Erratum to"Azimuthal asymmetry in electro-production of neutral pions in semi-inclusive DIS"published in Phys. Lett. B522 (2001) 37

We correct our analysis of the HERMES experiment for the azimuthal sin(phi)-spin asymmetry in semi-inclusive hadroproduction in DIS on longitudinally (with respect to the lepton momentum) polarized target because of discovered misprint in sign in the paper P.J.Mulders and R.D.Tangerman, Nucl. Phys. B 461 (1996) 197.

1 D 1 = (6.3± 2.0)% of DELPHI [7], such recalculation results in asymmetry values about twice smaller than the experimental data.A better agreement is, however, achieved with Correction to Fig. 2 in Ref. [5].Kinematics of the process lp → l ′ hX.Note the orientation of the azimuthal angle φ which corresponds to the convention of HERMES [9].In Refs.[1,2] the azimuthal angle is defined as (2π − φ).Correction to Fig. 4a in [5] Correction to Fig. 4b in [5] Correction to Fig. 4c in [5] Figure 3: Corrections to Figs.4a, 4b and 4c in Ref. [5].Azimuthal asymmetries A W (φ) U L (x, π) weighted by W (φ) = sin φ (solid line) and sin 2φ (dashed line) for the production of π 0 , π + and π − as function of x.The experimental data are from Refs.[8,9].Rhombs (squares) denote data for A sin φ U L (A sin 2φ U L ).The theoretical curves have an uncertainty due to the statistical and systematical error of the DELPHI result, eq.( 1), and the theoretical uncertainty of the model.
the "optimistic" value of DELPHI obtained from the whole available interval of polar angles 15 • < θ < 165 • in the DELPHI experiment [7].The results of these recalculations in comparison with the HERMES data are presented in Fig. 2 and Fig. 3 which replace Fig. 3c and Fig. 4 of Ref. [5].
It is interesting to note that the negative sign of the transversal contribution leads to a change of sign of asymmetries for x > 0.4.This is due to a harder behaviour of h 1 (x) with respect to h L (x) (as seen in Fig. 3b of ref. [5]).It should be noted that the prediction of A sin φ U L (x, π) = 0 at x ≃ (0.4 − 0.5) is sensitive to the approximation of favoured flavour fragmentation, which has been used in Ref. [5].In principle one could conclude from data, how well this approximation works.However, the upper x-cut is x < 0.4 in the HERMES experiment [8,9].
The corrected values for the totally integrated asymmetries are and replace the numbers in Table 1 of Ref. [5].The numbers in Eq.( 2) have an uncertainty due to the statistical and systematic error of the DELPHI result, Eq.( 1), and moreover an uncertainty of around 20% due to the theoretical uncertainty of results from the chiral quark soliton model.
The new estimate of the z-dependence of the analyzing power H ⊥ 1 (z)/D 1 (z) from the z-behaviour of experimental asymmetries, using as an input the transversities from the chiral-quark soliton model [10], is presented at Fig. 4 with a linear fit H ⊥ 1 (z) = (0.33 ± 0.06) z D 1 (z) and with average H ⊥ 1 / D 1 = (13.8± 2.8)% which is in good agreement with DELPHI result eq.( 1).
We would like to thank H.
z) vs. z, as extracted from HERMES data [8,9] on the azimuthal asymmetries A sin φ U L (z) for π + and π 0 production using the prediction of the chiral quark-soliton model for h a 1 (x) [10].The error-bars are due to the statistical error of the data.
b.The same as Fig. 4a with data points from π + and π 0 combined.The dashed line in both figures is the best fit to the form H ⊥ 1 (z)/D ⊥ 1 (z) = a z with a = 0.33.

Figure 4 :
Figure 4: Corrections to Fig.5 in Ref.[5].a. H ⊥ 1 (z)/D ⊥1 (z) vs. z, as extracted from HERMES data[8,9] on the azimuthal asymmetries A sin φ U L (z) for π + and π 0 production using the prediction of the chiral quark-soliton model for h a 1 (x)[10].The error-bars are due to the statistical error of the data.
[1,2]an and A. Kotzinian for stimulating discussions, and P. J. Mulders for conversations on signs in[1,2].The work of A. E. is partially supported by RFBR grant 00-02-16696, INTAS grant 01-587, by Heisenberg-Landau Program and by BMBF and DFG.This work has partly been performed under the contract HPRN-CT-2000-00130 of the European Commission.