We propose a far-IR study of a complete and well-defined sample of radio
sources from the B3VLA survey. The scientific goals of this investiga-
tion are to constrain the relative importance of AGN and starbursts in
high-redshift radio galaxies, and the unification models of quasars and
powerful radio galaxies. We wish to use the far-IR (FIR) luminosities
of these objects, in conjunction with the existing and soon to be obtai-
ned ground-based radio, optical, and near-IR data, in several statisti-
cal tests. While previous work based on the IRAS satellite data has
concentrated on the FIR properties of pure starbursts and QSOs, very
little is known about the FIR properties of radio galaxies, or indeed
radio-selected samples in general. Our proposed observations would
close this gap in a systematic manner, while addressing several inter-
esting scientific goals. They will also constrain the evolution of the
radio-loud populations over a substantial baseline in redshift, an
important measurement on its own. There is a real possibility of disco-
very of a new population of high-z starforming (proto)galaxies, similar
to the ultraluminous source IRAS F10214, or even some completely new
type of objects. Specifically, we propose to use PHOT for a photometric
survey of a subsample of 400 radiosources from the B3VLA sample, which
is complete down to 100mJy at 408 MHz. The subsample is at high lati-
tudes, and close to the Lockman hole, so that confusion due to cirrus
should be minimal. It has been extensively studied with VLA maps in 3
configurations, and flux measurements at several frequencies, ranging
from 0.15 to 10.6GHz, and with mm observations planned in 1995. Optical
IDs from POSS prints and new CCD work from Palomar exist for a large
fraction of this sample, with an ongoing efforts to identify the rest.
Redshifts have been measured for all quasars in the sample, and several
tens of radio galaxies, ranging up to z=3.2 so far. The sample spans a
large dynamical range in redshift and in radio power, which makes it
nearly unique for studies of systematic trends in both of those axes.