When work activities are anticipated, recognized, or found during an investigation (e.g., an outbreak investigation of an infectious disease) to involve risks to workers' health, preventive measures should be taken to control hazardous exposures. Respirators are often selected as a means of reducing workers' inhalational risks when engineering controls or administrative measures are insufficient or unavailable for controlling exposures to hazardous airborne contaminants, including infectious agents. Issues affecting the selection of respirators for reducing workers' exposures to infectious aerosols are reviewed in this paper. Infectious aerosols are dispersions of airborne liquid or solid particles capable of causing infection. The issues outlined in this paper are focused on workers in healthcare settings, but also apply in other settings where workers may be exposed to infectious aerosols. Methods that can be used to make a respirator selection are reviewed first. Choosing a selection method is the first important decision in a respirator selection process; the choices are the hazard ratio method, the risk analysis method, and the expert opinion method. Air concentration measurements and exposure limits applicable to infectious agents to which workers may be exposed are essentially nonexistent, and the absence of these essential data impedes the process of selecting appropriate respiratory protection. Until particle-size distributions and the viability and infectivity of particles comprising infectious aerosols generated in healthcare settings can be better characterized, the expert opinion method will likely continue as the method used most frequently to make respirator selections for healthcare workers. Specifying the rationale and all data inputs used in a respirator selection process is essential when using this method. Characteristics of infectious particles and the potential for airborne spread of infectious agents in healthcare settings are also discussed. The size of the particles comprising an infectious aerosol has received particular attention relating to the selection of respiratory protection for healthcare workers. Conflicting meanings of the term "droplet" are central to this issue. Evidence is presented in this paper supporting a position that 100 micrometers is the particle size defining the boundary between large-particle droplets and aerosol particles. Although workers caring for patients with contagious respiratory infections are at risk of exposure to large-particle droplets greater than 100 microm in diameter, their risks of inhalational exposure to infectious particles are likely to be predominantly to an aerosol consisting of a mixture of rapidly evaporating droplets and droplet nuclei that remain suspended in room air for prolonged time periods. Finally, information is provided demonstrating why surgical masks should not be worn as protection against infectious aerosols. The filter media of surgical masks allow penetration of small particles, and the poor fitting characteristics of their face seals allow the passage of particles at the edges of the masks. Thus, only NIOSH-certified respirators should be used as part of a strategy for protecting workers from inhalational exposures to infectious aerosols.