The distribution of pressure over a lifting surface in subsonic potential flow is investigated by means of a three-dimensional electrical potential analyser. It is shown that by using the linear perturbation theory it is possible to obtain the pressure distribution for small angles of incidence, as well as the slope of the lift-incidence curve, easily and rapidly. A comparison with results obtained by vortex theory calculation shows a good agreement in spite of the relatively coarse network which was used for the experiments. There are no practical limitations to the use of a network having a finer mesh, and problems concerning curved and twisted surfaces may also be studied with facility by the use of the potential analyser.

An old numerical method of solving fourth order differential equations is put in relaxation form. The higher order correction terms are included and the technique is illustrated by an example. The method has the advantage of being more rapidly convergent than the usual relaxation procedure for fourth order equations. Some comments are made on the numerical solution of the viscous flow equation.

A number of authors (e.g., Refs. 1 and 2) have dealt with the problem of uniform two-dimensional supersonic flow past a thin oscillating aerofoil, whose angle of incidence to the main stream is small. In this special case, the effects of vorticity are negligible, while the fields of flow on either side of the aerofoil may be treated on the same footing.

The object of this paper is to deal with the more general case in which the aerofoil performs small oscillations about a fixed position, in which its incidence to the main stream is appreciable.

The first object of this paper is to show the concordance between three different theoretical approaches to flame stabilisation and to compare published experimental data with the theoretical predictions. Good agreement between the theory and experimental results is obtained. The second object is to show that, at low pressures, the predictions of these theories must be qualified in view of the increasing importance of heat losses from the flame region. It is shown that there must exist a minimum pressure below which a flame cannot be maintained in a given combustible mixture, irrespective of the rate of flow of the mixture, unless heat losses are entirely eliminated.

A simple method is developed for computing the trailing vorticity which arises when a non-uniform stream is turned.

It is shown that, for a sudden and constant deflection of a non-uniform stream, no net trailing vorticity is set up in the exit flow and hence there is no secondary motion.

In the case of an impulse cascade of finite dimensions with constant turning, it is found that the trailing vorticity has three distinct components—the passage vorticity and two components which appear as vortex sheets springing from the trailing edges of the aerofoils. It is shown that for small angles of deflection there is no net circulation associated with the trailing vorticity downstream, of the cascade, and it is inferred that this should still be so for large deflections.