"Numbers are Essential": Victory in the North Atlantic Reconsidered, March-May 1943

by Rob Fisher


The Germans conceded defeat in the North Atlantic towards the end of May 1943 after sustaining heavy U-boat losses in a string of convoy battles since the beginning of the month. Allied forces destroyed forty U-boats during May after averaging only fourteen per month from January to April 1943.(1) The U-boat men called it "Black May"; others have called it the "Stalingrad at Sea". The magnitude of the German defeat was all the more surprising because February and March had witnessed several dramatic victories for the wolf packs against elite British and American escort groups.

What happened in the North Atlantic between March and May 1943 to cause this dramatic turnaround? There is no shortage of explanations. These months have been the most closely studied of the Battle of the Atlantic. Naval historians have attributed the Allied victory of May 1943 to a number of different factors: the development of sophisticated electronic equipment such as 10-cm radar and ship-borne High Frequency/Direction-Finding; improved tactics and training of escorts; special intelligence from breaking the Enigma codes; and more surface escorts and very-long-range or VLR aircraft. Undoubtedly all of these factors played an important role in the defeat of the U-boats in 1943. Most historians have been reluctant to point to a single factor that was decisive. Recent historians have emphasized that the victory of May was qualitative: obsolete U-boats were outclassed by the quality-technological and tactical- rather than the quantity of Allied surface and air escorts.(2)

I do not doubt the tremendous importance of Allied improvements in intelligence, technology, training, and tactics. Significant developments, however, did not occur at sea in these areas between March-the month of defeat-and May-the month of victory. These qualitative improvements took effect gradually and contributed to Allied victory over the long haul in the Atlantic campaign. None of them adequately explain the sudden change of fortunes that took place in April and May.

Decryption intelligence had played an important role since late December 1942 but although there were temporary blackouts no major shift had occurred since that time. Indeed, to some extent, the Germans had the intelligence advantage at this stage. The tremendous technological advantage conferred by centimetric radar and High Frequency Direction-Finding is also inadequate as an explanation for the sudden-ness of Allied victory. The British and American escort groups involved in the convoy disasters of February and March had carried this electronic equipment. It had been standard equipment on British escorts since early 1942. The impact of advances in anti-submarine training and tactics is more problematic to assess but the refinements in tactics since late 1942 seem to have been subtle rather than significant. Admiral Max Horton brought a renewed emphasis of training to Western Approaches Command in late 1942 which-as Marc Milner has demonstrated-resulted in the removal of the Canadian escort groups from the North Atlantic run in January.(3) Still, the prime beneficiaries of this training-the C groups-did not play a leading role in the May victory and the slow rotation of British escorts through the training establishments cannot explain the sudden change of fortunes.

What changed the balance between wolf packs and convoys? Numbers. Between late March and early May, the Allies reinforced the North Atlantic dramatically with large numbers of surface escorts, VLR aircraft, and escort carriers. Decisions made in March freed up surface escorts for the North Atlantic by suspending the dangerous Arctic convoys to Russia and scaling back the escort commitments for Operation Torch in North Africa. Western Approaches used this new-found wealth to enlarge the Mid-Ocean Escort Force and create six new support groups. In addition, two new VLR squadrons joined the lone VLR squadron operating in the North Atlantic. This reinforcement crushed the wolf packs in a sudden and convincing fashion.

Convoy Battles of Late 1942 and Early 1943

Time only permits a brief review of the major convoy battles of this period to demonstrate the impact of this reinforcement. In November and December 1942, two Canadian-escorted convoys suffered heavy attack. SC 107 lost fifteen merchant ships to a pack of sixteen submarines-Siegfried von Forstner in U-402 sank five ships alone. The close escort consisted of six escorts: one destroyer and five corvettes, all without 10-cm radar or HF/DF. It had to defend the convoy for several nights in the Greenland air cover gap without air or surface support. Canadian aircraft destroyed two U-boats at the outset of the battle but did not affect its outcome. Next month, the westbound convoy ONS 154 lost fourteen ships to a group of eighteen submarines. Its close escort consisted of six warships: one destroyer and five corvettes, newly-equipped with centimetric radar and HF/DF. HMCS St. Laurent destroyed one U-boat. Again, little support was offered. British authorities blamed the losses in both battles on the poor training of the Canadian escorts and had the C groups removed from the North Atlantic run.(4)

Allied code breakers cracked the four-rotor Enigma codes in mid-December 1942. From then until the climax of the battle in May 1943, both sides would have good signals intelligence. In early February, the Allies fared slightly better with convoy SC 118. It had an abnormally large British escort of nine warships: four destroyers, four corvettes, and one Coast Guard cutter. This powerful escort destroyed two U-boats but still lost eleven merchant ships. Again, it was von Forstner in U-402 who sank six ships in two attacks. Long-range air cover from Iceland and Northern Ireland ended the pursuit and destroyed one more submarine.

The Germans intercepted two convoys escorted by the American escort group A3 in February and March. ON 166 lost fourteen merchant ships to a pack of nineteen U-boats over four nights. The close escort had the latest equipment and consisted of seven warships: two large veteran US Coast Guard Cutters and five RCN corvettes. It destroyed one U-boat and VLR aircraft destroyed another. On its return to the United Kingdom with SC 121, the same group was hit hard again. This time it had only five escorts. In hurricane force weather, it lost twelve ships to a group of twenty-plus submarines. The arrival of aircraft and warships from Iceland put an end to the battle after three nights. Surface and air escorts did not sink a single U-boat.(5)

This Allied defeat set the stage in mid-March for two of the best known convoy battles of the war-SC 122 and HX 229. The British escort of SC 122 consisted of eight escorts: two destroyers, one frigate, and five corvettes. It was well-equipped with centimetric radar and HF/DF and, being British, was presumably well-trained. SC 122 lost nine merchant ships to a pack of about twenty submarines over three nights. Simultaneously, HX 229 was under attack by another fifteen to twenty U-boats. Its British escort consisted of only five escorts: two destroyers and three corvettes, similarly well-equipped. The pack sank thirteen ships. Intense air cover by VLR aircraft and the arrival of three warships from Iceland put an end to the slaughter of the two convoys. VLR and long-range aircraft flew eleven sorties on the 18th and destroyed one sub-the only loss to the enemy.(6)

The surface escorts had lost twenty-three ships without sinking a U-boat. It was a disaster of the magnitude of SC 107 and ONS 154. The much vaunted British escort groups-fitted with the latest electronic equipment and weapons systems-fared little better in the air cover gap than the American and Canadian escort groups had before them against slightly-better odds. The result proved that even a well-equipped and highly-trained close screen of six or seven escorts could not prevent heavy losses to a large wolf pack in the air cover gap. The defeat, understandably, caused dismay and soul-searching in the Admiralty. April and May would bring a much different result.

Reinforcing the North Atlantic

My argument hinges on numbers so we should consider the size of the forces arrayed against the wolf packs in the North Atlantic in March 1943. The Mid-Ocean Escort Force consisted of ten escort groups which normally sailed with two destroyers and four corvettes. The average for this period was 6.8 escorts per convoy, up slightly from 6.4 in November.(7) In terms of long-range and very-long-range air cover, Coastal Command had nine VLR Liberators in Iceland and nine more in Northern Ireland for a total of eighteen-none with 10-cm radar or Leigh Lights. Long range aircraft consisted of about fifty aircraft-mainly Catalinas and Fortresses-based in Newfoundland, Iceland and Britain.(8) In total, the North Atlantic defences consisted of about 75 surface escorts, eighteen VLR and about fifty long-range aircraft.

Decisions made in early 1943 caused a transfer of resources to this theatre in late March and April. The Casablanca Conference decided that "the defeat of the U-boat must remain a first charge on the resources of the United Nations". The Combined Chiefs of Staff agreed to several measures to bring this about, including (a) Provision of auxiliary escort carriers for the protection of Atlantic convoys at the earliest practical moment; (b) Provision of long-distance shore-based air cover as a matter of urgency for the transatlantic convoy routes; (c) Development of air fields in Greenland; and (d) Non-ocean-going escorts should be used for Operation Husky (the invasion of Sicily) to the maximum extent possible.(9)

Resources were not made available until late March. The suspension of the Arctic convoys to Murmansk freed up fifteen fleet destroyers and one escort carrier-HMS Dasher-for transfer to Western Approaches. These destroyers formed the basis of three new support groups in April. Furthermore, reductions in the supply convoys for Operation Torch released additional escorts-including the escort carriers HMS Biter and Archer-for the North Atlantic. The result was two more support groups composed of sloops and frigates. In addition, the return of the C-groups and adjustments to the convoy cycles allowed the Mid-Ocean Escort Force to be increased from ten to twelve close escort groups. The loss of HMS Dasher to an internal explosion on March 27th was a heavy blow but the loan of the carrier USS Bogue and its destroyer-escort to Western Approaches offset the loss. By mid-April, Western Approaches had six brand new support groups-three of them with escort carriers.(10)

This transfer of strength had an immediate impact. The twelve close escort groups of the Mid-Ocean Escort Force averaged about 7.9 escorts per convoy in April and May-as the force increased from about 75 to 95 escorts.(11) When we add the six support groups which averaged about four destroyers, frigates or sloops at sea, and the three escort carriers-the theatre experienced an increase in fighting force in the range of 70-90%. It is tempting to speculate that the heavy losses to A3 and the B-groups in February and March spurred the Allies to action. Still, it is clear that this reinforcement had support at the highest political levels in Great Britain and the United States-suspension of the convoys to Russia and transfers from Operation Torch would not have happened without it.(12)

The numbers of VLR aircraft increased similarly in the North Atlantic theatre in April. Coastal Command had eighteen Liberators in Iceland and Northern Ireland on March 15th. This figure had risen by mid-April to twenty-seven Liberators, and by mid-May to thirty-five VLR Liberators in three squadrons: an increase of almost 100%. Ten of the new aircraft had centimetric radar (partially contradicting my claims that there was no technological improvement in this period) but none had Leigh Lights. The doubling of the number of VLR aircraft was crucial because only they could operate in the Greenland air gap. Service and national rivalries, however, prevented the deployment of them to Newfoundland until the summer of 1943.(13)

In short, in a few weeks, Allied anti-submarine resources in the vital North Atlantic theatre had virtually doubled in terms of numbers and fighting force. The knowledge of this coming reinforcement produced optimism at Western Approaches in spite of the devastating losses of SC 122 and HX 229. Within days of those battles, Admiral Max Horton wrote to Rear Admiral R.B. Darke on March 23rd: "This job has been pretty sombre up to date, because one hadn't the means to do those very simple things for which numbers are essential, and which could quash the menace definitely in a reasonable time; but in the last few days things are much brighter and we are being reinforced, and I really have hopes now that we can turn to another and better role-killing them. The real trouble has been basic-too few ships, all too hard worked with no time for training and all that that entails". Horton continued that "The Air carried afloat is now turning up to an extent which may be almost embarrassing in the next few months....All these things are coming to a head just now, and although the last week has been one of the blackest on the sea, so far as this job is concerned, I am really hopeful".(14)

ONS 5 and the Convoy Battles of May 1943

This strength made an immediate difference in the North Atlantic. U-boat Command failed to mount a major convoy operation in April. Towards the end that month, however, the Germans had stationed over fifty U-boats in two patrol lines in the western Atlantic off Newfoundland and another fifteen boats in the eastern Atlantic (15) It was this smaller group that intercepted convoy ONS 5 on April 28th southwest of Iceland. The close escort, B7, consisted of nine British warships: two destroyers, a frigate, four corvettes, and two anti-submarine trawlers. Five or six U-boats made contact in bad weather and sank one merchant ship. Western Approaches ordered the newly-created support group EG 3, consisting of five destroyers, to sail from St. John's to reinforce ONS 5. They arrived on the 30th increasing the screen to fourteen escorts. In view of the disappointing results, U-boat Command called off the pack on May 1st. The heavy seas, however, had also scattered the convoy and caused the close escort to run low on fuel. Four of the destroyers-including the Senior Officer-had to detach for port leaving the convoy with a screen of three destroyers, a frigate, four corvettes, and one trawler.

The battle was not over. Ahead, two patrol lines northeast of Newfoundland lay in wait for an eastbound convoy. Instead, much delayed by the storm, on May 4th ONS 5 blundered into a patrol line of twenty-nine U-boats from the other direction. Western Approaches sailed a second support group of frigates and sloops, EG 1, from St. John's to bolster B7. An RCAF Canso from Newfoundland attacked U-209 on the 4th and damaged it so heavily that it sank a few days later. Indeed, VLR aircraft from Iceland provided protection at extreme range on the 5th, revealing how the Greenland air gap had closed. Even though the convoy had already reached the outer limits of air cover, U-boat Command ordered the pack to converge upon the convoy while there was still a chance. A wild and confused battle ensued. Initially, the Germans had the upper hand sinking twelve ships during the 4th and 5th of May while the escort destroyed one sub. The tide turned against them on the 5th, however, as heavy fog set in-bad conditions for U-boat lookouts-and the rough seas calmed-good conditions for Allied radar. The reinforced escort seized the initiative and destroyed five more U-boats on the 6th-turning defeat into victory. The close escorts took the honours, sinking four subs solo but the support groups had a hand in sinking two of the subs. The balance sheet for ONS 5 showed seven U-boats lost in return for thirteen ships sunk-a disastrous rate of exchange for the Germans. All told, about forty U-boats stalked the convoy which received protection from no less than nineteen warships-seven destroyers, four frigates, two sloops, four corvettes, and two trawlers. British authorities hailed the result as a decisive victory and historians have called it a turning point in the war at sea. It is well to remember that only eight escorts were present when most of the merchant ships were lost and things might have been much different had the convoy had to transit the air cover gap for several nights far from surface reinforcement.(16)

From here, things went from bad to worse for the Germans. Efforts to mount large wolf pack operations against several convoys in May cost them dearly. The pursuit of HX 237-defended by the seven escorts of C2 and support group EG 5 (including the carrier HMS Biter)-resulted in the destruction of three U-boats in exchange for three stragglers lost. HMCS Drumheller and RCAF Squadron 423 shared in the kills with other Allied surface and air escorts.(17) The balance sheet for SC 129-protected by a British group of eight escorts and reinforced by EG 5 with Biter-showed one U-boat destroyed and two merchant ships sunk-both by our old friend von Forstner in U-402. The attack on SC 130 was even less successful. Thirteen warships-nine from the close escort and four from the support group EG 1-prevented the pack from sinking a ship and destroyed one U-boat. On the last two days of the battle, VLR Liberators made twenty-seven U-boat sightings and destroyed one. German operations also failed against ON 184, defended by six close escorts and the four destroyers and USS Bogue of support group EG 6. The pack could not penetrate the powerful screen to sink a single ship and carrier aircraft destroyed one U-boat. The results against HX 239 resembled this action. Eight close escorts, four support group destroyers and HMS Archer prevented any losses from the convoy and carrier aircraft destroyed one U-boat.(18)

In two months, the Allies had solved the wolf pack. As the magnitude of "Black May" set in, U-boat Command conceded defeat. On May 24th it ordered submarines to withdraw from the North Atlantic to the central Atlantic beyond the range of shore-based air cover. Another order instructed them to remain submerged when conditions of poor visibility prevailed. With the benefit of hindsight, Admiral Karl Dönitz wrote in his memoirs: "We had lost the Battle of the Atlantic". U-boat Command blamed the defeat on Allied air cover and the performance of radar, and underestimated the number of U-boats destroyed during convoy battles.(19)

Allied forces destroyed forty U-boats in May; thirty-two of them in the North Atlantic. Surface and air escorts of transatlantic convoys destroyed twenty of these U-boats. Including shared kills, surface escorts had eleven kills, land-based aircraft had eight, and carrier aircraft had three kills. In return for the loss of twenty U-boats in the convoy battles, the wolf packs sank only nineteen merchant ships (including stragglers) from the transatlantic convoys, most of them in the opening battle for ONS 5. Not including ONS 5, the exchange was six merchant ships for thirteen U-boats.(20)

The key to the sudden reversal of fortunes was the increase in the strength of the surface and air escort provided to convoys. Not only were there more close escorts, but support groups reinforced twenty of the twenty-two convoys that sailed between April 6th and May 19th. In fact, four of these convoys received protection from two support groups. For all intents and purposes the three escort carriers and two additional VLR squadrons had closed the Greenland air gap. Admiral Horton wrote with confidence at the end of May: "We now know...what strength and composition of forces is necessary to deal with the U-boat menace against convoys"(21)


The results of the convoy battles of May are testimony to the decisive impact of numbers. But the German U-boat fleet was also growing rapidly in the spring of 1943. The front-line Atlantic U-boat fleet consisted of 167 attack boats on February 1st and 186 boats on May 1st; an increase of about 15% and enough to keep about sixty on active operations in the North Atlantic.(22) The steady growth of U-boat strength did not have the same impact as the increase in the numbers of surface and air escorts because of the dynamics of convoy battles and wolf pack tactics. U-boat Command stationed submarines twenty miles apart in patrol lines perpendicular to the convoy routes. The more U-boats, the longer the patrol lines, and the less likely that a convoy could be routed safely around the line. The disadvantage was that the longer the patrol line, the further from the point of interception would be the subs at the opposite end of the line.

The dynamics of convoy battles also favoured large escorts over large packs. Most of the sinkings in convoy battles of the 1942 to 1943 period were achieved by the two or three boldest and most aggressive U-boat commanders in the pack-men like Siegfried von Forstner. The initial attacks forced the escort to counter-attack and rescue survivors. This opened up the screen for the less aggressive commanders to move in and swamp the defenders. A close escort of six or seven warships would become disorganized quite quickly after two or three sinkings. An escort of ten to twelve warships including several destroyers or frigates and strong air cover was much less susceptible to this kind of disorder. The bold U-boat commanders found it much more difficult to penetrate the screen, and if they did manage to attack, the escort regrouped quickly to prevent further losses.

The Admiralty's Operational Research unit, led by the scientist P.M.S. Blackett, demonstrated mathematically the superiority of a large escort. Blackett found that "that the chance that a U-boat would penetrate the screen depended only on the linear density of escorts, that is, on the number of escort vessels for each mile of perimeter to be defended". Indeed, increasing the size of the convoy did not increase the size of its perimeter by the same proportion so that a convoy of sixty merchant ships and twelve escorts was in fact much more strongly protected than a convoy of thirty ships and six escorts. Blackett described convoy battles as "unstable equilibriums" where, when the advantage went to one side, it went completely to that side. In April and May 1943, Allied surface and air reinforcements tipped the balance decisively in favour of the defenders.(23)

Can we draw some larger conclusions from this analysis? Could these anti-submarine resources have been made available earlier? The Royal Navy had four escort carriers in October 1942: three assigned to Operation Torch and one to the Russian convoys. They did not share in one U-boat kill in either of these assignments between September 1942 and April 1943.(24) In addition, these routes absorbed scores of fleet destroyers and other escorts. The wastefulness of this deployment for Operation Torch was tacitly admitted at Casablanca by the decision that non-ocean-going escorts be used to the fullest extent possible for Operation Husky. Moreover, by October 1942 ships sailing in the coastal convoys of the United States received as much protection as those in the North Atlantic in spite of the much diminished threat in coastal waters and the close proximity of shore-based aircraft. It was the same story with air cover. Coastal Command assigned two new squadrons of long-range Liberators to the protection of the Torch convoys.(25) Scores of aircraft that could have tipped the balance against the wolf packs in late 1942 were lost in bombing raids over Germany. It is difficult not to conclude that had the Allied resources been dedicated to the task, victory in the North Atlantic could have come in the fall of 1942. It would have been a more gradual and less dramatic victory than that of May 1943 to be sure, but victory nonetheless.


1. Axel Niestlé, German U-boat Losses during World War II: Details of Destruction (Annapolis, MD, 1998), p.193.

2. David Syrett, The Defeat of the German U-boats (Columbia, SC, 1994), p.143, 263-264; and Michael Gannon, Black May (New York, 1998), pp.350, 393.

3. Marc Milner, North Atlantic Run: The Royal Canadian Navy and the Battle for the Convoys (Toronto, 1985), pp.185-213.

4. Marc Milner, North Atlantic Run, pp.177-213.

5. Clay Blair, Hitler's U-boat War: The Hunted, 1942-1945 (New York, 1998), pp.185-189, 252-254.

6. Martin Middlebrook, Convoy (New York, 1976) and Jürgen Rohwer, The Critical Convoy Battles of March 1943 (London, 1977), tell the story of these convoy battles.

7. Directorate of History and Heritage (DHH), 89/34, Convoy reports. Averages for a full cycle of 11 convoys in November-December 1942, and 10 convoys in February-March 1943.

8. DHH, 79/599, Great Britain, Air Ministry, Air Historical Branch, The RAF in Maritime War, Vol. 4, pp.26-27, 32.

9. NARA, RG 218, Box 6, file 13W3 13 15K, "History of the Joint Chiefs of Staff", Chapter 4, pp.103-104, 110.

10. Great Britain, Admiralty, Historical Section, The Defeat of the Enemy Attack on Shipping, Vol. 1A, pp.95, 293-294; Stephen Roskill, The War at Sea (London, 1956), II, pp.366-368, 373; and John Terraine, Business in Great Waters (London, 1989), pp.584-587.

11. DHH, 89/34, Convoy reports. Averages for a full cycle of convoys in April-May 1943.

12. DHH, 79/599, The RAF in Maritime War, Vol. 4, pp.23-26; and Stephen Roskill, The War at Sea, II, pp.364.

13. DHH, 79/599, The RAF in Maritime War, Vol. 4, pp.21-27; and W.A.B. Douglas, The Creation of a National Air Force (Toronto, 1986), pp.537-567.

14. W.S. Chalmers, Max Horton and the Western Approaches (London, 1954), p.188.

15. Great Britain, Admiralty, Tactical and Staff Duties Division, German Naval History: The U-boat War in the Atlantic, II, pp.102-103.

16. Michael Gannon, Black May, pp.115-240; W.A.B. Douglas and Jürgen Rohwer, "The Most Thankless Task Revisited", in James A. Boutilier, ed., The RCN in Retrospect, 1910-1968 (Vancouver, 1982), pp.228-233; and The U-boat War in the Atlantic, II, pp.104-106.

17. Michael Gannon, Black May, pp.334-349.

18. Michael Gannon, Black May, pp.353-377; and Clay Blair, The Hunted, pp.332-338.

19. DHH, 79/446, Befehlshaber der U-boote (BdU) War Diary, 22-24 May 1943; Karl Dönitz, Memoirs: Ten Years and Twenty Days (London, 1959), p.341; and The U-boat War in the Atlantic, II, pp.112-113.

20. Axel Niestlé, German U-boat Losses during World War II, p.193; and Jürgen Rohwer, Axis Submarine Successes, 1939-1945 (Annapolis, MD, 1983), pp.164-167. This figure includes the one merchant ship lost from ONS 5 at the end of April.

21. Stephen Roskill, The War at Sea, II, pp.376, 380-381.

22. DHH, 79/446, BdU War Diary, Feb.-May 1943; and Clay Blair, The Hunted, p.166.

23. Michael Gannon, Black May, p.111; and Marc Milner, North Atlantic Run, p.181.

24. Defeat of the Enemy Attack on Shipping, Vol. 1A, pp.293-294.

25. DHH, 85/588, United States Navy, Eastern Sea Frontier War Diary, October 1942; and DHH, 79/599, The RAF in Maritime War, Vol. 3, pp.514-521.

Return to home page

Copyright © Rob Fisher 2000