Spontaneous breaking of CP symmetry by orbifold moduli [Phys. Lett. B 357 (1995) 387]

Due to the unappreciated existence of flat regions of the effective potential (&>, which can simulate a minimum unless high precision numerical calculations are performed, the detailed calculations in this paper are wrong. Correct calculations of all of the cases presented yield results consistent with zero CP violation. However, the general qualitative features noted are observed in other calculations, to be presented elsewhere [ 11. These are that although a priori one might expect that if the T moduli have complex values at the minimum of xff then the CP violation would be large, this expectation is generically not realized, (a) because the minimum of V& occurs at a fixed point of the modular function g2, or at some other zero, or (b) because when Re T is large compared with its fixed point value, then ImC^2 is small.