Several of the concepts of sequence-stratigraphic models were tested by experiments in a large flume. Using different gradients for successive runs, a physical model of a fluvial drainage basin, coastal plain, continental shelf, and continental slope was constructed. In all experiments, initial gradients of the coastal plain and shelf were identical. Systematic raising and lowering of the water surface simulated the changes in relative sea level (i.e., base level) that are similar to most sequence-stratigraphic models. Base-level changes significantly affected the shelf area but they had little effect on the fluvial drainage basin. During base-level falls, fluvio-deltaic progradation occurred. Fluvial aggradation occurred only during periods of base-level stillstand or rise. Because of the slow rate of head ward erosion of incised valleys, a significant time gap existed from the time base level first fell below the shelfbreak and the development of a cross-shelf bypass valley. During this lag time, deposition occurred on the exposed shelf to form a fluvial braid plain. Concurrently, channels developed at the shelfbreak and grew by headward erosion at various rates. Continuous incision and widening of these channels formed valleys. These valleys were not linked directly to the main drainage basin until ultimately one captured the flow from that drainage. At that point, other valleys were abandoned and ceased to grow. During the subsequent base-level rise, the single, large, cross-shelf valley filled with coarser sediments while the other valleys partially filled with finer suspended-load sediments. The main cross-shelf valley also supplied coarser sediments to the slope.