Chapter 7. Pause, Counterattack and Victory: June 1943–May 1945

When Dönitz withdrew his U-boats from the North Atlantic in June 1943, he decided to deploy them against the routes used to reinforce the US forces in North Africa from the States. It was not a wise decision, as the USN used five escort carriers in the area, accounting for sixteen U-boats for the loss of only six ships from convoys. Another mistake was to order U-boats to cross the Bay of Biscay in groups on the surface by day, a decision that cost twenty-nine U-boats as the obvious counter of aircraft also working in groups was introduced. By the beginning of August, it was realised that the Metox radar detector was ineffective, and a new detector proved no better since it still did not recognise the 10-centimetre radar that was being used. Dönitz changed the transit rules, with boats crossing the bay independently, surfacing only by night to charge batteries and hugging the Spanish coast to confuse aircraft radar. Sinking in the bay transit dropped dramatically to about one per month.

Withdrawal from the principal North Atlantic convoy routes was intended to be only temporary, pending the introduction of new technology. Existing submarines were to have their AA armament greatly increased and, more importantly, they were to be equipped with torpedoes that would home on the noise from escort vessels’ propellers. New types of submarine were to be brought forward – first the fast battery boats, Types XXI and XXIII, and then the much faster hydrogen peroxide boats of Walter design.

The tonnage of merchant ships sunk was below 100,000, except in July (238,000). The total number of U-boats was fairly constant, at about 429. Roughly half were available for operations, the remainder being on trials or training. These figures equated to some fifty tons sunk per U-boat-day at sea.

Table 7.1: Causes of U-boat sinkings, June–September 1943

Surface vessels	16½
Aircraft	64½
Submarine	2
Mines	1
Accident and unknown	5

As can be seen from table 7.1, the outstanding feature of the period was the number of sinkings due to aircraft. This was partly a result of Dönitz’s ill-considered policy of fighting on the surface and crossing the Bay of Biscay in groups, but probably owed more to better aircraft, better weapons and better training.

From Counterattack to Victory: October 1943–May 1945

By September 1943, Dönitz felt confidence in the U-boats’ ability to re-enter the Atlantic battle. They had enhanced anti-aircraft armament and a new radar detector (Wanz) to protect them from air attack, and the new homing torpedo (T5) with which they could strike against surface escorts. The first big test came on 20 September 1943, when twenty-two U-boats launched an attack on the combined forces of ON202 and ONS18. By the time action was broken off because of fog on the 23rd, six merchant ships totalling 36,000 tons had been sunk, as had three U-boats.

Phase 6 of the Atlantic battle. The lack of merchantmen sinkings shows that the battle had been won. The Bay of Biscay was a major killing ground with many U-boats destroyed.

The attack on the escorts had interesting consequences. Twenty-four T5 torpedoes had been fired and it was claimed that twelve escorts had been sunk, with three more probables. This claimed figure was about what had been expected – and hoped for – and was accepted by Dönitz’s staff. Unfortunately, it was wildly in error. The confusion of a night battle, with torpedoes exploding in the wake and numerous depth charge explosions, combined with wishful thinking, led to exaggeration. The true figure was three escorts sunk and another damaged beyond repair.¹ The Germans, believing their own claims, saw the T5 as an excellent escort-killer.

On 15 October a pack attempted to attack ONS20 but a strong air escort kept them dived, making shadowing difficult and a concentrated attack impossible. Only one merchant ship was sunk but six U-boats were destroyed – four by aircraft, two by ships. This was the end of the wolf pack; for the rest of the war, single, submerged attack was adopted but only in small numbers. Through to May 1944 the total number of U-boats remained fairly constant at around 450, but the number at sea was halved from eighty-six to forty-three and the amount of shipping sunk per U-boat-day at sea fell to some twenty to thirty tons.

Cause of sinking	October–May 1944	May 1944 to the end of the war
Surface vessels	59½	85½
Aircraft	83½	163½
Submarine	3	8
Mines	3	20
Accident and unknown	13	41

Technical Developments

The withdrawal of the U-boats from the North Atlantic gave both sides the opportunity to develop new systems and technology. On the German side, there was the disastrous attempt to beat off air attack by increased gun power and the rather more successful tactic of attacking the escort vessels using homing torpedoes. Refuelling from milch cows came and went (chapter 5).² The RN measures were the introduction of depth-finding asdics and two successful ahead-throwing weapons. Additionally, countermeasures to homing torpedoes were brought into use.

Phase 7 of the battle shows a high proportion of U-boats sunk to merchant ship losses, reflecting the balance of power that had shifted decisively to the escorting forces.

The final phase of the submarine war, after the D-Day landings in Normandy, shows a dramatic shift of theatre to the waters around the UK and the western coasts of Europe.

Anti-Aircraft Armament of U-Boats

From 1940, the armament of Type VII U-boats was increased by mounting a single 20mm on the bridge. This was rarely used, as most COs preferred to dive if an aircraft was sighted. By mid-1942, the armament had been increased to two twin machine guns on the bridge and a 20mm on a platform below the rear of the bridge (to be replaced by a twin when available), By June 1943, no U-boat was allowed on operations unless it had at least two 20mm. From August that year, the minimum standard was ‘Conversion IV’, which involved two twin 20mm on the bridge and a quadruple 20mm on the platform below the bridge. The latter was to be replaced by a single 37mm when available from October 1943.³

There were a number of experimental fits. A very few boats were fitted as aircraft traps, with a very heavy armament of one 37mm and two quadruple 20mm. It is not often appreciated that additional topweight on a submarine has a twofold effect. The total weight of a submarine is fixed, equal to the weight of water displaced by the hull and appendages. This means that additional weight has to be compensated for by removal of ballast. Since the ballast is usually in the very bottom of the boat, this will cause another rise in the overall centre of gravity. Some offset was gained in April 1943, when removal of the low-angle 88mm was permitted. These additional appendages added to the drag when submerged, with a considerable loss of speed and endurance.⁴

Lagan was an early victim of the German homing torpedo (GNAT or T5). Few if any of the ships surviving were repaired.

The standard Conversion IV fit was a powerful armament; two or more U-boats in a group formed a dangerous target and the RAF paid a heavy price for their victories. However, aircraft could form groups as well, and wait till the submarine began to submerge – a slow process with the numerous gunners to get down a single hatch. It might also be possible to call up surface ships. Once the boat had submerged, the Type 24 mine could be used (see next section).

Homing Torpedoes

Pursuit of the distinction of being ‘the first’ is rarely profitable, as similar developments are often taking place in several establishments at the same time and the question of who is first is a matter of chance, or of who is willing to take the biggest risk. The German T4 Falke was the first homing torpedo to be used but it was slow, at twenty knots, and only suitable for use against merchant ships. Only about 100 were made, of which some thirty were used in action from January 1943.

The American Mk 24, often referred to for security reasons as the Mk 24 ‘Mine’ (informally as ‘Fido’), was a far more effective weapon and entered service in the summer of 1943. It was dropped from aircraft and had a 92lb Torpex charge and a range of 4,000 yards at a speed of twelve knots. It is said that 346 were dropped for 68 kills and 33 damaged.⁵ US aircraft dropped 142 for 31 kills and 15 damaged, the rest being by British aircraft. The first sinking appears to have been that of U-266 by an RAF Liberator of 86 squadron on 14 May 1943. It was three times as effective as an airborne depth charge. Dropped from 250 feet at 125 knots, it had a running time of ten minutes.

Affleck, another victim, seen laid up at Barrow after the war.

To maintain secrecy, very strict conditions on the use of the Mk 24 were imposed. It was not to be used near a coastline; it was not to be dropped if it could be seen by another vessel. No mention was allowed in the press, so that early post-war accounts did not mention it and, even now, the USA has not received the credit due for this very successful weapon.

The best-known homing torpedo, indeed the only one known to many people, was the German T5 Zaunkönig (code-named GNAT – German Naval Acoustic Torpedo), specifically designed to attack escorts. It had a speed of twenty-five knots on a sinuous path, which meant that ships moving at over eighteen knots were fairly safe. Ships moving at less than eight knots did not generate enough noise to attract the torpedo.

The T5 had a range of 5,700 metres. It was a very complicated device, with eleven valves and many metres of wire. Homing distance varied greatly but 450 metres for a fifteen-knot ship seems reasonable. The homing feature was inoperative for the first 400 metres of the run to protect the U-boat, and the firing pistol only sensed targets above the depth of the torpedo.

Some 640 were fired for only fifty-eight hits, much poorer than the 20 per cent success rate of other U-boat torpedoes. This was because it was often fired against difficult targets and because an effective decoy was soon introduced. The homing/firing mechanism worked well and hits were in the region of the propellers. Damage was severe; the stern was wrecked, the propellers and tail shaft were badly bent and, less obvious, the ship whipped, with buckling of the upper deck amidships in some cases. In view of the large number of escorts completing in British and US yards, few, if any, of the ships that survived a GNAT hit were repaired.

News of the German development of homing torpedoes reached the UK in the autumn of 1942 and work started on a decoy in September. ‘Foxer’ began to enter service soon after the introduction of the T5 but it took time to equip all escorts. Foxer consisted of a number of pipes that jangled together, making 10–100 times the noise of the ship. It was an awkward beast to handle and the noise that interfered with the GNAT also severely degraded the performance of the escort’s own asdic. The life of a Foxer unit was short because of its inherent self-battering. It was gradually replaced by ‘Unifoxer’, a simpler device developed in Canada.

Depth-finding Asdics and Squid⁶

Some interest had been shown in depth-finding asdic in 1939 and an experimental unit was fitted in the patrol sloop Kingfisher, but this was damaged at Dunkirk and removed. By November 1941, work had commenced on a different device. Two sets were developed in parallel – the Q attachment for fitting to existing sets similar to type 128, and the much more elaborate 147 for new construction.

The Q attachment had a wedge-shaped beam only 3º wide in the horizontal plane. It was mounted below the main transducer and could maintain contact with a submarine at a depth of between 300 and 700 feet at fairly close range, usually about 1,500 yards. It was tried in Vanquisher in February 1943. Production started in April 1943 and it entered service soon after, as ship-fitting took only two to six days. Ships with retractable domes needed docking; those with detachable domes did not.

Type 147 was a much more elaborate device. It was carried in a ‘sword’ ahead of the main dome. Its fan-shaped beam was 3º wide in the vertical plane and 65º in the horizontal. Ship trials took place in Ambuscade in May 1943. It would normally pick up the U-boat at about 800 yards and would put the correct setting on the Squid projectiles automatically.

The first operational set was installed in Hadleigh Castle early in September 1943. She visited the USA in December 1943, where her capability was much admired. Forty sets were supplied to the USN and 45 more by Canada.

Ahead-Throwing Weapons

After the failure of the early ‘stick bomb’ weapon tried in Torrid, there was little effort on ahead-throwing weapons until late 1939, when there was a proliferation of proposals. A study came up with two main types: a small charge (20lb) with a contact fuse and a much bigger version (200lb) with a depth-sensitive fuse, both of which would eventually enter service. Initially, there was no coordination of the various schemes and when, in late 1940, DNO was put in charge, there was a good deal of acrimony between proponents of different weapons.

The first serious proposal was the ‘Fairlie mortar’, in which two sets of ten mortars were fitted either side of the forecastle. The projectiles had a 20lb charge and formed a circular pattern. The mortars were trained by rocking about a horizontal axis. Though this weapon did not go into service, many of its features were adopted. In January 1940, Fairlie began work, with Vickers-Armstrongs, on a three-barrelled breech-loading mortar. There was also a proposal by Thornycroft for an array of five big mortars that would fire standard depth charges. This was fitted in Whitehall and tested in mid-1941. It appears that the trajectory of a depth charge in the air was too variable for operational use. Lastly, there was a spigot mortar originally intended for army use but adapted by DMWD to fire contact-fused projectiles.

Whitehall was fitted for trials with five large mortars firing standard depth charges. The mounting is seen here in A position.

The Hedgehog fired a salvo of bombs ahead of the ship. The bombs had a contact fuse so there was no ‘bang’ from a near miss. The charge weight was 30lb which was marginal but a change to Torpex filling was satisfactory. The drawing by John Lambert (opposite) shows a typical arrangement of a partially loaded mount.

All these weapons needed accurate information on the position, including depth, of the submarine, which became available from the 144/147 combination. The main objection to the small charge was that it only exploded on contact and hence there was no morale effect from a near miss; but on the other hand, asdic contact was not lost in the aftermath of the explosions. A more serious objection came from DNC, who maintained that a 20lb charge was insufficient to rupture a pressure hull.⁷ Goodall’s diary says that the original charge was proposed at 5lb but this may have been a misunderstanding. After heated argument, the charge weight was increased to 30lb and in service even this bigger ‘bang’ proved marginally effective. The smaller weapon had the advantage of a smaller recoil, needing less structural support, and was therefore more suitable for small ships.

Development of the spigot mortar was actively pursued during the first half of 1940. In final form there were twenty-four charges in four rows of six. Firing was by electricity, when a spring-loaded spigot would be forced up into the base of the projectile, igniting the propellant charge. This would send the projectile on its way and at the same time compress the spring for the next round. The projectiles were fired in pairs, further minimising the recoil loading on the deck. They fell in a circular pattern about 130 feet in diameter.

A post-war photo of two US Navy destroyers. The further ship has just fired a double salvo of Hedgehog and the splashes as they enter the sea may be seen ahead of her.

The first tests were held in February 1941 from the pier at Weston-super-Mare (HMS Birnbeck), with the advantage that the projectiles could be recovered at low tide. Sea trials followed, first in Enchantress and later in Westcott (the latter ship scored one of the first kills using Hedgehog on 2 February 1942, sinking U-581). Production followed quickly.

Initial results were very disappointing. Commanding officers were reluctant to use Hedgehog because there was no bang from a near miss.⁸ Hackmann tells the story of Lotus, which did not use Hedgehog until all her depth charges were expended and then sank her contact with the first salvo from her Hedgehog. There were mistakes in installation and a lack of maintenance information. It did not help when Escapade destroyed her own bridge with a misfiring salvo. Command even issued instructions that Hedgehog must be used. These problems were slowly overcome, though it was late in 1944 before the desired success rate of 20 per cent kills per salvo was achieved. The effectiveness of Hedgehog was soon increased by changing the explosive charge to Torpex, giving the equivalent of 50lb of TNT.

The single Squid fired a salvo of three projectiles each weighing 390lb with a charge of 207lb minol. The Loch class had a double mount firing a salvo of six bombs. The depth was set by asdic Type 147 at the instant of firing. It was by far the most effective A/S weapon of the war. (John Lambert)

Escapade with severe damage to her bridge due to a misfiring Hedgehog. (WSS)

Hedgehog was adopted by the USN where it scored its most spectacular success. In May 1944, the uss England sank six Japanese submarines with twelve salvoes in eleven days. The US developed a somewhat similar weapon for small craft, Mousetrap, firing six rocket-propelled projectiles with very little recoil.

Fears over the small size of the charge in Hedgehog projectiles led to a revival of the Fairlie mortar as ‘Parsnip’. This had two rows of ten mortars, firing projectiles containing 60lb of explosive. It was tried in the ubiquitous Ambuscade late in 1942 but was not adopted.

Instead, the Squid was developed, also owing much to the Fairlie mortar. The single squid, used in the Castle class, was a three-barrelled mortar firing 390lb projectiles with a 207lb charge of minol. The projectile entered the sea at a range of about 300 yards and had a sinking speed of 44ft/sec, twice that of Hedgehog, down to 900 feet. The three charges formed a triangle with 120-foot sides. A double Squid, as used in the Loch class, had two such systems set 60 feet apart in depth. The mortars could be tilted by 30º to allow for roll and yaw and small errors in course. It could be tilted to the horizontal for rapid power-loading.

The single Squid Mk IV mounting in a workshop setting. (John Lambert)

Sea trials took place in May 1943 from Ambuscade – again – and production sets entered service in Hadleigh Castle in September. In general, fitting was limited to new construction but Escapade was fitted after the destruction of her bridge by Hedgehog. The great advantage was that the depth setting on the charges was set and updated continuously by asdic Type 147 until the moment of firing. It pleased the operators as it provided an explosion from a near miss – though with a near 50 per cent success rate from a double salvo, there were not all that many near misses. One of the early successes was that of Loch Killin, which in July–August 1944 sank U-333 and U-736 with only three salvoes. Early in 1945, the system was modified to fire against surfaced submarines, under radar control and using a 20-foot depth setting.

The Loch Class Frigates

The Loch class frigates and their associated weapon system were the first design to incorporate both a full understanding of the requirements of the Battle of the Atlantic and the capacity of wartime industry. The result reflected great credit on the Assistant Director, A W Watson, and his constructor, A E Kimberley.⁹ Watson had designed the P boats of World War I.

Studies on two A/S designs began at the end of 1942 under A E Kimberley. One was basically an enlarged Flower with a single shaft and armed with a single Squid (see later section). The other was a twin-screw ship somewhat similar to the Rivers, with a twin Squid. The larger ship was preferred but, with the recognition that some smaller yards would not have slipways long enough to build the bigger ship, it was decided to build both designs. The Staff Requirements (that is, the official specification of requirements by the Naval Staff) were agreed and the design was frozen for two years from January 1943. The implications were a matter of concern; Goodall wrote,

This question brings us up hard at the onset against the difficulties attending anything like mass production of warships. The design must be fixed within the next month and by the end of the year [1942] no further changes can be accepted, though the ships will not commence to come out for many months and the last may not come out for more than two years from now.¹⁰

The first ship, Loch Fada, was laid down on 8 June 1943 and completed on 29 March 1944.

The design received Board Approval on 7 May 1943, despite efforts by the First Lord, Alexander, who wanted to replace the twin Squid by another 4in gun – in an A/S ship! – Goodall was not amused.¹¹ Table 7.3 shows the general particulars as designed. Initial armament was a single 4-inch, a four barrel pom-pom, two twin and two single Oerlikons; radar 277 (271 in a few early ships), asdic 144/147. Later increased with more 20mm and 40mm.

Loch Killisport entering harbour with her flags signalling her pendant number.

Displacement (tons): 1,435, 2,260 deep

Dimensions (feet): 307¼ x 38½ x 8¾, 13¼ deep

Shp and speed (kts): 5,500 = 20 (turbine 6,500 = 20.5)

Fuel (tons), endurance (miles) @ (kts): 730, 7,000 @ 15

Complement: 114

It was thought that at least 200 new escorts would be needed and first thoughts were for 120–45 of the twin-screw design (Lochs) and 70–80 single-screw ships (Castles). DNC’s production division under Hannaford then carried out a very careful analysis of the availability of slips of the lengths required. Matching slips and building times for individual yards, they came up with a definitive programme for 133 Lochs and 69 Castles. In December 1942, orders were placed for 226 sets of machinery. In total, 1,150 sets were completed during the war, many of the later ones being diverted to transport ferries (LST Mk 3).

Builder	Slips	Ships
Hall Russell	2	10
Robb	4	16
Inglis	1	2
Blyth	4	18
Smith’s Dock	6	37
C Hill	2	8

In building this new fleet, it was intended to use structural engineering works, mainly bridge builders, to the maximum extent. They were to prefabricate hull sections of up to two and a half tons in weight, limited by shipyard cranes, which would be taken by rail to shipyards to be built into the ship, which imposed dimensional limits on the sections (29ft x 8ft 6in x 8ft 6in). The transverse frames were arranged with riveted boundary angles joined to flanged plate inner sections. The boundary angles were interrupted at plate edges and longitudinals. Longitudinals were flat plates welded direct to the shell. Beams and girders were all simple sections, angles, flats or flanged plates.

The sections used welding or riveting, according to the experience of the contractor. Riveting was usually employed for the assembly of sections in the shipyard. Careful inspection was necessary to ensure that sections made in different places actually fitted to make a fair ship. This worked well, ‘parts for prefabricated frigates coming together beautifully’.¹²

Loch Insh. In the author’s view the Lochs were one of the all-time great designs.

Loch More in 1951, just after the author left her.

It is interesting that contemporary documents refer to prefabrication as ‘on American methods’. Few of these firms had the capability for much plate-bending, lacking bending slabs or facilities for hot smithing, particularly double curvature, so much attention was paid to simplifying the shape. The hollow forward waterlines of the Rivers were made straight and the sheer of the deckline was formed by three straight lines. The stern lines were curved in one direction only. Model tests at Haslar showed that these changes in shape made little difference to the power required. The above-water form was built with flat panels to a considerable extent.

The structural design was led by J L Adam of the British Corporation Classification Society (now merged with Lloyd’s Register). He used longitudinal frames with deep transverse frames about five feet apart. Adam said in 1947¹³ that the design was based on R Baker’s design for the much smaller minesweeper Seagull, the RN’s first all-welded ship.¹⁴ It was estimated that the cost was about 50 per cent greater than that of traditional structure, though there was a considerable reduction in building time. It was a very clever design for the day.

The first ship, Loch Fada, was built by John Brown, with the involvement of a special drawing office set up in Glasgow under the British Corporation and directed by the Warship Production Superintendent (Scotland). They converted the traditional shipbuilders’ drawings into a style that could be understood by structural engineers, quadrupling their number in the process. About half the fifty draughtsmen came from structural-engineering firms. The lines were laid off by the Henry Robb shipyard, who prepared many moulds (patterns). As table 7.5 shows, a very considerable number of firms were involved in this structural work. The table omits about fifty other headings and their subcontractors. It was found that very light items such as partition bulkheads were too easily distorted in transit and were better built by the shipyard. There were some 100 subcontractors and about 1,360 units per ship.

Photograph by the author of Loch Craggie from Loch More.

Item	No of firms involved
Keel	8
Shell longitudinals	1
Web frames	2
Plating	5
Stringer and sheer	4
Lower deck	6
Upper deck	7

Much of the equipment was ordered through a central office under ADNC(P) C J W Hopkins, RCNC. They supplied 500 tons of steel each week, complete offices for W/T, radar and asdic, pumps, rudders, watertight doors, valves, side scuttles, anchor and cable gear and other items too numerous to list here. The number of different types of valve was kept to a minimum and 2,000 were ordered each month.¹⁵ Pipe services, cable trays and ventilation trunking were designed and made by specialist firms. The schedule was geared to completing twelve ships per month. Careful scheduling was needed to ensure that the right bit arrived at the right time and place.

There were thirteen shipyards involved in building the Lochs, of which five acted as specialist outfitters. Six shipyards supplied complete bridges and superstructures. Shipyards of this era were very conservative in their practices and there were delays until opposition to prefabrication was overcome: ‘W Ayre doesn’t want bridge builders on the job – I got depressed, it looks as though the builders’ steady opposition to prefabrication is gaining ground.’ (4.1.43)¹⁶

It was recognised from the start that these ships were far more complicated than earlier ships and that outfitting would be a problem: ‘Engines and boilers now the limiting factor in frigates assuming we get labour for fitting out.’ (20.3.43) In particular, it was estimated that 400 additional electrical fitters would be needed. None of these men were found and, indeed, more were lost to the army: ‘It is sticking out a mile that presently we shall have the ships ready to be fitted out but insufficient labour.’ (20.4.43) Two specialised outfit centres were set up, one at Dalmuir – ‘A bleak spot’ (6.4.43) – for Clyde-built ships and the other at Hendon Dock for the north-east coast. Machinery installation was by NE Marine Engineering and George Clark at Hendon Dock and by John Brown at Dalmuir. At first all went well: ‘Hendon Dock – work is going well. We must use this place more, it is well laid out and Hunter is a go-getter.’ (27.1.44)

By 1945, these centres were failing to keep up with the delivery of hulls, mainly because of shortages of skilled labour, and the docks were full of incomplete hulls awaiting fitting out. By this date, the battle was almost over and delay in completion was not serious.

Altogether, twenty-eight ships were completed as Lochs, nineteen became the Bay class AA frigates, two were coastal forces depot ships, two became despatch vessels (C-in-Cs’ yachts), four were converted to survey ships and fifty-four were cancelled. A few cost breakdowns are available and, although there was considerable scatter, the following is reasonably typical: hull £90,000; machinery £35,000; profit £7,000; other £3,000.

On a personal note, I first became acquainted with the capability of the Lochs whilst serving in the (unmodernised) submarine Tabard in a big exercise with British and US ships off Gibraltar. We had no difficulty in evading any escort group until we met the Lochs, who had us very quickly. Soon after, I transferred to Loch More and learned how easy they found it to dispose of a World War II submarine. They could have coped with a Type XXI U-boat, albeit with difficulty, or a Soviet Whisky class. After an earlier spell living aft in Chivalrous, covered access between the fore and after parts of the ship was welcome and I swore that any ships for which I was responsible would be so designed. In my opinion the Lochs are one of the all-time great designs.

Knaresborough Castle. Note the Type 271 radar.

Castle Class Corvettes

The origin of the Castles has been mentioned previously, under the Loch class. The Lochs were the desired A/S ship but many smaller builders could not build ships of their length. On the other hand, the Flowers were too short for the sea-keeping needed in the North Atlantic and lacked the interior space for the later weapon and sensor systems – not to mention their crews. Smith’s Dock put up a proposal for a stretched Flower and this was developed by Watson’s Admiralty design team. The length was set at 252 feet overall and they were to have asdic Types 145 and 147B, controlling a single Squid with eighty-one projectiles.

After discussion with prospective builders, it was decided that they would be built by mainly traditional methods but using welding as far as yards’ equipment and experience permitted. Scantlings were based on the Flowers but lightened in some areas in the search for more speed. There was a limited amount of prefabrication of structure, and wireless and radar offices were supplied complete, as were lattice masts. The design was given Board approval in May 1943.

The form was developed by tank tests at the Admiralty Experiment Works (AEW), Haslar, and though the new ships were 420 tons heavier than the Flowers and had the same engine, they were half a knot faster, largely because of the increase of 37 feet in length.

Twenty-nine were completed; all Canadian orders were cancelled in December 1943 but twelve UK-built Castles were transferred to RCN.

Builders	Slips	Ships
J Lewis	3	9
J Crown	2	8
Smith’s Dock		2 (prototypes)
Fleming and Ferguson	4	14
Ferguson	3	10
G Brown	3	10
Inglis	4	14

Armament was a single 4in gun and two twin and two single Oerlikons, with provision for four more Oerlikons on the quarterdeck. Type 277 radar was fitted (early ships had 272), along with HF/DF, asdic Types 145 and 147B, one three-barrelled squid with eighty-one projectiles, two depth charge throwers, one rail and fifteen depth charges.

Displacement (tons): 1,580 deep

Dimensions (feet): 252 x 36½ x 13½ mean

Shp and speed (kts): 16.5

Fuel (tons), endurance (miles) @ (kts): 480, 6,200 @ 15 (clean bottom)

Complement: 99

The class’s principal defect was their speed of sixteen and a half knots, which was barely adequate for Type VII U-boats and would have been quite inadequate for Type XXI. Rayner says that they were underpowered, and excessive windage made them very hard to handle at low speed.¹⁷ They would not keep their bows to the sea but wallowed beam-on. Despite this, they were an effective use of available resources. Four of them served long as weather ships, a tribute to their sea-keeping, though many of the scientists who served in them would disagree. Five became rescue ships (qv).

Carisbrooke Castle.