We describe in a historical perspective the different procedures proposed to calculate the drift shell parameter L when the secular variation of the geomagnetic field is taken into account. We compare the secular variation of the average particle flux along low-altitude circular orbits, obtained with the AP-8 and AE-8 trapped radiation models by using these different procedures. It is shown that the secular variation of the geomagnetic dipole moment is not the only cause of the unrealistic secular increase of the predicted flux, as was first assumed. The effect of the secular variation of the quadrupole terms as well as of the higher-order terms cannot be neglected. In order to avoid unrealistic predictions of particle fluxes at low altitudes, the proper way to use the AP-8 and AE-8 models is to calculate (B,L) with the Jensen and Cain 1960 magnetic field model for AP-8 MIN, AE-8 MIN, and AE-8 MAX, and with the GSFC 12/66 model, updated to 1970, for AP-8 MAX, and to use McIlwain's value 0.311653 Gauss R3E for the magnetic moment of the Earth.