Parallel personal sampling was carried out with the open-face filter cassette and the IOM sampler for inhalable dust for nine types of organic dust. Parallel samples numbering 749 were obtained from 152 plants. Extremely large values and outliers were disregarded, and the remaining data for each type of dust were divided into subsets according to type of product or work task, and analyzed with the aid of linear regression. The coefficient of regression for each subset ranged between 0.2 and 0.7. Hypothetical occupational exposure limits (OELs) for inhalable dust were calculated based on the linear relation obtained between the dust concentrations measured with the open-face filter cassette and the IOM sampler. The fraction of person days with time-weighted average (TWA) concentrations exceeding the calculated hypothetical OELs for inhalable dust was obtained from the distribution of measured TWA inhalable dust concentrations. Based on the results of this study and the difference in sampling efficiency for large particles between the two samplers, it was concluded that the numerical value of the OEL for inhalable dust may be set at approximately twice the numerical value of the corresponding limit value for "total dust." Additional consideration of recently discovered health effects, and technical and economical factors may result in other numerical values of future OELs for inhalable dust.