Recent criticisms of the gauge factors usually employed in the 'testdipole ' method for the calculation of ' ring current' chemical shifts in conjugated molecules, are discussed. It is shown that, in a simple semiempirical scheme of the London-McWeeny type, insertion of a dipole contribution into the vector potential appearing in the gauge factor, whilst having no effect on the calculated ' ring current ' intensities, is algebraically analogous (and, at large distances from ring centres, numerically equivalent) to estimating the secondary field at the origin due to a set of classical line currents, as discussed originally by Salem ; these ' line currents ' are of the same magnitude as the quantum-mechanical 'bond currents' implicit in the ' ring currents' calculated using the simpler gauge factors originally due to London, but their contributions to the secondary magnetic field experienced by the peripheral protons, are estimated classically. Extensive numerical comparison is made between experimentally-observed proton chemical shifts in some conjugated hydrocarbons, and secondary fields estimated by this semi-classical formalism, and its predictions are found to correlate as well with experiment as do those of the original McWeeny approach. It is concluded that any further illegitimacy involved in the procedure of inserting a dipole contribution into the gauge factor, is evidently quite simply compensated for, numerically, by an appropriate empirical parametrization. Such empirical parametrizations are also thought to absorb errors due to all the other various approximations of the ' ring current ' theories (with apparently unwarranted efficiency), and they should, therefore, be treated with more scepticism than has previously been thought necessary.