A valuable framework for thinking about galaxy development was provided recently by Fukugita, Hogan and Peebles (1998). Taking a wide variety of data sources, they did a census of the baryons at both low and high redshift. While there are significant uncertainties associated with these estimates, they do provide a useful framework for constraining where the mass is in galaxies by type. For example, the low redshift census of the baryons in stars in galaxies shows that the dominant sink of baryons in galaxies integrated over all time t0 is spheroids. Ellipticals and the bulges of disk systems currently contain, in their stellar population, about 63% of the baryonic mass that is in galaxies (where this baryonic component is considered to be that in stars, stellar evolution end-products, or gas in the galaxy). All disks contain only about 21% of the baryonic mass, while the extremely numerous late type, and, typically, lower luminosity, galaxy population only has 2% of the baryonic mass in stars. The latter mass fraction is small even compared the gas that now remains in galaxies. The gas in galaxies comprises 15% of the total baryonic mass (integrated over all types). Thus a characterization of the star formation history of the universe is largely a description of the formation of spheroid populations, and secondarily, of disks, though if the timescales are quite different, one could still have a period where the dominant process is disk formation.