If the object of naval warfare is to control communications, then the fundamental requirement is the means of exercising that control. Logically, therefore, if the enemy holds back from battle decision, we must relegate the battle-fleet to a secondary position, for cruisers are the means of exercising control; the battle-fleet is but the means of preventing their being interfered with in their work. Put it to the test of actual practice. In no case can we exercise control by battleships alone. Their specialisation has rendered them unfit for the work, and has made them too costly ever to be numerous enough. Even, therefore, if our enemy had no battle-fleet we could not make control effective with battleships alone. We should still require cruisers specialised for the work and in sufficient numbers to cover the necessary ground. But the converse is not true. We could exercise control with cruisers alone if the enemy had no battle-fleet to interfere with them. If, then, we seek a formula that will express the practical results of our theory, it would take some such shape as this. On cruisers depends our exercise of control; on the battle-fleet depends the security of control. That is the logical sequence of ideas, and it shows us that the current maxim is really the conclusion of a logical argument in which the initial steps must not be ignored. The maxim that the command of the sea depends on the battle-fleet is then perfectly sound so long as it is taken to include all the other facts on which it hangs. The true function of the battle-fleet is to protect cruisers and flotilla at their special work. The best means of doing this is of course to destroy the enemy's power of interference. The doctrine of destroying the enemy's armed forces as the paramount object here reasserts itself, and reasserts itself so strongly as to permit for most practical purposes the rough generalisation that the command depends upon the battle-fleet. Of what practical use then, it may be asked, is all this hairsplitting? Why not leave untainted the conviction that our first and foremost business is to crush the enemy's battle-fleet, and that to this end our whole effort should be concentrated? The answer is to point to Nelson's dilemma. It was a dilemma which, in the golden age of naval warfare, every admiral at sea had had to solve for himself, and it was always one of the most difficult details of every naval war plan. If we seek to ensure the effective action of the battle-fleet by giving it a large proportion of cruisers, by so much do we weaken the actual and continuous exercise of control. If we seek to make that control effective by devoting to the service a large proportion of cruisers, by so much do we prejudice our chance of getting contact with and defeating the enemy's battle-fleet, which is the only means of perfecting control.