The Lyman-alpha forest clouds remain a valuable, but poorly understood, diagnostic of the thermal and dynamical history of the intergalactic medium (IGM). A long-posited model is that these clouds represent the cooler phase of a two-phase instability in the tenuous intergalactic gas and that they are pressure or gravitationally confined against evaporation in the hot pervasive medium. In confinement models, variations in the pressure of the hot confining medium or specific cloud mass ranges are used to explain the observed range of column densities at a fixed epoch. However, weak clustering of Ly-alpha absorbers compared to galaxies and the apparent rarity of detectable voids in the Ly-alpha forest, which implies that the pressure of the confining medium is approximately the same in the voids as well as in clusters, are difficult issues in such models. Gravitationally confined clouds are difficult to explain since the mass of such clouds must lie in a restricted range to maintain the gas in equilibrium against free expansion or collapse.

Recent models propose that the Ly-alpha clouds are stabilized by the gravitational potential wells of dark matter. Other models attempt to describe the dynamical evolution of the Ly-alpha clouds by relaxing the equilibrium requirements. However, arguments can be placed against such models, in that freely expanding clouds are not possible because the observed Ly-alpha line widths are too small for the expansion velocities of the clouds. Furthermore, collapse times are too short for self-gravitating clouds. It is generally agreed, however, that Ly-alpha forest lines arise from intervening clouds with a low heavy-element abundance that are cosmologically distributed.

The Ly-alpha forest clouds have clustering properties in velocity that set them apart from the metal-line systems. Recent high- resolution spectra show clustering at a level much weaker than that seen in nearby galaxies on velocity scales typically less than 500 km/s for at least some subsamples of Ly-alpha absorbers. Other observations find no evidence for low-velocity clustering in the Ly-alpha forest at redshifts in excess of 4 (on one line of sight), but such samples are subject to severe line blending. In general, the velocity splittings between adjacent Ly-alpha lines, c[dz/(1+|z|)] where dz=z2-z1 and |z|=(z1+z2)/2, is consistent with a random, uncorrelated Poissonian statistics. The heavy-element systems (e.g., C IV) tend to cluster strongly on galactic scales of less than about 300 km/s. An analysis performed by Tytler (1987) suggests a single population of Ly-alpha clouds and metal-line absorbers. Tytler claims that both cloud types fit the same power-law distribution of H I column density. However, if one considers the individual velocity components in data with higher velocity resolution, the distribution of column densities is not consistent with a single power-law fit, although the sample of higher column density absorbers on which this argument is based, is very small. Differences also remain in the clustering properties and redshift evolution that pose serious questions for the single-population picture.

The common absorbers shown above may be due to material ejected from each quasar along their respective lines of sight or physically associated intervening gas clouds separated by some 30 Mpc. There are physical problems involved in making either of these hypotheses work over such large scales.

Recent spectral observations reveal no common metal-line systems in any of the quasar pairs among the group of four quasars shown below. These findings are consistent with the scenario of single intervening galaxies along the lines of sight instead of large coherent structures of galaxies on supercluster scales that may be seen in common absorption. However, three of the quasars have one C IV system each, at redshifts 1.4566, 1.4146, and 1.4959 (see spectra below); the fourth has no C IV detected. The three lines of sight have a total path length of 2.1 (in redshift space) accessible to detecting C IV absorption. It is provocative that the three detected systems match in redshift to 0.08, or only 4% of the available redshift range. No firm conclusions can be drawn from the limited statistics of one group of lines of sight; however, observations of C IV systems in other wide quasar pairs may provide interesting information on large-scale structure. The consequences of coherent structures of galaxies on scales greater than 10 Mpc at a redshift of approximately 2 would be profound for theories of large-scale structure.