A modified version of the computer program written by Chan (1986) for the analysis of unbonded prestressed concrete members under third point loading was developed. The new program carries out the analysis more quickly and efficiently while still maintaining the required accuracy level. The program was used to analyse twenty-two unbonded beams with bonded steel tested in China and twelve unbonded slabs with bonded steel and a futher six unbonded slabs without bonded steel tested at the University of Canterbury. All analytical and experimental results were compared . The comparison revealed that the theoretical analyses gave good representation of the flexural behaviour of the unbonded concrete members both with and without bonded reinforcement. The report confirms the Chinese finding that the combined reinforcement index has very significant effect on the ultimate tendon stress increase and moment capacity for the unbonded members However, the effect of span-depth ratio on the flexural behaviour of unbonded partially prestressed concrete members was not significant. It is also confirmed that the unbonded members with bonded reinforcement behave in a more effective way than those without bonded reinforcement. The report recommends that bonded reinforcement should be used in design practice. Two recommendations based on experimental results are proposed for the determination of ultimate moment capacity of unbonded partially prestressed concrete members. The proposed expressions for member without bonded reinforcement is given just for comparison and it is not recommended for use in practical design. It is also recommended that the value of the combined reinforcement index chosen should be greater than 0.06 to prevent the occurrence of flexural instability and less than 0.305 for efficient design of the unbonded system. Value greater than 0.305 lead to a compression failure resulting in a low value of ultimate tendon stress increase and no increase in moment capacity.