When the contractile properties of single muscle fibres are studied, force is typically normalized by fibre cross-sectional area and expressed as specific force. We studied a set of 2725 chemically skinned human single muscle fibres from 119 healthy adults to determine whether specific force is the optimal way to express the relationship between single-fibre force and size. A linear mixed effects model was used to estimate the slope and slope variability among individuals of log-log plots of force and diameter. For type I fibres, the slope estimate was 0.99 (95% confidence interval 0.36-1.62), and for type IIa fibres it was 0.94 (95% confidence interval 0.77-1.11), indicating that force is proportional to fibre diameter, rather than to cross-sectional area. If force were proportional to cross-sectional area, the slope estimate would be 2.0. In future studies using the chemically skinned single fibre preparation, force may be normalized to fibre diameter rather than cross-sectional area. We propose that a new term, 'normalized force', be used for this variable, with units of newtons per metre. We demonstrate using our data set that when populations of single fibres are compared with one another, the determination of whether the size and force relationship is the same or different is dependent upon the method used to account for fibre size (i.e. specific forceversus'normalized force').