In order to produce papers with differing qualities, papermakers use pulps with varying fibre and fines characteristics. Since the early 1900’s, numerous studies have been conducted to investigate the relationships between pulp-fibre and paper properties. However, a common weakness was that previous investigations focused on relationships within restricted ranges of properties of pulp-fibre and paper. Forth is reason,the main goal of this investigation was to examine and to characterize a broad scope of fibre-paper relationships. In order to achieve this goal, a statistical approach was taken. This technique explained much of the total variation in paper properties by investigating a large number of pulp-fibre types, including fines,with a broad range of characteristics. It was discovered that robust and orthogonal non-linear multiple regression models could be developed to predict various paper properties. The multiple curvi-linear regression models reported here (based on the PhD thesis of the first author of this paper) could explain, on the average, 85 ± 10% of the -variance (R2)in the paper properties. Although the models generated are empirical, and thus lack fundamental interpretative meaning, they do clearly rank those fibre properties most important fora given paper property and allow prediction of the “form” of the relationship. These relationships confirm the expectation that there are no universally optimum fibre paper properties. Instead, compromises must be made to achieve an acceptable balance of properties. Such interactions are described in more detail in the paper.