perhaps--or at any rate in the case of France and Germany--the numbers
of the _active_ reserve immediately behind the conscript army in
peace.

The second method, which is better, but imperfect, is that which has
particularly appealed to technical writers. It consists in numbering
_units_; in noting the headquarters and the tale of army corps and of
independent divisions.

The fault of this method is twofold. First, that only actual
experience can tell one whether units are really being maintained
during peace at full strength; and secondly, that only actual
experience discovers how many new units can and will be created when
war is joined. In other words, the fault of this method (necessary
though it is as an adjunct to all military calculations) lies in its
divorce from the reality of numbers.

At the end of the retreat from Moscow each army corps of the Grand
Army still preserved its name, each regiment its nominal identity. And
the roll was called by Ney, for instance, before the Beresina,
division by division and regiment by regiment, and even in the
regiments company by company; but in most of these last there was no
one to answer, and there is a story of one regiment for which one
surviving man answered with regularity until he also died. What fights
is numbers of living men--not headings; and if five army corps are
present, each having lost two-fifths of its men, three full army corps
are a match for them.

The third method is that of commonsense. We must deduce from the
results obtained, from the fronts covered, from the energy remaining