Over the last half century, the widespread decline of bumblebees across the agricultural landscapes of Western Europe and North America has been well documented. This decline has undoubtedly been driven to a large extent by the intensification of agriculture, which has fragmented landscapes and removed large areas of suitable foraging habitat, nesting and hibernation sites. Consequently, some of the rarest Bombus species now persist only in isolated pockets of semi-natural habitat, which have been subjected to little agricultural intensification. Of the 25 Bombus species native to the UK, three species have gone extinct in recent decades and several others are severely threatened. Remaining populations of the UK’s rarest bumblebee, Bombus distinguendus, have become strongly associated with florally-rich machair grassland habitats found only in the North and West of Scotland and Western Ireland. Machair, a unique habitat that forms on soils rich in shell sand, has been maintained by rotational agricultural practices implemented by crofters. However, recent changes in crofting practices, which include the intensive grazing of machair in some areas, or conversely the abandonment of machair management all together in others, have resulted in sections of machair that have become degraded and consequently exhibit low floral abundance and species diversity. This has significant implications for species such as B. distinguendus, which have for the most part come to rely of the florally-rich swards of machair grassland. This thesis aimed to develop a greater understanding of how machair grassland habitats are utilised by foraging bumblebees, including B. distinguendus, and in turn examined the potential for restoring degraded areas of machair via a variety of methods. The research presented here examines the influence of current crofting practices on the abundance of bumblebees and their forage plant species and combines this information with a detailed exploration of the machair seed bank and potential machair restoration treatments. The specific foraging requirements of B. distinguendus were found to be similar to those of other long-tongued bumblebee species and the provision of plants from the Fabaceae family was found to be of particular importance. Current crofting practices implemented in the North and West of Scotland were, on the whole, found support low numbers of foraging bumblebees. Similarly, existing habitat management schemes, designed to provide early cover for corncrakes and foraging resources for bumblebees, were found to be largely ineffective in attracting B. distinguendus, when compared with florally-rich machair habitat. In addition, this research suggests that the existing machair seed bank is unlikely to provide a sufficient resource for reinstating florally-rich habitat to degraded areas of machair. However, this thesis has demonstrated that it is possible to implement seed mixes on machair which can reinstate species typical of machair plant communities and which also attract high numbers of foraging bumblebees. The findings of these habitat assessments and restoration trials are examined in full in the following chapters and implications for the future management of wildflower-rich machair are discussed throughout.

The machair is a coastal grassland system found only in parts of northern and western Scotland and Ireland. Despite its limited geographic distribution, machair landscapes have high ecological, geomorphological, and cultural significance, as recognised by numerous conservation designations and legislation. In January 2005 a severe storm caused extensive damage in the Outer Hebrides, drawing attention to the sensitivity of the machair coast to erosion. The aim of this research was to investigate variations and trends in the sensitivity of three field sites within the South Uist machair to soil and coastal erosion, and to interpret measured change alongside analysis of historic climate data. Two of the sites selected, Cille Pheadair and Staoinebrig, experienced some of the most dramatic geomorphological changes associated with the 2005 storm, while the third site, Milton, appeared to be more resistant to change. A combination of fieldwork, laboratory tests, and archive work was used to obtain and analyse information about sediment budgets, shoreline indicator change, and sediment erodibility, along with contextual climatic information. A key result of this work is the provision of a detailed framework of short-medium term cyclical changes and fluctuations in the coastal change, which provides a context for interpreting and responding to longer term trends in erosion and/or accretion. Results indicated high spatial and temporal variability in the erodibility of machair soils and landforms, with no clear relationship between climatic factors and rates of erosion. Considerable short-term variations in beach volume and the position of dynamic shoreline indicators caution against the relaibility of using ‘snap-shot’ historic datasets to infer long-term rates of change. It is proposed that the machair landscape currently functions in a state of highly dynamic equilibrium, which has been maintained over the last ~130 years. While storm events such as the January 2005 storm have locally dramatic consequences, they do not appear to have disrupted the overally physical and ecological functions of the system. This contribution is particularly timely given current concerns for the future of the machair landscape under predicted sea-level and climate change scenarios, and the potential for inappropriate hard-engineering responses to the perceived risk.