As has been mentioned in Chapter II., the first ships to be brought under a system of convoy were those engaged in the French coal trade and in the trade between Scandinavia and the United Kingdom.

In the case of the French coal trade, commencing in March, 1917, the steamships engaged in the trade were sailed in groups from four different assembly ports, viz.:

Between Southend and Boulogne and Calais the protection was given by the vessels of the Dover Patrol in the course of their ordinary duties, but for the other three routes special escort forces were utilized, and daily convoys were the rule.

Owing to the great demand for coal in France, sailing vessels were also used, and sailed under convoy from several of the south-west ports.

A large organization was required to deal with the trade, and this was built up under the supervision of Captain Reginald G.H. Henderson, C.B., of the Anti-Submarine Division of the Naval Staff, working under Vice-Admiral (then Rear-Admiral) Sir Alexander Duff, head of the Division, in conference with the Commanders-in-Chief, Portsmouth and Plymouth, under whose direction and protection the convoys were run. The immunity of this trade, carried out in the infested waters of the English Channel, from successful attack by submarines was extraordinary. No doubt the small size of the vessels concerned and their comparatively shallow draught were a contributory cause to this immunity. The figures for the period March to August, 1917, show that 8,825 vessels crossed the Channel under convoy, and that only fourteen were lost.

The history of the Scandinavian and East Coast convoys dates back to the autumn of 1916, when heavy losses were being incurred amongst Scandinavian ships due to submarine attack. Thus in October, 1916, the losses amongst Norwegian and Swedish ships by submarine attack were more than three times as great as the previous highest monthly losses. Some fear existed that the neutral Scandinavian countries might refuse to run such risks and go to the extreme of prohibiting sailings. Towards the end of 1916, before I left the Fleet, a system of "protected" sailings was therefore introduced. In this system the Commander-in-Chief, Grand Fleet, fixed upon a number of alternative routes between Norway and the Shetland Islands, which were used by all vessels trading between Scandinavia and Allied countries. The particular route in use at any given moment was patrolled by the local forces from the Orkneys and Shetlands, assisted when possible by small craft from the Grand Fleet. The Admiral Commanding the Orkneys and Shetlands was placed in charge of the arrangements, which were carried out by the Senior Naval Officer at Lerwick, in the Shetland Islands. At this period the intention was that the shipping from Norway should sail at dusk, reach a certain rendezvous at dawn, and thence be escorted to Lerwick. The shipping from Lerwick sailed at dawn under protection, dispersed at dark, and reached the Norwegian coast at dawn. Difficulties, of course, arose in the event of bad weather, or when the slow speed of the ships prevented the passage of about 180 miles being made in approximately twenty-four hours, and by April, 1917, it was evident that further steps were necessary to meet these difficulties, which were again causing heavy losses. Early in April, then, by direction from the Admiralty, a conference was held at Longhope on the subject. Admiral Sir Frederick Brock, Commanding the Orkneys and Shetlands, presided, and representatives from the Admiralty and the Commands affected were present, and the adoption of a complete convoy system to include the whole trade between the East Coast and Norway was recommended. This proposal was approved by the Admiralty and was put into force as soon as the necessary organization had matured. Escorting vessels had with difficulty been provided, although in inadequate numbers. The first convoys sailed towards the end of April, 1917.

The system may be described briefly as follows. The convoys all put into Lerwick, in the Shetland Islands, both on the eastward and westward passages, so that Lerwick acted as a junction for the whole system. From Lerwick, convoys to Scandinavia left in the afternoon under the protection of two or three destroyers, and, with some armed patrol vessels in company up to a certain stage, made the Norwegian coast at varying points, and there dispersed, and the destroyers then picked up the west-bound convoy at a rendezvous off the Norwegian coast shortly before dark, and steered for a rendezvous between Norway and the Shetland Islands, where an escort of armed patrol vessels joined the convoy at daylight to assist in its protection to Lerwick. From Lerwick convoys were dispatched to various points on the coast of the United Kingdom; those making for southern ports on the East Coast were escorted by a force composed of some of the old "River" class or of 30-knot class destroyers, and trawlers belonging to the East Coast Command based on the Humber, and those making for more northerly ports or ports on the West Coast were escorted merely by armed patrol vessels, as the danger of submarine attack to these convoys was not so great.

The main difficulty was the provision of the destroyers required for the proper protection of the convoys, and to a lesser degree the provision of armed patrol vessels of the trawler, whaler, or drifter types.

The conference held early in April, 1917, had reported that whilst stronger protection was naturally desirable, the very least force that could give defence to the convoys between Lerwick and the East Coast ports would be a total of twenty-three destroyers and fifty trawlers, whilst for each convoy between Lerwick and Norway at least two destroyers and four trawlers were needed. The destroyers for the latter convoys were provided by the Grand Fleet, although they could ill be spared. The total number so utilized was six. It was only possible to provide a force of twenty old destroyers and forty-five trawlers for the East Coast convoys instead of the numbers recommended by the conference, and owing to the age of a large majority of these destroyers and the inevitable resultant occasional breakdown of machinery, the number available frequently fell below twenty, although it was really marvellous how those old destroyers stuck to the work to the eternal credit of their crews, and particularly the engineering staffs. The adoption of the system, however, resulted during the comparatively fine summer weather in a considerable reduction in the number of merchant ships lost, in spite of the fact that great difficulty was experienced in keeping the ships of the convoys together, particularly at night, dawn frequently finding the convoy very much scattered.

It became obvious, however, that with the approach of winter the old destroyers of the 30-knot class would have the greatest difficulty in facing the heavy weather, and very urgent representations were made by Sir Frederick Brock for their replacement by more modern vessels before the winter set in. All that could be effected in this direction was done, though at the expense of some of the Channel escorts. Urgent requests for good destroyers were being received at the Admiralty from every Command, and it was impossible to comply with them since the vessels were not in existence.

Certain other steps which may be enumerated were taken in connection with the Scandinavian traffic.

The convoys received such additional protection as could be given by the airships which were gradually being stationed on the East Coast during the year 1917, and decoy ships occasionally joined the convoys in order to invite submarine attack on themselves. This procedure was indeed adopted on all convoy routes as they were brought into being, the rule being for the decoy ship to drop behind the convoy in the guise of a straggler.

Some of our submarines were also detailed to work in the vicinity of convoy routes in order that they might take advantage of any opportunity to attack enemy submarines if sighted; due precautions for their safety were made.

Among the difficulties with which the very energetic and resourceful Admiral Commanding the Orkneys and Shetlands had to contend in his working of the convoys was the persistent mining of the approach to Lerwick Harbour by German submarines; a second difficulty was the great congestion that took place in that harbour as soon as bad weather set in during the autumn of 1917. The weather during the latter part of 1917 was exceptionally bad, and great congestion and consequent delay to shipping occurred both at Lerwick and in the Norwegian ports. As the result of this congestion it became necessary to increase largely the number of ships in each convoy, thereby enhancing the difficulty of handling the convoy.

At the commencement it had been decided to limit the size of a Scandinavian convoy to six or eight vessels, but as the congestion increased it became necessary to exceed this number considerably, occasional convoys composed of as many as thirty to forty ships being formed. A contributory cause to the increase in the size of convoys was due to the fact that the trade between Lerwick and the White Sea, which had been proceeding direct between those places during the first half of 1917, became the target of persistent submarine attack during the summer, and in order to afford them protection it was necessary in the autumn to include these ships also in the Scandinavian convoy for the passage across the North Sea. Between the coast of Norway and the White Sea they proceeded independently, hugging territorial waters as far as possible.

It will be realized that the institution of the convoy system of sailing for the Scandinavian trade necessitated an extensive organization on the Norwegian as well as on the British side of the North Sea. For this reason Captain Arthur Halsey, R.N., was appointed in March, 1917, as Naval Vice-Consul at Bergen, and the whole of the arrangements in regard to the working of the convoys, the issue of orders, etc., from the Norwegian side came under him and his staff, to which additions were made from time to time. The position was peculiar in that British naval officers were working in this manner in a neutral country, and it says much for the discretion and tact of Captain Halsey and his staff and the courtesy of the Norwegian Government officials that no difficulties occurred.

Steps were also taken to appoint officers at British ports for the work of controlling the mercantile traffic, and as the organization became perfected so the conditions gradually improved.

By the end of September the bad weather prevalent in the North Sea had caused great dislocation in the convoy system. Ships composing convoys became much scattered and arrived so late off Lerwick as to prevent them proceeding on their passage without entering harbour. Owing to the overcrowding of Lerwick Harbour the system of changing convoy escorts without entering harbour had been introduced, and the delays due to bad weather were causing great difficulties in this respect. The question of substituting the Tyne for Lerwick as the collecting port was first discussed at this period, but the objections to the Tyne as an assembly port were so strong as to prevent the adoption of the proposal.

The system of convoy outlined above continued in force from April to December, 1917, during which period some 6,000 vessels were convoyed between Norway and the Humber with a total loss of about seventy ships.

There was always the danger that Germany would attack the convoys by means of surface vessels. The safeguard against such attacks was the constant presence of forces from the Grand Fleet in the North Sea. In view of the fact, however, that the distance of the convoy routes from the Horn Reef was only between 300 and 350 miles, and that on a winter night this distance could almost be covered at a speed of 20 knots during the fourteen or fifteen hours of darkness that prevailed, it will be seen that unless the convoys were actually accompanied by a force sufficient to protect them against operations by surface vessels, there was undoubted risk of successful attack. It was not possible to forecast the class of vessels by which such an attack might be carried out or the strength of the attacking force. The German decision in this respect would naturally be governed by the value of the objective and by the risk to be run. Admiral Scheer in his book states that on one occasion, in April, 1918, the German battle-cruisers, supported by the battleships and the remainder of the High Sea Fleet, attempted such an attack, but found no convoy. It was always realized by us that an attack in great force might be made on the convoy, but such risk had to be accepted.

The movements of the ships of the Grand Fleet were a matter for the Commander-in-Chief, provided always that no definite orders were issued by the Admiralty or no warning of expected attack was given to the Commander-in-Chief, and, prior to the first attack on the Scandinavian convoy, no special force of cruisers or light cruisers accompanied the convoy to guard it against attack by surface vessels, although a strong deterrent to attack lay in the frequent presence of forces from the Grand Fleet to the southward of the convoy routes, which forces would seriously threaten the return of any raiding German vessels. As the enemy would naturally make the northward passage by night we could hardly expect to sight his ships on the outward trip.

The first attack took place at daylight on October 17. The convoy on this occasion consisted of twelve ships, two British, one Belgian, one Danish, five Norwegian and three Swedish, and was under the anti-submarine escort of the destroyers Mary Rose and Strongbow, and two trawlers, the Elsie and P. Fannon. At dawn, shortly after 6.0 A.M., two strange vessels were sighted to the southward, and were later recognized as German light cruisers. They were challenged, but replied by opening fire at about 6.15 A.M., disabling the Strongbow with the first salvo fired. The Mary Rose steamed gallantly at the enemy with the intention of attacking with torpedoes, but was sunk by gunfire before she could achieve her object. The enemy vessels then attacked the convoy, sinking all except the British and Belgian vessels, which escaped undamaged. The Strongbow, shelled at close range, returned the fire, using guns and torpedoes, but was completely overwhelmed by the guns of the light cruisers and sank at about 9.30 A.M. The trawler Elsie effected very fine rescue work amongst the survivors both from the Strongbow and ships of the convoy, whilst under fire, and both trawlers reached Lerwick. The enemy sheered off soon after 8.0 A.M. Most unfortunately neither the Strongbow nor the Mary Rose succeeded in getting a wireless signal through to our own vessels to report the presence of enemy ships, otherwise there can be little doubt that they would have been intercepted and sunk. We had in the North Sea, during the night before the attack and during the day of the attack, a particularly strong force of light cruisers comprising four or possibly five squadrons (a total of not less than sixteen vessels), all to the southward of the convoy route, and had the information of the attack come through from the destroyers, these vessels would have been informed at once and would have had an excellent chance of intercepting the enemy. The extreme difficulty of preventing the egress of raiders from the North Sea at night, even when so large a force is cruising, was well illustrated by this incident, although a little reflection on the wide area of water to be covered, together with a knowledge of the distance that the eye can cover on a dark night (some 200 to 300 yards), would show how very great are the chances in favour of evasion.

This disaster to the Scandinavian convoy was bound to bring into prominence the question of affording to it protection against future attacks by surface vessels, for necessarily the protection against surface vessels differed from that against submarines, a point which was sometimes overlooked by those who were unfamiliar with the demands of the two wars which were being waged—the one on the surface and the other under the surface. It was very difficult to furnish efficient protection against the surface form of attack from the resources of the Grand Fleet if the practice of running a daily convoy was continued, because it was impossible to forecast the strength or exact character—battle-cruisers, cruisers or destroyers—of the attack; and the first step was to reduce the number of convoys and to increase correspondingly the number of ships in each convoy. A telegram was sent to the Admiral Commanding the Orkneys and Shetlands on October 26 asking whether the convoys could be conveniently reduced to three per week. A reply was received on the 29th to the effect that the convoy could be run every third day under certain conditions; the important conditions were the use of the Tyne instead of the Hurnber as a collecting port, and the provision of eight extra trawlers and nine modern destroyers. Sir Frederick Brock stated that he was assuming cruiser protection to the convoys and that the details would need to be worked out before the change could be made. He suggested a conference. He was requested on October 31 to consult the Vice-Admiral Commanding East Coast of England as to the practicability of using the Tyne as a convoy collecting port. Meanwhile Sir F. Brock had prepared a scheme for giving effect to his proposals, and on November 5 he sent copies of this scheme to the Vice-Admiral Commanding East Coast of England and other officers concerned for their consideration.

In forwarding proposals to the Admiralty on November 22, the Commander-in-Chief of the Grand Fleet stated that the destroyers asked for could not be provided from the Grand Fleet. Amongst other reasons it was pointed out that the destroyers required for screening the light cruisers protecting the convoys would have to be supplied from that source, thus bringing an additional strain on the Grand Fleet flotillas. He suggested the provision of these vessels from other Commands, such as the Mediterranean, and pointed out the manifest advantages that would result from providing a force for this convoy work that would be additional to the Grand Fleet flotillas. Consideration of the proposals at the Admiralty showed once again the great difficulty of providing the destroyers. It was impossible to spare any from the Mediterranean, where large troop movements needing destroyer protection were in progress, and other Commands were equally unable to furnish them. Indeed, the demands for destroyers from all directions were as insistent as ever. The unsuitability of the Tyne as a collecting port was remarked upon by the Naval Staff, as well as other objections to the scheme as put forward from Scapa. In order to decide upon a workable scheme, directions were given that a conference was to assemble at Scapa on December 10. An officer from the Naval Staff was detailed to attend the conference, to point out the objections which had been raised and, amongst other matters, to bring to notice the advantage of the Firth of Forth as a collecting port instead of the Tyne.

Meanwhile steps had been taken to furnish as much protection as possible from Grand Fleet resources to the convoys against attack by enemy surface vessels.

The conference of December 10 came to the conclusion that the Firth of Forth was the best assembly place, and that the port of Methil in that locality would offer great advantages. The conference made recommendations as to the provision of destroyers as soon as they were available, and, amongst other matters, mentioned the necessity for an increase in the minesweeping force at Rosyth to meet a possible extension of enemy minelaying when the new system was in operation.

On December 12 a second attack on the convoy took place. In this instance the attack was carried out by four German destroyers. Two convoys were at sea, one east-bound and one west-bound, the east-bound convoy being attacked. It was screened against submarine attack by two destroyers—the Pellew and Partridge—and four armed trawlers, and comprised six vessels, one being British and the remainder neutrals. The attack took place in approximately Lat. 59.50 N., Long. 3.50 E., and the action resulted in the Partridge, the four trawlers, and the whole of the convoy being sunk, and the Pellew was so severely damaged as to be incapable of continuing the action. At the time of this attack a west-bound convoy was at sea to the westward of the other convoy, and two armoured cruisers—the Shannon and Minotaur—with four destroyers were acting as a covering force for the convoys against attack by surface vessels. A wireless signal from the Partridge having been intercepted, this force steamed at full speed for the scene of the action, the destroyers arriving in time to pick up 100 survivors from the convoy and trawlers, but not in time to save the convoy. The 3rd Light Cruiser Squadron, also at sea, was some 85 miles to the southward and eastward of the convoy when attacked, but neither this force nor the Shannon's force succeeded in intercepting the enemy before he reached port. The short hours of daylight greatly facilitated his escape.

On receipt of the report of the meeting of December 10, and in view of the attack of December 12, the question of the interval between convoys was specially considered in its relation to the ability of the Grand Fleet to furnish protection against surface attack. It was decided that for this reason it would only be possible to sail convoys from Methil every third day so as to avoid having two convoys at sea at a time, a situation with which the Grand Fleet could not deal satisfactorily. The organization then drawn up actually came into effect on January 20, 1918, after my departure from the Admiralty, and was continued with certain modifications to the end of the war. The principal modification was an increase of the interval between convoys, first, to four, and later to five days in order to relieve the strain on the Grand Fleet arising from the provision of covering forces; the disadvantage of the resultant increased size of the convoys had to be accepted. Under the new system the Commander-in-Chief Coast of Scotland at Rosyth—Admiral Sir Cecil Burney—became responsible for the control of the Scandinavian convoys, the Admiralty selecting the routes.

The introduction of the convoy system for the Atlantic trade dates from the early days of May, 1917, when the prospect—for it was only then a prospect—of increasing assistance from the U.S. Navy in regard to destroyers and other small craft for escort duty as well as convoy cruisers for ocean work, made the system possible. Action taken with the U.S. authorities for the introduction of a system by which the trade from that country in neutral shipping was controlled enabled the ships of the 10th Cruiser Squadron to be gradually withdrawn from blockade duties and utilized as ocean convoy cruisers. Even with assistance from the U.S. Navy in the shape of old battleships and cruisers, the use of the 10th Cruiser Squadron, the withdrawal of the 2nd Cruiser Squadron of five ships from the Grand Fleet, the use of the ships of the North American and West Indies Squadron and of some of our older battleships from the Mediterranean, there was still a shortage of convoy cruisers; this deficiency was made up by arming a number of the faster cargo vessels with 6-inch guns for duty as convoy cruisers. These vessels usually carried cargo themselves, so that no great loss of tonnage was involved.

On May 17 a committee was assembled at the Admiralty to draw up a complete organization for a general convoy system. (The committee was composed of the following officers: Captain H.W. Longden, R.N., Fleet Paymaster H.W.E. Manisty, R.N., Commander J.S. Wilde, R.N., Lieutenant G.E. Burton, R.N., and Mr. N.A. Leslie, of the Ministry of Shipping.) This committee had before it the experience of an experimental convoy which arrived from Gibraltar shortly after the commencement of the committee's work, as well as the experience already gained in the Scandinavian and French coal trade convoys, and the evidence of officers such as Captain R.G. Henderson, R.N., who had made a close study of the convoy question.

On June 6 the report was completed. This valuable report dealt with the whole organization needed for the institution of a complete system of convoy for homeward and outward trade in the Atlantic. In anticipation of the report steps had already been taken to commence the system, the first homeward bound Atlantic convoy starting on May 24. A necessary preliminary for the successful working of the convoys was a central organization at the Admiralty. This organization—termed the Convoy Section of the Trade Division of the Naval Staff—worked directly under Rear-Admiral A.L. Duff, who had recently been placed on the Board of Admiralty with the title of Assistant Chief of the Naval Staff (A.C.N.S.), and who was in immediate control of the Anti-Submarine, Trade and Minesweeping Divisions of the Staff. Fleet Paymaster H.W.E. Manisty was appointed as Organizing Manager of Convoys, and the Convoy Section, comprising at first some ten officers, soon increased to a total of fifteen, and was in immediate touch with the Ministry of Shipping through a representative, Mr. Leslie. His function was to make such arrangements as would ensure co-operation between the loading and discharging of cargoes and convoy requirements, and generally to coordinate shipping needs with convoy needs.

The organizing manager of the convoys and his staff controlled the assembly, etc., of all convoys and vessels.

The routing of the convoys and their protection, both ocean and anti-submarine, was arranged under the superintendence of the A.C.N.S.

In addition to the central Admiralty organization, an officer with the necessary staff was appointed to each convoy port of assembly at home and abroad. This officer's duties comprised the collection and organization of the convoy and the issue of sailing orders and necessary printed instructions to the masters of the vessels, seeing that they were properly equipped for sailing in company, and forwarding information to the Admiralty of the movements of the convoy.

An essential feature of the system was the appointment of a convoy commodore. This officer was quite distinct from the commanding officer of the vessel forming the ocean escort, but acted under his orders when in company. The duty of the convoy commodore, whose broad pennant was hoisted in one of the ships, was, subject to instructions from the commanding officer of the escorting vessel, to take general charge of the convoy.

The convoy commodores were either naval officers, admirals or captains on the active or retired lists, or experienced merchant captains. The duties were most arduous and responsible, but there was no lack of volunteers for this work. Many of the convoy commodores had their ships sunk under them. The country has every reason for much gratitude to those who undertook this difficult and very responsible task.

By July we had succeeded in increasing the strength of the anti-submarine convoy escorting force to thirty-three destroyers (eleven of which belonged to the United States Navy) and ten sloops, with eleven more destroyers for the screening of troop transports through the submarine zone and for the protection of the convoys eastward from the Lizard, the position in which the other screening force left them. We had remaining twelve sloops, which, with trawlers, were engaged in protecting that considerable portion of the trade making for the south of Ireland, which we could not yet bring under convoy. It was intended to absorb these sloops for convoy protection as soon as circumstances permitted.

At this stage it was considered that a total of thirty-three more destroyers or sloops was needed to complete the homeward convoy system. The Admiralty was pressed to weaken yet further the Grand Fleet destroyer force in order to extend the convoy system, but did not consider such a course justified in view of the general naval situation.

In arranging the organization of the Atlantic convoy system it was necessary to take into consideration certain other important matters. Amongst these were the following:

1. The selection of ports of assembly and frequency of sailing. During the latter half of 1917 the general arrangements were as follows for the homeward trade:

Each port served a certain area of trade, and vessels engaged in that trade met at the port of assembly for convoy to the United Kingdom or to France.

The total number of merchant ships sailing thus in convoy every eight days in September, 1917, was about 150, in convoys comprising from 12 to 30 ships, and the total escorting forces comprised:

in addition to a considerable force for local escort near Gibraltar, consisting of sloops, yachts, torpedo boats, U.S. revenue cruisers, U.S. tugs, etc.

At this period (September, 1917) outward convoys were also in operation, the arrangement being that the outward convoy was escorted by destroyers or sloops to a position 300 to 400 miles from the coast clear of the known submarine area, and there dispersed to proceed independently, there being insufficient ocean escort vessels to take the convoy on; about twelve more were needed for this work. The escorting vessels used for the outward convoys were destroyers or sloops which were due to proceed to sea to meet a homeward convoy, the routine being that the outward convoy should sail at such a time as would ensure the homeward convoy being met by the escort without undue delay at the rendezvous, since any long period of waiting about at a rendezvous was impossible for the escorting vessels as they would have run short of fuel. It was also undesirable, as it revealed to any submarine in the neighbourhood the approach of a convoy.

It will be realized by seamen that this procedure (which was forced upon us by the shortage of escorting vessels) led to many difficulties. In the first place the homeward convoys were frequently delayed by bad weather, etc., on passage across the Atlantic, and, owing to the insufficient range of the wireless installations, it was often not possible for the commodore to acquaint the Admiralty of this delay in time to stop the sailing of the outward convoys. Again, outward convoys were often delayed by bad weather, resulting in the homeward convoy not being met before entering the submarine zone. As the winter drew near this was a source of constant anxiety, since so many of the vessels outward bound were in ballast (empty), and their speed was consequently quickly reduced in bad weather. The ships under these conditions became in some cases almost unmanageable in a convoy, and the responsibilities of the escorts were much intensified.

In September, 1917, the following was the position in respect to outward bound convoys:

About 150 vessels sailed every eight days in convoys varying in strength from 12 to 30 ships.

There was still a good deal of Atlantic trade that was not sailing under convoy. This comprised trade between Gibraltar and North and South America, between the Cape, South America and Dakar, and the coastal trade between North and South America. It was estimated that an additional twenty-five to thirty ocean escorts and eleven destroyers would be needed to include the above trade in convoy.

The Mediterranean trade is dealt with later.

The question of speed was naturally one of great importance in the convoy system. As has been stated earlier, the speed of a convoy like that of a squadron or fleet is necessarily that of the slowest ship, and in order to prevent delay to shipping, which was equivalent to serious loss of its carrying power, it was very necessary that convoys should be composed of ships of approximately the same speed. In order to achieve this careful organization was needed, and the matter was not made easier by the uncertainty that frequently prevailed as to the actual sea speed of particular merchant ships. Some masters, no doubt from legitimate pride in their vessels, credited them with speeds in excess of those actually attained. Frequently coal of poor quality or the fact that a ship had a dirty bottom reduced her speed to a very appreciable extent, and convoy commodores had occasionally to direct ships under such conditions to drop out of the convoy altogether and make their passage alone. Obviously this action was not taken lightly owing to the risk involved. Decision as to the sea speed of convoys was taken by the convoy officer at the collecting port, and he based this on the result of an examination of the records in the different ships. As a rule convoys were classed as "slow" and "fast." Slow convoys comprised vessels of a speed between 8 and 12-½ knots. Fast convoys included ships with a speed between 12-½ and 16 knots. Ships of higher speed than 16 knots did not as a rule sail in convoys, but trusted to their speed and dark hours for protection in the submarine area. The Gibraltar convoy (an exception to the general rule) contained ships of only 7 knots speed.

With the introduction of convoys the provision of efficient signal arrangements became a matter of importance. The issue of printed instructions to each master and the custom introduced of assembling the masters to meet the captain of the escorting cruiser before sailing, so that the conduct of the convoy might be explained, had the effect of reducing signalling to a minimum, but it was necessary that each ship should have a signalman on board, and the provision of the number of signalmen required was no easy matter. A good wireless installation was essential in the escorting cruiser and in the Commodore's ship in order that the course of the convoy could be diverted by the Admiralty if the known or suspected presence of submarines rendered it necessary, and also for the purpose of giving to the Admiralty early information of the position of a convoy approaching the coast, so that the escorting destroyers could be dispatched in time.

Fortunately for us, German submarines constantly used their wireless installations when operating at sea, and as a consequence our wireless directional stations were able to fix their positions by cross bearings. This practice on the part of the enemy undoubtedly went far to assist us both in anti-submarine measures and in diverting trade to a safe course.

The introduction of the convoy system rendered the provision of anti-submarine protection at ports of assembly a matter of great importance, owing to the very large number of vessels that were collected in them. Some of the ports were already in possession of these defences, but amongst those for which net protection was prepared and laid during 1917 were Halifax, Sydney (Cape Breton), Falmouth, Lamlash, Rosslare (on the south-east coast of Ireland), Milford Haven, Sierra Leone and Dakar. This involved extensive work, and was undertaken and carried out with great rapidity by Captain F.C. Learmonth and his staff, whose work in the production of net defences during the war was of inestimable value, not only to ourselves, but to our Allies, for whom large supplies of net defences were also provided. The U.S.A. also adopted our system of net defence for their harbours on entry into the war. Many anxious months were passed at the Admiralty and at the ports named until the anti-submarine defences were completed.

The escort of the convoys through the submarine zone imposed very heavy work upon the destroyers, sloops and other screening vessels. This was due partly to the fact that there were not sufficient vessels to admit of adequate time being spent in harbour to rest the crews and effect necessary repairs, and partly to the nature of the work itself and the weather conditions under which so much of it was carried out. It will be realized by those who have been at sea in these small craft that little rest was obtainable in the Atlantic between the west coast of Ireland and the mouth of the Channel and positions 800 to 400 miles to the westward, except in the finest weather. When to this is added the constant strain imposed by watching for the momentary appearance of a periscope or the track of a torpedo, and the vigilance needed, especially on dark and stormy nights, to keep touch with a large convoy of merchant ships showing no lights, with the inevitable whipping up of occasional stragglers from the convoy, some idea may be gathered of the arduous and unceasing work accomplished by the anti-submarine escorts.

It had been my practice during 1917 to call for returns from all commands of the number of hours that vessels of the destroyer and light cruiser type were actually under way per month, and these returns showed how heavy was the strain on the destroyers, particularly those engaged in convoy work.

For several months, for instance, the destroyers in the flotillas stationed at Devonport were under way on an average for just under 50 per cent. of the month.

This meant that several destroyers in these flotillas averaged quite 60 per cent. or even 70 per cent. of their time under way, as other vessels of the flotilla were laid up during the periods under review for long refits due to collision or other damage, in addition to the necessary four-monthly refit.

Anyone familiar with the delicate nature of the machinery of destroyers—which needs constant attention—and the conditions of life at sea in them will appreciate the significance of these figures and the strain which the conditions imposed on those on board as well as on the machinery.

It was evident in November, 1917, that the personnel and the machinery, whilst standing the strain in a wonderful manner, were approaching the limit of endurance, and anxiety was felt as to the situation during the winter.

Reports came in from the Grand Fleet indicating that the work of the destroyers engaged in protecting the ships of the Scandinavian convoy was telling heavily on the personnel, particularly on the commanding officers, and one report stated that the convoy work produced far greater strain than any other duty carried out by destroyers. No mean proportion of the officers were suffering from a breakdown in health, and since the whole of the work of the Devonport, Queenstown and North of Ireland flotillas consisted of convoy duty, whilst only a portion of the Grand Fleet destroyers was engaged in this work, the opinions expressed were very disquieting in their relation to the work of the southern flotillas.

However, the destroyers held on here as elsewhere, but it is only just to the splendid endurance of the young officers and the men who manned them to emphasize as strongly as I can the magnificent work they carried out in the face of every difficulty, and without even the incentive of the prospect of a fight with a foe that could be seen, this being the compensation given in their work to the gallant personnel of the Dover, Harwich and Grand Fleet flotillas. The convoy flotillas knew that their only chance of action was with a submarine submerged, a form of warfare in which the result was so very frequently unknown and therefore unsatisfactory.

Under the new conditions the Admiralty took upon itself responsibility for the control of the ships of the Mercantile Marine in addition to its control of the movements of the Fleet. Indeed the control of convoys was even more directly under the Admiralty than was the control of the Fleet. In the latter case the proper system is for the Admiralty to indicate to the Commander-in-Chief, Grand Fleet, or to other Commands the objective, and to supply all the information possible regarding the strength of the enemy, his intentions and movements and such other information as can be of use to the Commander-in-Chief, but to leave the handling of the force to the Commander-in-Chief concerned. This is the course which was usually followed during the late war. It was my invariable practice when at the Admiralty.

In the case of convoys, however, a different system was necessary owing to the difficulty of transmitting information, the great delay that would be caused were this attempted, and the impossibility of control being exercised over all convoys at sea except by the Admiralty. Consequently the actual movements of convoys for the greater part of their passage were directed by the Naval Staff. Owing to ships not showing lights at night, convoys were diverted clear of one another by wireless signal if they were getting into dangerous proximity; they were directed to alter course as necessary to avoid areas in which submarines had been located, and occasionally it became necessary to alter the destination of some ships as they approached home waters. The movements of all convoys were "plotted" from day to day, indeed from hour to hour, on a large-scale chart at the Admiralty, and it was easy to see at a glance the position of all the ships at any given time.

As the convoy approached home waters the ships came within the areas of the Commanders-in-Chief, Coast of Ireland, Devonport, and Portsmouth, and the Vice-Admiral Commanding the Dover Patrol, and were taken in charge by one or other of them. At each port a staff existed which kept a constant record of the movements of ships passing through or working in the Command, and enabled the Commander-in-Chief to take instant action if occasion arose.

The success of the convoy system in protecting trade is best shown by the figures relating to the year 1917 on the succeeding page (p. 144). In considering these figures the loose station-keeping of the ships in the Scandinavian convoy must be borne in mind. A large proportion of the ships in this convoy were neutrals, and it was naturally not possible to bring these vessels under discipline as was the case with convoys composed of purely British ships. Consequently there was much straggling, and the losses were proportionately heavier than in most of the Atlantic convoys. The comparatively heavy losses in the Gibraltar convoys were probably due to these convoys traversing two dangerous submarine zones. The extraordinary immunity of the French coal trade convoy from serious losses is remarkable and is probably due to the short passage which enabled most of the distance to be traversed at night and to the ships being of light draught.

The table on the following page would not be complete were no reference made to the heavy losses which were experienced during the year amongst ships which were unescorted through the danger zones, owing to the fact that no escorting vessels were available for the work.

In the Dakar convoy at the end of November and in the Halifax convoy 150 ships had been brought home without loss, whilst in the Sierra Leone convoy 1 ship had been lost out of 90 convoyed.

In the convoys starting from Queenstown 180 ships had been sent out up to the end of November without loss.

There were naturally loud complaints of these losses, but these were inevitable in the absence of escorting vessels, and no one realized the dangers run more than those responsible for finding protection; every available vessel was not only working at highest possible pressure, but, as has been mentioned, breakdowns from overwork amongst escorting craft were causing very considerable anxiety.

The following figures show the dangers which were run by unescorted vessels:

In considering the above table it should be pointed out that a large proportion of the losses shown under the heading "Ships unescorted" took place amongst ships which had either dispersed from a convoy or which were on their way to join up with a convoy at the port of assembly. It was unfortunately quite impossible to provide escorts for all ships either to their ports of discharge or from their loading ports to the ports of assembly for the convoy, as we had so few vessels available for this work. Thus, in the month of November, 1917, out of 13 vessels engaged in the main oversea trade that were lost, 6 were in convoy, 5 had left or had not joined their convoy, and 2 were not joining a convoy and were unescorted.

November was the month of smallest British losses during the period of unrestricted warfare in 1917, and it is of interest to examine the losses for that month. The total number of ships lost was 51. As many as 1,197 vessels entered or left home waters in overseas trade exclusive of the Mediterranean trade. Of this aggregate 87.5 per cent, were in convoy, and the total number of these vessels sunk (13) was divided amongst the following trades: North America, 1; Gibraltar, 5; West Africa and South America, 1; the Bay of Biscay, Portugal and Spanish ports west of Gibraltar, 5; Scandinavian, 1. In the same month there were 2,159 cross-Channel sailings and ten losses, nine of these vessels being unescorted.

Particulars of the locality of the total British losses of 51 ships for the month of November are as follows:

In order to give some idea of the great volume of traffic on the East Coast and the consequent difficulty of affording proper protection, it may be mentioned that in the month of October, 1917, the number of vessels passing between Spurn Head (River Humber) and St. Abb's Head (to the northward) was 740 going north and 920 going south. Of this total only 223 of the northward—and 413 of the southward-bound vessels were in convoy or under escort, the total losses being eleven, all amongst the unaccompanied ships.

Mention should be made here of the very serious situation which arose during the year 1917 owing to the success attending the attacks by enemy submarines on oil tankers bringing oil fuel to the United Kingdom for the use of the Fleet. A great many of these tank vessels were of great length and slow speed and presented the easiest of targets to the torpedo attack of a submerged submarine. So many vessels were sunk that our reserve of oil fuel became perilously low. Instead of a reserve of some five or six months we were gradually reduced to one of about eight weeks, and in order to economize expenditure of fuel it actually became necessary at one time to issue directions that the speed of oil-burning warships was to be limited except in cases of the greatest urgency. Such an order in war was a matter of much gravity; the great majority of our light cruisers and destroyers were fitted to burn oil fuel only, as well as our latest and most powerful battleships. The crisis was eventually overcome by drawing upon every source (including the Grand Fleet) for destroyers to escort the tankers through the submarine danger areas, and by the assistance given us by the Ministry of Shipping in bringing supplies of oil fuel to this country in the double bottoms of merchant ships. By the end of 1917 the situation had greatly improved.

The losses of shipping during 1917 were particularly heavy in the Mediterranean. Apart from the fact that the narrow waters of that sea render difficult a policy of evasion on the part of merchant shipping and give great advantages to the submarine, it was thought that the heavy losses in the early part of the year were partly due to the method of routeing the ships then in force, and in reply to representations made to the French Admiralty this system was altered by the French Commander-in-Chief. It should be noted that the Mediterranean outside the Adriatic was under French naval control in accordance with the agreement entered into with France and Italy. The cordial co-operation of the French Admiralty with us, and the manner in which our proposals were met, form very pleasant memories of my term of office at the Admiralty. During the greater part of the year 1917 Admiral Lacaze was Minister of Marine, whilst Admiral de Bon held office as Chief of the Naval Staff during the whole year. Nothing could exceed the courtesy extended to me by these distinguished officers, for whom I conceived great admiration and respect.

The result of the altered arrangement was a decided but temporary improvement, and the losses again became serious during the summer months. I then deemed it desirable that the control of the traffic should be placed in the hands of officers stationed at Malta, this being a central position from which any necessary change in the arrangements could be made more rapidly and with greater facility than by the French Commander-in-Chief, who was also controlling fleet movements and who, for this reason alone, was not in a position to act quickly.

A unified command in the Mediterranean would undoubtedly have been the most satisfactory and efficient system to adopt, but the time was not ripe for proposing that solution in 1917, and the alternative was adopted of British control of the traffic routes throughout the whole Mediterranean Sea subject to the general charge of the French Commander-in-Chief which was necessary in such an eventuality arising as an attempted "break out" of the Austrian Fleet.

Accordingly, with the consent of the French and Italian Admiralties, Vice-Admiral the Hon. Sir Somerset Gough-Calthorpe, K.C.B., was dispatched to the Mediterranean as British Commander-in-Chief; he was in control generally of all British Naval forces in the Mediterranean, and especially in charge of all the arrangements for the protection of trade and for anti-submarine operations, the patrol vessels of all the nationalities concerned being placed under his immediate orders for the purpose, whilst the whole of the Mediterranean remained under the general control of Vice-Admiral Gauchet, the French Commander-in-Chief. Admiral Calthorpe was assisted by French and Italian officers, and the Japanese Government, which had previously dispatched twelve destroyers to the Mediterranean to assist in the protection of trade, also gave to Admiral Calthorpe the control of these vessels.

In the requests which we addressed to the Japanese Admiralty I always received great assistance from Admiral Funakoshi, the Naval AttachÉ in London. His co-operation was of a close and most cordial nature.

The services of the Japanese destroyers in the Mediterranean were of considerable value to the Allied cause. A striking instance of the seamanlike and gallant conduct of their officers and men was furnished on the occasion of the torpedoing of a British transport by an enemy submarine off the coast of Italy, when by the work of the Japanese escorting destroyers the great majority of those on board were saved.

Admiral Calthorpe on leaving England was charged with the duty of organizing convoys in the Mediterranean on the lines of those already in force in other waters as soon as the necessary vessels were available, and a conference of Allied officers sat at Malta soon after his arrival, when a definite scheme of convoy was prepared. There had always, however, been a great scarcity of fast patrol vessels in the Mediterranean for this work. Divided control of the forces in that area was partly responsible for this. The Austrian destroyers were considered by the Italian Admiralty to be so serious a menace in the Adriatic as to render it necessary to keep in that sea the great majority of the Italian destroyers as well as several French vessels of this class. The situation at the eastern end of the Mediterranean necessitated a force of some eight British destroyers being kept in the Aegean Sea to deal with any Turkish vessels that might attempt to force the blockade of the Dardanelles, whilst operations on the Syrian coast engaged the services of some French and British destroyers. Continual troop movements in the Mediterranean also absorbed the sendees of a considerable number of vessels of this type.

Consequently there was a great shortage of fast small craft for escort and mercantile convoy work. It was estimated that the escort force required for the protection of a complete system of convoy in the Mediterranean was approximately 290 vessels, the total number available being about 215.

In spite, then, of the success of Admiral Calthorpe's work, the result was that convoys were not started in the Mediterranean until October, and they were then but inadequately protected, and losses were heavy, both from this cause and from the fact already mentioned—that the Mediterranean is a sea which, by reason of its confined nature, is particularly suited for operations by submarines against trade. Its narrowness at various points, such as the Straits of Gibraltar, the Malta Channel, the Straits of Messina, and the passages to the Ægean cause such convergence of trade as to make it a very simple matter for a submarine to operate with success. Evasion by change of route is almost impossible. Operations designed to prevent the exit of submarines from the Adriatic were difficult, because the depth of water in the Straits of Otranto militated against the adoption of effective mining and the laying of an effective net barrage.

For the above reasons the Admiralty was always very averse to the sending of a large volume of our Far Eastern trade through the Mediterranean, and strongly urged the Cape route instead; but the shortage of shipping, combined with the increased length of the Cape route, influenced the Ministry of Shipping to press strongly for the Mediterranean as opposed to the other route. A "through" convoy from England to Port Said was started in October, and by the end of November two ships had been sunk out of the thirty-five that had been under convoy. The return convoy; Port Said to England, was only started in December.

The losses of British merchant steamships per quarter in the Mediterranean during 1917 is shown below:

It is impossible to close this chapter describing the convoys without mention being made of the fine work accomplished by those upon whose shoulders fell the task of organizing and working the whole system. I cannot hope that I have succeeded in conveying to readers of this volume an adequate conception of the great and marvellously successful performance that it was or a full appreciation of what immense difficulties the staff had to contend with. They were very completely realized by me, who saw them appear day by day and disappear under treatment.

The head of the organization was, of course, Rear-Admiral A.L. Duff, the member of the Board and Staff immediately responsible also for the whole anti-submarine organization. Only those who witnessed Admiral Duff's work at the Admiralty during 1917 can realize the immense debt that the country owes to his untiring ability, patience, energy and resource. Capt. H.G. Henderson, who had been associated with the convoy system from its start, was an invaluable assistant, as also was Commander I.W. Carrington. Capt. Richard Webb, the Director of the Trade Division, and Capt. Frederic A. Whitehead, the Director of the Mercantile Movements Division, took an important share in the work of organization, whilst the work of Convoy Manager was carried through with quite exceptional skill by Paymaster-Commander H.W.E. Manisty. These officers were assisted by most capable staffs, and the Ministry of Shipping, without whose assistance the work could not possibly have been successfully carried out, co-operated most cordially.