Recognition of unsteadiness in flow velocities is an important factor in the identification of tidal deposits. One consequence of such unsteadiness is the production of reactivation surfaces -erosion surfaces developed within a cross-stratified set. Reactivation surfaces can be produced in a unidirectional flow system, by the migration of a faster-moving megaripple over a slower one ("overtaking"). The form of the surface is dependent upon the relative sizes of the two megaripples concerned and the flow conditions under which they migrate. Other types of reactivation surfaces are produced by the action of a subordinate current, which erodes the lee side of the dominant-current megaripple. The morphology of the reactivation structures depends on the size of the megaripples concerned (the examples considered here had heights in the range 0.2-2.5 m) and, most importantly, on the relative strengths of the dominant and subordinate currents. Where the subordinate current is relatively weak, only the tops of the dominant-current foresets are eroded. With increasing subordinate-current strength, the whole foreset height is subjected to erosion, and a large angular discordance is produced between the large-scale foresets and the reactivation surface. Such strong erosion drastically alters the megaripple relief--the lee-side slope is much reduced. Recovery of the lee side during the subsequent dominant-current phase is gradual, as can be seen from the development of small-scale structures immediately overlying the reactivation surfaces. The regular variation in the strength of the subordinate current during the neap/spring cycle is often reflected in the morphology of the reactivation surfaces. In such examples, the reactivation surfaces formed during spring-tide periods have a large angular discordance with the foresets, indicating that there was considerable erosion by the subordinate current. Reactivation surfaces formed during periods between spring and neap, in contrast, are developed only in the upper parts of the foresets, suggesting that subordinate-current erosion was relatively modest, while the deposits of the neap tides themselves show no evidence of any erosion by the subordinate current.