The exhaust modeling program for a fusion reactor based on a tokamak carried out for Next European Torus (NET)/International Tokamak Reactor (INTOR)/International Thermonuclear Experimental Reactor (ITER) in the years 1982 to 1991 during which the author was involved is open to criticism on at least two counts. The first is that although in general there are at least two plasma configurations in the divertor that balance the upstream plasma pressure and power flow into the divertor, only one solution was accepted. The other solutions were assumed to be nonexistent or unimportant. The second count is that the possibility was not considered that atoms backscattered from the plasma could deposit power in the divertor target; inclusion of this process would have enlarged the domain in which multiple solutions are important. In particular, a plasma in which the temperatures are low appears as a possible solution. Here the atomic and molecular properties of the fuel, which vary quite rapidly with the electron temperature, are very important, so obtaining this solution by an implicit procedure is difficult. The two-dimensional modeling programs referred to earlier were carried out with the use of the Braams plasma transport code, which relies on a “strongly implicit method” for its updating. Examination of this code shows that the techniques used to stabilize it are incompatible with the procedures required to find the low-temperature solution. These objections would remain in the case where a Monte Carlo code is used to trace the fate of the neutrals recycled in the divertor. Recent modeling work based on Monte Carlo codes suggests that the plasma temperatures of the higher temperature solutions are likely to be greater than previously thought. Thus, resolution of the problem of finding the lower temperature solutions becomes important.