This law is fundamental and any deviation therefrom may be con- 

 sidered as the result of errors in observation or measurement. In 

 the present study the maximum deviation is of the order of 4 percent; 

 the average error being 0.2 percent. In the evaluation of the data 

 presented hereafter the accuracy of the test as indicated by column 10 

 should be considered. 



The measm-ed values of wave velocity and wave length, for each 

 wave studied, are compared to the corresponding values computed 

 from theory, m table 2. Columns 2 and 3 list the wave velocity and 

 wave length as computed by the Gerstner (trochoidal) theory, where: 



Lt=CtT„, ^ (2) 



Tm is the measured period which is assumed to have been correctly 

 determined. Colunms 4 and 5 list the percentage differences between 

 measured and theoretical values. The averages given should be 

 considered only in conjunction with similar averages; the comparisons 

 for each test, when considered in company with values in column 10, 

 table 1, giving a more accurate picture of the relation. Note, for 

 example. Run 1, the accuracy of measurement is —4.4 percent, the 

 comparison of measured and theoretical velocity is +0.7 percent. 

 Inspection of the sign of the differences shows that if the measured 

 velocity was 4.4 percent too small, then the value in column 4, table 2, 

 when Cm is corrected by increasing it 4.4 percent, is of the order of 

 +5.0 percent. 



It will be noted that the agreement between the average measured 

 values and the average theoretical values derived from Gerstner's 

 formulae is of the order of 0.5 to 1.0 percent. 



The theoretical values for wave velocity and wave length computed 

 by the formulae of the Stokes-Levi-Civita theory where 



^' - 27r V+ L' + 2L' J ^^^ 



are given in columns 6 and 7. 



Again comparing average measured and theoretical values as above, 

 it is found that the agreement is of the order of 2,5 to 3.0 percent. 

 Since this theory requires the existence of mass transport, i. e., a trans- 

 port of water in the direction of wave travel, then in a closed tank (or 

 any finite body of water) there must be a return flow of water counter 

 to the direction of wave travel. It is assumed that this flow is uni- 

 form over the tank transverse cross-section, and reduction of the theo- 

 retical values by the velocity of the return flow has been made. 



