It is pointed out that, contrary to the view expressed by Nakanishi, there is no real reason why one should rigidly follow the so-called "real space-momentum" prescription when "pinching" occurs in a field theory with indefinite metric. Since by adopting the modification introduced by Cutkosky et al., a relativistic and unitary $S$ matrix can be obtained, that should be the obvious choice.

• Received 1 December 1970

DOI:https://doi.org/10.1103/PhysRevD.3.1046