|Annu. Rev. Astron. Astrophys. 1984. 22:
Copyright © 1984 by . All rights reserved
3.1 Current Global Star Formation Rates
Measures of Lyman continuum photon luminosities, e.g. as determined from the H or H emission lines or radio thermal fluxes, can be used to estimate the current rate of formation of massive stars. When coupled with an IMF such as from Salpeter (289) or Miller & Scab (245), this information can yield the total rate at which gas is condensing into stars of all masses (143a, 319). If the emission flux is that for the entire galaxy, then the global star formation rate is known. Galaxy-wide star formation rates for Irr galaxies have been estimated in this fashion by extrapolating from large-aperture spectrophotometry (191), from flux-calibrated H images (126), from H photometry (205, 206), from radio continuum observations (195), and from UV luminosities (29, 86). For one galaxy, NGC 1569, the number of Lyman continuum photons determined from the radio observations is about 3 times higher than the number found from H emission. This is probably due to extinction within HII complexes, and thus optical measurements will usually yield underestimates of star formation rates (205).
The star formation rates determined in this manner are found to cover a wide range, but the average rate per unit area in high surface brightness Irrs is comparable to the Milky Way disk (~ 5 x 10-9 M yr-1 pc-2; 330, 353, 364). If we consider the total rate of star formation per unit gas mass, the average high surface brightness Irr is again comparable to a typical spiral, although some Irrs are overachievers in this regard. (NGC 1569, for example, has a rate 30 times that of our Galaxy; 390.) It is important to keep in mind that the samples chosen for star formation rate studies have been intentionally biased toward high surface brightness, rapidly star-forming galaxies, and that there do exist low surface brightness dwarf Irrs that have very little current star formation activity (e.g. 57, 189, 288). Nevertheless, these studies show that Irrs exist that are uninfluenced by interactions or other outside perturbations and yet are actively forming stars. We must conclude, therefore, that spiral density waves are not necessary to a vigorous production of stars.
In order to understand the star formation mechanisms and the galactic characteristics that govern them, a search has been made for correlations between star formation rates and various global parameters. One might expect, for example, a correspondence between stellar birthrate and gas density such as in the Schmidt (308) empirical model in which star formation rate varies as the gas density squared. The higher the mean volume density in the interstellar medium, the higher the star formation rate would be. In practice it is difficult to ascertain from the observed projected density of interstellar matter the fraction that is in a proper physical state to produce stars. Furthermore, it is both the fluctuations and mean gas density that are important in producing stars, so we may expect some problems with this type of approach. From a study of spiral and Irr galaxies, Lequeux (222) concluded that the star formation rate actually decreases with increasing average gas density, while Guibert (142) found that if the star formation rate is proportional to a power of the gas density, the proportionality constant decreases with increasing gas fraction. Young & Scoville (391), on the other hand, suggest that in spirals the production rate of stars per gas nucleon is constant. In their study of Irrs, Hunter et al. (191) found no relationship between star formation rates and average HI gas density or gas mass, although Donas & Deharveng (86) do find a correlation with gas mass. Low CO fluxes from actively star-forming Irrs (95, 390) further indicate that no simple relationship exists between measurable global gas characteristics and current stellar production rates.
A few studies (86, 191) have also searched for relationships between star formation rates and other integral galactic quantities such as metallicity, total mass, the ratio of mass in stars to dynamical mass, and the HI mass per unit luminosity. No convincing correlations have been found. From this it is concluded that local parameters are probably more important than global ones in determining star formation patterns in noninteracting Irrs.
The manner in which these local processes interact, then, must set the global states of Irrs. Unfortunately, the mechanisms that provide coupling between star formation sites are not known and may not include the traditional local moderators of star formation processes: gravitational and magnetic instabilities. Galactic-scale magnetic fields probably are present in Irrs, as evidenced by organized interstellar polarization in the Magellanic Clouds (198, 309), and potentially play an important, although as yet poorly defined, role in large-scale star formation processes (96, 250, 251). Gravitational instabilities against axisymmetric perturbations have also been proposed as drivers of global star formation in disk galaxies (see 143, 255). Both gas and stars in Irrs, however, have typical velocity dispersions of ~ 10 km s-1 (105, 119, 182, 183) and therefore by the usual criteria (356, 357) are safely stable unless the dispersions are highly anisotropic. The most probable mode of interaction in star-forming processes within Irrs is thus through modifications of conditions in the interstellar medium, which can be induced by the young stars themselves.