Two experiments were designed to evaluate human sensitivity to elasticity. Elastic objects bend when a force is applied to them. Observers saw two computer-generated bending rods (defined by the motions of 50 dots) on any given trial and were required to judge which rod was more flexible. Elasticity difference thresholds were calculated for each observer for each of three bending conditions. The rods bent in a plane that was either frontoparallel or oriented 42.5 or 85 degrees from frontoparallel. The results showed that observers could precisely discriminate between bending rods of different elasticities, independent of whether the bendings occurred in the frontoparallel plane or in depth. To rule out the possibility that the ability to judge bending motion was based on the ability to judge 2-dimensional (2-D) speed a second experiment was conducted to obtain difference thresholds for 2-D speed. The observers' speed discrimination thresholds were not positively correlated with their elasticity discrimination thresholds, which suggests that the observers' ability to judge bending motion was not based on their capability to discriminate differences in speed.