5. SUMMARY AND SUGGESTIONS

We have explored in some detail the abundance patterns in spiral disks and elliptical galaxies as revealed through analyses of gaseous nebulae, stars, and integrated photometry of galaxies.

The principal points regarding abundance patterns in spiral disks are the following:

1. The metallicity as gauged by O/H in nebulae across the Milky Way disk decreases with Galactocentric distance, a finding supported by recent abundance results for disk stars. This negative gradient pattern is seen in most other spiral disks. A similar result is seen when luminous stars are used as abundance probes. Scatter at any particular galactocentric distance is consistent with observational uncertainty.

2. Global metallicity, taken as the abundance of oxygen at a standard galactocentric distance, is positively correlated with galaxy mass.

3. Metallicity at any location in a spiral disk appears to be positively correlated with the local total surface density.

4. Abundance gradients are steeper in normal spirals than in barred ones.

5. A plot of N/O versus O/H in spiral disks indicates that the production of nitrogen is dominated by primary processes at low metallicities and secondary processes at high metallicities.

6. C/O is positively correlated with O/H in spiral disks, suggesting that carbon production is sensitive to metallicity, possibly through metallicity-enhanced mass loss in massive stars.

7. Abundance ratios of Ne/O, S/O, and Ar/O appear to be universally constant across the range in metallicities observed, reflecting the idea either that the IMF is universally constant or that the stellar mass range responsible for producing these elements is relatively narrow, and thus these ratios are insensitive to IMF variations.

8. Stellar age and Galactocentric distance in the Milky Way show rough correlations with metallicity in the sense that metallicity decreases with increasing age and Galactocentric distance. However, all Galactic components (halo, bulge, thin disk, thick disk) have large scatter in abundance, and even the metal-poor halo is now thought to display an age scatter of several Gyr.

1. Abundance gradients are, on average, about a factor of 2-3 more shallow than in nonbarred spirals. This is well within the range expected from various formation pictures, including hierarchical mergers of smaller galaxies.

2. Nuclear or global metallic feature strengths (or colors) become stronger (or redder) in larger galaxies. The 1970s conclusion that larger elliptical galaxies must be more metal rich is reconfirmed, but every elemental species does not increase in lockstep.

3. Light elements N, Na, and Mg are enhanced relative to heavy elements Ca and Fe in the largest elliptical galaxies, implying a modulation of enrichment, plausibly due to variance of the Type II to Type Ia supernova ejecta, compared with smaller ellipticals, bulges, and disks.

4. The mean abundance near the nuclei of large elliptical and S0 galaxies is uncertain but is in the range [Z/H] = 0.0-0.4. Most of the difference in abundance between small and large galaxies is driven by the increasing abundance of elements lighter than those near the Fe peak, with [Fe/H] staying roughly constant for elliptical galaxies of all sizes.

5. The abundance distribution in elliptical galaxies, and, so far, every other well-studied large galaxy type, is strongly peaked like that of the solar cylinder, not broad like the simplest closed-box model predicts.

There are two overarching patterns which emerge from the combined results for spirals and ellipticals. First, there is a positive correlation between galactic metallicity and mass. This may be related to the greater retension of heavy elements ejected by supernovae by the stronger gravitational potentials of massive galaxies, or perhaps to the effects of galaxy mass on the star formation process. It is currently difficult to ascertain whether this relation is completely continuous across galaxy types; in other words, if one plotted global metallicity versus mass for a sample of galaxies containing both spirals and ellipticals would there be an unbroken straight line, or would the correlation for one type be offset from the other. The difficulty here is in directly comparing abundances between the two galaxy types. As we have seen, metallicity in spirals is generally gauged by observing oxygen in nebulae located in their disks. Yet in ellipticals, metallicity must be measured from integrated light using numerous photometric indices which are affected not only by metallicity but by age. Thus, no seamless technique exists for determining abundances consistently for spirals and ellipticals, and thus intercomparisons are problematic. This is made all the more complicated by the fact that we currently do not know how to represent the global abundance in a galaxy. Do we take the abundance at the nucleus, or at one effective radius, or at 0.4 optical radii?

The second pattern which has emerged is that abundance gradients appear to become flatter as one progresses from normal spirals to barred spirals to ellipticals. The difference between normal and barred spirals is currently explained by enhanced radial gas flows in the disks of barred spirals. To extend this model to ellipticals, it may be neccessary to invoke other radial mixing mechanisms, either during primordial formation or during later merging events. If the pattern is discontinuous between spirals and ellipticals, this might suggest that different processes operate to affect the gradients in the two galaxy types. Again, our lack of ability to intercompare spiral and elliptical abundances prevents further exploration of this pattern at present.

Understanding the broad picture of galactic chemical evolution will require us to firm up the links between elliptical and spiral galaxy abundances. While a common elemental yardstick may not exist because of the different elements which we observe directly in each galaxy type, it may be possible to tie the two types together abundancewise by observing elements in each which share the same production site nucleosynthetically speaking. An example might be oxygen and magnesium. In external spirals oxygen is taken as the metallicity gauge primarily because of its observability. Magnesium, which, like oxygen, is primarily produced in massive stars (Nomoto et al. 1997a, 1997b) may be measurable directly through a calibrated Mg₂ index. Then oxygen and magnesium might be linked by assuming a "cosmic" Mg/O ratio calibrated locally. Also, although the work is not started, it may be possible to construct an oxygen-sensitive photometric index for integrated light, perhaps revolving around the 2.3 µm CO feature in conjunction with the C₂-sensitive 4668 feature. As synthetic spectra and stellar abundances grow more precise, these speculative suggestions might take place, leading to a much more clear understanding of chemical enrichment and galaxy formation.

This collaboration was inspired by the 1997 October workshop "Abundance Profiles: Diagnostic Tools for Galaxy History" held at Université Laval, Québec. We are grateful to the organizers for giving all of us the opportunity to come together and share our ideas about what one participant, in an attempt to relabel the abundance field with a trendier and more attention-grabbing name, referred to as "bio-resources." We also thank Bill Blair, César Esteban, Mike Fich, George Jacoby, Yuri Izotov, Joachim Köppen, Walter Maciel, Dörte Mehlert, Anne Sansom, Jan Simpson, and Friedl Thielemann for promptly responding to inquiries with useful answers and information given generously. And finally, we are especially grateful to our referees, Dave Burstein, Karen Kwitter, and Bernard Pagel, for promptly and carefully reading the manuscript and making numerous constructive comments which have improved the paper tremendously.