Interaction curves for reinforced concrete columns, derived on the basis of Whitney's theory «> , or some variation thereof, can be plotted in dimensionless form for a given shape of column. This procedure is embodied in design handbooks such as that published by the American Concrete Institute( 2 ), and it has been extended to the case of prestressed columns by Zia and Guillermo (3>. When precast, prestressed sections such as single or double tees are used in load-bearing walls, they are subject to axial load and bending moment. Because of the irregular shapes of these members, their interaction curves do not lend themselves to dimensionless plotting. Nevertheless, various devices have been used to achieve the simplicity of the former problem, for example, substitution of an equivalent rectangular section with a radius of gyration equal to that of the actual section. It is the writers' contention that such devices are misleading and unnecessary, and that, in fact, the number of assumptions in the design procedure could and should be reduced rather than increased in connection with these members. This is based upon the following points: 1. The interaction curves of members such as prestressed T-sections are unusual in shape and are not well represented by drastic simplifications. In fact, the radius of gyration of the cross-section has no significance at all with respect to the interaction curve at ultimate strength, although it does, of course, affect the magnification factor associated with slenderness of beam-columns. 2. The justification for the use of such members as load-bearing wall panels is that they are massproduced in large numbers in precast plants. Therefore the economic pressure for highly simplified design procedures is reduced. 3. A consequence of Point 2 is that large numbers of cases are likely to be designed in advance for a plant as part of its promotional policy, and the use of a computer is strongly indicated. The need for simplifications is immediately removed. 4. The labor involved in calculating interaction curves, even by hand, is not, in fact, prohibitive if large numbers of members are to be manufactured. 5. Since there is no established