About 4900 values of 14CO2 activity have been measured on stratospheric air samples collected between 1953 and 1975 when the major nuclear weapon tests injected large amounts of 14C into the atmosphere. However, the validity of these data published in Health and Safety Laboratory reports was repeatedly criticized and their relevance is thus usually denied in model studies tracing the global carbon cycle with bomb 14CO2. To oppose this criticism, we perform here a comprehensive analysis of the measurements and calculate stratospheric bomb 14CO2 inventories for the period in question. We find out that the recognized weaknesses of the survey do not justify a general discrimination against the 14CO2 observations. Our 14CO2 inventories determined using numerical methods to interpolate the observations widely confirm the more hand-made results from a former study by Telegadas <1971> except in the northern poleward stratosphere. We are also able to clear away the reasons commonly advanced to call into question the stratospheric bomb 14CO2 inventories by up to 20%. These findings rehabilitate the most extensive data set of stratospheric 14CO2 observations and establish them, together with our corresponding bomb 14CO2 inventories, as a valuable observational constraint which should be seriously accounted for in global carbon cycle models and in other studies relying on an accurate simulation of air mass transport in the atmosphere. ¿ 2000 American Geophysical Union