18 BULLETIN 831, U. S. DEPARTMENT OF AGRICULTURE. 
by a well-defined range of velocities at the throat which can not 
exceed that resulting from a complete vacuum. The tendency of de- 
sign has been in most cases to assume that heads in excess of 34 feet 
were not adaptable to siphon installation, because it was accepted 
that the limit of vacuum draft had been reached at that point. As a 
matter of fact, siphons have been built with heads beyond 34 feet 
and in specific cases from 36 to 52 feet, but always, of course, with a 
resulting decrease in the coefficient of discharge due to added fric- 
tional resistance. It is generally accepted from a study of numerous 
designs that two forms of siphon tubes may be considered. The 
first of these will be treated as having throat and outlet of equal 
cross section, but not necessarily of uniform shape. Such a design 
has rarely been used in this country, but is rather common in Europe 
and is more or less followed in the development of the Ocoee Eiver 
installation in Tennessee. 1 
The second type has a contracted throat section, expanding with a 
divergent tube as the outlet and further modified to overcome local 
losses of head, or contributing defects. There are numerous exam- 
ples of the latter type, and their installation has been generally 
adopted in the United States and in Europe, so that most of the ex- 
periments conducted to determine efficiency, or the lack of it, have 
been on this type. Certainly all of the siphons which have been de- 
signed to date have been planned on the general principles of hy- 
draulics, and still there is a dearth of information by which the de- 
signer may be guided in selecting the proper proportions for the 
vital parts, or in departing from present practice with any assurance 
of increased efficiency or the accomplishment of a desired purpose. 
Nothing is definitely known of the relative head losses in the vari- 
ous parts of the siphon; the intake opening, the trance leg, the 
throat, the vertical or sloping outlet leg as the case may be, that due 
to shock or bends, or the curved discharge end. The effect on effi- 
ciency of a contracting or enlarging section beyond the throat is un- 
known, although such siphons have been built in this country. No 
experiments on working models have ever been conducted to deter- 
mine other than the over-all efficiency, as valuable as such informa- 
tion would be when it is considered that the slightest change in the 
form of some part of a hydraulic structure will very often make the 
greatest difference in efficiency. Of course, it is in the long run only 
the efficiency of the structure as a whole which is desired, and the 
problem is solved, other than the value of having such information 
as will assist in arriving at the most economical design or in making 
designs conform to some additional specification as to speed in bring- 
ing the structure into operation and maintaining certain variations 
in allowable water surface. 
1 See description Engineering Record, May 16, 1014. 
