Star formation is studied by astronomers not only because it produces new stars and planetary systems. It also generates copious amounts of ultraviolet light that heats dust which in turn causes the birth region to shine brightly in the infrared. Galaxies so far away, for example, that their light has been traveling for over eleven billion years have been discovered thanks to their bright infrared star formation activity. However the observation of star formation in other galaxies tends to encompass very large volumes; in our own galaxy, by contrast, research focuses on individual star forming molecular clouds because they are much closer and so appear much larger in angular size. A fundamental yet still only partially unresolved question is whether the same physical processes are at work in all cases. It could be, for instance, that large-scale effects in galaxies, such as inter-galaxy collisions, make their star factories completely different (on average) from those in local, relatively quiescent clouds. After all, the estimated rates of star formation in infrared galaxies are sometimes a million or more times that of local clouds.
CfA astronomers Charlie Lada and Jan Forbrich, with two colleagues, argue in a new paper that the basic processes are the same. They examined the relationship between the rate of star formation (as determined by numbers of young stars) and the density of molecular gas in the natal regions (as determined by radio measurements of diagnostic molecules). They found good evidence that in all cases the rate of star formation, across nearly a factor of a billion, is linearly proportional to the amount of dense gas present. Their result contradicts the earlier, more established idea that the relationship is non-linear with total gas abundance, but the authors offer a convincing explanation for why the earlier results were in error. The new paper helps to resolve the uncertainty about global star formation, and focuses future research on the question: what produces the dense gas that is responsible?