The complete load-deflection history of 18 beams tested during 1952-54 and 18 beams tested during 1959-61 at the Univ. of Ill. allowed a study of the factors which contribute to ductility. All tests involved simple spans with load applied through a column stub at midspan to simulate the beam-column connection in a reinforced concrete frame. This study involved only beams without axial load. Primary variables studied included depth, compression steel ratio, concrete strength, amount of transverse reinforcement, and ratio of compression steel area to tension steel area. The tests indicate clearly the contribution of compression steel to ductility of reinforced concrete beams. The modes of failure reflected the importance of confinement of concrete in the compression zone in insuring the capacity for large plastic rotation at the joint. In all cases the first crushing of concrete (ϵ c = 0.004), usually considered as ultimate, was found to be associated with a deflection much smaller than that corresponding to the ultimate load. For beams with compression steel the reverse ductility beyond first crushing was large, the deflection at ultimate load in some cases being more than ten times the first crushing deflection.