The flow of a jet directed normal to a uniform, steady cross-wind is considered. Experimental results show that for various jet strengths, the position of the jet in space, when stretched by the ratio of jet to cross-wind momenta, is described by a single function. Exceptions exist at very low velocity ratios where a shift of the potential core is evident. A natural system of axes is used to define important directions of the flow. The integrated equation of motion along the primary jet flow direction is made dimensionless after the general method of Morton (1961) and a virtual source is defined for the flow. It is shown that a single functional behaviour of the axial jet velocity exists for various velocity ratios if the jet is considered to originate from this source. Lateral velocity profiles show a similarity when scaled by appropriate lengths and velocities but true self-preservation is not attained.

The feasibility of an experiment to measure the recombination rate of oxygen is examined using the ideal-dissociating-gas model. The experiment is to be performed in a shock tube, the shock-heated (and dissociated) gas being cooled by passing it through a Prandtl-Meyer expansion, and then allowed to recombine in a constant-area channel. At appropriate densities and shock Mach numbers it is found that recombination takes place in a distance suitable for a laboratory experiment.

Using this technique, the recombination rate of oxygen has been measured at 2700 ˚K. To determine the recombination rate, the absorption of ultraviolet light at a wavelength of 2283 å measured 11 cm downstream of the expansion was compared with absorptions calculated for various values of the recombination rate constant.

The measured value of the recombination rate constant of oxygen is in agreement with values calculated from dissociation rate measurements.

The Wiener-Hopf technique is applied to solve the linearized problem of a two-dimensional compound gas jet, i.e. a jet embedded in a gaseous stream of finite width. The solution is found for all combinations of supersonic and subsonic flows in jet and stream. The general nature of the solution when only one of the flows is supersonic varies according as the value of a certain quantity mk, depending upon the gas constants, Mach numbers and widths of streams, is greater than or less than unity. When mk= 1 the solution appears to be invalid and it is suggested that, in this critical case, a steady flow (regarded as the limit in time of an unsteady flow) may not exist. It is further shown that the solution propounded by Pai (1952) for a supersonic jet embedded in a subsonic stream is simply the asymptotic form of the general solution. The findings of Pack (1956) for a supersonic jet in a supersonic stream are confirmed and extended.

When the dimensions of a convective system in a saturated porous medium are sufficiently great, diffusion effects can be neglected except in regions where the gradients of fluid properties are very large. A boundary-layer theory is developed for vertical plane flows in such regions. In special cases, the theory is equivalent to that for laminar incompressible flow in a two-dimensional half-jet, or in a plane jet or round jet, for which similarity solutions are well known.

A number of experiments have been performed using a Hele-Shaw cell immersed in water, with a source of potassium permanganate solution located between the plates. At very small values of the source strength, a flow analogous to that of a plane jet from a slit is obtained. The distance advanced by the jet front, or cap, is measured as a function of time, and the velocity is found to be nearly proportional to the velocity of the fluid on the axis of the steady jet behind the cap, as given by the similarity law of Schlichting and Bickley. At large values of the source strength, a two-dimensional ‘broad jet’ of homogeneous solution descending under gravity is produced; the shape of the flow region can be calculated with little error from potential theory, neglecting the effect of the mixing layers.

A possible example of a mixing layer observed in a geothermal region is examined. The theoretical form of the temperature distribution is calculated numerically, taking into account the large viscosity variation with temperature and also the possibility of a large permeability variation. These effects are found to have less influence upon the solution than might have been expected. Quantitative values obtained for the physical parameters are consistent with other geophysical observations.

An analysis of experimental results shows that the three most important terms in the energy balance of a mixing layer—the production, diffusion and dissipation terms—may be expressed in terms of an eddy viscosity, eddy diffusivity for turbulent energy, and a microscale, respectively, all being constant for a given cross-section. A first-order solution calculated for the resulting differential equation yields a turbulent intensity profile in qualitative agreement with experiment. The peak value of the turbulent intensity is found to depend on a dissipation constant and on the ratio of eddy diffusivity for turbulent energy to eddy viscosity. Experiments carried out with the intention of reducing turbulent intensities by means of extraneous vortices, generated by small delta wings, have shown slight increases in intensity, presumably because eddy viscosity increases with eddy diffusivity.

This study concerns the stability of the steady laminar boundary-layer flow of a homogeneous fluid which occurs in a rotating system when the relative flow is slow compared to the basic speed of rotation. Such a flow is called an Ekman boundary-layer flow after V. W. Ekman who considered the theory of such flows with application to the wind-induced drift of the surface waters of the ocean.

Ekman flow was produced in a large cylindrical rotating tank by withdrawing water from the centre and introducing it at the rim. This created a steady-state symmetrical vortex in which the flow from the rim to the centre took place entirely in the shallow viscous boundary layer at the bottom. This boundary-layer flow became unstable above the critical Reynolds number $Re_c = vD|v = 125 \pm 5$ where v is the tangential speed of flow, $D = (v| \Omega)^{\frac {1}{2}}$ is the characteristic depth of the boundary layer, v is the kinematic viscosity, and Ω is the basic speed of rotation. The initial instability was similar to that which occurs in the boundary layer on a rotating disk, having a banded form with a characteristic angle to the basic flow and with the band spacing proportional to the depth of the boundary layer.

The principal object of study is plane flow over bodies with a sharp apex at Mach numbers greater than unity. The magnetic field is assumed to be uniform, rectilinear, and parallel to the undisturbed stream. Flow behaviour near the apex of a wedge is investigated by the method of characteristics. It is found that for small wedge angles an attached shock attenuates initially with distance from the apex, but for larger wedge angles the shock grows stronger.

The structure of a slow magneto-gasdynamic shock is investigated for the case of strong magnetic field and small electrical conductivity. The streamlines are displaced within the shock although the initial and final flow directions are the same. An ordinary gasdynamic shock may occur on the upstream side of the transition. The shock structure theory gives a solution for the flow near the apex of a certain class of bodies.

For the study of slow shock structure, it is shown that the transition is described by a curve in the (F, H)-plane. F is the sum of pressure and momentum flux in the direction of variation; H is the sum of enthalpy and kinetic energy due to the velocity component in the direction of variation. General properties of the (F, H)-plane are found for a gas whose equation of state obeys the conditions of Weyl (1949). Flow behaviour on the transition curve is then determined. The theory of the (F, H)-plane can be used in the study of other one-dimensional processes in magneto-gasdynamics.

The wall jet with suction or injection is investigated; an analysis under conditions corresponding to similar flows is shown to reduce to an eigenvalue problem. Asymptotic solutions valid far from the surface are used to initiate the integration and circumvent the usual iteration associated with the two-point boundary-value problem. Typical solutions for various rates of suction and injection are obtained. It is found that the skin friction decreases with increasing rate of suction. Representative thermal solutions are obtained for Prandtl and Lewis numbers equal to one, under the special condition that the surface temperature is equal to the ambient temperature or that the enthalpy varies monotonically from the surface value to the ambient value.

In this paper, we consider the elementary processes involved in the appearance of a simple twin wake in the two-dimensional steady aligned flow of an electrically conducting viscous incompressible fluid in a uniformaly applied magnetic field. Fundamental relations between the twin wake and its elementary process are generally discussed, and the corresponding initial-value problem is formulated. Then, the solution of the initial-value problem corresponding to the above simple twin wake is obtained, and two limiting cases with the magnetic Prandtl number near one and zero (or infinite) are discussed in detail. In both cases, it is found that negative vorticity appears in between the main positive vorticity associated with the two Alfven spots. This fact gives some clue to the understanding of the property of twin wakes. As an application, the crossed-flow case is also discussed, and a conjecture upon the vorticity distribution is made.

The problem of the stability of plane Couette flow to infinitesimal disturbances is carried numerically to larger Reynolds numbers than heretofore. The flow is definitely stable up to R = 1430, although at Reynolds numbers in this vicinity plane Couette flow has been observed experimentally to be turbulent. Observations on boundary-layer instability suggest this to be the source of turbulence in plane Couette flow channels.