This thesis investigates past structure and dynamics of native Caledonian pine woodland, representing part of the western fringes of the northern European boreal woodlands. The biogeographical extent and Holocene history of the Scottish pine woods are well studied, yet questions remain at finer scales. This thesis is concerned with two factors over the recent Holocene oceanic period; (i) the long-term ecology within the woods; the spatio-temporal dynamics, the canopy structure, and community composition and continuity; and (ii) the former extent of the woods, especially the temporal pattern of the inferred easterly contraction in woodland over recent history. The importance of these factors - to ecologists, challenged with understanding the theories of equilibrium and non-equilibrium processes in long-lived woodland communities - and to conservation managers - challenged, by policy directives, with implementing the restoration and expansion of native woodlands - is discussed. Equally the identification and quantification of ecological detail over ecologically relevant temporal and spatial scales is an important challenge for palaeoecology. The thesis therefore applies fine spatial resolution pollen analyses over a network of sites, within, and at the western edge of, the extant woodland zone in east Glen Affric. Correlation between these individual site histories develops a detailed view of the grain and extent of woodland within the landscape, previously missing from western pine woods. Careful attention is paid to the interpretive potential and limitations of fine resolution palaeoecology; especially with regard (i) to techniques that can identify and spatially quantify stand-scale structure and community composition via reference to modern analogues; and (ii) to defining chronologies and elucidating rates and patterns of temporal change. The opportunities and limitations of the technique are explored and discussed, to ensure an understanding of the rigour and potential of the palaeoecological contribution to ecological research and to provide an evidence base for conservation. The application of 210Pb dating using the CRS and CIC models is explored. The CRS model is confirmed as suitable for peat deposits, but its use may mask fluctuations in peat sedimentation rate, which may be illustrated by the CIC model. The value of multiple 14C assays for each core and a need for a new approach to chronologies for application to fine-scale palaeecological studies is discussed. The relevant source area around the small basins in this semi-open pinewood is tentatively confirmed at 20 m from the pollen source. Tentative pollen productivity estimates for five key taxa in this ecosystem are presented. The temporal stability of native woodland in Glen Affric is confirmed to the eastern part of the extant woodland zone; but challenged to the west, where the open and semi-open landscape has a long history. Woodland diversity decreased over the last c. 200 years, and past woodland also shows a greater ground flora diversity. The western extent of the Caledonian woodland in this landscape may have changed little in the last c. 4000 years. Fluctuations, some clearly cyclical, in heath, and in woodland, communities are identified in the pollen record. The former may be aligned to changes in grazing regime or climatic shifts; and the latter to autochthonous shifts important in the maintenance of suitable edaphic conditions for the continuity of woodland. Former woodland is confirmed as likely to have been open in structure and mixed in tree species composition. The results presented here suggest that some caution should be applied to use of the term Caledonian, or native ‘pine’ forest: ‘Caledonian forest’ may better reflect the heterogeneity of past forests, particularly the importance, and persistence, of birch. The implications for conservation management and restoration are discussed. It may be difficult to establish a sustainable woodland to the west of the extant stands, and any pine woodland here may need to be mixed with stands of broadleaved trees to maintain or restore soil structure and ecological function. The landscape to the west may have been open for several thousands of years, and consideration of this is required when managing for the future to prevent loss of biodiversity.