Annu. Rev. Astron. Astrophys. 1997. 35: 389-443
Copyright © 1997 by . All rights reserved

Next Contents Previous


3.1. Number Magnitude Counts

Large format CCD and infrared-sensitive detectors have been used to extend ground-based number counts since Koo & Kron's (1992) review. The greater depth now available has confirmed a break in the count slope that occurs, respectively, at around B approx 25 (Lilly et al 1991, Metcalfe et al 1995b) and K approx 18 (Gardner et al 1993, Djorgovski et al 1995, Moustakas et al 1997) shown in Figure 5a. The slope, gamma = d logN / dm, ranges from gammaB = 0.47 to 0.30 and from gammaK = 0.60 to 0.25, but the surface density at the B break is approx30 times higher than that at the K break. If the two effects were manifestations of the same phenomenon, e.g. a decline in volume density beyond some redshift limit, then the mean B-K color should not change significantly across the break points. In contrast, the change in slope is accompanied by a marked increase in the number of galaxies with colors B-K < 5, which explains the steeper gamma B at fainter limits. Such blue galaxies were originally referred to as flat spectrum galaxies by Cowie et al (1989) because their SEDs approximate ones with constant f(nu).

Figure 5a
Figure 5b

Figure 5. (a) Differential galaxy number magnitude counts in the B and K passbands from the compilation of Metcalfe et al (1996) augmented with the Keck K counts of Moustakas et al (1997). The K counts have been offset by +1 dex for clarity. The two power law slopes (dashed lines) drawn have gamma (= d log N / dm) = 0.47,0.30 around B = 25 (Metcalfe et al 1995b) and 0.60,0.25 around K = 18 (Gardner et al 1993). The solid curve indicates a no-evolution prediction for an Einstein-de Sitter universe based on King & Ellis (1985) k-corrections and luminosity functions from local surveys with a normalization raised by 50% (see Section 4 for discussion). Beyond the limits drawn, the no-evolution prediction ceases to be reliable. (b) B-K versus K for galaxies from the Hawaii Bright Survey (fine dots, Huang et al 1997), the Hawaii Deep Surveys (Cowie et al 1996 and references therein), and HDF (Cowie 1997). Points with error bars refer to mean colors for various magnitude slices. The dashed line indicates the current limit for optical spectroscopy.

The break in slope and, more importantly, the disparate behavior between the B and K counts with respect to the no-evolution predictions makes it unlikely that a major portion of the excess counts arises via a nonzero cosmological constant, Lambda, as postulated by Fukugita et al (1990), Yoshii & Peterson (1991) (see Carroll et al 1992 for a full discussion). That such a dramatic excess should be seen in B but not K could only be consistent with the hypothesis of a nonzero cosmological constant if the B > 25 sources were significantly more distant than the K > 19 ones, which seems unlikely (Section 5). Note again that the K counts, by virtue of their insensitivity to the k-correction (Figure 1b), remain an interesting cosmological probe, although a satisfactory conclusion concerning Omega and Lambda will remain elusive so long as the evolutionary behavior is poorly determined (Djorgovski et al 1995).

The break in slope more probably signals a transition in galaxy properties at some redshift, with the bulk of the fainter sources being drawn from intrinsically less luminous sources at similar redshifts, a conclusion favored by Lilly et al (1991), Gardner et al (1993), Metcalfe et al (1995b). In this case, the abundance of fainter blue sources would suggest a high volume density of low mass galaxies and the slope of the B and K counts would directly reflect their relative contribution to the luminosity function at that time. Indeed, from Equation 5, clearly the break indicates the apparent magnitude beyond which the contribution of galaxies to the extragalactic background begins to converge. Specifically, in the B band that samples the rest-frame UV at the appropriate redshifts, the location of the break defines, albeit qualitatively, that era in which galaxies contribute most in terms of short-term star formation and associated metal production (Songaila et al 1990, Lilly et al 1996; Section 6).

Next Contents Previous