(b) whether a stable rolling-grain type ripple precedes 

 the appearance of the usual vortex-type ripple; 



(c) the rate of ripple growth, which is needed to define 

 and predict the onset of ripple equilibrium with the imposed 

 flow; 



(d) extension or modification of the results of Mogridge 

 and Kamphuis C1972) on stable ripple dimensions, especially 

 for natural sands with larger prototype values of a and T; 



(e) conditions of two- and three-dimensional bed forms; 

 and 



(f) conditions for the disappearance of bed forms. 



Also of interest is what, if any, increase in the water velocity is 

 needed for transition from one state of ripple development to the next, 

 and how the process is modified if the bed is initially disturbed, by 

 objects or by previous ripples. 



II. EXPERIMENTAL PROGRAM AND PROCEDURES 

 1. Nature and Scope of the Experiments . 



The experiments described in this study were designed to provide 

 additional observations of ripple development with special attention to 

 those aspects just enumerated. A further aim was to make the observa- 

 tions under near-prototype conditions. This end was served by the use 

 of three natural sands in an oscillatory water tunnel able to provide 

 values of a and T over wide ranges. There were 104 experiments of 

 which 24 began from a level bed. In these 24 experiments the entire 

 ripple development from initiation of grain motion to equilibrium pro- 

 files was observed. The other 80 experiments began with profiles pre- 

 viously established and with a or T, or both, changed from the 

 previous values. Here the response of the old profiles to the new 

 values of a and T was observed. Leveling the bed required that most 

 of the water be temporarily drained from the tunnel and that the top of 

 the test section be removed. In proceeding to a new experiment with a 

 previously rippled bed, these procedures were unnecessary. The amplitude 

 a, which in any case could not be changed during operation, was always 

 selected before an experiment. With previously rippled beds the period 

 was also selected, but when starting an experiment with a leveled bed, 

 T was gradually reduced until the onset of grain motion or spontaneous 

 ripples was observed, defining Tc or Tj- . After ripples had begun to 

 grow, development was allowed to proceed to the final equilibrium bed 

 form without further change in T. In most cases, experiments continued 

 long enough to ensure that equilibrium conditions had been established. 

 Some experiments were interrupted and continued on the following day. 

 Bed forms and their development were recorded with frequent photos 

 supplemented with written notes. 



27 



