The present work quantifies, through principal components analysis (PCA) the relationships among the variability of breath-by-breath ventilatory parameters [minute-ventilation (VE), tidal volume (Vt), and respiratory rate (FR)] during a maximal progressive exercise test. The results show that the first and second eigenvalues of the covariant matrix contains almost 90% of the variables' variance possible to see through the PCA, which means that the problem can be reduced by a two-dimensional analysis. The results show a close similarity between the global variability in two groups test, athletes and sedentary (control). For the athletes group, the parameter Vt is responsible for the high VE variability values while in the sedentary group the FR is more relevant for VE variability. The result improves the knowledge about respiratory variability during exercise, showing that Vt's and FR's variabilities contribute in different ways to global ventilation variability during a maximal cardiopulmonary exercise test in athletes and sedentary men.