Subject selection is an important factor in surveys of hearing and its influence is demonstrated. The self-recording audiometer appears to be a little more sensitive than the manual instrument but no more reliable and trace reading needs to be standardised. Audiometer calibration standards should not also serve as “normal” hearing levels. A presbyacusis standard for urban non self-selected males and a new technique for the assessment of “hearing loss” are proposed. The long term recoverability of acoustic trauma is demonstrated and modified definitions of TTS and PTS are suggested with regard to short and long term reversal. The question of the role of audiometry in the prevention of acoustic trauma is posed and left open.