Together with A. Boksenberg (Royal Greenwich Observatory) I recently completed observations of the absorption spectra of 60 QSOs with redshifts in the range 1.8 < zem < 3.56, primarily in order to study the clustering and evolution of heavy-element redshifts as represented by the easily identified CIV 1548,1550 doublet. The spectra were obtained with Boksenberg's IPCS detector mounted on the blue camera of the Double Spectrograph at the Cassegrain focus of the Hale telescope. The on-line display of the accumulated count level was used to obtain a uniform S/N ratio of 20:1. The resolution was 1.4 Å. Most of the spectra covered the wavelength range from the CIV 1549 emission line down to the Ly- emission line. In several cases the coverage extended into the Ly- forest and in some cases the resolution was 0.7 Å. The spectra have all been reduced, measured and redshifts have been determined. The results are now being assembled for publication (Sargent, Boksenberg and Steidel 1988). The results extend those obtained in similar, but less ambitious, surveys carried out by Young, Sargent and Boksenberg (1982) and by Foltz et al. (1986).
The main results of the CIV survey are as follows.
A total of 202 CIV absorption redshifts were identified in the 60 QSOs. A uniform sample was isolated of 130 redshifts with rest equivalent widths for both CIV lines greater than 0.15 Å. Some of these systems fall in clumps which are clearly not independent entities. Therefore, an additional sample (called the 'Poisson sample') was formed in which clumps containing more than one system on a scale of less than 1000 km s-1 were treated as one system. The Poisson sample contains 107 absorption redshifts.
In this sample, which is composed primarily of radio-quiet QSOs, there is no significant tendency for absorption redshifts to cluster around the emission redshift of the QSO. This behavior is in marked contrast to that exhibited by radio QSOs as Foltz et al. (1986) have clearly shown.
The distribution of number of absorption redshifts per QSO does not conform to a Poisson distribution as is expected for randomly distributed intervening absorbers (Bahcall and Peebles 1969). This is true at the 90 percent confidence level even for the Poisson sample. A likely explanation of this result is that over the small ranges in redshift (z ~ 0.6) available in any given QSO, the details of the large scale distribution of galaxies along the line of sight play an important role in the statistics.
The density of absorption systems per unit redshift range dN / dz decreases with increasing redshift in the range 1.4 < zabs < 3.4 according to an approximate law dN / dz (1 + z)-1.2±0.4. All three samples show the same qualitative behavior. On the other hand, a constant co-moving density of absorbers with constant cross-section should have dN / dz (1 + z) if q0 = 0 and dN / dz (1 + z)1/2 if q0 = 1/2. The observed behavior of the CIV doublet density is quite different to that shown by the Ly- clouds which increase in density as dN / dz (1 + z)2.3 over the same redshift range (Murdoch et al. 1986). Moreover, the 'Lyman limit' absorption systems and the MgII redshift systems are observed to increase with redshift at something like the rate expected for q0 = 0 (Tytler 1986, private communication).
The 2-point correlation function for the CIV doublets has been generated. It is generally flat as is expected for a randomly distributed sample of absorbers. However, there is significant clustering on scales less than 500 km s-1 in the rest frame of the absorbers. Analysis shows that the observed clustering is very unlikely to be due to motions of clouds within galaxies but is more likely due to the galaxian correlation function. These new results, which are still being analyzed, emphasize the differences between the heavy-element redshifts and the Ly- forest lines, which display little or no clustering tendency.