1999ApJ...518..226M

A satisfactory theory for the formation of globular star clusters (GCs) has long been elusive, perhaps because their true progenitors had not yet been guessed. In this paper, I propose a causal relationship between the strongly decreasing densities of Lyα clouds at high redshift and the formation of GCs–namely, that GCs were created by the accretion of Lyα clouds onto protogalaxies. I describe a scenario that involves an inherently stable and orderly cycling of compression and cooling in the central cores of these clouds during the extended period of dissipation in the outer regions of gas-rich protogalaxies, culminating in a burst of efficient star formation. I demonstrate that the comoving density of GCs is comparable to that of Lyα clouds at high redshift, that the energetic requirements for compression to core GC densities can be met, and that the timescale for cooling is within obvious limits imposed by dynamical stability. This dissipative process requires there to be a large column of dissipated gas about the attractor in order to form GCs. In addition, the energy requirements of compression to GC central densities requires attractor masses greater than that capable of sustaining circular velocities of ∼40 km.s^–1. If this scenario is supported by numerical simulations, then by implication the GCs were formed at modest redshifts of z∼1-3. This knowledge could help break the troublesome degeneracy between look-back time and redshift. This model is consistent with a picture of hierarchical galaxy growth over timescales of many billions of years.