An important step in the verification of the reliability of portal films as in vivo dosemeters is the evaluation of the agreement between exit dose profiles and optical density profiles measured on the portal film. To test the possibilities of a conventional portal film system in 60Co beams suitable for head and neck irradiation, we verified the agreement between relative exit doses (measured by ionization chamber) and relative optical densities, on cubic homogeneous phantoms, on an homogeneous "step" phantom and on a cubic phantom including air and aluminium inhomogeneities. The optical density profiles were corrected with the appropriate sensitometric curves. For an homogeneous phantom 10.8 cm thick and with the film in contact with the phantom, the agreement was found to be excellent with a mean deviation of 0.8% and a maximum deviation of 1.5%. The agreement was worse when the air gap between the exit surface of the phantom and the portal film was increased (with an air gap equal to 15 cm the maximum deviation was 4%), and when the thickness of the phantom was increased (for a thickness of 14.4 cm the maximum deviation was 3.1%). The agreement was found to be acceptable for the "step" phantom too, with a mean deviation around 1% and a maximum deviation within 2% (air gap equal to zero). When air and aluminium inhomogeneities were incorporated into the phantom a maximum deviation of 6% and a mean deviation less than 3% were found. Furthermore, the relative optical density profiles show an underestimate of measured off-axis exit dose values under a high density inhomogeneity and a small overestimate under a low density inhomogeneity. Results suggest the possibility of using conventional portal films for exit relative dosimetry in head and neck irradiation with 60Co beams if the air gap is kept as small as possible.