Two Problems in the Reduction of Thermodynamics to Statistical Mechanics