The most notable feature of the magnetohydrodynamic flow at large distances from a three-dimensional body is the formation of two wakes, within which vorticity and electric current are confined. In this paper results are obtained for the effective diffusivity and the relation between current and vorticity in each wake, for the balance between the strengths of the disturbances in the wakes and in the irrotational current-free flow outside, and for the lift and drag forces acting on the body. The final answers take the form of remarkably simple extensions of the corresponding formulae for non-conducting flow. In spite of the extra wake and the presence of a magnetic as well as a velocity field, the flow perturbation at large distances still has only three degrees of freedom.

This investigation is concerned with the phenomenon of a gas jet impinging on and penetrating into a liquid. The study has been restricted to the cases of circular and plane jets penetrating a liquid at right angles. Both ‘free streamline’ and turbulent jets were considered. The phenomenon was analysed from two viewpoints. The first, a stagnation-pressure analysis, related the depth of the surface depression or cavity to the stagnation pressure based on the centre-line velocity of the jet in the neighbourhood of the surface. The second, a displaced-liquid analysis, related the weight of the liquid displaced from the cavity to the momentum of the jet.

Numerous experiments were conducted in which the cavity depth, diameter or width, and peripheral lip height were measured. The role of surface tension in affecting the cavity depth was considered and the phenomenon of drop formation was examined. Some attention was given to the case of a plane jet impinging on a moving liquid. It was found that the experimental data fit into the framework of these analyses quite consistently.

Small spheres of the same size but of relative density varying from 0·92 to 1·25 were injected in turn into a horizontal water pipe, in which the flow was turbulent and the mean velocity was constant. A cross-section near the outlet was illuminated; the positions of the spheres as they crossed it were measured by photography, and the relation was established between the terminal velocity of the of the spheres in water and the vertical diffusivity. The velocity of the spheres along the pipe was found to be somewhat different in the galvanized steel and Perspex lengths of which the pipe was composed. The dispersion of the times of transit of the spheres increased slightly with their densisty. For purposes of comparison the theoretical velocity along the pipe was also calculated from the photographic measurements.

The concept suggested by Batchelor that motion of a marked particle in turbulent shear flow may be similar at stations downstream from the point of release is applied to a variety of diffusion data obtained in the laboratory and in the surface layer of the atmosphere. Two types of shear flow parallel to a plane solid boundary are considered. In the first case mean velocity is a linear function of log z (neutral boundary layer) and in the second case the mean velocity is slightly perturbed from the logarithmic relationship by temperature variation in the z-direction (diabatic boundary layer). Besides the parameters introduced in previous applications of the Lagrangian similarity hypothesis to turbulent diffusion, the ratio of source height to roughness length h/z0 is shown to be of major importance. Predictions of the variation of maximum ground-level concentration for continuous point and line sources and the variation of plume width for a continuous point source with distance downstream from the source agree with the assorted data remarkably well for a range of length scales extending over three orders-of-magnitude. It is concluded that results from application of the Lagrangian similarity hypothesis are significant for the laboratory modelling of diffusion in the atmospheric surface layer.

The impingement of a rocket jet, exhausting into a vacuum, onto a solid surface is examined using the Newtonian approximation to analyse the transonic region. The jet exhaust into an unbounded vacuum required for this analysis is simplified as a quasi-radial flow using an exact calculation as a basis. Two examples are calculated giving the shock detachment distance and the ground pressures as a function of the rocket altitude.

Resonances are discussed between electrons and protons and waves of whistler and hydromagnetic types. The conditions of resonance are investigated. Approximations for the effects on the particles are obtained. The occurrence of a ‘trapping’ phenomenon is demonstrated and the effect on velocity distributions is discussed.

It is shown that the turbulent wake of a self-propelled body moving in a fluid with a vertical density gradient is considerably different than of the same body moving in a fluid having no density gradient. In the uniform density case, the turbulent mixed fluid behind the body expands into an irregular conical shape. In the case of a density gradient, the initial expansion of the mixed fluid is quickly followed by a collapse in the vertical direction which is accompanied by a further spreading in the horizontal direction. This phenomenon is caused by the force of gravity. The volume of fluid behind the self-propelled body has a more or less constant density due to mixing. Thus, it is forced to seek its own density level in the undisturbed fluid.

The collapsing vertical wake is shown to be an efficent generator of internal waves, many of high order. These manifest themselves in surface movements.

The assumption that the internal waves are damped only by viscosity, not by turbulence, leads to results in general accord with the observations.

It is shown that the optimum design of duct in a magnetohydrodynamic generator is close to the one in which the flow Mach number remains constant. This constant-Mach-number generator is analysed in some detail and it is shown that the optimum Mach number can be defined to within a few percent. For a γ of 1·25, this optimum is near 0·85. For very short ducts, the maximum power output is obtained near matched-load conditions but for rather longer ones maximum total power output is obtained by working as close to short-circuit conditions as is practicable. Against this, the minimum compressor requirements are found by working as close to open-circuit conditions as is practicable, and so a compromise must be reached for optimum overall generator design as far as load conditions are concerned. This will probably give an internal ohmic loss in the fluid of about one-third of the total output. Curves are presented which enable the optimum Mach number to be determined with greater precision when the optimum load conditions have been selected.

A normalized formulation of the problem of isothermal bubble growth, dominated by viscosity and diffusion, reveals that in its full generality the problem involves eight dimensionless parameters. Limiting cases for extreme values of each of these parameters are investigated.

This paper describes a series of experiments carried out in a high-enthalpy stream of argon on materials that are known to sublime. The results confirm that an axisymmetric Teflon model ablates to a stable shape which is independent of the initial nose profile. The effect of changing the total enthalpy of the gas is simply to alter the recession rate of the nose. The experimental results show poor agreement with a simple theory which ignores the effects of mass transfer in the boundary layer.

The problem of the linear stability of a time-dependent shear flow is investigated and the effects of contraction and expansion of the layer are discussed. Models of flows observed prior to breakdown are constructed and used to investigate the resulting secondary instabilities which are shown to be extremely violent and in essential agreement with recent experiments. In relatively short time, one half period of the primary oscillation, the energy in the secondary disturbance increases by two orders of magnitude; the wave-number corresponding to maximum amplification is five times that of the primary wave.