Chemical and non-chemical contributions to breath-holding time (BHT) were directly determined by using a visual analog scale (VAS). These values were compared with those indirectly calculated from the method proposed by Godfrey and Campbell (1968). The magnitude of non-chemical factor at low PCO2 in our study was substantially less than the one obtained by the above investigators. We conclude that Godfrey and Campbell's model postulating linear augmentation of non-chemical sensation is inappropriate to explain dyspnea profile during breath-holding.