A wide range of rates of downstream fining in gravel-bed rivers has been reported. As a consequence, explanations of the phenomenon range from those that consider only abrasion to others that consider only selective sediment transport. This apparent confusion results in part from inconsistent definitions of downstream fining. Analysis of results from a small gravel-bed stream is used to show how the method of sediment sampling and the percentile used to characterize fining rates can affect the results. When fining is adequately defined, a theoretical relationship between fining rate and drainage-basin area is derived from established alluvial-fan relations. This relationship is remarkably consistent with a range of studies by numerous authors using often different methods. Such results suggest that fining is achieved through the mechanism of selective transport but that the overall control on fining patterns is the volume of sediment supplied from within the drainage basin. As such, the rate of size decline in sedimentary deposits such as river terraces has the potential to be used to infer past changes in climate, tectonics, and/or relative base level. Models of this are presented, and these are shown to be broadly consistent with observations made in an ancient sandstone body. Full use of changing rates of fining as paleo-environmental indicators requires further work on the rate of adjustment of fining profiles with respect to the rates of change in the forcing environmental conditions.