Denny’s Hole is a short multi-chambered cave, the largest of several on the lower slopes of Crook Peak. It is developed in the Clifton Down Limestone, which is here splintery, partially dolomitised and cut by a number of calcite/haematite and iron ochre mineral veins with some silicification. The base level for cave development is determined by the position of the marl facies in the Triassic Mercia Mudstone as indicated by the current outlet at Dunnet’s Springs at the south-east end of the ridge. The cave is predominantly controlled by the bedding, with anastomosing tube networks floored by platy breccia (’terrace breccias’) extending out from the First Chamber. The chambers are dome shaped, have blank up-dip terminations, walls with facets, phreatic pocketing, rock steps and unusual secondary carbonate accumulations, and appear to be incised into the bedding tube network. The links between the chambers are breached partitions with very small tubular passages and bridges. With the exception of material entering from the present entrance, there are no allogenic sediments in any of the caves, and there is an absence of typical vadose stream forms. The chambers are floored by a breccia also found in Unit 6 Picken’s Hole and Sandy Hole, which is characterised by a significant degree of calcite cementation and by the presence of iron-oxyhydroxides, terminated by banded flowstone (some of which is aragonite) and popcorn. The main phase of calcite cementation comprises pond deposits and includes unusual festoon panels of rhombohedral calcite growing from a central iron-oxyhydroxide layer, which is very similar to present deposits in Thornton’s Cave, an estavelle in west-central Florida (Florea, et al, 2011). The breccia has been formed by breakdown of secondary carbonates deposited on the wall and roof, and indicates a fluctuating environment with both vadose and sub-aqueous deposition, but the earlier parts of the sequence are not present in the cements of the terrace breccias. The features of the cave indicate a hypogene origin most probably by thermal waters rising from depth beneath the Triassic cover and mixing with more oxygenated local meteoric water. Cooling caused undersaturation and dissolution in an initial bedding tube network, with contemporaneous deposition of subaqueous calcite in the near surface zone of active degassing in Unit 6 of Picken’s Hole. The cave chambers were then incised into the bedding network as a result of near- surface condensation corrosion above a thermal pool in a closed cave environment with elevated PCO2 as base level fell. Reflooding and pond calcite deposition occurred when base level rose during interglacial conditions. Correlation with uranium series dated base level changes for the Cheddar Caves suggests that the higher parts of the system (Foxes Hole and Picken’s Hole) were active during Marine Isotope Stage (MIS) 9, with re-flooding during MIS 7, and abandonment prior to MIS 5 the last interglacial.