Laboratory experiments have been used to study a wide range of convective processes which are significant for the formation of rocks and the ores that form in them. Metal-rich solutions flowing through the sea floor produce 'black smoker' chimneys; both the size and structure of a solid chimney and the height to which the plume of effluent rises above it can be predicted using model experiments. Analogues of processes in solidifying magmas can also be studied using crystallizing aqueous solutions. The release of lighter or denser residual fluid as crystallization proceeds, and the resulting compositional convection, have a vital influence on the differentiation of initially homogeneous magma and on the evolution of stratified magma chambers. Mixing during the replenishment of magma chambers, and its dependence on the rate of filling and on the fluid properties, have also been studied on the laboratory scale. Experiments carried out in different geometries have been applied here to the formation of platinum and chromite ores and komatiite lavas. Larger-scale mantle convection, in particular the motion of subducting plates and plumes rising from the core- mantle boundary, also determine the tectonic settings in which ore bodies form. These phenomena too can be usefully investigated in laboratory experiments using very viscous fluids.