Should coolant accidentally be lost to the APT (Accelerator Production of Tritium) blanket and target, and the decay heat in the target be deposited in the surrounding blanket by thermal radiation, temperatures in the blanket modules could exceed structural limits and cause a physical collapse of the blanket modules into a non-coolable geometry. Such a sequence of unmitigated events could result in some melting of the APT blanket and create the potential for the release of mercury into the target-blanket cavity air space. Experiments were conducted which simulate such hypothetical accident conditions in order to measure the rate of vaporization of elemental mercury from pools of molten lead to quantify the possible severe accident source term for the APT blanket region. Molten pools of from 0.01% to 0.10% mercury in lead were prepared under inert conditions. Experiments were conducted, which varied in duration from several hours to as long as a month, to measure the mercury vaporization from the lead pools. The melt pools and gas atmospheres were held fixed at 340 C during the tests. Parameters which were varied in the tests included the mercury concentration, gas flow rate over the melt and agitation of the melt, gas atmosphere composition and the addition of aluminum to the melt. The vaporization of mercury was found to scale roughly linearly with the concentration of mercury in the pool. Variations in the gas flow rates were not found to have any effect on the mass transfer, however agitation of the melt by a submerged stirrer did enhance the mercury vaporization rate. The rate of mercury vaporization with an argon (inert) atmosphere was found to exceed that for an air (oxidizing) atmosphere by as much as a factor of from ten to 20; the causal factor in this variation was the formation of an oxide layer over the melt pool with the air atmosphere which served to retard mass transfer across the melt-atmosphere interface. Aluminum was introduced into the melt to investigate its effect upon the mercury vaporization rate in simulation of the aluminum structure in the APT blanket. No effect at all was observed for a case with an argon atmosphere. This suggests that there are no chemical effects of the aluminum on the vaporization kinetics. With an air atmosphere, the presence of aluminum in the melt reduced the mercury vaporization by a factor of six in comparison to the identical test but without aluminum present. This suggests that aluminum in the lead/mercury .melt retards the vaporization of mercury by creating a surface oxide layer in addition to the lead-oxide layer which increases the mass transfer resistance.