24 BULLETIN 831, U. S. DEPARTMENT OF AGRICULTURE. 
height of the throat, and this coefficient is based on the velocit} T at 
the throat (V). 
The friction loss in the outlet leg depends upon the material 
and class of workmanship therein, and is further dependent upon 
the cross section of the chamber with regard to the throat, which is 
assumed as constant and equal to the area of the throat. 
RESULTS OF TESTS. 
The writer has stated in another part of this paper that the only 
tests made to determine the losses on the different parts of the 
siphon are those of the small laboratory models, so that in summing 
up the results reference will be made to the tests of overall efficiency 
on working models built to discharge large volumes of water where- 
ever such information is available. Some points have been brought 
out incidentally in these larger tests, indicating the value which may 
be placed on the deductions drawn from the laboratory work. 
Taking these points up in the order in which they are listed in a former 
part of this paper, the following is a summary : 
(ay The theoretical and actual loss of head in the various parts of 
the structure as determined from the tests were not consistent for the 
various tests nor for the different models, but were of sufficient ac- 
curacy to warrant the use of the standard formulas until some more 
reliable data can be developed. The standard formula for the loss 
at entrance head 0.50H V for the type of opening for which the 
formula was developed ran both high and low in the tests, and may 
be considered as holding -good as an average, so far as any develop- 
ments in the laboratory results are concerned. Friction loss in the 
structure was indicated as being negligible in the larger sections of 
the tube, and was heaviest at the throat or contracted section. 
It was so small as to be neglected in the results. 
(I) The varied shapes of the discharge lip did not seem to affect 
the total efficiency, and since all of the models were of uniform design 
at the intake end, nothing developed in the tests at that point or in 
the bends from which to draw conclusions. Xo data from models of 
larger siphons are available with which to compare these. 
(c) The total efficiency for the various models for different air-inlet 
conditions ran 0.84, 0.98, and 0.983 for the three sets of tests when 
grouped and averaged. Similar tests on larger models, but without 
the introduction of varying air-inlet conditions, ran from 0.644 to 
0.805, and in a number of other siphons in this country and in Europe 
coefficients of discharge ranging from 0.T0 to 0.82 have been found. 
These points will appear in the descriptions of the individual cases 
hereinafter taken up. 
(d) The commonly accepted theory has been that the flow of water 
over the crest of the siphon would exhaust the air through the dis- 
