Using a sample of 165 X-ray selected QSOs from seven deep ROSAT fields we investigate the X-ray spectral properties of an ``average''radio-quiet broad line QSO as a function of redshift. The QSO stacked spectra, in the observers 0.1-2 keV band, in five redshift bins over the range 0.1<z<3.2 apparently harden from an equivalent photon index of 2.6 at z=0.4 to ~2.1 at z=2.4 as seen in other QSO samples. In contrast, the spectra in the 0.5-2 keV band show no significant variation in spectral index with redshift. This suggests the presence of a spectral upturn at low energies (<0.5 keV). Indeed, while at high redshifts (z>1.0) the single power-law model gives an acceptable fit to the data over the full energy band, at lower redshifts the spectra need a second component at low energies, a `soft excess'. Inclusion of a simple model for the soft excess, i.e. a black-body component (kT~100 eV), results in a significant improvement to the model fit, and yields power-law slopes of 1.8-1.9, for all redshift bins. This power-law is not inconsistent, within the error bars, with those of nearby AGN in the 2-10 keV band, suggesting that the same intrinsic power-law slope may continue from 10 keV down to below 0.5 keV. We caution that there is a possibility that the spectral upturn observed may not represent a real physical component but could be due to co-adding spectra with a large dispersion in spectral indices. Regardless of the origin of the soft excess, the average QSO spectrum has important consequences for the origin of the X-ray background: the average spectra of the typical, faint, high redshift QSO are significantly steeper than the spectrum of the X-ray background extending the spectral paradox into the soft 0.1-2 keV X-ray band.

We report on the discovery of a narrow-emission-line object at z=0.672 detected in a deep ASCA survey. The object, AXJ0341.4-4453, has a flux in the 2--10keV band of 1.1+/-0.27 E-13 erg/s/cm**2, corresponding to a luminosity of 1.8E44 erg/s (q0=0.5, $H0 = 50km/s/Mpc. It is also marginally detected in the ROSAT 0.5-2keV band with a flux 5.8E-15 erg/s/cm**2. Both the ASCA data alone and the combined ROSAT/ASCA data show a very hard X-ray spectrum, consistent with either a flat power law (alpha < 0.1) or photoelectric absorption with a column of nH > 4E22 cm**-2 (alpha = 1). The optical spectrum shows the high-ionisation, narrow emission lines typical of a Seyfert 2 galaxy. We suggest that this object may be typical of the hard sources required to explain the remainder of the X-ray background at hard energies.

We present additional optical spectroscopic identifications of sources identified in three deep ASCA GIS fields which also form part of a deep ROSAT survey. In total, 26 ASCA sources have been detected down to a 2-10keV flux limit of S = 5e-14 erg/cm/s. LDSS observations have increased the spectroscopic completeness of the survey to between 65 and 85 per cent, with identifications for up to 13 QSOs with broad emission lines and 6 objects with narrow emission lines. Combining these objects with the AGN identified in the LMA/HEAO-1 sample by Grossan, we find evidence for significant cosmological evolution in the 2-10keV band (<V_e/V_a>=0.63+/-0.03), fit by a pure luminosity evolution model; L \propto (1+z)^k, k=2.04^{+0.16}_{-0.22}$. The present-epoch 2-10keV AGN X-ray luminosity function, \Phi(L_X), is best represented by a two power law function: \Phi(L_X) \propto L^{-3.0}, L*>10^{44.1}erg/s; $\Phi (L_X) \propto L^{-1.9}$, L*<10^{44.1}erg/s. Depending on the extent to which we extrapolate the z=0 AGN luminosity function, we predict a total contribution to the 2-10keV X-ray background from QSOs and narrow-emission-line galaxies which ranges from 48 per cent (for AGN with L>10^{42}erg/s) to 80 per cent (L>10^{39}erg/s).

We summarise our recent work on the faint galaxy contribution to the cosmic X-ray background (XRB). At bright X-ray fluxes (in the ROSAT pass band), broad line QSOs dominate the X-ray source population, but at fainter fluxes there is evidence for a significant contribution from emission-line galaxies. Here we present statistical evidence that these galaxies can account for a large fraction of the XRB. We also demonstrate that these galaxies have significantly harder X-ray spectra than QSOs. Finally we present preliminary findings from infra-red spectroscopy on the nature of this X-ray emitting galaxy population. We conclude that a hybrid explanation consisting of obscured/Type 2 AGN surrounded by starburst activity can explain the properties of these galaxies and perhaps the origin of the entire XRB.

We present ASCA GIS observations (total exposure 100-200 ksec) of three fields which form part of our deep ROSAT survey. We detect 26 sources down to a limiting flux (2-10 keV) of 5x10^{-14} erg cm-2 s-1. Sources down to this flux level contribute 30 per cent of the 2-10 keV X-ray background. The number-count distribution, logN-logS, is a factor of three above the ROSAT counts, assuming a photon spectral index of 2 for the ROSAT sources. This suggests the presence at hard energies of a population other than the broad-line AGN which contribute to the ROSAT counts. This is supported by spectroscopic observations that show a large fraction of sources that are not obvious broad-line AGN. The average 1-10 keV spectral index of these sources is flat (0.92+-0.16), significantly different than that of the broad-line AGN (1.78+-0.16). Although some of the Narrow Emission Line Galaxies which are detected with ROSAT are also detected here, the nature of the flat spectrum sources remains as yet unclear.

We present ROSAT observations of two luminous L~10^44 erg/s EMSS galaxies, MS1019+5139 and MS1209+3917, previously classified as 'cooling flow' galaxies. MS1019+5139 does not appear to be spatially extended (<13 kpc) while its spectrum is well fit by a power law with Gamma = 1.73 +0.19-0.18; X-ray variability on a timescale of ~ years is also clearly detected. MS1209+3917 shows no evidence of spatial extension (<50 kpc) but it shows variability, while its spectrum can be fit with thermal bremsstrahlung emission (kT=1.8 +0.9-0.4 keV) or a power law model (Gamma = 2.50 +0.44-0.42, but with excess photoelectric absorption above the Galactic value). All the above argue against thermal emission from a group of galaxies or a galaxy but in favour of an AGN (possibly BL Lac) interpretation. We conclude that no 'normal' galaxies with high X-ray luminosities have yet been detected in the EMSS survey that could be significant contributors to the X-ray background.