EXPERIMENTS ON THE FROUDE. 97 
is 2y'5 inches in diameter and the hole in the shaft-log is about 3 inches in 
diameter; as the log is filled with oil there may be considerable fluid friction. 
Considering all the elements that enter into our investigations including 
the effect of the wind we feel justified in claiming a precision of 1 per cent. 
for results of tests at full power and speed (6 to 7 knots); at slower speeds 
and smaller powers the errors are liable to be larger, perhaps as much as 
2 per cent. at 4 knots. 
A corresponding degree of precision is claimed for such derived quan- 
tities as real-slip, wake, thrust-deduction and hull-efficiency; but it must be 
remembered that these are conventional functions which are derived in an 
arbitrary manner and that the interpretation of the meaning of these 
quantities is far from certain. 
An examination of the curves of Figs. 6 to 10, Plates 40-44, will show 
that the operation of the apparatus and the reduction of the observations 
are regular and certain. All the curves are fair and they go through all 
the observed points. Experiments could be repeated at will definitely and 
satisfactorily. It will be noted that the curves of thrust show clearly a 
bump at about 6.5 knots, which corresponds with the bump in the curve of 
Fig. 3, Plate 37; it is seldom that any determination of power and speed is 
made with sufficient precision to locate such a bump, unless it be for 
torpedo boats. 
There are certain anomalies especially at low powers and speeds, for 
which no explanation is offered. They are too well marked to be attributed 
to errors or uncertainties of observation or reduction, even should our 
estimate of errors be doubled. Among these are the curious variations of 
wake and especially the indication of negative thrust deductions. To my 
mind it appears possible that the degree of precision of our work may 
be sufficient to reveal real anomalies in the accepted conventions con- 
cerning the interaction of hull and propeller. 
To myself the most notable feature is the small wake computed for 
the normal setting of the propeller 1 inch from the stern-post, more especially 
as the wake for the 3-inch setting, while the propeller is still affected by the 
eddy behind the stern-post, is what we might reasonably expect. The gain 
in propulsive efficiency for the latter setting as indicated by a comparison 
of the shaft horse-powers is only 3 per cent. As I never saw the Manning 
in dock and as the drawings are somewhat ambiguous it is not certain where 
the propeller was actually set, but any possible variation is covered by our 
change from 1 inch to 3 inches. 
Our measurement of thrust was the most positive and certain of our 
measurements connected with power; consequently if the resistance of the 
