On the possible significance of the quadratic effect of age on lung-function measurements