Chapter 4: XX Bomber Command Against Japan
FOR its program of strategic bombardment XX Bomber Command borrowed from the Eighth Air Force many of its basic concepts and operational techniques. This was not unnatural: the Eighth had a richer experience in that mode of warfare than any other U.S. air force, and because many key figures in the Twentieth Air Force and its commands had served with the Eighth, that experience was easily available. A case in point is the method of combat reporting: XX Bomber Command’s tactical mission reports were patterned directly after those issued by VIII Bomber Command. Compiled at Kharagpur within a few weeks after each mission, the reports consolidated combat and technical information drawn from the lesser combat units and the various service and technical agencies concerned. Damage assessments were brought down to the date of issue but must, of course, be subject to constant reappraisal as new information becomes available, and certain types of statistics – notably on losses inflicted on enemy fighters – must be used with caution; but for much of the information given there is no need to go to the records of the subordinate units. The reports were made for command and staff personnel who needed a more precise record than that provided by spot intelligence summaries, but they have later proved valuable enough to the historian to warrant more than a passing word of thanks to the compilers. A complete file of the reports (numbered serially according to the missions) forms the basic source for MATTERHORN operations. Much as these useful (if somewhat desiccated) battle reports resemble those of the Eighth Air Force in form, however, a view of the whole series and of the voluminous Washington-Kharagpur correspondence
reveals two important differences between the MATTERHORN program and that of the Combined Bomber Offensive in Europe.
There was first the obvious difference in intensity. This was especially noticeable if the comparison be made as of the summer and fall of 1944, when the Eighth had reached its full strength. It was true also even if the figures of the Eighth’s early days are taken. For nearly 10 months (6 June 1944 to 31 March 1945) XX Bomber Command operated in the CBI, running 49 missions with a total of 3,058 sorties. During a like period at the beginning of its career (17 August 1942 to 11 June 1943) the Eighth Air Force had run 62 missions with 5,353 sorties in spite of a slow start.1 The difference came not so much from the size of the respective forces – it was only on its fourteenth mission that the Eighth was able to equal the 98 bombers airborne on XX Bomber Command’s maiden effort. It resulted rather from MATTERHORN’s peculiar logistics system, which required a long interval of transport operations to build up a fuel stock for each strike.
The second difference is to be found in the peculiar control system for MATTERHORN which left to a Washington headquarters the choice of targets and target dates (within the limits of possibility), and a great influence over tactical means employed. The far remove of this headquarters from its combat units and from the harsh realities of the theater made for an extensive, often protracted, correspondence by radio message, teletype conference, and courier over each separate mission. Those communications and the rarity of the strikes give to each mission a flavor of distinctiveness rare in the ETO and later in B-29 operations from the Marianas. The narrative which follows may reflect this flavor rather than the intrinsic importance of the strikes, which were seldom decisive.
The MATTERHORN plan as approved by the JCS early in April had derived its target objectives from a study submitted by the COA on 11 November 1943,* giving preference to six target systems, of which two – antifriction bearings and electronics industry – were passed over by the AAF planners; to the other four-aircraft industry, coke and steel, shipping in harbors, and urban areas-was added the refineries at Palembang as a compromise with those who supported POL targets. The COA had refrained from giving relative priorities but had showed a marked bias in favor of steel and coke, and this
* See above, pp. 26-28.
Twentieth Air Force planners were willing, from operational considerations, to accept as the target system to be hit first. Thus, in a revision of the air estimate and plan on 1 April 1944, they calculated the capabilities of the force for the first phase of operations (April to September) at 750 successful sorties out of China bases; of these, 576 were to be directed against coke ovens, 74 against shipping in harbors, and qzqz roo against urban areas. Palembang was to be hit by staging through Ceylon and the aircraft industry to be saved for a second phase of operations.2
This estimate was grossly optimistic, but in general the objectives held up with something like the relative importance indicated in spite of serious changes in the tactical situation in the CBI. Save for the attack on Palembang and small efforts against shipping and urban areas, it was steel – through coke ovens – which absorbed the bomber command’s efforts through September. But before the campaign against the Inner Empire opened there was need for a trial run.
First Phase
By many at AAF Headquarters the whole MATTERHORN project was considered a shakedown operation, one which would remove the kinks from the B-29 and its using organization before intensive operations were launched from the Marianas. But MATTERHORN had its own shakedown. XX Bomber Command’s staff thought of this process as involving three stages: the mass flight to the theater, the long weeks of hauling supplies over the Hump, and the first combat mission, staged against Bangkok on 5 June. The Bangkok raid was run without fanfare, its slight achievements being falsely credited for the moment to EAC’s B-24’s. The command called it a practice mission but it was more than practice, more than dress rehearsal; it was rather the New Haven tryout before the Broadway opening.
From the fly-out to India and from the Hump operations, crews had learned much about the B-29 and its R-3350 engine under varying climatic conditions. But the transport job had curtailed the more formal aspects of training, had absorbed indeed so much of the command’s energies that men had all but lost sight of the real mission, and a soldier could propose a toast in tepid mess-hall water to “the XX Bomber Command, a goddam trucking outfit.”3 The late delivery of B-29’s to the 58th Wing in its Salina period had left serious gaps in its training program for which no amount of gas-trucking would substitute:
notably in high-altitude formation flying, rendezvous, gunnery, and bombing, visual and radar.* Because of these deficiencies, General Wolfe decided to have his first go at the enemy at night, with planes bombing individually.
Wolfe signed the first field order on 17 May, with D-day slated for the 27th.4 This plan Washington vetoed on 19 May, General Arnold insisting that only a daylight precision attempt would provide the practice needed for the type of operations contemplated.5 By that date the command had piled up in the theater a total of 2,867 B-29 flying hours, of which 2,378 were on transport service, 50 on miscellaneous jobs, and only 439 in training activities, giving an average of less than 2 hours apiece for the 240 crews on hand.6 Wolfe postponed the strike and instituted a short, intensive training program. Bombardment runs were made at a range on Halliday Island, made available by the British; crews were given some training in formation flying; and even on the Hump run, B-29’s flew in battle formation in an uneconomical effort to make up for past deficiencies.7
The primary target assigned for the mission was the Makasan railway shops at Bangkok. These had been rendered especially important by recent damage to the shops at Insein and the related campaign against rail communications; destruction of the Bangkok shops would hurt Japanese efforts in north Burma. But the deciding factors were operational rather than strategic: the mission, staged from the Kharagpur area, would not cut into Chengtu fuel stores; the 2,000-mile round trip and the Japanese defenses at Bangkok would give a real but not too severe test. Secondary targets included the Malagan railyards and the Central Station at Rangoon.8
The AAF Proving Ground at Eglin Field ran off a simulated “Bangkok mission” and forwarded the test results to Kharagpur. Where operational details suggested on the basis of the test ran counter to experience in the CBI, Wolfe’s staff disregarded them. Bomb loads were lighter, fuel loads heavier than recommended: 5 tons of bombs (500-pound GP’s in three of the groups, 500-pound M18 incendiaries in the fourth) and 6,846 gallons of fuel for each B-29. The resulting gross take-off load of 134,000 pounds was too heavy for the makeshift runway at Charra, so that the 444th Group had to stage, in equal elements, from the three other bomber fields.9 Maintenance crews, working feverishly, had 112 B-29’s ready to go by D minus 1.
Take-off time was set at 0545* – dawn – to avoid high ground temperatures so dangerous for the R-3350 and to crowd the whole round trip into the daylight hours. Preliminary briefing was held on 4 June, the final briefing in the early hours of the next morning.10
The attack was launched approximately as planned in spite of an early ground mist. With planes leaving each base at one-minute intervals, ninety-eight were airborne within sixty-three minutes. At Chakulia, Maj. John B. Keller’s B-29 crashed immediately after take-off, killing all crewmen save one. Fourteen bombers aborted, and a few others failed to reach target.11 The field order called for an assembly and flight of four-plane elements in diamond formation.12 Low clouds and haze interfered; some planes joined the wrong elements and as weather thickened others broke formation and went on singly. The route out, a dog-leg which crossed the Malay Peninsula to come at the initial point (IP) from the Gulf of Thailand, was maintained with some help from radar. Approaching Bangkok, the B-29’s climbed from 5,000 feet to the stipulated bombing heights of from 23,000 to 25,000 feet.13
The first plane was over target at 1052, the last at 1232. The intervening qzqz roo minutes one navigator described as “Saturday night in Harlem.”14 It was not an orderly affair. Heavy overcast obscured the target and forty-eight of seventy-seven planes bombing did so by radar, and since few crews had received instruction in radar bombing, “learning by doing” proved a hard way. No effort was made to maintain designated formations, and bombs were dropped from as high as 27,300 feet and as low as 17,000, sometimes after repeated runs.15 Fortunately, Japanese opposition was too feeble to add much to the confusion. Heavy flak was barely moderate in quantity and was inaccurate, scoring only a holed rudder. To aid the mission, EAC had scheduled a dawn raid by B-24’s against Bangkok’s Don Muang airdrome but had scratched the attack because of weather. This failure hurt little. Fighter opposition hardly gave the B-29 gunners a decent workout: nine Japanese fighters made a round dozen of half-hearted passes while others coyly loafed along out of range. U.S. claims were correspondingly light-one probable, two damaged.16
The trip home was far more hazardous than the time over Bangkok, with the weather (it was the eve of the monsoon) and mechanical troubles proving more formidable than the Japanese. Maj. B. G. Malone’s
* Time, unless otherwise indicated, is local.
B-29, after some engine trouble, was short of gasoline leaving the target. Malone set a course for Kunming, nearest friendly base, but his tanks ran dry near Yu-Chi, sixty miles short of his goal. Ten of the crew parachuted safely and, after receiving good treatment from the Chinese, were fetched in by Capt. Frank Mullen of Air Ground Aid Service, Kunming. Another plane crashed at Dum Durn in a forced landing; others landed away from home – twelve at wrong B-29 bases, thirty at fields outside the command.17 Two planes ditched in the Bay of Bengal. One B-29 was headed for an emergency landing at Chittagong when its engines sputtered out. Capt. J. N. Sanders put the plane down into a smooth sea. A few minutes later Spitfires of Air Sea Rescue were hovering overhead and within forty-five minutes motor launches picked up nine survivors from rafts. Desperate searches by Sanders and his flight engineer failed to locate the other two crewmen nor were they found when the B-29 floated – repeat, floated – ashore next day.18
During the return another B-29 of the 40th Group experienced continued malfunctioning of its fuel-transfer system, a common ailment of the Superfort at that period. The pilot and radio operator were killed when the plane was set down in a rugged job of ditching, but ten men (there was a deadhead passenger aboard) crawled out or were blown free by an explosion, suffering injuries of varying degrees of severity. Eight of these rode out the night in two rafts and near noon picked up their two fellows, still afloat with no more aid than their Mae Wests and an empty oxygen bottle. Both were badly wounded, one incredibly so, and badly chewed by crabs. One, Sgt. W. W. Wiseman, had kept his weakened and delirious comrade, who could not swim, alive through a night of squalls only by most heroic and unselfish action. After another day and night of suffering the ten men were washed ashore near the mouth of the Hooghly River before dawn on the 7th. Two crewmen eventually contacted natives and through them the British, and an Air Sea Rescue PBY picked up the whole party. All hands credited the recovery of the wounded to a home-made survival vest designed by Lt. Louis M. Jones, squadron S-2, and worn by the flight engineer. Carrying essential supplies and drugs (the latter safely waterproofed in rubber contraceptives), the experimental vest had proved more practical than the standard E-3 kits. The whole story as it appears in the interrogations has much of the tone of a Nordoff and Hall sea saga.19
The command assessed the mission as an “operational success”; that is, it considered the loss of five B-29’s, with fifteen crewmen dead and two missing, as more than offset by the experience gained by the crews and the data obtained on B-29’s flying under combat conditions. Strategic results were less gratifying: bombing had been spotty. Photo reconnaissance on 8 June showed that some sixteen or eighteen GP’s had fallen within the target area, a few smack on the aiming point, the erecting and boiler shops. Four other bomb plots appeared at distances of 7,000 to 10,000 feet. The damage, to quote the tactical mission report, would cause no noticeable decrease in the flow of troops and military supplies into Burma.20 But XX Bomber Command had come out of its first test not too badly, and there was little time for holding postmortems on the Bangkok shakedown.
On 6 June, before all the errant B-29’s had been rounded up, Wolfe received an urgent message from Arnold: the JCS wanted an attack on Japan proper to relieve pressure in east China, where the Changsha drive was threatening Chennault’s forward airfields, and to assist an “important operation” in the Pacific. A maximum effort was needed: how many bombers could Wolfe lay on by 15 June? by 20 June?21 Previous policy had been to delay the first strike, and each subsequent “maximum effort,” until the Chengtu stockpile could support a hundred sorties, and Wolfe had tentatively set D-day at 23 June.22 Arnold’s message indicated an emergency compromise and perhaps some impatience, and it caught XX Bomber Command at an embarrassing time. Stockpiling had lagged behind schedule from the start. The Bangkok mission had interrupted freighting by the tactical B-29’s, and on 4 June General Stilwell, invoking emergency powers vested in him by the JCS, had diverted from MATTERHORN to Chennault the Hump tonnage (1,500 tons per month) guaranteed by ATC; Chiang Kai-shek had even wished Stilwell to take over the existing stockpile.* In view of these circumstances, Wolfe replied that he could put fifty planes over the target by 15 June, fifty-five by the 20th.23 Those figures did not satisfy Arnold, who insisted on a minimum of seventy B-29’s for the earlier date and called for more intensive transport efforts.24 But it was not only a matter of laying down fuel at Chengtu; Kharagpur could equip only eighty-six Superforts with the bomb-bay tanks needed for the long flight to Japan, and of that number some twenty-odd, on past performance, would fail to leave the forward
* See above, p. 87.
area, others fail to bomb.25 Wolfe nevertheless pushed his crews on the Hump line, cut down fuel consumption in the forward area, and put the 312 th Fighter Wing on a dangerously meager ration of gas.* Meanwhile, maintenance units and crews worked overtime to condition as many B-29’s as possible.26
Staging to the forward area began on 13 June and was completed only shortly before H-hour on the 15th. Of ninety-two B-29’s leaving Bengal, seventy-nine reached the China bases; one plane with crew was lost en route. With four bombers already forward, this gave Wolfe a potential striking force of eighty-three. Staging bases were assigned as follows: 40th Group, Hsinching; 444th, Kwanghan; 462nd, Chiung-Lai; 468th, Pengshan.27
The mission directive, dated 7 June, had designated as primary target the Imperial Iron and Steel Works at Yawata. This plant, most important single objective within Japan’s steel industry, had long held top priority for the first strike, and although Hansell preferred Anshan, in Manchuria, as more vulnerable, the existing priority held.28 This choice, as well as the timing, was influenced by the “important operation” in the Pacific, which turned out to be the assault on Saipan. It was fitting that the B-29’s give indirect help in the capture of a base area earmarked for their use, and a blow at a target on the island of Kyushu should prove more effective in that respect than one against the Manchurian city. But Yawata was important enough without tactical considerations. Target folders estimated Imperial’s annual production at 2,250,000 metric tons of rolled steel – 24 per cent of Japan’s total. This output was dependent upon three coke plants, of which the largest (the Minato-Machi with a capacity of 1,784,000 metric tons a year) was designated aiming point. The secondary target was Laoyao harbor, outlet for much coking coal, manganese, and phosphates.29
The B-29’s left Bengal battle-loaded, requiring only refueling in China. Each plane carried two tons of 500-pound GP’s, considered powerful enough to disrupt the fragile coke ovens by direct hit or blast. Washington, believing the B-29’s lacked range for a formation flight to Yawata and back, about 3,200 statute miles, had ordered a night mission with planes bombing individually.30 Bombing was to be done from two levels, 8,000 to 10,000 feet and 14,000 to 18,000 feet, and each group was to send out a few minutes in advance of the main
* See above, pp. 88-89.
flight two Pathfinder planes to light up the target. Take-off time was set at 1630 which would put the planes over enemy-held territory only during darkness.31
Everyone who could get orders cut and thumb a ride headed for China. Stringent regulations imposed by the gas shortage prevented a wholesale exodus from Kharagpur, but many a staff officer found urgent business in the Chengtu area and eight general officers had gathered there by D-day. Hitchhiking on to Yawata was harder. Wolfe, himself grounded for the mission by Washington but with full power otherwise over the passenger list, was chary with passes for the big brass: “Blondie” Saunders, in command of his wing’s first mission, was the only general to make the grade. Eight correspondents and three news photographers went along, briefed on Yawata and well primed with “background” after Bangkok.32
Take-off began a few minutes early, at 1616. Two groups approximated the scheduled two-minute intervals between departures, but the other two were quite slow in getting off. Seventy-five B-29’s were dispatched, sixty-eight airborne. One crashed immediately but with no casualties, and four were forced back by mechanical failures. Individual planes had little trouble in following the outward course, a long straight haul with only a single turn near the IP, Okino Island, which was readily identified on the radar scope.33
At 2338 (China time) the first B-29 over the target gave the signal “Betty,” meaning “bombs away with less than 5/10 cloud,” but Yawata was perfectly blacked out and haze or smoke helped obscure the city. Only fifteen planes bombed visually while thirty-two sighted by radar. Crewmen saw explosions but could not locate them in reference to the aiming point.34 The enemy was alerted long before the first Superfort arrived. Returning correspondents gave vivid firsthand descriptions of the battle over Yawata, but it was not a vicious fight.35 Sixteen enemy fighters were counted by crewmen but only three fired at the bombers, and they scored no hits. The Japanese put up heavy flak and automatic-weapons fire, both inaccurate, to give minor injury to six B-29’s. Searchlights, though spectacular and bothersome, gave little help to AA gunners.36
Forty-seven Superforts over Yawata jibed pretty well with Wolfe’s original estimate of fifty,and the rest of the sixty-eight airborne could be accounted for by the sort of operational calculus he had used. Besides the four aborting and the crack-up at Pengshan, there had been
a crash near Kiangyu, cause unknown, which wiped out the whole crew. Six other planes had jettisoned their bombs because of mechanical difficulties, two had bombed the secondary target at Laoyao, and five had bombed targets of opportunity.37 Two planes of the 468th Group were listed missing but were later tracked down with great difficulty by search parties led by Capt. H. M. Berry. Both had crashed, killing all on board including Robert Schenkel, a Newsweek nor respondent.38
The only known combat loss occurred during the return flight: Capt. Robert Root’s B-29 developed engine trouble, and about dawn he put the plane down at Neihsiang, a friendly Chinese airfield near the battle lines. He called in the clear for U.S. fighter cover and with Chinese aid tried to get his bomber ready for flight again. His message brought no Americans but more than enough Japs. Within half an hour their fighters appeared, then their bombers, and after a few unhurried passes they left Root’s plane a smoldering ruin. The crew, two of them wounded, were rescued by a B-25 from Hsinching. Harry Zinder of Time, who had ridden with them to Yawata and had been reported missing along with Schenkel, arrived with the crew in time to file a delayed story.39 One other loss, not officially charged to the mission, was a B-29 reconnaissance plane which crashed when going out to photograph bomb damage. In all, the command had lost seven planes and fifty-five men without much enemy activity.40
A diversionary raid against enemy airfields by the Fourteenth had been scheduled but was thwarted by weather.41 In spite of earlier fears, the Japanese made no retaliatory attack on the Chengtu fields. This was fortunate. Wolfe had elected to cut back gas deliveries to the 312th Wing in order to stage his maximum-effort mission, and in order to get all his planes back to India he had to borrow 15,000 gallons from the fighters’ limited supply.42 On the ground for several days, the B-29’s offered a fat target, but the enemy’s lethargy justified the gamble.
Photos made by the Fourteenth Air Force on 18 June indicated that bomb damage at Yawata had been unimportant. Only one hit had been registered on Imperial’s sprawling shops and that was on a power house 3,700 feet from the coke ovens. Some damage had been done to Kokura Arsenal, to miscellaneous industrial buildings, and to business-industrial areas, which were referred to as “hospitals and schools” in the Japanese reports. The steel industry, prime strategic target, was
still unhurt.43 Indirect results, the intangibles of war, were certainly more considerable, if incalculable. Timed with the Saipan assault, the first appearance of U.S. planes over Japan since the Doolittle raid brought home to ill-informed Japanese citizens something of the realities of the war. Enemy radio reaction was sharp enough to indicate deep concern. The size of the attacking force was minimized and claims of B-29’s (and B-24’s!) shot down were headlined. There was one curious report, a broadcast claiming the destruction of a B-29 and capture of its crew consisting of six lieutenant colonels and a major – a lot of rank even for a Superfortress! Names, ranks, and hometowns were given accurately, the only error being that none of the alleged POW’s had been on the mission. It took a lot of hasty telegraphing in the States to reassure next-of-kin.44
In the United States interest was more nearly consonant with the “firstness” of the mission than with its intrinsic importance. Once the bombs-away signal had been flashed, Washington headquarters had received a blow-by-blow account of the mission in a long series of cables which were relayed to Arnold, then in London.45 Next day, still 15 June by U.S. time, the Yawata strike and the public announcement of the existence of the Twentieth Air Force competed with news of the Normandy beachhead in the headlines. If reports of this and other early B-29 strikes sometimes gave an overly optimistic impression, the fault did not lie with the Twentieth’s public relations officers. The peculiar command system had dictated a policy of simultaneous releases by the Washington headquarters and by Kharagpur, through Stilwell, and there was inevitably some friction in its application.46 But in respect to tone the Twentieth had profited again by early experience of the Eighth Air Force, whose glowing headlines had sometimes backfired. Communiqués were factual and XX Bomber Command PRO’s were cautioned to hew to the line in releases to the press. Background materials should stress JCS control of the B-29 and the importance of the air-ground-sea team, rather than individuals, in ultimate victory. All concerned should use “extreme care” against overemphasis of B-29 accomplishments, recognizing that the plane and its organization were still in an experimental stage, and extravagances such as “the Wolfe pack” and “the dodo bird becoming an eagle” were to be eschewed.47 Unfortunately, this sound policy had little effect on headline writers stateside, and conscientious reporters in the field often found their factual stories blown up by rewrite men
at home. But, for 15 June there was news enough without inflation. Yawata ( ADMEASURE I to the encoders) was no great blow but was an earnest of what was to come.
Two days after the Yawata show General Arnold informed Wolfe that, despite the depleted fuel stocks in China, it was “essential” to increase pressure against Japan. Immediate objectives were: a major daylight attack on Anshan, small harassing raids against the home islands, and a strike against Palembang from Ceylon. When Arnold asked for an estimate of the command’s capabilities, Wolfe’s reply was none too hopeful.48 With low storage tanks at Chengtu (on 21 June XX Bomber Command had there only 5,000 gallons) he could not with his own resources build up for an all-out mission to Anshan before 10 August; if ATC would deliver 1,500 tons for the command in July, he could mount the mission by 20 August. Ceylon fields would not be ready before 15 July, and either the Palembang mission or the night raids would delay the Anshan attack.49
In spite of Wolfe’s cautious estimate, Arnold on 27 June issued a new target directive calling for a 15-plane night raid over Japan between 1 and 10 July, a minimum of 100 planes against Anshan between 20 and 30 July, and a 50-plane mission to Palembang as soon as Ceylon airfields were ready. To meet this schedule, Wolfe was admonished to improve radically the operations of C-46’s and B-29’s on the Hump run.50 Washington’s judgment of maintenance and operations standards was based on records of the Second Air Force in the United States which Wolfe did not think realistic. He outlined conditions necessary for fulfilling the directive: build-up of his B-29 force and a flat guarantee of ATC Hump tonnage.51 Even when it was decided that the command would get back its 1,500 tons for July, Wolfe’s operational plan of 30 June set up the Anshan mission for 50 to 60 B-29’s, not 100.52
Arnold received this plan on 1 July.53 On the 4th General Wolfe was ordered to proceed immediately to Washington to take over an “important command assignment,” and two days later he departed, leaving General Saunders temporarily in command at Kharagpur.54 Coming as it did after repeated delays in getting the B-29 over Japan and at a time when Wolfe’s estimates were consistently under those entertained by the Twentieth Air Force staff, this transfer had something of the appearance of a kick upstairs. Wolfe, with no combat experience but with an excellent engineering background and a thorough
knowledge of the B-29, had been a natural selection for the job of shaking down the plane. Now that a high-level reorganization was in process at Wright Field, he was going back to head the Materiel Command, which would carry two stars; his experience with the B-29 under combat conditions would be invaluable in his new primary mission of expediting production and improvements of that plane. Arnold’s own opinion may be best found in a letter to Spaatz some months later: “With all due respect to Wolfe he did his best, and he did a grand job, but LeMay’s operations make Wolfe’s very amateurish.”55
The change in command had no effect on the first July mission, the small night raid ordered by Arnold. D-day, 7 July, marked the seventh anniversary of the Sino-Japanese “incident” and Chinese considered the choice of that day a courteous gesture. In truth, however, it had been determined by moon phase, weather, and modification of camera-carrying B-29’s, not inter-Allied comity.56 Because the main intent was to impress the enemy with an early follow-up on the Yawata strike, the small force of B-29’s was to be split over Kyushu: primary targets included the naval dockyards and arsenal at Sasebo, the Akunoura Engine Works at Nagasaki, the aircraft factory at Omura, and steel works in the Yawata-Tobata area; Laoyao harbor, as before, was last resort target, Two B-29’s were assigned the task of photographing the Miike Dyestuffs Plant at Omuta and all other planes carried nine photo-flash bombs in addition to their eight 500-pound GP’s.57
Between 5 and 7 July twenty-four B-29’s assembled at the forward bases, and eighteen took off from China on the afternoon of the 7th. One aborted with engine trouble, but seventeen bombed some target. Eleven planes dropped in the general area of Sasebo, but a twelfth, its radar dead, was off by fifteen miles. Single planes struck at Omura Omuta and at Tobata, while the B-29 sent at Yawata bombed instead the secondary target at Laoyao. Two others, with fuel-transfer trouble, turned back to bomb Hankow, one missing it by twenty miles. Crewmen saw explosions in all the areas bombed, but because of undercast and defective photo-flash bombs, intelligence officers could learn little about damage from the strike photos.58 Certainly the mission achieved no great amount of destruction, but it may have accomplished its main objective by demonstrating to the Japanese the vulnerability of Kyushu. In any event, the mission was cheap: no plane
was hit by flak and only one received minor damage from eight Oscars and Tonys which attacked over occupied China. All seventeen effective sorties got back to the Chengtu bases safely and, thereafter, to Bengal.59
This was fortunate, for every plane would be needed for the “maximum effort” against Anshan. Arnold on 5 July had issued a supplement to the July target directive, specifically in regard to the Anshan and Palembang missions.60 Saunders would have preferred a night go at Anshan, but on the 7th Washington reiterated the demand for a 100-plane precision daylight attack, which must be done in July.61 Saunders, who based his estimates on “a realistic analysis of conditions,” thought he could squeeze in the Anshan strike on the 30th by postponing Palembang until mid-August; Ceylon fields would hardly be ready in July anyhow.62
One group commander described the command’s alternation of transport and combat operations as “getting money in the bank and then having our spree.”63 This time the bank would support a real Saturday night bust, for July was to prove a banner month for Hump tonnage. The bomber command got 976 tons from ATC and itself hauled 1,162 in C-46’s, 1,063 in tactical B-29’s, and 753 in B-29 tankers, for a total of 3,954 tons.64 The real shortage was in planes, not in fuel. Washington was promising substantial reinforcements, but with some B-29’s converted into tankers, Saunders would have on D-day only 127 combat planes in the theater.65 To get enough of these into commission and at the forward bases to launch a 100-plane attack would tax his maintenance facilities. He proposed to knock off Hump operations ten days before D-day (instead of seven days as for Yawata) and to start staging forward on D minus 5, giving time for aborts from Bengal bases to be repaired and redispatched, and for adequate maintenance in China. This procedure, and his suggestion to stage back more leisurely than before, increased the danger of Japanese counterattacks on the Chengtu fields but Washington indorsed his plan.66 The scheme paid off, and without drawing the retaliatory raids which Chennault feared. The B-29’s started moving forward on 25 July, and by the 29th, 106 had landed at the China bases. One plane had crashed near Midnapore, killing 8 crewmen – but only 4 of the 111 bombers dispatched from the rear area had been stopped short of China by mechanical troubles and 107 were available for the strike.67
Primary target was the Showa Steel Works at Anshan in
Manchuria – specifically, the company’s Anshan Coke Plant, producing annually 3,793,000 metric tons of metallurgical coke, approximately one-third of the Empire’s total. About half of this was used by Showa’s own steel works, second in size only to Imperial’s, and the rest for various industrial purposes in Manchuria, Korea, and Japan. Alternate targets tied in. The secondary was Chinwangtao harbor whence coking coal from the great Kailan mines was exported to Japan. Tertiary target was Taku harbor near Tientsin, which handled coal, iron ore, and pig iron. As last resort, bombers were to hit the railroad yards at Chenghsien, a possible bottleneck along a Japanese supply route. Aiming point at Anshan, as at Yawata, was to be a battery of coke ovens, and again the bomb load was set at eight 500-pound GPs per plane.68 With a change in weather threatened, D-day was moved up to 29 July. Rain during the previous night mired the runway at Kwanghan and the 444th Group was unable to get off the ground at H-hour, but the other three groups got seventy-two planes up out of seventy-nine dispatched. One B-29 fell a few minutes later, killing eight crewmen. Mechanical difficulties prevented eleven bombers from reaching Anshan, of which one bombed Chinwangtao, two Chenghsien, four targets of opportunity, and four failed to bomb.69 Sixty B-29’s, flying high over enemy-held territory, got over Anshan. Most of them were able to hold the tight four-plane diamond formation and to bomb from altitudes reasonably close to the designated 25,000 feet. Bombing conditions as they went in were nearly ideal, with clear skies and still air, but the first wave messed things up by dropping a stick of bombs on a by-products plant just off the aiming point, which was thereafter shrouded with a thick pall of smoke.70 Despite Anshan’s importance, enemy opposition was not too rugged. Heavy flak caused but minor damage to a few B-29’s and fighters scored only two unimportant hits. The Superfort’s speed in the bomb run, 182 to 212 m.p.h. indicated, made it hard for fighters to jockey into position for a shot, and there was no determined boring in. The B-29 gunners claimed three probables and four damaged.71
The only combat loss occurred near the last resort target. Capt. Robert T. Mills’ B-29, losing power in its No. 2 engine on the way out, shook out its bomb load over Chenghsien. Wounded there by heavy flak, the plane was jumped by five enemy fighters, including a P-40 with a Chinese-American Composite Wing (CACW) insignia, which shot out another engine. Mills gave the “abandon plane” command.
Eight men (not including Mills) parachuted into occupied China and with the aid of Chinese guerrillas walked out, reaching Chengtu a month later.72 Chinese also helped save another B-29 which, after bombing Chinwangtao, had made a forced landing at a CACW field near Ankang. The plane was on the ground for five days while an engine, spare parts, tools, and mechanics came in by C-46 from Hsinching to effect an engine change and other repairs. Air cover was furnished by Fourteenth Air Force fighters, who shot down a Lily bomber during a night attack. With full assistance from the Chinese and American garrisons at Ankang, the B-29 took off on 3 August and returned to Chiung-Lai.73
On D-day the wet strip at Kwanghan had dried enough by ten o’clock for the 444th to get twenty-four planes up. Nearly five hours behind schedule, the group was too late for Anshan and so headed for Taku. Sixteen planes bombed there without any interference from the enemy; three bombed Chenghsien.74
The day’s work, if not perfectly executed, was at least heartening to the command. Ninety-six B-29’s had been airborne in a close approximation of the 100-sortie mission directed. Eighty planes had reached target areas, and though mechanical and personnel failures had kept the weight of bombs dropped to 73 per cent of that dispatched, the bombing looked good. A comparison of strike photos and photographs taken by Fourteenth Air Force planes on 4 August seemed to indicate a substantial amount of damage at Anshan, including hits and near misses on several coke-oven batteries, other related installations, and the by-products plant. Damage at Taku and Chenghsien too seemed substantial. The command had learned much about running a daylight mission, and all in all, the loss of five B-29’s (three in China, two between India and China) seemed not exorbitant.75
The fifth and sixth MATTERHORN missions were run on the night of 10/11 August in a double-barreled strike at Palembang in Sumatra and Nagasaki in Kyushu, 3,000 miles apart. Palembang had been accepted as a target by the CCS at Cairo, and in the schedule of operations adopted by the Joint Chiefs in April it had enjoyed high priority as the only important POL target named.* Because of the extreme range of the target and the necessity of staging through an RAF base at Ceylon, the mission involved more planning and preliminary activities (save in the matter of fuel, which was furnished by
* See above, pp. 29-31.
the British) than any flown by the command, and in execution the operational success outweighed the strategic results obtained. According to target information available in early spring, the Pladjoe refinery at Palembang had seemed to be of highest strategic importance. With a reputed annual capacity of 20,460,000 barrels of crude oil, it was supposed to produce 22 per cent of Japan’s fuel oil and 78 per cent of its aviation gasoline. Shortage of tankers limited the amount of avgas that could be delivered to active fronts and prevented export of any of the motor gas produced concomitantly, but destruction of Pladjoe would put a serious crimp in Japanese military and naval operations.76 By mid-June some agencies had revised that appraisal. In Washington AC/AS, Intelligence and the COA, neither eager about the target earlier, thought that the changing tactical situation in the Pacific and the increasing shortage of enemy bottoms had robbed Pladjoe of its paramount importance. XX Bomber Command would have been quite happy to scratch a nasty mission which promised to hamper the prime effort against coke and steel objectives, but the JCS held firm: Arnold included Palembang in the target directive for July, ordering that the mission be flown as soon as the Ceylon fields were ready.77 Active planning for the mission, begun in May, had been complicated by those delays in airfield construction which have already been described.* During June and July, officers from XX Bomber Command, from AAF IBS, and from the CBI staff worked with the British in Ceylon to perfect arrangements. Earlier plans to use four, then two, staging fields were abandoned as construction lagged; it was finally decided to complete only one field, China Bay, and to run the whole mission through it. This involved extending facilities to two-group standards, with a fifty-six-point fueling system (brought in on loan from the CBI), fifty-six hardstands, and extra taxiways. Successive delays in this additional construction kept Palembang off the July schedule, but with completion assured by 4 August Arnold set a 15 August deadline and the 10th was finally named as D-day.78
Washington had stipulated a daylight precision attack at the command’s nominal strength, I12 aircraft. With only one field available, this would have meant staging in waves, a very ticklish job where minor variations from a tight schedule could spell disaster. Warned of probable high losses, Twentieth Air Force Headquarters had relented
* See above, pp. 71-73.
and on 27 June directed a dawn or dusk strike by at least fifty planes.79 Further negotiations by XX Bomber Command effected other changes in the directive so that the operational plan finally adopted called for a night radar attack.80 Part of the force, briefed by a WSN expert, was to mine the Moesi River, through which all Palembang’s exports were shipped; secondary target was the Pangkalan refinery and last resort target was the Indarung Cement Plant at Paclang, both in Sumatra.81 Field orders to effect these plans were completed on 1 August, and final preparations were rushed through with the friendly cooperation of the British. Fuel for the mission and construction costs might be charged on reverse lend-lease accounts, but the RAF went far beyond the bare essentials, virtually giving over the base to the Americans, with housing, messes, transportation, and with available whiskey rations and best wishes thrown in.82
On the afternoon of 9 August fifty-six B-29’s landed on China Bay’s 7,200-foot strip and wheeled onto allotted hardstands, directed in without radio and without an error by a control team recruited from XX Bomber Command. Next afternoon at 1645 a plane from the 462nd Group pulled up off the runway, and within eighty-four minutes fifty-four B-29’s were airborne with only one washout, a remarkable bit of flying on a strange and crowded field. Forty minutes later Capt. I. V. Matthews’ B-29 returned with a leaky engine, got patched up, and was again winging for Sumatra within a couple of hours.83
The bombers, proceeding individually, flew a straight track to Siberoet Island just below the equator, then bore eastward across Sumatra. A dozen planes failed, for various reasons, to reach a target, but two bombed Pangkalanbrandan, one an airdrome at Djambi, and thirty-nine reached their goal.84 Palembang had no lights and some undercast, and the one B-29 equipped with flares miscarried, but thirty-one planes bombed either by radar or visually through patch clouds. Crewmen later reported having seen explosions and fires through breaks in the undercast, but their fleeting observations were none too precise and the strike photos were too poor to be of much service.85 Eight planes of the 462nd Group found clear flying over the Moesi by dipping under a 1,000-foot ceiling and laid 16 mines in a good pattern with “excellent results.”86 The B-29’s met antiaircraft fire in various places and, for the first time, ground-to-air rockets. Crewmen reported seeing 37 enemy planes, some of which followed them back for 350 miles, but not a B-29 was scratched.87
Nor were operational losses as bad as had been feared. The mission had been coded BOOMERANG, perhaps in pious hope that the planes would come back from the long round trip of 3,855 air miles to Palembang, 4,030 to the Moesi. Because of the extreme range the command had lightened the useful load to a single ton of bombs or mines and had filled gas tanks to capacity; even so, the loss of some planes on the return trip was expected. In anticipation the British had set up an elaborate air-sea rescue force comprising submarines, a cruiser, destroyers, lighter vessels, and various aircraft types. Several Superforts turned back without bombing because of threatened fuel shortages, but only one went down at sea, and there rescue precautions paid off. The B-29 sent out an SOS giving its position ninety miles out of China Bay and somewhat off the return track. An intensive search was finally successful when planes and HMS Redoubt homed in on a “Gibson girl” signal from a life raft on the morning of the 12th. One gunner had been killed in the ditching but the other crewmen were picked up.88
Operationally the mission had been very successful; the skill with which it had been planned and executed was indicative of what XX Bomber Command had learned since the over-water flight to Bangkok. The command had never been keen on the assignment, and Washington’s insistence on a mission which under existing conditions had little chance of decisive results seems now to indicate a lack of flexibility in target priorities. The attack did little to speed the war. Photo reconnaissance was not flown until 19 September, by which time it was difficult to assess bomb damage accurately, but it appeared that in spite of earlier impressionistic reports of large fires, only one small building destroyed could definitely be credited to the strike. That and several probables were small returns for the effort.89
This disparity between effort and results XX Bomber Command realized, and on 24 August, long before damage assessment had been made, the command recommended the abandonment of China Bay as a staging base. Eventually convinced, Washington on 3 October gave permission for the command to remove all its own equipment, leaving behind for possible future use the fuel-service system which belonged to AAF IBS. But the B-29’s staged no other raids through Ceylon. The cost of developing China Bay into a VHB base for a single fruitless mission, whether figured in terms of effort and materiel or funds, is a glaring example of the extravagance of war.90
The reduction of BOOMERANG from a 112-to a 56-sortie mission left aircraft available for a strike elsewhere. Saunders had learned on 21 July that he would get an additional 1,500 tons laid down in China by ATC and next day informed Arnold of his capabilities for August: a small night incendiary attack on Nagasaki early in the month, a saturation incendiary attack on the Yawata-Tobata area on the zoth, and a daylight strike at Anshan on the 30th – these in addition to Palembang. Yawata-Tobata and Anshan would be “major” effort, calculated each at sixty-eight planes dispatched, fifty-five over target. Washington found the sortie rate “gratifying” but the effort too diffuse, and was willing to scratch Nagasaki to add weight to the other missions. This Saunders opposed, and on 28 July Twentieth Air Force Headquarters approved his schedule, but demanded eighty planes dispatched on the major strikes instead of sixty-eight.91
For psychological reasons the small night raid was synchronized with the Palembang mission. Nagasaki was to be chosen as target for the second atom bomb (9 August 1945) because that city had previously suffered little from air attacks. Yet in the summer of 1944 Nagasaki, with its crowded shipyards, docks, and military installations, carried high priority among urban industrial area objectives. Because of the reputed vulnerability of Japanese cities to fires, Twentieth Air Force planners had expected to combine saturation incendiary attacks, delivered by night, with precision bombing. Somewhat to the concern of Washington, the bomber command had so far relied wholly upon high-explosive bombs; now Saunders proposed to use fire bombs on Nagasaki, which Washington earlier estimated could be “saturated” with seventy-six tons.92 The Point Island military storage area at Shanghai was named as secondary target in spite of the possibility of criticism should strays hit in Chinese residential districts or POW camps nearby.93 The Hankow docks were chosen as last resort target. This was a gesture to Chennault, who had earlier asked that the B-29’s hit key positions in the Japanese supply line, and who was to repeat his request before the Nagasaki mission was run.” Staging for Nagasaki, as for Palembang, reflected increasing operational skill and resulted in a stronger effort than the twenty-five sorties promised. Of thirty-three bombers leaving Bengal, thirty-one arrived at forward bases; one was lost en route with two crewmen killed. Bomb load was heavier too, with 5,816 pounds of incendiaries
* See below, pp. 112-13.
and frags per plane.94 Twenty-nine planes got off in thirty-six minutes, of which two returned early, three bombed targets of opportunity, and the other twenty-four, flying out individually, released their bombs over Nagasaki. The city, blacked out and under light cloud cover when the first plane arrived, became progressively harder to see so that only eight planes bombed visually.95 As on other night missions, results could not be accurately estimated, but later intelligence was to show them not too significant. Enemy resistance, both by flak and by interceptors, was weak and not a single B-29 was scratched. Yet the air battle was memorable for one reason. T/Sgt. H. C. Edwards knocked off a Jap fighter at 600 yards with the first burst from his stern chaser 20-mm.; the fighter was seen to crash in flames and Edwards was credited with the command’s first official kill.96
On the way home Capt. Stanley Brown’s B-29 became lost after some mechanical trouble and early next morning, almost out of gas, landed at Hwaning, held by the Chinese but within easy reach of three enemy fields. The plane bogged down in mud at the end of the short strip, and Japanese strafers knocked out two engines. But the 312th Wing sent out a fighter cover which shot down three enemy planes and scored heavily on Jap fighters parked on one of the adjacent fields. Fuel, parts, and mechanics were flown in, and the B-29 was repaired and stripped to its bones. Meanwhile, the Chinese had jacked it out of the mud and slowly inched it down the strip, building a short runway by sinking 4,500 railroad ties in the soft spots. On 23 August the plane with only four crewmen aboard lifted off the ground and flew into Chiung-Lai to complete a most extraordinary job of salvage.97
The build-up for the seventh mission had begun well before the combined Palembang-Nagasaki strikes, and by 1 August Saunders had hopes of meeting Arnold’s demand for eighty sorties.98 The tactical situation in China threatened for a time to interfere. And, after vainly attempting earlier to have the B-29’s sent against Hankow’s waterfront,99 Chennault was now asking for 300 sorties against Hankow and Wuchang.100 Saunders, consulted by Arnold, declined to pass judgment on the strategic worth of Chennault’s plan but reported that such diversion of effort would make it impossible to comply with the bomber command’s current directive.101 When his second request was refused, Chennault on 10 August proposed that the command be shifted from current attacks on steel to a counter-air campaign or be
withdrawn from China.102 Arnold’s staff was interested in the aircraft industry and on 5 August had asked Saunders’ opinion about substituting the Omura Aircraft Plant for Anshan for the eighth mission. This Saunders was loath to do, preferring to finish off the Showa works, Yawata, and Penchihu before turning to the admittedly important airplane factory. Washington abided by this decision, perhaps a little suspicious of Chennault’s motive: Stratemeyer, after a visit to the Fourteenth’s headquarters, expressed to Arnold the opinion “that Chennault’s repeated requests for B-29 missions against Hankow are for use of those airplanes primarily from consideration of their own supplies being available in China.”103 At any rate, Saunders’ immediate objectives stood.
He had originally planned his seventh mission as a night incendiary attack on the Yawata-Tobata urban area. Within his staff a number of officers, perhaps the majority, had come to favor the employment of the Superfort exclusively as a night bomber. So far, crews had not been able to deliver rated bomb loads on lengthy missions flown by day in formation, and these staff officers believed a moderate bomb load dropped by radar at night was more effective than a minimum load carried by day. Some even wished to increase the weight of effort by converting more B-29’s into tankers, and with larger fuel supplies in China to send the whole force out more frequently, tankers sandwiched in with the tactical planes in night saturation attacks.104 Preliminary appraisal of photographs of Anshan taken after the 29 July strike seemed to strengthen the case for precision daylight bombing (though from the same pictures Chennault concluded that the bombing of the Showa works was futile), and on 4 August the command asked for permission to run the Yawata mission as a daylight attack on the Imperial Iron and Steel Works. The Twentieth acquiesced, “delighted with the change.”105
Flying in formation in a high-altitude approach through enemy territory and bombing from 25,000 feet, the B-29’s could carry only a light load. To avoid setting a standard load determined by the poorest crew’s performance, Saunders inaugurated a new policy by prescribing a one-ton minimum and allowing group commanders to set the loading according to the known efficiency of each plane and crew. This varied considerably – on the Yawata mission individual B-29’s would burn as little as 6,100 gallons of fuel, as much as 7,600 – and the
flexible loading scheme allowed the bombers to carry an average burden of 6.3 x 500-pound GP bombs.106
By D-day, 20 August, ninety-eight B-29’s had gathered in the forward area with one lost en route from India. At take-off, three groups got away without accident, but the eighth plane of the 462nd Group smashed up, blocking the south end of the runway. By afternoon it was possible, by lightening loads, to get eight more planes up over the wreckage and, joined by five aborts from other groups, they went on to Yawata for a night attack. With seventy-five B-29’s airborne for the day mission and thirteen for the night, the command had more than met requirements.107
On the daylight run, sixty-one planes dropped ninety-six tons of high explosives on the target area; six others hit the secondary (Laoyao) or last resort (Kaifeng) or random targets. Intense heavy flak over Yawata knocked down one B-29 with a direct hit and damaged eight. Fighter opposition was rated “moderate” but got three more Superforts, one with a combination of aerial bombing and gunfire and two in the first case of ramming experienced by the command. A Nick came in level from twelve o’clock, banked sharply, and drove its wing vertically into the outboard wing section of a B-29 flying wing position. Both planes disintegrated and the flying debris caught the No. 4 plane in that formation and sent it spinning down. Observers were uncertain but thought the ramming intentional. B-29 gunners claimed seventeen kills, thirteen probables, and twelve damaged (17/13/12).108
That night ten more B-29’s got over Yawata to drop an additional fifteen tons of bombs without being harmed by the enemy. Strike photos taken by the daylight formations seemed to show hits on two coke ovens, but according to Japanese records the damage was not serious.109 Losses, on the other hand, were heavy: besides the four B-29’s destroyed over Yawata, ten were lost to other causes and ninety-five airmen were dead or missing. Later it was learned that one crew had bailed out east of Khabarovsk. The U.S. embassy at Moscow had reported an earlier instance in which a wounded Superfort had been forced down near Vladivostok by Soviet fighters, and by the end of 1944 two others (including the much-publicized Gen. H. H. Arnold) had landed at that city. The Soviets, at peace with Japan, followed international law in interning the crews and planes,
but their subsequent conduct was not wholly consistent: they allowed the flyers to “escape” via Tehran, but kept the B-29’s, which after the war were to serve as models for a Red Air Force bomber, usually identified as the Tu-70.110
After the hasty reappraisal of MATTERHORN target objectives provoked by Chennault’s proposals, Saunders continued his plan for a return to Anshan.111 Scheduled for 30 August, D-day was twice postponed, once because of Chennault’s concern for possible Japanese attacks on Chengtu bases, once because of weather; final choice was 8 September.112 Plans for the strike were approved by Washington on 29 August.113 On that day Maj. Gen. Curtis E. LeMay assumed command of XX Bomber Command.114 LeMay, who had had an imposing record as a heavy bombardment officer in the ETO, had been slated for a B-29 job earlier, but had stayed on in Washington to work on the long-heralded reorganization of the command.115 His arrival did nothing to change plans for the Anshan show.
Saunders, in answer to Washington’s needling about his aircraft-over-target rate had announced a policy of dispatching every B-29 fit to fly.116 That policy, and improved maintenance, were reflected in Anshan II. By 8 September, 115 B-29’s had gathered in the forward area and 108 got off the runways. Of these, 95 reached Anshan to find good weather, 90 dropped 206.5 tons of GP’s at the Showa works, 3 bombed other installations, 5 hit at the Sinsiang Railroad Yards, and 3 others at various targets of opportunity.117 Enemy flak over Anshan was ineffective and Japanese fighters less aggressive than those encountered over Kyushu. Total losses for the mission were four: a crack-up near Dudhkundi on the way up; two forced landings in China, one destroyed on the ground by enemy planes and one partly salvaged; and a plane listed as missing. The crew of this last plane later walked out with the loss of only one man. In return, B-29 crews, their central fire-control (CFC) system working smoothly, claimed 8/9/10.118
A B-29 photo plane got some excellent shots on D plus 1 which showed a considerable amount of damage to the Showa plant. Of the sixteen batteries of coke ovens, three were thought to have been decommissioned for a year (one on 29 July, two on 8 September), and three for six months. Others would be hurt by damage to qzqz rdated installations, and the by-products plant had been hit hard. All in all,
command intelligence officers computed that the two attacks had cost Showa 35.2 per cent of its coking output, which would depreciate total Japanese rolled steel production by 9.3 per cent.119
Whatever the eventual results of the mission, its most immediate effect was to bring out the Japs in their first counterattack on the command’s forward bases. Both in Washington and the CBI there had been anxiety since the initial planning days that the enemy might hit those bases while they were crowded with B-29’s. The 312th Fighter Wing with its two groups (33rd and 81st) seemed capable of turning back any daylight raid but as yet had no night fighters. Shortly after midnight following the Anshan attack, Japanese bombers came over Hsinching and attacked Forward Echelon Headquarters, storage areas, and the parked B-29’s. Aided apparently by ground signals, the intruders made four runs, dropping frags and HE bombs to inflict minor damage on one Superfort and a C-46 and to wound two soldiers. No contacts were made by U.S. fighters.120
General LeMay had gone along on Anshan II, an interested observer of the crews and planes he commanded and of the enemy he faced. If what he saw was encouraging, it did not deflect him from his stated purpose of revising XX Bomber Command’s tactical doctrines and of instituting a thorough training program. Specifically, he intended to substitute for the current four-plane diamond formation a twelve-plane formation similar to one he had used with his heavies in the ETO. He proposed further to follow Eighth Air Force practice by subordinating night missions, so far numbering four of the command’s eight strikes, to daylight precision attacks. This would not mean the abandonment of radar bombing, so vital in variable weather. LeMay’s doctrine called for “synchronous bombing” in which both the bombardier and radar operator followed the bomb run in, with visibility determining who would control the plane during the crucial seconds before release.121
Precision bombing required training more sustained than the sporadic sessions which the command’s crews had undergone, and fortunately new arrangements for nourishing strikes out of China would release B-29’s and their crews from much of the Hump transport duty which had handicapped training. On 5 September LeMay had ordered each group to select six lead crews (later increased to eight) upon which other crews in a formation would drop. A week later a school was set up at Dudhkundi, occupied since early July by the 444th
Group, with instructors drawn from the command’s staff and from specialists brought out to the theater on TDY. Ground training and a simulated mission and critique on each of ten successive days made the eleven-day course at “Dudhkundi Tech” both strenuous and valuable. Meanwhile, the other crews of the four combat groups had been working with the twelve-plane formation and had made some progress when training was interrupted for the ninth mission.122
When LeMay took over at Kharagpur, the weight and target of that mission had not been determined. By the time Anshan II was run, he had decided that he could make only one other major strike in September, between the 25th and 27th. Headquarters, Twentieth Air Force, was engaged in revising target priorities and, anticipating an early report from the COA, asked LeMay to consider shifting from coke ovens to aircraft factories. Two days later, on 15 September, Washington temporarily tabled this suggestion by forwarding to Kharagpur a COA recommendation that the next mission be directed against coke ovens at Anshan and Penchihu.123 LeMay elected to finish off Anshan, still an important target and one well suited to an economical trial of the new bombing techniques.
LeMay had promised a 100-plane strike. On D-day, 26 September, he had 117 B-29’s forward, plus 1 photo plane, and counting a few late stragglers, 109 bombers were airborne that morning. Both figures represented improvements over Anshan II; not so the mission as a whole. Weathermen had predicted no worse than 4/10 undercast over Anshan, but a cold front moved in and blanketed the target, making it difficult for the bombers to maintain the new formation. Eighty-eight planes got over Anshan but only seventy-three bombed the Showa works, all by radar. Subsequent photographic coverage indicated absolutely no new damage.124 Two planes bombed Dairen, four Sinsiang, and nine bombed various targets of opportunity.125 Enemy planes over Anshan were very active but ineffective; not a single B-29 was lost from any cause on the mission and this was some solace for the bootless strike.126
Even so the enemy had the last word: that night a few bombers swept into the Chengtu area to drop three strings of bombs and damage five Superforts, two of them seriously. The Chinese warning net had tracked the Japanese planes in from Hankow airfields and the 317th Fighter Control Squadron at Chengtu had ample time to alert command personnel. But the one P-47 up could not make contact.127
The 312th Wing had suffered with other China-based units from lack of supplies, and in the interest of economy of fuel one of its P-47 groups had been exchanged for the 311th Fighter Group, equipped with P-51B’s. Chennault, reluctant to tie down two full groups for the static defense of Chengtu, had disposed part of the wing forward where the planes could take a more active part in the war, and events were to prove that this policy constituted no serious danger to the B-29 fields. The night raids of 8/9 and 26/27 September set the pattern for later Japanese raids, which usually followed B-29 missions. To guard against such sneak tactics, LeMay pressed for night fighters and 40-mm. AA guns. On 6 October (the eve of a third enemy raid), the first P-61’s of the 426th Night Fighter Squadron came to Chengtu, but it was mid-November before the 843rd AAA Battalion arrived to round out an integrated defense force. Japanese attacks were to continue until 19 December, but on the same light scale: in ten raids only forty-three enemy planes were counted and the damage done was more annoying than serious.128
Anshan III marked the end of the first phase of MATTERHORN operations. That fact is clearer in retrospect than it was in late September 1944, but even then there were indications of impending changes: a reorganization of the command, an improved logistics system, a shift in target priorities, and a closer coordination with operations in the Pacific. Had the command paused to take stock, it could have found little gratification in its strategic accomplishments. According to schedules hopefully concocted in Washington in April, coke and steel targets should have been destroyed. Only two missions had been really successful, Anshan I and II, and there had been no important dislocation of the Japanese steel industry. Yet, the command had learned much, as the operational record (as opposed to strategic results) of the later missions had shown. The shakedown was over, and with a revamped organization XX Bomber Command would in succeeding months more nearly approximate in the weight of its strikes the expectations of the original MATTERHORN planners.
Reorganization
From its inception the B-29 project had been an experimental organization, and this characteristic XX Bomber Command inherited along with part of the project’s personnel. AAF Headquarters looked on the command as a prototype for the XXI, XXII, and subsequent
VHB commands. At Salina and Kharagpur this attitude was as deeply ingrained as at Washington; XX Bomber Command’s historian reflected a widely held view when he referred to the command as “a great combat testing laboratory.”129 It was a crude sort of lab, where trial-and-error methods were more common than the scientist’s unhurried precision, but the essential spirit of testing results by such measurements as were available was not lacking. Combat testing involved two closely related problems, one tactical and one administrative. A new weapon, the B-29, had to be proved and modified, and a tactical doctrine formulated and refined; something of the command’s efforts in these respects has been told in the story of the first nine missions. Less spectacular but hardly less important to the success of combat operations were parallel efforts to perfect the administrative structure which supported the B-29 strikes. This process was a continuing one, but three distinct stages may be noted: the establishment of the 58th Wing in June 1943, the organization of XX Bomber Command in November, and a thorough reorganization which dragged through the following summer, At the end of September 1944, the date chosen to mark the end of the first phase of operations, that reorganization had been substantially effected, though another month or so was required to complete the task.
When a staff officers in Washington Salina had worked out the command’s original structure in the autumn of 1943, they had no exact precedents to follow. They starred with T/O’s borrowed from heavy bomber units and tried to expand them to fit anticipated needs of the B-29. In practice, these T/O’s proved less than perfect, in some particulars overmanned, in others undermanned. Remedial action was slow; here the fault seems more often attributable to AAF Headquarters than to the command. Inadequacies in maintenance and service units in particular were responsible for delays in combat operations. The need to step up the weight of attack spurred Washington to a belated correction of deficiencies which had been recognized in the Salina training period and had become more painfully obvious in India.
Two ideas had profoundly affected the original structure of XX Bomber Command: 1) it should be organized as a self-contained, independent force of great striking power, mobility, and flexibility, more akin to an overseas air force than to a conventional bomber command; and 2) it was to operate in a “primitive” theater.* These guiding
* See above, pp. 41 ff.
ideas had made for important differences in the command headquarters, in the internal structure of the bombardment groups, and in the functions and composition of the service groups. In each case, however, modifications made by the end of October 1944 tended to revert to more normal patterns. Changes in plans for deployment, which limited the command to a single wing and designated subsequent VHB units for the Marianas, made some consolidation of wing and command headquarters seem logical; here reorganization was a relatively simple task of compression. But in respect to the bombardment and service groups, where the command sought inflation rather than deflation, the increasing shortage of manpower in the United States imposed a formidable barrier. Reorganization was not accomplished in one sweeping act, so that it becomes convenient to treat separately of the service groups, the bombardment groups, and command headquarters, in that order, and of the transport system which had important bearings on all phases of the command’s administration.
In respect to its flight echelon, the 58th Wing was normal enough, with four bombardment groups of four squadrons each. Seven B-29’s per squadron gave the wing 112 UE aircraft and 50 per cent reserves brought the total up to 168 planes. Double crews and a crew of eleven, including five officers, made for a large number of rated officers, but this was a difference of degree, not of kind, from the standard B-17 or B-24 unit. It was in the maintenance and service elements that the innovations and the grief appeared.
Conventionally, first and second echelon maintenance, service, and supply were performed by the ground personnel of bombardment squadrons, third and fourth by service groups assigned to an air service rather than bomber command. In the interest of flexibility, mobility, and independence of theater support, this pattern had been discarded in XX Bomber Command. Each crewman was trained in some specialty other than his primary job in order to provide some constant maintenance of the plane. The ground personnel were separated from the flight echelon of the bombardment squadrons and formed into maintenance squadrons (sixteen in all) which could be moved about as needed to work on any B-29’s in the wing. For more advanced service, maintenance, and supply, there was to be for each VHB group a service group (special), presumably to be attached to the command rather than to a theater air service command; in addition
to its normal functions, this unit would perform housekeeping duties for the B-29 base.
At Salina there was no questioning of the basic concept of these experimental units, though experience had soon indicated that the composition of the maintenance squadrons would have to be modified. Two regular service groups, the 25th and 28th, were assigned to the command but were shipped out before they could be reorganized according to the new T/O’s. The tables had been based on tests made at the AAF Tactical Center at Orlando in February 1944; tentative copies sent to Salina for comments had elicited some suggestions for change, but Washington did little in the way of revision. It was expected that the two service groups would be split into four service groups (special), using as fillers personnel released by the contemplated reorganization of the maintenance squadrons.130
The 25th and 28th Service Groups, shipped early to set up house for the combat groups, were delayed en route by engine trouble on their transport and arrived in Bengal in May, six weeks late,131 only to be assigned to Stratemeyer’s ASC. This was not according to Arnold’s plan, and an appeal from Wolfe brought an answer specifically delegating to XX Bomber Command responsibility for third echelon service, maintenance, and supply, and inferentially control of the service groups. This policy was later described in detail in Stilwell’s GO No. 55, 7 June 1944, which gave the theater ASC responsibility only for fourth and fifth echelon functions.132
The T/O approved for the service group (special) on 15 April called for an organization of 710 officers and men.133 To create four such units from the 25th and 28th groups would require additional personnel. A delay in authorization for reorganization of the maintenance squadrons blocked that expected source of manpower, and on 9 May Wolfe asked Washington for some 97 officers and 453 men and a directive.134 AAF Headquarters refused to supply the personnel, save for central fire control and radar specialists, and funneled the directive through theater headquarters where it was delayed for weeks.135 Final authority for the changes was given by Stratemeyer on 23 June; Wolfe’s order went out on the 30th and was rapidly accomplished.136
Four service groups (special) were formed (25th, 28th, 80th, 87th), one for each VHB base. The internal structure of the group was
streamlined by regrouping the dozen or so existing units into three flexible squadrons – headquarters and base service, engineer, and materiel. No doubt the consolidation of units squeezed out fat which could be spared, but the bomber command considered that the new group was inadequate to its task of administration, if not of maintenance and supply. Actually, its tasks combined those of a service and of an air base group, including administration, mess, personnel classification, PX, special services, chaplains, etc.137 To man the new units from bulk allotments without specially requisitioned fillers required much ingenious juggling of personnel, but by reassignment within the command, reclassification, detached service, in-service training, and exchange with AAF IBS, the new tables were filled and the Bengal bases settled into a more orderly life. To provide for the four advanced bases, Washington allocated to each a bulk assignment of 150 officers and men.138
The reorganization of the bombardment groups went more slowly, though the composition of the maintenance squadrons had been challenged before they left Salina.139 The number of specialists in various MOS categories had proved unequal to the task of maintaining a complex and untried bomber; shortages of maintenance personnel had impeded the flight training program with an excessive planes-out-of-commission rate and remedial action was recognized as a “must.” Rather than interrupt the overseas movement, the command shipped out with the 16 maintenance squadrons unchanged but with a promise from Washington that a new T/O and about 550 additional personnel would be supplied when the command reached India. Men rendered surplus by the changes would supposedly be used as fillers in the new service groups.140
In India, command personnel officers found that though they had enough men to effect the desired augmentation of maintenance squadrons, they were still short in certain MOS categories, notably in mechanics for the temperamental R-3350. Washington, requested to fill the vacancies, replied that a radical revision of the VHB group was slated for July and that in the meanwhile alleviation would be provided only in the case of a few specialties, not including power plant mechanics.141 The “radical revision,” calling for a group of three VHB squadrons with ten UE aircraft each and for a merging of the maintenance and bombardment squadrons, had been approved by AAF Headquarters in a new T/O&E dated 17 April. It was the end of June
before Wolfe received a copy of these tables and even then he had no directive to adopt them. The bomber command had to struggle along with maintenance squadrons rendered “lame duck” by impending changes, their mechanics overworked as India’s hottest season played havoc with the R-3350 engine.142
The new T/O&E provided for thirty aircraft for three squadrons instead of twenty-eight for four, and would thus increase striking power and cut down on overhead personnel; here there could be no justifiable complaint. But the new squadron of 615 officers and men had a ground echelon of only 349 as opposed to the maintenance squadron’s 390, long recognized as inadequate. Group commanders, asked for comments, uniformly recommended substantially larger ground echelons. Some alleviation was promised in a change in the tables approved on 3 August, which authorized a squadron of 644 with a gain of nearly 30 ground personnel.143
The long delay in agreeing on and instituting the projected general reorganization stemmed in part from the distance between Washington and Kharagpur and from changes in the command’s leadership. The outlines of the reorganization had been worked out during Wolfe’s incumbency, and when he returned to Washington early in July, he was able to present the field point of view and to report back to Saunders the improvements contemplated. Later in the month Arnold sent his chief of staff, General Giles, to India to make detailed arrangements, and by 1 August an acceptable plan had been drawn up. In return, two of Saunders’ personnel officers went on TDY to Washington, but final action was still held up pending Saunders’ relief by General LeMay.144 LeMay arrived on 29 August but it took several radios to pry a final commitment out of AAF headquarters.145 On 20 September he was directed to effect the following changes: reorganization of the headquarters of the four VHB groups (90th, 444th, 462nd, 468th) according to the T/O&E of 29 June 1944 with certain augmentations; the disbanding of the sixteen maintenance squadrons and of four VHB squadrons (395th, 679th, 771st, and 795th); and the reorganization of the other twelve VHB squadrons according to the tables described above, as amended on 3 August.146
The changes were put through as rapidly as possible, not without some feeling that the command had too few maintenance personnel. The return to the standard-type bombardment squadron marked an abandonment of the concept of “flexible” maintenance now rendered
obsolete for the VHB program in general by the elaborate permanent installations being built in the Marianas. The new tables did away also with the doubly trained crews, capable of performing first and second echelon maintenance as well as fighting their plane. That training had shown up well in the Palembang mission and in cases where planes made emergency landings in China, but the extra cost in training (for example, forty weeks for an electrical specialist-gunner as opposed to five weeks for a gunner) was too great for mass production of VHB crews.147
To make most efficient use of the streamlined units LeMay issued on 10 October a new formula for the operation of the four VHB bases in India,148 a formula which owed much to his previous experience in the ETO. Each of the fields housed a bombardment group, a service group, a weather detachment, and an AACS detachment. The VHB group commander was in charge of the base, with the air executive, normally the deputy of the group, as second in command. The ground executive was usually the service group commander. Personnel on the base were regrouped functionally without loss of unit integrity: for example, the ground echelons of VHB squadrons and the engineer squadrons of the service groups were integrated to perform maintenance and service for B-29’s assigned to the base. Designed to spread the work more evenly, this system worked well; its success was indicated in the increasing weight of attacks and the decreasing rate of planes out of commission.149
Meanwhile, command headquarters had undergone extensive changes in structure and personnel. When Saunders succeeded Wolfe as commanding general of XX Bomber Command, no replacement was provided for the former’s previous job as commander of the 58th Wing. With Washington’s consent, Saunders on 13 July amalgamated the two headquarters, attaching personnel from the 58th to appropriate sections of the command’s staff. This marked the de facto, though not the formal demise of the 58th Wing.150 By a directive of 6 August, XX Bomber Command’s headquarters experienced a “functional realignment” to conform more closely with the current pattern followed in the conventional air force. This was in recognition of the fact that though possessing only four combat groups, the command performed some air force functions. The older staff system with four “A’s” reporting directly to a chief of staff was discarded; staff sections were regrouped under three deputy chiefs of staff (for administration;
operations; and maintenance, supply, and services), with some sections reporting directly to the chief of staff or commanding general.151 During the next two months authorized strength of the headquarters and headquarters squadron was set at 183 officers and 417 men. This showed a decrease of about 250 from the combined strength of the headquarters of XX Bomber Command and the 58th Wing (now formally deactivated), but it was still a numerous body, identical in size with that of the 3-wing XXI Bomber Command and justified only because of the XX’s ASC duties and of the necessity of maintaining a forward area headquarters.152
Changes in structure had been accompanied by a considerable turnover in personnel. The command had been fortunate in long maintaining a remarkable degree of homogeneity among its key personnel, but beginning in August many of Wolfe’s hand-picked officers went back to the States, presumably to be fed into new B-29 units, and their replacements took over. Saunders stayed on at Kharagpur for several weeks after LeMay’s arrival, though the elimination of the 58th Wing rendered him surplus. On 18 September Saunders was seriously injured in the crash of a B-25 during an administrative flight and was evacuated to the United States only after a slow recovery.153 Meanwhile, Hansell had left the Washington headquarters to take XXI Bomber Command out to Saipan and Brig. Gen. Lauris Norstad had succeeded him as chief of staff of the Twentieth Air Force. In spite of the numerous changes in organization which have been recorded above, there were throughout only minor fluctuations in the command’s total authorized and assigned strength, as the following figures show.154
| 30 June 1944 | 31 October 1944 | |||
| O | EM | O | EM | |
| Authorized strength | 2,214 | 12,798 | 2,193 | 12,940 |
| Assigned strength | 2,212 | 12,865 | 2,250 | 13,237 |
When LeMay informed Washington on 2 November that the reorganization of XX Bomber Command was “now practically complete,” he reported 75 officers and 484 men surplus and eligible for reassignment,155 but he was then negotiating for the exchange of surplus B-29 crews for a bulk allotment of about 900 men for air transport duty which would further modify the strength of his command.
The reorganization of XX Bomber Command, like its every other
activity, had been complicated by the necessity of operating a transport service to support strikes out of China. The original MATTERHORN logistics plan had failed, in part, as Wolfe insisted, because diversion of the 73rd Wing to Saipan had left him with too few B-29’s for self-support.156 Operations had been made possible only by the assignment to XX Bomber Command of three air transport squadrons (mobile) equipped with C-46’s. At the end of July, the situation was something like this: the command had turned over to the 312th Fighter Wing a monthly allocation of 1,500 tons, guaranteed by ATC in exchange for 18 C-87’s, and was quit of further responsibility to the wing; the command was operating both tactical and tanker B-29’s in transport flights direct from Kalaikunda to Chengtu; and it was operating 3 C-46 squadrons over the regular Hump route via Assam and Chengtu.” By these means the bomber command had hauled in July 2,978 tons, enough to support only 135 combat sorties according to current estimates of 22 tons per sortie. Deliveries could be stepped up as the command learned more about the transport business, but only by moderate degrees, and certainly not enough to allow the 225 sorties per month which the Twentieth Air Force had set as its minimum goal. In spite of the heavy investments already expended in the CBI, Arnold was considering the possibility of moving the B-29’s to another theater if the weight of attack could not be increased.157
Unable to find supplies for its regular monthly quota of missions, XX Bomber Command was committed also to a short-term campaign of air support for certain Pacific operations to be conducted by MacArthur and Nimitz in the autumn.† Stilwell, informed early in May of the support desired from the bomber command and the Fourteenth Air Force, had insisted he could build up the requisite stockpile in China only by increasing the lift potential of ATC’s India-China Division (ICD).‡.158 The emergency created by the Japanese push in east China accentuated the need for more Hump tonnage and the build-up of ICD became a matter of urgency. If accompanied by a firm guarantee of tonnage for regular missions and PAC-AID, as the support for the Pacific operations was coded, it would have solved the command’s supply problem.
† See below, pp. 275-88.
‡ On 1 August 1944 ATC’s India-China Wing became the India-China Division.
Neither Wolfe nor Saunders enjoyed the task of running a separate air transport line, and their control of the C-46 squadrons, originally a temporary expedient, inefficient, and in flat contradiction to the basic concept of ATC, had resulted from a conflict in principles of command. Stilwell, as theater commander, had the responsibility for allocating Hump tonnage, and the Joint Chiefs were reluctant to interfere by stipulating a flat guarantee for MATTERHORN. But Arnold, as commander of the Twentieth Air Force, feared that without a firm commitment XX Bomber Command would be squeezed by the more immediate tactical needs of the Fourteenth. In this dilemma, Arnold explored the two obvious possibilities, build-up of ICD with a firm monthly quota for MATTERHORN and the assignment of more transport planes to XX Bomber Command. It was only in November that a solution was reached.
The alternatives had been clearly delineated by Wolfe just before his recall to the States. Admonished sharply to improve his transport operations to allow for at least two maximum missions a month, Wolfe on 29 June listed three possible means of achieving his directive: by building the command’s strength to 180 B-29’s, 128 for combat, 52 for tankers; or by assigning 150 B-29’s plus 80 B-17 or B-24 tankers; or by having the JCS secure a firm allocation of 2,000 tons monthly from ATC.159 Under any plan, XX Bomber Command would retain its C-46’s. Back in Washington, Wolfe went over these proposals with LeMay and Arnold’s staff.160 The 180 B-29’s were already provided on paper in the imminent reorganization of the 58th Wing. On 7 July Arnold proposed to the JCS an increase of the Hump lift to 31,000 tons in December by adding to ICD’s resources;161 a week later he made the correlative recommendation that the Joint Chiefs guarantee to the bomber command enough of the increased tonnage to insure the designated 225 sorties per month.162 With the command’s own efforts, this amount was variously calculated at from 2,000 to 2,500 tons, exclusive of the needs of the 312th Fighter Wing.163 Both suggestions were referred to appropriate agencies for study, which meant no immediate action, but the matter of the B-17 or B-24 tankers could be handled within the AAF.
Saunders was informed of the proposals under consideration by the JCS and of the possibility that he might receive eighty B-17 tankers; Arnold wanted to know if he could furnish flight crews and maintenance for them.164 As to maintenance personnel, the command was
already shorthanded and eighty more transports would impose a heavy additional strain. Of flight personnel, however, there was no shortage. Original T/O’s had called for 2 combat crews per UE B-29, or 224 for the wing, but actually the 58th had come out with 240. Wolfe, and later Saunders, had complained of this plethora which made it impossible for all crews to get enough flying time to maintain efficiency and morale. Saunders wished to reduce the assignment to 1⅓ or 1½ crews per plane (160 or 180 for the wing under the new T/O) and send home the surplus crews.165 Arnold proposed to use excess crews to operate the additional tankers, which Saunders considered a waste of highly trained B-29 crews. He would have preferred to assign the tankers to ATC in return for a flat guarantee of perhaps 2,700 tons per month.166
General Giles went to the CBI in mid-July and there, at Arnold’s request, he held a conference of interested commanders to arrive at some agreement on the allocation of Hump tonnage for current needs and for PAC-AID. Calculating the total Hump lift with the expected augmentation of facilities at 21,320 tons, Giles on 2 August proposed that the JCS might allocate that sum as follows: to the Fourteenth Air Force, 10,000 tons; to other CBI agencies, 3,200; to XX Bomber Command, 6,300 firm including 1,550 tons for the 312th Wing; for PAC-AID stockpiling, 1,820. Of its 4,750 tons net, the bomber command would haul 1,600 tons in B-29’s and in 40 C-46’s; the other C-46’s (1 squadron of 20 planes) and the promised heavy bomber tankers would be operated by ATC.167 This plan was essentially what Saunders had earlier suggested but was not wholly acceptable to Arnold, who expected XX Bomber Command to operate the tankers (now designated as seventy C-109’s, B-24’s converted as tankers) and who repeated a sentiment recently expressed, that he would not tolerate the continued use of B-29’s as transports after the arrival of the C-109’s.168
Giles’ message seemed to indicate some willingness on Stilwell’s part to receive “additional guidance” from the JCS in regard to tonnage allocation. When asked by Marshall on 10 August, Stilwell thought it possible to maintain 6,300 tons for XX Bomber Command if ATC were expanded according to schedule; rather than a directive to that effect, however, he preferred that the JCS give him a statement of the relative importance of the several activities dependent upon the Hump lift.169 On 25 August the Joint Chiefs approved the increase in ICD’s potential as proposed by Arnold and provided a statement to guide
the allocation of tonnage.170 This went to Stilwell next day in a radiogram which suggested the following order of priorities: 1) maintaining the air link to China to insure operations and defense of bases needed for PAC-AID (supply of Fourteenth Air Force and stockpiling for PAC-AID); 2) implementing MATTERHORN at the rate of 225 sorties a month; and 3) requirements of Chinese air and ground forces.171 According to Arnold’s office, the intent of the JCS had been to insure for XX Bomber Command support for 225 sorties per month.172 The command, however, put little reliance in a directive that gave so low a priority to MATTERHORN strategic missions. When LeMay assumed command on 29 August, it seemed obvious that he would have to increase, not abandon, transport activities; there was no longer any thought of assigning the C-109’s to ATC, but rather of getting them to Kalaikunda and at work as soon as possible.173 A plan for operating the C-109’s had been approved on 25 August. A small cadre of administrative and maintenance personnel plus 244 enlisted trainees would be assigned to the “C-109 provisional unit” at Kalaikunda. The B-29 crews, less one bombardier-navigator and one radar operator each, to a total of seventy-two would be rotated on sixty-day TDY. To each C-109 would be assigned a flight crew of five and a ground crew of eight, drawn from the cadre mechanic trainees and surplus members of the B-29 crews.174 The provisional unit became a reality early in September when B-29 crews, ground personnel, and the first C-109’s arrived at Kalaikunda. Later, with the reorganization of the command, some 39 officers and 460 men were authorized for “cargo service units.”175
These arrangements did not bring immediate relief to the pressing need for more tonnage at Chengtu. Exclusive of requirements for fighter defense, the command needed some 4,950 tons for 225 missions, plus about 500 for overhead. According to their own figures,* XX Bomber Command in August had received at Chengtu 1,478 tons from ATC and had hauled 798 tons in C-46’s, 1,106 in B-29 tankers. At Arnold’s repeated insistence, the tactical B-29’s had been withdrawn from the wearing transport job, so that the over-all total was only 3,382 tons, nearly 600 short of the July haul. In September, with an early promise of 3,200 tons from ATC,176 the command received 2,141; with the tactical B-29’s back on the job and some help from
such C-109’s as arrived during the month, a total of 4,581 tons was brought in to the forward bases. Deliveries for October and November, when PAC-AID missions would be run, must be much heavier, but for those missions XX Bomber Command could draw upon Stilwell’s stockpile. LeMay’s estimate of 10,685 tons for October was almost exactly met with a total delivery of 10,830, of which 7,301 were by ATC planes.177 The C-109’s were slow in arriving – only thirty-three had been received by 5 October when deployment should have been almost completed – and in spite of valiant efforts at Kalaikunda it was still acknowledged by the bomber command that both C-109’s and C-46’s could be operated more efficiently by ICD.178
By mid-October, Arnold’s staff had come around to that point of view, concurring in a “feeler” sent out to Sultan by ATC at Washington to the effect that ICD take over on detached service one of the XX Bomber Command’s C-46 squadrons and twenty to thirty of the C-109’s. “We all agree,” LeMay was informed on the 17th, “that it would be desirable to get you out of the transport business but the main requirement is ... the insurance of ample tonnage.”179 Such an arrangement did not have to involve even temporary transfer of B-29 crews to ATC control, since the Twentieth Air Force had already proposed to cut down authorization of such crews from 2 to 1.3 per UE aircraft and to exchange for those rendered surplus a bulk allotment of 924 flight and maintenance personnel.180 Conferences in the CBI indicated that Stilwell was not averse to such an arrangement and that he would give a reasonably firm guarantee if the exchange were made.181 At LeMay’s suggestion, the details of the trade were modified to include all of the C-109’s and two squadrons of C-46’s, leaving him one squadron for hauling dry cargo. Permission to make the described reassignment was granted by the end of October, and the new arrangement was reflected in the transport records of November and December.182 B-29 tactical planes were taken off the Hump run in November, B-29 tankers in the following month, and deliveries were predominantly by courtesy of ICD. There would still be disagreement over each monthly allocation, but XX Bomber Command had been relieved of a task which had long absorbed much of its energies. The theory of the self-supporting bomber unit had been broken by the harsh realities of China–Burma–India.