undisturbed, the pivot piece remains in an intermediate position, and the crown wheel is locked by the double engagement of its pinion.
In this centrifugal balance, as in all apparatus of its type, the precise speed at which the whirling table is propelled is unimportant for the reasons already given. It is, however, necessary that the velocity should be steady, i.e., the whirling table should not be undergoing acceleration when the experiment is made. The reason of this precaution is that the pivot piece and bar possess moment of inertia, and change of speed of rotation involves a torque about the axis foreign to the conditions. This effect could be minimised by concentrating the mass as a bob weight, and making the lever and pivot piece carrying it as light as possible. The difficulty could be entirely eliminated by arranging a duplex apparatus, in which two sliding bars are employed, having opposite rotation, their preponderating weights being arranged at opposite ends.
§ 228. Dines' Experiments (continued). Aeroplane Experiments.—In the determination of aeroplane data, other than in the special case of the normal plane, the Dines method presents certain difficulties. It may have been noticed from the mechanical disposition of the parts, that it is the moment of the pressure reaction about the axis that is the quantity measured, consequently before the magnitude of the pressure reaction can be ascertained the position of the centre of pressure must be known. In order to avoid the necessity of independently determining the centre of pressure, observations are made with the adjustable arm in two complementary positions (Fig. 143 (a)), the angle of incidence being the same in both cases. It is evident that the arithmetical mean of the two readings will give the moment of the pressure reaction as if the total force were applied at the geometrical centre.
By investigating two further positions (Fig. 143 (b)), an attempt was made to compute the tangential force or skin-
344
