5.4.2. Hidden Velocity Structure
To obtain D/H we need to estimate the column densities of D and H. Column densities depend on velocity distributions, and when lines are saturated, it is hard to deduce these velocity distributions. Similar line profiles are made when the velocity dispersion is increased to compensate for a decrease in the column density. We mentioned above that this degeneracy is broken when we observe lines along the Lyman series. For Q1937-1009, which has the most saturated H lines of the quasars under discussion, Burles & Tytler  showed that the D/H did not change for arbitrary velocity structures, constrained only by the spectra. The same conclusion was obtained for Q1009+2956 . The favorable results for these two quasars do not mean that we will always be able to break the degeneracy. That must be determined for each absorption system.
There are two reasons why hidden velocity structure is not expected to be a major problem. First, we are concerned about hidden components which have high columns and low enough velocity dispersions that they hide inside the wider lines from lower column gas. Such gas would be seen in other lines which are not saturated: the D lines and the metal lines from ions with similar (low) ionization. Second, we search for D in absorbers with the simplest velocity distributions. They tend to have both narrow overall velocity widths and low temperatures, which makes it much harder to hide unseen components. Typically, the main component accounts for all of the absorption in the higher order Lyman lines, and these lines are too narrow for significant hidden absorption.