Drystone walls have been used extensively around the world as earth retaining structures wherever suitable stone is found. Commonly about 0.6m thick (irrespective of height), there are about 9000km of drystone retaining walls on the UK road network alone, mostly built in the 19th and early 20th centuries, with an estimated replacement value in excess of £1 billion[1]. Drystone wall design is traditionally empirical, based on local knowledge of what has worked in the past. Methods vary from region to region, driven by both custom and the nature of the materials available. Design is not necessarily optimised, and includes unknown margins of safety. There is a recognised need for guidance on the assessment and maintenance of dry stone retaining walls, as no suchdocumentscurrentlyexist. Thisthesisdocumentstheconstructionofaseriesoffull-scaletestsdesignedto provide sufficient information to validate current theoretical and numerical analysis techniques. The development of a unique test rig is detailed, in addition to the testing regime and results from a programme of five 2.5m high drystone retaining walls. The walls were subjected to localised surcharging and foundation movements, recreating the conditions that many in-situ walls are subject to. Movements such as toppling, bulging and sliding were observed, and recorded using a broad range of instrumentation. This has provided high quality, quantitative data relating to the factors which influence these mechanisms, and their affect on wall stability. Also documented are the associated laboratory tests which have been conducted to determine the mechanical properties of backfill and the walls themselves, as well as the manner in which they interact together. To assist in the analysis of these full-scale tests, a limit equilibrium program has been developed. This package allows the rapid generation of a wall of any size and constructed with any materials. The limit equilibrium program has then been used in conjunction with the data from the full-scale and laboratory tests to analyse observed drystone wall behaviour. These include the phenomena of toppling, bulging, bursting, sliding and individual block rotation. In each case, the underlying causes of such movements have been determined, and the critical parametersidentified.

Following a review of the paradigmatic context of broch towers in 2012, a revised standard model (the RSM) was defined. The then prevailing paradigm supports a view of broch remains as single monuments of highly variable form that continued in use over perhaps a millennium or more, without significant modification of their original tectonics i.e. their people/constructed-space relationships. This thesis challenges the pre-2012 paradigm by testing the hypothesis that brochs were built to the standard canonical form of the RSM and that their apparent diversity results from anthropic and, or natural modification, not design variability. The fieldwork tests could but did not find refutation of these hypotheses in the observable evidence and offered more profound interpretations of several surviving feature-types. The loading on the stone lintels of the entrance passage through the massively built outer wall and the structurally overladen inner wall created a major structural challenge, evoking a complex engineering solution. Its elements were individually noted pre-2012 but the significance of the engineering response to compression management had not been identified. This structural response was necessary for a tall structure with massive loads, and meaningless without one and its elements are therefore, jointly and severally, clear diagnostics of a broch tower. The entrance engineering was probably the inspiration of one individual or of a small group of master mason-types, not vernacular responses, contra the 2012 paradigm. Isolated stacked voids high in the inner wall are relict features indicative of significant modification of the inner wall. Other anomalous features are shown to be relict stacked void fragments. The East/West differences in brochs across Scotland have long been identified and these are generally attributed to their lithologies. Accepting that, this thesis argues that the principal differences are attributable to the social processes that gave rise to centralisation of settlement around, in and over brochs in the east and north, possibly during the first century BC, and the absence of centralisation in the west; perhaps also explaining the differences in the scale and composition of the artefact assemblages between the two zones. The canonical form facilitates calculation of the relative social costs of broch building for hard-rock and sedimentary stone types. This indicates that the costs of building, increase between 16-, and 32- fold over the buildable range of brochs. Constraints of design down-scalability, design weakness in ground loading, and design cost were major constraints on the mind of the broch builders. Canonicity and the limitations of drystone building technologies predicated specific forms of decomposition on the canonical broch, further complicating their autobiographies and their conservation: the main challenge now being that of finding ways to conserve the evidence for a sequence of processes while conserving the products of those processes.