Battleships and Flying Fortresses: The US Army Air Corps' Quest to Replace the US Navy
First published: 13th July 2014 | Dr. Arrigo Veligcogna
5.1. The Philippines
5.3. The Midway Interlude
In 1918, a little more than 10 years after its debut the airplane had moved from being an eccentric curiosity to a fully-fledged weapon of war. The flimsy flyer tested by the Wright’s brothers had been replaced by machines capable to perform different roles, from air to air combat, to observation, to strategic bombardment. Yet in 1919 with the Great War having just ended the airplane faced an uncertain future in the armed forces of the United States. Pacifism and worries about economy prompted a severe cut back of military expenses. The Army and the Navy not only faced severe losses in manpower but also in equipment and research. Against this background a small band of American aviators led by Brigadier General William ‘Billy’ Mitchell launched a campaign to prove the usefulness of airpower not only as an auxiliary element to land and sea operations but as independent element in itself. Furthermore General Mitchell argued that airpower had the potential to replace traditional weapons as the arm of decision in future warfare. Mitchell directed the bulk of his efforts against the US Navy and its admirals criticizing them both from the reliance on battleships and their proposal for carrier and naval aviation. Battleships were singled out as vulnerable to airpower and expensive, naval aviation as technically limited and incompatible with the tenet of a fully independent and centralized air arm.
According to conventional history Mitchell proved its arguments with the sinking of the old German battleship Ostfriesland during a joint Army-Navy manoeuvre breaking strict exercise regulation allegedly imposed to limit the effectiveness of airpower. After having been successful and having rallied public support to his crusade Mitchell was then punished by a conservative establishment and forced to retire. Yet his ideas soldiered on resulting in the creation of the US Army Air Corps and then US Army Air Forces and to the demise of expensive battleships and American success in the Second World War. Mitchell success is represented by the development and deployment of the first true mass produce strategic bomber, the Boeing B-17 Flying Fortress. Mitchell legacy was preserved giving his name to one of the most successful medium bombers of World War II, the North American B-25.
The aim of this paper is to correct this simplified version of history and instead describing how Mitchell’s efforts and its legacy, namely the support level bombing at sea resulted in dead end that instead hampered operations during wartime. First the paper will examine the sinking of the Ostfriesland and its technical value oppose to its propaganda value. Then it will chart the evolution of US heavy bombers between 1919 and 1941 with particular emphasis on the development of the B-17 to show that despite the Air Corps dogma no real effort was dedicated to tackle the problem of finding and sinking armoured warships at sea. Finally the paper will examine the record of B-17s employed in anti-ship missions in the first months of the Pacific War to show that they never lived to the expectations the theorists had placed in them.
Every sort of pretext, all kinds of people had tried to stop the tests. Strong pressure was brought to bear on President Harding, and the Congress, to withhold permission to designate the German ships as targets and thus block the experiment …
The test, held off the mouth of Chesapeake Bay in July, 1921 attracted widespread public interest. There, after naval aircraft in June had easily disposed of a surfaced U-boat, Mitchell’s First Provisional Air Brigade, hastily assembled and trained at Langley Field, attacked and sank three German ships — a destroyer, the cruiser Frankfurt, and the heavily compartmented Ostfriesland. Disputes arose as to the manner in which the experiment — directed by the Navy – had been conducted, and the Joint Board’s report tended to deprecate the effectiveness of aerial bombing. But the fact of the sinkings was indisputable, and Mitchell went on to clinch the validity of his claims by tests conducted with like results on obsolete US Battleships — the Alabama in September, 1921, and the Virginia and New Jersey in September, 1923. 
In the summer of 1921 the debate over the future of military aviation in the US had reached its peak. While no uniformed man was suggesting aircraft had to be removed from service the main issue was about control over them and suitable missions. The Army wanted aircraft for tactical missions in support of ground operations while the Navy needed them for reconnaissance and combat air patrol over the fleet. In direct opposition of these ideas General Mitchell was forcefully arguing on an independent air force with a strategic mission based on the British Royal Air Force example. With the peculiar strategic position of the United States separated by oceans from major potential adversaries the issue of the ability of aircraft to destroy ships became critical in the defence debate. At the same time the Navy was more than interested to test the ability of dreadnoughts to survive air attacks. A first series of tests was conducted in 1920 the USS Indiana using sandbag bombs to mark hits and then detonating demolition charges on her to simulate damage. The ship was moored and stationary. The Navy experts led by Captain William Leahy concluded that while damaged the Indiana could not have been sunk. General Mitchell was informed of the tests and started to accuse the Navy to have falsified it. The Congress’ interest was piqued by Mitchell approach and the Secretary of the Navy, Josephus Daniels, decided to prepare a series of public tests to better study the problem. The Army was invited to participate in the tests. It is important to highlight that the test had been organized by the Navy and not by General Mitchell.
The June and July test were laid out as a complex series of air to sea exercises involving US Navy, US Marine, and US army aircraft. The plans were laid out in January 1921 and presented to the Army as part of the invitation. The plans were further discussed in May with several officers, including General Mitchell, in attendance. Several captured German ships were provided as stationary targets while the old pre-dreadnought battleship USS Iowa, converted in a radio controlled moving target, would have provided a moving target and be attacked with dummy bombs. The tests on the Iowa were accuracy test designed to evaluate the accuracy of aircraft against moving targets at sea and the ability of planes to actually find the target in the first place. The other tests were designed to evaluate damage from different kinds of bombs on different targets. For that purpose inspection would have to be carried after each hit on each ship. From the minutes of the May conferences it appears that General Mitchell fully understood and approved the rules, furthermore he appeared very interested in theIowa test and stressed its importance. When the plans where finalized the schedule was the following:
21 June - U-117 bombing by aircraft
Before describing the actual tests, their results, and their aftermath it is important to quickly introduce the two main targets, especially the SMS Ostfriesland. The German battleships had been described as ‘unsinkable’ but in reality was not a top of the line design. She was an early dreadnought of the Helgoland class, the second oldest dreadnoughts of the Kaiser’s navy. She had been launched in 1909 and entered in service in 1911. The armour layout was old and concentrated in a vertical 300mm armoured citadel belt on its side tapering at 170mm on the upper and lower end. Bow and stern were less armoured with only 120mm (tapering to 100mm in the bow). Deck armour varied between 4cm and 6cm and turret armour was 30cm for the turret roofs. As comparison, the most recent US battleship, the Colorado class then under construction, had the following armour arrangements. The armoured belt was 343mm thick and tapered at 203mm at both extremities. Deck armour was 89mm and turret roof armour was 127mm. The Ostfriesland could hardly have been considered unsinkable or even modern. The SMS Frankfurt was a light cruiser commissioned in 1915 with much less armour. Her armoured belt was 60mm maximum tapering at 18mm at the bow, the stern was unprotected. Deck armour was 60mm forward, 40mm amidships and 20mm at the stern. During the exercise no ship, excluding the Iowa, would have been moving and manoeuvring. All bombing targets were moored and stationary. The main purpose of the tests was to evaluate effect of hits from the air, yet perception form the public was already skewed before the tests began. For the general public it was a test about the validity of Mitchell’s contention rather than a scientific investigation of a technical problem.
The initial phase of the test proceeded without a hitch with the submarines being duly sunk by aircraft and destroyers. Then things started to go awry. General Mitchell suddenly refused to let his planes taking part in the search and attack exercise against the Iowa, a reversal from his initial enthusiastic approach during the May conferences. While three army blimps were allowed to take part in the search no Army planes were allowed to attack her. The search was successful and the target was located but the attack proved a failure. According to Colonel Henry Arnold, future first Chief of Staff of the USAF, who was the main Army observer on the exercise, 85 bombs were dropped against the Iowa while she was manoeuvring at a speed of 6 knots and only two hits were scored by US Marine flyers. Arnold then stated that Mitchell wanted him to participate in a similar test the following day, but this test had never been agreed upon in the original schedule and never happened. Why General Mitchell refused to participate in the original test has never been properly explained. The Navy had routinely conducted firing tests on manoeuvring targets and the meagre percentage of actual hit was widely known. It is possible that Mitchell was aware of the limited publicity return of such an operation, Arnold himself describes the attacks on the Iowa as unspectacular, and wanted to avoid the embarrassment of so many misses. If this interpretation is accepted then his participation in the exercises takes the form of a mere advertising show to increase his public standing rather than a open minded approach to understanding a technical problem.
After the disappointing performance against the Iowa the G-102 was duly sunken by aircraft on 13 July as it was the Frankfurt on 18 July.Even with a stationary target percentage of hits was low with the Navy planes dropping 31 bombs for 8 hits and the Army 43 bombs for four hits; six of the Navy bombs failed to detonate. The bombing of theFrankfurt is also the moment where Arnold’s account of the experiments starts to differ greatly from the account of the Navy officer in charge of the Navy air operations. Arnold flatly stated that, after the first attack made by Navy and Army planes, the Navy wanted to wave off the Army bombers while Captain Johnson states that the Army bombers had appeared on the scene too early, when the inspection team was still aboard the cruiser. Arnold also insinuates that Navy surface ships were preparing to fire on the Frankfurt, a fact never mentioned in any other report and here is directly contradicted by Johnson that was aboard the observing ship, USS Shawmut, while Arnold was not on the scene.
The next and most critical phase of the exercises played on 20 July. TheOstfrieslandwas the target. She was moored and stationary; furthermore it appeared that watertight doors were not closed to allow for ventilation and facilitate inspections during the test. Arnold reminiscences of the attack cast it in a glorious light:
But it was not the battleship Ostfriesland, veteran of Jutland, the unsinkable battleship, that was the main thing.
Not only the Press, uniformly on Mitchell’s side, but other factors had heightened the public’s interest in the fate of the Ostfriesland.
The first attempt came on the evening of July 20th. Not understanding the rules limiting him to 600-pound bombs, the observers off the Virginia Capes thought Mitchell had failed because he did not sink the ship. It was no more than they had expected! The former German Flagship had four separate skins of steel, and every sort of unsinkable bulkhead. During the war, she had not only withstood the heaviest naval gun-fire in the battle of Jutland, but had made port after a mine had exploded directly under her.
Rules or no rules, Billy Mitchell had been out to sink that battleship. His first wave of Martin Bombers were loaded with two 1000-pound bombs apiece, and after a few hits, the ship went down …. Within a matter of hours, the Navy had protested against Mitchell’s tactics. The protests, however were drowned in the wave of excited headlines. Billy Mitchell had proved his point. His bombers had done what he said they would do.
Surviving documents flatly contradict this flamboyant prose. According to Arnold embellished account four Army Martin Bombers sank the ship on 21 July, but it was not the case. The actual schedule of the tests was the following:
The schedule was well known to the observers and in fact had been agreed upon during the May meetings. The exercise was divided in two days and, because of the adverse weather conditions on 20 July, only attacks 2 to 4 took place on the first day. Close inspection of the ship revealed very few hits. Contrary to Arnold claims the 600-pound bomb scored one hit (of 11 dropped) but that hit produced little damage. Navy and Marine aircraft had instead inflicted more hits (8 with 33 250-pound bombs dropped and 4 hits of 8 500-pound bombs). After the fourth attack ended the ship was clearly listing and taking water, but external damage was negligible. During the night between 20 and 21 July the ship took even more water and was already sinking, being listing five degree to port and being three feet deeper. No attempt at damage control was performed during the night. Captain Johnson account was telling: ‘In the morning, Ostfriesland was drawing 34 feet aft and 26 feet forward. With none aboard to rectify trim, or repair damage by pumping and counter flooding, or move the ship, and with the ship’s lower after airports close to the waterline, the Ostfriesland’s end seemed near at hand.’
Yet Mitchell had still to prove his boasts and the following day the test become farcical. The agreed procedure was that bombing would have been interrupted after each hit with 1,000 -pound bombs to allow observers to inspect the damage. It was considered extremely important to understand if 1,000 pounders were able to penetrate the armoured deck and the fuses were properly working after impacting armour. The Army bombers did not follow the rules refusing to wave off when instructed and continuing to bomb without pause forcing instead Navy bombers to wave off. Still even after several hits the Ostfrieselandarmour was not penetrated. Yet water continued to pour on it and by noon she was down feet by the stern and one foot by the bow. At 12:19 pm Mitchell’s bombers with 2,000 -pound bombs (carried according to exercise rules and not by the General’s initiative) appeared on scene and started their bombing runs without pause. They all missed. The last attack with 2,000 -pound bombs was unnecessary. By 12:30 pm the ships had already started to settle down rapidly and by 12:37 pm she was going down, capsizing at 12:40 pm.
The Ostfriesland was not sunk by Mitchell’s bombers. She was sunk by lack of damage control and unchecked flooding after cumulative attacks by Army, Navy, and Marine planes. But Mitchell had staked his reputation on sinking her so he immediately started to campaign in the media and in the Congress claiming the merit for her successful destruction. The Navy was clumsy in its attempt to counteract Mitchell distortions with evidence and soon the accepted truth was that Mitchell and his intrepid flyers had sunk the unsinkable Ostfriesland. The public pressure generated by his claims was sufficient to raise the status of the Flying corps and brought the demise of the Army Coastal Artillery (that had indeed sunk the USS Massachusetts in an earlier test). But amid Mitchell’s antics the problems remained. Mitchell had prevented the navy to test the 14” armour piercing bombs and he was careful to avoid reference to the fact that the Iowa was barely touched and that no penetration of the Ostfriesland’s armour has been made. Mitchell decided to present a different story to the press and the politicians as he did previously with the Indiana test. Mitchell was more adept than the Admirals to use publicity to further its own crusade so he won the battle of Virgina Capes. Soon the headlines of newspapers were filled with Mitchell’s exploits. Using the newly gained fame Mitchell pushed its case for strategic bombing and independent air force. While Mitchell’s own ego, arrogance, and taste for fame and flavour would led to his downfall few years later his boast had been taken at face value. The aviators now advocated their ability to sink armoured ships as a matter of fact and the distorted story was accepted as real for generations. Yet the question if the Army intrepid flyers and their flying machines would have been able to do the same to a real warship remained unsettled.
Mitchell’s dreams experienced serious setbacks in the years after the iconic Army-Navy joint exercise. While airpower fanatics in the States and abroad were boasting their accomplishment technology languished. The Martin NBS-1 used in the tests were primitive bombers by all accounts. They had a limited range (450 miles maximum) and a limited payload (2,000 pounds maximum), their speed was limited to 99 Mph. They lasted almost ten years in service before being replaced by the Keystone bombers starting in 1929. The Keystone series (B3A to B6A) was a limited improvement over the NBS-1. They were still biplane twin engine bombers with limited speed, range, and payload. Maximum speed was a paltry 130 Mph (yet almost a 50% increase over the NBS-1), the payload was just 2,500 pounds, and the range was 825 miles maximum but it dropped to 350 miles with full bomb load. Considering that 1,000 -pound bombs had only produced limited damage to the Ostfriesland the 2,000 pounders were thus the logical choice the range was severely limited. The Keystone represented only a marginal improvement over the NBS-1.
While part of the slowing down of aircraft developments were due to political and economic restrictions there was a strong neglect of appreciation of warship development from the part of the newly expanded US Army Air Corps. Three other battleships were attacked by Mitchell and his planes with less spectacular and less conclusive results than the Mitchell claimed. The USS Alabama, USS New Jersey, and USS Virginia were given to the Army to be sunk as targets. The Virginia and New Jersey were pre-dreadnought battleships (Virginia class) designed at the turn of the century, their construction authorized in 1899. They had a 13,500-tonne displacement and their armour was limited to vertical protection on the armoured belt (varying from 279mm to 203mm) and a sloped horizontal deck (76.2mm to 38.1mm). Furthermore theVirginia class layout had what the Navy considered a fatal flaw, a centreline bulkhead across the boiler rooms. This feature ensured that the ships were prone to capsize once water had started to flood one of the two sides. This is how the New Jersey was sunk. The Alabamawas even older, having being commissioned on 16 October 1900. Protection included a belt 419.1mm thick (maximum) and 69.85mm for the deck armour, albeit increasing at 76.2mm and 127mm forward and aft respectively. The Alabama was designed emphasizing vertical protection against relatively short range flat trajectory shell used in the late XIX century.
Mitchell used the tests not to develop further data but to continue to stoke his own fame. No mention of the obsolescence of these battleships was made in public statements for example. Despite his claims the new tests were much less clear cut than hoped. Several bombs failed to explode (the Navy had raised the question of fuses before the summer 1921 tests). The ships were stationary and without any form of defence or damage control. Furthermore they were old and without new armoured protection. These tests were roughly repetitions of theOstfriesland stunt with added pyrotechnics and more fanfare. Yet the novelty factor was limited and they did not created the positive publicity backlash Mitchell has hoped for. If the General boasted ‘the problem of the destruction of seacrafts by Air Forces has been solved and is finished’ the result were disappointing. The Alabama was attacked for three days and despite ‘smoke and flame’ there is no detailed report on the effectiveness of the bombing. The New Jersey and Virginialikewise were repeatedly attacked before significant hits were scored. The few pictures available did not show penetrations of the armoured decks.
At the same time the Navy extensively used surplus battleships’ hulls, resulting from the provisions of the 1922 Washington Treaty, for testing purposes. Of particular interests for the US Admirals were underwater and horizontal protection. An important series of tests was conducted on the partially complete USS Washington in November 1924.The ship was initially subjected to underwater explosions to simulate mines and near miss from aerial bombs. Then the hull was expended in a series of intensive gunnery and bombing exercises. The results were in stark contrast with the accepted version of the previous test. TheWashington sustained two direct torpedo hits followed by three 2,000-pound near misses and three days of heavy seas with only a three degree list. Subsequently several 14” shells were dropped on her from 4000 feet simulating armour piercing bombs but only one (of 14) penetrated. Then the USS Texas pounded her with 14” gunfire finally sinking her. The Washington test, while involving also aerial delivered ordnance, is usually glossed over and the discrepancy between the result of attacks on old and new ships is never mentioned.
In January 1931, after having agreed to transfer to the Air Corps full responsibility for coastal defence, the Navy and the Army agree again to perform joint tests on the feasibility of air attacks against ships. The Navy provided an old transport, the USS Mount Shasta, to be located and attacked by Army Keystone B3 and B5 bombers. The bombers proved unable to even locate the target. Aircraft technology has not advanced to the point where air attacks on ships could have been considered a reliable occurrence. Several factors contributed to this lack of capability. In this period aircraft technology was still linked to the use of heavier alloy making engines heavy compared to their horsepower rating and reducing overall performances. The light structure of aircraft precluded the use of the more accurate dive bombing approach together with the heavy bombs required to sink armoured warships. Finally there was a lack of any kind of sophisticated sensor except the proverbial ‘Eyeball Mark I sight’, both for search. Bombers were thus restricted to visual sighting without any external aid. Actual bombing was marginally better. Both the Navy and the Army had settled to use copies and modifications of the British Course Setting Bombsight MK III. An early attempt by Carl Norden to produce a better bombsight, the Mark XI, had ended in failure with a system having a Circular Error Probability (CEP, an imaginary circle where 50% of the bombs are supposed to fall) of 110 feet at 3,000 feet altitude. The system was also rated by pilots too complex and difficult to use. In the end accurately bombing relatively small targets like a ships was proving almost impossible, requiring more aircraft flying in tight pattern formations to compensate for the lack accuracy. The last factor affecting the ability of the Air Corps to live up its promises was the limited number of aircraft available. The production run of the NBS-1 (divided between Martin, Curtiss, Lowe-Willard-Fowler, and Aeromarine companies) was 130. The following Keystone bombers had a total production run of 145. The number of available aircraft was thus insufficient to ensure the proper mass for attacking a single modern battleship using inaccurate level bombing. By extension it was insufficient to provide any meaningful defence to the Continental United States and its colonies.
Yet by 1933 several technical developments promised a revolution in the capabilities of aircraft. Carl Norden had continued to work on his bombing system and by 1929 the Navy had awarded him a contract to develop a stabilized gyroscopic bombsight that promised to be able to increase accuracy in bombing. The system was also coupled to a simple autopilot and computing system reducing the number of separate activities required from the bombing officer. Then in 1930 Boeing Aircraft Company tested the Model 200 ‘Monomail’, an innovative all metal monoplane. The company quickly seized the initiative and produce a bomber version of it, the experimental YB-9, at its own expenses. The new bomber raised the maximum speed to 188 Mph a quantum leap compared to the Keystone B6A. Only seven YB-9 (1 YB-9, 1 Y1B-9, and 5 Y1B-9A) were procured because an even more impressive machine had appeared. Martin Aircraft Corporation had developed the XB-10 (model 123) at its own expenses and then sold the plane to the Air Corps. The Model 123 was a twin engines all metal monoplane with an astonishing (for the time) top speed of 207 Mph, superior to that of contemporary fighters. Its range was 600 miles with a full 2,260-pound payload. The XB-10 evolved in the YB-10 with enclosed cockpit for the entire crew and the production B-10B with rotating turrets for the defensive machine guns. 153 Martin Bombers were produced for the Air Corps from 1933 onward finally providing the aviators with a capable modern bomber.
At the same time the Army continued its demonstrations against ships. Lacking real targets to be sunk these exercises usually involved small formations of bombers trying to find a ship underway to prove the feasibility of air interceptions. The targets were usually US Army Transports and in several cases the location of these ships was established by their own communications (weather reports), by their schedules, and by reports from other ships. None of these exercise even approached war conditions and none was able to claim to prove that similar feats could be repeated in actual combat conditions.
In April 1934 the USAAC issued two new requests for bomber aircraft; an experimental project for a four engine long range bomber, Project A, and a requirement for a new heavy bomber to replace the new B-10B. Capitalizing on the new technologies available the USAAC was determined to push the capabilities of bomber aircraft to the extreme. Project A resulted in the innovative Boeing XB-15 prototype and paved the way for the future B-29 Superfortress, while the other requirement resulted in the Boeing Model 299 prototype, an all metal four engines bomber with retractable undercarriage, a maximum speed of 236 Mph, a payload of 4,800 pounds of bombs, and an impressive range of 3,101 miles. While the aircraft was exceptional it was lost in an accident during its second flight and the contract for the new bomber was awarded to Douglas with the more conventional twin engine B-18. Due to the exceptional characteristics of the Model 299 the USAAC decided to order 13 more aircraft for testing purposes and called them YB-17. From a technical standpoint the decision of the USAAC staff was correct. The YB-17 proved a better plane than the B-18, but in 1936 the Model 299 was too expensive (432,034$ compared to approximately 200,000$ for the B-18) and it was extremely complex. The B-18 was more reliable and in 1937 217 more, of the improved B-18A version, were ordered. The reduced cost of the B-18 allowed the Air corps to start to create and train more units. On the other hand the Air Corps was not blind to the potential of the Model 299 and continued to develop it. The 13 YB-17 were followed with a single YB-17A with supercharged engines and then by 39 B-17B in 1938, the first production aircraft. The B model evolved in the C model in 1940 (38 ordered) with more powerful engines and a different defensive armament arrangement. In the same year 42 D models were ordered representing the end of the initial limited production runs. The B-17D was replaced by the B-17E and mass production started. The E model was an improved product featuring rotating turrets opposed to machine gun blisters and countless refinements, the most visually striking was the replacement of the ‘shark fin’ tail with the rounded one traditionally associated to the Flying Fortress.
Even with a limited number of B-17 in service the USAAC was still determined to push its image of defenders of the coast. The best opportunity came in August 1937 when joint exercises were held off the Californian coast. The new remote controlled target of the US Navy, USS Utah, was to simulate a large fleet of two battleships, one carrier, and several other surface units. A large bombardment force was assembled including 30 B-10, 7 YB-17, and 4 B-18. An US Navy patrol wing of 30 planes under the command of Captain Ernst J. King provided reconnaissance and scouting for the fleet. During the first day of the exercise, 12 August poor weather hampered the operations but the Navy scout planes searched for the target. At 13:57 a Navy scout found the target duly relaying the position to headquarter. The Air Corps bombers took off after 16:03 and received tracking reports from Navy scouts during their flight, but were unable to find the USS Utah. According to King there was an error of one degree between the initial sighting report and the tracking reports. While the official USAF historian claimed that then there was no wonder that the bombers never found the ship the mistakes clearly illustrated the unreliability of search, report, and attack method and the kind of errors that could have happened in actual combat conditions. The exercise continued the day after with fog and bad weather still playing a critical role. The Utahwas located again but the exercise was interrupted and the target ship, at the time with crew, asked to move to a precise position for bombing exercise the day after. The Utah captain assumed no other activities were to take place. Still several YB-17s were able to locate her and attack her without the ship being able to perform any evasive action before the commencement of the bombing runs. Even in that condition the Utah proved no easy target. Her captain increased her speed and initiated several evasive manoeuvres.
‘At 1147 planes started bombing from about 600 feet. Planes well within gun range during approach, and could have been subjected to heavy Anti-Aircraft fire. Ship started to make radical and irregular changes in course. At 1159 last plane dropped bombs. Total number of splashes observed was 50.Total number of hits on ship 3.’
The day after the ship was again ‘bombed’ with water filled bombs. Again Air Corps YB-17s bombed after the exercise had been suspended. Some of the USAAC officers accused the Navy to have ‘hidden’ the Utah in the fog, but the Air Corps Chief of Staff, General Andrews, who flew as a passenger in one of the YB-17s, was adamant that any enemy force would have tried to use adverse weather to cover its approach. Again joint exercises proved inconclusive. Hits scored were minimal compared to the number of bombs dropped and, considering the data coming from the USS Washington, did not appear to have been sufficient to have inflicted damage on a real armoured warship. Yet they confirmed the idea in several Air Corps officers that they were right in their advocacy of the use of heavy bombers against ships. The Air Corps agenda was further boasted by another spectacular stunt in May 1938. The Italian liner Rex was the target of a trio of B-17s that overflew her and dropped a message container when the ship was approximately 1,000 miles from New York. The event was staged with camera crews and journalists on board the bombers, the liner was cooperating providing continuous position reports and following a fixed course. While the B-17 proved their ability to fly over open ocean and their navigators, including 1stLieutenant Curtis E. Le May, displayed excellent skills the real feasibility of similar action in wartime was not addressed.
By December 1941, despite 20 years having passed from the sinking of the Ostfriesland there was no real evidence supporting the claims of the Air Corps.
While the Air Corps finally had a proper bombardment force equipped with effective machines the doctrine behind it had never been fully tested. The 1921 and 1922 test had been performed against old ships with relatively obsolete protection scheme. Furthermore no evidence was available that even the 2,000-pound demolition bombs had penetrated armoured decks. The only test performed against an up to date armour scheme were not comforting, but there is no evidence that the Air Corps was ever interested in the data coming from the Navy tests. As far accuracy was concerned, despite the adoption of the Norden bombsight, accuracy against moving targets had not improved in significant way. In addition no tests had been conducted with opposing fighters or real antiaircraft fire. Summing up the condition of the battle between bombers and battleship on the eve of the US entry in the Second World War the reasonable conclusion was that, despite progress and innovation, Air Corps doctrine rested on flimsy concepts. Only actual tests in combat could have provided reliable validation – and such a test was about to happen.
US Army Air Forces (the Air Corps having changed its name in June 1941) claims to be able to stop a seaborne invasion of American territory were tested immediately after the Japanese attack on Pearl Harbour. On 8 December 1941 35 B-17Cs and Ds were stationed in the Philippines with the 19th Bombardment Group. While several of them were destroyed on the ground during the initial Japanese strike the survivors were immediately thrown in the fight of the islands. Five bombers, two operational, were in Luzon at Clark Fields. 15 more were ready at Del Monte on Mindanao. On 10 December 5 B-17s made the first attack on a Japanese invasion convoy. The bombers released 20 100-pound bombs for minimal result over a Japanese naval force in the port of Vigan. Later in the day a lone B-17C claimed to have sunk the Japanese Battleship Haruna. In fact the bomber, flown by Lieutenant Colin Kelly Jr., had attacked and damaged the minesweeper W-19 forcing it to beach. Two other B-17s attacked another task force from 25,000 feet, the first without reporting any result the second claiming one ship in flames. Japanese reports confirmed two merchant ships having being damaged by aerial bombing in the day. On 12 December a larger force of seven B-17s attacked Japanese shipping at Vigan again, without any result. On 14 December six bombers were sent against Legaspi on Luzon Island to attack a reported aircraft carrier. Three bombers aborted the mission and the other three were able to attack Legaspi claiming to have sunk two transports and damaged two more. Japanese sources admitted no losses and the US command decided to withdraw the remaining B-17s to Australia. By 19 December the entire bomber force, reduced to 13 operational B-17s, was concentrated in Australia having failed to affect the ongoing campaign or to sink a single large ship. A further raid was launched against Lingyaen, the site of the main Japanese landing on Luzon, with the bombers staging in the Philippines at Del Monte air base. Nine bombers were dispatched on 22 December and six launched the first raid but two aborted. Due to the distance involved and the need to carry additional fuel tanks in the bomb bay only four 500-pound bombs were carried by each aircraft. A single B-17 was able to drop its payload and no damage was reported. Two other inconclusive raids were launched one on 23 and another on 25 December but both failed. The B-17 contribution to defence of the Philippines was thus concluded.
The 19th Bomb Group was reorganized with additional aircraft (including several new B-17E) coming from the USA and dispatched to the Netherlands East Indies (NEI). Ten B-17s deployed at Malang in January 1942. From there eight of them attacked Davao on 4 December claiming a destroyer sunk, three hits on a battleship, and damage to several other ships in the harbour. In fact only the heavy cruiser Myokowas hit on its number 2 turret. The same cruiser was hit again in a follow on attack on 9 December by five B-17s (of a force of nine) while she was under repair in Davao. A further success was claimed on 19 January when a 10,000-tonne tanker was reported blown up and a cruiser damaged, but no damage was reported by the Japanese. On 24 January seven B-17Es and two Ds attacked Balikapan (Borneo) and reported to have left one transport blazing. Again no damage was reported by Japanese sources. On 26 January, B-17s made repeated attacks on Balikapan damaging the destroyer Hatsuhara. Further attacks on Balikapan were made on 29 and 30 January without results.
B-17s operating from Australia started to bomb the important port of Rabaul, now in Japanese hands, in February. A mission on 23 February ended in failure with only one bomber able to reach Rabaul and being severely damaged without scoring any hits on the transports seen in the harbour. A further strike on 18 March claimed to have ‘shred the bow’ of a large Japanese cruiser, yet the hit was not confirmed by any Japanese source. During the Battle of the Coral Sea again B-17s crews claimed hits, in this case on a carrier, presumably the Shoho, and a larger cruiser. The attacks were made in the same general time span of the attacks from carrier planes on the same targets. Japanese sources do not mention the B-17s at all.
In June 1942 the Army Air Forces face a new test of its ‘vaunted’ anti-surface capability during the critical battle of Midway. 17 B-17Es were dispatched at the end of May to the small atoll to bolster the defences and provide a long range anti-ship capability. The bombers were involved in the battle from 3 June with a strike on the invasion force commanded by Admiral Kondo Nobutake. Nine planes attacked in three groups from 8,000, 10,000 and 12,000 feet. The bombers claimed several hits, sinking a battleship and damaging a heavy cruiser. The next day 14 B-17s attacked the Japanese carrier force from 20,000 feet claiming to have sunk at least two carriers, if not more. After this mission the B-17s claimed another transport on 4 June and on 5 June they hit again a group of cruisers twice. B-17s also attacked the damaged Japanese carrier Hiryu from low altitude. While they did not scored any hits their machine guns were used to great effect. After every strike the B-17 crews reported great successes. In one case a crew claimed to have left a carrier ‘sinking and burning at the same time.’ Colonel Sweeney, the detachment commander, was supporting these claims even hinting of further damage inflicted to the Japanese but not recorded. Sadly the performance of the B-17s at Midway reached comedy proportions. With 300 bombs dropped no hits were achieved (an astounding 0% hit rate), two planes were lost, but Army pilots still reported incredible success including several carriers and battleships sunk. These claims would not be refuted officially until the final report of the JANAC commission after the war.
Midway is indeed important because it was the first instance when B-17s were employed in reasonable numbers and not in one or two-plane formations. According to pre-war assessment some hits had to be made based on plain percentages. Yet, while the Norden bombsight was indeed a technical improvement, the idea of hitting moving ships from high altitude with a strategic bomber was simply not working. Japanese ships simply waited after the B-17s had released their bombs and then started evasive manoeuvres, the time interval between drop and impact guaranteeing a successful evasion. Even the sophisticated Norden system was not able to predict the course a manoeuvring target. While the system was able to put the bombs in a reasonably limited area more often than not the targets had already vacated the impact area when the bombs arrived. It was not a technical shortcoming of the system as it was hinted earlier, but completely wrong approach. Technically high altitude level bombing against moving and relatively small targets was not a viable proposition, yet the USAAF was unable, or unwilling, to recognize the shortcomings of the B-17 employed in a naval environment.
Midway also made clear that there was no way to correctly assess result from high altitude attacks. With a Japanese order of battle that was known in advance it would have been easy to realize the over claims made by the pilots. Yet every crew that claimed hits reported distinct evidence including location of the hits and effects. An observer can only wonders how such precise claims as ‘three hits on the decks, four on the waterline’ could have been made from 22,000 feet with simple optical magnification and in the midst of splashes from near misses and antiaircraft fire. Furthermore the four carriers of the Kido Butai were a known element and by 6 June the US commanders knew that they had been sunk not by Army B-17s but by Navy SBDs. This fact had to be recognized and had to prompt a re-evaluation of the approach to aerial anti-shipping attacks, yet the USAAF stuck to its doctrine despite mounting evidence of unreliability of both the system of attack and the method to evaluate results.
During the campaign for Guadalcanal and the Solomons, the B-17s actually sank several ships. One destroyer, the Mutsuki was straddled by three bombs on 25 August: she was dead in the water providing assistance to survivors of other ships. On 15 October, B-17s sunk a Japanese transport anchored in front of the Japanese positions at Lunga Point. B-17s attacked the crippled battleship Hiei on 13 November, scoring a single 500 pdr hit without inflicting damage. Again there were instances of overoptimistic reports, as on 24 August when B-17s reported to have hit the Battleship Mutsu three times without any confirmation. These meagre results were acknowledged by the USAAF leadership several times. The official after action assessment made by the commander of the B-17s forces based at Espirutu Santo, Colonel Saunders, was extremely frank; 60 naval targets were attacked with a total of 828 bombs with 4 targets sunk and 15 more damaged. Additionally 9 other targets were claimed damaged by near misses. Actual results were even more scarce. Even the USAAF admitted that hit ratio were as low as 1.1 percent against moving targets and 12.5 percent against stationary ships. General Millard F. Harmon, the Army Air Forces commander in South Pacific, admitted that against moving targets, including probable hits and damaging near misses, the hit ration was around 2.5 percent. Clearly Harmon was disappointed with the results and urged colonel Saunders made efforts ‘to justify the type and volume of effort we are putting into our B-I7 operations for long range strike against enemy surface objectives.’ He also encouraged his subordinate to bomb from ‘dangerously’ low altitude. He instructed Saunders to be prepared to used low level skip bombing if necessary because engaging enemy shipping was the only way to use heavy bombers in offensive missions.
Such wastefulness can only be described as a product of an unrealistic, dogmatic approach to long range bombing. Advocates like Arnold had claimed that operating at extreme range would have been one of the great assets of the bomber force, enabling them to engage enemy ships far away from friendly bases. Now they were complaining they lacked a base closer to the action. Nor were poor hit ratios a surprise – airmen had maintained that they could have massed sufficient bombers to offset this, only to discover in-theatre bombers were too few. Finally, after having claimed their ability to find targets in any condition, they now complained an average of 78% of the available bomber strength was devoted to reconnaissance missions. Reality had shown pre-war claims to be hopelessly optimistic at best.
Worse yet, at the infamous Henderson Field, the IJN undermined the B-17’s image as an effective anti-ship weapon. In the night between the 13 and 14 October 1942 a surface force centred on the two fast battleships, Kongo and Haruna (the same one that the USAAF claimed to have sunk in December 1941), bombarded Henderson Field. After more than one hour and half of shelling the force withdrew and left the airfield and its tenants in shambles. Of eight B-17s present on the strip on 13 October two were completely destroyed and one damaged beyond repair. Two more were abandoned on the field and later drained by their fuel to keep fighters operational. At first light the surviving bombers evacuated.
While some B-17 operated again in the area around Guadalcanal taking off from Espiritu Santo they were not based again on Guadalcanal until the end of the battle for the island, out of fear of a new night foray. So devastating was the October bombardment to the aviation facilities that extraordinary measures were taken by the US Navy to prevent another occurrence. In the so-called ‘First Battle of Guadalcanal,’ a cruiser-destroyer force was sacrificed; Admiral Willis Augustus Lee and the fast and modern battleships USS Washington and USS South Dakota were later sent to the island in time for the last big surface battle of November 1942. To a certain extent it could be argued that during the Guadalcanal campaign battleships and surface forces asserted their primacy over carriers and airplanes.
During the same campaign B-17s were used more extensively in the reconnaissance role, a task for which their range and endurance were perfect. They were also used, with varying degrees of success, to interdict strategic Japanese targets in the area – namely the sprawling base complexes around Rabaul and Kavieng. Both the airfields and the port complex were attacked. While post-war analysis had rebutted the exaggerate claims of hundreds of planes and scores of transports hit made by both the flyers and 5th Air Force commander, General George C. Kenney, these attacks did affect Japanese operations. Yet these attacks were not without shortcomings. Too often the raids were conducted by small formations due to limited availability of bombers. The raids were usually made at night to avoid Japanese flak and fighters, thus reducing their effect. Even more importantly these raids were performed without fixed formation thus negating the air to surface tactics devised by the Air Forces’ planners. Yet even if these raids forced the Japanese to increase their defences and damaged and destroyed valuable aircraft on the ground, they never closed Rabaul as a shipping hub.
As a last, sad footnote, General Kenney ordered a concentrated attack on Japanese shipping at Rabaul on 6 January 1943. The aim was to prevent a convoy departing the harbour for New Guinea. General Kenney ordered a night raid, but the commander of the 5th Bomber Command, General Kenneth Walker, disagreed. Walker, a former instructor and one of the leading bomber theorist in the USAAF, decided to use the raid as an opportunity to prove the validity of the bomber doctrine. Around noon 12 B-17s from the 5th Bomber Command, led by General Walker himself, struck Rabaul. They attacked in daylight and in the close pattern as prescribed by the manuals. Walker put his faith in the defensive armaments of his bombers and the fact that a concentrated bombing pattern would have ensured significant results. Yet the results were meagre: only one transport was reported sunk and Walker died in the mission. Even with proper tactics, under favourable circumstances, the B-17s enjoyed only marginal anti-shipping value. While the bomber was about to play an important role in Europe, its Pacific value was far short of expectation.
The death of General Walker over Rabual was tragic, deservedly winning him the Medal of Honor for his personal bravery. He knew the risks involved and elected to lead the mission personally. During the inter-war years, he helped to shape the level bombing approach and remained faithful to it until the end. Yet it also showed the futility of more than 20 years of doctrine. Brigadier General Kenneth Walker died in the attempt to prove a concept that had proved unworkable beyond any reasonable doubt. In such light the entire mission was a waste of men and materiel, even more wasteful because better tactics were already being introduced. The Army Air Force dogma of high altitude precision level bombing blinded a whole generation of officers on its technical shortcomings.
The Boeing B-17 Flying Fortress was an excellent plane and amply proved itself throughout the Second World War. In December 1941 the B-17 was certainly one of the most advanced, if not the most advanced, heavy bombers in existence. It was reasonably fast, well-armed, well armoured, and possessed an excellent range and bomb load. Yet it failed to perform its critical intended role in the Pacific – to be a ship-killer. Even if the Second World War proved that warships, even heavily armoured ones, were vulnerable to air attack – the B-17 was not the weapon of choice.
In 1941, 1942, and 1943 B-17s attacked countless ships following the Air Corps and Air Forces doctrine and failed. As at Midway, these failure were spectacular but harmless. In some cases, as with General Walker over Rabaul, they were needlessly tragic. General Mitchell was proven wrong and, with him, the airpower fanatics. Yet history has remembered a different version of events, one in which General Mitchell played the heroic pioneer. Why was the case? The B-17 failed in his naval role for several reasons. Some were technical, some were organizational, and some had to do with personalities involved.
Technical problems are the easiest to pinpoint. There was an underestimation of both the effect of aerial bombs on warships and the accuracy of level bombing. Even with the Norden bombsight, ships were extremely difficult to hit. The bombsight did not compensate for evasive manoeuvres, winds, or errors in estimating target speed. These problems were known and had been encountered during the 1937 exercises. Still they were never addressed with the aviators relying on the results gathered from bombing stationary ships. Even in that occurrence the hit ratios were low. Armour penetration was a known issue from the start. The Navy had always been aware of this problem, but the air enthusiast had just glossed over it.
In their haste to craft a new strategic role for the Air Corps in the early twenties the airpower enthusiasts led by General Mitchell replaced science with will. In that they were helped by a cultural bias toward the aeroplane as a representative of the future against what was labelled as old technology – symbolised by dreadnought-type battleships. Perceived costs also played an important role in supporting the airpower fanatics. At face value a half dozen of planes were much cheaper than a battleship. Even if some Air Corps officers like General George Kenney were able to maintain a cold technical approach and realized the limitations of level bombing, Mitchell distorted the results of the exercises at Cape Hatteras for his own purposes. In later years the sinking of the Ostfriesland had become a sort of epic event and has been portrayed as the forlorn crusade of a noble man, General Mitchell, against the inability of an entire class, the Admirals, to recognize the shift in technology. This version of the events, despite the obvious shortcomings displayed during the war, was reprised by Arnold and the Air force when they secured independence in 1947.
Yet if there was a villain in this story it was Mitchell. He used the exercise as a public relation stunt for his own windmill chase about airpower’s total supremacy. It is not inconceivable to suggest that he purposely broke the rules of the exercise to diminish its validity and portraying the Navy as incompetent. He created a propaganda coup instead than accepting the mixed and inconclusive results of the exercise and concentrating on improving technology. He and his political allies presented the test as the irrefutable proof of the superiority of their own idea of continental defence. The idea that the same technique and the same kind of weapon was a solution to both coastal defence and, by extension, naval warfare, and strategic warfare became popular in political circles disinclined toward military expenses. But the same story also shown a narrowly minded individual with a limited grasp of broader technical, operational, and strategic problems. Rather than being a pioneer of air warfare Mitchell emerges from the battle off Virginia Capes as one of the main reasons behind the Army Air Forces failures in the early months of 1942.
Precision level bombing became a sort of final solution to all defence problems in the United States. The entire Air Corps, with some notable exceptions like Colonel Chennault and General Kenney, accepted it as an article of faith. Torpedo and dive bombing were eschewed in favour of level bombing. But because doctrine predicated everything on level bombing there was never a serious attempt to solve the related technical problems. Accuracy, bomb pattern dispersion, armour penetration, limited damage power of bombs, or even the problems of finding enemy ships in the first place were never addressed. Air Force histories had put a spin on the ineffective performances of the bombers in interwar exercises blaming it on Navy rules and ill will. This was both a poor substitute for honest technical research and for history. In practical terms the dogmatic approach to bombing resulted in the B-17 being unable to fulfil its coastal defence mission in the South Pacific. More importantly it a resulted in the wastage of lives of crew that braved enemy defences to perform a mission that was already widely recognized as technically impossible. In their fervour to sweep away the opposition the airpower fanatics had also swept away logic and technical limitations. It is worth to note that the Imperial Japanese Navy, the only other force that seriously considered, designed, and employed heavy bombers as ship-killing platform moved early on from bombs to torpedoes. The fact that Naval officers with an in depth knowledge of ships strength and weaknesses participated in the process ensure that, while still flawed and less effective than hoped, Japanese bombers were at least technically capable to sink ships as they demonstrated in several occasions.
In the end the story of the B-17 as a flying coastal fortress is a story of hubris, lack of intellectual honesty, and inability to understand limitations of available technology. It is also a story that is replete of lessons for future planners. When a new military technology appears it often spawns groups of zealots that preach the superiority of the new over the old and the necessity to immediately embrace the new without consideration for its technical maturity and usefulness. More often than not these zealots hinder technological developments because they took a dogmatic approach to technology that eschews flexibility and adaptation. In some case these zealots distort reality, capabilities, and hampers proper technical evaluation to push forward their visions of grandeur.
The same approach often permeates military history. In an effort to present a simple linear approach to military developments in which everything is based on easy to recognize revolutions and strong personalities often historian are also swept along by the fury of the zealots. Far from being a revolution in itself the battle off the Virgina Capes was just a step in the constant evolution of offensive and defensive means. It actually delayed the development of bombers capable to deal with enemy ships and the required ordinance, at least in the Army Air Corps and later Forces. Battleships had been declared obsolete before and after Mitchell boasts. Still battleships played an important role between 1939 and 1945. Number alones show more planes had been shot down or destroyed by battleships than vice versa. Land based airpower was unable to stop the Japanese from operating around Guadalcanal and only the proper application of land based air, naval based air, and conventional surface forces allowed the Americans to prevail in the struggle for the control of the space surrounding the island. In a somewhat surprisingly turning but fitting of events it was the sacrifice of the cruisers and destroyers under Admiral Callaghan and Scott, and the brilliant tactics displayed by Admiral Lee and his battleship task force, augmented with the armour and firepower of theUSS Washington and USS South Dakota, that saved Henderson Field and its aircraft from a repetition of the October disaster. While airpower in all its form was important in the desperate battles for Guadalcanal at the end it was the surface ships that decided the campaign. Even more importantly in that decisive clash the B-17, the wonder weapon that had been sold to the public as the definitive answer to America’s security needs, played only a marginal part.
If this story has a relevance for the future is that no single weapon system, notwithstanding how innovative, well designed, revolutionary, and ‘sexy’ is, will dominate the battlefield. Every weapon system has strength and weaknesses that have to be exploited as part of integrated system. It also show the importance for military historian to maintain a level head approach to the past and refuse to be drawn in inter-service turf wars and propaganda and be able to maintain an ability to recognize limits and potential of technologies.
 (Arnold, 1949)
 (Cate, et al., 1959 pp. 25-26)
 Admiral Johnson instead point out that the tests on the Indiana were not intended to prove or disprove the idea that battleships were invulnerable to air attacks, but rather to see the effect of damage from underwater and air dropped charges on specific section of the armour of the ships. Johnson pointed out that similar tests were performed in January 1921 against the old battleship USS Massachusetts with the same purpose (Johnson, 1959 p. 11).
 (Johnson, 1921)
 (Johnson, 1921)
 (Johnson, 1959 p. 17)
 (Cate, et al., 1959 pp. 25-26) (Arnold, 1949 pp. 103-104)
 (Staff, 2009 p. 39)
 (Friedman, 1985 p. 445)
 (Johnson, 1959 p. 19) (Johnson, 1921)
 (Johnson, 1921) (Arnold, 1949 p. 103)
 (Arnold, 1949 p. 103)
 (Arnold, 1949 p. 103)
 (Johnson, 1921)
 (Arnold, 1949 p. 103)
 (Johnson, 1921)
 (Arnold, 1949 p. 103) (Johnson, 1921)
 (Arnold, 1949 pp. 103-104)
 (Johnson, 1959 p. Appendix) (Johnson, 1921)
 (Johnson, 1921)
 (Johnson, 1921)
 (Johnson, 1921)
 (Johnson, 1921)
 (Johnson, 1959 p. Appendix)
 ( National Museum of the US Air Force, 2009)
 Data for the 1932 B6A (National Museum of the US Air Force, 2009)
 (Friedman, 1985 p. 429)
 (Friedman, 1985 p. 42)
 (Friedman, 1985 p. 428)
 (Maurer, 1987 p. 127)
 (Mitchell, 1923 pp. 22-26)
 (Friedman, 1985 p. 186)
 (Friedman, 1985 p. 186)
 (Moy, 2001 p. 86)
 ( National Museum of the US Air Force, 2009)
 (National Museum of the US Air Force, 2009)
 (Bowman, 1998 p. 8)
 (Bowman, 1998 p. 11)
 The B-18 had been developed as the low risk alternative to the Model 299. It was a cheaper plane based on the extremely successful Douglas DC-2 transport aircraft. While inferior to the Model 299 it was a better fit for the political and economic conditions of the time; 134 were produced. Both prototypes had been produced at the expense of their company. (National Museum of the US Air Force, 2011)
 (National Museum of the US Air Force, 2011)
 (Maurer, 1987 p. 405)
 (Maurer, 1987 p. 405)
 (Shores, et al., 1992 pp. 176-177)
 (Shores, et al., 1992 p. 182); Bowman claims the hit was made on the heavy cruiser Ashigara and he claims the bombing was made from 22,000 feet (Bowman, 1998 p. 38). Ashigara TROM refers to an american bomber attack but does not report damage. JANAC analysis confirm it was the M-19.
 (Shores, et al., 1992 p. 182)
 (Shores, et al., 1992 p. 189)
 (Shores, et al., 1992 p. 211)
 (Shores, et al., 1992 p. 212)
 (Bowman, 2003 p. 29) While the quotation is anecdotal in nature it shows the nature of the claims made by the USAAF crews.
 (Bowman, 2003 p. 32)
 (Bowman, 2003 p. 52)
 (Bowman, 2003 pp. 54-55)
 (Bowman, 2003 p. 54)
 (Bowman, 2003 p. 53)
 (Bowman, 2003 pp. 54-55)
 (Parshall, et al., 2005)
 (Morison, 2010 p. 105); the event is confirmed by the post war interview of Cmr. Okumyia Masatak, Interrogation of Japanese Officers I. 31.
 (Lundstrom, 2005 p. 222)
 Hq. 11th Bomb. Gp., Information on Operations in the Solomon Islands from 31 July to 30 November 1942, 3 Dec. 1942
 Hq. 11th Bomb. Gp., Information on Operations in the Solomon Islands from 31 July to 30 November 1942, 3 Dec. 1942
 Hq. 11th Bomb. Gp., Information on Operations in the Solomon Islands from 31 July to 30 November 1942, 3 Dec. 1942
 Letter Harmon to Saunders, 6 November 1942.
 Letter Harmon to Saunders, 6 November 1942.
 Letter, Harmon to Halsey, 20 November 1942
 (Lundstrom, 2005 p. 301)
 (Griffith, 1998 pp. 102-103)
National Museum of the US Air Force. 2009. Martin NBS-1 Fact sheet.2009.
Arnold, Henry H. 1949. Global Mission. New York : Harper and Brother, 1949.
Bowman, Martin. 2003. B-17 Flying Fortress Units of the Pacific War.Oxford : Osprey, 2003.
Bowman, Martin W. 1998. Boeing B-17 Flying Fortress. Ramsbury : Crowood Press, 1998.
Cate, J. L. and W.F, Craven, [ed.]. 1959. The Army Air Forces in World War II; Volume 1 January 1939 to August 1942. Washington : Department of the Air Force, 1959.
Friedman, Norman. 1985. US Battleships, An Illustrated Design History.Annapolis : US Naval Institute Press, 1985.
Griffith, Thomas E. Jr. 1998. MacArthur’s Airman, General George C. Kenney and the War in the Southwest Pacific. Lawrence : University Presso of Kansas, 1998.
Johnson, Alfred W. 1959. The Naval Bombing Experiments off the Virginia Capes June and July 1921. Washington : Department of the Navy, 1959.
Johnson, Alfred W. 1921. The Naval Bombing Experiments: An Account of the Bombing. U.S. Air Service. 1921, October.
Johnson, Alfred W. 1921. The Naval Bombing Experiments: Minutes of the Pre-Bombing Conferences, May 10 and May 18 1921. Washington DC : Department of the Navy, 1921.
Lundstrom, John B. 2005. The First Team and the Guadalcanal Campaign, Naval Fighter Combat from August to November 1942.Annapolis : Naval Institute Press, 2005.
Maurer, Maurer. 1987. Aviation in the US Army, 1919-1939. Washington DC : Office of Air Force History, United States Air Force, 1987.
Mitchell, William Brig. Gen. 1923. Report on Bombing Maneuvers, 5 Sept. 1923; AFHRC 248.222-71. 1923.
Morison, Samuel Elliot. 2010. History if United States Naval Operations in World War II: Volume 5: The Struggle For Guadalcanal, August 1942 – February 1943. Annapolis : United States Naval Institute, 2010.
Moy, Timothy. 2001. War Machines: Transforming Technologies in the U.S. Military, 1920-1940. Lubbock : Texas A&M University Press, 2001.
National Museum of the US Air Force. 2011. Douglas B-18. 2011.
—. 2009. Keystone B3A, B4A, B5A, B6A Fact Sheets. 2009.
Parshall, Jonathan B. and Tully, Antony P. 2005. Shattered Sword, the Untold Story of the Battle of Midway. Washington D.C. : Potomac Books, 2005.
Shores, Cristopher, Cull, Brian and Izawa, Yasuho. 1992. Bloody Shambles Volume I. London : Grubb Street, 1992.
Staff, Gary. 2009. German Battleship 1914-1918 (1); Deutschland, Nassau and Helgoland Classes. Oxford : Osprey Publishing, 2009.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
© 2015 The Phoenix Think Tank