The external CO2 dead space volume (Vd) in facial visors and respiratory protective devices is difficult, but important, to measure in human subjects. The lack of proper methods for its assessment has hampered the development of standards and the improvement of visor and device design. We have improved and evaluated a method for measuring Vd and the mean inspired fraction of CO2 (FI,CO2) in human subjects wearing facial visors or respiratory protective devices. The method is based on indirect measurements of inspiratory volumes using a calibrated respiratory inductive plethysmograph (RIP), and measurements of FCO2 with a mass spectrometer. The accuracy of the RIP method was assessed in eight subjects and its repeatability was studied during nasal and oral breathing, at rest and after bicycle ergometer work. We related the Vd,RIP results to two known external VdS (with the effective Vd equal to the geometric volume; 15 ml and 320 ml), through which the subjects breathed, using RIP and a pneumotachometer (PTM) simultaneously (Vd,RIP and Vd,PTM). The repeatability of Vd and mean FI,CO2 results was determined from duplicate recordings during the wearing of a welding visor. Initial RIP calibration was accepted if the inspiratory tidal volume error was < or = 10%. This resulted in an acceptable Vd,RIP error (< or = 20%; Vd = 320 ml) in six eight subjects. The validation technique allowed us to separate the Vd,RIP error into a volumetric error and an error related to CO2 measurement. Poor RIP volumetric accuracy over the initial portion of inspiration was detrimental to Vd accuracy. Using the welding visor, Vd and mean FI,CO2 were less at rest than after work and less during nasal breathing than during oral breathing. The intra-individual variability of Vd and mean FI,CO2 were lower during nasal breathing than during oral breathing. To summarize, the improved RIP based method can accurately (< or = 20% error) assess Vd and mean FI,CO2 in facial visors and respiratory protective devices in standardized work situations. A meticulous RIP calibration procedure and repeated validations of RIP volume and CO2 measurement accuracy must, however, be applied.