We consider a new version of the Einstein-Podolsky-Rosen (EPR) paradox
in which the measurements have limited accuracy, so that a photon num
ber is specified only to plus or minus a macroscopic number-of photons
. In our example the minimum quantum uncertainty product for the two r
elevant conjugate quantities is itself a macroscopic photon number. Th
is new version of the EPR situation is stronger than the original micr
oscopic versions, which were based on the validity, now disproved by B
ell experiments, of 'local realism' at a microscopic level. The new ve
rsion is based on the validity of 'macroscopic local realism', which h
as to date not been questioned.