The very sensitive new FIRAS CMB-deviation limits severely constrain parametric translational models for galaxy evolution at wavelengths longer than 500µm. It follows that the evolutionary rates for infrared-bright galaxies implied by such models in the local universe cannot continue to cosmological redshifts for the class as a whole, unless exponential luminosity evolution is adopted. Most likely the current generation of parametric models is too simple in approach to adequately model this complex situation; in particular the likely strong evolution of the dust content of early type galaxies is not taken into account by these models.
The analytical chemical evolution models of Wang (1991a, b) address the evolution of the dust content. His accretion models are not in conflict with the current COBE limits but the prompt initial enrichment model is closely constrained. Similarly, the models of Franceschini et al. (1993), which are based on the closed box population synthesis models of Mazzei et al. (1992, 1993) incorporating chemical and dust evolution, are presently consistent with, but close to being constrained by, the COBE data. The simulations of deep galaxy fields of Chokshi et al. (1994, in preparation), which are also based on the population synthesis models of Mazzei et al. (1993), are also likely to provide interesting constraints (see Chokshi et al. 1993).
The FIRAS limits also constrain models for infrared-bright protogalaxies. If a significant fraction of the light created in primeval galaxies emerges in the far-infrared due to large dust optical depths then the FIRAS limits restrict the epoch or formation to z < 5 for low , and/or require relatively warm dust temperatures. A model based on the luminous, warm, z = 2.286 IRAS galaxy F10214+4724 can satisfy these requirements.
DIRBE will be able to detect the backgrounds expected from evolving galaxies and from infrared-bright protogalaxies, unless these objects are much less dusty than the models assume.
ISO and SIRTF will be able to detect objects like F10214+4724 to redshifts approaching 10. SIRTF will be able to resolve all of any background that DIRBE can expect to detect at 3 microns, and most of a DIRBE background at 60µm.
I thank Paola Mazzei, K. Matthews and B. T. Soifer for allowing me to reproduce figures from their papers, and Seb Oliver for making his model data files available. I am grateful to Mike Hauser for discussions regarding the DIRBE observations, Mike Werner and Peter Eisenhardt for discussions regarding SIRTF's potential, and Brick Young for permission to use his calculations of ISO and SIRTF point source sensitivities. I thank Ned Wright for allowing me to quote his unpublished derivations of SIRTF's ability to resolve infrared backgrounds. Rick Shafer kindly estimated the correction to the FIRAS limit for galactic foreground emission at the meeting. I thank George Helou and Chas Beichman for allowing me to show their revised evolutionary models in Figure 4, and the foreground solar system and galactic emission models in Figure 1. I thank Joe Mazzarella for discussions on the nature of F10214+4724, Boqi Wang and Mike Fall for bringing my attention to some important work, and Perry Hacking for comments on the manuscript. The research described in this review was carried out by the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration.