The prime minister; Hankey; Sir Edward Carson, the first lord; Jellicoe; and Duff discussed these ideas at breakfast on the 13th, but there were no immediate results. Lloyd George was probably preoccupied with the press of other business and perhaps not yet ready for a serious dispute with the Admiralty. The horrendous losses of April changed the situation. On 25 April the War Cabinet concluded that the prime minister should investigate all the means used at present in antisubmarine warfare on the grounds there was not sufficient coordination in present efforts. The Admiralty moved in a similar direction, and on 26 April Admiral Duff, head of the Anti-Submarine Division, submitted a minute in favor of convoying all vessels bound from the North and South Atlantic to the United Kingdom. Jellicoe approved the following day and also recommended to the first lord that the British expeditionary force at Salonika be withdrawn. This would reduce the need for escorts in the Mediterranean and would be an important means of providing sufficient escorts for convoys. The crucial decision concerning convoys had therefore been made before Lloyd George made his celebrated visit to the Admiralty on 30 April, and was not, as is sometimes stated, something he forced on a reluctant Admiralty that day.82
The first trial convoy sailed from Gibraltar the evening of 10 May. It consisted of 16 ships arranged in three columns escorted by the Q-ships Mavis and Rule. Three armed yachts escorted the convoy through the danger zone to 11°W. The station keeping about which the Admiralty had such anxiety was reported as “on the whole satisfactory.” There were some problems; the convoy turned up 20 miles west of its rendezvous with the destroyers sent from Devonport and their meeting was delayed eight hours. There were, however, no submarine attacks. The five ships bound for the west coast were detached off the Scillies with an escort of two destroyers. The remainder of the convoy proceeded to Plymouth on the 20th and then proceeded up the Channel in three divisions, escorted by twenty-four drifters that met them off Poole. The convoy reached the Downs safely on the 22d. The naval staff study declared: “The trial had been an entire success, and from that moment it may be said that the submarine menace was conquered.”83
This may have been true in the long run, but the convoy system took time to establish, and things were not always so clear to those involved. Sims became a hearty supporter of the concept, but Admiral Benson and the U.S. Navy Department remained skeptical, and the Americans denied an Admiralty request to form a convoy of 16 to 20 ships sailing from Hampton Roads to be escorted by one of the groups of American destroyers on their way to Queenstown. The Navy Department did not consider such a large number of ships in convoy to be practicable and recommended groups of no more than four. The destroyers sailed separately. Almost a month later, Sims received a cable from the Navy Department stating that they considered American vessels having armed guards were safer when sailing independently than when in convoy.84
The U.S. Navy Department was slow to learn. The first trial convoy of twelve ships sailed from Hampton Roads on 24 May escorted by the British armored cruiser Roxburgh. On the evening of 6 June on the other side of the ocean it was met at 50° N, 16° W by eight destroyers from Devonport. The convoy separated into west and east coast sections south of Ireland, and by the 10th all ships had arrived safely at their destinations. Unfortunately two ships had been compelled to drop out because they were unable to maintain the required speed, and one of them was torpedoed and sunk.85 As for the convoy as a whole, the Roxburgh’s captain reported the merchant captains had been attentive to signals, kept good station, and zigzagged in a satisfactory fashion, and he would be prepared to escort as many as thirty ships instead of twelve.86
The convoy system grew out of the deliberations of a convoy committee nominated on 15 May to devise a program. The committee’s report of 6 June resulted in the formation of a separate Convoy Section at the Admiralty with Fleet Paymaster H. W. E. Manisty appointed as “organising manager of convoys.” The Convoy Section would work in close liaison with the Ministry of Shipping and with the Intelligence Division. The latter was particularly important, and there was now some loosening of the great secrecy that had surrounded the work of Room 40 and that may have cost the British a decisive victory the morning after Jutland the year before. Previously, information derived from Room 40’s intercepts could not even be plotted on the charts of the enemy submarine section of the Intelligence Division. Now, however, a pneumatic tube connected the enemy submarine section and direction-finding sections with the convoy section, and the information derived from intercepts and directionals could be plotted on the convoy chart so that the latest intelligence on the location of submarines could be compared to the convoy track and, if necessary, the convoy could be diverted by means of a wireless message to the convoy commodore’s ship. All convoy commodores were now in ships equipped with wireless.
The use of wireless for diversions was a great improvement over the routing systems used in the past. In the judgment of the naval staff study, “Part of the undoubted efficacy of convoy lay in the simplicity of ‘Carrington’s Chart.’” The latter referred to Commander John Carrington, who kept the 7-foot-high convoy chart hung on the west wall of the drawing room of Admiralty House. Three navigating lieutenants kept a constant watch on it, using red flags to indicate today’s positions, blue flags yesterday’s, and grey flags the day before yesterday’s. A thread of red silk represented the position of each convoy.87
The system quickly showed results. There were five HH convoys from Hampton Roads in June, each with a single armored cruiser or armed merchant cruiser as ocean escort until the rendezvous with the Devonport destroyers. A total of 71 ships representing 363,170 tons arrived safely. Only one ship was torpedoed in the Channel, and she was salved. No ships were lost.88 The Admiralty had been forced to institute the HH convoys from Hampton Roads and run successive convoys in June because the stocks of oil fuel for the Grand Fleet had fallen below requirements. There was as a result a pressing need to protect oilers from North America and include as many oilers as possible in these convoys.89
Admiral Sims by mid-June had become an eloquent partisan of the convoy system, but the U.S. Navy Department remained hesitant and still regarded arming merchantmen with adequate guns and trained crews a suitable answer to the submarine. Sims countered that this would merely force the U-boats to use torpedoes to sink ships.90 There would always be a distinct difference between the British and Americans: the British were primarily concerned with protecting the flow of supplies to the United Kingdom; the American emphasis would be on protecting the transport of American troops to France. Josephus Daniels, secretary of the navy, informed Sims at the beginning of July that the “paramount duty” of American destroyers in European waters was the protection of American troop transports and that “everything is secondary” to having a sufficient number of escorts to protect those troops.91 This position had the potential to become a major controversy in 1918 when American troops were brought across the Atlantic in large numbers. The issue was still minor in July 1917. The Allies had yet to survive the submarine onslaught.
The success of these experimental convoys led the Admiralty on 15 June to institute a regular system of HH convoys sailing from Hampton Roads every four days, bound alternately for the west or east coast of England. The east-coast convoys were met by escorts from Devonport and proceeded around the south coast, whereas the west-coast convoys were met by escorts from Buncrana and proceeded around the north of Ireland. The need for fuel oil was so pressing that whatever their original destination all oilers, if ready when a convoy departed, were included in that convoy. On 22 June and 6 July, respectively, the Admiralty ordered the system extended to vessels sailing from Canadian ports and North American ports, with New York as the point of assembly. The HS convoys sailed from Sydney at intervals of eight days, except from November 1917 to July 1918 when, with the closure of the St. Lawrence in the winter months, the port of assembly was altered to Halifax. The New York convoys were designated HN. Both HS and HN convoys alternated between
west and east coasts, although eventually the HN convoys sailed only to the east coast. The U.S. Navy provided cruisers as ocean escorts for the HN convoys, and the American destroyers at Queenstown were switched from their sterile patrols to meet and escort HS and HN convoys bound for the east coast. The other North Atlantic convoys had as ocean escorts either British cruisers or armed merchant cruisers drawn from the Tenth Cruiser Squadron. They were supplemented by commissioned escort ships, which were merchantmen carrying cargo but fitted with three or four 6-inch guns. There was a retired flag officer in each commissioned escort ship, and when in the company of a convoy they wore the white ensign—that is, naval flag—and had the status of a warship.
On 26 July Gibraltar was brought into the system when the first HG convoy sailed, to be followed by regular convoys alternating between the west and east coast at four-day intervals. The escorts that met the HG convoys came from either Falmouth, for the east coast, or Milford Haven for the west coast and were usually trawlers rather than destroyers. The ocean escorts of the Gibraltar convoys were also a mixed bag. They were Q-ships at first, and later small American light cruisers, U.S. Coast Guard revenue cutters, old American gunboats, and, in 1918, a pair of small British light cruisers. The last of the initial system, the South Atlantic convoys, were ordered on 31 July. The HL, or fast, convoys (10 knots or more) assembled at Sierra Leone, whereas the HD, or slow, convoys would assemble at Dakar. The ocean escorts were armed merchant cruisers from the Ninth Cruiser Squadron (West Coast of Africa Station) supplemented by the Tenth Cruiser Squadron.
The system was modified and refined as the war went on. In August 1917 the HX convoys—fast (12½ knots or more) combined troopship and merchant ship convoys for Canadian troops—were introduced using Halifax as the port of assembly. In October and November, the HE and OE “through Mediterranean” convoys were introduced (see chapter 12). The need to transport large numbers of American troops caused further changes in 1918. The minimum speed for ships in the HX convoys was raised to 13 knots (with the ability to steam 13½ knots through the danger zone), and their port of assembly was shifted from Halifax to New York. In return a series of intermediate speed (11½ knot) convoys—designated HC—were instituted from Halifax. Troop convoys such as the HX and HC had either an armored cruiser or old predreadnought battleship as ocean escort. The Americans also reinforced them with a U.S. predreadnought and destroyer. The need to supply American troops in France also led in April 1918 to the introduction of HB convoys between New York and French ports in the Bay of Biscay. French armored cruisers contributed to the ocean escort of these convoys.
It should be noted, however, that with all convoys it was invariably much easier to find ocean escorts than it was to provide the destroyers and other small craft to bring them through the submarine danger zone as they approached the British Isles. Destroyers were the limiting factor, and it is hardly surprising the naval staffs came to consider them as worth their weight in gold. When the needs of the Grand Fleet were added to these considerations, it is little wonder Jellicoe informed the war cabinet in November 1917, “It is not too much to say that the whole of our naval policy is necessarily governed by the adequacy of our destroyer forces.”92
The size of convoys steadily increased as fears associated with station keeping and the “too many eggs in one basket” argument diminished and confidence in the system grew. The first HH convoy from Hampton Roads had been limited to 12 ships. The limit was increased to 20 by the end of June 1917. In September the Admiralty required special permission for convoys larger than 26 ships, but the limits expanded and soon the Admiralty insisted on special permission only for convoys larger than 36. They tried to keep this as a maximum for convoys to the west coast, which went north around Ireland and into the Irish Channel. The largest Atlantic convoy, HN.73 in June 1918, had 47 ships and was reported to be completely successful.93
The convoy system was extended gradually, and the initial emphasis was on homeward-bound vessels. These were the ships carrying vital supplies; the U-boats also had concentrated their attention on them. In the disastrous month of April 1917, 18 percent of homeward-bound ships had been sunk, compared to only 7 percent of those outward bound. This changed after the introduction of convoys. The U-boat commanders now found convoys hard to find and difficult to attack. The presence of escorts worked against attack on the surface by gunfire, and with the convoy zigzagging it proved difficult to get into a favorable firing position, particularly as those same escorts made maneuvering on the surface dangerous.94 The outward-bound ships remained, however, much easier targets, and before long losses of outward-bound ships rose steadily. It was obvious that a ship sunk outward bound was as much of a reduction in available tonnage as the loss of a ship bringing cargo to the United Kingdom. On 11 August 1917, the Admiralty arranged for convoys to take outward-bound ships through the submarine danger zone. There were eventually a series of outward convoys that matched the homeward convoys.
But where would the escorts for these outward convoys come from? The Admiralty tried to solve the problem by making the escorts do double duty. That is, the escorts took the outward convoy through the submarine danger zone and at the dispersal point parted company with the convoy, generally at dusk or after dark. They then steamed during the night to a rendezvous with a homeward convoy. Each H convoy therefore had a corresponding O convoy. These arrangements called for precise schedules and certainly put a heavy load on the escorts, which instead of merely steaming to meet the homeward convoy now had to spend about two days zigzagging with the outward convoy before the rendezvous. The senior naval officers of the destroyer escort were instructed that in case of delay with the outward convoy, the governing factor must always be meeting the homeward convoy at the correct time and place. It should also be noted that the outward convoys were only convoys through the submarine danger zone. The ships were then dispersed. This was modified in October 1917 when the Falmouth (OF) and Milford (OM) convoys, the east and west coast branches of the Gibraltar convoys, were to be kept together all the way to Gibraltar with any ships bound for the North or South Atlantic being detached en route. The OE/HE “through Mediterranean” convoys were also, as their name implies, kept together after their introduction in the fall of 1917.95
The statistics compiled by the Technical History Section at the Admiralty after the war demonstrated that the system worked. The percentage of loss of ships in the homeward Atlantic convoys had been reduced to 1.43 in the weekly return of 8 September 1917, and the loss in the outward convoys had fallen to .88 percent. By the end of the war, there had been 9,250 ships convoyed safely in homeward Atlantic convoys and 104 sunk, giving a loss of 1.11 percent. There had been 7,289 ships convoyed safely in outward convoys and only 50 sunk, for a loss of .68 percent. The combined homeward and outward Atlantic convoy loss was .92 percent, and the preceding figures include 16 ships sunk by marine peril and 36 ships sunk when not actually in contact with the convoy.96
These figures refer to convoys, but convoys did not operate everywhere, nor were all ships in convoys. There were also areas where losses were heavier than average, particularly, as we shall see in the following chapter, in the Mediterranean. The German bid for victory was based on the destruction of tonnage, that is, cargo-carrying capacity. Individual ships did not really matter, the tonnage they represented did. How then did the introduction of the convoy system show up in the worldwide shipping losses and in the losses among U-boats themselves? The amount of tonnage sunk throughout the world by submarines had fallen from a peak of 860,334 in April to a still-high 696,725 in June. With convoys in operation, worldwide sinkings fell to 555,514 tons in July, 472,372 tons in August, and 353,602 tons in September. Losses rose to 466,542 tons in October, only to fall to 302,599 tons in November, ending with 411,766 tons in December.97 The monthly success of U-boats in the last five months of 1917 was therefore well below their success in the first five months after the introduction of unrestricted submarine warfare.
&
nbsp; As for U-boat losses, there were 6 in all theaters in July, 5 in August, 10 or 11 (it is not clear whether one of the U-boats was sunk in September or October) in September, 5 or 6 in October, 8 in November, and 8 in December. Mines were or probably were responsible for about 21 to 22 of the 43 U-boat losses. Five of the U-boats had been sunk by escorts.98 The exchange ratio, the numbers of ships sunk for every U-boat loss, had been as high as 167 in April, and the average was as high as 73 in May, June, and July when the first convoys were being introduced. It fell to only 16 in the six-month period from August 1917 to January 1918. Nevertheless, submarines continued to cause substantial losses, and at the end of the second six months of unrestricted submarine warfare, the total shipping available was still declining. World shipping losses had been roughly 2.25 million tons, of which only somewhat more than 1.5 million tons had been replaced by new building and about a half-million tons of interned German and Austrian shipping, which was seized and brought into Allied service. Admiral Hezlet points out that imports were down 20 percent by the end of 1917 from the preceding year, and that it was only by rigid controls, the exclusion of nonessentials, and strict rationing that the situation was “just kept in hand.”99
A Naval History of World War I Page 63