Bosonic string scattering amplitudes on the disc and the projective plane are compared with amplitudes from the effective field theory derived by demanding consistency of string propagation in background massless fields. Path integral (as opposed to BRST) quantization is used throughout, with the ghost fields integrated out. We point out a difficulty in defining a covariant dilaton vertex operator in this formalism. A nonstandard graviton vertex operator is constructed, by which we compute the vacuum energy for the two topologies and find that it has the correct ratio to the dilaton tadpole, as predicted by the effective Lagrangian. Factorization of tadpole divergences in N-point amplitudes leads to the same result, modulo a paradox discovered by Fischler, Klebanov, and Susskind, which is reviewed. We also show that the dilaton two- and three-point functions agree with the exponential form of the dilaton potential in the effective action, despite quadratic divergences that initially appear in these amplitudes.
Read full abstract