Thus it «s better that very many different thermometers are used by many different people following different
practices than that a few
ships using one thermometer each make the
measurements using procedures that vary only slowly, unless accurate data is available on the calibration of those few thermometers and the procedures used.