Progressive threshold loading (PTL) is frequently used to assess inspiratory muscle endurance in health and disease. We and others have noted a systematic increase in endurance with the first few exposures to the task in subjects previously naïve to PTL, which may not be related to conditioning of the muscles themselves. The purpose of this study was to investigate the mechanisms responsible for this increased endurance by examining the ventilatory responses to 3 PTL tests, each > 24 h apart, in 18 healthy subjects. During PTL, threshold pressure (Pth) was increased by approximately 10% every 2 min until task failure. Subjects were allowed to adopt any breathing pattern. Respiratory muscle strength (maximal inspiratory pressure [PImax]) was unchanged over successive tests while maximal Pth (Pthmax) during PTL increased (69 +/- 17, 77 +/- 16, and 86 +/- 11% of PImax, respectively, p < 0.05) (mean +/- SD), indicating that the increased Pthmax could not be attributed to improved respiratory muscle strength. Breathing pattern changed with successive tests, so that for comparative loads inspiratory time (TI), respiratory frequency (f ), and duty cycle (TI/Ttot) decreased. This change in breathing pattern did not alter respiratory muscle efficiency (respiratory muscle V O2/work), which was similar in each test (2.4 +/- 2.2%), but perceived effort (Borg Score), which was maximal at task failure in each test, decreased at comparative loads with successive tests. Thus, Pthmax during initial tests appeared to be limited by discomfort rather than respiratory muscle function. These findings suggest that the increased Pthmax with successive tests is a consequence of differences in the breathing pattern adopted, reflecting neuropsychological rather than respiratory muscle conditioning. Measurements from PTL should only be used to assess respiratory muscle performance after allowing time for learning.