaces on the germination and establishment of J.trifidus were examined experimentally using a range of gravel sizes and depths at four sites. The benefits of a gravel covering at high altitude were clear. Differences between treatments in numbers of seeds germinating, germination rate, survival, growth and development were sometimes difficult to interpret, but clear altitudinal trends existed. In general, the results supported observations on seedling occurrence in terms of differential effects of surfaces on germination and establishment of J.trifidus? Six Open Juncus trifidus heath sites were used to study the effects of disturbance level on the vegetation and the structure of the J.trifidus population. Significantly lower total plant and J.trifidus cover were only found where the level of disturbance was high. However, moderate disturbance was associated with an increase in smaller J.trifidus individuals, and the number of these and other sized individuals decreased with increasing disturbance. These results were supported by the findings of two studies of footpath disturbance. Wind tunnel experiments showed that surfaces modify microenvironmental conditions with consequent effects on the germination and early growth of J.trifidus. Changes in rates of water loss were important, and a thin layer of gravel benefited J.trifidus establishment. In general, disturbance was thought to influence the recruitment of individuals into the J.trifidus population and thereby modify population structure. The overall dynamics of Open Juncus trifidus heath were discussed and suggestions made for future work.

Snowbed vegetation in Scotland is dominated by a distinctive bryophyte flora. This type of vegetation is on the edge of its European range in Scotland. Due to its dependence on late snow lie, snowbed vegetation may be threatened by climate change. Snow is a very efficient scavenger of atmospheric pollutants. Due to its melt dynamics these pollutants are released to the underlying vegetation in a highly concentrated 'acid flush'. A survey of changes over this century in the tissue nitrogen content of K. starkei, a snowbed bryophyte, showed this to reflect increases in NOx emissions. A regional survey of tissue nitrogen content of K. starkei across Scotland did not reflect mapped patterns of NO3 deposition. This is because pollutants are concentrated in snowbeds as compared to other systems. Present day tissue nitrogen concentrations in K. starkei are up to 50% higher than those of other upland bryophytes. Sensitivity of K. starkei to acidic pollutants in subnivean conditions and at various stages subsequent to exposure from under the snow has been investigated. Results demonstrate that these pollutants, when received in realistic present day amounts, result in physiological damage to the bryophytes whether they are received by them under the snow, just emerged from the snow, or out of snow cover. Damage was greatest in plants from out of the snow. However, the ability to recover following damage is also much greater in these plants than those which become damaged while still under snow. Damage is both direct, and indirect, affecting the ability of the plants to withstand other stresses. Because of enhanced exposure to pollutants, snowbed vegetation is unlikely to be protected by the current critical loads approach to pollution control. The possible combined effects of pollution and climate change on snowbed bryophyte communities are discussed.

It has been suggested that wood mice <I>Apodemus sylvaticus</I> (L.) inhabiting maritime sand-dunes in NE Scotland "<I>work harder</I>" (Gorman 7 Akbar 1993; p395) than those in nearby deciduous woodland. The main objective of this study was to investigate the daily energy budgets of free-living adult male <I>A.sylvaticus</I> from these two habitats. Live-trapping was used to compare the population dynamics and demography at the two sites. Although similar temporal fluctuations in density were found at both sites, <I>A.sylvaticus</I> in the woodland maintained higher densities. Body masses of wood mice were significantly heavier in the breeding compared to the non-breeding season and in the woodland compared to the sand-dunes. <I>Apodemus sylvaticus</I> from the woodland had lighter kidneys, longer small intestines and longer stomachs than mice from the sand-dunes. It was suggested that these variations reflected site differences in food availability and diets of the mice. Apparent absorption efficiencies of dry matter and energy for <I>A.sylvaticus</I> fed mealworms, wheat grain or "All-bran" diets did not differ across sites. Thus, the larger gut capacity of mice living in woodland may facilitate a faster ingestion rate, possible due to the higher food availability, while maintaining constant absorption efficiency. The heavier kidney mass and greater water flux of <I>A.sylvaticus</I> on the sand-dunes were probably associated with the relatively higher water content of their diet, which was dominated by invertebrates, compared to the seed diet of mice in the woodland. At the woodland site, the breeding season was associated with an increase in home range size and distance travelled relative to the non-breeding season. Although the total time spent active did not differ between seasons, diurnal activity was only exhibited during the breeding season (<I>ca</I>. 60 mins).

This study investigated the ecology of red and roe deer in a protected area of Central Apennines in Italy in 1992 and 1993, with particular reference to the habitat selection in relation to food resources and the botanical composition of the diet of the two deer species and their inter-relationships. In addition, deer population densities were estimated and their impact on woody vegetation was assessed. Habitat selection by red and roe deer in relation to food resources of habitats was investigated by pellet group counts. Red deer used a wider range of habitats, preferring Deciduous Wood in winter and Rocky and Steep Meadow in summer and autumn, while roe deer selected woods with thick undergrowth and conifers, with no marked seasonal variation. Both deer species avoided Beech Woods, which provided the least undergrowth and cover. Red deer showed a higher use than roe deer of all habitats except of Mixed Wood where it was similar. The patterns of habitat selection shown by the two deer species seemed to reflect their ecological requirements, although seasonal interspecific competition was suggested for the use of some habitats. The diet of red and roe deer was investigated by microhistological analysis of faeces. An anatomical key of food plants was created to identify food categories. Red deer ate mainly Monocotyledonous Herbs, especially in winter, as well as Deciduous Trees (in autumn) and Conifers (in spring). Dicotyledonous Herbs and Shrubs were the least frequent food category, eaten mainly in summer and autumn. Roe deer diet was composed relatively equally by all the food categories, with little seasonal variation. However, Deciduous Trees and Conifers, combined, were the most consumed foods. Within the tree food categories, roe deer showed a preference for Conifers, whereas the selection for them by red deer probably depended on the seasonal availability of Deciduous Trees.

This thesis investigates and compares the population performance and diet of roe deer (<I>Capreolus capreolus</I>) inhabiting farm woodland and large commercial forestry plantations in North East Scotland, and investigates roe deer use of agricultural fields surrounding farm woodland. Feeding damage by roe deer to cereal fields and conifers was considered, along with the potential of Wolf (<I>Canis lupus</I>), Bear (<I>Ursus arctos</I>), and Lynx (<I>Lynx lynx</I>) odours as roe deer feeding repellents. There was found to be no significant difference in the reproductive performance between farm woodland and forestry plantation habitats. No difference was found between habitats, in the number of ovulations (<I>Corpora lutea</I> present), pregnancy rates (embryos present), and number of kids present per doe. Reproductive performance was found to be related to maternal body weight, and no significant difference was found in either the body weights, or body condition, between habitats. It was concluded that farm woodland and forestry plantation habitats are of similar quality for roe deer, and that new farm woodland is likely to be rapidly colonised by roe deer dispersing from farm woodland or forestry plantation habitats. Rumen and faecal analysis, however, indicated that the diet of roe deer inhabiting farm woodland differed from roe deer inhabiting forestry plantations. The forestry diet was similar to that found by previous British studies, and was characterised by forbs, heaths, and few graminoids. In contrast roe deer occupying farm woodland predominantly ate agricultural cereals, with few forbs and heaths. Cereals were the single most important component of the farm woodland diet, and formed 36.8% over summer, and up to 68.8% over winter, of the volume of the rumen contents. Conifers were eaten in similar proportions in both habitats, contributing up to 4% over the summer, and up to 12% over the winter, of the volume of the rumen contents.

The biology of juvenile Atlantic salmon and brown trout was studied within a small section of the Girnock Burn, a tributary of the Aberseenshire Dee. Salmon parr and brown trout were captured, measured, weighed, individually marked, and repeatedly recaptured over the period 1985-1988. Physical resources were mapped in detail within the study section, and were found to be patchily distributed. Patch size was small, typically a few square metres. Significant differences in micro-habitat existed between young and old fish, between sympatric Atlantic salmon and brown trout, and between seasons. All fish avoided shallow waters and fine substrates. Sites that combined low bottom velocities with access to fast surface currents were significantly selected by both salmon parr and brown trout. These sites supported the highest local densities of fish, and held the largest individuals. As drift-feeders, salmon appear to choose sites during the growing season that maximise food intake and minimise energy expenditure. Growth of juvenile salmon was related to water temperature, being rapid from March to June, slowing down from July to September, and in some fish ceasing altogether from September to March. In the fall, salmon 0&'43 moved to parr habitat, resources became more limited, and some parr left the stream. At low temperature, juvenile salmon hid within the substrate in riffles, while brown trout overwintered around large rocks and in a spring-fed tributary. Habitat overlap between salmon 0&43 and older parr increased from summer to fall while habitat overlap between salmon and trout decreased. Juvenile salmon were spatially aggregated. Highest densities were found at favourable sites near surface boulders. The area used by fish and the proportion of fish found at unfavourable habitats increased with fish density. Habitat selection and the spatial distribution of the population appeared to be density-dependent. Site-fidelity was shown both by salmon parr and brown trout, though in general fidelity was stronger in salmon parr. The majority of recaptures of salmon parr and brown trout were close to previous capture locations in the stream. Long-term site-fidelity was also shown by salmon parr from one year to the next. Two groups of fish, differing in their degree of mobility, were apparent both for salmon and trout: a large, stationary group, and a smaller, more mobile one. The groups were not distinguished by size, season, or stage of sexual maturity. Movements out of the burn were confined to the fall and spring. Immature parr, mostly females, migrated downstream in the fall, but were not yet adapted to full-strength seawater. Mature males performed complex upstream and downstream movements in the fall, and some of these fish homed back to their summer sites after spawning. For many males, maturation in the fall was followed by smolting in the spring, but repeated maturation was also common. Homing success in displaced fish was related to fish size and to the direction of displacement. Homing was better among larger than smaller fish, and it was better among fish displaced downstream than those displaced upstream. Fish rendered anosmic showed a greater tendency to stay where released, and homed less successfully than intact fish. The home areas of salmon parr were smaller than those of brown trout. The home areas of individuals overlapped and were unrelated to fish size; they appeared to be inversely related to substrate size and to current velocity. By spending most of their lives in a small area of the stream, juvenile salmon and brown trout may only interact with a few conspecifics. The population appears to consist of a number of small social units, rather than a single freely mixing group. The fate of individuals may be determined by very localised environmental and social conditions.

This study investigated the ecology of three species of Zygaenid or Burnet moth, the Narrow bordered five spot burnet or Talisker burnet (<I>Zygaena lonicerae jocelynae</I>) a sub-species confined to three locations on the Isle of Skye, the Transparent burnet (<I>Zygaena purpuralis</I>) restricted to between 25 and 30 'colonies' on the west coast of Scotland (also described as a sub-species) and the Six spot burnet (<I>Zygaena filipendulae)</I>, which although common is restricted to coastal areas in Scotland. The burnets of Scotland have been the subject of some conservation concern given their limited distributions in Britain and the discovery that the New Forest burnet colony in Argyll was under severe pressure from an increase in sheep grazing. This concern has focused attention on the other restricted burnets. All three species occur at Talisker, Isle of Skye, where autecological studies of the species were undertaken. This data could then be examined in terms of the requirements of the three species. Grazing levels are crucial to the ecology of all three of the species and it is the variation in the vegetation structure occurring at each specific area within Talisker which allows the three species to co-occur there. The Talisker burnet is restricted to ungrazed undercliff grassland where its foodplant, <I>Lathyrus pratensis </I>grows in abundance. Within the ungrazed areas there is considerable variation in the vegetation heights available and in areas were the vegetation is short the larvae are found relatively high in the vegetation, as are the pupae of this species. Genetic analysis of populations of <I>Z. purpuralis </I>indicate that they are quite diverse, having maintained a high rate of genetic variability between populations which are relatively close and indicates that populations greater than 10 kilometres apart have very little interchange. To conservation population of <I>Z. purpuralis</I> grazing is required on several <I>Z. purpuralis </I>sites to maintain the short vegetation required, while at Talisker <I>Z. lonicerae</I> requires early successful areas which are created by natural rockfalls and soil slippage.

A constitutive <I>lux</I>-marked biosensor, <I>E. coli</I> HB101 pUCD607, was developed as a novel, acute, 'fingerprinting' technique by which individual pollutants, and whole effluents, could be identified and quantified. In this way, the advantages of both chemical analysis and toxicology were encompassed in one technique. An assay technique for obtaining temporal response curves was developed by injecting the biosensor into toxic samples, and then measuring luminescence continuously for 5 minutes. An algorithm was formulated, and implemented as a simple computer program, by which 'unknown' temporal response-curves could be compared against a database of 'reference' pollutant response-curves. Two different techniques were derived for analysing the output data; a binomial (yes or no) 'best-fit' technique, and a continuous 'comparative-fit' technique. These techniques were validated by 94% identification success for inter-comparisons of 7 pollutants, each at 5 concentrations, and 100% identification success of four pollutant spikes in three environmental samples. The technique was tested on three complex environmental effluents. Cu was identified as the main pollutant in a distillery effluent, and Zn the main pollutant in a metal processing effluent. Identification was successful despite the presence of complexing substrates and co-pollutants. The technique could also differentiate between three complex effluents and four artificial complex effluents without the need for actual pollutant identification. In conclusion, this technology can be used to 'fingerprint' any form of pollution that is bioavailable to the biosensor, without the need for prior knowledge of sample chemistry. This represents a potentially valuable tool for toxicity assessment and screening.