SPILLWAYS FOE KESEEVOIES AND CANALS. 25 
charge leg and under the discharge lip, provided the outlet leg was 
sealed at that point and that the air becoming rarefied would encour- 
age an increase in this flowing head until full siphonic action was 
accomplished. The tests on the laboratory model indicated a tend- 
ency to. develop a back pressure in the crown by this rise of the water 
surface at the inlet, and the suppressed effort of the air to escape 
under the discharge lip when the seal was complete. 
In fact, it was shown that with the discharge lip submerged 0.1 to 
0.2 foot, and fully out of water, resulting in the tube being unsealed, 
the siphon primed easier and quicker than when the outlet lip was 
submerged to increasing depths to effect more complete seal. The 
tests on the Yuma siphons indicated that this is probably true, but 
the head on the discharge lip during the Yuma tests was not reduced 
sufficiently to disclose its action with sealing heads of 1 foot or under, 
or when flowing freely into air. The Ocoee River installation of the 
Tennessee Power Co., however, does bear out the tests of the small 
models, although it has been argued that the shape of the outlets of 
these siphons in itself forms a seal. Conclusive proof that a sealing 
basin is not necessary on siphon spillways or that the efficiency of the 
structure is not decreased by the absence of it, would lead to a great 
reduction in their cost and would permit of the elimination of another 
weak point in freezing weather. 
(e) The placing of air inlets and the submergence of the outlet 
end are so closely related that the points are difficult of separation 
in summing up the results. One question to be solved is to determine 
the proper depth of submergence of the discharge lip to produce maxi- 
mum siphonic action, and another is to determine the shortest period 
of time necessary to produce complete action under the same condi- 
tions. To enable the plotting of a curve to show these relations, various 
conditions were produced by the manipulation of improvised air in- 
lets, made possible by utilizing the tubing for connecting the gage 
glasses. With all air inlets or outlets closed over the throat section, it is 
intended that the siphon will come into full action just as the water 
flows over the crest. The condition as developed with a complete 
seal of the siphon tube and with the water rising in the inlet and 
outlet legs produced an air compression in the tube which tended to 
retard the increase of head on the overflow crest, and consequently 
required a longer period of priming because of the resulting decreased 
flow of water necessary to produce expulsion of the confined air in the 
tube. The shortest period of priming was obtained by using the 
gage-tube connection at the throat as an automatic air valve, allow- 
ing all compression in the siphon to be relieved when the water rose 
to produce it, and then to be instantly closed when a reverse pressure 
was started. 
(/) This suggested that the delays in priming were shortest when 
a relief valve was provided ; that the heads on the crest of the throat 
