When the Arcadia Conference convened in Washington on December 23, there was a new face in the senior American group: Admiral Ernest J. King, age sixty-three, former commander of the Atlantic Fleet. In the aftermath of Pearl Harbor, Roosevelt had appointed King (effective December 30) to the newly recreated post of Commander in Chief, U.S. Fleet, a position comparable to that of the Royal Navy’s First Sea Lord. King exercised operational and administrative control of the Atlantic Fleet, commanded by his successor, Royal E. Ingersoll; the Pacific Fleet, commanded by Chester W. Nimitz, replacing Husband E. Kimmel; the Asiatic Fleet, commanded by Thomas C. Hart; and the United States Coast Guard, which President Roosevelt had transferred from the Department of the Treasury to the Navy by executive order on November 1, 1941. For the time being, Admiral Harold R. Stark retained the title of Chief of Naval Operations, but his days in Washington were numbered and his responsibilities were limited to administrative matters and “long-term planning.”
King was a brilliant, salty fighter, so flinty and tough, the story went, that he “shaved with a blow torch.” In his forty-year professional career he had served on surface ships and submarines, had qualified as a naval pilot, and had been a pioneer in naval aviation and carrier operations. No one knew the strengths and weaknesses of the Navy—and its senior officer corps—better. No one was more offensive-minded or more eager to avenge the treachery of Pearl Harbor, which had left the Navy in shock and despair. Although King was approaching retirement age, Secretary of the Navy Frank Knox had recommended him to Roosevelt as the admiral most qualified to shake the Navy out of its post-Pearl Harbor paralysis.
King had only just arrived in Washington. He was still in the process of setting up an office and recruiting a staff. “Nothing was ready,” he said later. “I had to start with nothing.” However, he had participated in the meetings at the Atlantic Conference in Argentia the previous August and thus he knew most of the American and British luminaries. He said little and the British delegates, including First Sea Lord Dudley Pound, gained the impression that King’s mind was focused more on the war in the Pacific than on the war in Europe and the Mediterranean.
This was only partly right. Having just left command of the Atlantic Fleet, King was as acutely aware of naval problems in the Atlantic theater of war as was any British delegate. In particular King, like the British, was deeply concerned about a possible sortie to the Atlantic by the super battleship Tirpitz, to be joined by the battle cruisers Gneisenau and Scharnhorst, the “pocket” battleship Admiral Scheer, and the heavy cruisers Prinz Eugen from Brest and Hipper from the Baltic. An Atlantic raid by those six German big ships would present a daunting challenge to the depleted British Home Fleet and the American Atlantic Fleet and would imperil troop and supply convoys. The danger would be increased if the Vichy French warships in Martinique elected to reinforce the Germans or to sail independently against Allied naval forces or the Panama Canal.
Throughout the Arcadia Conference, a flood of profoundly upsetting bulletins arrived from the Pacific and Far East, announcing one Japanese victory after another. The conferees were thus compelled to spend much time devising emergency measures to help beleaguered Allied forces in that theater. Nonetheless, they hewed to the main purpose, which was to lay out a global strategy for winning the war and to formulate war-production schedules to implement the strategy. All delegates, including most emphatically Admiral King, again affirmed the earlier agreements to crush Germany and Italy first, Japan second, but there was much uncertainty and disagreement over how this was to be done and at what point the Allies were to shift from the defensive to the offensive in the Pacific to minimize the consolidation of Japanese conquests.
Underlying all discussion was a unanimous desire to launch offensive action against Germany and Italy at the earliest possible time. Churchill proposed a plan (Gymnast) for an Allied invasion of French Northwest Africa, designed to trap Rommel and satisfy Soviet demands for a “second front.” Still skeptical of Churchill’s “Mediterranean strategy,” and believing the Allies should strike directly for the German heartland, Army Chief of Staff George C. Marshall counter-proposed an Allied invasion of Occupied France (Sledgehammer) in late summer of 1942. Admiral King enthusiastically approved Marshall’s early invasion plan, in part because he believed it might result in the evacuation of German U-boat bases in France, forcing the Atlantic U-boats back to more vulnerable and inconvenient bases in Norway and Germany.
Not wishing to delay the commitment of American forces against the Germans a day longer than necessary, Roosevelt overrode Marshall and approved Gymnast, but he also authorized a buildup in the British Isles (Bolero) for Sledgehammer, or its larger alternative, Roundup, to take place in 1943. Others proposed that the Allies seize and fortify islands in the Atlantic off Africa (the Azores, the Canaries, Cape Verdes), but when King pointed out bluntly that “[w]e cannot do all these things,” the last proposals were tabled.
In the end, the conferees settled on the following major courses of action
• Support to the fullest extent possible the Soviet Union, which, with the assistance of the bitterly cold winter weather, had repulsed the Germans at the gates of Moscow and Leningrad. The aid for the Soviets was to go the long way via the Cape of Good Hope to the Persian Gulf and also by the shorter route to Murmansk on convoys, sailing from Iceland and fully supported by Allied naval forces, including, when and if necessary, capital ships.
• Carry out Gymnast, the Allied invasion of French Northwest Africa, by May 25.
• Commence immediately an American troop buildup (Bolero) in the British Isles for Sledgehammer in 1942, or if that operation proved to be unfeasible, for Roundup in 1943. Five or six American infantry and tank divisions which had completed most of their training were to embark for Northern Ireland and England as soon as possible. These forces would also serve as a deterrent to a German invasion of the British Isles, still believed to be a possibility.
• Stop the Japanese short of Australasia. For that purpose, the Americans were to hold open a “line of communications” running from Hawaii to the South-west Pacific islands of Samoa, Fiji, and New Caledonia and commence a military buildup in those places and in Australia for counterattacks.
• Substitute the American Army’s 5th Infantry Division for the American Marines and for the British forces in Iceland, enabling the Allies to redeploy those forces to active battlefronts.
The Arcadia decisions generated a huge demand for shipping: troopships, freighters, tankers, and large and small landing craft. Merely to meet their first and most urgent assignments, the Americans were to deploy over 100,000 men and a million tons of supplies to overseas bases in January and February alone. The later, larger-scale operations (Gymnast, Sledgehammer) were to require staggering numbers of troops, weaponry, landing craft, and tens of millions of tons of supplies. However, it soon became evident that there was nowhere near enough shipping to expeditiously carry out all these operations. The new chief of Army plans, Dwight D. Eisenhower, expressed the situation simply but forcefully in his diary: “Ships! Ships! All we need is ships! What a headache!” Churchill later put it more eloquently: “Shipping was at once the stranglehold and sole foundation of our war strategy.”
As an emergency step, it was agreed to pool British- and American-controlled merchant shipping. Although the British merchant marine had suffered heavily in the twenty-eight months of war, as related, the losses had been offset to a considerable extent by new construction and acquisition of foreign shipping. Hence it was still very much a formidable force, amounting to some 20 million gross tons, but it was stretched thin to meet British commitments, and many of the ships were worn out or laid up awaiting battle-damage repairs. Sorely neglected in the prewar years, the American merchant marine numbered only about 1,500 ships of about 8 million gross tons. About a third of the ships were old and decrepit and another third were tankers unsuitable for troop and cargo movements. However, when pooled, the two merchant fleets (including tankers) amounted to almost 30 million gross tons.
Pooling assets helped, but it did not solve the acute shortage of shipping. As a result, the planners had to cancel or defer many of the courses of action agreed upon at the Arcadia Conference. One of the first casualties was Gymnast, the early invasion of French Northwest Africa. It was postponed indefinitely. The deployment of American Army troops to Iceland (Indigo) and to Northern Ireland (Magnet) had to be trimmed back and stretched out. Although the Americans continued to urge the buildup (Bolero) for a 1942 invasion of Occupied France (Sledgehammer), it, too, had to be canceled, replaced by the alternate plan, Roundup.
For the long term it was agreed at the Arcadia Conference to build merchant ships on an unprecedented scale. The U.S. Maritime Commission was ready.* At the time of Pearl Harbor, the six British “jeep” carriers under construction were Areher, Avenger, Biter, Dasher, Tracker, and Charger. The last was retained by the U.S. Navy for training air crews, replacing its only “jeep” carrier, Long Island, which was sent to the Pacific. Ten of the twenty-five newly ordered “jeep” carriers were to go to the Royal Navy, raising the number of Royal Navy “jeep” carriers under construction in American yards to fifteen. Later, American shipyards produced another twenty-three “jeep” carriers for Great Britain, making a total of thirty-eight, two of which went to the Canadian Navy. British yards produced four. To January 1, 1942, the Maritime Commission had produced 185 ships (54 tankers), of which 103 (27 tankers) were built in 1941. Before the Arcadia Conference, Roosevelt had directed the commission to build 12 million tons of new shipping: 5 million in 1942 and 7 million in 1943†. During the Arcadia Conference on January 3, Roosevelt raised the goal to 18 million tons: 8 million in 1942 and 10 million in 1943. After the conference, on February 19, Roosevelt raised the goal yet again to 24 million tons: 9 million in 1942 (750 ships) and 15 million in 1943 (1,500 ships). When—and if—achieved, the last program was to increase the combined Anglo-American merchant fleet to nearly 50 million tons, not counting projected losses.
The early, urgent troop movements arising from the Arcadia Conference imposed a great strain on the U.S. Navy. Coming on the heels of the disasters in the Pacific and Far East, the loss of a troopship at sea would be not only another tragic setback but also an intolerable shock to the American public. King therefore continued in force the policy that all troopships were to sail in convoys and were to be very heavily escorted whenever possible by battleships and cruisers, as well as numerous destroyers, which were to maintain a continuous sonar watch and form virtually impenetrable walls of steel around the troopships.† The naval historian Thomas B. Buell wrote that the new commander of the Pacific Fleet, Chester Nimitz, “used almost the entire Pacific Fleet” to escort the first contingent of Marines to Samoa in mid-January.
The number and availability of American destroyers for convoy escort in the Atlantic in late 1941 and in 1942 soon became controversial issues and led to much acrimony between London and Washington and also between the American Army and Navy. Owing to the propensity of popular writers and even naval historians—especially British historians—to accept uncritically the assertions of one side or the other in these debates, an inexcusably distorted record has evolved. The facts are as follows.
Since the outbreak of war in Europe, the United States had been building warships as fast as possible. Owing to the long lead times required, Washington assigned highest priority to building ten new battleships, thirty-one Essex-class fleet carriers, and numerous heavy and light cruisers and submarines. Shortly after Pearl Harbor, partly at the urging of London, the Navy added twenty-five “jeep” carriers for convoy escort to the urgent list. High priority had also been granted for the repair, modernization, and refitting of British warships, for the construction of six British “jeep” carriers, and about 300 other British vessels suitable for convoy escorts.*
As a result of these construction priorities and the transfer of the fifty four-stackers to Great Britain and Canada, when America declared war the Navy confronted an acute shortage of destroyers. In total it had 177, of which only about 100 were modern (post-1934). The others were four-stackers that had been in continuous service since World War I or had been demothballed and recommissioned recently.† The American destroyers were divided three ways in December 1941: ninety-two in the Atlantic, seventy-two in the Pacific, and thirteen with the Asiatic Fleet.
At the time of Pearl Harbor, the ninety-two destroyers in the Atlantic were working very hard on various assignments. The hard work and rough North Atlantic seas resulted in the need for repairs, refits, and overhauls. Thus at any given time, about fourteen destroyers (15 percent) were in shipyards and not immediately available. The others were engaged in various special duties in December 1941, some of which bear repeating:
• North Atlantic convoy escort between Canada and Iceland. Although the American destroyers rotated in and out of that assignment frequently, at all times about thirty were carrying out this task. About twenty-five of these made up the five American escort groups; the others were at Argentia or Iceland.
• As related, in December eight American destroyers were assigned to escort a special British troop convoy, WS 12X, from Halifax to Cape Town.† These ships were tied up in that operation for most of December 1941, and several required refits thereafter.
• As part of Task Force 19, composed of Arkansas and Nashville, six American destroyers were assigned to escort Troop Convoy 16 (TC 16) from New York to Iceland. Two of three American troopships aborted with mechanical problems; three British troopships joined from Halifax.
• As related, in December four American destroyers were assigned to escort the carrier Yorktown and the light cruisers Richmond and Trenton to the Pacific.
• In January, five American destroyers were assigned to escort the three older battleships to the Pacific. Two other destroyers went there as well.
• Numerous American destroyers were assigned to escort the remaining battleships, carriers, and heavy cruisers of the Atlantic Fleet, some of which were on standby to intercept a sortie of the super battleship Tirpitz and/or other big German ships or the Vichy French warships in Martinique.
There was to be no dramatic increase of the Atlantic destroyer force in the months ahead. Construction of new destroyers was not yet in high gear. The Navy commissioned only two new destroyers in December 1941 and three in January 1942. Two more were to be commissioned in February and three in March. The new destroyers had to undergo weeks of shakedowns and workups before joining the fleet.
Worse yet, the Navy had no ships other than destroyers suitable for blue-water convoy escort. That it did not was a noteworthy and regrettable lapse. President Roosevelt and, later, naval historian Samuel Eliot Morison and others blamed the Navy.* But in his postwar memoir, King blamed Roosevelt. The official documentation of that era supports King overwhelmingly.
Early in the European war, the Navy’s General Board—a group of senior admirals facing retirement, including King—recognized the need to acquire large numbers of vessels suitable for escorting convoys. As King wrote, the board agreed that fleet destroyers were “all right” for this purpose, but the board (like the Admiralty) deemed destroyers to be overpowered, overgunned, and too costly to produce in quantity for that task. Hence the General Board, King went on, sought a design “less elaborate than the destroyer.” Believing it was “essential to build something at once,” at King’s urging, the board recommended, as an emergency measure, quantity production of the new and proven 327-foot Treasury-class Coast Guard cutter, but this proposal was rejected by the Navy and by President Roosevelt. Although they had excellent sea-keeping characteristics and crew habitability, the Navy considered them too big, too expensive, and a little too slow for convoy escort.
In London, meanwhile, the members of Admiral Ghormley’s mission and the naval attaches closely analyzed British efforts to produce an ideal oceangoing convoy escort to serve in place of the Royal Navy’s fleet destroyers and the inadequate corvettes. As related, the first effort—the Hunt-class destroyer—proved to be a flop. The second effort—a 290-foot, 1,100-ton, 21-knot “destroyer escort,” or in British naval terminology, “frigate”—which was to be built in Canada and the United States under Lend-Lease, held much greater promise.
In early 1941, one of the senior assistant naval attachés in London, Edward L. Cochrane, a noted naval engineer, returned to Washington to take over the Bureau of Ships. On February 1 that year, Cochrane directed the design staff to rush blueprints for a slightly larger and improved version of the British frigate. With a length of 306 feet, a displacement of 1,400 tons, and a speed of 23.5 knots, this “destroyer escort” perfectly met the General Board’s recommendation for a ship “less elaborate than the destroyer.”*
When the blueprints were completed, the then Secretary of the Navy, Charles Edison (son of inventor Thomas Alva Edison), and Admiral Stark presented them to Roosevelt for approval. To their dismay, Roosevelt approved a British request to build fifty frigates, but denied authority for the Navy to produce destroyer escorts for itself. As King recalled in his memoir, when he—as Atlantic Fleet commander—asked Stark why “nothing was being done to provide adequate escort vessels,” Stark replied that “presidential approval could not be obtained.” Furthermore, Roosevelt continued to reject requests from Edison’s successor, Frank Knox, and from Stark to build destroyer escorts right up to—and beyond—America’s entry into the war.
There were several reasons behind Roosevelt’s position:
First, as King put it, Roosevelt was “something short of realistic in assessing the submarine menace.” Based on his experience in World War I, Roosevelt had a “predilection for small antisubmarine craft,” which could be mass-produced cheaply and quickly when the need arose.† At his urging the Navy had contracted for prototypes of two such vessels: a 110-foot, 14-knot, wood-hull submarine chaser (SC) and a 173-foot, 22-knot, steel-hull patrol craft (PC). But neither was suitable for hunting modern U-boats in rough North Atlantic waters.† King’s successor as Commander of the Atlantic Fleet, Royal Ingersoll, put it this way in his oral history:
The submarine-chasers of which the Navy [eventually] had a lot were not very good. They were one of Mr. Roosevelt’s fads: he was a small-boat seaman himself and loved to cruise on little things like the [165-foot presidential yacht] Potomac and he liked small ships. But the submarine-chaser was no craft to combat the submarine on the high seas.
Second, a mindless theory had taken deep root in Washington—fostered by Jerry Land at the U.S. Maritime Commission and others—that one way to defeat the U-boat was simply to produce merchant ships at a much faster rate than U-boats could sink them. Hence, in the early days of the war, Roosevelt awarded a higher priority to building merchant ships than to any convoy escorts other than the small SCs and PCs.
Third, in his desire to pit American soldiers against the Germans at the earliest possible time (Gymnast and/or Sledgehammer) and to meet requirements in the Pacific, Roosevelt awarded a very high priority for landing craft. The estimated numbers required were prodigious: 8,000 to 20,000 (large and small) for Sledgehammer and not less than 4,000 for the Pacific. Production of these vessels in American yards also took precedence over the British frigates and American destroyer escorts throughout most of 1942.
When America entered the war, Roosevelt ordered fast, mass production of SCs and PCs. The SC program was known as “Sixty Ships in Sixty Days.”* The Navy had no objection to this directive. As anticipated, the little SCs and PCs proved to be necessary school boats for the tens of thousands of reservists coming into the Navy and for convoy escort in inshore waters and in the Gulf of Mexico and Caribbean Sea. But as the official papers of Knox, King, Stark, and Ingersoll reveal, the Navy deeply resented Roosevelt’s stubborn and quixotic refusal to authorize mass production of the destroyer escorts required for rougher waters where enemy aircraft might be encountered. In one angry letter to Stark on the subject, Knox aptly characterized Roosevelt’s policy as “Blind Folly.”†
It is abundantly clear from these and other contemporary records—and it bears repeating—that the American Navy was fully alive to the urgent need for convoy escorts and repeatedly urged President Roosevelt to authorize them. He, not the Navy, was responsible for the prolonged delay in the production of the British frigates and American destroyer escorts. Roosevelt vastly underestimated the U-boat threat in Atlantic waters and authorized only the wrong kind of ships (PCs, SCs) to deal with the threat, both unfortunate—and ultimately scandalous—miscalculations for which Roosevelt avoided any blame, then or later. Statements to the effect that Admiral King in particular failed to anticipate the need for Atlantic convoy escorts or delayed the construction of same in favor of weaponry to fight in the Pacific are gross distortions.
Although the main task of the Coast Guard in prewar years had been search and rescue, it had served ably in the hunt for seagoing bootleggers in the 1920s. Therefore its 30,000 personnel (25,000 military; 5,000 civilians) and 168 named ships one hundred feet in length or larger were a welcome addition to the Navy. Most of these larger vessels, along with several big yachts, were to be pressed into service as convoy escorts. The sixty-seven larger Coast Guard vessels based in the Atlantic and Pacific were:
The most formidable of these vessels was the aforementioned 327-foot Treasury class, named for persons who had held the post of Secretary of the Treasury.* A modification of the Navy’s discontinued Erie-class gunboat, the Treasury or Secretary class had a top speed of 19.5 knots and a range of 7,000 miles at 13 knots cruising speed. Designed originally to accommodate a small seaplane for search and rescue purposes, the vessel was beamy (41 feet) and thus quite roomy. Some vessels were armed with two single 5/51 caliber guns; others with three single 5/51 caliber guns. All had depth-charge tracks; some had Y-type throwers.
A naval officer, Andrew G. Shepard, thought these seven ships made excellent convoy escorts. “They are considerably more roomy, so that they can carry a large number of survivors. They are better sea boats than destroyers, and lend themselves better to boat operations and rescues. In connection with picking up people, their hospital accommodations are superior to those of destroyers.”