The authors commence by reviewing the most recent literature on methods for analysing postures and come to the conclusion that two factors are vital for a proper study of working postures. These are the description and the assessment of posture tolerability. The descriptive models most recently put forward in the literature enable the information required for subsequent assessment to be collected and further efforts to improve the descriptive capacity of these would be of slight benefit. The major difficulties in posture analysis arise during assessment. In the first place there is still a need for further basic knowledge to throw light on the natural and clinical development of changes due to unsuitable postures and to isolate the most significant individual or workplace-related factors. In the second place, the most widely used assessment methods (electromyography, study of discal pressures, biomechanical analysis, etc.) are capable of providing useful information on the individual apparatus considered, whereas a posture turns out to be tolerable only when all the apparatuses involved are comprehensively safeguarded. Lastly the above-mentioned assessment methods are not always capable of providing both information on the appearance of short-term postural discomfort and enabling predictions to be made of the appearance of damage in the long term, while both requirements must be met for a working posture to be considered as tolerable. For these reasons the authors are in favour of the combined use of several methods for the assessment of working posture tolerability. Practical examples provide guidelines to the authors' approach and conclusions are given.