A curious distinction attaches to the work of naval coast defence. Usually the line of observation lies in advance of the line of resistance. In coast defence the line of observation is in rear of the line of resistance. So far as the patrol flotillas are concerned, the British coasts are themselves the only true and certain line of observation. The approach of an enemy may be undetected by the cruising squadrons or by the patrolling flotillas. But it ought to be certain that his first contact with the coast at any point is reported to the Admiral of Patrols, and that that officer will have his available forces massed at convenient points from which an attack can be at once delivered. The Admiral of Patrols must treat his problems selectively and recognise that absolute certainty is out of reach, that his flotillas are for fighting purposes, and that their rôle of scouting is secondary. It is of very little use reporting the approach of an enemy when one has not the forces with which to strike him. The patrol flotillas are therefore kept in hand at the best strategic points, neither scattered nor exhausted, and a system of land observation by outposts, cyclists, aircraft and signal stations, all connected by telephone, ought to be perfected, from which accurate information can be transmitted to the points where the patrol flotillas are massed. Dalesvoe (Shetlands), Fort Ross, Firth of Forth, North Shields, Grimsby, and Yarmouth are the bases of the patrol flotillas, and a force of fourteen or fifteen vessels would, on the average, be available for each. It is upon this disposition that the Admiralty rely to interrupt the disembarkation of any considerable force. It is of vital importance that the watching of the coast-line from the shore should be taken up from the earliest moment and in advance of general mobilisation. The effectiveness of the work of the patrol flotillas and consequently the restriction of possible landings depend upon early information being received of any disembarkation. The size of any raiding party that could be landed will, of course, be accurately proportionate to the delay. It would no doubt be impossible or undesirable to put the whole system of coast watches into operation in the precautionary period. No doubt the arrangements made after war had actually begun would be much more thorough, and larger numbers of cyclists and watchers would be available. But a system of watching likely landing-places ought to be devised which could be brought silently into operation as soon as the precautionary period is declared or, if necessary, immediately before, just in the same way as the watch over the magazines and other vital points can unostentatiously be improved.