Annu. Rev. Astron. Astrophys. 1984. 22: 37-74 Copyright © 1984 by Annual Reviews. All rights reserved

2.4 Stellar Populations

The blue colors of Irrs are generally taken to mean that a proportionally larger component of the stellar population consists of early-type stars (247, 248). This is consistent with the fact that the high surface brightness Irrs are endowed with numerous HII regions and are actively forming stars. Bagnuolo (15) and Huchra (178), for example, have fit the colors of Irrs with a composite between old and young stellar populations, while continuous star formation models have been computed by Searle et al. (313), by Huchra (178) and by Code & Welch (59).

Specifies of the OB star populations have been explored through IUE spectra that are now available for a variety of high surface brightness Irr galaxy family members (27, 28, 29, 179, 180, 219, 377; see also 47, 196). Generally the 1150-2000 Å UV spectra of Irrs show features consistent with rich OB star clusters having near normal IMFs, but there are important exceptions. On the basis of UV spectra, a case was made for a very massive superstar (~ 2 x 10³ M) in the core of the giant 30 Doradus HII complex in the LMC (52, 107, 305) and more recently for the somewhat similar NGC 604 giant HII region in the nearby spiral M33 (237). Even if we do not accept the presence of superstars in these systems (245a), it is clear that extraordinary concentrations of high-mass stars (~ 10² M) must be present to meet the ionization requirements and to fit the observed spectral characteristics. Giant HII regions are common in Irrs (169), and their possible relationship to very massive stars is now receiving careful scrutiny.

The low surface brightness dwarf Irrs present more of a problem with regard to stellar content, since they are also fairly blue but do not have the many obvious star-forming regions that characterize the high surface brightness systems. In general, the dwarfs seem to have stellar population mixes similar to those of the larger Irrs (81, 82, 160, 189), but the numbers of luminous stars are down in a manner qualitatively consistent with a lower total star formation rate (173, 185, 228, 295). Some dwarf systems have extremely blue colors and high surface brightnesses, perhaps indicating that bursts of star formation have recently been completed (158, 177, 178, 232, 313); detailed population studies of a few resolved galaxies support these viewpoints (57, 288).

The Irrs, therefore, are correctly noted for their young stellar population, but they also contain older stars. Only a few of the extreme ``intergalatic HII regions'' seem to be without a possible older stellar component (cf. 179, 312, 349, 377), but the nature of this older population and the number of previous generations of stars are not so clear. In the dwarf Irr VII Zw 403, for example, the metallicity of the gas is sufficiently low and the level of current star formation sufficiently vigorous that one is forced to doubt whether the galaxy could have formed stars in this vigorous way at an earlier stage (361). Nevertheless, the existence of a diffuse stellar component implies that an older population is in fact present. (Star formation histories are discussed in a later section.)

Many Irr galaxies are close enough that they can be resolved into individual stars. The LMC and SMC are, in fact, the best systems outside our own for studying stellar populations and support the concept of a nearly constant IMF in disk galaxies (see below). Beyond the MC, only intermediate- to high-mass stars can be individually observed at present, and aside from very luminous supergiants, only colors and magnitudes are available (187, and references therein). In more luminous Irr dwarfs, observed color-magnitude diagrams are similar in form, with pronounced blue and red supergiant branches well separated by a Hertzsprung gap (184, 185, 186, 202, 295, 296, 301). Massive stars evidently are present with relatively constant properties in Irrs, and thus the door is open to modeling the light from young stellar populations in terms of fairly standard components (as in 76, 180, 189, 191, 225).

Differences in OB stellar content between galaxies are largely explainable in terms of statistical effects, which can be quite severe in faint dwarfs, where OB star formation probably involves a series of time-disconnected discrete events (172, 288, 313). At the upper extremes of stellar mass, the situation is, as we have seen, less well defined; for example, the presence of many Wolf-Rayet stars in a small galaxy like Tololo 3 (216) might be due either to statistics or to special processes (e.g. very massive stars) in very large star formation events that are seen as giant and supergiant HII regions (169, 237). Finally, we point out that intermediate-mass stars become very luminous during the AGB evolutionary phase (see 193) and may be seen as resolvable stars in the diffuse light of Irrs (140), as long-period variables of interest to the extragalactic distance scale (387), and as major contributors to the infrared luminosity (272).

Star clusters provide further important clues to stellar populations in galaxies (e.g. 56), and currently they can be detected to distances of several Mpc in Irrs as a result of the open structures of the parent galaxies (155, 188). While numbers, sizes, and richnesses of star clusters vary from galaxy to galaxy (160, 167, 369), the clusters themselves are found to be remarkably similar in their integral optical stellar properties within such diverse Irr systems as the LMC (370a), M82 (257), and NGC 6822 (165, 372). The main variables affecting integral cluster observables are well known to be IMF, age, and chemical composition, although stellar richness can also be a significant factor (279). From color-magnitude diagrams of individual Magellanic Cloud star clusters it has been possible to calibrate approximately variations in global cluster parameters as a function of age for metallicity levels appropriate to most Irrs. Unfortunately, some disturbing inconsistencies remain in the details of the age scales (170, 249), and subtle differences exist between clusters and stellar evolution model predictions (26, 112, 113). Still, the analysis of LMC cluster photometry in the classic work of Searle et al. (314), as well as Rabin's (277) investigations of individual cluster spectra, assures us that among younger clusters age is the major determinant of spectral properties, while in very old clusters metallicity is a primary factor. Star clusters thus are a comparatively reliable means for unraveling the stellar age/metallicity strata that hold the histories of galaxies.

Studies of the Magellanic Clouds and other nearby Irrs reveal star clusters covering a full range of age classes, and thus these galaxies have been actively producing stars for at least several billion years (162, 167). Recently photometry has been obtained by Stryker (336, see also 338) down to the main sequence turnoff in the red LMC halo cluster NGC 2257. As this cluster contains a well-defined horizontal branch and RR Lyrae variables, the preferred age calibration method, developed by Rood & Iben (284; see Iben 192) and based on the distance-independent luminosity difference between main sequence turnoff and horizontal branch, can be applied to show that NGC 2257 is as old as Galactic globulars. Evidently the LMC produced or obtained star clusters from the same early epoch as the Milky Way. In contrast to the Milky Way, however, the Magellanic Clouds are still making globularlike star clusters, the ``populous blue clusters'' or ``blue globular clusters'' (156). Although there has been some resistance to considering these as total parallels to young globular clusters, LMC blue globular cluster masses lie in the range of 10⁴-10⁵ M and therefore overlap with true globulars (58, 114, 118, 148, 254). These clusters are not unique to the Magellanic Clouds, and the luminous, near-stellar knot seen in actively star-forming regions of Irrs such as NGC 1569 or NGC 5253 (2, 84, 188, 189, 367, 370) may be populous stellar clusters in early evolutionary phases when OB stars and circumstellar gas are still present. It is therefore not appropriate to attribute the production of globular star clusters only to unique conditions in the early Universe (e.g. 85), but rather there may be a variety of channels for the formation of dense, spheroidal star clusters.

Not all Irrs, however, display the same small-scale spatial patterns of star-forming activity. At one extreme, the low astration rate dwarfs often lack rich clusters and pronounced OB associations, even when massive stars are present (e.g. 288, 301, 369). At the other extreme, some rapidly star-forming amorphous Irrs are also quite smooth in their optical appearances (125, 294, 300) and are pervaded by diffuse optical emission lines from ionized gas (84, 157, 188). It is quite clear that these galaxies may contain large complements of massive young stars, as evidenced by their high H luminosities and hot IUE ultraviolet spectra (219). The optically distinct, large star-forming complexes (OB associations, HII regions, etc.), which are the hallmark of most Irrs, are, however, missing. Perhaps in these systems the individual star-forming sites are overlapping or the stars are forming via a different mechanism than in most Irrs. But in either case, the amorphous Irrs illustrate that kinematically similar galaxies do not necessarily follow identical evolutionary paths.