Nov. 8, 1915 Use of Current Meiers in Irrigation Canals 225 



From these experiments it appears that the coefficient is constant for 

 different vakies of n for cross-section areas over about 35 square feet. 

 The rate of variation of the coefficient is greatest for the smaller chan- 

 nels. The observations for cross-sectional areas over 100 square feet 

 were too few in number to give dependable averages for canals of larger 

 size, but both these results and Bazin's formula indicate that the coeffi- 

 cient is practically constant for such larger cross sections. 



Similar cun^es were also obtained based on the value of the coefficient 

 and the discharges. These were similar in form and indicate that the 

 velocity within the limits of the experiments did not materially affect 

 the ratio of maximum surface to mean velocity. These values are not 

 given, as the coefficients based on canal areas are more convenient to use. 



The results were further classified by the shape of the channel. Ap- 

 parently the coefficient does not vary with the form of cross section, as 

 the coefficient from the curves agrees fairly well with the observations 

 when the proper values of n are used, whether the canal is rectangular 

 or irregular or whether the section is deep or shallow relative to its width. 



The average variation of the observed coefficients from the curves was 

 0.045. The average of all observations agreed with the curves, the plus 

 variations equaling those of minus sign. Expressed as a percentage, the 

 average variation was 6. For any single observation the observed value 

 of the velocity coefficient is as likely to differ from the mean curve by 

 less than 0.045 3-s it is to differ by more than this amount. For the 

 average values of the coefficient this amounts to a variation of 6 per 

 cent. In 17 of the 92 experiments, or 18.5 per cent of the total number, 

 the observed value differed by over 10 per cent from the curves. 



The more usual practice where such methods of measurements of 

 velocities by floats are made is to use some general value of the coefficient, 

 usually 0.80 or 0.85. These experiments, as well as the observations 

 given by Kutter, clearly indicate that the coefficient varies quite ma- 

 terially for different-sized canals and for different values of n. These 

 results give values for the coefficients which are less than 0.80 for all 

 values of 71 over 0.016, becoming as low as 0.60 for small canals having 

 high values of w. 



The value of n for any given canal is, of course, uncertain to some 

 extent. The coefficient varies most rapidly with the lower values of n. 

 An error of 0.002 in selecting the value of n makes a difference of 5 per 

 cent in the value of the correct coefficient to be used for low values of n, 

 and less than 2 per cent for the higher values. 



The coefficients to be used should be selected from the cross-section 

 area and the value of n. The character of the canals corresponding to 

 the different values of n given in Table IV can be secured from the 

 general list following : 



