Annu. Rev. Astron. Astrophys. 1992. 30: 429-456 Copyright © 1992 by Annual Reviews. All rights reserved

7. SUMMARY AND FUTURE WORK

There are three observational ``handles'' with which to identify the origin of the XRB: the spectrum, the isotropy, and the identified source content. They do not yet lead to a clear statement on the origin, although several classes of objects are ruled out and several others are found unlikely. A ``soft excess'' to the XRB at energies below about 3 keV is probably dominated by quasars. Direct counts of these sources in ROSAT images confirms earlier work from the Einstein Observatory ruling out any simple extrapolation to a surface density which integrates to give the whole XRB even at these energies. At least 2000 sources per square degree are needed to explain the XRB around 1 keV. This is a higher density than that of quasars, but does not rule out a lower luminosity class of AGN.

A much lower surface density of extragalactic sources has been identified in the 3-100 keV band where the bulk of the energy density of the XRB occurs. Quasars cannot make a major contribution unless their spectra flatten (by a change in power-law index of about unity) above 3 keV in the observed frame and unless they are less clustered than optical work indicates. No unevolved population that is clustered like nearby bright galaxies can contribute more than about 13% of the XRB in the 2-10 keV band. No known class of object has a mean spectrum similar to that of the XRB, although some classes of highly-absorbed objects such as Seyfert 2 galaxies in which there is a wide range of low-energy cutoffs might be acceptable (the last point means that strong evolution would also be required). A population of AGN, moderately X-ray luminous at a redshift of about 2, in which reflection is dominant, does predict the overall spectrum in a reasonably parameter-free manner. Whether this prediction is confirmed by observation must await the next generation of X-ray telescopes.

The XRB is relatively smooth at all energies. It is even smoother when the granularities from known classes of sources are removed. The current studies are reaching the levels in which spectral and anisotropy signals are beginning to emerge. ROSAT data will define the soft X-ray spectrum of the XRB, the mean shape of quasar spectra, and quasar source counts. The high fraction of quasars detected in ROSAT images offers the exciting possibility of mapping the spatial distribution of complete (i.e. flux-limited) quasar samples on scales of many tens of Mpc, and presumably thereby mapping the underlying mass distribution on those scales. Fluctuation and ACF analyses will measure the clustering of all sources including quasars and hopefully reveal other structures in the soft XRB. (It is to be hoped that absorption by Galactic ``cirrus'' does not seriously complicate such studies.) The spectrum of the XRB above 3 keV will be well-determined by the detectors on the recently-flown BBXRT which has also provided us with a taste of the results expected from ASTRO-D, Spectrum-X-, AXAF, and XMM. These next generation telescopes should make X-ray studies in the 0.1-12 keV band of both young AGN and the large-scale structure of the Universe an important research field.

ACKNOWLEDGMENTS

We thank our collaborators in the Ginga studies, F. Carrera, J. Butcher, K. Hayashida, H. Inoue, G. Stewart, and R. Warwick, together with K. Jahoda and D. Gruber for much help and B. Boyle and O. Lahav for discussions. ACF thanks the Royal Society for support and XB acknowledges partial financial support from the CICYT.