A high density of massive dark objects (MDOs), probably supermassive black
holes, in the centres of nearby galaxies has been inferred from recent obse
rvations. There are various indications that much of the accretion responsi
ble for producing these objects took place in dust-enshrouded active galact
ic nuclei (AGNs). If so, then measurements of the intensity of background r
adiation and the source counts in the far-infrared and submillimetre waveba
nds constrain the temperature of dust in these AGNs. An additional constrai
nt comes from the hard X-ray background, if this is produced by accretion.
One possibility is that the dust shrouds surrounding the accreting AGNs are
cold, about 30 K. In this event, the dusty AGNs could be some subset of th
e population of luminous distant sources discovered at 850 mum using the SC
UBA array on the James Clerk Maxwell Telescope, as proposed by Almaini, Law
rence & Boyle. An alternative is that the dust shrouds surrounding the accr
eting AGNs are much hotter (>60 K). These values are closer to the dust tem
peratures of a number of well-studied low-redshift ultraluminous galaxies t
hat are thought to derive their power from accretion. If the local MDO dens
ity is close to the maximum permitted, then cold sources cannot produce thi
s density without the submillimetre background being overproduced if they a
ccrete at high radiative efficiency, and thus a hot population is required.
If the dust-enshrouded accretion occurred at similar redshifts to that tak
ing place in unobscured optical quasars, then a significant fraction of the
far-infrared background radiation measured by COBE at 140 mum, but very li
ttle of the submillimetre background at 850 mum, may have been produced by
hot dust-enshrouded AGNs which may have already been seen in recent Chandra
X-ray surveys.