Previous
abstract 
Next
abstract
We find that HST Lyman continuum spectra of 27 redshift one QSOs are completely flat from 850 to 1000 A. We do not see any absorption or excess emission around 912 A, nor is there any change in the continuum slope. There are important implications for both the source of the Lyman continuum radiation, and for the broad emission line clouds.
The continuum radiation is either (1) non-thermal, as Rauch \& Blanford argued is required to explain the rapid microlensing variation of Q2237+020, or (2) it comes from extremely dense clouds with $10^{17}$ atoms per cm$^3$ which Ferland \& Rees 1988 showed emit like black bodies (in LTE). Such clouds can not have cool atmospheres which would absorb in the Lyman continuum. We do not favor emission from thermal accretion disks or free-free emission because in both cases the temperature must be extremely high to avoid Lyman continuum absorption, and such high temperatures are probably ruled out by the observed continuum slopes.
It is well known that the power emitted by the broad emission lines is equivalent to about 10\% of that in the Lyman continuum. We know that the broad line clouds are not much larger than the Lyman continuum source because we have never seen a strong intrinsic Lyman limit systems which could arise in a broad line cloud (BLC). Our new result also excludes clouds which are much smaller than the continuum source and which are isotropically distributed, because we would then expect to see 10\% absorption in most QSOs. There remain two possibilities. (1) The broad line clouds are restricted to a two-dimensional distribution, such as a disk. We also require that lines of sight which pass from the continuum source and through the BLC are blocked, so that we do not see much UV radiation. We never see absorption in the BLCs because such QSOs are not UV bright. This picture is consistent with the general theme of unification, in which AGNs are not spherically symmetric, and there is obscuration. Alternatively, (2) the BLC are not powered by UV radiation, which is harder to accept because line equivalent widths are similar over a very large range of continuum luminosity.
