Nov. 8, 1915 Use of Current Meters in Irrigation Canals 229 



accurate results as the more detailed observations. The use of the 

 0.6-point method gave results averaging 4.8 per cent too high, and the 

 use of from 3 to 5 verticals in flumes and lined sections gives an average 

 of 3 per cent too small. Apparently where few verticals are to be 

 observed, the use of the 0.6-point method may be preferable, as the 

 errors will tend to balance. This may be expressed by saying that about 

 the same relative detail should be used in measuring the velocities in the 

 verticals that is used in the number of verticals observed. 



The results are obtained by using the verticals taken in the detailed 

 measurements and selecting every alternate or every fourth vertical and 

 computing the discharge that would have been obtained had only these 

 verticals been observ^ed. It is possible that gagings where the lower 

 numbers of verticals were to be observed could be made to give closer 

 results by using some means for the selection of the location in the canal 

 section at which the verticals should be taken. It has been previously 

 shown that the use of velocity measurements at the 0.2- and 0.8-depth 

 points will give very nearly the same results as measurements at 6 or 

 more points in the vertical and that a single observ^ation at 0.6 gives 

 results within 5 per cent of being correct. 



If one or two points can be found in the vertical velocity curves the 

 velocities of which can be used to determine the average velocity of the 

 whole vertical, it would seem probable that perhaps 2 verticals on the 

 horizontal velocity curve could be found which could be used to give 

 the average velocity in the whole cross section. Such points, or index 

 verticals, as they may be called, would be useful in the rougher meas- 

 urements often needed in canal operation, and information as to the 

 relative accuracy of such methods should be of value. 



Two such selected verticals may be used to determine the discharge 

 in two ways. In one the velocities only might be used and the cross- 

 section area more carefully determined, if not known from previous 

 observation. In the other the observed verticals may be used to obtain 

 not only the mean velocity but also the depths at these verticals, and 

 the width of the section may be used to determine the cross-sectional 

 area. 



The use of such index-vertical methods is, of course, most applicable 

 to canal sections such as flumes which have practically uniform depths, 

 as the error in determining the cross section is largely eliminated. 



The measurements were examined to see whether such index verticals 

 could be found. The horizontal velocities and cross sections were 

 plotted on a sufficiently large scale so that the velocity and depth at any 

 point could be read from the curves. Such index verticals would be 

 most easily used if their distance from the sides is some definite propor- 

 tion of the water-surface width. Verticals located at different points 

 were tried. The different types of canal cross sections are discussed 

 separately. The general results are given in Table VI. 



