Detailed field mapping (Map, Appendix B) has been conducted in and around the boundaries of a 14x18km, volcanic complex 35km northeast of Molteno in the Eastern Cape Province, South Africa. The structure is interpreted as a subsidence structure, and is filled with two volcaniclastic breccias, numerous lava flows, a number of sedimentary facies, and lies on a base of Clarens Formation overlying Elliot Formation rocks. This is an important study because 'widespread, voluminous fields of basaltic breccias are very rare (see Hanson and Elliot, 1996) and this is the first time that this type of volcanic complex and its deposits have been described. Detailed analyses of the two volcaniclastic breccias revealed changes in colour, clast types, clast sizes, and degree of alteration over relatively short distances both vertically and laterally within a single breccia unit. The variation in clast sizes implies a lack of sorting of the breccias. The lower of the two volcaniclastic breccias fills the subsidence structure, and outcrops between the Stormberg sedimentary sequence and the overlying Drakensberg basalts and was produced from phreatomagmatic eruptions signalling the start of the break-up of Gondwanaland in the mid-Jurassic. The upper volcaniclastic breccia is interbedded with the flood basalts and is separated from the lower breccia by up to 100m of lava flows in places, it is finer-grained than the lower volcaniclastic breccia, and it extends over 10km south, and over 100km north from the volcanic complex. The upper breccia is inferred to have been transported from outside the study area, from a source presumably similar to the subsidence structure in the volcanic complex. The pyroclastic material forming the upper breccia was transported to the subsidence structure as a laharic debris flow, based on its poorly sorted, unwelded and matrix-supported appearance. However, both breccias are unlikely to have been derived from epiclastic reworking of lava flows as they contain glass shards which are atypical of those derived from the autoclastic component of lava flows. The breccias are therefore not "secondary" lahars. There is also no evidence of any palaeotopographic highs from which the breccias could have been derived as gravity-driven flows. Based on the occurrence of three, 1m thick lacustrine deposits, localised peperite, fluvial reworking of sandstone and breccia in an outcrop to the south of the subsidence structure, and channel-lags encountered only in the upper units of the Clarens Formation and only within the subsidence structure, the palaeoenvironment inferred for the subsidence structure is one of wet sediment, possibly a shallow lake, in a topographic depression fed by small streams. Magmatic intrusions below the subsidence structure heated the water-laden, partly consolidated Clarens Formation sandstones, causing the circulation of pore fluid which resulted in the precipitation of minerals forming pisoliths in the sandstones. Intruding magma mixed, nonexplosively, with the wet, unconsolidated sediments near the base of the Clarens Formation (at approximately 100m below the surface), forming fluidal peperite by a process of sediment fluidisation where magma replaces wet sediment and cools slowly enough to prevent the magma fracturing brittly. Formation of fluidal peperite may have been a precursor to the development of FCIs (Fuel Coolant Interactions) (Busby-Spera and White, 1987). The breccias may represent the products of FCIs and may be the erupted equivalents of the peperites, suggesting a possible genetic link between the two. The peperites may have given way to FCI eruptions due to a number of factors including the drying out of the sediments and/or an increase in the volume of intruded magma below the subsidence structure which may have resulted in a more explosive interaction between sediment and magma. Phreatic activity fragmented and erupted the Clarens Formation sandstone, and stream flows reworked the angular sandstone fragments, pisoliths and sand grains into channelised deposits. With an increase in magmatic activity below the subsidence structure, phreatic activity became phreatomagmatic. The wet, partly consolidated Clarens Formation, and underlying, fully consolidated Elliot Formation sediments were erupted and fragmented. Clasts and individual grains of these sediments were redeposited with juvenile and non-juvenile basaltic material probably by a combination of back fall, where clasts erupted into the air fell directly back into the structure, and backflow where material was erupted out of the structure, but immediately flowed back in as lahars. This material formed the lower volcaniclastic breccia. A fault plane is identified along the southwestern margin of the subsidence structure, and is believed to continue up the western margin to the northwestern corner. A large dolerite body has intruded along the inferred fault plane on the western margin of the structure, and may be related to the formation of the lower volcaniclastic breccia, either directly through fluidisation of wet sediment during its intrusion, or as a dyke extending upwards from a network of sill-like intrusions below the subsidence structure. Geochemical analysis of the Drakensberg basalt lava flows by Mitchell (1980) and Masokwane (1997) revealed four distinct basalt types; the Moshesh's Ford, the Tafelkop, the Roodehoek, and the Vaalkop basalts. Basalt clasts sampled from the lower volcaniclastic breccia were shown to belong to the Moshesh's Ford basalt type which does not outcrop in situ within the subsidence structure. This implies that the Moshesh's Ford basalts were emplaced prior to the formation of the lower volcaniclastic breccia, and may have acted as a "cap-rock" over the system, allowing pressure from the vaporised fluids, heated by intruding basalt, to build up. The Moshesh's Ford basalt type was erupted prior to the resultant phreatomagmatic events forming the lower volcaniclastic breccia.