The Extension of the Hubble Diagram. Il New Redshifts and Photometry of Very Distant Galaxy Clusters: First Indication of a Deviation of the Hubble Diagram from a Straight Line

Redshifts are given for 50 brightest cluster galaxies, extending as far as z = 0.75; BVR photometry is given for 33 clusters. These data are combined with earlier data of a similar kind in order to investigate several effects. The measured B V and V R colors as a function of redshift are well represented by Whitford's standard-galaxy K corrections, as far as these are defined (to · z = 0.28 in B V and z = 0.48 in V R). This suggests both that the K corrections are valid over these ranges of z and that no major color change of the galaxies has occurred over the last 4-5 x 109 years. At larger redshifts, the colors, which start out being monotonically redder with z, turn over and become bluer with z. The data at large z seem to follow the prediction based upon ultraviolet photometry of NGC 4486 (M87), which is one extreme of a range of galaxies measured by Code and Welch. Other standard corrections to the measurements are discussed, and formal least-squares values of q0 are computed. To explore the possibility that one might be stalemated by the fact that the value of q0 must, in principle, be known a priori to compute the aperture correction, the data are corrected separately for assumed q0 values of 0 and + 1. The subsequently computed values of q0 differ by only 0.2 between the two cases, which indicates that a simple iterative procedure will converge to produce a self-consistent value of q0• The new data do not significantly change earlier discussions of the corrections for cluster richness and Bautz-Morgan contrast type. If these two corrections are not made, the effect on the present data is to significantly increase the positive curvature of the Hubble diagram, as well as the dispersion in apparent magnitude. To within the limits of the present data, the absolute magnitude of the brightest cluster galaxy does not, in general, depend on whether it is also a bright radio source. The present sample, cut off at z = 0.4 to avoid selection effects and uncertainties in the data, shows the first significant evidence for curvature of the Hubble diagram, with V and R magnitudes giving similar results. The formal value of q0 (with galaxy evolution ignored) is + 1.6 ± 0.4. The dispersion in absolute magnitude is less than 0.3 mag, with Mv = -23.28 ± 0.03 and MR= -24.09 ± 0.03. The use of the Hubble diagram in cosmology now depends on a knowledge of brightness changes in galaxies, on the one hand, or of q0 from other evidence, on the other. For example, if it were known with certainty that there has been no significant change in elliptical galaxy luminosities during the last 4 x 109 years, then the present data are nearly good enough for one to say definitively that the universe is closed and finite, with a finite lifetime. At the other extreme, if it were known with certainty from other evidence that the universe was nearly empty (q0 ::::: 0), then the present data set the constraint that net galaxy luminosities have decreased by .-0.5 mag during the last 5 x 109 years, with a net color change A(B V) ~ 0.1 mag. It seems possible at present to construct a self-consistent model with q0 ::::: 0 that satisfies the known data, but the case is not yet settled. Subject headings: cosmologygalaxies: clusters of galaxies: redshifts