An experimental program and an extensive theoretical study on the strength and behaviour of masonry beam-columns were conducted. Thirty-nine large scale masonry walls including three plain and thirty-six reinforced specimens with pinned support conditions were tested under pure axial loading and combined axial and lateral loading. Masonry strains measured at both tension and compression surfaces of wall specimens were used to define curvatures at mid-height of walls. Moment vs. curvature curves were then established and used to determine effective flexural rigidities, EI_eff, at the time of failure of walls under various loading combinations.

Concurrent with the experimental progamme, an analytical technique based on a finite element model was developed and implemented to investigate the effects that geometric, material and loading parameters have on the strength and behaviour of masonry walls. The computer model accounts for tensile cracking of masonry joints, compressive crushing of masonry, amount and position of steel reinforcement, and the non-linear stress - strain relationship of masonry in compression. With this technique, the effects of varying flexural rigidity, EI, due to cracking and stress increase, as well as the stability of a deteriorated member are included directly in the analysis.

After verification of the finite element model by comparing analytical results with test datą obtained from this research and other reported literature, the computer model was implemented to conduct a detailed parametric study to investigate how various parameters affect the strength and behaviour and the ultimate effective flexural rigidity of masonry walls.

Elerr values at failure were obtained analytically and compared with values suggested in the Canadian masonry code, CSA S304.1-M94(1994). It was concluded that S304.1- M94 tends to underestimate Elerr values and thus leads to a conservative design over a range of parameters. This underestimation is most significant in the case of low eccentricities and high slenderness ratios.

Based on approximately 500 computer model tests, a lower bound bilinear limit for the effective rigidity of masonry walls was established as:​ EI_eff = [1-(2.375-0.0175h/t)e/t] EI₀ for 0.0 ≤ e/t ≤ 0.4 and EI_eff = (0.05+0.007h/t)EIo for e/t ≥ 0.4. This limit is based on a realistic analysis of stability and material failure including the effects of cracking and stress variation along the wall height and it is believed that it provides a more accurate and more realistic estimate of EI_eff than is currently provided by CSA S304.1-M94.