can determine the wavelength very accurately. Since it is easy to measure 

 frequencies with far higher accuracies than required by this experiment, 

 this method is well suited to high quality laboratory apparatus, and it 

 could, for example, form the heart of a superior calibration facility for 

 velocimeters under the assumption, of course, of no dispersion. Some ex- 

 periments to compare these different methods have been done: a progressive 

 wave at 2.6 MHZ was sampled by leadscrew and phase sensitive detector and 

 agreement with time-distance methods, but slight disagreement with inter- 

 ferometry, was found (Reference 36). Also a "freefield" propagation exper- 

 iment in a large tank a 1 MHZ yielded velocities about 0.6 m/s lower than 

 "sing around" velocimeter data (Reference 37). These discrepancies, however, 

 are not necessarily due to some real effect, such as dispersion, but prob- 

 ably rather illustrate the difficulties of setting up a clean experiment: 

 Reflections must be avoided in propagation methods and corrections for walls, 

 posts, etc., must be included in standing wave methods. 



39 



Arthur l.lLtttlc.Knc. 



