114 Journal of Agricultural Research voi. ix. No. 4 
was not possible to control the flow of water as perfectly in the long chan¬ 
nel as in the laboratory proper, still the regulation of the head was quite* 
good, and in all experiments the discharge was determined volumetrically. 
Of the 22 experiments made for check purposes, 15 were comparable 
directly with original experiments made in 1914. They show that the 
gate caused an increase in the discharge, but, when allowance is made 
for probable experimental error, this increase is a maximum of approxi¬ 
mately 2 per cent. The 1916 experimental results do not show the 
original data to have been in error an impractical amount due to the 
close proximity of the control gate to the orifice. 
