Galactic evolution: Do globular clusters belong?

Abstract

Do globular clusters belong? Virginia Trimble Do the globular star clusters really belong to the galaxies they now inhabit, in the sense of having formed in and with them? Yes, according to some of the results presented at a recent symposium*, and no, according to others. The issue arises because these clusters are a bit like dinosaur teeth -small hard things that could have survived a long time under harsh conditions. Thus, they might, alternatively, have formed separately before their host galaxies or as part of other systems that amalgamated to make those galaxies. The issue is important because the globular clusters, unlike dinosaur teeth, are not just very old, but are actually the oldest subjects in the Universe available for study at close range. They ought, therefore, to tell us something of when and how galaxies and stars began. Globular clusters have so far been seen in about 40 galaxies (spirals, ellipticals and irregulars) and studied closely in half a dozen. The 125-or-so clusters known in our own Milky Way contain, for instance, about 1 per cent of the stars of the spherical halo. Evidently, indicators are needed to show whether the clusters resemble the stars around them more than they resemble each other, and whether their properties (individually or as populations) are correlated with those of their host galaxies. Resemblance and correlation would suggest formation in situ, and the opposite would suggest a more complex process. The evidence, unfortunately, is contradictory and rather evenly balanced. Starting with our own Galaxy, the globular clusters (according to R. Zinn, Yale University) do generally resemble the other halo stars in their distribution of numbers as a function of radius , in their average heavy-element abundance and its gradient with galactocentric distance, and in their dynamical properties (net rotation , velocity dispersions in three dimensions and average orbital eccentricity). Differences exist , however, in the periods of variable stars as a function of their composition (G. Wallerstein , University of Washington) and in the metallicity distribution, in the sense that the field stars have a larger representation at very high and very low values than do the clusters (J .B. Laird , University of North Carolina) . Other indicators that the clusters 'belong' to their hosts are: the likelihood that galaxies more massive than the Milky Way, including M31, NGC5128 and M87 , have redder, more metal-rich clusters (F. Fusi Pecci, University of Bologna; H.C. Harris, US Naval Observatory; J. Huchra, Center for Astrophysics, Cambridge, Massachusetts), though the high value for M87 was disputed by J . Nemec (University of British Columbia); the two- standard-deviation result that central (massive, cD) galaxies in clusters have more than their fair share of globulars (W.E. Harris, McMaster University); and the undisputed fact that gas-rich galaxies, including M33 and the Magellanic Clouds, unlike the Milky Way and similar spirals, are still producing massive clusters (though not quite such massive ones as the true, old globulars; E. Olszewski, Steward Observatory; C. Christian, Canada- France-Hawaii Telescope). In contrast, several other observations suggest that the clusters did not originate with the galaxies they now inhabit. First, the number of clusters at a given luminosity, the function N (L), seems to be much the same, a gaussian with the same width and a peak at M, = - 7.4, in nearly all galaxies examined; at any rate, the assumption that this is so leads to a consistent distance scale out at least as far as the Coma cluster (D.A. Hanes and D.G.