5 Colors of high redshift radio galaxies

Based on our previous work on GPS host galaxies in the near infrared (De Vries et al. 1998a, 1998b), we know their galaxy colors to be consistent with those predicted by stellar synthesis models (e.g., Bruzual & Charlot 1993). Since we are imaging GPS galaxies around a redshift of 1, the expected R-K color is around 5 to 6. Indeed, the sources 2121-014 and 2128+048 both have an R-K color of $\sim$5.7 and measured redshifts of 1.16 and 0.99 respectively. These points lie close to the passive evolutionary track (cf. Fig.  3), however, in the case of 2121-014 with its large uncertainty in the published R band magnitude ($\sim$1 mag, Snellen et al. 1996) this may be coincidental. Based on the R-K colors of the two other sources, $\sim$5.4, and $\sim$6.0 for 1942+772 and 2322-040 respectively, their most likely position in the plot is indicated by the overplotted box. The implicit assumption here is that the sources have colors matching those of an old, passively evolving galaxy, which was found to be the case for our near-IR and NICMOS samples (De Vries 1999).

Judging by their position in the plot, these sources are probably going to be among the highest redshift GPS galaxies currently known. The remaining empty fields (e.g. 0008-421, cf. Paper I) are presumably at higher redshift still. This extension of the redshift baseline for our GPS sample is much needed to distinguish for instance between the various evolutionary models. Note how in Fig. 3 the tracks start to diverge around the z=1mark. Currently 2121-014 is among one of the very few z>1 GPS galaxies, and not very well constrained in R magnitude. Obviously moving to large aperture telescopes and into the near-IR in order to minimize possible non-stellar color contamination (e.g. De Vries et al. 1999), will improve significantly on the yield of color-redshift plots like these.

Figure 3: Predicted galaxy colors based on stellar synthesis models. Dashed lines indicate a fixed mean stellar age, 10 Gyrs for the top track, 5 Gyrs for the bottom one. The solid lines are for passively evolving galaxies, with an initial burst of starformation at z=6 (top) and z=3 (bottom). All tracks are computed assuming a mean solar metallicity. The points are R-K colors for 2128+048 and 2121-014 (left and right data point, respectively). The overplotted box indicates the range where the data points of 1942+772 and 2322-040, both without redshift information, are most likely to land on the plot

Figure 4: Finding charts for the GPS sources, with accurate radio positions overplotted

Figure 4: continued

Figure 5: Keck K band images of 4 optically faint GPS galaxies

The published images for these sources are not very deep, and often barely detect the host galaxy (cf. notes on individual objects). Our Keck images not only unambiguously show the galaxies, but also provide morphological information. The source 2121-014, for instance, has a double structure (cf. Fig. 5) reminiscent of two closely interacting galaxies.

To summarize, in order to complete the identification and redshift determination of the faint end of our GPS sample, we need to use facilities with comparable collecting areas, i.e., 8 meter class telescopes. In an upcoming paper we will report on results from our first VLT program.