This paper presents a synthesis of geologic and geotechnical concepts to present a unified model of conditions controlling the development of cover-collapse sinkholes and associated ground subsidence. Appropriate engineering response to the hazards associated with collapse and subsidence requires a full understanding of the underlying mechanisms that produce such effects. The geotechnical characteristics of the overlying clay mantle and occurrence of the associated cover-collapse features are not random, but rather are directly tied to the underlying water flow routes and their development through time. The clay mantle and underlying epikarst are two components of a single system, each of the components influencing the other. This paper brings together these two aspects in terms of the author's personal experience and observations as a geologist, geotechnical engineer, hydrogeologist, and caver. A summary of the basic model follows. Much of the clay mantle and pinnacled upper surface of the epikarst form while surface drainage still prevails. At this stage, the karst underdrains are insufficiently developed to transport soils, although some subsidence into cutters occurs because of dissolutional rock removal. Soil arches and macropore flow routes associated with cutters have developed by this stage. As competent deep conduits extend into the area by headward linking, the cutters with the most favorable drains are linked to the conduits first and act as attractors for the development of a tributary, laterally integrated drainage system in the epikarst. Once the most efficient cutter drains become competent to transport soils, the depressed top-of-rock and ground surfaces characteristic of dolines develop. A given doline underdrain is likely to have multiple tributary drains from adjacent cutters, which vary in soil transport competence. Soil stiffness in the clay mantle over the limestone varies as a result of the pattern of stresses imposed as the underlying rock surface is lowered by dissolution and later as soil piping locally removes soils. In the absence of karst, these soils would have developed a laterally uniform, stiff to very stiff consistency. Where soil near the soil–bedrock interface is locally removed, however, the weight of the materials overlying this void is transferred to abutment zones on the pinnacles by soil arches. Local soil loading in the abutment areas of these arches would increase at least on the order of 50% in the case of an isolated cavity. In some cases, multiple closely spaced cutters whose soil arches have narrow, laterally constrained abutment zones bearing on the intervening pinnacles may produce substantially higher soil abutment stresses. If the clays in the abutment zones do not fail, they would respond to this increase in stress by consolidating: stiffening and decreasing in volume. The cutters spanned by the soil arches accumulate raveled soils that are "under-consolidated", the soft zones noted between pinnacles by Sowers. A simple integral of stresses analysis makes it obvious, however that no continuous soft zone exists. It is the transfer of load to the pinnacles through the stiffened abutment soils that allows these locally soft areas to exist. Soil stiffness profiles from borings substantiate this pattern. Cover-collapse features develop where soil transport through cutter drains is sufficient to remove the soils from beneath these arched areas. Two types of collapse have been observed: type 1 collapses have an upward-stoping open void whose rubble pile is removed by transport as fast as it is generated, producing a deep, steep-sided final collapse. In some cases, multiple voids in clusters can form with narrow abutments separating them. Large collapses may involve a progressive failure of several members of a cluster, including intervening pillars. Type 2 features are soil-filled voids limited in their rate of upward growth by the rate of soil removal, have little open void space, and migrate to the ground surface as a column of soft soils, finally producing a shallow depression. The type 2 features have geotechnical significance because of their effect on settlement under imposed loads. A single underdrain system may service both types of features, the behavior of particular voids being dependent on the relative efficiencies of their drains. This behavior can also change with time because backfilling of the underdrains with soil or flushing out of the soil filling can occur with changes in hydrologic or erosional regimes.