My task is to discuss the galactic context of star formation and to consider how large-scale phenomena in the Galaxy can influence the processes of star formation. This review, and the whole of the first session, are supposed to set the stage for the later consideration of protostellar and prestellar objects and smaller-scale effects in general.

To attempt to understand star formation without knowing the physical state of the dense interstellar molecular gas from which stars are made is an almost impossible task. Star formation has developed late as a branch of astrophysics largely for lack of observational data, and in particular, has lagged badly behind the study of the atomic and ionized components of the interstellar gas because spectroscopic techniques which work well at low density have an unfortunate tendency to fail when the density is high. Optical spectroscopy, which has been applied to the interstellar medium for over 70 years, has made little progress in regions of high density because of obscuration, and the same is true a fortiori of spacecraft spectroscopy in the UV; radio 21-cm and recombination line observations, although unhampered by obscuration, are unsatisfactory because the dense condensations are almost entirely molecular in composition.

Star formation is widely considered as resulting from the collapse of interstellar molecular clouds. The purpose of this paper is to review the observational evidences for collapse in dense clouds, and also for the factors which can play against collapse (turbulence, rotation, magnetic field). We shall also examine to which extent the maser sources (OH, H2O, SiO) can be related to star formation. An overlap with the review papers given by P. Thaddeus, P.G. Mezger, and to some extent by C.G. Wynn-Williams and L. Mestel appears unavoidable.

This review is divided into three parts. The first section gives a brief introduction to the different infrared wavelength ranges and to the various kinds of infrared objects seen in regions of star formation. The second section reviews the recent progress in infrared observations, concentrating on the three years since the review by Wynn-Williams and Becklin (1974) was written. The third section describes in more detail four varied examples of star formation regions.

A wealth of new observational data made available through the advent of modern infrared and microwave techniques, have provided astronomers with their first glimpses into the current-epoch stellar wombs - dark interstellar clouds. As a consequence, the past ten years have witnessed major advances in our understanding of the early evolution of stars and the mechanisms which trigger star formation. A future, more detached view of this period might suggest that we have achieved

To bring out the basic problem of star formation, consider a non-rotating cosmical gas cloud which is free to contract and so to convert most of its gravitational energy into kinetic energy of mass motion. Fluid dynamical systems in general tend to dissipate their kinetic energy, and a diffuse gas cloud easily radiates excess heat. If the cloud did not break up into fragments, then the steady conversion of gravitational energy into radiation would lead ultimately into the formation of a compact massive body - perhaps a quasar - but not of a star cluster, which is maintained at a comparatively low mean density by random stellar kinetic energy. Thus the transition from a gravitationally-bound cloud of gas and dust into a bound star cluster - or a fortiori into an expanding O-B association - requires that gravitational energy released during collapse be conserved as the random macroscopic kinetic energy. It is very likely that star formation is often triggered by processes that are essentially dissipative, such as shocks following passage of the spiral wave, or cloud-cloud collisions; however, it is equally important to emphasise that we require at some stage a cut-off in kinetic energy dissipation. And indeed, once a cloud has been able to fragment into a group of self-gravitating blobs, each of small geometrical cross-section, the chance of inelastic collisions between blobs is sharply reduced.

Although we now possess a wealth of information about the properties of molecular clouds and the circumstances in which stars form, we still have little direct information about the crucial question of how the material in molecular clouds actually becomes condensed into stars. In this report we shall discuss briefly the current status of theoretical attempts to understand this problem, based on calculations of the dynamics of collapsing clouds and protostars.

I have no plan of trying to sum up or to abstract the observational aspects of this conference. It would seem more profitable to discuss a few points that deserve special praise, or seem in need of critical attention, and to point to a few things I found puzzling and to tell you how I tried to explain them to myself.