|Annu. Rev. Astron. Astrophys. 1992. 30:
Copyright © 1992 by . All rights reserved
Attempts to model the changes in the integrated properties of evolving stellar populations - their luminosities and intrinsic colors - have been made by Tinsley (1972), Bruzual (1983), Arimoto and Yoshii (1986), Rocca-Volmerange and Guiderdoni (1988), and subsequent papers by the same authors based on similar premises. The main adjustable ingredients are the shape of the initial stellar mass function and the time history of the rate of star formation. The characteristics of late stages of stellar evolution are in practice also adjustable, but are more explicitly subject to known astrophysical constraints. Some models have included effects of chemical evolution or internal extinction by dust (Wang 1991).
Such models are powerful tools for the interpretation of data, as they allow one to explore quantitatively some of the possible behavior of high-redshift galaxies. However, since the initial mass function and the time history of the rate of star formation are still poorly constrained by independent data or by astrophysical arguments, none of the models should be taken too seriously in terms of predictive power. The shape of the initial stellar mass function is conventionally taken to be that of the solar neighborhood. Although such an assumption does not lead to immediate contradictions when applied to field galaxy samples, its adoption is by default. Similarly, the history of the star formation rate is modelled as some function of the amount of gas available to form new stars; this is not unreasonable, and it is computationally convenient, but none of the formulations has a compelling physical basis.
The amount of blue light (or far-infrared flux) from nearby galaxies is a measure of the young population, but the rate of star formation in the same galaxies billions of years earlier cannot be reliably determined by extrapolation in time. The metallicity, mass-to-light ratio, and gas-mass to total-mass ratio at the present epoch do impose important boundary conditions on the past history of star formation, but the prevailing large uncertainty in galaxy ages illustrates that these boundary conditions do not provide stringent constraints, nor lead to a unique solution for the past history of star formation.
Since one can adjust the redshift of formation and other critical parameters relating to galaxy ages, and since there is freedom to make adjustments in the mix of spectral energy distributions so long as the present-epoch color distribution is reproduced, a rather extensive palette is available to the modeller. One can learn for example which of the model input parameters is likely to be of greatest importance, but the significance of a model fit is often difficult to establish.