We propose to use ISO to identify high redshift (z>1) galaxies from
  the IRAS Faint Source Survey.  The strong signature of the redshift
  in changing the far infrared spectrum towards peaking at longer
  wavelengths makes ISO multiwavelength photometry a powerful tool
  for identifying high redshift galaxies.  The discovery of IRAS
  F10214+4724 at z = 2.23, with a bolometric luminosity of 5e14 solar
  luminosities, has been enermously stimulating for submillimetre
  continuum and molecular line astronomy, offering the prospect of
  studying galaxies close to their moment of formation.  However
  it has proved extremely difficult to find other galaxies of this
 type in the IRAS catalogues.
     We have followed up the IRAS FSS 60 mu sources in two areas
  of the northern hemisphere each 700 sq deg in area, a total
  sample of over 2800 IRAS galaxies.  For the first area (FSS1) we have
  identified and obtained redshifts for all but 53 of the 1400 sources
  in the area and for most of these 53 unidentified sources we have
  obtained VLA maps, and where necessary CCD images.  Most have either
  very faint identifications on the sky survey plates or CCD frames
  or are blank, and are thus likely to have z > 0.2 and hence 60 mu
  luminosities > 1.e12 solar luminosities.  For the second area (FSS2)
  the identifications are 90% complete and we have selected 71 objects
  which are the best candidates for having faint, high redshift
  idenstifications.
  The purpose of the ISO observations is (1) to confirm the reality
 of the IRAS FSS source, (2) to measure the SED of this interesting
  sample, the bulk of which will be ultraluminous galaxies, (3) to
  identify high redshift galaxies from their SED for further intensive
 study with large ground-based telescopes.