Hunters and hunted: the sinking of SS Arandora Star within the wider context of the Battle of the Atlantic 1939–1940

Abstract

On 2 July 1940, the ocean liner SS Arandora Star was torpedoed and sunk by German submarine U-47, with the loss of around 805 lives; over half of these were British-Italian civilian internees. This article approaches the event from the arena of Second World War military history, contextualising the sinking within the early Battle of the Atlantic. In so doing, it shifts the customary focus away from government internment policy and discussions of cultural legacy towards examining British and German naval strategies and realities. Tactical and logistical considerations of the conflict are investigated, the explication of which allows more detailed discussion of the sinking controversies and enables delivery of ‘answers’ to the persistent ‘questions’ of why Arandora Star was sailing unescorted and without Red Cross insignia. The broad perspective offered engages with transgression and culpability, and overall the article seeks to advance Arandora Star scholarship with its distinctive maritime focus.

Italian summary

Il 2 luglio 1940, il transatlantico SS Arandora Star fu silurato e affondato dal sommergibile tedesco U-47, causando la perdita di circa 805 vite, di cui più della metà erano internati civili italo-britannici. Questo studio si inserisce nell'ambito della storia militare della Seconda Guerra Mondiale, contestualizzando l'affondamento nelle fasi iniziali della Battaglia dell'Atlantico. Di conseguenza, si allontana dalla convenzionale analisi sulla politica di internamento del governo britannico e dalle discussioni sull'eredità culturale dell'Arandora Star per concentrarsi sulle strategie navali britanniche e tedesche. Attraverso l'esame delle considerazioni tattiche e logistiche del conflitto marittimo, è possibile approfondire alcune controversie relative all'affondamento e rispondere alle domande riguardanti la navigazione senza scorta e senza le insegne delle Croce Rossa dell'Arandora Star. L'approccio adottato affronta questioni di trasgressione e responsabilità, con l'obiettivo di far progredire la ricerca sull'Arandora Star partendo da considerazioni prettamente ‘navali’.

Introduction

The sinking of SS Arandora Star on 2 July 1940 was the worst civilian loss of life in a single sinking of a British ship, thus far in the Second World War. Of a total of over 1,550 men on board, some 805 were killed, more than half being British-Italian civilian internees. This level of civilian loss was eclipsed only three more times in British ships throughout the rest of the conflict. Despite the absolute human tragedy, her casualties formed part of total losses of approximately 72,200 sailors and civilians killed at sea (White 2008, 2). Arandora Star represented just one of many large British passenger ships lost and one of approximately 3,800 Allied ships sunk in total during the Battle of the Atlantic, her 15,500 tons contributing to 17 million tons lost (Ellis 1993, 268). The struggle in the theatre of the Atlantic between Allied shipping trying to keep supply lines open and German U-boats trying to destroy them was not only the longest campaign, but also ‘the dominating factor all through the War’, ultimately influencing the outcome (Churchill 1948, 529).¹ For the present investigation, it is impossible to ignore the part of the other vessel – Type VII U-boat, U-47 – commanded by Kapitänleutnant Günther Prien. By July 1940, Prien was already a Nazi celebrity and Kriegsmarine poster-boy for his audacious sinking of the Royal Navy battleship, HMS Royal Oak, on 14 October 1939. Arandora Star was one of 30 Allied ships sunk by U-47 between 1939 and Prien's own demise in the north Atlantic in March 1941 (Joshua 2000a).

The most researched aspects of the Arandora Star disaster pertain to government internment policy, its implementation and the treatment of internees (Gillman and Gillman 1980; Stent 1980; Lafitte 1988; Cesarani and Kushner 1993; Carr and Pistol 2023) and the impact and cultural legacy of the tragedy for British-Italians (Colpi 1993, 2020, 2023; Sponza 2000; Chezzi 2014), with post-generation memory work still occurring (Palladino this issue). In contrast to these trends, this article situates the sinking within the military arena, specifically the Battle of the Atlantic, for the first time. In so doing, the aim is to offer new insight into the context, controversies and questions surrounding the event. Controversies around the sinking itself, for example, why casualties were so high, are interrogated using predominantly untapped survivor accounts. Through examination of the strategic and logistical naval circumstances, firstly British and then German, not only is crucial contextualisation gained, but also responses to persistent ‘unanswered questions’ (Capella 2015, 8–14). Why was Arandora Star sailing alone without escort? Why was there no International Red Cross insignia? The broad perspective offered allows for consideration of the tragedy within the rules of maritime warfare set out by Naval Treaty and the Hague and Geneva Conventions and whether any of these were broken, either by Britain in the marine transport of civilian internees or by Germany in attacking and sinking a civilian ship.

Sinking of SS Arandora Star

The 15,500-ton cruise liner Arandora Star was requisitioned by the British government in December 1939.² Immediate assessment by the Admiralty appraised her as too top-heavy and unsuitable for conversion to an ‘armed merchant cruiser’, for defending merchantmen against surface raiders (Figure 1). She was instead outfitted as an experimental, anti-torpedo test ship, the hull equipped with anti-torpedo nets and paraphernalia. While the torpedo tests were considered successful, this equipment was never put into widespread usage. By 1940, it was decided to convert Arandora Star into a troop transport ship. Hickey and Smith describe the conversion:

The lavish furnishings were torn out and tiers of crude metal bunks were fitted in her holds to increase the passenger accommodation from 400 to 1,600. The liner was armed fore and aft with a 4.7-inch cannon and a 12lb anti-aircraft gun. Edward Moulton [the Captain] watched as his beautiful cruise liner was transformed before his eyes into a camouflaged grey vessel of the war machine. (1989, 62)

Still captained and crewed by civilian officers and ratings, Arandora Star was painted grey for camouflage at sea, the same as any other ship of war, and fitted with defensive armaments. Pressed into service as a troop transport, under Admiralty command, she now flew a Blue Ensign defaced with a gold Admiralty anchor (ADM 199/2133), rather than the Red Ensign of the Merchant Navy.

Figure 1. The Blue Star Line ocean liner SS Arandora Star © IWM Q 71901

In the early morning of 2 July 1940, Arandora Star was steaming independently at a top speed on 15-knots (17 miles/28km per hour) in a zigzagging course, approximately 125 miles north-west of Ireland. Both lone ships and convoys would sail in zigzags to confuse enemy U-boats as to their actual course and hinder them aiming torpedoes. The Germans recognised zigzagging as typical Allied shipping behaviour. She was en-route from Liverpool for St John's, Newfoundland with a consignment of ‘enemy alien’ civilian internees and prisoners-of-war. A precise total for the number of men aboard remains unestablished, published sources giving different figures: officers and crew between 174 and 182; German and Austrian POWs and internees, 475–479; British-Italian internees 712–734, (revised to 707, Pacitti this issue); and British soldiers between 200 and 257. An early, but still cited, total of 1,673 (Dorling 1973), is discounted by several authors who give the total aboard as between 1,564 and 1,615 (Gillman and Gillman 1980; Cesarani and Kushner 1993; Pistol 2017; Pacitti this issue).³ Being July and at a northerly latitude, the sun had risen before 05.00 and according to one survivor, the sea was ‘calm and still, glass-like, like a mirror’ (Duxberry 1990). Just before 07.00, Arandora Star was struck amidships on her starboard side by a single torpedo, fired by U-47. No warning of attack had been given. The explosion tore through the hull, rapidly flooding the boiler room and lower decks. The precision of the torpedo hit and the speed of flooding had a devastating effect on the ship's stability, ensuring a sinking amongst the fastest for large vessels during the Battle of the Atlantic. Within 20 to 30 minutes, Arandora Star had capsized and sunk, sliding vertically into the ocean, stern first.

The rapid sinking is confirmed by reports and statements from surviving ships’ officers collected by Admiralty representatives a couple of days after the sinking (ADM 199/2133; FO 916/2581), reports by surviving military personnel (HO 213/1722; WO 361/4), and by internee survivor testimony, published variously. Close interrogation of these sources allows greater light than previously to be shone on the controversies around the sinking.⁴

According to Arandora Star survivor ship's chief officer F.B. Brown, the torpedo impact was ‘just abaft of the funnel on the starboard side, in the after-end of the engine room’, one of the largest open areas below deck, and making the critical difference. Brown reported that the doors of the engine-room had been blown off, and that within two to three minutes the water inside was ‘practically level with the water outside’ (ADM 199/2133). With 1,550–1,600 men aboard, and the difficulty this posed for the crew moving around the ship, it is perhaps unlikely that the best practice of keeping all the hatches closed-down to prevent flooding had been followed. Moreover, military escort survivor Captain F.J. Robertson stated that ‘internees had discovered unguarded hatchways, companion ladders etc., and were roaming about visiting other decks’ (HO 213/1722).

Two of the main controversies surrounding the sinking have been that more men might have been saved had emergency procedures, and both provision and handling of safety equipment, been better. That no emergency drill or training for the crew, military escort, or passenger-prisoners had been conducted has been cited as a reason for the high death toll (Gillman and Gillman 1980, 200). Embarkation of internees occurred from around 14.45 on 30 June, with departure from Liverpool at 04.00 on 1 July (ADM 199/2133; HO 213/1722), giving 27 hours sailing time available for emergency drills before the torpedo struck, but only around 14–15 hours of daylight. In his report Captain Robertson stated that at 11.00 on 1 July a ‘conference’ of officers took place, during which holding a ‘Boat Drill’ later that day, at 16.00, was discussed. However, the proposed drill had to be postponed due to the upheaval in re-quartering ‘200 dangerous Germans’ (the POWs who were merchant seamen) that afternoon from the ballroom to cabins below deck, well away from the ship's bridge. In any event, Robertson further reveals

… owing to the insufficient number of [life]boats it was considered advisable not to include the Internees in the drill so as not to create an uneasy feeling, and as this entailed leaving the guards on duty, the Drill was to be confined to ship's crew and some NCOs of the escort. (HO 213/1722)

Thus, had time and circumstances permitted, any drill would not have included the prisoners, and even basic instruction on how to use the cork life-jackets was lacking as recounted by survivor Vittorio Tolaini:

No boat drill, nor any other life-saving information had been given. Life belts were not distributed, but left here and there on the deck. No instructions on how best to abandon ship were issued and this lack of information caused many casualties. A great number of deaths were caused by people sliding down ropes with their bare hands thus causing terrific burns and, letting go of the ropes, plunging into the sea and to certain death. Another cause of death was the old-fashioned lifebelt, which if not properly secured resulted in breaking one's neck. … Nobody knew what to do excepting the German seamen who no doubt had been trained for such emergencies. (Tolaini 1982)

Reference to the capabilities of the German merchant seamen in the evacuation is also implied by survivor Corporal Ivor Duxberry, who said: ‘There were some very helpful Germans’ (Duxberry 1990). Such statements acknowledging the contribution of the dangerous-turned-helpful-German merchant seamen allude to the logistical breach they stepped into. As Capella notes, the guards aboard were military rather than naval (2015, 13), and similar to the majority of their prisoners had little or no experience at sea, and likewise, in the absence of any drill training were unprepared for the emergency. Captain Robertson highlights the inadequacy of these men with ‘general ignorance of ships and their ways’, recommending it ‘advisable that escorts proceeding overseas might be formed or contain a proportion of RNVR [Royal Navy Volunteer Reserve], or men who are fully at home on the sea’, adding that the guards were ‘of course useless with the [life]boats’. In addition to incompetence at sea, the military escort had been drawn from multifarious regiments of the army,⁵ presumably with no history of acting together as a unit or team. Furthermore, regarding the crew, Robertson stated ‘Eighty per cent … had been signed on the morning we sailed’, implying they were not yet fully familiar with the ship (HO 213/1722). Thus, the highly heterogeneous military guard was serving alongside a potentially less than fully competent crew. Finally, it can also be noted that the Merchant Navy generally was considered poorly trained. According to Lloyd's:

A cursory merchant navy approach to survival training in the first part of the Second World War, contributed to heavy loss of life in the Battle of the Atlantic, with seafarers dying for lack of skills to escape from sinking ships’. (Lloyd's Register Foundation)

Two surviving army officers corroborate a clear-cut contributory mistake. Captain C.M.C. Lee wrote in his statement that ‘many of the rafts served no useful purpose as they were thrown overboard in the early moments whilst the ship still had considerable way and they consequently drifted behind’ (WO 361/4). Captain Robertson concurred:

I put on a coat and lifebelt and went on deck, where members of the crew and escort were throwing rafts overboard. This proved to be an error as the ship had a considerable way on her and when she eventually stopped and sank the rafts were a considerable distance away. (HO 213/1722)

Thus, while the rapid sinking of Arandora Star undoubtedly caused major issues for the ship's evacuation, the lack of emergency training and the inefficiency and ineptitude of both army escort and crew contributed substantially to the heavy losses.

Alongside the shortcomings of the military escort and crew, the other main concern influencing the high death toll was the extent of lifeboat and raft provision. Whilst the official position, as stated by the shipping minister, was that ‘lifeboats and liferafts more than sufficient to accommodate all passengers and crew were provided’ (HC 09 July 1940), as noted above, the number of lifeboats was considered ‘insufficient’. Statements vary on whether there were 12 or 14 originally. That one lifeboat was smashed by the torpedo, one crashed and was lost during lowering, and men from ten were successfully rescued (DeWolf 1960; HO 213/1722), giving a lifeboat capacity for 700–800 persons from over 1,550–1,600 is, however, generally agreed upon. It is likely the ten lifeboats were full to capacity and beyond, although again there is disparity in the statements. Chief Officer Brown specified the lifeboat capacity was 84 persons (ADM 199/2133), yet Captain Robertson said of his lifeboat ‘We numbered 58 persons and there was not an inch of space left’ (HO 213/1722). Captain Lee, on the other hand, describes going back to the wreckage to pick up men until ‘in danger of sinking’, with 112 men in his boat (WO 361/4). On the question of the liferafts, the ‘compiled account’ from several officers estimated that about half of the 90 available had been thrown overboard ‘at once’, when the ship was still moving (FO 916/2581). Chief Officer Brown opined that these made no contribution to saving lives because ‘the aliens refused to use them’ (ADM 199/2133); the joint officers’ report similarly stating ‘none of them [the aliens] would go over the side to the rafts’ (FO 916/2581). The implication that the prisoners would not jump into the sea and attempt to swim to the rafts to try and save themselves appears to echo the official line of there being sufficient provision.

Other rafts were possibly inaccessible due to barbed wire, although its quantity and placement, used as much to separate the German and Italian prisoners, as for security from escape, has never been conclusively established and accounts differ (Gillman and Gillman 1980, 199–200). During the voyage of the first internee transport, the Duchess of York, one internee was shot dead by a British guard and two others were wounded, ensuring increased security on subsequent transports (Stent 1980, 100). Chief Officer Brown reported ‘we had protested before we left Liverpool [about] the barbed wire entanglements’, saying these greatly ‘impeded’ the crew (ADM 199/2133). While in her capacity as a troop transport, Arandora Star could not be considered overloaded during her final voyage (Dorling 1973, 40–42), in the eventual circumstances that unfolded requiring fast and decisive action, clearly any amounts of barbed wire could have been a significant hindrance.

Survival rates amongst the different groups aboard differed greatly, with the German merchant seamen the highest, around 80 per cent,⁶ and the Italians the lowest, at 37 per cent. Amongst the prisoners, it was no coincidence that many of the survivors were fit, young men, who were sleeping out on deck or in the public rooms when the ship was torpedoed. Those quartered in the lowest decks, which included a large proportion of the Italians (Gillman and Gillman 1980, 198), were disadvantaged in escape upwards due to the blackout caused by the torpedo explosion. Those who managed to reach open deck found the lifeboats either already launched or full. Sources describe Italians on deck who ‘still’ refused to jump into the sea even when it was clear the ship was about to sink (FO 916/2581). Gino Guarnieri, a 27-year-old waiter from Soho, scrambled into a lifeboat before it was lowered, and witnessed the disaster unfolding.

There was about 20 of us in the boat. One of the officers, he had a hatchet in his hand. He cut the rope and he said ‘Now start rowing. Every one of you, start rowing’. So we said ‘Look, there's still a lot of people on the ship’. He said ‘I'm in charge here, whatever I say goes. If anybody doesn't obey me, I'll chop their head off. When I say row, I mean row!’ So we rowed about 200 yards and then he said ‘Now we stop, and I'll tell you why we had to row away. If we had stopped there everybody would have gone down, because the ship is going to sink any minute now. We're going to pick up survivors after. I had to think of the boat, because if there isn't any boat, nobody will survive’. (Guarnieri 1990)

Accounts of the short time between the torpedo hit and the ship sinking reveal chaotic disorder, although uncontained panic was largely discounted (ADM 199/2133; FO 916/2581). Captain Robertson perhaps best sums up the chaos of those 20–30 minutes:

It is difficult for anyone not an eyewitness, to realise the chaos that reigned on this disastrous mismanaged trip or the exasperation of Officers, men and Internees, and the needless suffering inflicted by the complete absence of any forethought or organisation. (HO 213/1722)

As the only surviving officer not previously asked to report, in his statement to the War Office in March 1941, Lieutenant J.F. Constable stated that the fast sinking, which he puts at 20 minutes, the fact that the crew had not been ‘practised in Emergency Stations’ and that ‘Military Personnel and Internees had not been instructed in the use of life jackets or told of Emergency Stations’ led to the ‘indescribable chaos and heavy casualties resulting’. Clearly aware of the controversy surrounding the sinking, Constable mentions the Parliamentary Inquiry (Snell 1940), and concludes:

… at this juncture … a full statement of the FACTS might be misunderstood; therefore, as a good soldier, having kept it ‘under my hat’ for nine months, I can, if necessary, continue to do so, giving you only a ‘useful’ but innocuous story. (WO 361/4)

Neither the ‘facts’ nor even a ‘useful story’ were subsequently sought and no further correspondence between the Lieutenant Constable and the War Office was found.

Admiral Sir Dudley Pound, naval engagements, and convoys

One of the greatest controversies within the British-Italian community around the sinking of the Arandora Star is that she sailed alone, and unescorted. This clearly left the ship vulnerable to enemy attack. As Phillips O'Brien attests ‘… World War II would demonstrate, any unescorted merchant ship sailing in a submarine-patrolled area was in danger, even if there were only a handful of submarines on operational duty’ (2015, 109). Speculation around this issue led to accusations from victims’ families that conceivably Arandora Star had, at best carelessly, and at worst deliberately, been left to her fate (Capella 2015). To address this question, the state of the British fleets, both merchant and naval, at the beginning of the war and the wider, strategic context in which the First Sea Lord, Admiral Sir Dudley Pound, Commander of the Fleet of the Royal Navy made his decisions, are of relevance (Figure 2).

Figure 2. Admiral of the Fleet Sir Dudley Pound, who acted as chairman of the British Chiefs of Staff Committee 1939–1942; © IWM A 20791

Alongside his contemporaries, Pound was influenced by experience of the First World War. Lessons had been bitterly learned by the Royal Navy and consequently few argued against the convoy system as the best strategy to protect merchant and passenger ships from U-boat attack. Convoying acted both to protect and attract U-boats, giving the escorts a chance of counter-attack (Keegan 2011, 85). Pound knew that the next world war would be a ‘total war’ and understood that convoy escort would be one of the Navy's primary activities in this ‘war of imports’. Yet in 1939, Britain was woefully underprepared to engage in total war at sea. Plans prepared by the Admiralty ‘to institute convoy should war break out with Germany revealed that an insufficient number of surface ships for escort duty were available’ (Schofield 1981, 119).

At the start of the war, the British merchant fleet was by far the largest in the world, with a gross register tonnage (GRT) of 17.9 million spread across the empire, out of a global total of 68.6 million. Although shipbuilding in Britain was also strong, with over 40 per cent of the world's merchant ship tonnage still built in British yards in the 1920s and 1930s (Mawdsley 2020, 85–86), the construction of new escort vessels and the training of crews would take time. As Mawdsley contends, the measures taken from early 1939 to strengthen the Navy's escort force, such as ordering the ‘Flower’ and ‘Hunt’ classes, or reviving the ‘Corvette’ classification of small ocean-going warships, were ‘belated steps’ (2020, 94). Restricted in building new ships by the London Naval Agreement of 1930, the Royal Navy fleet was outdated, slow, and short of modern capital ships in comparison to a newer, more advanced and faster, albeit significantly smaller, German fleet, which had been built almost ‘from scratch in the 1930s’ (Levy 1999, 95–96). Overall, British naval resources were severely stretched due to the scale of commitments in the Mediterranean, the Far East and the trans-Atlantic trade routes.

By 10 June 1940, the day Italy declared war on the Allies, Norway and Denmark had fallen in April and the French government had abandoned Paris in the face of the lightning German invasion in May. Against this rapid turn of events and increasing threat of invasion the War Cabinet secretly finalised the policy of deportation overseas for the most ‘dangerous’ internees (Lafitte 1988, 123–125). To defend the coastal waters the Admiralty stripped the Home Fleet and the Atlantic convoys of destroyers and corvettes to maintain the strongest possible anti-invasion patrols along the south coast of England. Meanwhile, upon Italy's entry to the war, the Regia Marina, by far the most capable Italian arm-of-service, on paper, immediately went to war with the Royal Navy in the Mediterranean. Mussolini thought that by defeating the Navy he would restore ‘mare nostrum’, the mortar binding his dream of a neo-Roman empire (Abdulafia 2011, 601–612). By 2 July 1940, Admiral Pound had deployed units to the Mediterranean to engage the Italians and also prevent defeated French ships falling into Axis hands, unavoidably taking further badly needed ships, aircraft and resources away from home waters and the battle in the Atlantic (Mawdsley 2020, 111–115).

Clearly, the fundamental issue facing Britain during the early Battle of the Atlantic was the shortage of escort vessels to protect merchant shipping convoys, the strategic sinew in the total war, upon which Britain depended for all her necessary imports of food and raw materials. By March 1940, the Navy only had 135 vessels capable of convoy protection in home waters, compared to 449 in 1917 when Germany started unrestricted submarine warfare (O'Brien 2015, 126). Ships in British coastal waters were given priority, but for deep-sea ships there were only sufficient escorts to convoy them as far as the Western Approaches, 90 miles west of Ireland, in and out (Edwards 1996, 19). Convoys were only as safe or as fast as the slowest vessel. Fast convoys, of 8–13 knots (15 miles/24km per hour) were used mainly to transport troops while slow convoys, 7 knots or slower (8 miles/13km per hour), were for cargo. Up until May 1940, Phillips O'Brien calculates ‘a total of 503 British, Allied or neutral merchant vessels were lost as a result of enemy action in all waters of the globe’, with only 36 of these sunk while in convoy (2015, 127). From May until October 1940, U-boats enjoyed their first ‘Happy Time’ when the overall rate of unescorted merchant sinkings increased, aided by the capture of the Norwegian and French coastlines allowing U-boats to enter and exit homeports more freely.

Crucially, for the Arandora Star, there were simply not enough escort ships available in the summer of 1940. Vessels faster than 13 knots, such as the 15-knot Arandora Star, were normally destined to sail without convoy, and unescorted. Arandora Star was considered too fast and too big to sail as part of merchant convoys, her greatest defence being ability to outrun U-boats. Although Type VII U-boats, such as U-47, had a top surface speed of 17 knots, this could not be sustained for long (Keegan 2011, 91). Nevertheless, as historian Vice Admiral Brian Schofield pointed out, ‘the loss rate of independently sailed ships with the exception of a very few capable of speed in excess of 20 knots was consistently much higher than for those sailing in convoy’ (1981, 121).

In her short career between late 1939 and 1940 as a merchant vessel in wartime service, Arandora Star sailed both in fast convoys transporting troops and goods, and independently carrying both troops and civilians. Before conversion to a troopship, in September 1939 she sailed in Convoy HXF 1, the first of 17 fast trans-Atlantic convoys, as one of eight large and similarly fast ocean liners returning from Halifax to Liverpool (ADM 199/2184/1). Sailing at 15 knots, the convoy was also escorted by Canadian destroyers on the way out and by two heavy cruisers near home waters. Between December 1939 and May 1940, Arandora Star made a number of short independent voyages fetching and carrying between ports around the British Isles, but in early June 1940, she sailed again in convoy to help evacuate 10,000 Allied troops from Norway (Gillman and Gillman 1980, 188). Later in June, amongst other vessels sent to rescue Allied troops and British refugees stranded in south-west France, Arandora Star undertook several solo missions – to Saint-Nazaire, Quiberon and St Jean de Luz in the Bay of Biscay. On 24 June, at Bayonne, she embarked 3,000 Polish Army and Air Force personnel. Sailing for Plymouth alongside MV Ettrick, one of the other troop transports soon to be deployed for internee deportation, the two vessels were escorted by the destroyer HMS Harvester, arriving 26 June (Ettrick 1978). On 27 June, Arandora Star returned to Liverpool (Gillman and Gillman 1980, 189).

Due to the fast-paced exigencies of war and relentless tonnage losses from U-boat attack, logistical issues concerning availability, speeds and convoying were under constant review between the Admiralty and Ministry of Shipping.⁷ The Arandora Star was certainly not the only large, civilian-carrying, liner attacked. For example, on 17 September 1940, after departure of the convoy escort, a similar tragedy occurred when the 11,100-ton, 15-knot liner, SS City of Benares was torpedoed by U-48. Of the nearly 500 aboard, 90 were British child evacuees being transported to safety in Canada. Although the liner stayed afloat, 260 people were killed, including 77 children. Two of these children were re-assigned survivors from SS Volendam. Torpedoed on 30 August 1940, carrying 320 child evacuees, Volendam had remained afloat with all but one passenger rescued (Heiligman 2019). War cabinet meetings in June 1940 disclose that discussion of ship assignment for evacuation of children from Britain to safety overseas was interwoven with that of internee deportation (CAB 65/7/69; CAB 65/7/70). Fear of ‘enemies within’ when invasion threatened put the internee transports ahead of the children's evacuation. The comparable transport of these two different groups to the Dominions indicates not only the normality of such government policies but also the belief that deported internees were being subjected to no more danger at sea than other segments of the population, including children.

In summary, Arandora Star was sailing alone and unescorted for the prosaic reasons that she was too fast to be effectively used in merchant convoys, was considered fast enough to outrun U-boats, and the fact that there was a severe shortage of escort vessels in summer 1940. As one of five ships deployed to deport POWs and enemy alien civilian internees, Arandora Star's sinking did, however, directly impact policy on subsequent transports. Both MV Ettrick and HMT Dunera, which departed on 3 and 10 July respectively, were partly escorted by destroyers in the dangerous home waters. The very fast, 20-knot Ettrick reached Canada unscathed, but the slower Dunera, destined for Australia with a speed similar to Arandora Star, was attacked and narrowly missed by two U-boat torpedoes in the Irish Sea (Pearl 1983, 24–26). Both the 17-knot SS Duchess of York and the 18-knot MS Sobieski, the first and last internee transports, safely reached their Canadian destinations. Clearly, at 15-knots, the ‘ill-fated Arandora Star’ was not fast enough after all.

Admirals Erich Raeder and Karl Dönitz, and Captain Günther Prien

The immediate cause of the Arandora Star disaster was that she was targeted and torpedoed by enemy submarine, U-47, commanded by Günther Prien. His job, as demanded by his superior officers in the Kreigsmarine, was to sink the greatest tonnage of Allied shipping possible on each patrol. A fervent Nazi (Prien 1954, 196), in June 1940 Prien was aged 32 and already a veteran of five U-boat missions. His daring and skilful sinking of HMS Royal Oak earned him the Knight's Cross – the first U-boat commander to receive the highest order of the Iron Cross – and the name ‘Der Stier von Scapa Flow’ (‘the bull of Scapa Flow’) (Figure 3).

Figure 3. Kapitänleutnant Günther Prien being greeted by Vice Admiral Dönitz after sinking HMS Royal Oak on 14 October 1939 at her moorings in Scapa Flow; © IWM HU 2226

Prien's superiors, Grand Admiral Erich Raeder, Head of Naval Command, and Admiral Karl Dönitz, commander of the U-boat arm, had come of age as senior officers during the ‘unrestricted’ submarine warfare of the First World War. Despite Hitler's wish for a short, quick war, Raeder understood this would be a total war. Although Hitler aspired to a grand surface fleet to rival the Royal Navy, Dönitz thought resources should be focussed on submarines to destroy Britain's merchant fleet (Dönitz 1958, 116). In 1940, Hitler was fearful of provoking the United States into joining the war as the sinking of RMS Lusitania in 1915, followed by unrestricted submarine warfare, had done in 1917. In 1936, he signed the ‘Submarine Protocol’, as part of the 1930 London Naval Treaty, applying, albeit impractically, the same rules to submarines as to surface vessels. The Treaty bound both parties to sink only warships, which could be sunk without warning. Passenger ships could not be sunk. Merchant ships could be sunk ‘only in accordance with international law – that is, with warning if … sailing unescorted and without warning only if escorted or protected in convoy’ (Keegan 2011, 86). Submerged submarines could not attack merchantmen with torpedoes without warning (Carroll 2009, 408). These ‘cruiser rules’ or ‘prize laws’ requiring U-boats to surface, give enemy merchant ships warning to permit their crews to evacuate and, unfeasibly for submarines, be put in a place of safety, which did not include lifeboats, before sinking, did not, however, prevail for long. Ambiguity surrounded ‘definite instructions’ as submarine commanders tried to obtain ‘clarification of orders and, at the same time, a lifting of the seemingly impossible limitations’ (Mallmann Showell 1989, 11). Armed merchantmen were considered de facto naval auxiliaries, deemed outside the prize laws, and as such could be sunk without warning. Whilst prize laws were German official policy, the Navy's U-boat arm had never fully supported them and, ‘step by step’ found them increasingly impossible to observe ‘in response to breaches … by the enemy’ (Levie 1998, 317). In this war of attrition, Dönitz considered every enemy ship a legitimate target (Dönitz 1947).

Demonstrative of this attitude from high command, on the first day of war, 3 September 1939, the SS Athenia, a 13,500-ton, 15-knot unarmed passenger liner, was torpedoed and sunk without warning by U-30, with no attempt to assist survivors, effectively setting the precedent for unrestricted submarine warfare (Carroll 2009, 410). Leutnant Fritz-Julius Lemp ignored the prize laws, and specific orders from Hitler, and doctored his logbook to cover this. Dönitz accepted Lemp's explanation that it was a case of mistaken identity – a passenger ship taken for an armed merchant cruiser – and Lemp not only escaped serious punishment but was awarded the Iron Cross (2^nd Class) and promoted to Kapitänleutnant (Mawdsley 2020, 3–4).

Yet, in the early stages of the war, U-boat captains tried to keep within international law, negotiating a fine line between ‘enemy’ and ‘fellow seamen in distress’, often assisting crews after sinking an enemy vessel (Bennett 2011, 19–20). Due in part to this practice, and also the unsustainable loss of nine U-boats by December 1939 (Edwards 1996, 19), Dönitz insisted on ruthlessness and issued his Standing Order 154:

Rescue no one and take no one with you. Have no care for the ship's boats. Weather conditions and the proximity of land are of no account. Care only for your own boat and strive to achieve the next success [enemy sinking] as soon as possible. We must be hard in this war. The enemy started the war in order to destroy us, therefore nothing else matters. (Keegan 2011, 88)

The blank cheque for unrestricted submarine warfare against Allied shipping had effectively been signed by German Navy high command and by August 1940 total abandonment of international law reigned on the seas (Mallmann Showell 1989, 11).⁸

During September 1939, Günther Prien adhered to orders, keeping attacks by U-47 within the Submarine Protocol. He gave warnings and assisted lifeboats in his sinkings of unescorted vessels SS Bosnia and SS Rio Carlo (Bennett 2011, 19). But for ‘ace’ U-boat captains like Prien, GRT of enemy shipping sunk measured success, with tonnage scores contested between rivals, and his sixth patrol proved a ‘Happy Time’.⁹ He had already sunk seven merchant ships, beginning with Balmoralwood of Convoy HX47 on 14 June 1940 and including several oil tankers (Joshua 2000b), when, with his last torpedo, he sank Arandora Star on 2 July.

In his captain's log for 2 July, position 56° 20N 10° 40W, at 06.29 hours, Prien records sighting a passenger steamer – ‘passagierdampfer’ – moving at fairly high speed (Prien 1940).¹⁰ He dives and sounds the alarm. Drawing closer and watching his ‘opponent’ for 25 minutes, at 06.54, he estimates her speed as 15 knots, identifying her as hostile by the zigzagging course. In favourable conditions with ‘gute sicht’, the unescorted ‘15,000 ton’ ship, estimated almost exactly by Prien, presents an easy target. At close-range, of 2500m,¹¹ he gives no warning before firing at 06.58.28. After confirming a hit on the forward funnel 97 seconds later, and observing a water column at the rear funnel, at 07.15, Prien logs that the ship lies stopped and listing, with lifeboats partly launched, partly still being lowered. Still submerged, at 07.40, he observes no sign of the ship through his periscope. He surfaces at 07.48, and still unable to see anything, he continues his ‘retreat’ and assumes the ship sunk although, now 5.2 nautical miles away, he is not entirely sure. He completes his log, reporting that his ‘opponent’ was armed with a ‘clearly identifiable’ larger calibre cannon at the stern and another one apparently at the bow. Thus the ship's size, zigzagging course, and her weapons furnished sufficient information for attack. It is possible he understood the steamer to be a troop transport, eligible for attack without warning and within orders, even international law (Levie 1998, 316). However, Prien made no attempt to offer aid, violating the rules; his confirmation of lifeboats in the water, with perhaps an assumption of some survivors, seemingly adequate for his superiors. Not until 3 July, by radio, did Prien learn the ship's name. It is unknown when he discovered the majority of passengers were German and Italian civilians, a detail concealed from the German public (Joshua 2000c).

The Red Cross and the Hague Convention 1907

An issue long voiced within the British-Italian community is that the Arandora Star should have been marked with Red Cross insignia to indicate she was carrying civilians (Capella 2015, 1, 13–14). This concern is, however, misguided and erroneous. Whilst it is correct to reason that in wartime vessels marked as ‘Hospital Ships’ cannot be attacked or captured, Articles 1–3 of the 1907 Hague Convention make clear the definition of Hospital Ships as those

… constructed or assigned by States specially and solely with a view to assisting the wounded, sick and shipwrecked, the names of which have been communicated to the belligerent Powers at the commencement or during the course of hostilities, and in any case before they are employed … (International Committee of the Red Cross 1907; Schindler and Toman 2004, 399)

Article 5 of the Convention goes on to state how Hospital Ships are to be recognised. In the case of private ships, officially commissioned for this specific purpose, they are ‘distinguished by being painted white outside with a horizontal band of red about a metre and a half in breadth’.¹² Civilian Hospital Ships

… shall make themselves known by hoisting, with their national flag, the white flag with a red cross provided by the Geneva Convention, and further, if they belong to a neutral State, by flying at the mainmast the national flag of the belligerent under whose control they are placed … (International Committee of the Red Cross 1907; Schindler and Toman 2004, 400)

Thus, it is clear that designated Hospital Ships were solely for the transport and treatment of wounded civilians and could not be used for civilian transport. Furthermore, any use of such marked ships, including false and unauthorised flying of flags, as anything other than Hospital Ships, was forbidden by Article 21 of the Hague Convention (Schindler and Toman 2004, 401–402). In response then, to the oft-posed question, why was the Arandora Star not flying the flag of the Red Cross on 2 July 1940, the answer is straightforward. Having not been designated a Hospital Ship, she was not entitled to be painted in the white and red livery of a Civilian Hospital Ship. Nor was she entitled to fly the flag of the Red Cross. Arandora Star was a military transport ship and as such, ‘flying a blue ensign with the Admiralty anchor’ (ADM 199/2133). When the Foreign Office was challenged by Italy, through the American Embassy in Rome, about not informing the International Red Cross Committee that the Arandora Star was about to sail, the Admiralty countered by saying it could not ‘give the German Government details of the course and/or date of sailing of these British [internee] ships’ (FO 916/2581 568).

While the 1907 Hague Convention is unambiguous as to the designation and status of Hospital Ships, it does not cover the position of wartime civilians aboard liners, troopships or cargo vessels. By 1940, as the only applicable international agreement, the Convention was woefully out of date. It revised and enlarged the Geneva Convention of 1899 concerning adaptation to maritime warfare but was still using precedents of requisitioned merchant ships dating to the Franco-Prussian War of 1870. Written before submarine warfare became a universal feature of naval conflict, the status of civilian ships requisitioned and converted by governments for war, was unclear. The Convention could only speculate on the status of civilian ships and whether they constituted being ‘Privateers’ (Schindler and Toman 2004, 1065–1068).¹³ Furthermore, the details of the 1907 Convention pertaining not only to civilian ships and civilians transported thereon, but more specifically to civilian internees, were far from explicit. Civilian internees were denied prisoner-of-war status and protections even as set out in the provisions of the later 1929 Geneva Convention (Carr and Pistol 2023, 3). Unfortunately, the enemy alien civilian internees aboard Arandora Star were in a convention ‘no-man's land’; effectively treated as prisoners-of-war by the British, and as the company of an enemy ship by the Germans. The status and protections of both civilian crews of ships and civilian internees, as prisoners of war, were only finally confirmed in the Fourth, 1949 Geneva Convention (Levie 1979, 85; Carr and Pistol 2023, 9).

Conclusion

The sinking of SS Arandora Star and the resulting high loss of life of British-Italian civilians has evoked strong emotion and heartache for both Italians living in Britain and for their communities of origin in Italy over the last 84 years (Colpi 2023). Few British-Italians were untouched by the disaster – if not directly affected, many knowing friends or colleagues who lost a family member. The shock and anger caused by the scale of the tragedy is understandable, considering it was a culmination of the indiscriminate and haphazard internment policy (Cesarani and Kushner 1993; Sponza 2000; Pistol 2017). Many Italians, long-settled in Britain and whilst patriotic and still culturally connected to Italy, were loyal to their adopted country and had little love of fascism. Arrested without assessment on 10 June 1940, the attitude and deportation policy of the government against potential ‘fifth columnists’ evoked resentment and bitterness amongst the Arandora Star bereaved. Some descendants of victims, even today, see this as evidence of the cruelty and indifference of the government towards the internees carelessly put in harm's way. Traumatised by the heavy losses, the British-Italian community fell into a long ‘silence’ of some 50 years (Colpi 2020), during which time there was little accessible information due to closed government files. Unexplained and unknown aspects of the tragedy held mythological power and were woven into the narrative, which partly addresses why answers to relatively straightforward questions remained illusory. Apportioning ‘blame’ for the deaths was mainly laid at the government's feet (Capella 2015), and at Mr Churchill's in particular (Chezzi 2014, 386).

Yet, the historical facts and realities of the early Battle of the Atlantic, particularly the lack of naval resources to protect shipping, and a Britain facing invasion in summer 1940, indicate a disaster that was both the result of questionable government internment policy and especially its imperfect implementation, and German unrestricted submarine warfare tactics. As a military transport ship equipped with defensive armaments, Arandora Star was seen as de facto a Royal Navy auxiliary. As such, in German interpretation, she was outside the provisions of the 1936 Submarine Protocol for merchantmen. Yet, for the unescorted ‘passenger steamer’, that Prien identified, he failed to give warning of attack and made no attempt to aid survival, thus ignoring both international law and cruiser rules. Evaluating chance in history, given Arandora Star's evasive zigzagging, Prien was ‘lucky’ to sink her with just one torpedo, his last, which had been subject to adjustment on 1 July (Prien 1940). Alternatively, she was extremely ‘unlucky’ to cross U-47’s path, one of relatively few U-boats patrolling the Atlantic at that time, and to be dealt such a fatal blow, causing her to sink so quickly. Had the British authorities chosen another, arguably safer, means of transport for the Italian and German internees, such as a convoyed passenger liner, there would have been no guarantees of safe passage, as the sad example of SS City of Benares evidences. Had more care been taken in the organisation of the transport, more lives may possibly have been saved. Arandora Star was a fast, relatively new and well-maintained ship that had accomplished several wartime missions. Nevertheless, at 15-knots she was the slowest and therefore most vulnerable of the five transports used to deport enemy aliens. However, in sending these men overseas, Britain broke no international conventions since civilian internees were afforded little protection under the still extant 1907 Hague Convention. Furthermore, contrary to popular belief, international law neither entitled nor permitted Arandora Star to fly the Red Cross, indicating a Hospital Ship, as protection against attack.

It is possible to dismiss the loss of life on 2 July 1940 as just another inevitable and tragic fact of war. However, this diminishes its particular tragedy. As so many Italians were lost, their bodies never recovered, there are no large and well-attended graveyards of those who perished and public memorialisation only began in the 2010s (Colpi 2020). Thus, it is easy to be unaware of the immense impact of wartime losses at sea. The loss of 442 British-Italian civilians and over 350 others, should be a rallying call to politicians and historians to give all those lost at sea in mass sinkings, both in naval action or by accident, the historical dignity and status they deserve. With the notable exception of the sinking of the RMS Titanic in 1912, mass loss of life at sea is in danger of becoming forgotten and unacknowledged. For example, the sinking of the MS Estonia in the Baltic Sea in 1994, with the loss of 852, seems barely remembered in recent European history, and the drowning of some 500 migrants in the Messenia disaster off Greece, in June 2023, is already ‘forgotten’. We owe it to those lost on Arandora Star, and all others lost in maritime disaster, to keep the memory alive.