A panoramic view of radio galaxy evolution from a redshift of 0 to a redshift of 4.3

This paper is a continuation of the observational study of high-redshift (z > 2) radio galaxies (HZRGs) started by Eales & Rawlings (1993). In this paper we present infrared imaging and spectroscopy of two samples of HZRGs: a large inhomogeneous sample and a smaller statistically complete sample. Many of our basic results are simply confirmation of similar results in the earlier paper. The dispersion in the ratio of [O III] lambda 500.7 emission to radio emission is remarkably small, and HZRGs lie on the continuation of the relation between emission-line luminosity and radio luminosity seen for low-redshift radio galaxies, implying that the emission lines from an HZRG are from gas that is being photoionized by the active nucleus rather than by stars. The Ly alpha/H alpha ratios are much less than the values predicted by photoionization models and less than the ratios for low-redshift radio galaxies, implying, if resonant scattering is unimportant, that there is a large amount of dust reddening in these objects. Finally, there is a significant contribution to the K-band emission from emission lines, and this means that age estimates from the uncorrected colors are often much too high. In a companion paper, we show that the apparent increase in infrared luminosity of 3C radio galaxies between the present day and a redshift of similar to 1 is not the result of stellar evolution, but instead arises from an intrinsic correlation between infrared and radio luminosity. In this paper we examine the question of the evolution of radio galaxies by estimating the rest-frame optical luminosities (M(V)) and colors (U-V) for both the HZRGs in our study and the radio galaxies in two other samples. Since these quantities can be estimated over the entire redshift range in which radio galaxies are found, this procedure gives us a panoramic view of how the optical properties of radio galaxies have changed between a redshift of similar to 4 and the present day. We conclude that between z similar to 1 and z similar to 4 we are finally seeing genuine evolution in the optical luminosities of radio galaxies. Although it is possible to think of other causes, the most likely explanation of the increase in luminosity is the large change in the luminosity of a galaxy expected around the time most of the stars in the galaxy are forming.