Tests are described in which short, rectangular steel columns filled with concrete are loaded to failure under axial load. Part of the load is applied at typical beam connections within the column length; and part is applied at the top, representing the load from upper stories of a building. The length of the column near the connection over which slip takes place is examined, and the ultimate strength of the columns is compared with that of stub columns loaded uniformly over their length. A small reduction in strength is found to occur, which depends upon the proportion of total load applied at the connection. An empirical means of accounting for this strength reduction is given. Key words: columns, composite construction, connections, tubular columns, axial compression.

Compression molding of randomly oriented strands (ROS) composites offers the possibility to manufacture complex parts with a fast processing cycle. In this paper, effects of pressure, strand size, temperature, and material placement in the mold cavity on the quality of a T-shape part were studied experimentally with carbon/polyether ether ketone ROS composites. Minimum filling pressure was obtained for the consolidation of a 25-mm deep cavity. Parts processed at filling pressure showed a void content no greater than 1.2%. Increasing pressure to 70 bar resulted in decreased void content as low as 0.03%. Mechanical testing of the T-shape showed similar strengths for parts processed at filling pressure and higher (70 bar). At the component level, initial strand placement greatly affected mechanical performance as the presence of a knit line was responsible for a reduction of 60% in tensile strength. The main findings show that processing a complex feature at filling pressure P fill was sufficient to reach nominal mechanical properties. This suggested that moderate porosity was not detrimental to the mechanical performance for the given tests where fatigue performance was not evaluated.

At 75/25 concentration ratio, bisphenol a polycarbonate (PC)/styreneacry‐lonitrile copolymer (SAN) blend has poor impact resistance compared to PC/ABS. A rubber phase methacrylate‐butadiene‐styrene (MBS) of core/shell type was dispersed in PC/SAN blend. The morphology of the unmodified and modified blend was investigated. The influence of the acrylonitrile ratio in the SAN on the microstructure was studied. It clearly shows that core/shell resides at the interface between PC and SAN. It seems that core/shell particles enhance the adhesion between the different phases. Their presence influences the interface mobility; i.e., the coalescence of the dispersed phase observed in pure PC/SAN is considerably reduced when the MBS particles are added. The impact resistance of the samples was correlated with the morphology.