Sampling a galaxy's interstellar medium directly provides a snapshot of the current abundance picture at the location being tested, in contrast to stellar abundances which for the most part are indicative of interstellar abundances at the time that the star formed. The most straightforward way of determining interstellar abundances is through the analysis of emission spectra produced by gas heated by nearby hot stars with continua rich in photons having wavelengths shortward of 912 Å, i.e., the ionization edge of hydrogen. Such stars have effective temperatures exceeding 30,000 K and spectroscopically belong to the O and early-B classes. Object types with these conditions include H II regions and planetary nebulae, reviews of which can be found in Shields (1990), Vila-Costas & Edmunds (1992), and Zaritsky, Kennicutt, & Huchra (1994) for H II regions and Peimbert (1990), Henry (1990), Perinotto (1991), Clegg (1993), and Habing & Lamers (1997) for planetary nebulae. Old supernova remnants in which the stellar ejecta have completely mixed with the interstellar medium in principle represent a third type of emission-line probe, since in this case the interstellar gas is heated by the shock wave producing emission lines. Often, however, full abundance studies are precluded by limited spectral coverage even within the optical (W. P. Blair 1999, private communication), and thus there are far fewer abundance results available.
Ionized gases of the types just mentioned maintain temperature equilibrium in most cases by radiating photons at discrete wavelengths following recombination or collisional excitation processes involving ion-electron encounters. Spectra of these objects can then be analyzed to provide abundance, temperature, and density information. Measured from the ground in most cases, the strengths of the resulting emission lines can be converted to ionic and elemental abundances of He, C, N, O, Ne, S, and Ar using the techniques described in Appendix A, which includes a table listing a number of prominent emission features. While the resulting abundances refer to levels in the gas phase only, Savage & Sembach (1996) indicate that none of these elements is expected to be heavily partitioned into the solid phase in the form of dust. Thus, gas-phase abundances should represent total values reasonably well.
Due to their size and therefore their accessibility in external galaxies, most of the abundance data from emission-line systems useful in chemical evolution studies relate to H II regions, which because of their association with recent star formation are located in spiral disks and irregular galaxies. This section focuses on abundance patterns in spirals.
It should be noted that abundances discussed are taken directly from the sources listed; no attempt has been made to homogenize them by recalculating the abundances in a consistent way. In general, differences in techniques and atomic data employed produce ranges in abundances which are smaller than observational uncertainties in the line strengths, and we believe that presenting unhomogenized data still provides a realistic representation of patterns and an opportunity to see the big picture.