The McWeeny ‘ring current’ theory is tested critically against a set of consistent, accurate experimental chemical shifts for 85 protons in 16planarunsubstituted condensed benzenoid hydrocarbons, and a regression equation is derived relating quantities calculated from the theory to experimentalτvalues for the non-hindered protons. Systematic deviations of different types of protons from this line are discussed, as also are sigma-bond anisotropy effects and discrepancies observed for overcrowded protons. Some of the assumptions about the geometry of planar benzenoid hydrocarbons normally made in these calculations are relaxed. The McWeeny theory gives a good account of the chemical shifts of non-hindered protons in planar molecules, and the predictions of the theory are not significantly altered if the calculations are based on experimental x-ray geometry. Sigma-bond anisotropy effects are unimportant for non-hindered protons, and, at most, make only a partial contribution to the down-field shifts of hindered ones.