46 AERONAUTICS IN RELATION TO NAVAL ARCHITECTURE. 
They also should not introduce disturbing influences on the center of pressure 
of the aeroplane, as this is also involved in the equilibrium of the aeroplane in flight. 
All of the preceding qualities must be attained to as high a degree as practi- 
cable in floats, which at the same time must be rugged enough to stand severe pun- 
ishment on the surface and still be of the lightest possible construction. The solu- 
tion is necessarily a compromise. Buoyancy and stability require forms and dimen- 
sions which conflict with the best aerodynamic forms and the requirements of mod- 
erate resistance and weight. Planing in flight also requires a form of bottom 
which does not readily meet stream line requirements. 
At the model basin at the Washington Navy Yard a speed of over 15 knots is 
available; this admits of the use of one-ninth size models for a get-away speed of 
45 miles per hour, and this size model is generally used. 
The first condition to be provided is that of loading the model to the displace- 
ment corresponding to the speed. The assumption is made that the lift of the wings 
increases with the square of the speed. This is only approximately correct, but 
serves well for the purposes of comparison. It is not exact, for the attitude of the 
aeroplane is dependent on the amount of air control existing, and whether it is suffi- 
cient to overcome or modify the effect of the forces acting on the floats and thus ob- 
tain control of the angle at which the air surfaces act. As these effects cannot be 
anticipated, or the control be determined or applied, this assumption does not ap~ 
pear unreasonable, and the comparison of model and full-size performance appears 
to justify this and the other assumptions made. 
The model is usually set at a trim of 3 degrees by the stern from the normal 
water line, as numerous experiments indicated this to be the average condition. A di- 
rect comparison was made of a one-quarter size float under the artificial conditions 
assumed and the same float fitted to a model aeroplane to the same scale. Except at 
low speeds where the tip floats, which are not used in model tests, introduced in- 
creased resistance, and at high speeds where the complete model was not under fly- 
ing control, the agreement was very satisfactory, and at the point of maximum re- 
sistance, which is of the greatest interest, was practically in exact agreement. 
The models are first towed from a parallel motion system, which allows the 
model to rise or fall under the influence of the planing or suction effect while main- 
taining a constant trim relative to the surface. The rise or fall from the initial con- 
dition is recorded and indicates the planing power of the model. The gear is so 
arranged that the model is readily counterweighted to the desired degree for each 
condition. 
Plate 32 is a complete record of test of model No. 1844. This test is that of a 
model of a twin-float design of a 2,000 pound aeroplane. The model was one- 
ninth size and the corresponding speeds, therefore, were one-third those for the full 
size. Get-away was assumed at 45 miles per hour. 
Curve B shows the gross resistance of the model and towing gear, and curve C 
the net resistance of the model. 
Curve D shows the change in trim at each condition due to the planing power. 
Curve F shows the modification of the resistance curve when the model is 
