An extensive examination of combustion gases containing trace amounts of mercury shows unambiguously mercury's propensity for heterogeneous chemistry. Although additional mechanisms for the oxidation chemistry of mercury have been implied by the continuing inadequacy of modeling attempts, details of the specific chemistry have remained unknown. Now it is shown that mercury can efficiently chemi-deposit onto surfaces encountered in practical combustors. If sulfur is present, condensed mercuric sulfate forms momentarily. This is then converted by gaseous HCl to HgCl2 that may sublime into the flow or be retained. This elusive and efficient noncatalytic mechanism most likely explains the observed fractional conversions to the dichloride observed in coal combustors. A receptive surface acts solely as an intermediary, facilitating the conversion while disguising its role. Without sulfur, a corresponding mechanism occurs but via HgO that is similarly converted to the dihalide. Such heterogeneous dynamics have significant repercussions for both full-scale combustors and bench-type experiments, which data have been reassessed and reviewed. Conclusions imply that observations concerning mercury will be system dependent and no two combustors can be exactly alike. The re-examination of prior work provides significant support for these conclusions. This fundamental understanding now lays a foundation for meaningful interpretations and program planning. It has indicated also the extreme care needed in sampling and monitoring the speciation of mercury in such combustion flows for reliable results. It now points to a simple low-cost surface-induced mitigation method for effectively converting the mercury in flue gases to the water-soluble dichloride. It is in essence no more than an optimization of the natural process that is currently occurring in combustors but to only limited degrees.