1. INTRODUCTION

A normal galaxy is one that derives its luminosity primarily from nuclear burning in stars, and is at neither the high nor the low extreme of the luminosity distribution. In such galaxies, the visible and ultraviolet is the only direct window onto the photospheres of stars, whereas the rest of the spectrum reflects reprocessed light from stars (Figure 1). Dust in the interstellar medium reprocesses part of the stellar luminosity into infrared emission from a few µm to 1 mm or longer wavelengths. The radio emission at millimeter wavelengths derives from thermal emission from ionized plasmas in HII regions, whereas the synchrotron emission from cosmic ray electrons (CR e^-) trapped in the magnetic field of the galaxy fill in the cm-wavelength spectrum. In the X-rays, the luminosity is dominated by very hot plasmas at T 10⁶ K created by shocks from supernova explosions. The global spectrum of normal galaxies has some very stable signatures, and other quite variable aspects. Much of this article is an attempt to understand what makes for stability in the spectrum, and what significance to attach to the variable parameters.

Figure 1. The full spectral energy distribution for three representative star-forming galaxies. Note in particular the variation in the relative importance of the infrared and the visible-ultraviolet bands. Spectra are normalized to the same ordinate at 60 µm.

Astration, the cycle of star formation, nuclear burning and ISM enrichment in heavy elements, is the primary secular change and main evolutionary process affecting the chemical make-up and energy balance of the Universe. Because the luminosity function (e.g. Kim & Sanders 1998) falls fast enough that the integral over the population is dominated by the low end of the luminosity interval, over 90% of star formation in the local Universe takes place in normal disk galaxies, rather than the spectacular extremes. Thus, normal galaxies essentially cause and host Cosmic nuclear evolution.

Normal galaxies form a unique bridge between our detailed understanding of the Milky Way and our global understanding of the Universe and its history. Empirically, the integrated emission from galaxies must be compatible with the properties of Galactic objects. Our physical models of the local energy balance and of large-scale processes in the ISM must be consistent with the observed global behavior. On the other hand, galaxy luminosity and spectral properties are a crucial ingredient to modeling the observed faint extragalactic source counts and infrared background. Moreover, modeling the counts and background requires more than a census of the passive population (cf. Puget's article in this volume), and our models of galaxy behavior must be consistent with population evolution required to explain deep counts. The enrichment in individual galaxies must also add up to the observed current-day abundances. Thus achieving an accurate representation of normal galaxies would verify directly the validity of underlying models of the Milky Way ISM, and is a necessary step to the interpretation of cosmological results.

The InfraRed Astronomy Satellite (IRAS; Beichman et al. 1986) ushered in the era of space-based infrared astronomy in a dramatic fashion, revealing a stunningly rich infrared sky, unanticipated from the bits of infrared data previously and heroically collected from the ground. IRAS conducted the first unbiased all-sky survey at 12, 25, 60 and 100 µm, as well as a spectroscopic survey with a resolving power of 20 between 7.5 and 23 µm, for objects brighter than 10 Jy or so. IRAS had a profound influence on astronomy in general, not just the infrared, because it represented such a very large gain in sensitivity and spatial coverage, comparable perhaps to going from attempting visual astronomy in daylight to observing in a dark night (Beichman 1987; Soifer, Houck & Neugebauer 1987). Another significant factor in this influence was the dissemination of data products from IRAS, including source catalogs, image atlases and sky brightness estimates, generated with great care and well characterized and documented as to reliability, completeness, flux accuracy, and other statistical attributes.

The Infrared Space Observatory (ISO, Kessler et al. 1996) took infrared astronomy in space to new levels of sensitivity and into a new realm of spectroscopy across the whole range of 3 to 200 µm. Its versatile instrumentation took to the rich and fascinating extragalactic infrared sky first revealed by IRAS, and moved our knowledge of that sky an order of magnitude further and deeper.

In what follows, methods traditionally used in the study of normal galaxies are described, and outstanding questions currently pursued in the field are stated. The most prominent results from the IRAS survey are reviewed in Section 3. Contributions by ISO in the field of broad-band photometry are then presented (Section 4), followed by results in spectrospcopy (Section 5). Normal galaxy studies not directly concerned with the ISM are reviewed in Section 6. The outlook and challenges in pursuing the interpretation of infrared data on the ISM are discussed in Section 7.