The wall of a large tank or reservoir breaks, sending fluid against a secondary containment dyke. The impact of the surging fluid against the safety barrier is studied. The results of theoretical analysis and numerical simulation (for vertical dykes) are in good agreement with experimental data concerning overflow and total spillage as well as the fluid motion after collision, including the development and formation of a strong shock. The dependence of spillage on the inclination of the dyke is also determined by experiment.

Free shear layer stability measurements with a hot wire revealed that the probe itself can trigger and sustain upstream instability modes like the slit jet-wedge edge tones. The flow fields associated with the free shear layer tones induced in axisym-metric and plane air shear layers by a hot-wire probe and by a plane wedge were then explored experimentally, and found to be different in many ways from the widely investigated jet edge tone phenomenon.

As many as four frequency stages have been identified, there being a fifth stage associated with the subharmonic attributed to vortex pairing in the free shear layer. No evidence of hysteresis could be found in the shear layer tone. In the interstage jump (i.e. bimodal) regions, the tone occurred in only one mode at a time while intermittently switching from one to the other. Frequency variations in each stage are shown to collapse on a single curve when non-dimensionalized with the initial momentum thickness θe or with the lip-wedge distance h, and plotted as a function of h/θe.

Phase average measurements locked onto the tone fundamental show that the phase velocity and wavelength of the tone-induced velocity fluctuation are essentially independent of the stage of tone generation; in each stage, both phase velocity and wavelength decrease with increasing frequency but undergo jumps at starts of new stages. The measured amplitude and phase profiles, as well as the variations of the shear tone wavenumber and phase velocity with the Strouhal number, show reasonable agreement with the predictions of the spatial stability theory. The wavelength λ bears a unique relation to h, this h, δ relation being different from the Brown-Curle equation for the jet edge tone.

Shear layer tones would be typically induced in near-field shear layer measurements involving invasive probes, and can produce misleading results. A method for determining the true free shear layer natural instability frequency is recommended.

Non-parallel shear flows of an inviscid, incompressible, density-stratified fluid are considered. The stability is studied in terms of the possibility of complete mixing within a horizontal layer of given thickness. It is assumed that the energy interchange between the mixed region and the external fluid can be neglected. It is also assumed that the required turbulent energy is greater than the energy needed to invert the region to be mixed. The term ‘stable’ as used here means that the kinetic energy released by making the velocity constant over the layer thickness is not sufficient to provide the required turbulent energy for that layer thickness.

If each horizontal fluid plane has a translational velocity of the same magnitude and shear is produced by rotation of the velocity vector with increasing height, then ‘stability’ increases strongly with increasing layer thickness.

On the basis of a stability analysis of finite amplitude, two-dimensional convection, we have determined the dimensions of boxes containing fluid-saturated porous material in which convection is necessarily unsteady or steady and three-dimensional. For certain box sizes, convective rolls are unstable at Rayleigh numbers Ra lower than 380, the value below which rolls are stable forms of convection between infinite parallel planes. For Ra = 100 and 200, it appears unlikely that there are any box dimensions for which there is not a stable (possibly multicellular) two-dimensional steady motion. At Ra = 340 and 400, boxes in which rolls are unstable have heights which range from one to five times their horizontal dimensions.

The problem of the response of a porous elastic bed to water waves is treated analytically on the basis of the three-dimensional consolidation theory of Biot (1941). Exact solutions for the pore-water pressure and the displacements of the porous medium are obtained in closed form for the case of waves propagating over the poro-elastic bed. The theoretical results indicate that the bed response to waves is strongly dependent on the permeability k and the stiffness ratio G/K’, where G is the shear modulus of the porous medium and K’ is the apparent bulk modulus of elasticity of the pore fluid. The earlier solutions for pore-water pressure by various authors are given as the limiting cases of the present solution. For the limits G/K′ → 0 or k→ ∞, the present solution for pressure approaches the solution of the Laplace equation by Putnam (1949). For the limit G/K′→ ∞, the present solution approaches the solution of the heat conduction equation by Nakamura et al. (1973) and Moshagen & Tørum (1975).

The theoretical results are compared with wave tank experimental data on pore-water pressure in coarse and fine sand beds which contain small amounts of air. Good agreement between theory and experiment is obtained.