We examine topographical coupling between the Earth's fluid outer core and solid overlying mantle as a possible explanation of decadal period variations in the length-of-day. The topographical torque can be calculated straightforwardly, provided that the flow at the core surface and the topography of the core-mantle boundary are known. Such calculations generally give torques which are larger than required to explain the length-of-day variations; we show however that these calculations are extremely sensitive to the flow, and so cannot be taken as providing any observational evidence either for or against topographical coupling. However, observations of the magnetic field do provide useful constraints on the problem of topographical coupling, in particular constraining the timescale and the nature of the response of the core to topographical coupling.