A thorough investigation of scale effects arising from refraction over a variable topography was conducted with the resultant conclusion that sufficient scale effects would be present in a ripple tank to preclude any quantitative investigation of refraction effects. Even qualitative studies can be misleading due to errors in the direction of wave propagation arising from refraction scale effects. A brief discussion of possible ripple tank scales for quantitative studies leads to the conclusion that although the topography might be molded to a sufficient accuracy, it would be costly and time-consuming. The measurement of ripple tank wave heights of sufficient accuracy to be useful for design purposes would be extremely difficult, if not impossible. (Modified author abstract).