"He came to the conclusion which all subsequent observations have confirmed, that no sensible difference can be detected by Eudiometrical analysis between the purity of different specimens of atmospheric air. It was universally such "that the quantity of pure air in common air is 10/48" or... the per centage by volume of oxygen in air is 20.83. This... is remarkable... accuracy, when we consider how totally the great majority, not only of Cavendish's contemporaries, but also his successors, even among living philosophers, failed to obtain any constant results with nitric oxide eudiometers."