S U M M A R YThe variation of ocean depth with age in the absence of crustal thickening and dynamic support places valuable constraints on the thermal and rheological properties of the lithosphere and asthenosphere. We have attempted to estimate this variation using a global data set of shiptracks, with particular emphasis on young ocean floor. In this respect, this paper extends a previous study published in this journal by the same authors, which concentrated on the older parts of the ocean basins. We find that, prior to 80 Ma, subsidence patterns are reasonably consistent, with gradients of 325 ± 20 m Ma −1/2 and zero-age depths of 2600 ± 200 m. There is a strong inverse correlation between zero-age depth and the gradient of depth with the square root of age which is unrelated to local variations in dynamic support. Global depth-age trends to 160 Ma are not significantly different to those for the individual ocean basins. Within corridors of similar basement age, gravity-topography correlations are consistently 30 ± 5 mGal km −1 . Simple isostatic theory and numerical modelling of mantle plumes suggests that, if the minimum depth of convection is defined by the base of the mechanical boundary layer, the admittance should be a function of plate age. The observation that it is not implies that the active convective upwelling beneath young lithosphere ceases at the same depth as it does beneath old oceanic plates. This result is consistent with geochemical modelling of melts near mid-ocean ridges. We have examined the relationship between residual topography and gravity worldwide, and have found that good spatial correlations are restricted to the Atlantic, North Pacific and youngest Indian ocean basins. By contrast, residual topography and gravity are poorly or negatively correlated in the South and young North Pacific Ocean and in the older Indian Ocean. Away from regions of thick crust and flexure, histograms of residual topography and gravity have symmetric distributions about zero. We then use this residual topography to estimate the volume and buoyancy flux of seven major plume swells. In Hawaii, the clear correlation between melt and swell volumes in discrete age corridors is evidence that the horizontal velocity of the hot plume material far downstream from the plume is similar to the plate spreading velocity and that the plume pulses over time. Finally, comparison with seismic tomographic models suggests that the long-wavelength (>2000 km) residual topographic and gravity anomalies have an origin deeper than 250 km. This result is consistent with observations that the admittance is approximately constant at wavelengths longer than 800 km.