Author: Hiroyuki Shindo

From February 1942 until July 1944, a war of attrition was fought by the air forces of the United States, Australia and Japan in Papua, New Guinea, the Bismarck Archipelago and the Solomon Islands. Although this period comprises more than half the length of the war in the Pacific, somehow more attention seems to be paid in popular histories to other aspects of that war, such as the actions of the carrier fleets of Japan and the United States. The air campaign in the South Pacific, however, was of extreme importance, not just to the persons of all sides who fought there, but to the outcome of the war. This is because it was in, over and around the island of New Guinea that the Japanese Army and Navy, and their air forces, were first stopped, worn down and finally pushed back. Since it is obviously impossible to look at every aspect of that campaign in a single article, the following discussion concentrates on the major strategic and operational issues on the Japanese side. It also focuses on Japanese air operations over the main island of New Guinea, even though in actuality these operations, when seen from the Japanese side, were closely intertwined with activities undertaken in the Solomon Islands.

Broadly speaking, the Japanese air campaign over New Guinea may be divided into four phases. The first phase was the Japanese Navy’s offensive campaign against Port Moresby, from the spring through fall of 1942. Next, from early 1943 until about June of that year, the Japanese Army filled in for the Navy (whose air forces were increasingly committed to the Solomons campaign) and fought a campaign which was intended to be offensive but became increasingly defensive in nature. The third phase was a short period in the summer of 1943 in which the Japanese Army assigned a more positive role to its air forces in New Guinea, only to see the bulk of that force destroyed in a single air attack. Finally, the Japanese Army air forces fought an unglamorous defensive campaign for approximately a year, from the summer of 1943, before the Japanese Army was pushed out of New Guinea and the war itself shifted to the Marianas and the Philippines.

Occupation of Rabaul and the start of air operations against Port Moresby

When the Japanese Army and Navy developed plans for a forthcoming war with the United States and Great Britain in 1941, the islands of the Bismarck Archipelago, particularly New Britain, were included in their target list from early on. This was because Japanese planners (especially those in the Navy) saw Rabaul on New Britain, with its excellent natural harbour, as a potential threat to Truk Island in the Carolines. Situated only 1100 kilometres from Rabaul, Truk was the site of the Japanese Navy’s most important base in the central Pacific Ocean. The seizure of New Britain was therefore necessary to protect the base at Truk, and the surrounding islands had to be controlled as well, in order to secure Rabaul. While the main efforts of the Army and Navy would be aimed at the southwest Pacific area – i.e. the Philippines, Malaya, Singapore, Netherlands East Indies and other surrounding areas that made up the so-called “Southern Resources Area” – both the Army and Navy therefore assigned units to operations in the South Pacific.

In the plans which received imperial sanction on 5 November 1941, the Army’s South Seas Detachment was assigned the task of seizing first Guam, then airfields in the Bismarck Islands, in order to eliminate the threat posed by the enemy. Since the operation in the South Pacific was seen as the Navy’s responsibility, however, the Navy’s Special Landing Forces were to take over the occupation of Guam at an appropriate time, after which they would cooperate with the Army in occupying Rabaul. It was further specified that the South Seas Detachment would be replaced by the Special Landing Forces and withdrawn to the Palau Islands as soon as possible. The Navy’s Fourth Fleet, based in Truk, would be responsible for supporting both the Guam and the Rabaul-Bismarcks operations. A sea-borne assault on Rabaul was accordingly made on 22 January 1942 by the Special Landing Forces, with the support of both land-based and carrier-based aircraft, and the town quickly captured.

While these operations were being carried out, the Japanese Army and Navy, faced with the unexpectedly rapid success of their operations in the Philippines, Malaya and Netherlands East Indies, had to decide upon their next steps. Briefly speaking, Japan’s options included continuing westward into India; invading the Australian mainland; invading New Guinea, the Solomons, Fiji and Samoa, in order to cut off Australia from the United States; and driving eastward towards Hawaii. The Navy High Command wanted to invade Australia, in order to eliminate it as potential springboard for a counter-offensive by the Allies, but the Army balked at this because it would require an excessive commitment of manpower. The Navy therefore settled for the lesser option. The Fourth Fleet, with the XI Air Fleet (the Navy’s land-based air force in the Pacific theatre), was ordered to assault Lae, Salamaua in New Guinea, Port Moresby in Papua, and the Solomon Islands.

In compliance with this strategy, Zero fighters of the Chitose Air Corps moved to Rabaul on 31 January. Shortly afterwards, on 2 and 5 February, Kawanishi Type 97 “Mavis” flying boats of the Yokohama Air Corps bombed Port Moresby for the first time, and the air war over New Guinea was underway. On 9 February, Gasmata (on New Britain’s southern coast) was occupied, and work begun on an airstrip. To carry out further operations, the 4th Air Corps was newly created, with a nominal strength of 27 fighters and 27 bombers. It was placed under the command of the 25th Air Flotilla, and its headquarters was located in Rabaul. On 24 February, aircraft from the 4th Air Corps began bombing Port Moresby.

On 7 March 1942, the Japanese High Command decided upon the operations which would follow the so-called “First Stage Operations,” which had been aimed at the occupation of the Netherlands East Indies and other areas of the Southern Resources Area. The second stage of operations which was thus adopted called for the continuation of major offensive operations in order to secure a long-term, unbeatable politico-strategic situation, after which additional active measures would be taken aimed at forcing Great Britain to capitulate and making the United States lose the will to fight.1 As part of this new strategy, the decision was made to continue the advance in the Solomons and New Guinea area, with the aim of eventually cutting off the supply route between the United States and Australia. The 7 March decision therefore confirmed what was already being executed. Lae and Salamaua on the northeastern New Guinea coast were occupied on 8 March. Two days later, the Tainan Air Corps (one of the fighter units deployed to Rabaul as part of the new strategy) sent eleven of its Zero fighters to Lae, which became an exceedingly busy advanced airbase.2

Until the end of July 1942, the naval air units based at Rabaul and Lae became intensely involved in flying missions over the Owen Stanley Range to attack Port Moresby, or other Allied bases on the New Guinea mainland. Such operations consisted of either bombing missions with fighter escort, or sweeps by fighters alone. The Japanese fighter units at this time were also kept extremely busy intercepting Allied air attacks on the Japanese bases. This phase of the air war was characterized by the lack of clear superiority by either side. Although the Australians and Americans often lost more aircraft in individual air battles, Allied air strength did not diminish significantly. On the other hand, the Japanese, although suffering fewer losses, saw a slow decline in the quality of their forces as highly-trained and experienced pilots were lost and replaced by less and less experienced ones. This period was, therefore, somewhat of a stalemate, as the Japanese could not batter the Allied air forces enough to drive them out of New Guinea.

The commitment of Army air forces to the South Pacific

The next stage in Japanese air operations over New Guinea involved the deployment of Japanese Army air forces in the region. After the Americans landed on Guadalcanal Island in the Solomons on 7 August 1942, the Japanese air forces based in Rabaul became involved in a two-front operation. Now they were forced to make increasingly greater efforts in the Solomons, while continuing their campaign against New Guinea (which still principally meant Port Moresby). The air battle in the Solomons would be fought principally by naval aircraft, and, as this commitment grew, the Army’s air forces would play a greater role over New Guinea.

The American counter-attack at Guadalcanal came as a complete surprise to the Japanese. In their estimate of the world strategic situation in March 1942, the Japanese High Command had concluded that the Americans would not mount a major counter-attack before 1943. On 12 August the Army and Navy High Commands agreed that the American landing on Guadalcanal was only a local tactical move, and the island could be easily recaptured. This view began to change after the Battle of the Tenaru River on 21 August, when the Japanese Army’s Ichiki Detachment, numbering some 900 men, made the first attempt to retake Henderson Field and was almost completely wiped out. The Japanese Navy High Command now realized that it would take more than a simple operation to retake Guadalcanal, and that controlling the air around the island was a prerequisite for success. The Navy felt, however, that it alone was unable to make the necessary commitment in terms of air forces, due to the depletion of its air units in Rabaul. In late August it therefore asked the Army to send some of its forces to reinforce the air effort in the Solomons and New Guinea area.3

The Army considered the Navy’s request, and almost immediately rejected it. First of all, the Army was not inclined to make a major commitment in the South Pacific area, because it still felt that its traditional area of responsibility was the Asian mainland, while the Pacific Ocean area was the traditional responsibility of the Navy. This had been the understanding since the establishment of the Japanese Army and Navy in the late 19th Century, and while there was no formal written agreement to this effect, the Army and Navy’s doctrines, training, tactics, strategy and equipment were all based upon it. Over the years the Army’s air units had been prepared almost solely for fighting a war with Japan’s traditional enemy on the continent, Russia. The Army had never even considered the possibility of conducting air operations in the New Guinea-Solomons area, and recognized that it knew almost nothing about the geographic, climatic and other conditions of the South Pacific. Not only was the Army extremely reluctant to commit its air forces in such circumstances, but it suspected that the Navy actually did have additional air forces in quieter areas of the Pacific which it could commit to the South Pacific.

The Army formally replied to the Navy in late August. Besides the reasons given above, the Army explained that its air forces were spread out all over Manchuria, China, Burma and Sumatra, and in addition had to defend the Japanese homeland from possible threats from the North Pacific area. There were, it claimed, no personnel or aircraft that could be spared for the South Pacific. The Army also pointed out that its air forces had been trained and prepared for mainland operations, and were ill-suited for conducting operations or deployments over large expanses of water, which would be the case if they were committed to this new theatre. Furthermore, it was mentioned that there were no overland air routes to the South Pacific front which Army air units might follow, and that airfields and other necessary facilities were too unprepared there for them to operate properly. Thus, no agreement was reached at this stage concerning the deployment of Army air forces to the South Pacific.

In mid-September a further attempt to retake Henderson Field which the Army mounted, using the Kawaguchi Detachment, also failed. As preparations commenced for yet another attempt involving the 2nd Division, the need for regaining command of the air around the Solomons was keenly recognized by Army planners. Gradually, opinion within the Army High Command began tilting towards the deployment of Army air forces to the region. In late October a proposal was produced by the Operations Section of the Army General Staff for the assignment of two fighter and two light bomber sentais (air groups) to support Army operations in New Guinea. This, however, was only intended as a temporary measure to help out the Navy while its air units were committed to the Guadalcanal campaign. (In fact, by then the Navy’s major bombing missions against Port Moresby from Rabaul had all but ceased, except for night raids by one or a few planes at a time, as more and more Navy planes were sucked into the Guadalcanal campaign.) As soon as Port Moresby was captured, it was intended that the Army air forces would be pulled out again.

The issue was settled shortly afterwards. From 28 October, Takushiro Hattori of the Operations Section, Army General Staff, visited the New Guinea-Solomons front, including XVII Army headquarters on Guadalcanal. His report dated 11 November called for, among other things, the deployment of army air forces to the region in order to regain air superiority. In mid-November, the Army and Navy High Commands and the Government produced an estimate of future American air power which projected that the Americans would have in the South Pacific some 11,000 army and navy first-line planes by December 1942, and 24,500 by December 1943. It was recognized that the most urgent need facing Japan was to increase her air power.

Faced with looming defeat on Guadalcanal, and with setbacks in their drive on Port Moresby where Japanese forces had begun withdrawing from Kokoda on 26 October, the Army finally decided to commit some of its air forces to the South Pacific. On 18 November an “Army-Navy Central Agreement on Operations in the South Pacific Area” was signed, and the 6th Air Division was committed to the New Guinea front. This, it should be noted, was only a temporary arrangement to support the Navy’s operations – that is, the Army units would be “loaned” until the Navy’s air forces recovered or certain victorious conditions (such as the occupation of Port Moresby) were attained. In accordance with the agreement, sixty Nakajima Type 1 “Oscar” fighters of the 11th Sentai reached Rabaul via Truk on 18 December, and almost immediately became involved in air defence operations. By the end of the month they were flying missions against targets on mainland New Guinea, such as Buna, some of these as joint missions with Navy aircraft. On 29 December, heavy bomber units of the Army were ordered to deploy from Burma to New Guinea.

Six weeks later, on 3 January 1943, another Army-Navy Central Agreement was signed which designated the areas of responsibility of the Army and Navy’s air units. The Army air forces were given the mission of supporting the ground forces on New Guinea and providing their air defence, and supporting the transport of supplies to New Guinea. The Navy air forces would be responsible for air operations in the Solomons, and for air operations in New Guinea other than those assigned to the Army.4

While the decision to withdraw from Guadalcanal was made at the end of 1942, on 4 January the Army and Navy High Commands ordered that operations on New Guinea were to be continued. The purpose of Japan’s operations in the South Pacific was to be to “secure a position of superiority”. Lae, Salamaua, Wewak and Madang on New Guinea were to be strengthened or occupied, and the area north of the Owen Stanleys was to be secured so that it could function as a base for operations aimed at Port Moresby.5 The Japanese then had 164 Army and 190 Navy aircraft on their bases at Rabaul and the surrounding area. At this stage, therefore, the Japanese had decided to abandon the southern Solomons but still intended to continue offensive operations in New Guinea, aimed at the eventual capture of Port Moresby.

Thereafter, Japanese air operations over New Guinea were conducted principally by the Army, operating out of Wewak and other bases. In addition to bombing raids on Allied airbases, the Japanese air forces also had to take on ground support missions (although never conducted as closely as was the case with the Allied air and ground forces), combat air patrol over their own airbases and escort missions of ship transports operating around New Guinea. As was the case with the Japanese Navy aircraft operating against Port Moresby in mid-1942, however, the Japanese never really gained the initiative in the air, and instead were gradually pressed into the defensive.

The Japanese were acutely aware of the gradually growing strength of the Americans during this period. At the end of January 1943, the Japanese Army estimated that American strength in the New Guinea area was 300 aircraft – the same as the combined strength of the Japanese Army and Navy. They estimated, however, that the Americans were producing 4,000 aircraft per month, of which the South Pacific would receive 480 (and the New Guinea front at least 80) aircraft per month. Judging that the Americans would suffer losses of 50 per cent per month, the Japanese Army projected that the Americans would have 700 aircraft in the New Guinea area by June, and 950 by September. Compared to this, the combined strength of the Japanese Army and Navy in the New Guinea-Solomons area was projected to grow to possibly 350 in June, but was not expected to exceed that number at any time.6 The Japanese were therefore clearly aware that if the current situation continued, the Americans would gradually attain a sheer superiority in numbers through 1943.

After the Japanese successfully withdrew from Guadalcanal in early February 1943, the next major setback which befell them was the Battle of the Bismarck Sea on 2-3 March that year. In this action, an entire convoy of transports carrying the 51st Division and a large amount of heavy weapons and supplies earmarked for the reinforcement of Lae, was sunk in the Dampier Straits, along with four out of eight escorting destroyers. As a result of this disastrous defeat, which occurred even though the Japanese had provided the convoy with what they believed was an adequate fighter escort, the Japanese were forced to realize that they could no longer run major convoys of transports in areas where the Americans were able to operate their bombers more or less freely. This posed obvious problems for the supply of the Japanese troops on New Guinea.

Following the Bismarck Sea battle, the Army reconsidered its entire New Guinea strategy, including the possibility of abandoning New Guinea outright and withdrawing to a new defensive line. This option was not adopted, however, because the Navy’s operations – not only in the New Guinea area but also throughout the South Pacific – would be constricted by such a pullback by the Army. In addition, no new defensive line to which the Army might safely fall back was prepared, and in any case, such a retreat would only expose the Philippines, Celebes and other Japanese-held islands in the South Pacific to air attack. Joint Army-Navy studies which were held on 14 March confirmed this policy, and the Army continued to try to hold its current positions in New Guinea.7



Further commitment and destruction of Army air forces

The lopsided defeat in the Bismarck Sea was an enormous shock to the Army High Command, which was forced to recognize even more the importance of the New Guinea front to Japan’s war efforts, and that the situation there was extremely critical. As a result, the Army had to rethink its strategy for the Pacific War, and decided to make a further substantial commitment to the area, in both ground and air forces. On 25 March 1943, a revised Army-Navy Central Agreement was concluded between the Army and Navy High Commands.8 Regarding the South Pacific, it was agreed that the main weight of operations would be placed on the New Guinea front, i.e. the defence of Lae and Salamaua and the strengthening of bases along the northern coast, while a holding or delaying operation would be fought in the Solomons. The assault on Port Moresby, while officially still a long-term objective, was for all practical purposes abandoned, although lip service was still paid to it to avoid damaging the morale of the troops fighting in New Guinea.

Reflecting this revised strategy, the Army decided to strengthen its air forces in the New Guinea area. Earlier, on 18 March, the Army High Command had decided to strengthen its air forces in the South Pacific generally, starting with the replacement of the fighter units in the 6th Air Division. The newly assigned units were the 68th and 78th Sentais of the 14th Air Brigade, which flew the Kawasaki Type 3 “Tony” fighter. These units started arriving in Rabaul in late April, and were duly deployed to New Guinea. On 2 April the Army further decided to dispatch to the New Guinea front, among other units, the 13th Sentai (flying the Kawasaki Type 2 “Toryu” twin-engine fighter) and the 24th Sentai (flying the venerable “Oscar”). These units began arriving in Rabaul in the latter half of May.

The situation changed further when the Japanese discovered that the Allies were constructing airfields in the New Guinea highlands, at Mount Hagen and Bena Bena, which would threaten the Japanese airfields at Madang and Wewak. Ground operations to meet these new threats were immediately planned. As a further measure it was decided to send to New Guinea the 7th Air Division, which was a relatively new formation, having been formed in late January 1943. It had been previously assigned to bases in eastern Java, Timor and other islands in the East Indies, and, along with the defence of this area, had been scheduled to carry out air raids on the Port Darwin area. In order to meet the extremely pressing situation in New Guinea, however, it was ordered to deploy to Wewak on 19 June.

Thus, the Japanese Army was busily deploying new air units to the New Guinea front in the spring and summer of 1943. While the actual movement of these forces was delayed due to the insufficient readiness of bases in the New Guinea, among other reasons, Army air strength – at least on paper – was steadily reinforced during this period. The Army made another organizational change during this period, when it created the Fourth Air Army to exercise overall command over the 6th and 7th Air Divisions. The new air army was formed in mid-June, and its headquarters began deploying to Rabaul from 25 July through 10 August.

These moves were important because they indicated a shift in the Army’s stance concerning its air forces in the New Guinea area. Until then, the Army had considered its air forces in the South Pacific as a temporary measure, a backstop for the Navy’s efforts in the region. Over the spring and summer of 1943, however, the Army came to recognize fully that the defence of northern New Guinea against the Allied advance was its responsibility, and began to deploy additional air forces from Manchuria and elsewhere to the New Guinea front. Eventually about a quarter of the Army’s air forces would be committed to the South Pacific. Considering that most of these units were the best that the Army had, and would be subjected to losses of approximately 50 per cent per month, this was indeed a major commitment by the Army.9

As mentioned above, the 7th Air Division was ordered to deploy to New Guinea on 19 June 1943. The major 7th Air Division units which were transferred at this time were the 59th Sentai (fighters), 5th Sentai, 7th Sentai (heavy bombers) and 61st Sentai (heavy bombers). While the Eighth Area Army, which commanded the Army ground and air forces in the Solomons-New Guinea region, wanted the move to occur as soon as possible, the 7th Air Division resisted, citing the lack of readiness of the airfields to which they would deploy. Ultimately, it was not until mid-July that the fighter units were transferred, with the bomber and other units following in late July and beyond.10 Meanwhile, on 9 July, the 6th Air Division also moved its headquarters to Wewak.

With these new deployments, it seemed that the Army air forces had overcome the growing disparity between the Japanese and the Allied air forces. However, in mid-June, the Army finally had to face reality, and the Operations Section of its High Command agreed upon an “Outline for Operations Guidance in the New Guinea Area” with the Operations Sections of the Navy High Command, which called for a holding strategy in New Guinea.11 While this meant that the assault on Port Moresby was finally and formally shelved, the Army air forces in New Guinea still had the task of neutralizing the Allied air bases at Mount Hagen, Bena Bena, Wau, Salamaua and other places. In addition, they had to defend their own airbases, and provide fighter escort for convoys attempting to supply the Japanese garrisons on New Guinea.

Defensive tasks sapped much of the Japanese Army’s air strength on New Guinea. In fact, by this time the Army air forces were fighting a largely defensive campaign. Out of 1,308 sorties flown in July 1943, 494 were convoy escort, 84 were intercept and 190 were ground support. Such missions meant that fewer aircraft and pilots were available for mounting air attacks on Allied airbases, even though the need for such air attacks was well-recognized as necessary for regaining air superiority.12

In early August 1943, the Fourth Air Army had an operational strength of 130 aircraft (the Navy’s operational strength in the Solomons-New Guinea area at this time was approximately 220 aircraft).13 This was just one-third of what the Fourth Air Army was supposed to have on paper, and was also only an operational rate of approximately 50 per cent. The major causes of this low operational rate were widespread illness among the aircrews, along with, of course, the lack of aircraft replacements.

Nevertheless, the Japanese attempted to carry out their plan to regain air superiority. On 12 August, the Fourth Air Army began to carry out air raids on the Allied airbases at Hagen, Bena Bena, Wau, Salamaua and elsewhere. This effort came to naught, however, when on 17 August 1943 Allied air armadas mounted a surprise air attack on Wewak, which was the Japanese Army’s principal airbase on New Guinea. As a result of this action, over 100 Japanese aircraft were destroyed, and the Fourth Air Army was reduced to an operational strength of just over 30 planes. The destruction of the Japanese air forces meant that the Allies could now conduct air operations more or less freely as far northwest as Aitape, whereas until then they had been held to Madang.14

The Army took this defeat quite seriously, since it meant a virtual end to air operations in New Guinea for the time being. The major cause for the scale of the defeat, as stressed by Eighth Area Army, was said to be the unpreparedness of the Japanese airbases. In particular, it was pointed out that the lack of sufficient aircraft shelters had left the Japanese planes vulnerable. In addition, the lack of a sufficient warning system was also identified as a culprit. The Japanese were still relying almost completely on a visual warning system, which did not provide enough time for aircraft on the ground to either scramble or be hidden. This problem was made worse by, again, the primitive condition of the airfields, which did not allow the quick scrambling of a large number of aircraft.

Yet another major cause stressed by the Fourth Air Army was the fact that all of its planes had been based on Wewak and other airbases which were, quite literally, in the front line. The Fourth Air Army, along with the 6th and 7th Air Divisions, had wanted to deploy its units more in depth, with a certain number of aircraft based at airfields further to the rear. However, both the Army High Command and the Eighth Area Army had insisted that as many aircraft as possible be based at airfields at the front, to make them easier to use in operations. While this was to a certain extent inevitable, since the Japanese did not have the necessary number of airbases in the rear, it also meant that the bulk of their aircraft would be based in locations extremely vulnerable to large-scale enemy air attacks.

The end of the Japanese air forces in New Guinea

After the debacle at Wewak, the Army tried to replenish its losses. At the same time, it began to strengthen its airbase at Hollandia, in order to provide more depth to the Wewak position. The Allies, however, would not wait while these efforts were being made. In September, the Allies pushed the Japanese back from their positions at Lae and Salamaua, and on 22 September they landed near Finschhafen, on the northeastern coast of New Guinea. Despite their desperate defensive efforts, therefore, the Japanese continued to slowly be pushed back in New Guinea, and in the Solomons as well. In late September, the Army air forces had at their disposal only 60 or 70 operational aircraft with which to oppose the Americans in New Guinea, even though they had two air divisions, the 6th and 7th, in the area.15

Faced with these developments, the Japanese High Command was finally forced to make a major strategic change. On 15 September 1943 the concept of the Absolute National Defence Zone was adopted as policy, finally revising the strategy of 7 March 1942, mentioned above, which had called for the continuation of major offensive operations in order to secure a undefeatable strategic situation. Under the Absolute National Defence Zone concept, a delaying action would be continued in the South Pacific while a new line of defence was prepared along the Marianas-Carolines-Philippines arc. Here it was hoped that the Americans would be smashed when they tried to assault it. Counter-offensives would follow, and, if the situation permitted, offensive operations in New Guinea would be resumed in mid-1944 or later.16

The Army and Navy High Commands adopted new strategies based upon the new concept on 30 September. Regarding the New Guinea area, the Second Area Army was transferred from Manchuria on 29 October and placed in charge of defending the East Indies and the western half of New Guinea up to 140 degrees east longitude. On 30 October, the 7th Air Division was transferred from the Eighth Area Army to the Second Area Army, and its headquarters pulled back to Ambon Island on 21 November.

During this time the 6th Air Division, using what few planes it had (since the Japanese air forces had not recovered from the August air raid), continued its mostly-defensive operations out of Wewak. When the Allies landed at Saidor on 2 January 1944, thereby sealing off the Dampier Straits, the Fourth Air Army launched an all-out counter-attack against the invasion areas. By this time, however, it had less than 100 operational aircraft, and was able to fly only 160 sorties in all on five separate missions. Not only did this have a negligible effect on the Allied landing force, but by the time this effort was over, the Fourth Air Army was down to less than 50 operational planes.17 The Japanese Navy also launched attacks against the Saidor invasion force, since the loss of the Dampier Straits would mean that Rabaul and New Guinea would be cut off from each other. Even when the Navy aircraft operating out of Rabaul were included, however, the Japanese were down to approximately 100 operational planes for the entire New Guinea-Solomons area by mid-January, which made even a holding operation extremely difficult and indicates the desperate situation of the Japanese air forces in this entire area at this time.

In order to improve its position, the Army High Command decided at the end of January 1944 to move the Fourth Air Army’s centre of operations further westward than Hollandia, in order to give its positions more depth and enable it to conduct a more flexible holding operation. These orders were held up by Eighth Area Army, however, because it felt that the necessary airbases in the rear, to the west, were not ready. The High Command also decided at this time to temporarily reinforce the Fourth Air Army with some of the Second Area Army’s air units. While they would remain under Second Area Army control, these units were to “cooperate” with the Eighth Area Army. The units to be loaned in this fashion were the 33rd and 77th Sentai (fighters), 45th and 75th Sentai (light bombers) and the 60th Sentai (heavy bombers). These had been slated as Second Area Army reserve, to be based in the East Indies, but were thus committed to the front line.

In the meantime, the Japanese Navy’s air force was finally knocked out of New Guinea and the Solomons as well. On 17 February 1944, U.S. Navy carriers launched a massive raid on the key Japanese base in the Central Pacific, Truk Island, destroying over 200 Japanese aircraft and inflicting heavy damage. The Japanese Navy had tried to hold out at Rabaul because its loss would threaten Truk, but since Truk was now directly threatened, the Japanese Navy had to replace its air losses with the only air forces available, which were those at Rabaul. Three days after the attack, all of the remaining naval aircraft of the Japanese II Air Fleet at Rabaul were flown to Truk, and the Japanese naval air arm’s presence in the New Guinea-Solomons area came to an end.18

On 25 March 1944, the Fourth Air Army was transferred to the Second Area Army, and its headquarters arrived in Hollandia. As an airbase, however, Hollandia was still not adequately prepared. For example, facilities needed for dispersing aircraft and supplies, and for storing supplies, were still not ready. In addition, radar and other warning and intelligence networks were also only being prepared. The same situation which had plagued the Japanese at Wewak thus existed at Hollandia. Nevertheless, by the end of March the Japanese Army had managed to assemble approximately 300 aircraft at Hollandia; of these, however, only about 150 were operational.

In an eerie repeat of the earlier disaster at Wewak, the bulk of Japanese aircraft at Hollandia were wiped out in an air raid carried out over two days, 30-31 March. Over 150 planes were destroyed on the ground. The commander of the 6th Air Division and the chief of staff of the Fourth Air Army, among others, were relieved, because for them this was the second time that so many Japanese aircraft had been caught by surprise on the ground. This time, the Japanese were not given any time to recover. On 22 April 1944, the Americans landed near Hollandia. The personnel of the 6th Air Division, including its remaining pilots, were forced to retreat overland to the west, arriving in Sarmi by early May, but had to leave their aircraft behind. As a result, the 6th Air Division’s remaining 100 aircraft were lost. The division was never reconstructed, ultimately being disbanded in August 1944.

This left the 7th Air Division, operating at this time primarily out of bases in the East Indies. But this force, too, was severely depleted. On 25 May 1944, it had an operational strength of only 87 aircraft.19 When the Americans landed on Biak Island two days later, the 7th Air Division tried to provide air support for the defenders, but since it had so few aircraft – most of which had to be used for convoy escort and air defence missions – it was unable to do so effectively. By early July the Biak garrison had been wiped out. This was to be the last major action in which Japanese air forces operated over New Guinea. Thereafter, the major action in the Pacific would shift to the Marianas, Palaus and then the Philippines, and what remained of the Japanese Army and Navy air forces was committed to these areas. The air war in New Guinea was effectively over, having ended in an Allied victory.

Causes and implications of the Japanese defeat

There was no single, direct cause for the Japanese defeat in the air battle over New Guinea. Since this issue has been covered extremely well in many works in both English and Japanese, the causes will be enumerated only briefly here.20 First of all, the numerical advantage of the Allies was undoubtedly a major factor. While this was not a decisive advantage until the latter half of 1943, its effect was ultimately crucial. The Allies’ aircraft strength grew steadily, while the Japanese air forces in the area barely grew, and instead often steadily decreased, in number. This applied to aircrews and ground support staff. The Allies, and especially the Americans, were able to send increasing numbers of such personnel to the front, and as the war progressed they were better trained too. In comparison, the Japanese were hard-pressed to replace their pilots and ground staff, and the flying and combat abilities of what replacement pilots were sent to the front gradually decreased in quality.

The technical quality of the aircraft used was also a factor. While Japanese aircraft in this period, especially those of the Navy, were quite adequate (and in some respects excellent) for certain purposes, they were not as suited for an aerial war of attrition when compared to the Allies’ aircraft. For example, Japanese machines, lacking good self-sealing fuel tanks and pilot armour, were more susceptible to damage. In addition, Japanese types were often more difficult to manufacture, which means that they were not as suited to mass production as those operated by the Allies.21

A third factor is the generally low level of technology of Japan at this time, in comparison with the Allies. For example, as has often been pointed out, the Japanese constructed most of their airfields with manual labour. The Japanese Army recognized this as a problem as early as 1942, and had studied the issue – even establishing a base construction school in Japan22 – but little if anything had changed by 1944. This meant that the Japanese could not construct airbases quickly enough to maintain some semblance of parity with the Allies.

Another technological area in which Japan was deficient, and which is sometimes overlooked, is aircraft engine technology. To the end of the war, the Japanese were unable to produce a reliable 1,500-2,000-horsepower engine which would also be small enough to be used on fighter aircraft. Because of this, they were unable to equip their fighters with pilot armour, heavier armaments and better self-sealing fuel tanks. It is often said that the Japanese sacrificed these things because they wanted to lighten their aircraft and improve their offensive capabilities. While this is true, it is also the case that the Japanese wanted to install these items, especially later in the war, but could not do so because their additional weight would worsen aircraft performance to an unacceptable degree. The Japanese were unable to develop an engine large enough to handle the weight increase, so they had to leave such items off and make do with relatively unprotected aircraft.

Individually, each of these or any of the other factors which brought defeat to the Japanese air forces may not have been decisive; combined, they proved to be fatal. Faced with such a combination of negative factors, it is doubtful whether any nation could have overcome them and attained victory.

Finally, what did defeat in the South Pacific air war mean for the Japanese? The obvious implication is that it exposed the Marianas, Carolines and other central Pacific islands to Allied counter-attack, as well as the East Indies and Philippines. More importantly, it destroyed the air forces of the Japanese Navy, and to a lesser degree the Army as well, in terms of pilot quality. According to one estimate, the Japanese Navy’s losses between August 1942 (the start of the Guadalcanal campaign) and February 1944 (the end of its presence at Rabaul) amounted to more than 7,000 aircraft and aircrew members.23 This was the equivalent of several years’ worth of pilots.24 Because the quality of pilot training declined year by year, this meant that the ranks of experienced pilots within the Japanese Navy was severely depleted by early 1944. Since the Japanese Army committed some of its best air units to the New Guinea front, the large losses these suffered meant that the Army, too, had lost a large number of its best pilots.25 It is, therefore, not much of a surprise that the Japanese air forces, especially those of the Navy, suffered such lopsided defeats in the 1944 battles in the Marianas and the Philippines.

The Japanese were able, within a few months, to partially replenish the numbers of aircraft and pilots it had awaiting the Allies in 1944. What they could not replace were the qualitative losses they suffered in the South Pacific. Notwithstanding that Japanese aircraft were becoming outmoded by 1943, right up to the war’s end these could be very effective in the hands of a good pilot. But when the quality of the pilots declined, the result was disastrous. Without an effective air force, the Japanese would have had an exceedingly difficult time holding off the Americans in 1944, especially since the Americans’ air forces had steadily improved, both in quality and quantity, during 1943 and early 1944. The greatest implication of the Japanese defeat in the South Pacific therefore may be that it eliminated Japan’s ability to achieve air superiority anywhere in the Pacific during the remainder of the war.

© Hiroyuki Shindo

The author

Hiroyuki Shindo is currently an assistant professor in the Military History Department of the National Institute for Defense Studies, Tokyo. His special areas of interest are US-Japan diplomatic and military relations in the 1930s and 1940s, and the military history of the Second World War.


1 Takushiro Hattori, Daitoa Senso Zenshi (A Complete History of the Great East Asia War) (Tokyo: Hara Shobo, 1965), p.294.

2 The activities of the Navy’s air forces based in Rabaul and Lae during the spring and summer of 1942 have been excellently depicted in Saburo Sakai, Sakai Saburo Kusen Kiroku (Sakai Saburo Air Combat Records)(Tokyo: Shuppan-Kyodosha, 1953) (English translation: Samurai! (Dutton, 1957)) and many other works.

3 For a detailed account of the process by which the Army air forces were committed to the South Pacific, see Boeicho Boeikenkyusho, ed., Senshi Sosho Tobu Nyu-Ginia Homen Rikugun Koku Sakusen (Army Air Force Operations, Eastern New Guinea) (Tokyo: Asagumo Shinbunsha, 1967), pp.20-63.

4 Boeicho Boeikenkyusho, ed., Senshi Sosho Rikugun Koku no Gunbi to Unyo (3) Daitoa Senso Shusen made (Equipment and Operation of Army Air Forces (3) Until the end of the Great East Asia War) (Tokyo: Asagumo Shinbunsha, 1976), p.72.

5 Boeicho Boeikenkyusho, ed., op. cit. Senshi Sosho Tobu Nyu-Ginia Homen Rikugun Koku Sakusen, p.155.

6 Boeicho Boeikenkyusho, ed., op. cit. Senshi Sosho Rikugun Koku no Gunbi to Unyo (3) Daitoa Senso Shusen made, p.75.

7 Boeicho Boeikenkyusho, ed., op. cit. Senshi Sosho Tobu Nyu-Ginia Homen Rikugun Koku Sakusen, pp.185-90.

8 Takushiro Hattori, op. cit., pp.405-407.

9 Boeicho Boeikenkyusho, ed., op. cit. Senshi Sosho Rikugun Koku no Gunbi to Unyo (3) Daitoa Senso Shusen made, p.95.

10 Among other reasons, the slowness with which the Army transferred its air forces around the Pacific has been given as one of the reasons why the Army air forces were relatively ineffective against the Allies, because that prevented them from deploying and operating rapidly in areas where they were most needed. Masatake Okumiya, Rabaul Kaigun Kokutai (Rabaul Naval Air Corps) (Tokyo: Asahi Sonorama, 1998), pp.169-170.

11 Boeicho Boeikenkyusho, ed., op. cit. Senshi Sosho Rikugun Koku no Gunbi to Unyo (3) Daitoa Senso Shusen made, pp.93-95.

12 Ibid., p.104.

13 Ibid., p.101.

14 Ibid., pp.107-108.

15 Takushiro Hattori, op. cit., p.504.

16 Takushiro Hattori, op. cit., p.498.

17 Boeicho Boeikenkyusho, ed., op. cit. Senshi Sosho Rikugun Koku no Gunbi to Unyo (3) Daitoa Senso Shusen made, p.152.

18 Okumiya, op. cit., pp.366-369.

19 Boeicho Boeikenkyusho, ed., op. cit. Senshi Sosho Rikugun Koku no Gunbi to Unyo (3) Daitoa Senso Shusen made, p.187.

20 For example, see Eric Bergerud, Fire in the Sky (2000), for an excellent, interesting and in-depth analysis of the air war in the South Pacific.

21 For an analysis of Japan’s aircraft technology and its effects on the war as seen from the Japanese side, see for example Kunio Yanagida, Reisen Moyu (Zero Fighter Aflame) (Tokyo: Bungei Shunjyu, 1984-1990).

22 Boeicho Boeikenkyusho, ed., op. cit. Senshi Sosho Rikugun Koku no Gunbi to Unyo (3) Daitoa Senso Shusen made, p.135.

23 Bergerud, op. cit., pp.667-668.

24 An exact equivalent is difficult to calculate, because the Japanese were expanding their pilot training programs every year even before the Pacific war began.

25 According to one estimate, nearly 800 aircraft were lost by the Japanese Army in the South Pacific. Bergerud, op. cit., p.668.

The Siege of Tsingtau 1914: Aircraft

Launching a Maurice Farman two-seater from the Wakamiya during the Siege of Tsingtau.

“Plüschow in front of the city wall of Haizhou in the province of Jiangsu, on November 6, 1914. On that day he had escaped the besieged Tsingtao by his plane, and after a flight of ca. 200 km to the southwest, landed at Haizhou, because the plane had no fuel anymore.”

Both sides had elements of an air component; the Japanese Navy had the Wakamiya Maru with its complement of four Maurice Farman floatplanes, whilst the army detachment, initially consisting of three machines, deployed from an improvised airstrip near Tsimo on 21 September. Japanese aviation was, as was the case in every other nation, a recent phenomenon in terms of powered aircraft. Previous interest in aeronautics had centred on the use of balloons for reconnaissance, the first Japanese military balloons were sent aloft in May 1877, and an advanced kite type was designed and constructed in 1900 and successfully used during the Russo-Japanese War. A joint committee, with army, navy and civil input, was created on 30 July 1909 to investigate and research the techniques and equipment associated with ballooning; the Provisional Military Balloon Research Association or PMBRA.

Nominated by the Japanese Army to serve on the PMBRA were two officers with the rank of captain, Tokugawa Yoshitoshi and Hino Kumazo. Both had some experience with aviation. Hino designed and constructed a pusher type monoplane with an 8hp engine that he unsuccessfully, the engine was underpowered, attempted to get off the ground on 18 March 1910. Tokugawa was a member of the balloon establishment during the Russo-Japanese War. Both were sent to Europe in April 1910 to learn to fly at the Blériot Flying School at Étampes, France. Having passed the rudimentary course, they purchased two aeroplanes each and had them shipped to Japan; Tokugawa obtained a Farman III and a Blériot XI-2bis in France whilst Hino purchased one of Hans Grade’s machines and a Wright aircraft in Germany.

The first flights of powered aeroplanes in Japan occurred on 19 December 1910 at Yoyogi Park in Tokyo. Tokugawa flew the Farman III, powered by a 50hp Gnome engine, for 3 minutes over a distance of some 3000m at a height of 70m. Hino followed him immediately afterwards in the Grade machine, powered by a 24hp Grade engine, which flew for just over a minute and covered a distance of 1000m at a height of 20m.

A naval member of the PMBRA, Narahara Sanji, started on designing and constructing an aeroplane, with a bamboo airframe and a 25hp engine, during March 1910. Because of the low powered engine the machine failed to lift off when this was attempted on 24 October 1910, but with a second machine, the ‘Narahara Type 2’ powered by a Gnome engine similar to that used in the Farman III, he managed a 60m flight at an altitude of 4m on 5 May 1911. This flight, at Tokorozawa, in Saitama near Tokyo, the site of Japan’s first airfield, is considered to be the first Japanese civilian flight as Narahara had left the navy when he made it. It was also the first flight by a Japanese-manufactured aeroplane.

The first military flight by a Japanese-manufactured machine took place on 13 October 1911, when Tokugawa flew in a ‘PMBRA Type (Kaisiki) 1’ of his own design, based on the Farman III at Tokorozawa. These pioneers, whilst they had made astonishing progress, did not however possess the necessary research and technological resources to take Japanese aviation further. Because of this the Japanese decided to import aviation technology from Europe, though the navy established the Naval Aeronautical Research Committee in 1912 to provide facilities to test and copy foreign aircraft and train Japanese engineers in the necessary skills. Via this system, the foundations of a Japanese aviation industry were being laid; in July 1913 a naval Lieutenant, Nakajima Chikuhei, produced an improved version of the Farman floatplane for naval use. Nakajima Aircraft Industries, founded in 1917 after Nakajima resigned from the navy, went on to massive success.

The French were the world leaders in military aviation, with 260 aircraft in service by 1913, whilst the Russians had 100, Germany 48, the UK 29, Italy 26 and Japan 14. The US deployed 6.55 It comes as no surprise then to note that during the campaign against Tsingtau all the aeroplanes deployed by both Japanese services were French. Four Maurice-Farman MF7 biplanes and one Nieuport 6M monoplane formed the Army’s Provisional Air Corps, flying eighty-six sorties between them, whilst the navy deployed one Maurice-Farman MF7 floatplane and three Henri-Farman HF7 floatplanes. The navy planes flew 49 sorties and dropped 199 bombs.

A floatplane from the Wakamiya Maru flew over Tsingtau on 5 September, causing something of a surprise to the defenders, on a reconnaissance and bombing mission, releasing three bombs that caused no harm. It wasn’t the first aerial bombing ever – that had taken place during the 1911–12 Italo-Ottoman War – but it was a total surprise to the defenders. The reconnaissance element of the mission was more useful, being able to ascertain that Emden was not in harbour but that there were several other warships present. It was to be the first of several visits by both army and navy aeroplanes, against which the Germans could offer little defence, though the fact that the defenders had their own air ‘component’ was to result in what was probably (there are other contenders) the first air-to-air combat in history. Indeed, despite the remoteness of the campaign from the main theatre, this was only one of a number of such ‘firsts’.

Aviation on the German side was represented by one man and one machine; Kapitänleutnant Gunther Plüschow and his Rumpler Taube. Plüschow had served in the East Asiatic Cruiser Squadron, at the time under the command of Vice Admiral Carl Coerper, as a junior officer aboard SMS Fürst Bismarck in 1908. Assigned to the Naval Flying Service in autumn 1913, he arrived on 2 January 1914 at Johannisthal Air Field near Berlin to commence pilot training, and, having first taken to the air only two days previously, acquired his licence on 3 February 1914. The Naval Air Service, which had been created in 1912 and divided between aeroplane and airship sections the following year, was, in 1914, something of a misnomer; naval aviation was in a greatly underdeveloped state with sole assets consisting of two Zeppelin airships, four floatplanes and two landplanes. This was largely due to Tirpitz, who, despite his later assertions that, prior to 1914, he saw the aeroplane as the weapon of the future as against the airship, was not prepared to, as he saw it, divert funds from the battle-fleet in order to develop the technologies and techniques required.

Having not seen it for some six years Plüschow arrived in Tsingtau by train on 13 June – an extremely long and undoubtedly tedious journey across the Siberian steppe – whilst two Rumpler Taube aeroplanes with 100hp engines, especially constructed for service in China, travelled by sea arriving in mid-July 1914. The second machine was to be piloted by an army officer assigned to the III Naval Battalion, Lieutenant Friedrich Müllerskowski, and the arrival of the two aviators and their machines took the total ‘air force’ available in the territory to three men and aircraft. The third aviator was Franz Oster, a former naval officer who had settled in Tsingtau in 1899, but returned to Germany in 1911 and learned to fly. He returned to the territory in 1912, via Ceylon (Sri Lanka) complete with a Rumpler Taube equipped with a 60hp engine. During his sojourn in Ceylon he attempted a flight at Colombo Racecourse in a Blériot Monoplane on 30 December 1911 that ended in near disaster; the machine was wrecked and Oster was hurt. He nevertheless, after having returned to the territory and replaced the engine in his Taube with a 70hp Mercedes unit, made a series of flights from the Tsingtau racecourse, the first being on 9 July 1913.

Plüschow and Müllerskowski took charge of reassembling the two aeroplanes delivered by sea and the former successfully made several flights from the extremely small and dangerous landing ground at the racecourse on 29 July 1914. A further two days were needed to get the second aeroplane constructed, and on the afternoon of 31 July Müllerskowski set off on his first flight. It ended in disaster; after only a few seconds in flight, and from an altitude estimated to be 50m, the machine lost control and plunged over a cliff onto the rocks below. Müllerskowski was seriously, though not mortally, wounded and the Taube completely wrecked.

Whether it was just bad luck or whether there were atmospheric conditions pertaining at the time that made flying problematical we cannot know, but it would seem to have been a combination of the two that afflicted Plüschow on 3 August. Having taken off successfully and flown a reconnaissance mission over the territory, his first ‘important’ sortie, he was experiencing difficulty in attempting to land when his engine failed and he crash-landed into a small wood. He was unhurt but the Taube was badly damaged and, upon accessing the spare wings and propellers sent out with the aeroplanes, he discovered that the replacement parts had rotted away or suffered moistureinduced damage during the voyage. He was fortunate that the engine, for which replacement parts could only have been extemporised with difficulty, was still serviceable and that there were skilled Chinese craftsmen available; the latter fashioned him a new composite propeller from oak. Despite this device having to be repaired after every flight, having been assembled with ordinary carpenter’s glue it exhibited a disconcerting tendency to revert to its component parts under the strain of operational usage, it remained serviceable throughout the rest of the campaign.

Plüschow’s machine was out of action until 12 August, but on the 22 August an attempt was made to augment it with Oster’s aeroplane; he attempted to lift off from the racecourse in his older craft but stalled and crashed, occasioning damage necessitating several days of repair though remaining unhurt personally. Another attempt was made on 27 August with the same result, though this time the damage was more severe with the aeroplane ‘completely destroyed’ to such an extent that ‘reconstruction was no longer viable’. It seems however that Oster did not concur as the diary entry for 13 October 1914 made by the missionary Carl Joseph Voskamp, records Oster once again, and apparently finally, attempting and failing to take off, and notes that this might be due to unfavourable atmospheric conditions.

Plüschow and his Taube, by default the sole representatives of German aviation, could not of course provide anything much in the way of air defence against the Japanese. Nor could they achieve a great deal in the way of keeping open communications with the world outside the Kiautschou Territory. What was possible though, within the operational capabilities of man and machine, was reconnaissance, and the clearing up somewhat of the weather on 11 September allowed an aerial sortie to take place two days later. Plüschow flew northwestwards to investigate rumours of the Japanese landing and advance, and discovered their forces in some strength at Pingdu; the marching elements of the Japanese force reached Pingdu between 11 and 14 September. He also received his ‘baptism of fire’ from the infantry, returning with around ten bullet holes in his plane and resolving not to fly below 2000m in future in order to preserve his engine and propeller.


Reconnaissance duties devolved then on to the air component, represented by Plüschow and his Taube. There was also a balloon detachment consisting of two observation-balloon envelopes and the necessary ground infrastructure. The German observation balloons of 1914 were known as Drachen, a name commonly adopted for all sausage-shaped kite-balloons, and had been developed by Parseval-Sigsfeld. Adopted for use in 1893, they represented a significant investment in terms of equipment and manpower for the Tsingtau garrison, the standardised balloon section in 1914 consisting of 1 balloon (plus 1 spare envelope), 4 observers, 177 enlisted ranks, 123 horses, 12 gas wagons, 2 equipment wagons, a winch wagon and a telephone wagon. The balloon had made several ascents from Tsingtau during the course of the conflict, but the observer had been unable to see anything of value. In order to attempt to remedy this the device was moved closer to the front and another ascent made on 5 October. It was to be the last such, as the Japanese artillery immediately found its range with shrapnel shell and holed it in several places. A ruse involving the spare balloon was then tried; it was sent up to draw the attackers’ fire and so reveal the position of their guns. According to Alfred Brace:

It contained a dummy looking fixedly at the landscape below through a pair of paste-board glasses. But there happened to arise a strong wind which set the balloon revolving and finally broke it loose and sent it pirouetting off over the Yellow Sea, the whole exploit, I learned afterwards, being a great puzzle to the British and Japanese observers outside.

Plüschow flew reconnaissance flights every day that the weather, and his propeller, permitted, sketching the enemy positions and making detailed notes. He achieved this by setting the engine so as to maintain a safe altitude of over 2000m, and then steering with his feet, the Taube had no rudder and horizontal control was achieved by warping the wings, whilst peering over the side of his cockpit. The Japanese had extemporised a contingent of antiaircraft-artillery – the ‘Field-Gun Platoon for High Angle Fire’, with the necessary angle achieved by dropping the gun-trail into a pit behind the weapon – and although the shrapnel barrage thus discharged proved ineffective it was deemed by Plüschow to be troublesome nevertheless. Where he was at his most vulnerable was on landing, and a battery of Japanese artillery was tasked specifically with destroying the Taube as it descended to the racecourse, which of course was a fixed point at a known range. Little more than good luck, and what he called ‘ruses’ such as shutting off the engine and swooping sharply to earth, saw him through these experiences, but remarkably both man and machine came through without serious injury.

Whatever inconveniences Plüschow and the fortress artillery might inflict upon the force massing to their front, they could do nothing to prevent the landing of men and materiel at Wang-ko-chuang and Schatsykou, nor could they prevent the deployment of these once landed. The previous efforts by the navy in terms of minelaying did though still pay dividends, as when the Japanese ‘aircraft carrier’ was badly damaged. As the report from the British Naval Attaché to Japan put it in his report of 30 November:

[…] a few minutes after 8 a.m. [on 30 September] the ‘Wakamiya Maru’ struck a mine in the entrance of Lo Shan Harbour, and had to be beached to prevent her sinking; her engines were disabled owing to breaking of steam pipes, No. 3 hold full, and one man killed – fortunately no damage done to aeroplanes though it is feared that a spare engine may be injured. […] As the Aeroplane establishment is all being moved ashore at this place this accident will not affect the efficiency of the Aeroplane Corps.


In order to achieve the enormous amount of digging that the plan required, the organic engineering component of the 18th Division, the 18th Battalion of Engineers, was augmented by two more battalions; the 1st Battalion of Independent Engineers (Lieutenant Colonel Koga) and the 4th Battalion of Independent Engineers (Lieutenant Colonel Sugiyama). The infantry that would man the siege works were also provided with weaponry specifically suited to trench warfare; two light platoons and one heavy detachment of bomb-guns (mortars).

In order to gain detailed knowledge of the defences the naval and army air components were tasked with flying reconnaissance missions over the German positions. They also flew bombing missions, which caused little damage, and attempted to discourage their single opponent (though they were initially uncertain how many German aircraft they faced) from emulating them; the latter with some degree of success. Plüschow records that he was provided with extemporised ‘bombs’ made of tin boxes filled with dynamite and improvised shrapnel, but that these devices were largely ineffectual. He claimed to have hit a Japanese vessel with one, which failed to explode, and to have succeeded in killing thirty soldiers with another one that did. It was during this period that he became engaged in air-to-air combat, of a type, with the enemy aeroplanes. Indeed, if Plüschow is to be believed, he succeeded in shooting down one of the Japanese aeroplanes with his pistol, having fired thirty shots. It would appear however that even if he did engage in aerial jousting of the kind he mentions the result was not as he claimed; no aircraft were lost during the campaign. The Japanese did however do their utmost to prevent him reconnoitring their positions, as they were in the process of emplacing the siege batteries and, if the positions became known, they could expect intense efforts from the defenders to disrupt this process. Experience showed the Japanese that the time delay between aerial reconnaissance being carried out and artillery fire being concentrated on the area so reconnoitred was around two hours.

Indeed Watanabe insisted that his batteries were emplaced during the hours of darkness, despite the inconvenience this caused, and carefully camouflaged to prevent discovery. That the threat from Plüschow, albeit indirect, was very real had been illustrated on 29 September; he had overflown an area where the British were camped and noted their tents, which were of a different pattern to the Japanese versions. This had resulted in heavy shelling, causing the camp to be moved the next day to the reverse slopes of a hill about 1.5km east of the former position. He also posed a direct threat though perhaps of lesser import; on 10 October he dropped one of his homemade bombs on the British. It failed to explode, but the unit concerned moved position immediately – such an option was not available to 36-tonne howitzers that required semi-permanent emplacement.


The Japanese Navy began sending in vessels to shell the city and defences again. On 25 October the Iwami approached, though staying outside the range of Hui tsch’en Huk. By listing the ship to increase the range of her main armament, Iwami was able to fire some thirty 305mm shells at Hui tsch’en Huk, Iltis Battery and Infantry Work I. The next day the vessel returned in company with Suwo and the two vessels bombarded the same targets. On 27 October Tango and Okinoshima replaced them, and the same ships returned the next day to continue the assault. Because of the distance involved, some 14.5km, this fire was inaccurate in terms of damaging the specific installations in question, but nevertheless was destructive of the nerves of the trapped garrison. It was particularly frustrating in terms of the gunners at Hui tsch’en Huk who were unable to effectively reply.

In addition to this display of naval force, Japanese air power had been much in evidence over the period, their operational activity increasing with sorties over the German lines and rear areas.

Almost every day these craft, announcing their approach by a distant humming, came overhead, glinting and shining in the sun as they sailed above the forts and city. At first they were greeted by a fusillade of shots from all parts of the garrison. Machine guns pumped bullets a hundred a minute at them and every man with a rifle handy let fire. As these bullets came raining back upon the city without any effect but to send Chinese coolies scampering under cover, it was soon realised that rifle and machinegun fire was altogether ineffective. Then special guns were rigged and the aeroplanes were subjected to shrapnel, which seemed to come nearer to its sailing mark each day but which never brought one of the daring bird-men down. One day I saw a biplane drop down a notch after a shell had exploded directly in front of it. I looked for a volplane [glide with engine off] to earth, but the aviator’s loss of control was only momentary, evidently caused by the disturbance of the air. During the bombardment these craft circled over the forts like birds of prey. They were constantly dropping bombs, trying to hit the ammunition depots, the signal station, the Austrian cruiser Kaiserin Elizabeth, the electric light plant, and the forts. But […] these bombs were not accurate or powerful enough to do much damage. A few Chinese were killed, a German soldier wounded, tops of houses knocked in, and holes gouged in the streets, but that was all. The bombs fell with an ominous swish as of escaping steam, and it was decidedly uncomfortable to be in the open with a Japanese aeroplane overhead. We are more or less like the ostrich who finds peace and comfort with his head in the sand: In the streets of Tsingtau I have seen a man pull the top of a jinrickisha over his head on the approach of a hostile aeroplane and have noticed Chinese clustering under the top of a tree.

They also managed an aviation first on the night of 28–9 October when they bombed the defenders’ positions during the hours of darkness. Attempts to keep Plüschow from effectively reconnoitring were largely successful, even though the efforts to dispose of him or his machine permanently were ineffective. However, because problems with the Taube’s homemade propeller kept him grounded on occasion, and because the Japanese positions were worked on tirelessly, when he did take to the air he found the changes in the enemy arrangements – ‘this tangle of trenches, zigzags and new positions’ – somewhat bewildering and difficult to record accurately. Precision in this regard was not assisted by the Japanese attempts to shoot him down or otherwise obstruct him.

The artillery coordinating position on Prinz-Heinrich-Berg reported itself ready for action on 29 October and Kato sent four warships in to continue the naval bombardment whilst acting under its direction. Between 09:30hr and 16:30hr Suwo, Tango, Okinoshima and Triumph bombarded the Tsingtau defences, adjusting their aim according to the feedback received from the position via radio. They withdrew after discharging some 197 projectiles from their main guns, following which SMS Tiger was scuttled during the hours of darkness.

Plüschow managed to get airborne on the morning of 30 October and was able to over-fly the Japanese positions before the enemy air force could rise to deter him. This might have been lucky for him as one of the aeroplanes sent up had been fitted with a machine gun. He was able to report the largescale and advanced preparations of the besieging force, information that the defence used to direct its artillery fire. This was repaid when Kato’s bombarding division returned at 09:00hr to recommence their previous day’s work. Despite the communication channel working perfectly, and the absence of effective return fire from Hui tsch’en Huk – they had established the maximum range of this battery was 14.13km and accordingly stayed just beyond its reach – the firing of 240 heavy shells again did little damage.

The 31 October was, as the defenders knew well, the birthday of the Japanese Emperor and by way of celebration Kamio’s command undertook a brief ceremony before, at about 06:00hr, Watanabe gave the order for the siege train to commence firing, or, as one of the correspondents of The Times put it: ‘daylight saw the royal salute being fired with live shell at Tsingtau’. The Japanese fire plan was relatively simple.

On the first and second days, in addition to bombardment of the enemy warships, all efforts would be made to silence the enemy’s artillery so as to assist the construction and occupation of the first parallel.

From the third day up until the occupation of the second parallel (about the fifth day) the enemy artillery would be suppressed, his works destroyed and the Boxer Line swept with fire in order to assist in the construction of the second parallel.

Following the occupation of the second parallel, the majority of the artillery fire would be employed in destroying the enemy’s works, whilst the remainder kept down hostile infantry and artillery that attempted to obstruct offensive movement in preparing, and then assaulting from, the third parallel.

After the Boxer Line had been captured, the artillery would support the friendly troops in securing it from counter-attack and then bombard his second line; Iltiss, Moltke and Bismarck Hills.

There were, roughly, twenty-three Japanese artillery tubes per kilometre of front, a density that was comparable to that attained during the initial stages of the war on the Western Front, though soon to be dwarfed as artillery assumed the dominant role in positional warfare. The land-based artillery was augmented, from about 09:00hr, by the naval contribution as Kato once again sent his heavy ships into action.

The combined barrage soon silenced any German return fire because, even though they had refrained from pre-registering their siege batteries, the Japanese knew where the fixed defensive positions were and shortly found their range; fire was also brought to bear on any targets of opportunity. The German batteries were suppressed less by direct hits than by their positions being submerged in debris from near misses. This was to prove of some importance for the defenders were able in several cases to return their weapons to service, largely due to the relative antiquity and thus lack of sophistication of much of the ordnance, without the need for extensive repairs. Indeed, despite the crushing superiority enjoyed by the attackers, the defensive fire was to continue to some degree throughout the day and into the night. The most obvious sign of the effects of the bombardment, at least to those observing from a distance, were the huge plumes of smoke caused by hits on the oil storage tanks adjacent to the Large Harbour. Two of these, owned by the Asiatic Petroleum Company – the first Royal Dutch/Shell joint venture – and Standard Oil respectively, had been set afire early on and their contents in turn caused other fires as they flowed around the installations, these proving beyond the capacity of the local fire brigade to control. In fact the destruction of the Standard Oil installations was accidental. The General Staff History records that a note had been received via the Japanese Foreign Office from the US government asking that they be spared. Accordingly the objective was struck out of the plan but to apparently no effect, perhaps demonstrating the relative inaccuracy of the fire.

There were a number of independent observers of the operations at this stage; correspondents from various journals and foreign military observers had arrived in the theatre in late October. Though the Japanese were intensely secretive they could not conceal the fact of their bombardment or the plainly visible results.

The thunder of the great guns broke suddenly upon that stillness which only dawn knows, and their discharges flashed readily on the darkling slopes. The Japanese shooting, it is related, displayed remarkable accuracy, some of the first projectiles bursting upon the enormous oil tanks of the Standard Oil Company and the Asiatic Petroleum Company. A blaze roared skywards, and for many hours the heavens were darkened by an immense cloud of black petroleum smoke which hung like a pall over the town. Shells passing over these fires drew up columns of flame to a great height. Chinese coolies could be seen running before the spreading and burning oil. Fires broke out also on the wharves of the outer harbour.

Many of the Japanese shells, no doubt due to the lack of pre-registration, were over-range and landed in Tapatau and Tsingtau, though the former received the worst of it. It has been estimated that at least 100 Chinese were killed during this period and a deliberate targeting exercise carried out later in the day on the urban areas. The bombardment continued with varying levels of frequency throughout the daylight hours of 31 October, and at nightfall the Japanese gunners switched to shrapnel – by bursting shrapnel shell over the defenders’ positions they made it difficult, if not impossible, for repairs to be carried out. Such fire also covered the forward movement of the Japanese engineers as they extended their saps towards the Boxer Line and began constructing the parallel works some 300m ahead of the advanced investment line.

At daylight on 1 November the high-explosive barrage resumed, again concentrating primarily on the German artillery positions though many of these were now out of action. The secondary targets were the defences in the Boxer Line, particularly the infantry works and the extemporised defences between them. The ferro-concrete redoubts withstood the bombardment without any serious damage, and, though they were scarred and badly battered externally, none of the projectiles penetrated any vital interior position. The communication trenches and other intermediate field works were however obliterated and this, together with the destruction of much of the telephone system, isolated the personnel manning the works, both from each other and from the command further back. The targeting of the signal station further hindered communication of every kind, and with the bringing down of the radio antenna even one-way communication from the outside world was terminated. Shutting this down was probably a secondary objective; the primary reason for targeting the signal station was to prevent it jamming and otherwise interfering with Japanese wireless communications which had become a problem.

After nightfall the sappers returned to their task of advancing the siege works whilst infantry patrols went forward to reconnoitre and probe the defences. One such probe crossed the Haipo and a four-man party entered the ditch near Infantry Work 4, which was under the command of Captain von Stranz, and began cutting the wire. They remained undetected for some time, indicating the lack of awareness of the defenders who stayed under cover, but were eventually heard and forced to retire with the loss of one man after machine-gun fire was directed at them. A second patrol took their place a little later and, under the very noses of the defence, completed the wire-cutting task before they too were detected.

The defenders conceived that an assault in strength was under way and called down artillery fire in support from Iltis Battery and moved a reserve formation made up of naval personnel, whose ships had been scuttled, towards the front. The Japanese patrol withdrew, leaving the defenders under the erroneous impression that they had defeated a serious attempt at breaching the line, rather than, as was the case, an opportunistic foray. However, what the probe had revealed to the attackers was that the defenders were remaining largely inside the concrete works, leaving the gaps between them vulnerable to infiltration. This was confirmed by the experiences of a separate patrol that reconnoitred near Infantry Work 3, also known as the ‘Central Fort’ to the Japanese; the knowledge gained being of some potential worth. Also of value was an understanding of the nature of the barbed-wire obstacles. These were permanent fixtures, with extra heavy wire holding barbs ‘so closely together that it would be difficult to get a pair of pliers in a position to cut it’. It did prove possible to cut the wire, but the stakes it was strung on, made from heavy duty angle-iron secured to a square baseplate some 300mm per side and sunk into the ground to a depth of about half a metre, proved almost impossible to dislodge. Initial intelligence had indicated that the wire was charged with 30,000 volts, but direct examination showed this not to be the case.

Apart from repelling the Japanese attack, as they thought, the defenders spent the night of 1–2 November destroying further equipment that might be of use to the besiegers. Chief amongst this was Kaiserin Elisabeth; shortly after midnight, having fired off her remaining ammunition in the general direction of the Japanese, the vessel was moved into deep water in Kiautschou Bay and scuttled. Explosive charges extemporised from torpedo warheads ensured that the ship was beyond salvage even if the wreck was located. The Austro-Hungarian cruiser was only one of several vessels scuttled that night, including the floating dock, which was seen to have disappeared the next morning. Only the Jaguar remained afloat at sunrise at which point the rapid rate of advance of the attackers, up to the edge of the Haipo River between Kiautschou Bay and the area in front of Infantry Work 3, was revealed to the Germans.

Revealed to the Japanese, by their action during the hours of darkness, was the precise position of the Iltis Battery and this was promptly targeted and put out of action by counter-battery fire. The remorseless battering by the siege train also resumed, and the German inability to respond effectively due to the accuracy of Japanese return fire began to be exacerbated by a shortage of ammunition. The Japanese bombardment, though intense, was perhaps not as destructive as it could have been. What seems to have mitigated the effect to some extent was the high rate of dud shells. One press correspondent that entered Tsingtau after the conclusion of operations noted the proliferation of ‘giant shells, some three feet long and a foot in diameter, [that] were lying about on side-walk and street still unexploded’. Burdick, working from contemporary German estimates, calculates that between 10 and 25 per cent of the Japanese ordnance failed to explode. This shortcoming, being attributable to faulty manufacture, played a major role in sparing the defenders a worse ordeal than they had to endure anyway.

Also sparing the defenders to some extent on 2 November was the onset of rain, which affected the assailants more than the defence inasmuch as the attackers’ diggings became waterlogged and collapsed in some cases. Further alleviation was attributable to the reduction of the rate of fire of the 280mm howitzers. Their temporary emplacements suffered from the immense recoil and had the potential to render firing both dangerous and inaccurate. The remainder of the siege train concentrated its fire on the Boxer Line, particularly in an attempt to destroy the wire and obstacles in the trench and thus mitigate the need to resort to manual methods with their inevitable human cost. The power station was also targeted, with the result that the chimney was brought down in the evening, thus rendering the city dependent on primitive forms of lighting.

The 3 and 4 November saw further progress in the advancement of the siege works and continuing bombardment, though useful targets for this were now at a premium as little of the defences remained other than the concrete Infantry Works. The defenders had begun destroying their batteries on 2 November as they ran out of ammunition, and in any event returning the Japanese fire was a hazardous business due to the rapidity and accuracy of the response. The lack of defensive fire allowed some reorganisation of the siege artillery and several of the batteries were moved forward and swiftly re-emplaced with the minimum of disruption. On the far right of the Japanese line the sappers attached to 67th Infantry Regiment had advanced their works to within a short distance of the Haipo River, and thus close to the city’s water pumping station situated on the eastern bank. The decision was taken by 29th Infantry Brigade to attempt to take the station on the evening of 4 November and a company sized unit, comprising infantry and engineers, was assembled. It had not been bombarded by the siege artillery; the idea seemingly being to preserve it for future use by the occupying force. So the engineers cut through the defensive wire with Bangalore Torpedoes, thus allowing the infantry to surround the place, whilst a box artillery barrage insulated it from any attempted relief.

Despite its relative isolation the pumping station was actually a wellfortified strongpoint. The machinery rooms, stores and personnel quarters were located underground and well protected by ferro-concrete. The whole was surrounded by a bank of earth some 6m in height, itself protected by a ditch, about 12m wide and 2m deep at the counterscarp, that was filled with barbed-wire obstacles. The leader of the platoon investigating the area, 2nd Lieutenant Yokokura, later reported that the personnel manning the station had locked themselves inside behind ‘iron doors’ and were still working the pumps, but when they realised that the enemy were upon them they opened the doors and surrendered. The haul amounted to one sergeant major, twenty rank and file, two water works engineers and five Chinese, together with twenty-five rifles. The station was immediately fortified against any counter-attack and with its loss Tsingtau was without a mains water supply and thus dependent on the several, somewhat brackish, wells within the city.

Elsewhere along the line the nocturnal ‘mole warfare’ techniques advanced the saps and trenches ever nearer to the ditch in order to construct the third parallel – the final assault line. This progressed everywhere apart from the British sector of front, where enemy fire prevented the final approach being made. As Barnardison reported:

On 5 November I was ordered to prepare a Third Position of attack on the left bank of the river. This line was to a great extent enfiladed on both flanks by No. 1 and 2 Redoubts, especially the latter, from which annoying machine-gun fire was experienced. The bed of the river […] had also to be crossed, and in doing so the working parties of the 2nd Battalion South Wales Borderers suffered somewhat severely, losing 8 non-commissioned officers and men killed and 24 wounded. The 36th Sikhs had only slight losses. Notwithstanding this a good deal of work was done, especially on the right flank. I considered it my duty to represent to the Japanese Commander-in-Chief the untenable nature, for permanent occupation, of the portion of the Third Position in my front, but received a reply that it was necessary for it to be held in order to fit in with the general scheme of assault.

Though most diplomatically phrased, it is possible to distinguish in the final sentence of this quotation a hint of asperity in the relations between the Allies. Indeed, though suppressed for political reasons at the time, the British military contribution did not impress the Japanese in any way, shape or form. Reports from the front revealed the perception that the British were reluctant to get involved in the fighting and ‘hard to trust’. More brutal opinions had it that they were no more than ‘baggage’ and ‘decoration’ on the battlefield. The nature of these observations filtered through to the Japanese press, one report stating that: ‘Only when nothing happened were British soldiers wonderful and it was like taking a lady on a trip. However, such a lady can be a burden and lead to total disaster for a force when the enemy appears.’

Daylight on 5 November saw three Japanese aeroplanes overfly the German positions dropping not explosive devices, as might have been expected, but rather bundles of leaflets carrying a message from the besiegers:

To the Respected Officers and Men of the Fortress.

It would act against the will of God as well as humanity if one were to destroy the still useful weapons, ships, and other structures without tactical justification and only because of the envious view that they might fall into the hands of the enemy.

Although we are certain in the belief that, in the case of the chivalrous officers and men, they would not put into effect such thoughtlessness, we nevertheless would like to emphasise the above as our point of view.

On the face of it this message seemed to clearly indicate that the besiegers, perceiving that they would shortly be in occupation of the city, desired that as much of it be preserved as was possible. If so they adopted a rather contradictory attitude inasmuch as shortly after dispensing it a naval barrage, delivered by Mishima, Tango, Okinoshina and Iwami from Hai hsi Bay to the west of Cape Jaeschke, was directed onto the urban area of Tsingtau. Backed by the land batteries, this bombardment caused great damage to the city, though one shot, apparently misaimed, struck one of the 240mm gun positions at Hui tsch’en Huk, destroying the gun and killing seven of the crew. Without a mains water supply the possibility of fire-fighting in Tsingtau was greatly reduced and several buildings were burned down, though because of the relative spaciousness of the city fires did not jump easily from building to building and so there was no major conflagration. Tapatau, the Chinese quarter, was not constructed on such generous proportions, though the relatively smaller size of the dwellings and their less robust structural strength meant they collapsed rather than burned, and it too was spared an inferno. Because the German artillery was now virtually silent the sapping work continued during daylight hours without fear of interruption, and the third parallel was completed during the day close up to the defensive ditch. Unable to effectively counter these moves the defenders, also ignoring the Japanese plea as contained in their airdropped leaflet, began putting their coastal artillery batteries out of action, which in any event, other than Hui tsch’en Huk, had proved mostly ineffective. It was clear to all that the end was not far off, and only the ferro-concrete infantry works remained as anything like effective defensive positions, though a report from one to Meyer-Waldeck ‘reflected the universal condition’:

The entire work is shot to pieces, a hill of fragments, without any defences. The entire trench system is knocked out; the redoubt still holds together, but everything else, including the explosives storage room, is destroyed. Only a single observation post is in use. I shall hold the redoubt as long as possible.

Given the impossibility of offering effective resistance to the besiegers, Meyer-Waldeck was under no illusions as to the length of time left to the defence force. Evidence for this may be adduced from his ordering Plüschow to make a getaway attempt the following day. He was to carry away papers relating to the course of the siege and several symbolic items such as the fastenings from the flagpole, as well as private letters from members of the garrison.

The attackers saw the elusive Taube take to the air the next morning, and, according to Plüschow himself, make a last circuit of Tsingtau before setting off southwards; ‘never,’ as the Japanese General Staff history put it, ‘to come back’. Though the Japanese artillery made what was to turn out to be their final efforts to shoot him down, hostile aircraft did not attempt to follow and he made good his escape towards neutral China, eventually reaching Tientsin where he was reunited with the crew of S-90. As he left his ground crew destroyed any remaining equipment, but his place over the city was soon taken by the Japanese aeroplanes who sortied in force, dropping numerous bombs onto the defenders’ positions, as a less than effective adjunct to the efforts of the artillery. As the bombardment from land and air went on the Japanese infantry began to move into their final assault positions in the third parallel. The British however, still troubled by the fire from the German machine guns, only occupied their sector with a thin outpost line. The Governor, noting the proximity of the attackers and expecting an imminent assault, ordered a general alert for the afternoon.

Kamio now had all his infantry where he wanted them, with the exception of the British contingent, and all his equipment in place. His orders for the night of 6–7 November did not however call for a general assault, but rather stipulated small scale, though aggressive, probing of the Boxer Line to test for weak spots, together with the usual artillery barrage. He emphasised flexibility and the exploitation of success. As darkness fell the sappers dug forward from the third parallel and, using mining techniques, burrowed through the counterscarp before blowing several breaches in it. This allowed the infantry direct access to the ditch without the need to leave the entrenchments. It was also discovered that the ditch in front of Infantry Works 1 and 2 differed somewhat from the saw-tooth version already noted, being a conventional channel in section. It was also found to be subject to flanking fire from the Central Fort (Infantry Work 3).

Wire that had remained intact following the previous attention of the artillery was cut or covered, allowing more or less unrestricted access within the ditch, and patrols moved across it and out onto the German side as darkness fell. At around 23:00hr a firefight broke out around Infantry Work 2 as a patrol from the 56th Infantry Regiment attempted to infiltrate and bypass it. The defenders were more alert than they had been on 1 November and sallied out to meet them. Eventually, after about an hour of fighting, the Japanese retreated and called down an artillery barrage onto the defenders for their pains.

More or less concurrently Infantry Work 3 (Central Fort) under the command of Captain Lancelle was the object of similar tactics. The results were however rather different. Engineers from the 4th Independent Battalion, preceding units from the 56th Infantry Regiment, discovered that they met with no resistance whatsoever when they began cutting two ‘roads’ through the entanglements in both the inner and outer ditches in front of the fortification. Accordingly this work was completed expeditiously, and the information on the apparent passivity of the defenders in that sector passed up the chain of command. Major General Yamada, commander of the 2nd Central Force, immediately decided to attempt an assault to take the work, but sought the sanction of Kamio before so doing. The division commander concurred, so a company sized unit under Lieutenant Nakamura Jekizo of the 56th Infantry Regiment crossed the ditch at about 01:00hr.

The plan required some courage on the part of the participants, who were all volunteers. They comprised twenty engineers and six infantry NCOs who were armed with hand grenades, whilst further infantry units, complete with mortars, stood ready in immediate support. The whole regiment was also awaiting developments and was ready to advance at a few minutes’ notice. Formed into two detachments, the raiders used ladders to climb into the ditch away from the breach and, unseen, safely reached the German parapet which they scaled before moving to reform. The plan called for a stealthy advance until the occupants of the redoubt opened up on them, and then a charge forward throwing the grenades in an attempt to disable the defenders and damage the machine guns. This procedure was modified when it became apparent that the fortification was, effectively, unguarded and Nakamura instead sent his men left and right around it to the rear (or ‘gorge’ in fortification terms) where they occupied the shelter trenches.

Detailing ten grenadiers and an NCO to resist any German potential counter-attack, he used the rest of his men to block the redoubt’s exits and then sent for reinforcements. Before these could arrive however the Japanese were detected by defence posts on the flanks of the fortification, whereby a hot fire with machine guns was opened. Several volleys of grenades stemmed that, and in the meantime the redoubt’s telephone wires were cut and access was forced into the signal room where, after the occupants were overcome, the power was cut. This plunged the whole work into darkness and prevented any further telephoning or signalling.

By this time two platoons of reinforcements had begun to arrive; half of them formed a defensive line behind the redoubt whilst the rest broke into it. It was claimed that they found the occupants in bed, but whatever the truth of the matter Lancelle immediately surrendered the work complete with its complement of about 200 men to Nakamura. It had taken forty minutes and been, to quote Burdick’s words, ‘ridiculously easy’. It was undoubtedly a famous and daring victory, and Nakamura was awarded the Order of the Golden Kite (4th class) which he undoubtedly deserved.

In practical terms however, there was now a large and growing gap in the very centre of the Boxer Line, and word quickly got to Meyer-Waldeck who ordered his reserves to counter-attack under cover of a German barrage. Whilst such a move was theoretically sound, it was, practically, almost impossible. There was simply not enough artillery left to provide an effective bombardment, and precious little manpower, particularly in comparison with that available to the attackers, to seal the breach. The effort was made, but the counter-attackers, including a contingent of Austro-Hungarian sailors landed from Kaiserin Elisabeth, were simply too weak to throw back the rapidly reinforcing Japanese.

The Boxer Line, being a linear defence, was vulnerable to being ‘rolled up’ from the flanks once breached at a given point. The Japanese having made the breach now proceeded to widen it by moving against Infantry Works 2 and 4 on either side. Both works held out for some hours, assisted by the Jaguar, the last German warship afloat, which fired away her remaining ordnance in support. The outcome however could be in no doubt and both works surrendered after about three hours of resistance. The Boxer Line was now useless, for with no defence in depth the penetration meant the route to Tsingtau was now as good as wide open. The Japanese infantry surged through the gap and began a general advance on Tsingtau and various strategic points, such as Iltis and Bismarck Hills. The batteries on the former fought the attackers for a time before surrendering, whereas the artillerymen on the latter, having fired away the last of their ammunition, set charges to destroy their guns and vacated the position at about 05:00hr. This final destruction of land-based artillery had its counterpart on the water; Jaguar, after attempting to repulse the infantry attack, had been scuttled in Kiautschou Bay.

At 06:00hr Meyer-Waldeck held a meeting at his headquarters in the Bismarck Hill Command Post where the latest information was assimilated. It had long been an unwritten rule of siege warfare that a garrison could honourably surrender following a ‘practical breach’ being made in their defences. The Japanese, using classic siege warfare methodology, had now achieved just such a breach. Whether the Governor was aware of the ‘rule’ is unknown, but he had now only two options; surrender or a fanatical ‘fight to the last man and last bullet’ scenario. Meyer-Waldeck was no fanatic. Brace put it thus:

If the governor had permitted the unequal struggle to go on his men would have lasted only a few hours longer. It would be an Alamo, and the name of the German garrison would be heralded throughout history as the heroic band of whites who stood against the yellow invasion until the last man. On the other hand the governor had with him a large part of the German commercial community of the Far East which Germany had built up with such painstaking care.

The Governor ordered the white flag hoisted on the signal station and over the German positions and composed a message to Kamio: ‘Since my defensive measures are exhausted I am now ready to enter into surrender negotiations for the now open city. […] I request you to appoint plenipotentiaries to the discussions, as well as to set time and place for the meeting of the respective plenipotentiaries. […]’ The carrier of this message was Major Georg von Kayser, adjutant to Meyer-Waldeck’s Chief-of-Staff, naval Captain Ludwig Saxer – the latter being the Governor’s appointee as German plenipotentiary. Despite Barnardiston’s contention that all firing ceased at 07:00hr Kayser had difficulty getting through the lines in safety, but he was eventually allowed to proceed under his flag of truce to the village of Tungwutschiatsun, some 4km behind the Japanese front line, more or less opposite the celebrated Central Fort (Infantry Work 3).81 It was agreed that a general armistice would come into play immediately, and that formal negotiations for the capitulation would commence that afternoon at 16:00hr in Moltke Barracks.


The Battle of Shanghai, 1937

Thirteen years after Commodore Matthew Perry steamed into Tokyo Bay, a nationalist rebellion overthrew the conservative Tokugawa Shogunate, installed the Emperor Meiji in power, and implemented a program of sweeping national reform. In an act no less stunning than the revolution itself, nearly all of the former ruling families voluntarily surrendered their power to the emperor, declaring, “We therefore reverently offer up all our feudal possessions so that a uniform rule may prevail throughout the Empire. Thus, the country will be able to rank equally with the other nations of the world.”

By 1895, Japan achieved its goal. Japanese armies had seized Korea and defeated China in the first Sino-Japanese war. Ten years later, Japan sent shockwaves through the world when the Imperial Japanese Navy (IJN) sank or destroyed most of the Imperial Russian Fleet in the two-day battle of Tsushima (27–28 May 1905)—at the time the most significant naval action since Trafalgar. Japan’s warships were newer with more modern guns and range-finders than those of their Russian opponents. Whereas the Russians used wireless communication sets produced in Germany, the IJN used sets that had been manufactured in Japan.

To the Western powers, Japan’s meteoric rise to strategic prominence in northeast Asia demonstrated that the Japanese were fundamentally different from the Chinese, and, indeed, from all the peoples of the Far East. John Pershing, who served as the U.S. military attaché in Tokyo during the Russo-Japanese War, described the Japanese he saw as “a strong, virile, aggressive race, with an ambition and determination that will carry them very far in the contest of nations for power.”

Britain lost no time in securing an alliance with Imperial Japan to contain Russian expansion, as well as to compensate for Britain’s strategic dilemma of “imperial overstretch.” Under the terms of the original 1902 Anglo-Japanese Treaty, Japan agreed to secure British commercial and territorial interests in the Far East, allowing Britain’s Royal Navy to concentrate in the North Sea for its future war with Germany while Britain recognized Japan’s de facto conquest of Korea. The treaty lapsed in 1923 after the signing of the Washington Naval Treaty, but in the run-up to World War I, the treaty placed Japan on the winning side, a position that rewarded Japan with control of Germany’s Chinese and Pacific territories—lucrative gains for Japan’s modest role in the war. Far more important, Japan gained Western recognition as a great power.

For a nation that rose from obscurity to great power in less than fifty years, the military triumphs over China and Russia were intoxicating. When an incident occurred on the night of 7 July 1937 involving Japanese and Chinese troops near the Lugou bridge, known internationally as the Marco Polo bridge, the temptation for Tokyo to conquer yet again was too strong to resist.

Major General Ishiwara Kanji, the dynamic Japanese army officer who masterminded the Mukden Incident (the alleged Chinese bombing of a Japanese railway) that wrested control of Manchuria away from Nationalist China in 1931, was cautious. He urged restraint, telling the general staff of the Kwantung Army, “If we act now against China, the sky will fall in. Let’s keep the incident from developing further and have the local command settle the issue.”

Concurrently, Emperor Hirohito approved the mobilization of five Japanese divisions for a campaign against the Chinese that his minister of war claimed “would be finished up within two to three months.” When Ishiwara heard the news, he was more pessimistic. He told his colleagues, “We may find ourselves with a full-scale war on our hands. The result would be the same sort of disaster which overtook Napoleon in Spain—a slow sinking into the deepest sort of bog.”

Chiang Kai-shek, the leader of Nationalist China, responded by ordering the Chinese army to attack the Japanese garrison in Shanghai, eventually sending 600,000 Nationalist Chinese troops to fight for the city of 3.3 million people, which in 1937 was the fifth largest port in the world and Asia’s financial capital. Chiang was well aware that Shanghai was a city where fabulously rich Chinese and European “colonials” lived like kings inside a special “international zone” next to millions of impoverished Chinese. Under the circumstances, Chiang decided to make the fight for Shanghai in Zhabei, the poor, industrial section of the city on the north side of the international zone, hoping to produce an incident that would rally the Europeans and Americans to his side in the war with Japan.

For reasons that seem obtuse today, Japan’s military and political leaders believed control of China, a nation torn by civil war with hundreds of millions living in poverty, would add to Japan’s margin of victory in future wars. Japan’s national military and political leaders equated industrialization and access to markets and resources with the control of territory and peoples.


The appearance of Commodore Matthew Perry’s “Black Ships” in Japanese waters in 1853 forced the Japanese to face a disconcerting reality: Japan’s military power and the economic strength to support itself were inferior to those of the West. Without a rapid transformation into a modern state, Japan itself might not survive contact with the West.

To modernize and catch up with the Western nations, Japan embraced “raw, unbridled capitalism.” Japan may not have had Calvinism and the Protestant culture that launched Great Britain, Germany, the Netherlands, and the United States on the economic fast track to prosperity and power, but Japan possessed the time-honored values of integrity and hard work, as well as a culturally and racially homogenous population with a deeply ingrained sense of duty and a collective obligation in all aspects of life.

The same cultural values of energy and intelligence that underpinned Japanese economic modernization also catapulted the Imperial Japanese Army (IJA) into modernization. As always, revolutions in military affairs involve much more than new technology. For a national military strategy to be effective, its goals and tenets must align in harmony with the respective nation’s cultural norms, geographic position, and economic potential.

In Japan, the ruling elites looked carefully at the various ways in which other great powers sought to harmonize these factors—culture, geography, and economy—within the framework of national military strategy. Thus, the Japanese embraced the American model for industrialization, the British model for ship-building and maritime power, and the German military model for land power. But it was Japan’s embrace of Otto von Bismarck’s Prussian-German concept of “rich country, strong military” that most profoundly influenced Japanese thinking about national security.

In time, two competing strategic views emerged in Japanese thinking about security and commercial trade, a continental approach and a maritime approach. The continental IJA faction argued for Japanese expansion to the “north,” through Korea and into Manchuria, Mongolia, and, eventually, eastern Siberia. The maritime IJN faction urged expansion to the “south,” the soft underbelly of Asia and the Pacific basin. In the first decades of the twentieth century, the IJA overpowered the IJN. Eventually, the IJA drove Japan into war first with China, then tsarist Russia.

In 1924, an important figure in Japanese military history, General Ugaki Kazushige, stepped into the middle of this contest for the hearts and minds of the Japanese people. Ugaki served as minister of war from 1924 to 1927 and again from 1929 to 1931. He lived and studied abroad, serving twice in Germany as Japan’s military attaché, from 1902 to 1904 and again from 1906 to 1907. During this service, Ugaki developed a strong affinity for the German people, their patriotic spirit, and their cultural values of integrity and hard work. On completing his service in Germany, he envisioned Japan’s future strategic cooperation with Germany “as a means to keep Russia down in case our force will not be [strong] enough. I would rather prefer a Japanese-German alliance than a Japanese-Chinese one. Suppressing Russia and China from two sides, East Asia will come under our hegemony, Western Europe under Germany’s.”

Believing control of Manchuria and, if possible, eastern Siberia was essential to Japan’s long-term security and prosperity, Ugaki made a detailed examination of the massed Japanese bayonet charges and frontal assaults against Russia’s fixed fortifications during the Russo-Japanese War. He was repelled by the Japanese losses and found the Japanese generals’ callous disregard for human life distasteful.

Ugaki also grasped the most important lesson of the confrontation with Russia: Japan’s victory was more the result of Russian incompetence and the inability of Russian forces to maneuver against the Japanese than of Japanese superiority. General Kodama Gentaro, the IJA chief of staff during the Russo-Japanese War, confirmed this insight, declaring, “This greatly simplified matters for us. It also made the result of battle far greater than we had anticipated.” Ugaki concluded that the views of many younger IJA officers were correct. In the future, the IJA would need the mobility and firepower to conduct sweeping flank attacks, enveloping or encircling the Russian enemy.

When Imperial Japan was invited by President Woodrow Wilson in 1918 to join the Western allies in a joint attack on the new Bolshevik state, General Ugaki, now deputy chief of the Imperial General Staff, was assigned to plan Japan’s military intervention. Ugaki welcomed the opportunity and drafted the IJA’s plans for the seizure and occupation of eastern Siberia. He saw intervention in Russia as an immense opportunity to reinvigorate Japan’s northern strategy and expand its foothold on the Asian landmass. His far-reaching plans utilized the railways all the way to Lake Baikal and recommended their expanded use to move Japanese forces farther west if the opportunity arose. However, in Russia, the IJA encountered a new Russian enemy in the form of Bolshevik cavalry forces—mobile guerrilla armies that operated over thousands of square miles, often under horrible weather conditions.

In four years of hard fighting, the Bolshevik armies scored few military triumphs, but they did wear down the IJA and push Japan’s limited industrial capacity and economy to the point of exhaustion, imparting a strategic lesson Ugaki would not forget: Japan’s army and its supporting scientific-industrial base were not prepared to meet the requirements of modern warfare. Ugaki resolved to change this condition by infusing the Meiji-era IJA with new thinking, a new organization, and new forms of armored mobility, firepower, and aircraft. The question for Ugaki was how to finance and implement his plan.

After the IJA’s four-year intervention in Siberia ended, the politics of economic stringency confronted the army with a severely constrained defense budget. A new internal debate raged regarding how, where, and against whom to fight. Once again, Ugaki’s eyes fixed on Manchuria, not the Pacific, and his relations with the admirals of the IJN quickly deteriorated. He viewed Japan’s enormous investment in naval power as a diversion of resources the IJA would need for the unavoidable collision that would decide Japan’s strategic future: a land war with the Soviet Union.

As chief of staff, Ugaki embodied the fight for change inside the Imperial Japanese Army, which was always intertwined with its contest with the IJN for resources. Ugaki’s faction consisted of the so-called revisionists, IJA officers who believed strongly that future wars would commit the army to protracted conflicts against more advanced Soviet and Western opponents, particularly the Western colonial powers in Asia. The revisionists believed that modern armaments and new organizations for combat, not numbers or morale, were the keys to victory in future warfare. Under Ugaki’s leadership, the revisionist program reorganized the IJA into smaller, triangular (three-regiment) divisions and introduced new technology in the form of tanks, mobile artillery, and aircraft—all paid for with savings from an overall reduction in IJA manpower.

On the other side of the debate were the IJA traditionalists, officers convinced that numerically large conscript forces could always compensate for deficiencies in weaponry and technology so long as they were imbued with strong combat spirit. The traditionalists argued that future wars would look like Japan’s first and brief war with China in 1898 or the more intense Russo-Japanese War. Since Japanese troops were always outnumbered by Bolshevik insurgent forces in Siberia, the traditionalists cited the IJA experience in Siberia as evidence for the importance of numbers rather than mobility and firepower. Major General Horike Kazumaro, who opposed Ugaki, expressed the traditionalist view, asserting, “We made studies, but putting it bluntly, Japan’s industrial capacity at that time could not carry out all these things we’ve spoken of, like mechanization of the army, the development of tanks, and the use of aircraft in group formations. If we overstrained in trying to do it, it would have entailed a third and a fourth force reduction, and the army would have been broken up.” Though the IJA’s share of the defense budget fell from 18.8 percent in 1919 to 16.2 percent in 1922, the traditionalists made sure the IJA grew smaller, but they also resisted change in its organization, equipment, or thinking about warfare.

However, Ugaki’s fortunes changed in 1923, when his patron and mentor Tanaka Giichi, now Japan’s prime minster, appointed Ugaki as minister of defense. He was finally in a position to drag the IJA through a second revolution in Japanese military affairs. The essential features of Ugaki’s reform package were to:

Reduce the army budget. Reductions included the disestablishment of four infantry divisions to offset the costs the Japanese government incurred during the four-year Siberian intervention and the Great Kanto earthquake, an approximately 7.9 magnitude quake that transformed Tokyo into a blackened wasteland of death and destruction. Discharged officers were sent to middle schools and high schools to become teachers.

Retire general officers opposing reforms. Ukagi removed the IJA’s top eleven generals, prompting a Japanese journalist to record, “There was no way to treat these stone heads other than to replace them.” Ugaki’s allies were thus put into key positions in both the army general staff and national command structure during 1925.

Change the force structure. Ugaki set forth his program to reorganize Japanese divisions from square divisions into triangular divisions. The square division was downsized by removing one regiment and skipping the brigade as an intermediate level of command between regiment and division, thereby achieving more savings in manpower without a loss of fighting power. The smaller division retained the same number of supporting arms—artillery, engineers, and related elements—leaving the formation just as effective, but more mobile and less vulnerable to concentrated enemy fires.

Modernize the force. Ugaki secured a reduction in the IJA budget to 12.4 percent in 1927 that partially funded the purchase and eventually the development of new tanks, artillery, aircraft, and automatic weapons for the IJA. He established the bureau of supplies and equipment to supervise the IJA’s modernization.

In the two years after Ugaki left office in 1927, many of his reforms were predictably delayed or halted entirely. Reactionaries in the senior ranks of the IJA reasserted their influence to slow or halt Ugaki’s efforts to modernize the army at the expense of the numbers of men serving in the IJA. Simultaneously, interservice rivalry between the army and the navy worsened, further poisoning the contest for resources and bureaucratic dominance. In later years recalling the events of his two terms as minister of war, Ugaki said, “I tried to seize the initiative, but the tendency of the army was to go in the opposite direction.”

It would take another nine years for Major General Ishiwara Kanji, another Germanophile in Japanese uniform, to push through Ugaki’s reform program. In 1936, Ishiwara succeeded in persuading the IJA leadership to complete Ugaki’s reforms by reorganizing all of the IJA’s divisions into smaller triangular formations equipped with more modern weapons. The commitment of resources and manpower to Japan’s invasion of China in 1937 postponed Ishiwara’s implementation plan until 1940, too late for the IJA’s decisive battle with the Soviet armed forces in 1939 at Nomonhan. Dismayed by Tokyo’s rush to war with China, Ishiwara cabled the minister of war with the following message concerning the decision to invade China proper: “Tell the Prime Minister that in the two thousand years of our history no man will have done more to destroy Japan than he has by his indecisiveness in this crisis.”

None of the IJA’s infantry divisions that fought in the battle of Shanghai were reorganized until after 1940, long after the tactical and operational advantages of the smaller, triangular division over the large, unwieldy World War I square division were undeniable. Though Ugaki’s reforms did not succeed in transforming the IJA divisions into smaller, more mobile formations, Ugaki did succeed in equipping Japanese divisions for the battle of Shanghai with twice the number of fighting men, three times the number of rifles, and double the number of crew-served weapons (machine guns and mortars) and artillery than were present in the Chinese divisions that opposed them. Despite the setbacks to modernization, thanks to Ugaki, the IJA deployed two tank brigades plus one additional tank battalion to fight in the battle. As it turned out, the presence of just twenty-four tanks inside each Japanese infantry division was decisive in the battle for Shanghai.


Tokyo’s deliberate escalation of the dispute with China over the Marco Polo bridge incident into full-scale war on 7 July 1937 hurled Japan’s Kwantung Army into action. In mid-August, Japanese forces struck south from Manchuria to seize Beijing and Chahar. General Chiang Kai-shek suddenly confronted a difficult situation. Nationalist Chinese military strength in the north was thin, and China was still weak from years of civil war between his Nationalist forces and Mao Zedong’s Communists.

Confronted with a similar strategic dilemma in 1933, Chiang opted to consolidate his strength in the south and concentrate China’s military power against the Chinese Communist Party, at the time a much more serious threat to Nationalist China than Japan. Under the terms of the Tanggu truce that ended Japanese hostilities with China, the price paid for this temporary retreat was humiliating but small. The humiliation entailed the formal recognition of Japan’s conquest of Manchuria and the loss of Rehe, a portion of Inner Mongolia controlled by China. Yet Manchuria and Mongolia were not historically part of China, and the regions had few Chinese inhabitants. Chiang chose to husband his resources and defend what he considered to be most important: China’s core Han population. Shanghai was an entirely different matter.

Chiang could play for time in the north, but Shanghai, only forty miles from Nanjing, was the capital of Nationalist China, meaning it could not be surrendered without a fight. In mid-July, Chiang ordered General Zhang Zhizhong, a forty-two-year-old political confidant and commander of Chinese forces in Shanghai and the 9th Nationalist Chinese Army Group, to prepare his troops for an attack to drive the Japanese garrison out of Shanghai.

Under the terms of the Tanggu truce that ended China’s previous hostilities with Japan, a demilitarized buffer zone was established between China and the city of Shanghai. The Chinese military presence in the city was restricted to a Peace Preservation Corps, essentially a large paramilitary police force, but Japan was allowed to maintain a military garrison in Shanghai together with the French, British, and Germans. As soon as war broke out in July, both sides began moving additional men and equipment into Shanghai: the Chinese into the ranks of Shanghai’s Peace Preservation Corps and the Japanese into their fortified Shanghai garrison.

Shanghai and Its Surroundings

The Japanese commander tasked with the mission of seizing and securing Shanghai was sixty-year-old Matsui Iwane, a distinguished Russo-Japanese War veteran of samurai heritage who had retired from active duty just four years earlier in 1933. Matsui had served as IJA chief of intelligence under Ugaki and supported his military reforms on the German military model.

Like many officers of his generation, Matsui believed Japan’s mission was to liberate Asia from Western colonial rule. Early in his military career, Matsui was strongly sympathetic to China’s nationalist movement. In his younger years, he even befriended Sun Yat-sen, the leader of China’s republican revolution who led the movement to overthrow China’s last emperor in 1911. In 1937, in a sad turn of events for Matsui, he was tasked to attack a country he had once hoped to liberate.

On reporting for duty in Tokyo, Matsui was presented with a new operational plan to invade and occupy northern China with nine to fourteen divisions while three divisions seized Shanghai and two more divisions landed in Hangzhou Bay, south of Shanghai. Once Shanghai was secured, the plan directed Matsui to advance on Nanjing and compel the surrender of China’s Nationalist government. In many ways, the plan was a Japanese version of Germany’s Schlieffen plan from World War I, aimed at knocking out China quickly, then turning north to refocus Japanese military power on the real threat to Japan: the Soviet Union. Moving beyond Nanjing to secure China’s vast interior was never mentioned or anticipated in the original plan.

Until General Matsui and the Shanghai Expeditionary Force (SEF) arrived, the 3rd IJN Fleet under the command of fifty-four-year-old Rear Admiral Hasegawa Kiyoshi would fight the Chinese on land, on sea, and in the air. Hasegawa was a veteran of the Russo-Japanese War and had served on Admiral Togo Heihachiro’s flagship, the battleship Mikasa, when the IJN destroyed tsarist Russia’s fleet in the Straits of Tsushima. He also served as Japan’s naval attaché in Washington and in 1945 provided the final report on the condition of the IJN to Emperor Hirohito, persuading him to surrender.

For the IJA to succeed in its fight for Shanghai, Admiral Hasegawa knew that Japanese positions ashore in the city of Shanghai must be retained at all costs. On his own initiative, Hasegawa moved quickly to reinforce the small Shanghai Special Naval Landing Force (SNLF)—literally, Japanese sailors taken off ships, issued rifles, and used as infantry on land with additional sailors. In the ten days before the IJA arrived, Hasegawa expanded the SNLF to ten thousand sailors. They would provide an important margin of victory in the battle by defending Japanese positions ashore against the attacking Chinese Nationalist army until the IJA arrived.

Anticipating the decisive role that Japanese naval gunfire and airpower would play, Hasegawa also requested additional naval power. Tokyo responded by sending the 8th Cruiser Division and 1st Destroyer Squadron to reinforce the fleet. By 11 August, Admiral Hasegawa had a fleet of thirty warships in the 3rd IJN Fleet, including the 16th Destroyer and the 11th Battleship divisions plus two aircraft carriers.

Three ships of the 11th Battleship Division were anchored in Shanghai harbor to provide fire support to Japanese forces in the city. The rest were located up and down the Yangtze River, where they operated without interference from the Chinese. The Nationalist Chinese had a fleet of fifty-nine ships and twelve patrol boats equipped with torpedoes designed for riverine or coastal operations, but without the ability to mine the approaches to Shanghai harbor and the Yangtze River, the vessels were of little use in this battle.

Knowing that all of the IJN’s forces could not operate from Shanghai harbor, Hasegawa directed the Kamoi, 1st Carrier Division, 1st Combined Air Unit, 22nd Air Unit, and the 12th Sea Plane Division to set up bases of operations in the vicinity of the Shengsi Islands some thirty kilometers off the coast of Shanghai. These units added a total of thirty-eight Type 96 attack/reconnaissance planes and twenty Type 95 fighters, all vastly superior in range and striking power to China’s antiquated and numerically inferior air force.

On paper, the Chinese National Revolutionary Army fielded a force of 1.75 million soldiers. As is often the case in war, the numbers were impressive but misleading. At most, perhaps 300,000 Chinese soldiers were trained and organized in any meaningful way.

Chiang Kai-shek regarded the troops that his German advisers trained as his best. Like the Japanese army officers, Chiang Kai-shek held the German military in very high esteem. His closest military advisers were Germans. One of the most famous, Colonel General Hans von Seeckt, travelled to China and provided him with a white paper that was instrumental in persuading Chiang to stand up an elite force of 80,000 men under German advisers. In August 1937, these troops were better trained and equipped than the rest of the Nationalist Chinese forces, but they were still in an early stage of development and not ready to take on the IJA in pitched battle.


Japanese troops reaching the destroyed North Station in downtown Shanghai

General Zhang, China’s 9th Army Group commander, was a realist. He knew his troops were outgunned and outranged in every key category of military capability. Zhang’s most promising margin of victory depended on how effectively he could exploit the Nationalist Chinese Army’s numerical superiority to drive Shanghai’s Japanese garrison into the sea before the IJA could land in force. By 9 August, Zhang had assembled 4 infantry divisions, a separate infantry brigade, 2 artillery regiments with 400 to 500 guns, 2 heavy mortar battalions, 2 anti-tank gun batteries, and 1 light tank battalion with 30 to 40 light tanks—a force of roughly 50,000 for the mission.

Zhang’s scheme of maneuver was not complex. Like many generals in Chinese history, he made virtue of necessity. He decided to exploit China’s one true advantage over the Japanese: numbers of soldiers. Zhang reasoned that if he could achieve surprise, he could concentrate overwhelming numbers of Chinese troops against the Japanese defenders, eventually driving the retreating Japanese into the Huangpu River. His plan called for an assault by 20,000 troops or two divisions. One division would strike at the IJA headquarters near Zhabei, the poor industrial section of Shanghai. The other division would attack the IJA headquarters located inside a textile mill. Once these objectives were taken, Zhang planned to deploy additional Chinese forces along the coast to defeat Japanese attempts to land reinforcements.

Unfortunately for Zhang, his initial attack did not go as planned. During the morning of 13 August, Chinese troops in the Shanghai Peace Preservation Corps opened fire on members of the Japanese garrison, alerting the Japanese to impending action. Realizing that the element of surprise was lost, Chiang directed Zhang to begin his deliberate assault on the Japanese targets as soon as possible. For Zhang, that meant the next day. Zhang knew that a successful Chinese attack depended heavily on surprise, but he followed orders.

When the two Chinese divisions attacked their objectives on 14 August, they made a shocking discovery. Not only had the Japanese forces been alerted, but also the two Japanese headquarters were protected by concrete defensive barriers that were impervious to even the largest Chinese artillery (150-millimeter howitzers). Additionally, Japanese armored cars and machine guns were placed along the streets and corridors where Chinese attacks were expected to pass, and Japanese naval gunfire pulverized Chinese troops in the areas surrounding the Japanese strongpoints.

Undeterred by the failure of the first attacks, the Chinese generals sent their men into action, shoulder to shoulder, with fixed bayonets. Predictably, losses were catastrophic. Chinese flesh and bone performed poorly against Japanese steel, concrete, and machine guns. Even more disappointing, the Chinese tanks of British, French, and Italian manufacture lacked the armored protection to withstand fire from Japanese heavy weapons and provided practically no offensive capability at all. Fire from Japanese heavy machine guns penetrated the light armor and killed or wounded the Chinese tank crews.

Frustrated and desperate to destroy the Japanese strongholds in the city, Zhang decided to throw in two of Nationalist China’s German-trained divisions, the 87th and the 88th, in an operation he named Iron Fist. On 17 August, these elite infantry divisions attacked the Japanese strongholds using the German infiltration tactics they had practiced under the supervision of German officers on contract to train Chiang’s army.

Despite the demoralizing wall of fire that greeted them, the heroic Chinese soldiers fought their way to the forward edge of the Japanese defenses. It made no difference. The attacking Chinese troops simply lacked the firepower and explosives to penetrate the defensive works, and their artillery and machine gun support was insufficient. Once more, the Japanese sailors of the SNLF held their ground thanks to the concrete defensive works and the fleet’s accurate, devastating naval gunfire.

In the interim, Nationalist Chinese air and naval forces struck the Japanese ships in Shanghai harbor. On 14 August, forty Chinese aircraft flew a total of two hundred sorties, attacking Japanese ships as well as troops in the city. Once again, the attacks were ineffective. Chinese river patrol craft launched a successful torpedo attack against the battleship Izumo, scoring two hits, but did not do serious damage to it. Bombs meant for the Japanese fleet fell instead on a British warship and on Shanghai’s international zone, accidentally injuring or killing three thousand civilians inside an area now brimming with more than a million people seeking refuge from the fighting. Japanese forces ashore sustained no losses at all.

In the first forty-eight hours of the fight for Shanghai, Japanese airpower fought the weather, not the Chinese. When the weather cleared on 16 August, the 3rd Air Fleet launched air strikes from Taipei, Cheju-do Island, Shengsi Islands, and its two carriers, Hosho and Ryujo, hit Chinese troop concentrations and airfields from Shanghai to Nanjing. After decrypting the Nationalist Chinese military’s encoded messages on 19 August, the Japanese established air superiority over Shanghai. By the time Matsui’s SEF of 80,000 men arrived on 23 August, Admiral Hasegawa’s actions had retained Japan’s foothold ashore, securing Japan’s margin of victory. The scales were tipping in favor of the IJA.


General Matsui’s first contingent comprised two divisions, the 3rd and 11th Infantry Divisions. To augment the capabilities of these two divisions, Matsui was given additional machine gun, tank, and mortar battalions, a heavy field artillery regiment, and a siege artillery battalion, adding about 100 tanks and 300 to 400 artillery systems of different calibers to the SEF for a total of 80,000 troops. Separate radio platoons were attached to major regiments and divisions, giving the SEF a significant communication advantage over the Chinese forces, which were equipped with comparatively few radios. The lack of communication capabilities became a chronic weakness. Chinese defensive operations could not keep up with the frequent attacks launched by the Japanese let alone effectively monitor the actions of their own forces.

A few days later, an IJA fighter squadron from the Provisional Air Group operating in north China was added to reinforce Japanese airpower ashore. By 10 September, China’s small air force was largely out of action, and Japanese forces enjoyed air supremacy for the rest of the campaign.

Matsui’s task, however, was by no means easy. Geography still favored the Chinese defender. The maze of waterways and buildings inside Shanghai, together with the swamps, rivers, and lakes on the city’s outskirts, severely restricted the movement of ground forces on both sides. If skillfully used, these waterways would favor the Chinese defender. On the north side of Shanghai is the great Yangtze River, five to ten miles wide and sixty-five to eighty feet deep. In the center of the peninsula is Suzhou Creek, an eastward flowing stream that makes a sharp northward turn into the Huangpu (formerly called the Whampoa) River, eventually converging with the Yangtze. The source of the Suzhou Creek is an area of swamps, and farther west is Taihu Lake, a body of water larger than the modern city of Shanghai itself (869 square miles). To the south of Shanghai is Hangzhou Bay, an elongated thirty-six-mile-wide inlet that gradually narrows as it reaches the Qiantang River and the city of Hangzhou.

Aside from his age, Matsui was hardly the typical army four-star general. He was remarkably fit with an exceptionally sharp mind and no shortage of personal physical courage. More important, Matsui would do early in the campaign what seemed impossible given the serious interservice rivalry that afflicted relations between the IJA and the IJN. He would make peace with the Japanese admirals.

In a discussion on 21 August aboard Admiral Hasegawa’s flagship, Matsui listened carefully to the recommendations of his own staff, as well as to those of the Japanese naval officers already engaged in the battle for Shanghai. Rear Admiral Nagumo Chuichi urged Matsui to conduct two landings, which would force the Chinese defenders to fight in two directions simultaneously. (Nagumo would later command the IJN’s 1st Air Fleet at Pearl Harbor and Midway.)

Ultimately, Matsui overruled the IJA staff officers who urged him to make a single landing in one location in favor of Nagumo’s plan for two. Matsui decided to conduct the landings in two echelons at two different locations roughly ten miles apart. The northernmost landing would place the 11th Infantry Division in the vicinity of Chuanshakou, while the 3rd Infantry Division landed farther south near Wusong. The Japanese conducted amphibious landings along a forty-kilometer front, from Liuhe northeast of the Shanghai metropolitan area to the Shanghai docks on the Yangtze.

Matsui instructed his commanders to execute a double envelopment using the 3rd Infantry Division to capture the city of Wusong while the 11th Infantry Division pushed inward to the southwest and capture Luodian. The capture of that city—a transportation center connecting Baoshan, downtown Shanghai, and several other towns with highways—would facilitate the encirclement Matsui wanted by cutting off Chinese forces inside an area of about approximately fifty square miles. The encirclement was part of Matsui’s larger strategic goal of forcing the Chinese 15th Army Group to withdraw eight miles from the coast. If successful, the withdrawal would expose the left flank of the Chinese 9th Army Group, which was still fighting in metropolitan Shanghai, to Japanese assault and, potentially, encirclement.

To support General Matsui’s plan, Admiral Hasegawa directed the entire 3rd Air Fleet to thoroughly reconnoiter the area and then provide the attacking Japanese divisions with close air support. When the IJA formations began landing in waves on the Chinese coast, the entire might of the Japanese fleet would assemble to systematically pulverize Chinese troops and defense works.

In view of Chinese weaknesses in airpower, artillery, and tanks, the Chinese had little choice but to compensate with numbers. By the time the Japanese landings started, Chiang had moved the Nationalist Chinese 15th Army Group, a force of 150,000–200,000 men under the command of General Luo Zhuoying, into hastily prepared defensive positions along the coast. Luo’s mission was to stop the Japanese landings. To do so, he had 17 infantry divisions, a separate infantry brigade, and a separate artillery regiment with 600 to 800 artillery systems. Chinese infantry divisions were generally 10,000 strong, or half the size of the 22,000-man Japanese square divisions.

The landings began on 23 August with the Japanese SEF coming ashore near the coastal towns of Liuhe, Wusong, and Chuanshakou. Resistance near or on the beaches was sporadic, but IJA 3rd and 11th Infantry Divisions encountered tough opposition from Chinese forces as they advanced inland to capture towns and villages. Neighborhoods and buildings that the IJA seized during the day with the aid of Japanese naval gunfire were lost to ruthless Chinese counterattacks at night. Even with the systematic bombardment of Chinese defensive positions and assembly areas, the fighting was harder than Matsui had anticipated, and he went ashore to find out why.

Japan’s Shanghai Expeditionary Force Lands in China, 22 August–23 September 1937

Upon arriving at the front, Matsui discovered that his general staff had directed the two attacking divisions to leave behind their tanks and artillery. Japanese general staff officers assumed the soft sand and ground on the coast plus the network of canals and streams surrounding Shanghai would not support the effective employment of tanks and artillery. Japanese infantry attacks bogged down due to the absence of Japanese firepower and mobility.

Matsui was deeply angered by his staff’s omission, and because of the lack of mobile armored firepower, five days would pass before the 11th Infantry Division would capture Luodian and the 3rd Infantry Division would capture Wusong. However, far more important for the larger Japanese campaign was the fact that without the tanks and artillery, the SEF opportunity to envelop the Chinese 15th Army Group was irretrievably lost.

The town of Baoshan finally fell on 4 September. For its Chinese defenders, the battle was a fight to the death. The few Chinese units that survived that battle escaped to establish new defensive positions in and around Luodian, barely sixteen miles from the center of Shanghai. Taking Luodian would now require a deliberate assault, something Matsui had hoped to avoid.

To defend Luodian, Chiang concentrated 300,000 troops, including the survivors of Baoshan, to stop 100,000 Japanese soldiers supported by tanks, artillery, aircraft, and naval gunfire. The Chinese were too poorly equipped and trained to halt the IJA for long, but as the battle progressed, the Chinese found innovative ways to slow or derail the Japanese advance.

Under cover of darkness, Chinese soldiers mined the roads leading into their defensive positions around Luodian and launched raids to isolate and cut off Japanese outposts. To reduce their losses from the massive Japanese artillery and air strikes, the Chinese kept the forward lines lightly manned, a tactic the Germans had employed against the British and French armies in the last two years of World War I. Then, when the Japanese infantry closed in, the Chinese would appear to spring up from nowhere and attack them at close range. These tactics worked well for the Chinese, who used them repeatedly.

When one defensive line was wiped out and taken by the Japanese, the Chinese would fall back to another, over and over, until few Chinese soldiers were left alive. These diehard tactics had a profound effect. On 18 September, the Japanese attack was halted, and a thousand Japanese soldiers were killed or wounded. It is no accident that in Chinese history, the battle for Luodian is referred to as the “grinding mill of flesh and blood.”

For the Japanese to regain momentum, Matsui concluded that more drastic measures were needed. He recalled Japanese air forces from bombing runs on Chinese airfields and supply columns and ordered them to focus on Chinese forces holding up the Japanese attacks. On 10 September, Matsui launched a massive air assault with aircraft from the IJN 3rd Fleet and 2nd Combined Air Unit to support the IJA’s infantry divisions. With the relentless and massive application of airpower, the SEF was finally able to dislodge the determined Chinese from their last defensive positions in Luodian before the end of September.

By this point, the IJA had dramatically improved its coordinated use of aircraft, artillery, tanks, and infantry in daylight attacks. The result was not quite a blitzkrieg, but it was close enough in the fighting with the infantry-centric Chinese. For the Chinese, the experience was sheer terror.

Imperial Japanese Army attacks were planned and executed with precision; they began at first light with massive air attacks supported by observers in aircraft who identified and targeted Chinese positions for the artillery and naval gunfire. When smoke appeared on the battlefield, IJA infantry began advancing with tanks in the lead while fighters flew forward to search for and attack Chinese reinforcements racing to the sector under attack. This was a pattern that the IJA would repeat with great effectiveness.

Nevertheless, Japanese losses at this point in the fight for Shanghai were much higher than Tokyo anticipated; 8,000 casualties, or nearly 30 percent of Japanese losses during the entire battle for the city. Even with the reinforcement of 40,000 troops from the IJA’s 13th and 9th Divisions, the battle for Shanghai was beginning to resemble a bloody stalemate on the World War I model. The lack of mobile armored firepower inside the Japanese forces meant that Matsui’s troops could not move and concentrate quickly enough to envelop and trap the Nationalist Chinese forces as he had originally intended.

Determined to regain the initiative from the Japanese no matter the cost, Chiang now committed all of his military commanders to the battle for Shanghai, directing them to fight literally to the last Chinese soldier. On 21 September, he reorganized his forces into area defense formations. Chiang assigned specific areas to Chinese armies and army groups with defined boundaries that established command responsibility for what contemporary American military leaders call “defend to retain” missions. The river defense forces, Nanjing’s capital garrison forces, and the Nationalist Chinese 23rd Army Group (as a separate command) were directed to establish an area defense south of Hangzhou Bay, far from Nanjing. Seventy Nationalist Chinese infantry divisions would eventually be involved in the operation, but only the forces designated “Right Wing Forces,” “Central Forces,” and “Left Wing Forces” would be directly engaged in the battle of Shanghai.

Generalissimo Chiang Kai-shek’s 21 September Plan for the Defense of Shanghai and Its Surroundings

Chiang’s arrangement was problematic. His reorganization inserted an extra layer of command and control that slowed decisionmaking, restricted the flow of information, and reportedly caused some unnecessary confusion. Moreover, instead of utilizing all nine Chinese army groups to attack and overwhelm the IJA, he initially committed only two (and later, a third) army groups to directly engage the IJA at any one time.

Chiang later said the new defensive scheme was necessary to accommodate the growing number of Chinese troops committed to defend Shanghai and its surroundings. While this may have been true, there was another reason. The IJA’s war of ruthless extermination against its enemy cultivated extreme hatred in the Chinese, evoking a particularly visceral response from the Chinese soldier. Chiang wanted to put this hatred of the invader to good use.

Nationalist Chinese soldiers now killed Japanese prisoners of war as readily as the Japanese killed Chinese prisoners of war. Major General Wei Li-huang, the commander of the 67th Division, Chinese 15th Army Group, noted, “It’s impossible to have a prisoner delivered to headquarters although we pay from 50 to 100 yuan upon delivery, and there are severe punishments for not doing so. The soldiers say that the prisoners die along the way.” Chiang could defend Shanghai and its surroundings to the last Chinese soldier because the “war of armies” was becoming a war between two peoples.


A National Revolutionary Army machine gun nest in Shanghai


On 30 September, the IJA and IJN signed Memorandum 519, which set down the agreed responsibilities of army and navy air forces. This memorandum was crucial for attacking Japanese ground forces in gaining access to air support from both the services. Under the terms of the agreement, the 2nd Combined Air Unit, a mixed formation in northern China consisting primarily of IJN aircraft, was tasked to support the 3rd and 9th Infantry Divisions. The 1st Combined Air Unit halted raids against Nanjing and instead agreed to coordinate with the IJA forces in Shanghai on a daily basis until 27 October or whenever the Shanghai phase of the operation ended.

The same day the memorandum was signed, Matsui set in motion an attack southward from Liuhang through Wusong Creek, toward Zoumaotang Creek. He hoped the attackers would quickly seize two stone bridges west of Dachang being defended by the Nationalist Chinese 33rd and 18th Divisions.

The Japanese employed the 101st, 3rd, and 9th Infantry Divisions, with the 11th Infantry Division covering the right flank from possible spoiling attacks by the Chinese 15th Army Group. The IJA 13th Infantry Division was held in reserve. Chinese accounts also mention the presence of Taiwanese and Manchu units intermingled with IJA reinforcements, totaling 100,000 troops including more than 300 artillery systems, 300 tanks, and 200 aircraft.

The Battles between the Creeks, 24 September–26 October 1937

On the receiving end of this juggernaut was the Nationalist Chinese 9th Army Group, commanded by General Zhang Zhizhong, which was now tasked with protecting the north side of Shanghai. At this point in the battle, General Zhang’s troops were all that stood between the attacking Japanese and the city of Shanghai.

The 9th Army Group had 4 infantry divisions and a separate brigade, a force of 60,000 men. However, unlike other Chinese army groups, the 9th Army Group had substantially more artillery, around 400 guns, and even a separate battalion of 50 light tanks, a unique formation in the Nationalist Chinese Army. General Zhang also continued to command the German-trained 87th and 88th Infantry Divisions, the best formations in the Nationalist Chinese army.

Japanese prime minister Konoe Fumimaro was unaware of Chiang’s intentions and was totally reliant on the advice of Matsui and other IJA commanders, with whom he had an uncomfortable relationship. To the extent that he was able, Konoe restricted the IJA’s political influence. When he agreed to integrate and escalate the combat in the north China and central China theaters by launching an October offensive, he did so reluctantly. He was skeptical that the operation would compel Chiang’s capitulation and bring the conflict to a close.

By early October, the IJA’s strength in the Shanghai region approached 250,000 men. After several days of fighting that was often hand to hand, the IJA 9th Division broke through on 5 October and secured the far side of Wusong Creek. Here, however, the Japanese attack broke down. The Chinese defenders were waiting in carefully prepared trench works that utilized barbed wire, mines, machine guns, and artillery. From 7 to 13 October, heavy rains intervened to provide a respite from the fighting, which allowed time for both sides to replenish supplies. Then the Japanese offensive continued with the goal of eliminating all Chinese resistance north of Wusong Creek.

Chiang Kai-shek had another card to play: Nationalist China’s Guangxi Army of roughly 70,000 to 80,000 soldiers. Given the success of Zhang’s defense, Chiang judged the timing to be right for a Chinese counteroffensive that might fully consolidate Chinese control of Dachang. With the timely arrival of the Guangxi Army on 17 October, Chiang was certain that Nationalist China’s Central Army would be strong enough to regain control of the Yunzaobin River bank, the key to frustrating Japanese efforts to maneuver Chinese forces out of Shanghai. With the combined strength of the 9th and 15th Army Groups, the Nationalist Chinese would enjoy an advantage in personnel numbers over the IJA and a slight advantage in the number of artillery systems but would still labor under major handicaps in the number and quality of tanks. The Chinese also had no real defense against Japanese air strikes.

While Chiang planned his counteroffensive, the Nationalist Chinese 8th Army Group remained on the defense at Pudong, east of Shanghai city and north of Hangzhou Bay, while the 10th Army Group stayed in its defensive positions south of Hangzhou Bay. For reasons that have never been satisfactorily explained, Chiang did not employ these formations to reinforce the 9th Army Group’s flank in the north. Their mission was to simply defend their assigned areas. As a result, nine divisions and four separate brigades that could have joined the fight and diverted IJA forces from defending against the Chinese counterattack did nothing. Whether these might have tipped the balance in favor of the Chinese offensive is unknown, but the fact that they were not massed to support it is peculiar.

The counteroffensive was not well coordinated, and as a result the Chinese attackers were decimated by superior Japanese firepower. As in previous operations, the lack of experience, poor communications, woefully inadequate firepower, and incompetence in the senior Chinese ranks constrained Chinese freedom of action. Chiang’s counteroffensive petered out.

On 25 October, the Japanese proceeded with their planned assault on Dachang. This offensive is worthy of attention because it was probably one of Matsui’s finest operations. The Japanese attack was supported by 700 artillery systems of various calibers and 400 aircraft including 150 bombers. It was also spearheaded by just 20 tanks. In a rare event in Japanese military history, the tanks were concentrated as envisioned by the Germans and, as Ugaki always urged, not broken into small groups to operate as fire support platforms for light infantry.

The Chinese 33rd and 18th Divisions sustained 90 percent casualties and the bridges across the river were captured, but the 26 days of fighting also cost the Japanese 25,000 casualties. When Dachang fell, the 9th Army Group’s flank collapsed, and the door to Shanghai was flung wide open. The operation General Matsui had hoped to complete in a week had instead lasted for almost a month.

Chiang Kai-shek was despondent, writing, “I had hoped that after the troops from Guangxi entered the fighting, we could hold out. The military situation is shaky. It is very discouraging.” Chiang’s fight for Shanghai was always a desperate gamble with an underequipped and inadequately trained force. Now, he began to realize he would have to salvage what he could for a war that would last for years.


Reeling from heavy losses, Chinese forces were totally unprepared when Japanese troops forced their way across the Suzhou River on 30 October, putting Chiang Kai-shek’s remaining troops in grave danger. The combination of the loss of Dachang on 26 October and the Japanese breakthrough on 30 October meant Nationalist Chinese troops in Shanghai would have to withdraw or face certain destruction through methodical Japanese encirclement. If the Chinese army was going to survive to fight again, Chiang Kai-shek’s troops would have to abandon Zhabei and Jiangwan along with other heavily defended Chinese positions. On 30 October, the last remnants of the heroic 9th Army Group, soldiers who had held the line against the IJA through seventy-five days of unimaginable hardship, finally retreated. It was the end of the 9th Army.

Chinese forces now outran the advancing Japanese. There was little the Japanese could do to catch the retreating Chinese forces. The IJA lacked the tracked armor to outpace them. When the Japanese 10th Army, composed of units from northern China, landed in Chingshaweicheng on 5 November, they encountered no real opposition. However, the retreating Chinese were able to establish effective blocking positions with the 19th Army Group in the area north of Taihu Lake and the 10th Army Group south of the lake. These forces successfully covered the majority of retreating Chinese troops.

Within ten days, the IJA 6th Division would advance some thirty miles west and capture Chiahsing in an attempt to cut off retreating Chinese forces, while the 9th Division occupied the city of Suzhou and linked up with the rest of the SEF on 19 November. Another landing by the IJA 16th Division north of Paimaokou designed to cut off retreating Chinese forces came too late to be effective. The operation failed, and the Chinese escaped yet again.

The battle for Shanghai was over. General Matsui Iwane relinquished his command of the SEF and assumed command of a new, larger force, Japan’s Central China Expeditionary Army, which included both the SEF and the 10th Army. The great offensive to smash remaining Chinese resistance was about to begin.

As the war moved inland and away from Shanghai, the Japanese occupation authority in Shanghai restored order, repairing and reopening many of Shanghai’s factories and administrative buildings. But 70 percent of Shanghai’s productive capacity lay in ruins. Thousands of residential buildings, factories, and workshops were damaged or completely demolished. Tens of thousands of homeless Chinese swarmed the city’s streets, and hundreds of thousands more slept in stockrooms, warehouses, temples, and parks. Before the end of 1937, 101,000 corpses, mostly civilian, were recovered from the city’s ruins.

Shanghai’s banking community fared somewhat better. Though Japanese occupation cut off Shanghai banks from economic interaction with the rest of China, the banks still functioned without too much regulatory interference from the occupying authorities. After December 1941, this situation ended, and the Japanese imposed strict controls on all facets of life in Shanghai from monetary policy to food.

The Arrival of the Japanese 10th Army and the Chinese Retreat

While Shanghai settled in for a long Japanese occupation, the fight for control of Nanjing transformed an already bitter conflict into a prolonged series of atrocities on a scale not seen since the days of Genghis Khan. Almost nine years later, Matsui, the man who once befriended Sun Yat-sen, would be tried and executed by the Allied Tokyo War Tribunal for war crimes committed against the Chinese people by his troops in Nanjing.

After Nanjing, Japan’s war with China evolved into the bog that Ishiwara had originally predicted—war without end. Chinese military strategy methodically adhered to the formula described in a 1936 letter to Nishi Haruhiko, who would be Japan’s foreign minister when the Japanese attacked Pearl Harbor: “If war breaks out between Japan and China, the Nationalist Government has made secret plans to resist to the bitter end. If their Shanghai-Nanking [Nanjing] defense line is broken, they will withdraw to a Nanchang-Kiukiang [Jiujiang] line. If that line cannot be defended, they will retreat to Hankow [Hankou]. If the Wuhan defense collapses, they will shift to Chungking.”

With the fall of Nanjing, Japan’s military campaign to subjugate China dragged on without result. In response, Japan’s political and military leaders turned to airpower for an answer. Chinese cities were leveled and millions of Chinese were killed, but the Chinese fought on. The Japanese never seemed to comprehend the power of Chinese nationalism, which was as fierce as their own.


In the history of China’s war with Japan, the fight for Shanghai involved more Chinese soldiers and matériel and incurred higher human losses than any other single engagement. Before it ended, the fight for the city embroiled nearly a million soldiers in combat. For the Chinese Nationalist forces, the outcome was catastrophic. An estimated 270,000 Chinese soldiers were wounded or killed. The exact number of Chinese civilians killed in the fighting is unknown, but numbers range from 250,000 to 450,000. Chiang Kai-shek’s decision to commit his best units to fight for Shanghai devastated China’s best military manpower. In 3 months of intense combat, Chiang Kai-shek sacrificed thousands of his army’s most highly trained junior officers, compelling his forces to fight for years without the competent tactical leadership they desperately needed.

Japanese casualty figures were much lower, estimated to be around 40,000. The disparity in Chinese and Japanese losses highlights the impact of Ugaki’s modest modernization efforts and the high quality of Japanese troops and their leadership, but the struggle for control of Shanghai was harder and bloodier than it should have been.

Whereas Sir Richard Haldane succeeded in changing the British army just enough to play its vital part in the opening battles of World War I, the IJA failed to change enough to achieve a true margin of victory. Rather than downsize the IJA and use the savings to invest in the war-winning technologies of modern warfare—tanks, artillery, and aircraft—the Japanese generals chose to rely on masses of manpower.

The Japanese generals’ preference for numbers in uniform over capability produced the bloodbath at Shanghai and the “grinding mill of flesh and blood” at Luodian. Their determination to retain the large, ponderous World War I square division structures—ideal for the attritional character of trench warfare but unsuited to rapid and decisive offensive operations—was a costly mistake. In the months it took to seize Shanghai, the generals compensated for the lack of offensive capability in their large infantry formations by committing their entire arsenal of tanks and aircraft to move forward and drive the Chinese defenders out of Shanghai.

Tokyo celebrated the fall of Shanghai as a great victory, but the Japanese victory concealed the truth that the IJA was still critically short of completing its transformation into a military force suitable to modern warfare as practiced by the Germans and Soviets. The essential lesson that decisive operations required large, mobile armored forces supported by thousands of advanced fighters and bombers went unlearned. Japanese generals such as Matsui were ferocious and courageous, commanding from the front whenever and wherever conditions demanded decisive leadership. The Japanese generals made use of the tanks and artillery they had, but the IJA still remained organized for World War I until World War II ended.

Despite the gradual introduction of newer and better tanks before war’s end in 1945, the IJA never fielded more than the equivalent of three armored divisions in combat, a force woefully inadequate for operations in the Chinese and southeast Asian theaters of war. Japanese tanks and self-propelled guns never attained the quality and capability of comparable equipment in the U.S., German, or Soviet arsenals. Serious weaknesses in mobile armored firepower and logistics allowed the inexperienced and ill-equipped Chinese armies to escape, regroup, and resume their attacks on Japanese occupation forces time and again.

One reason the Japanese generals failed to learn from their operations in Shanghai was the weakness of their opponents in the early stages of World War II. In Malaya, where the British opponent could not melt away into an expansive interior, the Japanese conducted a masterful campaign involving the coordinated use of tanks, artillery, engineers, infantry, aircraft, and ships, defeating a British army in February 1942 that was twice as large as the attacking force. Against the British troops defending Singapore, the quality of IJA training and aggressive leadership compensated for the IJA’s numerical inferiority on the ground. The conquest of the U.S.-held Philippines in May 1942, though more challenging than the campaign to take Singapore, simply reinforced Tokyo’s illusion of Japanese military superiority.

Japan’s modest military advantage did not last long. Two years after the fall of Shanghai, Japan’s Kwantung Army was decisively defeated by Soviet armed forces on the plains of Nomonhan. Japanese airpower fared better than the ground forces; superb Japanese pilots fought their Soviet opponents in the air to a draw, but airpower could not compensate for the IJA’s acute weaknesses in mobility, armor, firepower, and organization. Soviet army small arms, tanks, artillery, and organization were demonstrably superior to anything the IJA could field. Japan sued for peace, releasing Soviet forces in Siberia to move west in time to defend Moscow in the winter battles of 1941–42.

As American air and naval power closed the ring around Japan in early 1945 and Japanese cities were incinerated from the air, 1 million of the Imperial Japanese Army’s 5.9 million soldiers were still in China, and another 780,000 were in Manchuria. Chiang Kai-shek’s decision to trade lives for space exacted a high price from the IJA. The defenders of Shanghai prevented the Japanese from striking directly into central China, slowing the Japanese advance long enough for the Nationalist government to begin moving a portion of its defense industries deeper into China, toward China’s new wartime capital, Chongqing (Chung King).

Japan’s war with China not only delayed and disrupted the IJA’s modernization, it also fatally crippled Japan’s northern strategy to defeat the Soviet Union, while putting Japan on a collision course with Britain and the United States. In August 1945, the calamity that Ugaki, Ishiwara, and their supporters had feared the most struck the IJA in Manchuria, when Japanese forces were annihilated in less than two weeks by Soviet armored and mechanized units consisting of 5,556 tanks together with several thousand self-propelled guns and supporting armored fighting vehicles.

The hard lesson that mass and athleticism do not equate with real military capability was not lost on the postwar Japan Self Defense Forces (JSDF). Less than twenty-five years after World War II ended, the much smaller JSDF, only 4 percent of the size of the IJA at its height, already possessed thirty times more firepower than its predecessor. Emperor Hirohito’s comment that “our military leaders put too much emphasis on [fighting] spirit and forgot the significance of science” still resonates in Japan. Technology, not manpower, combined with superior organization and leadership, is now widely recognized as Japan’s future margin of victory.

Ugaki Kazushige won his fight for change inside the IJA after all, albeit at the cost of a lost war. Today, the Japan Ground Self-Defense Force (JGSDF) fields significant armored forces and more tanks and armored fighting vehicles than the British or French armies. Japanese soldiers are highly educated, trained, and disciplined. At this writing, the JGSDF is upgrading its armored force to the new Type 10 tanks while retiring the older Type 74 tanks. It is also likely that production of the Type 89 Infantry Fighting Vehicle will resume in the near future. The JGSDF is following a familiar pattern: shed old equipment first, then modernize with initial low rate production runs until the new equipment performs to expectation. In the West, this process is called rapid prototyping, and it is something much discussed but seldom seen in practice.

China’s emergence as a great power in the twenty-first century must also be viewed through the lens of its war with Japan, especially the battle of Shanghai. China’s inability to defend its near seas, great rivers, and coastal cities, and its vulnerability to attack from Formosa (Taiwan), Japan’s unsinkable aircraft carrier during the battle for Shanghai, and the seizure of all of China’s coastal cities traumatized the Chinese people. These experiences not only shape current Chinese military modernization efforts, they also figure prominently in Chinese thinking about the potential conflict with Japan and the United States, Asia’s two great maritime powers.

China’s recent demarcation of an air identification zone that includes the Senkaku and Diaoyu Islands must also be seen as part of a larger Chinese anti-access/area denial (A2AD) strategy with its roots in a tragic, costly war against Japanese invasion that began with the battle of Shanghai. Americans would do well to keep this history in mind before leaping to conclusions about China’s alleged belligerence toward the United States and its commercial interests in Asia.

N1K1J “George” fighter

From the standard reference by Rene Francillon, Japanese Aircraft of the Pacific War (UK: Conway Maritime Press, 1987; also in US by Naval Institute Press), pages 323-25:

“… [despite engine troubles] … In combat, however, the Shiden was a superlative combat aircraft and experienced pilots had little difficulty in engaging American aircraft and, under the code name GEORGE, it was considered by Allied personnel to be one the best Japanese aircraft.

“… In operation the N1K2-J revealed itself as a truly outstanding fighter capable of meeting on eqial terms the best Allied fighter aircraft. Its qualities were demonstrated spectacularly by such pilots as Warrant Officer Kinsuke Muto of the 343rd Kokutai who, in February 1945, engaged single-handed twelve US Navy Hellcats, destroying four (of them) and forcing the others to break off combat. Against the high-flying B-29s the Shiden Kai was less successful as its climbing speed was insufficient and the power of its Homare 21 (engine) fell rapidly at high altitudes.”


This engine was one of the main problems with flying the Shiden, since it could fail to run up to its full power for combat. The undercarriage was relatively weak as well. But once in the air and at full power it was a dangerous adversary.

There are three surviving N1K2-J examples left today, all in the US. One is in the Smithsonian Institution’s National Air and Space Museum near Washington, another in the US Air Force Museum in Ohio, and the third in the New England Air Museum in Connecticut.

In 1993, the Smithsonian’s Shiden was sent to the Champlin Fighter Museum in Arizona for its restoration. There had been a special job of rewiring the engine to match the original six-color, 12- and 16-gauge wiring and its braiding. The experts had enthused that even this job alone was almost a work of art by itself, a common sentiment in rare warplane restoration and maintenance.



Kawanishi N1K1-J Shiden George 11:


One Nakajima NK9H Homare 21 eighteen-cylinder air-cooled radial rated at 1990 hp for takeoff, 1825 hp at 5740 feet, 1625 hp at 20,015 feet.

Performance: Maximum speed 363 mph at 19,355 feet, 334 mph at 8040 feet.

Cruising speed 230 mph at 6560 feet, service ceiling 41,000 feet, cruising speed 230 mph at 6600 feet.

Climb to 19,685 feet in 7 minutes 50 seconds.

Normal range 890 miles at 230 mph at 13,120 feet, maximum range 1580 miles.

Weights: 6387 pounds empty, 8598 pounds loaded, 9526 pounds maximum loaded.

Dimensions: wingspan 39 feet 4 7/16 inches, length 29 feet 1 25/32 inches, height 13 feet 3 27/32 inches, wing area 252.95 square feet.

Armament: Two 7.7-mm Type 97 machine guns in the fuselage, two 20-mm Type 99 Model 2 cannon in the wings, two 20-mm Type 99 Model 2 cannon in underwing gondolas. Two 132-pound bombs or one 88 Imp gall drop tank could be carried externally.


N1K2-J Shiden Kai George 21:


One Nakajima NK9H Homare 21 eighteen-cylinder air-cooled radial rated at 1990 hp for takeoff, 1825 hp at 5740 feet, 1625 hp at 20,015 feet.

Performance: Maximum speed 369 mph at 19,355 feet, 359 mph at 9840 feet.

Cruising speed 230 mph at 9845 feet, service ceiling 35,300 feet, cruising speed 230 mph at 6600 feet.

Climb to 19,685 feet in 7 minutes 22 seconds.

Normal range 1066 miles at 219 mph at 9840 feet, maximum range 1488 miles with 88 Imp. gall. drop tank.

Weights: 5858 pounds empty, 8818 pounds loaded, 10,714 pounds maximum loaded.

Dimensions: wingspan 39 feet 4 7/16 inches, length 30 feet 7 29/32 inches, height 12 feet 11 29/32 inches, wing area 252.95 square feet.

Armament: Four 20-mm Type 99 Model 2 cannon in the wings. Two 551-pound bombs or one 88 Imp. gall. drop tank could be carried externally.


Minoru Genda’s elite 343 Kokutai flew the Shiden-Kai. Training began in January 1945 at Matsuyama Airfield, and the base of operations later moved to Kanoya (April 4), Kokubu (April 17), and Omura (April 25) in Kyushu.

During the Battle of Okinawa, the 343 Kokutai had the task of trying to clear the way for kamikaze planes as they flew from southern Kyushu to Okinawa during the Kikusui operations from April 6 to June 22, 1945. The Shiden-Kai pilots fought several fierce battles with American fighters over Amami Oshima and Kikaigashima. When American planes bombed Kyushu airfields to try to stop kamikaze attacks in April and May 1945, the 343 Kokutai at times engaged enemy aircraft although the Shiden-Kai was not intended for high-altitude interception of B-29s.

Genda’s Blade: Japan’s Squadron of Aces: 343 Kokutai by Henry Sakaida and Koji Takaki.

All English-language names for Japanese fighters derived from Western Allied identification codes, in which male names were given to enemy fighters and female names to Japanese bombers. The Japanese Naval Air Force (JNAF) Zero, or Mitsubishi A6M “Reisen” (Zero-Sen), was the best fighter available in the Pacific in 1941. It was lighter, faster, and more maneuverable than American land-based aircraft. It also had a much greater range and more nimble handling than any U.S. carrier-based fighters. That gave the Imperial Japanese Navy a critical advantage in early carrier vs. carrier fights such as Coral Sea. The Japanese Army Air Force (JAAF) flew three models of the Nakajima Ki-43 “Hayabusa” (“Falcon”). Designated alternately as “Jim” or “Oscar” by the Western Allies, these land-based JAAF fighters saw most service in China and Southeast Asia, flying cover over ground forces. They faced handfuls of older Soviet and other fighters in China until the arrival of American pilots and modern aircraft of the American Volunteer Group, or “Flying Tigers” (“Fei Hu”). Japanese pilots in Hayabusa also faced RAF Spitfire and Hurricanes in Malaya and over Burma. Western pilots were initially shocked at the excellent performance of the Hayabusa, whose characteristics were not known to British or American military intelligence. The JAAF also flew the very fast “Hein,” which reached speeds above 400 mph. The “Frank” (Nakajima Ki-84-Ia “Hayate”), introduced in 1944, and the excellent “George” (Kawanishi N1K1-J “Shiden”), introduced in 1944–1945, were also well-known to Allied sailors, troops, and flyers. But as improved as those aircraft were, neither model could match Western Allied fighters by that point in the war: the Japanese planes were relatively underarmored and undergunned, and by 1944 were usually flown by inexperienced, young pilots. However, over Japan the Hayate’s ceiling of nearly 38,000 feet and rocket weapons did pose a threat even to American B-29 bombers.

The USN F4F Wildcat was overmatched by Zeros in nearly all ways, an often-fatal disadvantage not overcome by introduction of new American fighters for the first two years of the Pacific War. But the USN controlled the skies of the Pacific after powerful Pratt & Whitney engines were put into its heavily armored F6F “Hellcats” and F4U “Corsairs.” The combination of power, climb rate, ceiling, and arms and armament allowed those aircraft to master the fast but lightly armored Zero and to splash hundreds of slow IJN and Japanese Army bombers. The USAAF also had inadequate and mostly short-range fighters at the start of the war. But by war’s end, the USAAF boasted several of the finest and most effective fighters in the world. Many U.S. fighters were shipped to the Soviet Union under Lend-Lease, including 4,700 Bell P-39 “Airacobras” personally requested by Stalin. The P-47 “Thunderbolt” and P-51 “Mustang” dominated the skies of Italy, France, and Germany almost as soon as they were introduced in 1943. The P-51 may have been the fi nest fighter of the war. It was equipped with long-range drop tanks that permitted it to escort strategic bomber formations deep into Germany and to the home islands of Japan. Both the P-47 and P-51 were also fitted with rockets and used in a “tank buster” role. In combination with late-war deterioration in Japanese aviator skills, better trained American pilots with new and better tactics in much improved machines achieved a 10:1 or higher kill ratio in Pacific War dogfights.


N1K1 Kyofu

  • N1K1: only standard type as floatplane, which was used from early 1943.
  • N1K2: reserved name for an intended model with larger engine, not built.

N1K1-J Shiden

  • N1K1-J: Prototypes: development of fighter hydroplane N1K1 Kyofu, 1,357 kW (1,820 hp) Nakajima Homare 11 Engine, 9 built
  • N1K1-J Shiden (“Violet Thunder”) Navy Land-Based Interceptor, Model 11: first production model: 1,484 kW (1,990 hp) Homare 21 engine with revised cover, armed with two 7.7 mm (.303 in) Type 97 machine guns and two 20 mm Type 99 cannons. Modified total-vision cockpit.
  • N1K1-Ja, Model 11A: Without frontal 7.7 mm (.303 in) Type 97s, only four 20 mm Type 99s in wings
  • N1K1-Jb, Model 11B: Similar to Model 11A amongst load two 250 kg (550 lb) bombs, revised wing weapons
  • N1K1-Jc,Model 11C: definitive fighter-bomber version, derived from Model 11B. Four bomb racks under wings.
  • N1K1-J KAIa: experimental version with auxiliary rocket. One Model 11 conversion.
  • N1K1-J KAIb: conversion for dive bombing. One 250 kg (550 lb) bomb under belly and six rockets under wings.

N1K2-J Shiden-KAI

  • N1K2-J Prototypes: N1K1-Jb redesigned. Low wings, engine cover and landing gear modified. New fuselage and tail, 8 built
  • N1K2-J Shiden KAI (Violet Thunder, Modified) Navy Land Based Interceptor, Model 21: first model of series
  • N1K2-Ja,Model 21A: Fighter-bomber version. Four 250 kg (550 lb) bombs. Constructed by Kawanishi: 393, Mitsubishi: 9, Aichi: 1, Showa Hikoki: 1, Ohmura Navy Arsenal: 10, Hiro Navy Arsenal: 1.
  • N1K2-K Shiden KAI-Rensen (Violet Thunder Fighter Trainer, Modified) Trainer version of N1K-J Series with two seats, operative or factory conversions

Further variants

  • N1K3-J Shiden KAI 1, Model 31 Prototypes: Engines displaced to ahead, two 13.2 mm (51 in) Type 3 machine guns in front, 2 built
  • N1K3-A Shiden KAI 2, Model 41: Carrier-based version of N1K3-J, project only
  • N1K4-J Shiden KAI 3, Model 3: Prototypes, 1,491 kW (2,000 hp) Homare 23 engine, 2 built.
  • N1K4-A Shiden KAI 4, Model 4: Prototype, experimental conversion of N1K4-J example with equipment for use in carriers, 1 built
  • N1K5-J Shiden KAI 5, Model 25: High-Altitude Interceptor version. Project only

Total Production (all versions): 1,435 examples.

Factory Production

The Sinking of the Musashi

The sinking of the superbattleship Musashi by US carrier aircraft in the battle of the Sibuyan Sea (24 October 1944) gives an idea of Japanese anti-aircraft capability and may be compared to the sinking of Prince of Wales about three years earlier. Musashi was considerably larger than the British battleship, and she absorbed much more damage. The sixteen bombs which hit her did not contribute directly to her sinking, but they did help reduce her anti-aircraft effectiveness and thus indirectly contributed to the success of the torpedo bombers which sank her. The Japanese initially reported that she was struck by twenty-one torpedoes, including two duds, but the US Navy concluded from interrogation of survivors and other Japanese naval personnel that only ten hits and four possible but not probable hits could be identified. Analysis suggested that ten hits equally divided between both sides in the forward three-quarters of the ship would have been enough to sink her. Analysis was complicated because the Japanese did not produce war damage reports comparable to those produced by the US Navy or the Royal Navy, nor were commanding officers required to submit war damage reports. However, both the executive officer and the engineering officer of Musashi kept detailed notebooks, which survived the war.

When she was attacked, Musashi was part of a large Japanese surface action group, Admiral Kurita’s Center Force. It included both superbattleships and three older battleships (Nagato, Haruna and Kongo), plus numerous cruisers and fifteen destroyers. The ship’s CO, Rear Admiral Inoguchi, was a gunnery officer who reportedly placed great faith in the special shrapnel/incendiary shells her main battery could fire. Like Admiral Phillips off Malaya three years earlier, Admiral Kurita asked for fighter cover, which would have been provided by the large naval air force ashore in the Philippines. It was later claimed that ten fighters had been kept aloft over his force, but US attackers saw only four of them, which they quickly shot down. All the losses to the attackers were due to anti-aircraft fire. Musashi had been upgraded considerably since completion, and her numerous 25mm mounts were all controlled by directors comparable to those on board Prince of Wales in 1941. Probably the most significant difference in the two actions was that Musashi was not unfortunate enough to suffer early hits which put her electrical power out of action.

As in the earlier case, prior to the attack the ship was shadowed by a search plane, in this case from the carrier Intrepid. The ship tried unsuccessfully to jam the aircraft’s radio. About two hours later, the first strike (estimated by the Japanese as thirty aircraft, which would be equivalent to Intrepid’s combined torpedo and dive bombing force) arrived. Some aircraft came from the light carrier Cabot. The attack began with eight SB2C dive bombers, which caused minor damage. They were followed by three Avengers, one of which hit the ship amidships, slightly abaft the bridge. The shock of this hit jammed the main battery director, so the ship was unable to fire her 46cm Type 3 shells. Two of the three Avengers were shot down. During this attack the ship fired forty-eight 155mm (low-angle) and sixty 127mm shells. After this attack the ship switched to her after main battery director; changeover of this type was awkward in Japanese ships due to synchro and switchboard design.

About half an hour later the ship’s air-search radar detected a second raid 81km out. A few minutes later the aircraft were sighted, and another eight Helldivers from Intrepid attacked, this time scoring two bomb hits and five near-misses. A bomb fragment which penetrated the muzzle of one gun in No. 1 turret detonated a Type 3 shell which had just been loaded, disabling the turret. Nine Avengers delivered a hammer-and-anvil torpedo attack, eight of them dropping torpedoes. Three hit the port side amidships, flooding one engine room. The director changeover made it possible to fire fifty-four 46cm Type 3 shells. In addition, the ship fired seventeen 155mm and 200 127mm. Bomb damage to an engine room slowed the ship, and she was left down by the bow. The attacking US pilots had never encountered Type 3 shells before, and they were impressed that the Japanese would fire against them at ranges of 25,000 to 30,000 yds, at which the relatively slow train and elevation rates of large-calibre turrets would not be a problem. ‘The fire was surprisingly accurate and somewhat distracting, though no damage was sustained by the planes so attacked.’ US pilots thought the shells were loaded with phosphorus.

About an hour and a half later twenty-nine aircraft from Essex and Lexington attacked, including two strafing Hellcats. Four Helldivers made two hits near starboard amidships and abeam the after 46cm turret, causing casualties among the 25mm crews. Other Helldivers made four bomb hits on the port side. Another hammer and anvil attack, this time by six Avengers, made four more torpedo hits, two on each side. The ship fired another thirty-six 46cm Type 3 shells, plus seventy-nine 155mm and over 500 25mm. The ship was now further down by the bow, reduced to 20kts and thus lagging behind Kurita’s 22kt force.

About two hours later eight Hellcats and twelve Helldivers from Essex attacked two of the other four battleships, Yamato and Nagato. The bomb damage they inflicted had no real effect. At this point the CO of the accompanying cruiser Tone suggested that the ships of the force provide anti-aircraft support for Musashi.

A fifth attack carried out by sixty-nine aircraft from Enterprise and Franklin made four hits with 1000lb AP bombs, three in the bow area, and three torpedo hits. The pilots reported that the ship was dead in the water, heavily down by the bow and smoking. After they left she managed to increase speed to 16kts (soon reduced to 13kts) and she corrected her starboard list.

The sixth and final attack on Kurita’s force was mounted by seventy-five aircraft from Intrepid (thirty-four), Franklin (thirty) and Cabot (one); thirty-seven of them attacked Musashi. They made a total of ten bomb hits, some of which wiped out 25mm guns. It is not clear how many torpedo hits were made, since totals given by different sources vary. A battle narrative gives a total of nineteen torpedo hits (ten to port, nine to starboard), seventeen bomb hits and eighteen near-misses. However, most current Japanese accounts give eleven torpedo hits, ten bomb hits and six near-misses.

A total of 259 US carrier sorties was flown, and eighteen US aircraft were shot down during the attacks, for a loss rate of 6.9 per cent, better than that inflicted by Prince of Wales and Repulse during their final battle. US pilots were unimpressed by the Type 3 shells, and fire by 127mm guns seems to have been limited. The main defence was 25mm guns, for which US pilots had respect. Dive bombing attacks could not sink the ship, but they certainly could destroy the light anti-aircraft guns which were beating off the torpedo bombers. At the least they could help saturate the ship’s anti-aircraft fire control channels. Note that virtually all attacks were combinations of dive bombers and torpedo bombers. It is also obvious that the Japanese had not adopted US-style circular formations with their interlocking fields of anti-aircraft fire. Only at the end did Musashi receive support from other ships (at the end she was attended by the cruiser Tone and the destroyers Shimakaze and Kiyoshimo, neither of them an anti-aircraft destroyer).

This was the first major US air strike against a Japanese capital ship since 1942. In retrospect it seems surprising that attacks were not spread effectively over the rest of Kurita’s force. One answer lies in the way the attacks were carried out, in succession by different Task Groups. No pilots from any one such attack knew what his predecessors had hit, and it was easy to concentrate on one spectacular target. Moreover, pilots in successive waves seem to have thought they were hitting different ships.

Musashi absorbed enormous punishment and in so doing seems also to have absorbed the air striking power available to Task Force 38. Kurita cannot have intended it that way, but because the pilots concentrated on her, they were unable to inflict significant damage on the rest of Kurita’s large surface force. Task force commander Admiral Halsey later commented that the attack showed just how difficult it still was for aircraft to sink a large surface combatant. In effect that was a post-battle justification for his unwillingness to form a battle line (Task Force 34) when he went north to engage the Japanese decoy carrier force. The attack also showed how misleading pilots’ reports could be. They exaggerated the damage they had inflicted on the other battleships (two bombs each on Yamato and Nagato, five near-misses on Haruna). Given their claims, they were too ready to report that they had turned Kurita back. They interpreted ships milling around to support damaged units as ships stopped and ready to retire. Kurita did retire temporarily, calling for strikes by land-based aircraft (which could not materialise) to precede him. On the night after the battle, he turned back towards Leyte Gulf, which had been denuded of capital ship protection on the basis of the exaggerated strike reports.

The most interesting lesson is that battle-damage assessment is the most difficult part of an attack. Issued in March 1945, the Cominch compilation of ‘Battle Experience’ for Leyte Gulf reflects after-action reports. It seems clear that the pilots reported that they had crippled Kurita’s force. TG 38.2, which made more than half the attacks (146 sorties), dropped 23 tons of bombs and twenty-three torpedoes. Its pilots reported that they had hit Yamato with three torpedoes and hit a sister ship (possibly the same ship) with one torpedo and two bombs; that they had hit the battleship Nagato with a torpedo and a bomb; that one Kongo class battleship had been hit by two torpedoes and six bombs; a Mogami class cruiser had possibly been sunk by a torpedo; the cruiser Nachi had been hit by one torpedo. Task Group 38.3 reported one bat-tleship badly hit, two others damaged, and four heavy and two light cruisers damaged. Task Group 38.4 reported one battleship (Musashi?) hit by a torpedo, on fire, down at the bow and probably sunk, one Yamato class battleship hit by one to three torpedoes and two bombs; a Kongo class battleship hit once by a bomb, one light cruiser sunk, one destroyer sunk, one destroyer probably sunk and four destroyers damaged. Although some of this information was not immediately relayed to Admiral Halsey, the impression that great damage had been done was unmistakeable. Halsey was convinced; in his after-action report he wrote that the enemy had turned back to attack off Samar out of blind obedience to an Imperial command to do or die.

It was soon obvious that the pilots had exaggerated about as badly as the Japanese had when they reported sinking the US fleet several times over after attacking the Task Force off Formosa just before Leyte Gulf. In March 1945 Cominch credited the pilots with having damaged Musashi and sunk the heavy cruiser Haguro. In fact the cruiser was quite intact, having steamed south, but Musashi had been sunk. Unlike the Japanese, who orbited the crippled British capital ships to make sure they were sunk, the US carriers did not maintain anyone over the battle scene to be sure of what happened. That was partly due to range (the battle was at extreme strike range for the Task Force) and probably also because the need for such assessment had not been driven home.

The Cominch combat analysis emphasised the problems pilots faced in evaluating their results. In a secret letter, CinCPAC Admiral Nimitz pointed out the problem, and Cominch clearly felt it had to be repeated. Nimitz quoted a report after an air strike early in the war: one 15,000-ton transport (AP) on fire and beached; one transport (AP) sunk and burning; one transport or cargo ship beached and probably sunk; one transport or cargo ship sunk, bottomed in shallow water, and listing; one Mogami class cruiser blown up and sunk; one Kinugasa class cruiser afire and headed for the beach, believed sunk; one light cruiser headed for the beach, believed sunk; one seaplane tender (Kamoi class) damaged and stopped; one destroyer listing, afire, and sinking fast; two other destroyers probably sun; one gunboat set afire and severely damaged; one minesweeper stopped and burning fiercely, probably sunk. Confirmed sinkings were actually three cargo ships of 4000 to 6000 tons. No warships were sunk.

Nimitz did not want to ruin his pilots’ enthusiasm, but he did want them to know that there could be a gap between good-faith reports and reality. The worst problem was that it was generally unwise to remain in an area to assess results, as long as any ships and their AA crews survived to keep firing. With so many aircraft involved, reports would necessarily be duplicated, and they might be difficult to disentangle. Pilots tended to be over-optimistic about the effects of their attacks: there could be tremendous explosions topside, yet a ship might still get underway and get home. Pilots could also be over-optimistic about near-misses: if near enough, they could certainly do tremendous damage, but then again they might not. Similarly, there was over-optimism as to fire and smoke: a small and possibly harmless fire could produce a great deal of smoke. Even ships afire from stem to stern could survive. There was over-valuation of ships ‘beached and sunk’. A damaged ship might well beach herself lightly until the attack was over – but she would survive. Finally, Nimitz cited the ‘lack of familiarity with ships on the part of many pilots, which handicaps them in distinguishing types and tonnages and in estimating the seriousness of the damage or the probability of a ship sinking’. This last point applies to nearly all the examples of air-sea combat already quoted.

The attack on the battleship Yamato in April 1945 contrasts with that against Musashi. The US Navy seems to have realised that the ship’s sheer capacity to absorb aerial punishment ensured that strike aircraft would keep coming back to attack her rather than distribute their fire among the ships in the Japanese surface strike force. This time the strikes were very differently organised. Each of the groups launched by a Carrier Task Group had a coordinator. The effect of coordination shows in that considerable numbers of torpedoes were devoted to the other ships in the force. It probably helped that Yamato was the only large ship in her task force, the others being the light cruiser Yahagi and eight destroyers. Concentrating on the one large ship did not have the unfortunate effects of the concentration on Musashi which left the other ships of the surface strike group effectively undamaged to fight the battle off Samar the next day.

Task Force 58 was alerted on the night of 6–7 April by two US submarines (Threadfin and Hackleback) patrolling the Bungo Suido Channel (reportedly the Japanese intercepted the uncoded sighting reports). At dawn the Task Groups launched a total of forty aircraft, all fighters, in groups of four, to search to a depth of 325nm. At 08.22 an Essex search aircraft reported one battleship, probably Yamato class, two cruisers, and eight destroyers making 12kts. The fighter could not contact the carrier directly via her VHF line-of-sight radio, but linking aircraft had been launched. She radioed via them (100 and 200nm away). The next step was to shadow the enemy group, so that a strike group could be vectored to them. At 09.56 a tracking and covering force of sixteen fighters was launched, followed at 10.00 by strikes from Task Groups 58.1 and 58.3 and 45 minutes later by a strike from Task Group 58.4.

All three Task Groups were to have launched together, but the Hancock strike (12 torpedo bombers, 15 dive bombers, and 24 fighters) was 15 minutes late on take-off and failed to join up (and hence to find the targets). This was an immense force, totalling 386 aircraft: 113 from TG 58.1 (52 fighters, 21 dive bombers, 40 torpedo bombers), 167 from TG 58.3 (80 fighters, 29 dive bombers, and 58 torpedo bombers), and 106 from TG 58.4 (48 fighters, 25 dive bombers, 33 torpedo bombers). All the torpedo bombers carried torpedoes. The dive bombers (Helldivers) carried 1000lb SAP and underwing 250lb GP bombs. Each fighter had a 500lb GP bomb and a long-range drop tank. This huge strike left enough fighters with Task Force 58 to deal with enemy attacks (a Kamikaze crashed into the carrier Hancock while the strike was away). The launch position was about 250nm from the estimated position of the Japanese force (i.e., aircraft would have to fly about 240nm). All of the searching was needed: the enemy force unexpectedly turned north, to be found again by a land-based search aircraft from Okinawa. It shadowed the enemy force for the rest of the day, but shadowing reports failed to get through to the Task Force Commander. It turned out that it did not matter very much because the strike leaders were soon in touch with the shadower. Final homing was by APS-4 radar on Helldivers, which picked up the enemy force at 32nm from 6000ft.

When first sighted, the Japanese force was 70nm from the first sighting position. A combination of poor weather and the sheer size of the attacking force made the attack difficult to coordinate. Anti-aircraft fire was heavy but ineffective, and the three Task Groups attacked roughly in sequence, TG 58.1 and 58.3 first and then TG 58.4.

The attack developed in three phases, of which the first consisted of two almost simultaneous attacks. The first two strike groups hit not only Yamato but also the cruiser and three destroyers. The attacks began with strafing to suppress her light anti-aircraft battery. This time the torpedo bombers concentrated on one side of the ship. The first two bomb hits around around No 2 turret. wrecked a 12.7cm anti-aircraft mount and many light anti-aircraft guns. Another two, inflicted a few minutes later, wrecked the after secondary battery director and exploded above the protective deck, starting a fire which was never extinguished. The after 15.5cm mount was gutted. At least the first two hits seem to have been by 500lb GP bombs rather than AP bombs. A second wave of attacks began 40 to 45 minutes later, inflicting three or four torpedo hits on the port side and one on the starboard side. There were no bomb hits. About thirty minutes later a third and last attack began. The ship took two more torpedoes to port and one more to starboard (some Japanese officers thought there were additional hits, but the post-war US analysis discounted that). Altogether Yamato seems to have taken at least nine torpedo hits (plus three possible, but improbable), of which seven were on her port side and two on her starboard side. The ship also took at least four hits from dive bombers.

Yamato capsized to port 20 to 30 minutes after the last three torpedo hits, her magazines exploding as she rolled over. Both the Assistant Gunnery Officer and the Chief of Staff told US officers after the war that they believed that the fire aft ignited the magazines of the after 15.5cm mount, passing to them as the ship rolled over. A study of main battery shell fuses militated against the alternative explanation, given by the ship’s Executive Officer, that as the ship rolled over her HE and incendiary AA shells fell out of their racks in all three 46cm magazines, hit their noses on the deck, and exploded.

US losses amounted to ten aircraft (four dive bombers, three torpedo bombers, and three fighters) and twelve aircrew (four pilots and eight crew).

Assessed results were Yamato and a light cruiser (Yahagi) sunk, as well as four destroyers, plus one badly damaged (Akizuki class) and one left burning. In fact the light cruiser was sunk (she took, among other damage, six torpedoes) and the initial wave sank the destroyer Isokaze and damaged two others so badly that they had to leave the area (one of them, Hamakaze, later sank). Later waves sank the destroyer Asashimo and damaged Kasumi so badly that she sank. Three destroyers rescued survivors and returned to Japan. Thus the assessment in the after-action report was far more accurate than it had been in the October 1944 action, perhaps as a result of Nimitz’ comments at the time. A search and a fighter sweep (thirty-two fighters armed with bombs) the next day failed to find the surviving Japanese ships.