This report describes a revision of the model used in "ICRP Publication 30" to calculate radiation doses to the respiratory tract of workers resulting from the intake of airborne radionuclides. This revision was motivated by the availability of increased knowledge of the anatomy and physiology of the respiratory tract and of the deposition, clearance, and biological effects of inhaled radioactive particles, and by greatly expanded dosimetry requirements. To meet fully the needs of radiation protection, a dosimetric model for the respiratory tract should: provide calculations of doses for individual members of the populations of all ethnic groups, in addition to workers; be useful for predictive and assessment purposes as well as for deriving limits on intakes; account for the influence of smoking, air pollutants, and respiratory tract diseases; provide for estimates of respiratory tract tissue doses from bioassay data; and be equally applicable to radioactive gases as well as to particles. In addition, large differences in radiation sensitivity among the respiratory tract tissues and the doses they receive from inhaled radionuclides argue for calculating radiation doses to specific tissues of the respiratory tract. To use these tissue doses effectively for radiation protection purposes, they must be compatible with the ICRP dosimetric system. Addressing all of these requirements has resulted in a dosimetry model that is more complex than previous models. This complexity is reflected in the structure of this report, which includes chapters on respiratory tract morphometry, physiology, and radiation biology or health detriment. Other chapters deal with deposition and clearance of inhaled radioactive particles, inhalation of radioactive gases, and application of the model to estimate respiratory tract doses. The model provides most of the flexibility needed to calculate doses to the respiratory tract for a wide range of exposure conditions and for specific individuals. This flexibility also allows for revision of reference parameter values as new information becomes available without changing the model. The wide availability of personal computers allows for easy use of the model.