An in vitro study was designed to define the optimal oxygenator design for loading perfluorocarbon emulsions (PFCEs) with oxygen. A circuit was designed to circulate fluid between a deoxygenating device and a test oxygenator. Protocol I defines the increment in oxygen transfer rate attributable to a PFCE added to a crystalloid solution. Protocol II compares the oxygen transfer rates for four oxygenator designs (two hollow fiber microporous membrane oxygenators, one silastic true membrane oxygenator, and one bubble oxygenator). The null hypothesis is that the oxygen transfer rates to a crystalloid PFCE mixture are identical for all oxygenator designs.