2000ApJ...530..277E

Observations suggest that star formation occurs in only one or two crossing times for a range of scales spanning a factor of ∼1000. These observations include (1) measurements of embedded cluster ages in comparison with the cloud core dynamical times, (2) measurements of the age difference versus separation for clusters in the Large Magellanic Clouds in comparison with the crossing time versus size correlation for molecular clouds, (3) the hierarchical structure of embedded young clusters, and (4) the high fraction of dense clouds that contain star formation.

Such a short overall timescale for star formation implies that sources of turbulent energy or internal feedback are not required to explain or extend cloud lifetimes and that star and protostar interactions cannot be important for the stellar initial mass function. Stars appear in a cloud as if they freeze out of the gas, preserving the turbulent-driven gas structure in their birth locations. The Galaxy-wide star formation rate avoids the Zuckerman-Evans catastrophe, which has long been a concern for molecular clouds that evolve this quickly, because the multifractal structure of interstellar gas ensures that only a small fraction of the mass is able to form stars. Star formation on large scales operates more slowly than on small scales, but in most cases the whole process is over in only a few dynamical times.