hat an army may, with the view of destroying the enemy, pass the line
of these forts,--always, however, leaving a force to observe them; 3d,
that an army cannot pass a large river, like the Danube or the Rhine,
without reducing at least one of the fortresses on the river, in order
to secure a good line of retreat. Once master of this place, the army
may advance on the offensive, leaving detachments to besiege other
places; and the chances of the reduction of those places increase as the
army advances, since the enemy's opportunities of hindering the siege
are correspondingly diminished.

7. While large places are much the most advantageous among a friendly
people, smaller works are not without importance, not to arrest an
enemy, who might mask them, but as they may materially aid the
operations of an army in the field. The fort of Koenigstein in 1813 was
as useful to the French as the fortress of Dresden, because it procured
a _tete de pont_ on the Elbe.

In a mountainous country, small, well-located forts are equal in value
to fortified places, because their province is to close the passes, and
not to afford refuge to armies: the little fort of Bard, in the valley
of Aosta, almost arrested Napoleon's army in 1800.

8. It follows that each frontier should have one or two large fortresses
as places of refuge, besides secondary forts and small posts to
facilitate military operations. Walled cities with a shallow ditch may
be very useful in the interior of a country, to contain depots,
hospitals, &c, when they are strong enough to resist the attacks of any
small bodies that may traverse the vicinity. They will be particularly
serviceable if they can be defended by the militia, so as not to weaken
the active army.

9. Large fortified places which are not in proper strategic positions
are a positive misfortune for both the army and state.

10. Those on the sea-coast are of importance only in a maritime war,
except for depots: they may even prove disastrous for a continental
army, by holding out to it a delusive promise of support. Benningsen
almost lost the Russian armies by basing them in 1807 on
Koenigsberg,--which he did because it was convenient for supply. If the
Russian army in 1812, instead of concentrating on Smolensk, had
supported itself on Dunaburg and Riga, it would have been in danger of
being forced into the sea and of being cut off from all its bases.

The relations between sieges and the operations of a