Duke (1985b) argues that ‘most examples (of hummocky cross-stratification) were formed by tropical hurricanes.’ His statement is based on the assumption that ‘hurricane-generated surface gravity waves form powerful oscillating or multidirectional flows at the water-sediment interface which do not possess a significant unidirectional component.’ It is true, as one of us has previously stated, that hurricanes are rapidly-moving, short-lived, localized, and infrequent systems as compared with mid-latitude storms; midlatitude storms are consequently more efficient in coupling with the shelf water-column than are hurricanes. However, Duke's argument that hummocky cross-stratification may be the result of purely oscillatory flow is untenable. His reasoning contradicts established theory about oscillatory bedforms, and his numerous examples of hummocky cross-stratification come largely from continental shelf settings where the storms (tropical or otherwise) would have created concurrent alongshelf undirectional flow as well as wave oscillatory motion. There is no theoretical or observational basis for the belief that water movement on the sea-floor during hurricanes is qualitatively different from water movement during mid-latitude storms. Consequently, hummocks are no more liable to form beneath hurricanes than they are beneath mid-latitude storms.