more than about one part in 100,000 of the volume displaced by the pis- 

 tons. The tests were performed for S/S^ = 0.33 and 0.64, and in the 

 ranges, T between 3 and 25 seconds and U between 30 and 150 centimeters 

 per second. 



In the permeability experiments [Lofquist, 1975) the tunnel provided 

 flows over naturally rippled sand with a mean diameter of 0.56 millime- 

 ter. Values of S/S^ remained small, about 0.07, 0.10, and 0.14 (each 

 corresponding to a different ripple size); T ranged from 3 to 14 seconds. 

 Values of U were kept below about 35 centimeters per second to prevent 

 excessive sand motion and erosion of the ripple profiles. During the 50 

 experiments, each lasting about 6 hours, the tunnel performed without 

 serious incident. 



VII. CONCLUSION 



The tunnel described in this study may be compared with four other 

 oscillatory water tunnels that have been built. Two tunnels are of 

 vertical U-tube design and use pneumatic devices to maintain the oscil- 

 latory flow. The first tunnel, described by Carstens and Neilson (1967), 

 operates only at a resonance period of about 3.6 seconds where it main- 

 tains a nearly sinusoidal flow. The second tunnel, described by Lundgren 

 and Sorenson (1957), can operate over a wide range of periods above 

 3 seconds; however, it is unclear how nearly sinusoidal its basic flow 

 remains at periods removed from a resonance period of 9 seconds. 



The other two tunnels are of a horizontal closed-loop design, and 

 operate by the nearly positive displacement of a close-fitting (though 

 not tight) piston in one side of the loop. In one tunnel, described by 

 Dedow (1966), an electronically controlled hydraulic drive can provide 

 almost any specified flow-time relationship. In the other tunnel, 

 described by Brebner and Riedel (1973), a motor drive operating over a 

 wide range of periods provides motion as nearly simple harmonic as is 

 compatible with an eccentric and connecting rod. 



The test cross sections of all four tunnels exceed that of the pre- 

 sent tunnel, and the two horizontal closed-loop tunnels are massive in- 

 stallations. Thus, aside from its unique capabilities for the permeabil- 

 ity study (the paired pistons and reservoirs; Lofquist, 1975), the tunnel 

 described in this study has no single capability not found or exceeded in 

 the other tunnels. However, the present tunnel requires very little power 

 (3/4 horsepower, as compared to the 25 -horsepower motor of the closed- 

 loop tunnel described by Brebner and Riedel, 1973), and yet is able to 

 provide a sinusoidal flow over wide ranges of period and amplitude. Thus, 

 the present tunnel may be regarded as a proven design compatible with 

 limited space and resources. 



25 



