be sacrificed. But in what proportion? If we confine ourselves to the view
that command depends on the battle-fleet, then we shall attach to it such a
number as its commander may deem necessary to make contact with the enemy
absolutely certain and to surround himself with an impenetrable screen. If
we knew the enemy was as anxious for a decision as ourselves, such a course
might be justified. But the normal condition is that if we desire a
decision it is because we have definite hopes of success, and consequently
the enemy will probably seek to avoid one on our terms. In practice this
means that if we have perfected our arrangements for the destruction of his
main fleet he will refuse to expose it till he sees a more favourable
opportunity. And what will be the result? He remains on the defensive, and
theoretically all the ensuing period of inaction tends to fall into his
scale. Without stirring from port his fleet is doing its work. The more
closely he induces us to concentrate our cruiser force in face of his
battle-fleet, the more he frees the sea for the circulation of his own
trade, and the more he exposes ours to cruiser raids.

Experience, then, and theory alike dictate that as a general principle
cruisers should be regarded as primarily concerned with the active
occupation of communications, and that withdrawals for fleet purposes
should be reduced to the furthest margin of reasonable risk. What that
margin should be can only be decided on the circumstances of each case as
it arises, and by the personal characteristics of the officers who are
responsible. Nelson's practice was to reduce fleet cruisers lower than
perhaps any other commander. So small indeed was the margin of efficiency
he left, that in the campaign already cited, when his judgment was ripest,
one stroke of ill-luck--a chance betrayal of his position by a
neutral--availed to deprive him of the decision he sought, and to let the
enemy's fleet escape.