Introduction

Determination of the physics behind the process of star formation, in addition to being of importance in its own right, is an integral part of the larger question of galaxy formation and evolution. Indeed, given that we use luminous galaxies as tracers of large-scale structure, even our understanding of cosmology ultimately depends on an understanding of what causes stars to form. Normal late-type disk galaxies are the sites of a significant fraction of the star formation in the local Universe and understanding the rate of star formation in these systems is of obvious significance. The effects of star formation on the interstellar medium (ISM) are also highly relevant to understanding galaxy formation and evolution.

The current paradigm for galaxy formation and evolution invokes hierarchical clustering, in that gravitational collapse occurs first on small scales, and subsequently these systems merge to form galaxies, with baryons being only around 10% of the total gravitating matter (e.g. Silk & Wyse 1993). The merging history of `galaxies' and the rate at which the baryons turn into stars must play some role in the origins of the Hubble Sequence. In particular, feedback from massive stars may well determine the low efficiency of star formation in disks compared to that inferred for elliptical galaxies, and the difference in angular momentum distributions of disks and of ellipticals. The angular momentum of galaxies is believed to be generated by tidal torques between density perturbations, prior to gravitational collapse and is essentially independent of overdensity, scale, environment etc. Thus one is required to explain the origins of both rotationally-supported disk galaxies and `pressure'-supported elliptical galaxies (the `pressure' is random stellar motions) from the same initial values of the angular momentum parameter. This requires significant angular momentum re-arrangement in proto-ellipticals, while the angular momentum distribution is conserved in proto-disks (e.g. Zurek et al. 1988). High density stellar substructure will transport angular momentum most efficiently as it merges, via dynamical friction, whereas low density, gaseous substructure may not. The latter then could form a disk galaxy. Hydrodynamic simulations of galaxy formation without star formation and feedback, but with gaseous cooling and viscosity, cannot form extended disks (e.g. Navarro & Benz 1991). Attempts to include feedback from massive stars in such simulations are as yet incomplete, but it remains the most plausible solution to the problem of excessive concentration of proto-disks (e.g. Navarro & Steinmetz 1997). Feedback from massive stars in the rather fragile potential of a disk will also keep the star formation efficiency low, allowing star formation to last a Hubble time.

We discuss here a program to address these questions from an observational study of a sample of nearby late-type disk galaxies. The spatial variation of present-day/recent massive-star formation is traced by H emission, while the past star formation is constrained by broad-band photometry, sensitive to older stars, and by spectroscopy of HII regions, to determine their chemical abundances. Special effort was made to study the outer regions of these galaxies, given that many of the `laws' of disk galaxy star formation diverge most in their predictions in the outer regions, and several aspects of present day outer disks, such as gas fraction, may be similar to those of the early stages of disks in general.

© Copyright Astronomical Society of Australia 1997