This paper shows that when students are introduced to the derivation of one of the most important
physical formulas—the canonical distribution—they are exposed to assumptions which may be confusing
because they contradict physical reality. The paper provides an alternative derivation of the
canonical distribution. Our derivation takes into account internal physical processes in
macrosystems leading to the canonical distribution and does not require any physically unjustified
assumptions. The article is intended for teachers and students of statistical physics, and for
general physicists.