The study of Part 2 showed that the possibility was very small of realizing the fluorescent lamps with a general color-rendering index of R_a =100, whose spectral power distribution had line spectra similar to ordinary fluorescent lamps.
This conclusion was based on the special solution of R_a =100 which was obtained for the constraint of minimum deviation σ between both spectral power distributions of the test illuminant and the reference.
However, there might be other methods of optimization leading to spectral power distribution with positive powers everywhere in the spectrum as suggested by G. Wyszecki.
In order to discuss this problem, characteristics of a general solution of eqs.(10) and (11) in Part 2 were discussed in the present paper.
Examination of a general solution showed the impossibility of getting the expected illuminant for any other method of optimization which permits the use of “constraints” on the solutions.

View full abstract