The quality and quantity of saliva may be important in defending against pathogens transmitted via the buccal cavity. The aim of the present study was to examine the effect of cycling to exhaustion at moderate and high intensity on various salivary parameters and the time course of recovery. Eighteen male subjects of mixed physical fitness took part in the study. Subjects performed two bouts of exercise on separate occasions at least one week apart. Following an overnight fast, subjects cycled on an electrically braked cycle ergometer at a work rate equivalent to 80% VO2max until exhaustion. On another occasion they cycled on the same ergometer at 55% VO2max for 3 h or to fatigue (whichever was sooner). The order of the rides was randomised. Timed, unstimulated saliva samples were collected pre-exercise, during exercise, at cessation of exercise and at 1, 2.5, 5 and 24 h post-exercise. Saliva samples were analysed for IgA, total protein and osmolality. Saliva flow rate was significantly reduced by exercise (P < 0.01). Saliva IgA concentration, secretion rate and ratio to osmolality increased during exercise (P < 0.01). IgA to protein ratio did not change significantly during exercise. Since saliva protein secretion rate increased during exercise (P < 0.01) it appears that correcting for loss of saliva water by expressing IgA relative to protein is misleading. IgA secretion rate and IgA to osmolality ratio are more appropriate measures and neither parameter was lowered by exercise. The results of this study indicate that exercise may detrimentally affect the quantity of saliva produced, but not the quality of saliva. Furthermore, when exercise is to exhaustion, the intensity of the bout does not appear to influence the saliva response. Neither exercise protocol had any long term effect on saliva as all variables recovered within 1 h post-exercise.