364 S$. H. BARRACLOUGH AND T. P. STRICKLAND. 
earlier experiments the times were measured with a stopwatch, 
but as its rate was found on close examination to be irregular, a 
pendulum clock marking seconds, was substituted and its rate 
determined by daily comparison with the time ball of the Govern- 
ment Observatory. As an alternative method of measuring the 
aaa of ni the ae of discharge at various heads for the 
d, as detailed below, and as the method 
was found to involve at toast i in the case of orifice 2, which was 
the one used in the great majority of the experiments, only 4 
small, and quite negligible error, it was adopted in many of the 
later experiments and provided a useful check on the earlier ones. 
In the following table h is the head on the orifice in centimetres, 
and ¢ is a coefficient representing the value, at the respective 
heads, of the expression 
cub. em. per second. 
vh 
Table IV. 
Onfice 1, | Orifice 2. Orifice 3. 
5714 58 
7°06 84°44. 
55°86* “99 
57°39 84-01 
79°83* | 83°90 
83°38* | 83°90 
93°08 83°67 
* These were observations taken merely for determining the 
coéfficient, + This was an experiment in SeriesI. The remain- 
der form part of Series II. 
In one or two of the experiments at the greatest hydraulic 
radii and slopes, it was found necessary to open two of the orifices 
together, in order to supply sufficient water, and it was assumed 
in reducing the retults for such cases that the orifices did not 
appreciably interfere the one with the other. 
8. Measurement of slope.—At each station down the channel 
from B to G, bench marks were established, 749:5 millimetres 
apart, immediately over the centre of the channel, and levels were 
