On the technique of obtaining samples of alveolar air for estimation of the carbon dioxide tension