Before examining the May convoy battles, it is worth noting another result of the ORU’s work. (Blackett had become Director, Naval Operational Research.) It was appreciated that the number of escort ships with a convoy reduced the loss rate and that air cover of about eight hours a day decreased ship losses by a third. However:
Since it was by no means safe to rely on the increase of air support to stop the crippling ship losses of the autumn of 1942, an energetic search was made for some other measures which could be put into operation quickly. Detailed attention was given, therefore, to the organizational aspects of the Atlantic convoy system. Perhaps some alteration in the organization of the convoys might conceivably improve the situation.
Hitherto, organization of convoys and escorts had been ‘a matter of chance’. The Admiralty had defined some broad principles, including the belief that large convoys were more dangerous and that, therefore, the optimum was about forty vessels. Sixty was the maximum and larger convoys were not permitted. A rough guide to escort numbers was also in place; known as the 3 + N/10 rule, this stipulated a minimum of three escort vessels for a very small convoy with an additional ship for every ten merchantmen in the convoy: a convoy of twenty would have the minimum three escorts plus two; in this case the value of N was twenty. Likewise, the largest convoy, of sixty, would have nine escorts: three plus six, the value of N.
The basic assumption was that every convoy, irrespective of size, was equally safe or, at least, likely to suffer the same percentage rate of loss. Where the 3 + N/10 rule originated, no one seemed to know, but Blackett points out that it ‘could be shown to be not consistent with the view that small convoys were safer than large’. That inconsistency is demonstrated by his theoretical example of running three twenty-ship convoys, each with the five escorts of the 3 + N/10 rule, against pooling those convoys and their escorts. By applying the rule, a sixty-ship convoy should enjoy only a nine-ship escort whereas pooling the three convoys and their escorts would produce an escort of fifteen.
Examination of records of ships lost in differing-sized convoys over the previous two years showed clearly, and surprisingly, that larger convoys had suffered relatively smaller losses.
The figures were startling. Dividing convoys into those smaller and those larger than forty ships, it was found that the smaller convoys, with an average size of thirty-two ships, had suffered an average loss of 2.5 per cent, whereas the large convoys with an average size of fifty-four ships, had suffered only a loss of 1.1 per cent. Thus large convoys appeared to be in fact over twice as safe as small convoys.
Although the calculations appeared reliable, Blackett knew that the Admiralty would be reluctant to introduce larger convoys, and so ORU began gathering evidence to strengthen the case for a change. Accounts by captured U-boat personnel proved very enlightening, and after several weeks of intensive work sufficient evidence had been gathered. It was discovered that the chances of any individual merchantman being sunk during any voyage depended on three factors:
(a) the chance that the convoy in which it sailed would be sighted; (b) the chance that, having sighted the convoy, a U-boat would penetrate the screen of escort vessels around it; and (c) the chance that when a U-boat had penetrated the screen the merchant ship would be sunk. It was found: (a) that the chance of a convoy being sighted was nearly the same for large and small convoys; (b) 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; and (c) that when a U-boat did penetrate the screen, the number of merchant ships sunk was the same for both large and small convoys – simply because there were always more than enough targets.
The researchers concluded that, given the same linear escort strength, the same absolute number of sinkings could be expected, irrespective of convoy size, and the percentage of losses would be inversely proportional to size. Thus the number of convoys sighted should be reduced by decreasing the number of convoys run, which could be achieved by increasing convoy size. After ‘some weeks of earnest argument’ new orders were issued in spring 1943; before long convoys of up to a hundred ships were crossing the Atlantic.
Coastal Command was receiving ASV III for its aircraft towards the end of 1942 and aircraft so fitted became operational early in 1943. Terence Bulloch, who was resting from operations, took part in testing ASV III and was impressed with its quality which ‘was many times easier to interpret and presented the information in a readily acceptable form, the Planned Position Indicator’. ASV III’s beams could not be detected by Metox and so the advantage conferred by that equipment waned. Boats crossing the Bay by night had felt much safer with Metox to warn of the presence of an ASV II-equipped aircraft, but, once again, they found themselves being attacked without warning. Surfaced U-boats elsewhere were taken by surprise when a Coastal Command aircraft appeared from low clouds. Although U-boats had been fitted with anti-aircraft armament, the conning tower and deck of a submarine was not a steady gun platform, but U-boats that fought it out on the surface could inflict serious damage, often bringing down aircraft.
With more VLR and LR aircraft available, the air gap was being reduced steadily. The advent of escort carriers and MAC-ships finally removed it completely. During May the U-bootwaffe would learn that there was no safe area in the North Atlantic. Land-based aircraft and flying boats operated from Newfoundland, Iceland and Northern Ireland, and the US Navy had opened bases in Greenland, from which some B-24s could operate in favourable conditions; the Greenland Fleet Air Group included detached elements of US Navy squadrons which, in April 1943, included two PBY-5A Catalinas from HQ Squadron Fleet Air Wing 7, and three Lockheed Venturas and three Catalinas from VB-126.7 Coastal Command disposed almost thirty squadrons, including four-engined Liberators, Halifaxes, Fortresses and Sunderlands, and twin-engined Catalinas, Wellingtons, Whitleys and Hudsons. The Handley Page Halifax was one of Bomber Command’s ‘heavies’ but some had been lent to Coastal and proved excellent long-range aircraft.
Some of these aircraft were called into action again in support of B7 which, having fought HX231 through to the UK, left Londonderry to escort ONS5 to North America. The convoy, forty-two ships, left Liverpool on 21 April with B7 joining next day. HMS Duncan, Gretton’s own ship, was back and he also had Tay and the corvettes Loosestrife, Pink, Snowflake and Sunflower. Although Vidette remained with the group, she had sailed earlier for Iceland to escort three ships joining the convoy. As well as two rescue trawlers, Northern Gem and Northern Spray, there were two tankers to refuel escorts; one would prove of no use since its hose system was canvas rather than rubber. Initially the voyage was uneventful but for the weather, which was so bad that two ships collided with one making for Iceland for repairs.
Apart from a false alarm on the 24th there were no signs of U-boats until four days later when HF/DF intercepted a signal from straight ahead and close to the convoy. In between, in the afternoon of the 24th, a Fortress from Benbecula-based No.206 Squadron attacked and sank U-710, which was lying ahead of the convoy on its course. It appeared that ONS5 might escape interception but the boat from which the signal was heard on the 28th proved to be U-650, which shadowed the convoy all day, awaiting another fourteen boats which had been ordered to join her and attack that night. With inclement weather preventing aircraft interdicting the gathering wolfpack, it became obvious that a battle loomed.
The other boats being called up were U-533, U-386, U-231, U-532, U-378, U-192, U-258, U-552, U-954, U-648, U-209, U-413 and U-710, forming the Star wolfpack, which included the boats of the former Meise group and six that had deployed with Meise two days earlier. They had been patrolling between Iceland and Greenland, just south of the amended course Western Approaches Command had advised the convoy to take that morning, a course intended to keep the convoy as far as possible from known U-boat concentrations. Oberleutnant Ernst von Witzendorff, of U-650, could only see six ships and, from his sighting report, U-boat headquarters assumed that he had spotted part of ONS6, which had yet to leave Liverpool. Since von Witzendorff was forced to submerge several times by Catalinas of VP-84 from Iceland, his failure to see more ships can be understood. Nonetheless his signal brought four boats to join him that day; escorts attacked the U-boats that night.
HMS Sunflower depth-charged U-386, damaging her, while U-532 and U-650, having launched torpedoes at Snowflake and Duncan, were also depth-charged. During the 29th U-532 was attacked again, this time by Tay, while U-258, in a daytime attack, sank an American merchantman and U-528, damaged by a Catalina of VP-84, had to make for home. Both U-386 and U-532 sustained such serious damage that they also had to break off. The destroyer Oribi was ordered to leave SC127 to augment B7 while 3rd Support Group, with the destroyers Offa, Impulsive, Penn and Panther under Captain James McCoy DSO, an Irishman with an Italian wife, was ordered from St John’s to support ONS5. The weather was so bad that McCoy’s group had difficulty finding the convoy and did not make contact until 8.00pm on 2 May.
With poor weather and reduced visibility on 30 April and 1 May, the U-boats lost contact after an unsuccessful night attack by U-192. Conditions were such that the convoy had to heave to in gale-force 10 winds, which resulted in some ships becoming separated from the main body; escorts could not refuel due to heavy seas, the threat from icebergs, and pack ice. Six ships straggled, but Northern Spray tried to keep them together, while a further half dozen were gathered together to be shepherded by HMS Pink.
After 3rd Support Group joined, the convoy came out of the ice but scattered widely. Had the weather been good the stragglers would have fallen victim to U-boats but these had also suffered and had been ordered to break off operations late on the 1st.14 On the 3rd Duncan, short of fuel and with another gale raging, had to make for St John’s and Lieutenant Commander Sherwood of Tay assumed the duties of senior officer B7. Next day, two of McCoy’s group, Penn and Panther, had to leave to refill their tanks. Western Approaches Command then ordered 1st Support Group to reinforce the escort; this group, under Commander Godfrey Brewer, included the sloop HMS Pelican, the cutter Sennen and four Rivers, Rother, Spey, Wear and Jed.
German attention switched for a time to SC128 with boats of the Star and Specht groups forming a patrol line across its expected route. This patrol line, Fink, was augmented by another, Amsel, formed by boats newly-arrived from France and further divided into Amsel 1 to 4. In total, more than forty boats deployed. The German plan was foiled, thanks to an Ultra decrypt that allowed SC128 to be re-routed away from the waiting wolfpacks. Allied aircraft also deployed and a Royal Canadian Air Force Canso (the RCAF name for the Catalina) attacked and damaged U-209; another Canso damaged U-438. U-209 subsequently disappeared and may have sunk as a result of the damage. At the time a Canso was believed to have sunk U-63016 but this boat was actually destroyed by Vidette on 6 May.
While SC128 escaped the planned ambush, ONS5 ran into the concentration, meeting the patrol line from the other side. Weather conditions meant that the convoy had only progressed by twenty miles. When U-628 made its sighting report the Fink and Amsel 1 and 2 groups were ordered to attack ONS5. ‘The real battle was joined after dark on the 4th’. By then the weather had improved, allowing some thirty ships to re-assemble. Pink and Northern Spray were still escorting stragglers, one of which, the SS Lorient, became the first victim of the renewed assault. Lorient was torpedoed and sunk by U-125 which was destroyed two days later by gunfire from Snowflake, having been rammed by Oribi.
That night Oribi, Snowflake and Vidette depth-charged and damaged three boats, which withdrew, although U-514 resumed its patrol some days later. The U-boats pressed home their attacks with U-707 attacking from the front, diving under, passing below, and attacking and sinking a straggler. U-628 penetrated the screen to fire five torpedoes at five targets but only damaged a single ship, which the same U-boat subsequently finished off. At much the same time, U-264 made a similar attack with five torpedoes, four of which found targets: the American West Maximus and the British Harperley were sunk. After that U-358 fired three torpedoes sinking Bristol City and Wentworth.
Attacks continued next day. Dönitz had taken command, exhorting his commanders to seize any opportunity. Serial submerged daylight attacks followed, putting considerable pressure on the escorts. U-638 sank the steamer Dolius but, hunted down by Sunflower and Loosestrife, was sunk by the latter. Another multiple-torpedo attack, by U-266, claimed three ships with four projectiles, the Norwegian Bonde and the British Gharinda and Selvistan. Offa damaged U-266 which was sunk ten days later by a Halifax of No.58 Squadron. An American steamer from the stragglers with Pink was torpedoed by U-584 in the afternoon while Pink was engaging U-358, which had made the first attack on the stragglers, thus allowing U-584 a clear run. U-358 had to return to France.
That evening a Liberator of No.120 Squadron from Iceland spent a short time overhead but was at the limit of its endurance and could not loiter, even though it was a VLR machine. Undeterred by the Liberator, the U-boats continued gathering. Before the light faded Tay had spotted seven boats but there were no fewer than fifteen already in contact. Dönitz continued encouraging commanders to greater efforts as he ‘anticipated that the night would bring some hard fighting, but also considerable success’.
With the escorts running low on depth charges, it threatened to be a very bad night. However, as evening was falling, ONS5 sailed into a thick fogbank, which worked to the advantage of the escorts and merchantmen since the former could still find the U-boats with their radar, which Metox could not detect. Nonetheless, attacks continued throughout the night, about twenty-four being made from every direction except ahead, before the attackers eased off at 4.20am on the 6th. In a very confused situation the escorts, well trained and very experienced, had gained the upper hand. Dönitz’s headquarters’ war diary reads:
A golden opportunity had thus been ruined by fog; no further success was scored by any U-boat. During this fog period alone fifteen boats were attacked with depth charges and six of them were located by destroyers, surprised on the surface and engaged with gunfire. The lack of any means of counteracting this radar location undoubtedly left the boats in an inferior and, indeed, hopeless position.
Seven U-boats were lost, including U-531 and U-630, both sunk by Vidette. Loosestrife depth-charged U-575 without success before obtaining a radar fix on U-192. As the corvette loomed out of the mist at about 500 yards, the U-boat launched two torpedoes at her, both of which missed. A pattern of depth charges, set for shallow detonation, destroyed the still-surfaced submarine. Snowflake chased off U-107 with depth charges before beginning a search for another four boats that had appeared on her radar. While Snowflake was engaging some of those with gunfire, Oribi appeared from the fog and rammed U-125 which the corvette then sank with gunfire. Offa, which had made five attacks before midnight, damaged U-223 with gunfire and depth charges. Also damaged was U-533 but both escaped destruction; U-223 was damaged again on the 11th when she was depth-charged to the surface by Hesperus which then rammed her; she survived the encounter and limped home. While escorts and U-boats were fighting, Brewer’s 1st Support Group arrived and joined in the fray. Pelican found U-438 by radar and closed to within 300 yards before being spotted. The boat crash-dived too late as a fusillade of shallow-set depth charges sent her to the bottom. Meanwhile Sennen raced to join Pink and her charges, en route attacking both U-650 and U-575 with depth charges and Hedgehog. Neither sustained serious damage. Spey fell into station behind the convoy and drove off U-634 with gunfire; the boat was hit by two rounds but not damaged seriously.
Success, let alone ‘considerable success’, had eluded Dönitz. When he realized the scale of his losses he called off the engagement. Although twelve merchantmen had been sunk, not including the vessel on 29 April, he ‘regarded this convoy battle as a defeat’. He writes that the 10-centimetre radar with which the escorts were equipped had ‘a direct and extremely adverse effect on the fighting of the individual U-boat’. For the Royal Navy, the ONS5 battle was remarkable in that most defensive work had been done by surface escorts who, in spite of what looked like overwhelming numbers, fought off a very large wolfpack; in all, as many as fifty-five U-boats deployed against ONS5. The US Navy Intelligence Service estimated that, on the evening of 5 May, fifteen boats were in contact with the convoy while another ten to fifteen were no more than fifteen nautical miles away. Captain James McCoy stated that ‘the convoy was threatened with annihilation’. ONS5 was also memorable as the last time so many merchant ships were lost in convoy.
Dönitz had not given up the fight and when B-Dienst provided details of the next two eastbound convoys, HX237 and SC129, he ordered thirty-six U-boats to attack. The convoys were to take more southerly routes, passing not far from the Azores. HX237 was protected by C2 under Lieutenant Commander Evelyn Chavasse DSC, which included HM Ships Broadway, Lagan and Primrose and HM Canadian Ships Chambly, Drumheller and Morden with HM Trawler Vizalma and a tug. Thus half of this RCN group was made up of British ships and the Irish Chavasse was a Royal Navy officer. For its journey across the Atlantic HX237’s escort was reinforced by 5th Support Group with Biter and the destroyers Opportune, Obdurate and Pathfinder; commanding the group was Captain Conolly Abel Smith, whose maternal grandfather was the Irish VC John Augustus Connolly.
At first Abel Smith refused to place his carrier within the convoy from where his destroyers could ‘reinforce the very sketchy close screen’ and act as a striking force. Instead, he planned to operate between twenty to fifty miles from the convoy. That was not the only friction that Chavasse had to face at the start of the journey:
On 6 May C2 … sailed from St John’s [although] the Vizalma, the tug and one merchant ship sailed late to intercept, and Biter and his boys sailed from Argentia. The usual thing happened, as so often in May: fog closed down. The convoy itself, unknown to me, became almost completely scattered and disorganized …, and we had the greatest difficulty in finding it. Most of the Local Escort had lost touch, and the situation was most confused. Homing on to them by radar was useless, and we had to resort to a lot of chatter by R/T (radio/telephone) in an effort to make contact. No doubt U-boats were avidly listening and licking their chops in anticipation …; and in fact I was rebuked from shore for using too much wireless. But there was no alternative. At dusk on the 6th we did find a few ships of the Western Local Escort … but no convoy, and throughout the night we chugged along together on convoy course and speed.
As dawn broke, on the 7th, the weather was clearer. The Biter, who was to the northward of us, put up an air search, found the convoy, and signalled to me a course to steer, which turned out to be wildly wrong. I put all my ships on an extended screen at visual distance from each other and, by the greatest good fortune, the ship at the extreme end of the screen sighted the convoy on a totally different bearing. … By afternoon we were in touch.
Abel Smith’s refusal to bring the support group, including Biter, into the convoy was causing problems. On 8 May, a slightly misty day, Biter flew off aircraft which failed to find the convoy ‘and were therefore quite useless’ to Chavasse in detecting and reporting U-boats since the Swordfish crews did not know the convoy’s position. Since the position of Biter and her destroyers relative to the convoy had not been established, the carrier could not be used to ‘fix’ any high-frequency signals that might be intercepted. With deteriorating weather on the 9th, Biter was unable to fly off any aircraft. That was the day the first U-boat appeared. At the time Broadway was refuelling.
Broadway received what was called a ‘close-range B-bar’, in other words a wireless transmission from a U-boat, dead astern of the convoy, and not far away. We were temporarily connected by hosepipe to the tanker, and while we were hastily disengaging I ordered Primrose, who was stationed astern of the convoy, to search and, if possible, attack. At the same time I informed Biter, but she could not do anything useful, as she didn’t know where we were. Primrose, however, actually spotted the U-boat, which hastily dived, but she did not make asdic contact and she later regained her station on the screen astern of the convoy.
By this stage, the Admiralty had recognized the folly of Biter operating so far from the convoy and, on 10 May, Abel Smith was told to take 5th Support Group into the convoy, placing himself under the orders of the escort commander. The commodore made space for the ships in the heart of HX237. From that point on, ‘we never looked back’ as Biter did magnificent work, flying off aircraft in spite of foul weather and never once refusing a request to send aircraft out on a sortie; the carrier’s embarked unit was No.811 NAS, with a mix of Swordfish bombers and Wildcat fighters, still known as Martlets in British service. In a series of ‘superb feats of sea-plus-airmanship in the most difficult conditions imaginable’, Biter made a tremendous contribution to the safety of HX237. C2 and Abel Smith’s group had become that most valuable asset, a co-ordinated team as Swordfish kept watch over the seas around the convoy while the fleet destroyers, each capable of over 30 knots, chased down any submarine spotted near the merchantmen and C2, ‘perhaps a little more experienced in these matters’, provided the final close protection.