OCEANIC hotspot tracks generally indicate that the mantle plumes that give rise to them are fixed in position (relative to a lower-mantle reference frame) for timespans of similar to 100 Myr(1-3). It has not been clear whether continental hotspots are equally stable; for example, there is evidence that the velocity of the Yellowstone hotspot has been quite irregular with respect to the North American plate over the past 16 Myr(4-7), but measurements of its velocity based on caldera locations have been too uncertain and irregular to be definitive. Here I describe a new method of obtaining the plume velocity, which uses the broader lithospheric effects of the hotspot rather than the locations of individual calderas. For the past 10 Myr, the results indicate a constant North American plate velocity of 2.2 cm yr(-1), which agrees with independent estimates from global plate motion inversions(8,9). This velocity is considerably lower than previous estimates obtained from caldera locations, and clearly differs from the much higher (but less well constrained) velocities of the previous 6 Myr (refs 5, 7).