INVESTIGATION OF FLOW OF WATER IN UNIFORM CHANNELS. 371 
As previously stated the stream did not attain a steady con- 
dition of flow immediately on entering the channel but was 
subject to a marked acceleration for some distance from the 
channel inlet. In the case of Series II., where the cross section 
of the stream was small, this distance proved to be comparatively 
short, the flow being practically steady at Section B, and, except 
at extremely low slopes, it appeared to be but very slightly 
affected by the slope. The distance in which the acceleration is 
complete depends however on the size of the section, and in 
Series I., where the mean hydraulic radius varied from 1°5 cm. 
to 1-7 cm., the velocity was found to slightly increase throughout 
the length of the channel. On the assumption, probably justi- 
fiable, that each experiment would be subject to the same 
percentage error, the line obtained in the logarithmic plot will 
remain parallel to the true line, and hence the index of the slope 
in the foregoing expression so obtained, will be almost, if not 
quite, correct. The necessary quantities for Series I. are given 
in Table VI., and the two curves marked “I” are shown in the 
lower and upper portions of Fig. 3, as already described for 
Series IJ. In this case, as in Series II., a straight line curve 
Table VI. 
Serres I.—Mean hydraulic radius =1-5cm. Mean temperature 
=13:9C 
— in | an Veloety, hoe. = icp Log of Slope. Channel 6 Rections 
2 | 518 | 17143 | 00595 | 3°7745 | E, F 
3 | 7990 | 1:9025 | 0138 | 21899 | D, EF 
5 | 9665 | 19852 | 0280 | 23617 | #, F 
1 | 1102 |.20422 | 0270 | 24314 |» » 
6 | 1169 |.20679 | 0358 | 25539 |» » 
__7a | 1501 | 21764 | 0504 | 27024 |. » 
best represents the logarithmic plot, the inclination of the line 
being approximately -47. The expression connecting slope and 
velocity, all other conditions being constant, may therefore be 
written 
U=k' I* 
