We consider an interpretation of the recent lattice data on the current-cur
rent correlators. in the x-space. The data indicate rather striking differe
nce between (axial)vector and (pseudo)scalar channels which goes beyond the
predictions of the standard non-perturbative models. We argue that if the
difference is to be explained by power corrections, there is a unique choic
e of the form of the correction. We discuss the emerging picture of the pow
er corrections. (C) 2001 Published by Elsevier Science B.V.