Polarization of Charmonium in \pi N Collisions

Measurements of the polarization of $\jp$ produced in pion-nucleus collisions are in disagreement with leading twist QCD prediction where $\jp$ is observed to have negligible polarization whereas theory predicts substantial polarization. We argue that this discrepancy cannot be due to poorly known structure functions nor the relative production rates of $\jp$ and $\chi_J$. The disagreement between theory and experiment suggests important higher twist corrections, as has earlier been surmised from the anomalous non-factorized nuclear $A$-dependence of the $\jp$ cross section.