FLOW OF WATER IN DREDGED DRAINAGE DITCHES. 99 
SOUTH FORKED DEER RIVER NEAR JACKSON. 
Discharge measurements of this channel were made from a single- 
span skew highway bridge on the Bolivar levee road about one-half 
mile from Jackson. The upper end of the slope course was located 
about 75 feet below the bridge and extended downstream 952 feet 
in a straight course. The cross-sectional area was fairly uniform for 
high stages, but for low stages several abrupt changes in size occurred 
(figs. 5B,and6B). The effect of roughness and irregularities in the 
lower portion of the channel upon the value of n is revealed in the 
results obtained (Table 3). Although the channel was free from 
vegetation and obstructions, yet the values of n obtained are con- 
siderably higher than those obtained for the channel at Roberts. 
The irregularities in the channel were left at the time of construction, 
the bottoms and sides having never been smoothed up properly 
(see Plate VIII, fig. 1). 
SOUTH FORKED DEER RIVER NEAR HENDERSON. 
The gauging station was located on the single-span highway bridge 
about 14 miles east of Henderson, Tenn. A straight course 624 feet 
in length, below the gauging station, was selected for slope measure- 
ments. Experiments were made on this channel during both 1916 
and 1917, and there was very little difference in the channel for the 
two sets of experiments or in the values of n obtained for the corre- 
sponding stages (Table 3). It is believed that the results obtained 
are quite accurate. The channel as a whole was in very good con- 
dition. The side slopes were slightly irregular, the bottom was 
fairly even except for a few depressions, and there was practically 
no vegetation in the channel (see Plate VIII, fig. 2; and figs. 5 ©, 
6 C, and 6 D). 
NORTH FORKED DEER RIVER NEAR TRENTON. 
A single-span skew bridge on the Huntingdon levee road about 
one-half mile from Trenton was used as a gauging station. The slope 
course was straight and extended upstream from near the bridge for 
a distance of 699 feet. The variations in the cross-sectional area 
were quite large, but there were no abrupt changes in section for the 
higher stages. Owing to the fact that the size of the channel 
increased gradually from the upper to the lower end of the course, 
it is not believed that the differences in the area of the cross section 
had any appreciable effect upon the-value of n for the course (see 
figs, 5 D,6 EH, and 6 F). There were practically no vegetation | or 
other Ee ictions in the channel-(Plate IX, fig. 1). 
Referring to measurements 28 and 37 in Table 3, 1t is seen that the 
values of n obtained for the highest stages tates 1916 and 1917 
were 0.0271 and 0.0267, respectively. The increase in size of the 
