We intend to obtain photometry of a large, high-redshift
  sample of Ultraluminous IRAS galaxies,  over the full spectral
  range offered by ISO.  We will then be able to explore the
  evolution of the spectral energy distribution with redshift for
  this very important class of galaxies.

  Direct comparisons with local high-luminosity samples would reveal
  any changes in the SED with redshift which might indicate either
  a change in the dominant source of power, e.g. an increase in AGN
  activity over star-burst activity, or the dust distribution, e.g.
  higher optical depth, different geometries.

  We would be able to deconvolve the observed SED into components
  due to AGN and starburst and investigate how their relative
  contributions change with both luminosity and, when compared with a
  low-z sample, redshift.

  Radiative transfer modeling of these SEDs would provide vital
  ingredients to models of AGN and starburst evolution.

  This kind of study represents a significant advance over traditional
  methods of studying evolution which are usually limited to studying
  changes in the luminosity function which can often be ambiguous.

  ISO data is vital because such galaxies emit the vast bulk of
  their energy at far-infrared wavelengths not accessible to ground
  based instruments.