Eight adult manual resuscitators were compared with regard to weight, dimensions, number of parts, possibility of misassembly, durability, self-reinflation time, average stroke volume with male and female operators using one hand and two hands, maximum cycling rate at room temperature and low temperature, delivered oxygen concentration, per cent demand ventilation, and ease of use. Of the devices tested, the Robertshaw and the Hudson Lifesaver had design problems in the oxygen delivery system, reducing the effective cycling rates. The Ohio Hope I delivered inadequate oxygen concentrations, and the Ohio Hope II delivered inadequate oxygen concentra- tions when used without a reservoir. The Hope II was also found awkward to use because of the bulky reservoir. The Puritan PMR delivered low oxygen concentrations and its valve required a screwdriver for disassembly and cleaning; a possible mis- assembly hazard was also noted. The Air-Shields Ambu Mark II, the Laerdal II, and the Penlon units rated high with respect to the tested variables, with the Penlon and Laerdal II devices achieving oxygen concentrations greater than 95%.