|Annu. Rev. Astron. Astrophys. 1984. 22:
Copyright © 1984 by . All rights reserved
3.4 The Initial Mass Function
The lack of understanding about the IMF (initial mass function) is a serious stumbling block in the study of galactic star formation histories (see 307 for a comprehensive review). The Magellanic Clouds are the only systems that are sufficiently near to check the IMF outside of the Galaxy, although as we noted earlier, massive star populations are relatively similar in most nearby Irrs. An inventory of Magellanic Cloud supergiants by Dennefeld & Tammann (69), for example, indicates that the stellar component of masses 9 M is not radically different from the Galaxy (see also 223, 224, 296). Deep luminosity functions measured in the LMC by Butcher (46) and Stryker & Butcher (337) extend this conclusion to ~ 1-3 M stars. The presence of classical Cepheids (269) and RR Lyrae variables (138, 139) in the Magellanic Clouds further shows that stars covering a range in mass from 1-10 M have evolved off the main sequence. Cepheids also have been observed in several other Local Group and nearby Irrs, but only in IC 1613 are the numbers sufficient to probe the stellar content (291), which is found to be like the Magellanic Clouds (25, 240). Thus the stellar mixes of the Magellanic Clouds are surprisingly similar to those of the disks of spiral Local Group members (144, 145), although there are hints from the luminosity functions that the star formation rates are not smooth functions of time in either the LMC or SMC (147, 337).
If the IMF is to be different in the Irrs, we require that high surface brightness systems with extreme blue colors and strong emission lines be overabundant in high-mass stars, while the low surface brightness Irrs have fewer massive stars. In making such comparisons, however, one must distinguish between changes in the form (e.g. slope) of the IMF and statistical effects, i.e. very large star-forming events naturally will produce many massive stars. The similarities between the large HII complexes in Irrs and those in other types of galaxies give no reason to expect that local environments are sufficiently diverse that a different IMF would result. Hence, we are reluctant to invoke an unusual IMF when there is no compelling evidence for it. Terlevich & Melnick (347, 348) have argued that there is a systematic variation of the IMF with metal abundance, such that more metal-poor systems have flatter IMF slopes, but their approach to this problem is still controversial (cf. 124). Some effect of this type may, however, be necessary to explain the correlation between ionization level and metallicity that exists in HII regions. Below, we discuss further evidence that a normal IMF is at least consistent with the evolutionary histories of typical Irrs. However, we do not know over what time or spatial scales it is necessary to average the stellar populations in order to obtain the ``normal'' IMF. It is possible that during galaxy-wide bursts of star formation, conditions may be such that this spatial- or temporal-averaging process is disrupted, and a peculiar IMF results.