Ground Improvement using vibro stone columns is gaining increasing acceptance on marginal soft clay sites as a sustainable foundation solution, particularly for lightly loaded low-rise structures supported by shallow, narrow footings. Most experience in this context however has been with widespread loads and use of the wet top-feed stone column technique, which has now been largely superseded, on environmental grounds, by the dry bottom-feed technique, and for which no significant published field trial data currently exists in deep soft clay deposits in the context of shallow, narrow footings. This research is therefore principally concerned with evaluating both the ground response to installation of partial depth vibro stone columns using the dry bottom-feed method in a deep moderately sensitive soft clay soil, together with the influence of parameters such as stone column spacing and length, founding depth within a thin surface 'crust', and also foundation shape on the performance of narrow footings subsequently constructed and subjected to incremental loading, over the installed stone columns, at the Bothkennar soft clay research site in Scotland. Comparisons are made with footings constructed within the surface 'crust' at Bothkennar without stone columns. Whilst stone columns were satisfactorily constructed with the dry bottom-feed technique at Bothkennar, it was evident that the vibroflot should not remain in the ground for longer than is necessary, in order to avoid excessive soil disturbance. For this reason construction of partial depth stone columns to a more uniform diameter, without construction of an 'end bulb', is advocated. Stress ratio was found to increase significantly with increasing length of stone column and also applied load, up to a maximum value of around 4.0. Moreover, for a trial footing founded at the base of the 'crust', stresses attracted by the columns were higher than all other columns where founding depth (level) was at shallower depth in the crust. A significant stress transfer was also measured beneath the toe of columns intentionally installed shorter than the minimum design length predicted by the Hughes and Withers (1974) approach at all iii applied loads, but not for columns equal to, or longer than minimum design length, confirming the predictions of this laboratory-based approach at the field scale. The stress measurements recorded by the field instrumentation demonstrate that the behaviour of the composite stone column-soil-foundation system is complex, with simultaneous and interdependent changes in pore pressures, soil stress ratios and resulting stiffness of both soil and columns. Whilst observed settlements exceeded those predicted, with larger foundation settlements observed at low applied loads over stone columns than at the same loading level in untreated ground, principally due to soil disturbance and accelerated consolidation effects during initial loading, at higher applied loadings however the stone columns significantly reduced the rate and magnitude of settlement compared to a foundation in the untreated 'crust'. It is therefore clear that the stone columns 'reinforced' the weak soil, providing a significantly increased factor of safety against bearing failure.