Maximal inspiratory pressure (MIP) is often relied upon as an index of inspiratory muscle strength, and reproducibility of MIP taken to indicate maximal effort. This study was designed to determine whether reproducibility is a valid indicator of maximal effort.Ten normal subjects were studied, all of whom were familiar with the MIP test but none was an experienced subject. They were told that the purpose was to measure how accurately they could generate 50% of their MIP. Each performed nine MIP efforts and nine submaximal efforts. Means and coefficients of variation of peak negative inspiratory pressure (Pmax) and the ranges of the best three efforts were calculated for each type.Mean (SE) Pmax averaged-93.8 (6.0) cm H2O for the maximal efforts and -60.6 (7.7) cm H2O for the submaximal trials, with coefficients of variation averaging 8.71 (1.75)% and 14.58 (2.63)%, respectively and the ranges averaging 6.5 (1.1)% and 13.4 (3.5)%, respectively. There was no clear separation between the coefficients of variation or ranges of maximal and submaximal efforts. In four cases the ranges of the best three submaximal efforts were less than 5 cm H2O and less than 5% -criteria that have been used to validate MIP results. These four subjects had lower ranges for submaximal than maximal efforts, even when expressed as percentages of the means.Reproducibility should not be relied upon to indicate a valid MIP test, especially for research purposes when relatively small changes in inspiratory muscle strength must be discriminated.