Introduction

In April 1917, German submarines sank 860,334 tons of Allied and neutral shipping. Britain still controlled the surface of the North Sea, with the High Seas Fleet behind the screen of the Royal Navy's Grand Fleet. Yet that command no longer guaranteed the passage of food, fuel, munitions, and raw materials. The Grand Fleet could defeat an enemy battle fleet, but it could not be everywhere along the routes taken by thousands of merchant voyages. The losses of that month suggested that it might run out of shipping before it found an answer.1

Because no new weapon could restore the lost tonnage quickly, the answer adopted that spring would protect merchant ships by changing how they crossed the Atlantic. They would assemble outside selected ports, cross the ocean in groups, and enter the submarine zone behind a screen of warships. At first, that arrangement appeared merely to concentrate danger, for a convoy spanning several miles could throw up a bank of smoke visible on the horizon; once attacked, slow merchantmen might collide or scatter. Naval officers who raised those objections were reasoning from genuine shortages and familiar dangers, although some of their facts and assumptions were wrong.

Those objections described dangers that convoy could not eliminate, yet they missed the geographical advantage created when forty ships shared one route instead of sailing separately. A single group reduced the number of encounters available to each U-boat, brought escorts to the place where an attack occurred, and allowed intelligence officers to route concentrated traffic around reported danger. Making that formation useful also required accurate shipping data, assembly ports, commodores, sailing schedules, signals, local escorts, and cooperation across navies. Britain did not defeat unrestricted submarine warfare by suddenly making the U-boat harmless; it prevented the U-boat from stopping enough cargo in time.

A Deadline Written in Tonnage

Behind the tonnage crisis lay a timetable that German naval leaders had placed before their government four months earlier. In a memorandum of December 22, 1916, Admiral Henning von Holtzendorff argued that unrestricted attacks could sink 600,000 tons of shipping each month. Neutral shipowners, frightened from British trade, would magnify the physical destruction, while shortages and higher freight costs pressed on a country dependent upon seaborne imports. Holtzendorff predicted that five months of such losses would force Britain to peace. The United States might enter the war, he acknowledged, but it could not exert decisive military power before the calculation had done its work.2

When Germany put that calculation into operation on February 1, a U-boat commander no longer had to surface, stop a merchant vessel, and make provision for its crew before sinking it. He could attack without warning inside the declared war zone, including neutral ships whose flags had once offered some protection. The new orders expanded the possible targets while leaving the submarines widely dispersed. Their strategic effect came from accumulating tonnage across many brief encounters, most of them far from any fleet action.

By March, the accumulated sinkings had exceeded Holtzendorff's monthly target, and in April they rose higher still. British officials also had to count ships damaged beyond economical repair, marine accidents, delayed arrivals, and the imports never loaded because owners withheld vessels. New construction could not replace the spring losses at the same pace. Each weekly return therefore narrowed the interval in which the Admiralty could convert patrol craft, build weapons, and learn how the enemy was operating. Germany had made merchant tonnage lost the measure of the campaign; British officers would have to answer by counting cargoes that arrived.

Hunting the Approaches

To keep those cargoes arriving, the Royal Navy had to rethink the geography built into its antisubmarine plans. Before the war, the expected commerce raider was a surface cruiser whose position could be reported and whose need for coal restricted its movements. The Royal Navy had defeated that threat in the opening months of the war by hunting German cruisers and clearing the trade routes. A submarine appeared to invite the same response at closer range: patrol the focal areas outside British ports, meet the attacker where shipping converged, and preserve merchantmen's freedom to sail independently.3

That logic weakened as U-boats acquired the range to attack farther into the Atlantic. Patrol vessels spread across the Western Approaches found little in broad water, especially when early hydrophones could not search effectively while their own engines were running. A submarine that sighted a patrol could submerge or alter course; a submarine that found an unescorted steamer could choose the time and angle of attack. Adding more armed trawlers, yachts, sloops, and destroyers enlarged the search, but it did not make the ocean small.4

As the search widened without finding enough U-boats, the Admiralty added armed merchant ships, mines, air patrols, depth charges, disguised Q-ships, restricted sailing times, and protected cross-Channel and Scandinavian routes. These measures imposed risks and sometimes sank submarines, yet none removed enough lone merchantmen from the ocean spaces where a U-boat could meet them without an escort nearby. The patrol system still placed the larger burden of search upon the defender.

The Convoy Britain Thought It Could Not Form

Convoy promised to reverse the patrol system's burden of search, yet putting that principle to sea challenged the Royal Navy's experience and resources. Troop transports already crossed under escort, and coastal or cross-Channel sailings could be controlled along short routes, but ocean trade posed a larger problem. Every destroyer assigned to merchant protection was unavailable to the Grand Fleet or another patrol; short-legged escorts also needed fuel before they could reach far into the Atlantic. German surface raiders had not vanished as a possibility, so an ocean convoy seemed to require both a cruiser against a raider and destroyers against submarines.5

Alongside this naval demand, merchant masters and shipping officials supplied a second set of objections. Ships arrived from different companies with different speeds, turning circles, signalling equipment, and masters. A group might be held for its slowest member, while concentrated arrivals could overwhelm tugs, pilots, berths, and railway sidings. In February 1917, experienced merchant captains told the Admiralty that ordinary masters could not maintain the close station expected of a battle squadron. That warning carried weight because naval planners had borrowed their idea of formation keeping from the Grand Fleet, whose columns maneuvered for battle under common training and command.6

Even if merchantmen could hold formation, the number of ships to be escorted appeared to settle the question. Published customs returns counted repeated movements by coastal and cross-Channel vessels as though each were a distinct ocean-going arrival or departure. When Commander Reginald Henderson and Captain Kenneth Dewar sought figures from the new Ministry of Shipping, they found much less overseas traffic than the totals implied. Their later accounts compete for priority and differ in detail, but both reveal the institutional weakness: the Admiralty had evaluated convoy without a reliable registry of the ships that would need it.7

Once Henderson and Dewar had separated ocean-going arrivals from repeated coastal movements, their corrected numbers changed the scale of the practical problem without abolishing escort shortages or port congestion. Trials on the French coal route and between Britain and Scandinavia had also shown that merchantmen could follow instructions less exacting than fleet maneuvers. The practical question was no longer whether every arriving ship could receive its own destroyers, but how many groups could be protected as they crossed the most dangerous water.

Three Days and a Long Argument

During the final week of April, pressure for ocean convoy moved through several meetings rather than producing one dramatic conversion. Prime Minister David Lloyd George pressed the issue again at the War Cabinet on April 23, and the Admiralty agreed to report. On April 26, Rear Admiral Alexander Duff, director of the Anti-Submarine Division, submitted a plan for experimental ocean convoys; First Sea Lord John Jellicoe approved it the next day. Lloyd George visited the Admiralty on April 30, after that approval but before the experiment had demonstrated anything on the Atlantic.8

Those overlapping steps later became a struggle over credit because each participant could point to a different moment in the process. Lloyd George presented civilian intervention as the force that broke professional resistance, while Jellicoe emphasized Admiralty planning already under way. Admiral William Sims described American advocacy and support as essential to the new system. The sequence does not sustain a story in which one visitor invented convoy on April 30. It shows political pressure, junior officers' corrected evidence, Duff's staff work, Jellicoe's authorization, and the prospect of American destroyers converging during the same crisis.9

The late-April authorization sent convoy to sea first as an experiment, beginning with seventeen ships that left Gibraltar on May 10. Two dropped out, one of them was sunk, and the ships that held the formation reached Britain. A second trial of twelve ships left Hampton Roads on May 24. Its merchantmen received an ocean escort for the passage east, then met destroyers near the submarine danger zone. By summer, regular assembly routes were extending outward from ports in North America and the Mediterranean.10

More Than Ships in Columns

Turning those successful trials into a routine Atlantic service required work that began before any convoy left port. Ships had to arrive at the same port, be sorted by speed, receive routes and recognition signals, and place themselves under a designated commodore. Masters who had spent their careers exercising independent command now followed standardized orders for station, zigzag, fog, breakdown, and attack. A straggler might expose itself and delay no one; a ship unable to keep the convoy's speed belonged in another group. These rules translated naval protection into actions that merchant crews could perform.

Once a convoy cleared its assembly port, protection changed with the geography of the voyage. Across the open ocean, where a German surface raider remained a concern and U-boats were less concentrated, the escort could include a cruiser or armed merchant vessel. Destroyers and sloops sailed west from British or Irish bases to meet the approaching ships before they entered the danger zone. The same local escorts then carried an outward convoy toward its release point. Because their endurance was limited, rendezvous times and positions had to fit refuelling, weather, and the movement of other groups.11

To bring those changing escorts together at the proper times, Jellicoe's central Convoy Section coordinated the Naval Staff, the Ministry of Shipping, port officers, and overseas authorities. Its charts tracked expected arrivals and escorts; printed instructions established common procedure; signals changed routes when intelligence indicated danger. William Sims compared the schedules to a railway, an interested but revealing image for a system that moved vessels through timed junctions, where individual captains no longer entered the ocean alone.12

The schedules also gave intelligence officers something the patrol system had lacked: a known body of shipping whose course could be changed before it entered reported danger. Room 40 read German naval signals, while direction-finding stations, patrol reports, and merchant sightings helped plot the probable positions of U-boats. Officers responsible for submarine tracking compared those plots with the convoy charts; messages could then divert an approaching group or alter a destroyer rendezvous. The resulting route was never perfectly informed, but a mistake now threatened one moving group; independent ships no longer followed the same line in a daily procession.13

The Ocean Empties

Once traffic moved on schedules and in groups, a U-boat searching an Atlantic route had far fewer chances to find a target. Forty independent ships traced forty paths across the areas a U-boat might patrol; forty ships in convoy presented one moving patch of sea. The smoke from that patch could be conspicuous at close range, but the submarine first had to be near enough to see it. Between October and December 1917, German submariners located only 39 of 219 Atlantic convoys. For long intervals, an individual boat found the traffic lanes unexpectedly bare.14

Finding a convoy did not restore the conditions in which submarines had hunted independent ships. A U-boat on the surface still had the speed and visibility to maneuver ahead of an independent merchantman, use its deck gun, or arrange a torpedo approach. Destroyers near a convoy could compel it to submerge, reducing both speed and sight. A boat that penetrated the screen might fire once or twice before the formation passed and the escorts began dropping depth charges. Convoy protection did not require an escort to detect and sink every attacker; interruption could be enough for the cargoes to continue.

Across nearly 84,000 ships convoyed between February 1917 and October 1918, 257 were sunk, a rate of 0.30 percent. The loss rate among independent sailings in the same compilation was 5.93 percent. Those aggregate figures cover different routes and phases of a changing campaign, so they should not be treated as a controlled experiment. They nevertheless match the operational record: most ships lost were sailing outside convoy, while escorts were present for the attacks concentrated traffic did attract.15

On August 13, while the system was still incomplete, Sims told Secretary of the Navy Josephus Daniels that most trade in the danger zone was then moving in convoy. He counted twelve mercantile convoys approaching British waters. His report carried an advocate's enthusiasm, yet the movements he described can be checked against the expanding schedules and the later loss returns. More merchant ships were reaching port under a procedure that had barely existed three months earlier.16

A System, Not a Single Cure

Even as convoyed traffic multiplied, the spring crisis did not end with the arrival of the first escorted groups. The system expanded gradually, and U-boats continued to sink an average of nearly 400,000 tons per month into early 1918. Approaches to port created new concentrations after an ocean convoy dispersed, while coastal shipping and vessels still sailing independently remained exposed. Escorts were scarce, bad weather broke rendezvous, and a merchantman that could not hold its place sometimes finished the voyage alone.17

While escorts protected individual voyages, the Ministry of Shipping managed the national balance between imports and losses. It controlled sailings and allocated tonnage; import restrictions reduced demands for less essential cargo; shipyards and repair facilities returned hulls to service; railway and port management cleared arrivals. At sea, mines, depth charges, armed merchantmen, patrols, and intelligence continued to impose costs on U-boats. Aircraft rarely destroyed a submarine, but their presence could force one to submerge and lose the speed needed to reach a convoy's track.18

That wider system also depended on Allied ships and bases that the Royal Navy could not provide quickly. American destroyers based at Queenstown strengthened the danger-zone escort, while naval authorities on both sides of the Atlantic organized departures, exchanged intelligence, and adjusted rendezvous. French, British, and American responsibilities differed by route, and disputes over allocation did not disappear. Coordination across the ocean allowed those imperfect contributions to support one another without placing every vessel under one national command.19

With convoy and its supporting measures operating together, monthly Allied losses declined from 696,725 tons in June to 353,600 in September. After an October increase, the November toll was 302,600 tons, though monthly losses did not remain below 300,000 until the spring of 1918. By then Germany had more capable submarines operating across a wider area. What it lacked was the uninterrupted succession of targets required by Holtzendorff's five-month calculation. The British import system bent under the losses, but its ships and cargoes continued to arrive.20

Conclusion

Since the Allies could neither sweep the Atlantic nor reveal every submerged hull in 1917, convoy changed the distribution of the shipping that submarines hunted. Concentration reduced encounters, escort turned an attack into a hurried contest, and intelligence could move a known body of ships around a reported danger. Each advantage depended upon work ashore that assembled vessels, corrected schedules, and brought an escort to the appointed patch of water.

Because protection depended on this organization ashore as well as action at sea, the April argument over credit cannot identify a single author of the result. Political pressure helped open the way, as did better data, naval staff work, merchant practice, and American reinforcement; the decisive evidence accumulated afterward, in arrivals through the rest of 1917. Germany had planned to make Britain count down five months of disappearing tonnage; by autumn, individual U-boats were instead spending days searching an emptier ocean for the next convoy.

Endnotes

  1. Jan S. Breemer, Defeating the U-boat: Inventing Antisubmarine Warfare, Newport Paper 36 (Newport, RI: Naval War College Press, 2010), 1–3, open-access PDF. ↩
  2. Henning von Holtzendorff, “Unrestricted Submarine Warfare,” memorandum, December 22, 1916, translated in German History in Documents and Images, accessed August 23, 2026, document and translation; Christopher B. Havern Sr., Ensuring the Lifeline to Victory: Antisubmarine Warfare, Convoys, and Allied Cooperation in European Waters during World War I (Washington, DC: Naval History and Heritage Command, 2020), 3–11, open-access PDF. ↩
  3. Robert W. H. McKillip, “Undermining Technology by Strategy: Resolving the Trade Protection Dilemma of 1917,” Naval War College Review 44, no. 3 (Summer 1991): 19–24, journal record and PDF. ↩
  4. Breemer, Defeating the U-boat, 44–47; Havern, Ensuring the Lifeline, 18–20. ↩
  5. John Rushworth Jellicoe, The Crisis of the Naval War (London: Cassell, 1920), 102–19, digitized edition; McKillip, “Undermining Technology,” 24–30. ↩
  6. Breemer, Defeating the U-boat, 51–55; Havern, Ensuring the Lifeline, 20–22. ↩
  7. Breemer, Defeating the U-boat, 47–50; McKillip, “Undermining Technology,” 28–31. Breemer notes the close similarity and unresolved priority of Henderson's and Dewar's retrospective accounts. ↩
  8. McKillip, “Undermining Technology,” 30–31; Breemer, Defeating the U-boat, 58–61. ↩
  9. Breemer, Defeating the U-boat, 58–61; Havern, Ensuring the Lifeline, 20–23. Havern concludes that American advocacy probably accelerated British action but cannot by itself establish that the Admiralty would otherwise have refused convoy. ↩
  10. Havern, Ensuring the Lifeline, 22–25; William S. Sims and Burton J. Hendrick, The Victory at Sea (London: John Murray, 1920), 73–98, digitized edition. ↩
  11. Jellicoe, Crisis of the Naval War, 119–35; Sims and Hendrick, Victory at Sea, 86–98. ↩
  12. Jellicoe, Crisis of the Naval War, 123–51; Sims and Hendrick, Victory at Sea, 101–16. Both are retrospective participants' accounts; they are used here for procedures also described in the secondary works, not to assign individual credit. ↩
  13. Robert Metcalfe, “Submarines and Naval Layered Defence in the First World War” (PhD diss., Brunel University London, 2024), 168–70, university repository record and PDF; Breemer, Defeating the U-boat, 65–66. ↩
  14. Havern, Ensuring the Lifeline, 62–63; Breemer, Defeating the U-boat, 62–65. ↩
  15. Breemer, Defeating the U-boat, 62–65. The figures cover ships convoyed from February 1917 through October 1918, including protected traffic beyond the homeward Atlantic convoys begun in May. ↩
  16. William S. Sims to Josephus Daniels, August 13, 1917, in Naval History and Heritage Command Documentary Histories, accessed August 23, 2026, transcription. ↩
  17. Breemer, Defeating the U-boat, 61–64; McKillip, “Undermining Technology,” 31–32. ↩
  18. McKillip, “Undermining Technology,” 32–35; Alexander Howlett, “The Royal Naval Air Service and Anti-Submarine Warfare in the North Sea, 1917–1918,” British Journal for Military History 4, no. 2 (February 2018): 32–36, open-access PDF. ↩
  19. Havern, Ensuring the Lifeline, 15–25, 60–64. ↩
  20. Havern, Ensuring the Lifeline, 62–64; Breemer, Defeating the U-boat, 61–64. ↩

Bibliography

Primary Sources

  • Holtzendorff, Henning von. “Unrestricted Submarine Warfare.” Memorandum, December 22, 1916. Translated in German History in Documents and Images. Accessed August 23, 2026. Document and translation.
  • Jellicoe, John Rushworth. The Crisis of the Naval War. London: Cassell, 1920. Digitized edition.
  • Sims, William S. Letter to Josephus Daniels, August 13, 1917. In Naval History and Heritage Command Documentary Histories. Accessed August 23, 2026. Transcription.
  • Sims, William S., and Burton J. Hendrick. The Victory at Sea. London: John Murray, 1920. Digitized edition.

Secondary Sources

  • Breemer, Jan S. Defeating the U-boat: Inventing Antisubmarine Warfare. Newport Paper 36. Newport, RI: Naval War College Press, 2010. Open-access PDF.
  • Havern, Christopher B., Sr. Ensuring the Lifeline to Victory: Antisubmarine Warfare, Convoys, and Allied Cooperation in European Waters during World War I. Washington, DC: Naval History and Heritage Command, 2020. Open-access PDF.
  • Howlett, Alexander. “The Royal Naval Air Service and Anti-Submarine Warfare in the North Sea, 1917–1918.” British Journal for Military History 4, no. 2 (February 2018): 32–49. Open-access PDF.
  • McKillip, Robert W. H. “Undermining Technology by Strategy: Resolving the Trade Protection Dilemma of 1917.” Naval War College Review 44, no. 3 (Summer 1991): 18–37. Journal record and PDF.
  • Metcalfe, Robert. “Submarines and Naval Layered Defence in the First World War.” PhD diss., Brunel University London, 2024. University repository record and PDF.