1. INTRODUCTION

There are certain more or less well or badly determined integral constraints on the past history of star formation in the universe. These include

• Cosmic baryon density _B, now fairly well determined both from primordial deuterium (O'Meara et al 2001) and from the CMB fluctuation spectrum (e.g. Turner 2001).

• Cosmic mass density of stars _*, rather less well determined as it involves a combination of luminosity-density measurements with an IMF and evolutionary population synthesis models.

• Extragalactic background light (EBL) intensity, now known within a factor of 2 or so from COBE (FIRAS and DIRBE) and galaxy counts in the optical and near IR (e.g. Gispert, Lagache & Puget 2000).

• Cosmic abundance of "metals", _Z, due to the heavy-element content of stars, the interstellar medium and the intergalactic medium, a quantity that is very poorly known and largely a matter of guesswork. In this talk I shall nevertheless make some guesses, so that at least one can see more easily how things relate to one another. In particular, the metallicity Z_IGM of the intergalactic medium has tended to be either neglected or underestimated in models of the past star formation rate, and it is of interest to ask about its relation with EBL.