The reason why radio galaxies are so important is that they are the only stellar systems available in reasonable numbers at redshifts greater that one for which studies of their stellar populations can be made. Another reason why radio galaxies are so useful is that they form (almost) complete statistical samples with well defined selection criteria with the result that although there might be astrophysical biasses within the sample, because the samples are (nearly all) completely identified, the selection criteria are well defined.
The other key point is the importance of the near infrared waveband for these studies. First, the optical spectrum of a giant elliptical galaxy peaks at about 1 m and therefore, when observed at large redshifts, most of the energy is shifted into the infrared waveband, 1-2 m. Thus, it is relatively easier to detect very distant radio galaxies at 2 m as compared with the optical waveband. The second point is astrophysical and that is that the stars which contribute most of the light in the infrared waveband are stars belonging to the cool red giant branch in the Hertzsprung Russell diagram. These stars are derived from the oldest stellar populations of the galaxy. Therefore, when the integrated light of the galaxy is measured, the evolution of the stellar population of the galaxy is averaged over cosmological timescales. This contrasts strongly with what is observed in the optical region of the spectrum in which much of the light can be contributed by young stellar populations which are still undergoing their main sequence evolution. Thus, the optical light of a galaxy can be strongly influenced by bursts of starformation occurring throughout the life of the galaxy, whereas the infrared observations sample the majority old stellar population.