For pedagogical reasons we compute the caloric curve for 11 particles in a
3(3) lattice. Monte Carlo simulation can be avoided and exact results are o
btained. We compare canonical and microcanonical results for the caloric cu
rve. Even down to this small system, agreement between the canonical model
and the microcanonical model is surprisingly good. We point out that the in
troduction of kinetic energy in the nuclear lattice gas model modifies the
results of the standard lattice gas model in a profound way. The model is a
lso used to test the accuracy of the saddle-point approximation for density
of states.