INVESTIGATION OF FLOW OF WATER IN UNIFORM CHANNELS. 365 
taken to these marks with an 8” theodolite by Troughton and 
Simms, which was placed very nearly in line with the channel so 
that the telescope had to be turned only through a small arc. A 
level staff was constructed of a boxwood millimetre scale, a metal 
pin being attached to its end, and a plumb-bob and thread at the 
back to ensure its being held vertically. It was found possible to 
read the staff to ‘01 cm., so that as will be shown more fully in a 
subsequent paragraph, the error in the determination of the slope 
is a satisfactorily small one. At least two sets of readings of the 
levels were taken for each experiment, usually by two different 
observers, and the results in nearly every case proved to be practi- 
cally indentical. If any discrepancy occurred between the two 
sets of readings, they were always repeated. The vertical distance 
from the bench mark to the zero line of the scale from which the 
distances to the water surface were measured, is as follows :— 
Section A Ps eee, * eee 2 E P decree o fern « & 
Distance (mm) 56 55 67 58 615 56 63 63 
9. Method of making an experiment.—The routine of an experi- 
ment may be described in general as follows. The channel. 
was adjusted to about the required slope, was approximately 
levelled in a transverse direction, and any slight lack of uni- 
formity in the gradient from section to section, as revealed by 
the hand-level before referred to, was as far as possible removed 
by minor adjustments of the supports. The appropriate orifice 
was then opened and the water allowed to rise in the supply-tank 
till the necessary depth was attained in the channel, when the 
head was adjusted by means of the overflow pipe and by a corres- 
ponding regulation of the inlet valve. A set of levels was then 
taken, during the course of which the head on the orifice was 
frequently noted in order to detect any tendency to fluctuation. 
The actual experiment was now begun by reading the scale of the 
gauging tank and then turning the water so as to flow into it, 
carefully noting the time at the same moment. The distances of 
the water surface in the channel from the zero line of the scale 
