ABSTRACT
The majority of geomorphological techniques seek either directly or indirectly to provide a rate of process operation in time or space. The advent in time of an ever-improving ability to be more precise over measurement of nearly every type has broadened the base from which process statements either forwards or backwards in time can be made. D. Walling has outlined errors in the systematic sampling in time of a continuous variable. However, the advances in techniques and technical sophistication may well demand the retraction by physical geographers of statistical modelling in favour of deterministic models in most areas other than time-series approaches. To operate at one point in continuous time is the easiest way to meet this requirement, but it is not the most desirable in the context of physical geography. By contrast, experimental studies of hillslope hydrological behaviour can achieve a much closer one-to-one correspondence between experiment and nature, particularly since time and spatial scales can be identical.
