Running speed is limited by a mechanical interaction between the stance and swing phases of the stride. Here, we tested whether stance phase limitations are imposed by ground force maximums or foot-ground contact time minimums. We selected one-legged hopping and backward running as experimental contrasts to forward running and had seven athletic subjects complete progressive discontinuous treadmill tests to failure to determine their top speeds in each of the three gaits. Vertical ground reaction forces [in body weights (W(b))] and periods of ground force application (T(c); s) were measured using a custom, high-speed force treadmill. At top speed, we found that both the stance-averaged (F(avg)) and peak (F(peak)) vertical forces applied to the treadmill surface during one-legged hopping exceeded those applied during forward running by more than one-half of the body's weight (F(avg) = 2.71 +/- 0.15 vs. 2.08 +/- 0.07 W(b); F(peak) = 4.20 +/- 0.24 vs. 3.62 +/- 0.24 W(b); means +/- SE) and that hopping periods of force application were significantly longer (T(c) = 0.160 +/- 0.006 vs. 0.108 +/- 0.004 s). Next, we found that the periods of ground force application at top backward and forward running speeds were nearly identical, agreeing to within an average of 0.006 s (T(c) = 0.116 +/- 0.004 vs. 0.110 +/- 0.005 s). We conclude that the stance phase limit to running speed is imposed not by the maximum forces that the limbs can apply to the ground but rather by the minimum time needed to apply the large, mass-specific forces necessary.