FLOW OF WATER IN DREDGED DRAINAGE DITCHES. 3 
the velocity was determined at the surface, mid-depth, and the bottom 
of the stream. In a few instances the velocity was measured at 0.2 
and 0.8 depths. The best methods were observed in the use and care 
of the current meter, and it is believed that the results are entirely 
dependable. 
Carefully made soundings were obtained during low water stages 
at the velocity-measuring points and wherever a decided change 
in the perimeter of the channel took place. As a check on these 
measurements, and to detect any changes due to silting or erosion, 
soundings were also made at the time of the velocity measurements. 
The depths were measured to the nearest tenth of a foot. 
SLOPE OF WATER SURFACE. 
The slope of the water surface was measured along the straightest 
and most uniform course of channel available near the gauging station, 
except in two instances where it was desired to ascertain the values 
of n for curved and crooked channels. The courses chosen were such 
that no tributary streams entered the channel and no decidedly abrupt 
changes in the bottom grade of the channel occurred along the courses, 
and were so located as to exclude such obstructions as logs, fallen 
trees, piling, and bridge piers. In one experiment, however, courses 
were chosen with a view to determining the influence of such obstruc- 
tions upon the flow in the channels. Where the course of the channel 
was not limited in length by the conditions named above, the length 
of the course was made at least 10 times the top width of the channel. 
It was apparent that representative values of n could not be obtained 
for a channel if a long course was chosen without regard to the con- 
ditions mentioned above. For instance, during a high stage in the 
Bogue Hasty channel, in Bolivar County, Miss., drift accumulated at 
a pile trestle bridge just above the slope course, causing a fall of 0.3 
foot from the upstream to the downstream side of the bridge. The 
actual slope of the water surface for this channel was less than 1 foot 
per mile, from which it is apparent that had a long slope course been 
chosen which included this bridge, decidedly different values of n 
would have been obtained. Also where long slope courses are used, 
incorrect values of n are often obtained, due to the entrance of surface 
water along the course. 
In these experiments the slope of the water surface was determined 
by making vertical measurements from permanently set reference 
points to the surface of the water. These reference points were 
established on horizontal arms extending from vertical posts set 
firmly in the ground at each end of the course. Variations in water 
level of from 3 to 4 feet could be measured from each post, and a 
sufficient number of posts were set on the side slope of the channel 
so that all fluctuations in water surface could be measured. The 
elevations of the points on the arms of the posts were determined 
