In 1812 Gauss, in a letter to Laplace, proposed without proof a formul
a explaining the statistical regularity of continued fractions. There
has since been speculation concerning the manner in which Gauss arrive
d at this formula. In this article we present a plausible explanation,
which at the same time gives an elementary proof of the full ergodic
nature of the underlying dynamical system. The method seems to be of i
nterest for other number-theoretic expansions.