How did homogeneously distributed matter in the early universe condense into galaxies? To address this question empirically we must observe the galaxy formation epoch, then trace the development of large scale structure in the distribution of galaxies as a function of redshift. We must understand how individual galaxies relate to the processes that explain these structures, and on a more local scale, we must fit the details of our own Galaxy into this overview.

How do cosmic backgrounds relate to the large scale structure of the universe?

Results from the COBE mission will provide an initial opportunity to look for evidence of primordial galaxy formation. The COBE all sky maps from 1 µm to 1 cm will be the first step toward a determination of the true cosmological background radiation. The best windows on the extragalactic universe are at 3 µm to 4 µm, between the scattered and re-radiated components of the zodiacal dust, and at 200 µm to 500 µm, between the thermal emission from Galactic dust and the cosmic microwave background. These windows are the most promising in the entire electromagnetic spectrum for the detection of radiation from primordial galaxy formation. Such radiation can provide cosmological information even in the absence of observations of individual primordial galaxies. Extragalactic infrared backgrounds can set important limits on the epoch and nature of galaxy formation as long as known galaxies at lower redshifts are accurately subtracted. This requires accurate knowledge of all populations contributing to the backgrounds. These populations will be accounted for by the infrared surveys of the 1990's. Folding in the deep source counts of near-IR and mid-IR observations, combined with estimates of their contributions in the far-IR should allow us to tell unambiguously if infrared background radiations arise from the glow of the universe at the time galaxies first formed.

Infrared sky surveys in the 1990's will be essential to probe the distribution of matter in the nearby universe. The near infrared is the optimum region to obtain a mass census of galaxies because infrared radiation is both insensitive to the extinction within galaxies (including our own) and sensitive to the stellar component which dominates the luminous mass. The "great attractor", for example, is in a direction greatly affected by obscuration from the Milky Way Galaxy. Any mass concentration there may be detectable only in the infrared. Infrared observations will be the only way to detect cool, solid objects in galaxy halos. Systematic observations of galaxies with old stellar populations and the deepest possible searches for faint matter are essential to understand the mass distribution in large scale structures.

When did galaxies form?

Recent observations in the optical have established that the epoch of initial nucleosynthesis, star formation, and galaxy formation is at a sufficiently high redshift that the initial formation epoch can only be observed in the infrared, because the redshift must be greater than 5. Pushing back the limits on this formation epoch, or finding it, is a major imperative for observational extragalactic astronomy.

A quasar is now known at a redshift of nearly 5. If a significant population of quasars exists much beyond this, such a population cannot be measured without observations in the near infrared. Although quasars currently represent the highest redshift objects observed, their relation to primordial galaxy formation is unknown because their ultimate luminosity sources seem unrelated to stars. Direct evidence, or limits on the existence of galaxies made of stars, is also an observational requirement before the formation of stars and galaxies in the early universe can be understood. Determination of a redshift based on the bluest line, Lyman alpha, requires infrared spectroscopy for redshifts greater than 7, while observations of the old stellar population require measurements at wavelengths of about (l+z) µm.

Such observations of old stellar populations in distant galaxies will set limits on the epoch of star formation in the universe, but direct observation of the first generation of star formation is the ultimate goal. Depending on the redshift of the formation epoch, near to mid-infrared observations will be required to detect the intrinsic ultraviolet luminosity from hot, young stars. IR observations have already demonstrated that extensive star formation in local galaxies is invariably accompanied by dust absorption at UV and optical wavelengths and re-emission in the IR. Is this the case also at high redshift epochs? Will HST observations of young galaxies at great distances be affected by dust attenuating the ultraviolet continuum?