Size-selective sampling is a health-related method to collect airborne particles based on penetration of inhaled particles into different regions of the human respiratory tract; thus, it is the most relevant sampling method to correlate health risks with occupational exposure. The current practice of sampling asbestos and other fibers is not a size-selective method. The thoracic size fraction, defined as the portion of inhaled particles that can penetrate through the larynx, has been suggested as the most relevant size-selective sampling method for fiber aerosol. The thoracic fraction is based on 1-deposition of inhaled spherical particles in the human extrathoracic airways for mouth breathing and corrected for the particle inhalability. There is no comparable information for fiber aerosols; therefore, there is no technical basis to ascertain whether the current thoracic fraction definition is suitable for fiber aerosols. No human data are available from controlled experiments of inhaled fiber aerosols for the obvious reason that most fiber materials are potentially hazardous when inhaled. Our approach was to measure penetration of fiber aerosol in realistic human oropharyngeal airway replicas and to compare that with data from spherical particles. We showed that realistic human oral airway replicas (including the oral cavity, pharynx, and larynx regions) provided useful spherical and fiber particle deposition in the human head airway. These data could be used to test the thoracic fraction curves. The spherical penetration is in agreement with human in vivo data used to establish the thoracic fraction curve. Fiber penetrations through the larynx of two human oral airway replicas were higher than those for spherical particles for the same aerodynamic diameter using the same replicas. The thoracic curve as defined for spherical particles, therefore, may not include some fibers that could penetrate to the thoracic region.