This is an experimental study of the plastic yielding of steel columns. It consists of tests of four specimens; two tension coupons, a 12-inch stub column and a 21-inch stub column. The tension tests were used to determine the physical properties of the material while the stub columns demonstrated the effects of residual stress upon the initiation of yielding and the propagation of the yielding. The test procedure used in both the tension tests and the compression tests was the same. The specimen was loaded in increments up to a load just, below the yield point. Beyond that point the specimen was strained at a constant rate in order to neutralize the effects of creep. This test procedure required that all the data for each set of readings be adjusted to give the strain at any gauge on the specimen at the same instant of time. This was done by considering the differences between consecutive sets of readings and adjusting all strains to the strain at the gauge showing the greatest change during the set. This was done by means of a computor [sic] program. During the tests of the stub columns local buckling of the flanges and web was restrained by a system of bars and bracing frames. The system used in the test of the 21-inch stub column proved to be particularly successful. Data was collected from strain gauges in the tension tests and strain and dial gauges in the compression test. The dial gauges in the compression test provided confirmation that the strain gauges accurately represented the state of strain in the specimens and also rotation of the upper end of the stub columns during the tests. There are several general conclusions which, it must be emphasized, are based upon a small number of tests. The yield stress of the tension specimens was greater than of the compression specimens. Yielding initiates at several independent foci rather than propagating from one single point. Yielding, once commenced, generally continued without stopping well into the strain hardened region. Therefore, there are both unyielded material and strain hardened material in the specimen at the same time. The initiation and progress of plastic deformations in the specimen are greatly affected by even small local irregularities. The effect of residual stresses upon the initiation of yielding is not uniform. The strain hardening modulus of the specimens used has been found to be only about one-half of that expected. Wide variation of this property is common. / Applied Science, Faculty of / Civil Engineering, Department of / Graduate