Erratum: Interference fragmentation functions in electron-positron annihilation

Equation (19) of Phys. Rev. D 67, 094003 (2003)PRVDAQ0556-282110.1103/PhysRevD.67.094003 turns out to contain mistakes, as pointed out recently in Phys. Rev. D 97, 074019 (2018)PRVDAQ2470-001010.1103/PhysRevD.97.074019 by H. H. Matevosyan et al. In particular, Eq. (19) contains an erroneous explicit dependence on the angle φl, and the terms linear in g^ have an opposite sign with respect to what one would deduce from Fig. 1. As a result, the expressions for the projection of the cos(2(φR-φR̄)) azimuthal dependence are incorrect. Specifically, Eqs. (34), (37), and (38) should all vanish, which means that the cos(2(φR-φR̄)) projection is not a measure of the di-hadron fragmentation function G1, as concluded in the paper. The vanishing of the projection explains the nonobservation of the related asymmetry in the recent experimental search by the BELLE Collaboration (A. Abdesselam et al., arXiv:1505.08020). We refer to Phys. Rev. D 97, 074019 (2018)PRVDAQ2470-001010.1103/PhysRevD.97.074019 for the correct version of Eq. (19) and for an alternative nonzero projection of the asymmetry that does probe the function G1.