The trace metals that have the greatest potential value as tracers for circulation and mixing in the ocean are markedly non-conservative, with short oceanic residence times. Their distributions strongly reflect the influence of sources either at boundaries or within the water column itself. The signals arising from these inputs can act as indicators for sources and transport pathways. These characteristics are well exemplified by dissolved manganese, which shows variations in concentration within the ocean that reflect the advection and mixing of waters that have acquired enhanced concentrations as a result of three principal processes. First, injection at the sea floor in hydrothermal fluids leads to anomalies in deep-water concentrations that are detectable over about 1000 km in parts of the Pacific Ocean. At the Mid-Atlantic Ridge, the advection of hydrothermal inputs is largely restricted by the topography of the axial rift valley, but the signals have proved valuable in exploration for hydrothermal sites. Secondly, pronounced maxima are associated with the highly oxygen-deficient waters of the oxygen minimum in parts of the eastern North Pacific Ocean. They are accounted for by in situ remobilization from particulate forms of manganese and by lateral advection from sources at the ocean margins. Interpretation of the data for the manganese maximum layer with two-dimensional box models has illustrated the potential to estimate lateral advective velocities on the basis of fluxes of manganese. Thirdly, the occurrence of maxima in the surface mixed layer of the ocean, which is particularly marked in areas where aeolian inputs appear to be an important source for dissolved manganese, suggests its potential use as a tracer for mixing and transport in the upper ocean. In all three cases, a more detailed and quantitative knowledge of the geochemical processes that determine the observed features will be needed if the potential for tracer applications is to be fully realized. Dissolved aluminium varies in concentration in deep-ocean waters by about two orders of magnitude, suggesting that it may be useful as a tracer for the movement and mixing of water masses. Systematic differences in concentration occur in water masses in the northwest Atlantic Ocean, but the origins of the differences remain to be clarified. Beryllium may also serve to identify water masses in the Atlantic Ocean. The few data available for cobalt suggest that this metal also may find applications as a tracer.