Detection and classification of faint images by eye has traditionally encountered systematic errors faintwards of 20th mag on Schmidt plates and 22nd mag on 4-meter plates. Automated classification of Schmidt plate images has pushed the classification limit to 22 mag (Kibblewhite, et al., 1975). Automated detection and classification of faint 4-meter limit plate images has recently led to statistical studies of galaxy numbers and clustering at redshifts where cosmology and galactic evolution dominate over local effects. Here we report on some aspects of the FOCAS (Faint Object Classification and Analysis System) automated classifier (Tyson and Jarvis, 1979) and compare our results of number counts in SA57 with those of Kron, 1979. Differential galaxy counts in six high latitude fields and evidence for galaxy evolution are briefly discussed.

The color distribution of faint galaxies is an observational dimension which has not yet been fully exploited, despite the important constraints obtainable for galaxy evolution and cosmology. Number-magnitude counts alone contain very diluted information about the state of things because galaxies from a wide range in redshift contribute to the counts at each magnitude. The most-frequently-seen type of galaxy depends on the luminosity function and the relative proportions of galaxies of different spectral classes. The addition of color as a measured quantity can thus considerably sharpen the interpretation of galaxy counts since the apparent color depends on the redshift and rest-frame spectrum. To a first approximation two colors for a galaxy can determine a redshift and a spectral class, because redshift loci in a color-color diagram run roughly parallel to each other, and roughly perpendicular to the zero-redshift galaxy “main sequence” (Tinsley 1977a, Pence 1976) for small redshift. This game becomes more and more uncertain at higher redshift, because the systematics of galaxy UV spectral energy distributions are not well known, and what is known is not well understood (Code and Welch 1979). Redshifts for some random sample of faint galaxies is required to pin down the color-redshift relations; steps in this direction have already been taken by E. Turner. The reason for stressing colors, as opposed to redshifts, is that colors can be obtained relatively easily for large samples of faint galaxies if panoramic detectors are used: indeed, colors are not much more difficult to obtain than magnitudes.

Results are presented of counts of galaxies over the apparent magnitude range 21 ≤ B ≤ 25, which have been obtained at the 6–meter telescope using McMullan's electronographic camera. These data do not indicate evidence for rapid luminosity evolution of galaxies. It is noted that the faintest galaxies, on the whole, have bluish color and a small correlation function amplitude.

The combination of deep photography using fine-grain emulsions and fast automated plate-measuring machines is proving to be a valuable tool in studying galaxy evolution. Until recently, the favoured method for monitoring evolution was the spectroscopic study of one type of galaxy at various redshifts (see Spinrad 1977 for a review). It is considerably more economical, however, to derive evolutionary information from the statistical properties of the numerous faint images detectable on deep plates. The drawbacks are that the data at each apparent magnitude involves galaxies of different types seen over large redshift ranges. Also to interpret the statistics, we need to know the properties of large numbers of galaxies, for example their luminosity function, ultraviolet spectra and, particularly, the morphological variations in such properties.

An investigation of the clustering of galaxies making use of correlation function techniques is discussed. Correlation functions are defined and the relations between the angular and spatial functions are presented. The results obtained from application of the method to three samples of galaxies are described and some conclusions are drawn. Finally, some directions for future research are suggested.

I discuss four optical methods to locate standard candles (giant E galaxies), which should eventually lead us to reasonably large samples at high z. Redshift criteria and determinations are briefly discussed. The observed faint galaxy colors fit well into a simple model of stellar (galactic) evolution for systems with star-formation confined to a short time interval. Finally a summary of the statistics of blue galaxies in Coma-like clusters (Butcher-Oemler effect) is presented and interpreted.

New data on the colors of galaxies in distant, rich clusters of galaxies are presented. Of seven clusters examined having z<0.20, all exhibit color distributions indicating that a significant proportion of the member galaxies are undergoing star formation. The one cluster studied with z<0.20 clearly shows a reduced fraction of such galaxies.

Most models of galactic evolution include a high luminosity epoch of bright star formation at a redshift 2 < Z < 100. Past searches have failed to detect galaxies in their “primeval” state, but recent advances in detector technology will soon enable much improved searches that will very seriously constrain the evolutionary models. Using the CCD detectors on the Space Telescope offers the opportunity for a thousandfold improvement over present search limits. We review here several models of primeval galaxies and comment on their observability in present and future experiments.

Relatively little is known about superclusters. Observations of superclusters should increase our understanding of these grandest aggregates of matter and may tell us a great deal about the large scale distribution of matter in the Universe. We report here observations of two superclusters chosen from the list by Murray et al. (1978, Ap. J. [Letters], 219, L89).

In a search for very distant clusters of galaxies on a redsensitive (6° × 6°) Schmidt plate taken at Siding Spring in the region of the south galactic pole, we found some 49 very faint and compact clusters. We tried to reduce the number of chance configurations, by limiting our survey to (a) very compact and isolated groups of galaxylike images outside crowded regions; (b) some degree of concentration could be observed towards the brightest galaxy; (c) the size of the central object exceeded by several times the seeing disk diameter; (d) a minimum of 15 galaxies could be counted within a circle of ~ 30 arcsec in diameter. The red magnitudes of the brightest galaxies in each clusters are estimated to be fainter than 19.0 magnitudes. From the Hubble diagram for red magnitudes we estimate the redshifts to be about z ≥ 0.5. The integrated red magnitude of the first ranked galaxies in the faintest clusters are about 21.5. Taking into account seeing effects, a more realistic formula for the surface brightness at the centers of galaxies (rather than the [1 + z]−4) is derived. It is shown that, assuming no evolution, the limit of our survey corresponding to the faintest clusters is about z ≥ 1.0. A list of clusters coordinates, magnitudes of the first ranked galaxies and finding charts can be obtained from the author.

Digital SIT spectra have been obtained in collaboration with J.E. Gunn and W.L.W. Sargent for a sample of 49 field galaxies chosen from Kron's survey of selected area 68. Most have apparent visual magnitudes in the range 19 to 21. Redshifts have been measured for 48 of the 49 galaxies. The redshift (or, equivalently, luminosity) distribution for the sample is anomalous in that ≥ 10% of the sample has z ≥ 0.3 (and, thus, L ≥ 6 L*). These high z and L galaxies are among the bluest objects in the sample. These data require substantial evolution of at least some galaxies at small z. It is tempting to speculate that this luminous, blue field galaxy population is related to the Butcher and Oemler blue cluster galaxies found in the same redshift range.

We discuss the distribution in depth of different kinds of quasars: quasi-stellar radio sources with steep radio spectrum, those with flat radio spectrum, and optically selected quasars. All exhibit an increase of space density with distance to a different degree. The optically selected quasars, in particular, show a steep increase of surface density with magnitude. The steepness of the increase is inconsistent with a uniform distribution of quasars in the local hypothesis. In the cosmological hypothesis the co-moving space density of optically selected quasars increases by a factor of 100,000 to a redshift of 2, and by factors of 1000 and 10 for steep-spectrum and flat-spectrum radio quasars, respectively.

Recent results with the slitless spectrum technique are reviewed and the total surface densities from five independent surveys are compared. The observational base for extending the technique to search for quasars with z > 3.5 is now large enough that either the apparent cutoff at z = 3.5 will be confirmed in new surveys or quasars with z > 3.5 will be found.

It is generally assumed that quasi-stellar objects represent phenomena taking place in galaxies. There are at least three lines of reasoning that lead to this viewpoint. First, there appears to be a continuity of properties extending from Seyfert galaxies and N systems to QSOs with the differences between the various groups principally a question of the contrast between the luminous central object and its surrounding galaxy. Since Seyfert galaxies and N systems are definitely in galaxies, it is concluded that QSOs must also be in galaxies which are not directly visible because of the high luminosity of the central object. A second argument is based on the result that QSOs do not appear to differ dramatically in abundances of elements from those which are typical of normal galaxies. With the prevailing view that virtually all the elements heavier than helium originated as a result of stellar processes, it would be concluded that QSOs must be associated with galaxies of stars that produced the heavier elements. A third, rather indirect argument is based on the results of Stockton (see his paper in this volume) and others that QSOs are often found in groups of galaxies and therefore are likely to be located in galaxies themselves (guilt by association). But the fact is that direct evidence that QSOs are in galaxies is sadly lacking.

Evidence that most presently known low-redshift QSOs are associated with groups of galaxies has led to a program aimed at determining the typical properties of these groups. For the 3C273 field, observations of essentially all galaxies with mB ≤ 21.5 over a region 10.1′ (∼1.5 Mpc) in diameter yield four galaxies having redshifts close to that of the QSO. Consideration of the present evidence concerning the richness of groups associated with QSOs leads to the conclusion that, for at least those QSOs that are radio sources, their galactic environment is similar to that of radio galaxies, except that they seldom, if ever, occur in the centers of rich clusters of galaxies.

All of the speakers at this meeting except me have assumed, or will assume, without question that redshifts are measures of distance. This reflects a point of view which is not borne out by all of the evidence. I shall attempt to give you some of the flavor of the evidence which so many wish to ignore. It comes from individuals who perhaps not surprisingly, have not been invited to speak.

A classification scheme for QSO absorption line spectra is described which ascribes the origin of the lines to at least four mechanisms: (A) Explosive ejection of material at speeds up to 0.1 c. (B) Absorption by highly ionized material moving in a rich cluster in which the QSO is embedded. (C-1) Cosmologically distant intervening material with ‘normal’ abundances, probably associated with large galactic halos. (C-2) Cosmologically distant intervening material consisting of primordial uncondensed gas. Examples of each type of spectra are given and their ionization and other spectral characteristics discussed. The similarity between the development of novae spectra and a possible evolutionary sequence of the explosive ejecta of type A is striking and suggestive. Several difficulties and unsolved problems involving this scheme are noted. Finally, we speculate on the interpretation of two interesting objects (PKS 0237-23 and the ‘twin quasars’ 0957+56A,B) in the context of this scheme.