We addressed the inconsistency between the electron mass anisotropy ratios determined by the far-infrared experiments and DC conductivity measurements. By eliminating possible sources of error and increasing the sensitivity and resolution in the far-infrared reflectivity measurement on the single crystalline and on the polycrystalline La1.84Sr0.16CuO4, we have unambiguously identified that the source of the mass anisotropy problem is in the estimation of the free electron density involved in the charge transport and superconductivity. In this study we found that only 2.8 % of the total doping-induced charge density is itinerant at optimal doping. Our result not only resolves the mass anisotropy puzzle but also points to a novel electronic structure formed by the rest of the electrons that sets the stage for the high temperature superconductivity.