A deviation between the angle of convergence during visual work and the resting level of the vergence system (vergence discrepancy) should be associated with costs; subjects should tend to reduce these costs by increasing the observation distance. The remaining costs should contribute to visual fatigue. Both hypotheses were tested using a sample of 40 subjects, mainly students. In agreement with the first hypothesis, subjects with large initial vergence discrepancy increased their observation distance more in the course of 1 1/2 hours of visual work than subjects with smaller initial vergence discrepancy. Visual fatigue was assessed using a set of questions developed in two pilot studies. Contrary to the second hypothesis, visual fatigue did not depend on effective vergence discrepancy, that is, the deviation between the resting level of the vergence system and the actual angle of convergence. In addition, such a relationship was not found for further dependent variables (performance, BMS questionnaire, critical flicker frequency, evaluation of conditions of work, postural complaints). The results are discussed with respect to the role of tonic and dynamic components of the vergence system for visual fatigue.