We present a table giving the “present-day” (average over most recent ∼5 m.y.) azimuths of tracks for fifty-seven hotspots, distributed on all major plates. Estimates of the azimuth errors and the present-day rates for those tracks with age control are also given. An electronic supplement contains a discussion of each track and references to the data sources. Using this table, the best global solution for plates moving in a fixed hotspot reference frame has the Pacific plate rotating about a pole at 59.33°N, 85.10°W with a rate that gives a velocity at this pole's equator of 89.20 mm/yr (–0.8029 °/m.y.). Errors in this pole location and rate are on the order of ±2°N, ±4°W, and ±3 mm/yr, respectively. The motions of other plates are related to this through the NUVEL-1A model. The large number of close, very short tracks in the Pacific superswell region precludes all hotspots being rooted near the core-mantle boundary. In general, we think the asthenosphere is hotter than the mantle just below it (in the sense of potential temperature). Asthenosphere is very hot—it is brought up from the core-mantle boundary by plumes. The mantle is cooled by downgoing slabs, and a convective stability is established whereby mantle rises only at plumes and sinks only at trenches. We propose that this normal mantle geotherm is overwhelmed by much-larger-than-average mantle upwelling in superswell areas, making many short-lived instabilities in the upper mantle. Because soft asthenosphere so decouples plates from the mantle below, instabilities in the upper mantle (even above the 660-km discontinuity) are relatively fixed in comparison to plate motions. With the mantle velocity contribution being minor, tracks are parallel to and have rates set by plate velocities.