Although this paper is a regional study of some aspects of solutional processes, karst hydrology and limestone landform development, some systematic problems are also considered in the context of the southern Pennines. For instance, it seems that by water sampling at several points it is possible to estimate mean calcium carbonate concentrations for hypothetical catchments of 100 per cent and 50 per cent limestone. These calculations may be more useful for comparative purposes than the observations made at just one point in a study area. Despite wide variations in drainage area and proportions of limestone in these areas, there is a close correlation between drainage area and the estimated mean discharge. This suggests that the underground watershed could approximate to the watershed on the land surface in more places than might be anticipated in a limestone terrain. The limestone appears to return to the surface streamflow at least as much precipitation as the gritstone and impermeable shale outcrops, possibly because the younger strata encircling the limestone outcrop produce a ‘ponded karst’ effect. A denudational loss from the limestone of 75-83 cubic metres per square kilometre per annum is suggested as reasonably typical of present-day processes. Estimates of denudational loss have little bearing on the interpretation of relief unless some idea can be gained of the location of the zone of greatest removal. One of the primary concerns of the paper is the consideration of five main possibilities. It seems that solution of valleys by the near-saturated river water is unlikely. As more limestone is being removed in a single year than the volume of known cave systems, it seems improbable that cave formation is the dominant expression of solutional activity. The implication of high carbonate contents, and of the presence of cave formations at shallow depth in the limestone, is that percolating water does not produce evenly distributed widening of joints and bedding planes to an unlimited depth. The restricted occurrence of ‘clints’ and the slow flow-through rates of karst water rule out the possibility of joint widening at the surface. It is concluded that a considerable proportion of the limestone dissolved is probably removed from the surface, and four independent lines of evidence support this conclusion. One implication of the scale of surface lowering of the limestone is that it may be difficult to recognise relict features in the present-day limestone relief. Doubt is therefore cast on the assumption that erosion surfaces on limestone are well preserved because absence of running water restricts surface erosion. The possible scale of surface lowering could reveal at the surface former underground routes. It is suggested that the undulating plateau surface of the limestone tract of the southern Pennines, although showing many examples of development with little regard to structure, is explicable without reference to Davisian hypotheses of landform evolution. The land surface may have developed by differential solutional lowering during progressive exposure from beneath a protective shale cover. This suggested evolution is analogous to the development of the pedestal of ‘perched blocks’: both depend on the protection which a shale cover affords the underlying limestone. However, if impermeable strata were thin or locally absent, water percolating through the cover may have caused at some points, solution of the unprotected limestone. Lowering of the limestone may therefore have started in some localised areas long before the limestone itself was exposed at the surface. The hollows of the ‘pocket deposits’ could have developed as ‘covered karst’ features in this way.