While a number of authors over the years have discussed the star formation history of the universe, one of the clearest discussions of this issue has been provided by Madau (see, e.g., Madau et al. 1996; Madau 1998; Madau, Pozzetti and Dickinson 1998). The form that was used by Madau (1998) is shown in Figure 1, and gives the star formation rate per comoving volume vs redshift. This was one of the early papers from Madau and shows a rise from z ~ 5 to z ~ 1-2, and then a comparable drop to the present day. This form of the SFR vs redshift figure has become known as the ``Madau plot''. If plotted versus time, instead of redshift, the star formation history is more symmetric with a clear peak around a lookback time of ~ 60% of t0.
Figure 1. The star (and element) formation history of the universe, from the HDF (Madau 1998). The early data used for this figure indicated a strong peak in the SFR at about z ~ 1-2. See below for the form of the ``Madau plot''just 2 years later.
However, a number of developments (e.g., dust corrections, submm results, new estimates of the SFR at low redshift) are changing the shape of the SFR vs time, and making the variation since z ~ 5 much less. For example, one of the most significant developments of the last few years has been the recognition that dust extinction is skewing our view of distant galaxies. In particular, the rest frame UV SEDs (spectral energy distributions) that are measured from optical and near-IR observations invariably have slopes indicative of modest reddening. Since even small amounts of reddening significantly affect the the UV fluxes that we observe the resulting corrections can be large (a factor of 2 or more). This important development will be discussed further below.