Differences in the x-ray diffraction pattern of polyacrylonitrile (PAN) have been observed, both here, as well as in the work reported in the literature. Various workers have cited grossly differing unit cells for PAN and this has never been explained. In this work, it is shown that in dry samples, the diffraction pattern contains only two equatorial peaks, d1 almost-equal-to 5.3 angstrom and d2 almost-equal-to 3.0 angstrom. This is consistent with a hexagonal packing of rod-like chains of PAN with no chain-axis order. However, in solvent-plasticized films, the diffraction pattern is altered. It is proposed that certain solvents cause a hexagonal to orthorhombic transition. However, the orthorhombic polymorph appears to be a solvated form, where the solvent has co-crystallized with the polymer. After removal of solvent, the structure reverts to hexagonal packing. Hence, it is suggested that the discrepancy in cell parameters cited in the literature, particularly when large cell dimensions are quoted, may be due to the inadvertent presence of solvent which could have been introduced during sample preparation. Similarities with solvated crystals in other polymeric systems are also briefly discussed.