|Annu. Rev. Astron. Astrophys. 1988. 26:
Copyright © 1988 by . All rights reserved
7.3. Reassessment of the N(m) Count Evidence for Luminosity Evolution at z > 0.5
Faint multicolor photometry of survey fields for the galaxy number count data is beginning to produce data that will eventually tell if the N(m, q0, E) excess for m > 21 is due to luminosity evolution, i.e. to the E(z) term in Equation 35. Ellis (1988) reviews the enigma that the colors become significantly bluer than expected at magnitude fainter than mJ ~ 21 (see also Kron 1980a, Hamilton 1985), where mJ is the J-band magnitude. Furthermore, this is just brightward of where the N(m) counts begin to show an excess from the predicted curves (Figure 2), an excess that grows to a factor of ~ 10 over the model at mJ ~ 26. Recall that this excess has been considered to be strong evidence for luminosity evolution.
However, the mean redshift at mJ ~ 21 for field galaxies is only z ~ 0.4, which is so small that no appreciable luminosity evolution is expected in any reasonable galaxy evolution model. This circumstance gives a clue to what is happening. Spectroscopy is not yet available for complete galaxy samples in the critical magnitude range of mJ > 21.5 so as to determine the redshift distribution. However, in an initial redshift survey that is complete between 20.0 < mJ < 21.5, Broadhurst et al. (1987, as summarized by Ellis 1987) found a subset of the complete 200 field galaxy sample that was blue and had strong emission lines. A further subset of these blue galaxies has a slope of d log N(m) / dm = 0.6 ± 0.2 for the counts, suggesting that they are nearby and intrinsically faint. If the blue galaxies are removed from the complete sample, the slope for the remainder of the counts is d log N(m) / dm = 0.34 rather than 0.44, causing the excess counts at faint magnitudes in Figure 3 to disappear. Clearly redshifts for the entire blue subset are required to make a stronger case. However, these data suggest that luminosity evolution for field galaxies at z 0.4 may not be needed to explain the N(m) counts. But if so, one of the stronger cases for evolution would disappear.