A 1000 m thick early Mississippian carbonate supersequence, the “Carboniferous Limestone” of southwest Britain, consists of three third-order depositional sequences. These comprise parasequences in various configurations, and the whole forms a carbonate ramp stack. Within this framework five major oolitic carbonate sandbodies developed: (a) Castell Coch Limestone, (b) Stowe Oolite, (c) Brofiscin Oolite, (d) Gully Oolite, and (e) High Tor Limestone. The depositional regime was storm- and wave-dominated throughout and the major sandbodies represent a range of progradational carbonate beaches, barriers and detached subtidal shoals. Analysis of the three-dimensional shapes and distribution of these five examples shows that they evolved to produce three major carbonate sandbody geometries: (a) strings, (b) sheets, and (c) wedges. These geometries are characterised using the five field examples and offered as a template which may assist in the exploration and reservoir modelling of petroleum-rich high-energy ramp systems. Progradation, for up to 40 km, of barrier islands (Stowe Oolite) and beach-ridge plains (Gully Oolite Formation) generated strings, and “thick” sheets individually up to 10–20 m thick. “Thin” shoreface-retreat carbonate packstone/grainstone sheets up to 5 m thick (High Tor limestone) developed during transgressions as veneers across flooding surfaces. These are comparable with sheet sands developed in siliciclastic shelf depositional systems. Progradation, for up to 30 km, and vertical aggradation of shoreline-detached oolite shoals (Castell Coch limestone, Brofiscin Oolite), generated basinwards-expanding or thinning wedges up to 30 m thick. Tectonically controlled stacking of strandplain sheets produced a composite carbonate sandbody up to 80 m thick (Gully Oolite). The intrinsic (sedimentary) and extrinsic (eustacy, tectonism, climate) factors which controlled these sandbody geometries are addressed. Establishing the positions of the sandbodies accurately within depositional sequences allows them to be located within inferred seismic sequence geometries and provides one possible solution to the difficult problem of predicting carbonate facies distribution in subtle stratigraphic plays. In this ramp system, the most homogeneous sandbodies (up to 30 m grainstones), with greatest reservoir facies potential, are represented by shoal-belt wedges. Potential grainstone reservoir facies in the prograding shorelines are limited to the upper parts of individual shoreface sequences (max. 10 m grainstones). For shoreline carbonate sandbodies, the greatest reservoir and stratigraphic trapping potential exists in the earliest ramp parasequences where enveloping offshore sediments are siliciclastic mudstones. In later stages, potential seals are likely to be less reliable, low-porosity outer ramp carbonates.