In order to measure the thermal emission from high-redshift z > 3
 QSOs, we are proposing to do photometry by using ISOCAM and PHT-C
 at the LW2 (6.7 um), LW10 (11.5 um) C_90 (90 um), and C_160 (160 um)
 band.  From this observation, the mass of dust in a circumnuclear
 region and a disk  could be estimated.  If the formation of galaxies
 took more than 1 Gyr as suggested by a star formation rate measured in
 several high-redshift radio-quiet galaxies, galaxies were still
 forming at z = 4.  Hence, it is natural to search for the thermal
 mission from dust in high-redshift (young) QSOs, and the mass of
 dust would provide a clue to star formation at high redshifts where
 dust formed through supernovae and destryed by shocks and intense
 radiation.  Our primary goal is to detect QSOs at the LW2 (1.3 um
 in the rest frame at z =4) and C_90 (18 um at z = 4) bands.
 The C_90/LW2 luminosity ratio will be used as an index of emission
 from dust.  The sample consists of 27 - 30 QSOs with z > 3 (27 QSOs
 for spring and 30 QSOs for autumn).  Most of them
 are optically selected, NE_8Band so probably radio-quiet QSOs,
 the infrared continuum of which would be dominated by the thermal
 emission from dust as in the case of low-redshift PG QSOs.