INVESTIGATION OF FLOW OF WATER IN UNIFORM CHANNELS. 357 
more satisfactory determination of the degree of accuracy with 
which such, or similar formule, may be made to represent the 
velocity of flow in channels, when the conditions are varied over 
a wide range. The suggestion of this investigation is due to Mr. 
Knibbs, and was the result of his critical study of the history of 
the subject, and analysis of the deductions of earlier investigators. 
We take this opportunity of acknowledging our indebtedness to 
him not only for the fullest access to all his notes, but also for his 
generous codperation and counsel in planning and carrying out 
the experiments. The experimental work itself was made possible 
by the kind assistance of Prof. Warren, who, when the need of 
the work was discussed, at once undertook the suitable equipment 
of his laboratory for the prosecution of these and similar hydraulic 
experiments. For this, and for his cordial assistance and counsel 
during the whole course of the work we desire to place on record 
our grateful thanks. 
2. Objects of experiments.—The conditions which may be 
assumed as having the most marked influence on the rate of flow 
in open channels are— (a) slope, (6) hydraulic radius, (c) tempera- 
ture, (@) roughness, and (e) form of channel section. The present 
experiments were carried out in one channel so that the last two 
conditions were constant (unless it be supposed that roughness 
must be considered in relation to the absolute dimensions of the 
channel) and the first three only were included in the enquiry. 
The objects in view may be thus stated :— 
i. While keeping the temperature and hydraulic radius as 
invariable as possible, to determine the relation between velocity 
and slope, over.the greatest attainable range of slope. 
ii. At certain fixed slopes and with the temperature as invari- 
able as possible to determine the relation between velocity and 
hydraulic radius. 
iii. At certain fixed slopes and with the hydraulic radius as 
invariable as possible, to determine the relation between velocity 
and temperature. 
