A few remarks on the relative cost of ships and forts, and the economy of their support, and we will close this discussion. We do not regard this question, however, as a matter of any great importance, for it can seldom be decisive in the choice of these two means of defence. No matter what their relative cost may be, the one cannot often be substituted for the other. There are some few cases, however, where this might be taken into consideration, and would be decisive. Let us endeavor to illustrate our meaning. For the defence of New York city, the Narrows and East River must be secured by forts; ships cannot, in this case, be substituted. But let us suppose that the _outer_ harbor of New York furnishes no favorable place for the debarkation of troops, or that the place of debarkation is so far distant that the troops cannot reach the city before the defensive forces can be prepared to repel them. This outer harbor would be of great importance to the enemy as a shelter from storms, and as a place of debarkation or of rendezvous preparatory to a forcible passage of the Narrows; while to us its possession would not be absolutely essential, though very important. Strong fortifications on Sandy Hook, and one of the shoals, might probably be so constructed as to furnish a pretty sure barrier to the entrance of this outer harbor; on the other hand, a naval force stationed within the inner harbor, and acting under the protection of forts at the Narrows, might also furnish a good, though perhaps less certain protection for this outer roadstead. Here, then, we might well consider the question of relative cost and economy of support of the proposed fortifications, and of a home squadron large enough to effect the same object, and to be kept continually _at home_ for that special purpose. If we were to allow it to go to sea for the protection of our commerce, its character and efficiency as a _harbor_ defence would be lost. We can therefore regard it only as a local force--fixed within the limits of the defence of this particular place--and our estimates must be made accordingly.