Japanese Carriers and Carrier Doctrine, 1920–1941
It is important to understand that the various naval air tactics and technologies outlined in the last chapter had a symbiotic relationship to the composition of the Japanese carrier fleet.1 The construction of that fleet began at a time when carrier design in the three major navies was being shaped not only by the universal problems of propulsion, hull structure, seakeeping, crew accommodation, and compatibility with shore facilities, but also by the growing recognition of the ways in which the embarkation and flight operations of aircraft enormously complicated the task of naval architects. Above and beyond the matter of appropriately proportioned flight decks for air operations—takeoffs and landings—there were problems of the proper siting of control centers for such operations, provision of adequate aircraft storage, the location of maintenance facilities and magazines for aircraft ordnance, and the arrangement of storage and delivery systems for large quantities of aviation fuel, along with the distribution of physical safeguards to minimize the resultant hazards of fire and explosion. In addition to these new challenges, the naval limitation treaties signed after World War I restricted the displacement tonnage of individual carriers to 23,000 tons, although in the case of the Japanese and American navies it allowed the conversion of certain capital ships to carriers (the Akagi and Kaga and the Lexington and Saratoga).2
But two problems in particular bedeviled carrier design in this period. The first was rapidly changing aircraft technology, particularly the increasing weight and wing loading of aircraft, which required longer flight decks for takeoff. The second was the still uncertain function of the carrier and its place relative to the battle line. These were ambiguities that, for a while at least, seemed to argue for the installation of heavy surface weapons on carriers for defense against cruisers and destroyers. Given these problems, it is not surprising that carrier design in these years was a matter of trial and error, and that in all three major navies the design of the first warships of this type included certain features that proved impractical in the long run.
JAPANESE CARRIER DESIGN IN THE TREATY ERA
These matters relate particularly to the design and construction of the Hōshō, Japan’s first and essentially experimental aircraft carrier.3 Viewed from the perspective of later carrier design, the Hōshō, at less than 8,000 standard tons displacement, with a flight deck of only 168 meters (552 feet) was a very small carrier. The design also incorporated certain features that demonstrated the uncertainties of her designers in confronting problems involved in flight operations. As completed in 1922, the Hōshō was fitted with a rudimentary island-type bridge, on the starboard side of the flight deck about 45 meters (150 feet) from the bow, from which the ship was conned and flight operations controlled. Just aft of the island were the carrier’s three stacks, which were hinged so as to fold over into a horizontal position during flight operations. Forward of the island, the flight deck itself was inclined slightly downward. Two small aircraft elevators, one forward and one aft, led down to the single narrow hangar deck that began forward of the island bridge, widened amidships, and at the fantail took up the full beam of the ship.4
After repeated experimentation in launching and recovering aircraft in the year after the Hōshō was commissioned, the navy decided to remodel the ship. Because the island superstructure, its mast, and the three stacks reduced the 23-meter (75-foot) flight deck and to some extent obscured the pilot’s visibility, they were all removed or rearranged in 1924. With the removal of the island bridge and its mast, the Hōshō became a flush-deck carrier. Flight operations were controlled from an operations platform beside the flight deck, a decision that influenced Japanese carrier design for over a decade. But the positioning of the carrier’s boiler uptakes proved to be troublesome. Even in their horizontal position, the three stacks caused smoke and stack gas to drift across the path of flight operations, and the stack wells encroached too far into the already narrow flight deck. For this reason, in 1934 they were permanently fixed at an angle downward over the ship’s side. In 1931 the lengthwise arresting gear, which had proved inadequate and had thus contributed to a number of landing accidents, was replaced by a system of transverse cables across the deck.5
As originally designed, the Hōshō could carry approximately twenty aircraft. The first types embarked after the ship was commissioned were Mitsubishi 1MF1-5 (Navy Type 10) carrier fighters and Mitsubishi B1M1 (Navy Type 13) torpedo planes, some of which were used for reconnaissance.6 During its first decade, the Hōshō’s small size posed no limit to her usefulness, given the slow speeds and light naval aircraft of the day. But by the late 1930s, as larger and faster attack aircraft were developed, the Hōshō could accommodate only half her original complement. For this reason, for the rest of her period of service she was used for training or, at best, defensive cover for the fleet. She was relegated to auxiliary status by the end of the Pacific War and was one of the two operable carriers left to the Japanese navy at the end of that conflict. Yet long before that, the Hōshō had proved of value as a laboratory for carrier design, construction, and flight operations.7
If the small size and modest aircraft complement of the Hōshō prevented her from being a significant element in the navy’s strategic planning in the 1920s, the completion of the fleet carriers Akagi and Kaga,8 and the more than 120 aircraft they added to the offensive capability of the fleet, did indeed create new strategic as well as tactical possibilities for the navy. Yet their initial design incorporated features that demonstrated continued uncertainties about the function of carriers and an underestimation of the rapidity of change in both the technology and the tactics of naval aviation. The Akagi was begun as a battle cruiser whose construction was halted because of the capital ship limitations imposed on all signatories by the Washington Naval Treaty. While the essential battle cruiser hull structure was retained, the warship’s new function as a carrier and her consequent need for hangar space necessitated a number of modifications. The armor belt was lowered and reduced in thickness, and the torpedo bulges were modified. Both modifications improved stability and helped compensate for the topside weight of the double hangar deck.9 A major design problem in both the Akagi and the Kaga was the disposal of boiler exhaust. In the Akagi this was done by trunking the boiler uptakes over to starboard, so that the exhaust was vented through one large funnel that was canted downward and through a small upward funnel immediately abaft of it.
At the time of her initial construction in 1923, the Akagi was a flush-deck carrier (i.e., she had no island bridge). Her most unusual feature was the multiple-flight-deck arrangement—somewhat similar to that of the British carrier Furious and the Courageous-class carriers constructed several years earlier—which was thought to accelerate flight operations by making it possible to launch and retrieve aircraft simultaneously. For this purpose, the Akagi had three separated, vertically arranged flight decks. An upper flight deck, 190 meters (624 feet) in length, was the longest and was slightly sloped from a point amidships toward the bow and toward the stern to facilitate landings and takeoffs for the underpowered aircraft of the day. Beneath it was a middle and very short (18-meter [60-foot]) takeoff deck for the smallest and lightest of the carrier’s aircraft. Beneath that, on the main deck, was a somewhat longer (49-meter [160-foot]) flight deck for launching torpedo bombers. The sixty aircraft that the carrier was originally designed to embark were housed in three hangars. As originally completed, there were two long hangar decks at the same levels as the middle and lower flight decks, both of which were used for operational aircraft. There was a third short hangar deck aft and below the two longer hangar decks. The aft elevator serviced the upper flight deck and all three hangar decks. To transport these planes to and from the hangars, the upper flight deck incorporated two elevators, one toward the bow and a smaller one toward the stern.10
Just as these flight-deck and hangar arrangements in the Akagi represented assumptions about the limitations of aircraft performance at the time that she was designed, the carrier’s armament, like that of the American carriers Lexington and Saratoga, reflected contemporary and ultimately mistaken ideas about the heavy surface weapons installed on carriers for use in possible gun action with enemy warships.11 The Akagi carried ten 8-inch (20.3-centimeter) guns, four in twin turrets on each side of the middle flight deck, roughly amidships, and three in single mountings in casemates on each side of the ship. While this firepower was numerically superior to that of the Lexington and Saratoga, the placement of the Akagi’s main batteries actually put her at a comparative disadvantage, since the American carriers mounted their eight 8-inch guns in twin turrets fore and aft of their island superstructures, allowing them to bring all eight guns to broadside, while the Akagi could bring a maximum of only five guns to bear on any single target.
It can be argued that the arming of carriers with heavy surface weapons was logical at the time, considering the tactical and technological situation of the mid-1920s, when carriers had to operate close to the battle line (because of the limited radius of their aircraft) and thus within the range of enemy surface guns.12 Time and a change of opinion by strategists and designers in all the major naval powers were to indicate both the futility and the redundancy of such heavy armament: the vulnerability to shellfire of any aircraft carrier, particularly the flight deck, would make it suicidal for such a ship to participate in a gun action; the ship’s greatest defensive protection came to be seen as the offensive capabilities of her own attack aircraft.13
The layout of the Akagi’s contemporary, the Kaga, was similar to that of the Akagi in many respects. But because the Kaga was originally laid down as a battleship rather than a battle cruiser and thus had a shorter and broader hull, her upper flight deck was 19.5 meters (64 feet) shorter and slightly wider than that of the Akagi. The main flight deck was perfectly horizontal for its entire length. Her propulsion system was smaller and thus allowed a slower maximum speed (a little over 27 knots). Boiler exhausts were vented through two tubes 30 meters (100 feet) long on each side of the ship, an arrangement quite similar to that of the British Furious. These tubes extended horizontally toward the stern, where the funnel mouths were canted downward and outward.
Within half a dozen years of the commissioning of the Akagi and Kaga, advances in naval aircraft, changes in naval opinion, and actual operational experience had led to the need to make drastic alterations in some of their most innovative features as conceived early in the 1920s. By the early 1930s the considerable increase in the size and power of naval aircraft required far more distance for takeoff than that provided by the two lower flight decks on both ships. Surface armament was now viewed as less important for a carrier than the size of her air group. A flush deck may have been preferred by some carrier pilots, but not by others, who considered islands to be useful reference points during landing operations, and certainly not by those who had to conn the ship or control her flight operations. And finally, the funnel arrangements for the Kaga, hotly disputed at the time of their design by naval architects, proved unworkable, since hot gases and smoke still obstructed or disturbed flight operations at the stern of the ship.14
Source: Fukui, Shashin Nihon kaigun zen kantei shi, 1:337.
For all these reasons, the two carriers underwent reconstruction during the mid-1930s at the Sasebo Naval Yard.15 Because of her slower speed, impractical funnel configuration, smaller flight deck, and thus general inferiority to her sister ship, the Kaga was the first to be modernized, even though she had been commissioned a year later than the Akagi. Her extensive reconstruction was completed in one year, from June 1934 to June 1935, while modifications for the Akagi, though far less extensive, took three years, largely because of budgetary delays.16
The single most important modification for both ships was the elimination of the three-stage flight-deck arrangement and its replacement by a single extended upper flight deck that projected over the stern and almost to the bow. This alteration made possible far greater hangar space and thus a considerable increase in aircraft capacity: ninety-one for the Akagi and ninety for the Kaga. The two longer hangar decks—which formerly had also been flight decks—were now extended forward almost to the bow, and the lowest hangar deck was enlarged. The original elevators remained in place and serviced the flight deck and hangar decks as in the ships’ original layout, but a third elevator was added forward and serviced the flight deck and the two long hangar decks. Navigation and control of flight operations were now conducted from a modestly scaled island bridge, constructed on the port side of the flight deck for the Akagi about halfway along the flight deck, and on the starboard side for the Kaga.17
The difference in the location of the island superstructures of these two carriers needs some explanation. Along with the carrier Hiryū (see below), then under construction with her sister ship Sōryū, the Akagi was one of only two aircraft carriers in the world to have an island on the port side. Initially it had been intended that the bridge on both the Akagi, then undergoing modernization, and the Hiryū, then building, would be located on the starboard side of the flight deck toward the bow. A study issued by the Naval Aviation Department, however, suggested that this particular configuration would produce an unacceptable amount of turbulence for flight operations, and thus the island superstructure should be placed amidships. But because the boiler uptakes on the Akagi and Hiryū were located at that spot on the starboard side, it was necessary to shift the bridge to the port side, an arrangement that had the advantage of helping to provide a counterweight to the funnel. Pilots, however, complained that the air turbulence over the deck was actually worse with this arrangement, and they prevailed to the extent that subsequent Japanese carriers had their islands on the starboard side. The Akagi and Hiryū were not modified, probably because there was insufficient evidence that their island configurations caused a significant problem.18
Other modifications common to both the Akagi and the Kaga were the substitution of completely oil-fired for coal-fired and mixed-firing boilers, a substantial increase in fuel oil bunkerage, the trunking of the boiler uptakes into a single funnel that projected downward over the starboard side, the removal of the twin forward 8-inch turrets (in order to extend the single flight deck), and an increase in antiaircraft armament. The Kaga, however, required specific additional improvements: the lengthening of the hull to improve the ship’s drag coefficient, the addition of a new propulsion system of four sets of Kampon equal-pressure geared turbines and new propellers, and the raising of the ship’s metacenter—as part of a wide-ranging effort to correct the serious instability of a significant number of Japanese warships that was diagnosed in the mid-1930s—by adding an additional torpedo bulge that increased the beam. In the Akagi the twin 8-inch turrets were removed, leaving the remaining six 8-inch guns in casemates as the main armament, three on each side abaft the bridge; in the Kaga the four guns were moved to casemates, so that all ten 8-inch guns were retained. This last arrangement demonstrated the continuing contemporary uncertainty surrounding the function and place of carriers in combat. Okada Heiichirō, who as an engineering lieutenant helped draft the plans for the modernization of both the Kaga and the Akagi, tells us that the plans, particularly those concerning the removal of the forward turrets, created a heated debate within the navy. Led by Adm. Yamamoto Isoroku, the Naval Aviation Department, wishing to transform the two ships from functional hybrids into true carriers, urged removal of all the main batteries. The Navy Technical Department, dominated by gunnery-minded officers, argued that removal of such armament would leave the carriers vulnerable in any surface engagement. The compromise seems to have been reached by simply removing the forward turrets so as to conform to the new flight-deck arrangements.19
Soon after the Akagi and Kaga went into service in the late 1920s, fleet maneuvers made obvious the offensive potential of carrier aviation. Staff college war-gaming further underscored the possibilities of the carrier weapon, as did well-publicized advances in American carrier aviation. Thus, Japanese navy leaders became convinced of the need for more carriers. But budgetary limitations, and the fact that the construction of the Akagi and Kaga had already used nearly 54,000 of the 81,000 tons allocated to Japan for carriers by the Washington Treaty, meant that only one or two fleet carriers could be constructed within the remaining 27,000 tons. Accordingly, the navy attempted to build an effective carrier at less than 10,000 tons, in order to have it exempted, along with the Hōshō, in the calculation of Japan’s allowed carrier tonnage. These considerations governed the design of the Ryūjō, the navy’s third-generation carrier and the second Japanese carrier originally designed as such.20 Her designers, determined to keep her displacement under 10,000 standard tons, drew up plans that originally called for an 8,000-ton ship. Such size limitations resulted in the design of a flush-deck carrier with boiler uptakes trunked over the starboard side, no heavy surface armament, almost no armor protection, and provision for only one hangar, capable of housing twenty-four aircraft.21
But even while the design was on the drawing boards, various studies had determined that such a small number of aircraft would comprise the effectiveness of the air group. For this reason, the design was altered to include a second hangar deck above the first, so that forty-eight aircraft could be embarked. This addition increased the Ryūjō’s displacement to almost 12,500 standard tons—a fact kept secret at the time—and, what was worse, added too much weight above the waterline, making her less stable. A year after her launching in 1933, a well-publicized capsizing incident in the navy, involving the torpedo boat Tomozuru, brought home the danger posed by this defect.22 The Ryūjō was therefore brought into the Kure dockyard for the required modifications, which included incorporation of new hull bulges, the addition of a heavy ballast keel, and removal of some antiaircraft batteries. But even these modifications proved insufficient to improve the ship’s seaworthiness. Caught in the great storm that battered the Fourth Fleet off the coast of Honshū in 1935, the Ryūjō shipped huge amounts of water in the heavy seas because of her low forecastle, a fact that again raised questions about her stability.23 Once more the Ryūjō entered drydock, this time to have the front of her bridge reshaped to lower wind resistance and to have her forecastle deck raised by adding an extra deck.24 While these changes did little to improve the overall efficiency of this “minimum carrier” (the words are Norman Friedman’s), her defects of size did provide salutary experience for the design of the navy’s next set of carriers.
The two ships of the Sōryū class had their origin in the desire of the Navy General Staff to circumvent, legally, the treaty limitations on total carrier construction. Taking advantage of the definition of an aircraft carrier as a warship designed primarily for aircraft operations, the staff conceived of a hybrid warship that was as much a cruiser as it was a carrier, thus not to be counted in the total carrier tonnage allocated to Japan. Drawn up in 1931–32, the initial designs for this class had therefore incorporated not only a flight deck but also cruiser-class (6-inch [15.2-centimeter]) armament mounted in barbettes forward of the flight deck.25
But even as these plans were being drafted, Japan had begun to consider withdrawal from the treaty system by the end of 1936. As Japanese renunciation of the treaties would make irrelevant the question of keeping naval tonnage within any limitations, the question of the 81,000-ton limit on carrier construction would become moot. On this assumption, the Navy General Staff set forth its requirements for two new carriers, using the general hull configuration and flight-deck arrangement of the hybrid design, but eliminating the cruiser armament.26 The Sōryū, laid down in 1934 and completed in 1937, met most of these requirements, and though she exceeded the planned displacement by 5,000 standard tons, her hull was very light for a ship of her size. Her powerful cruiser propulsion system provided an extraordinary maximum speed of 35 knots. The Sōryū’s design retained the double-deck hangar system capable of accommodating sixty-eight aircraft (fifty-one operational and seventeen reserve), but the lower and shorter of the two hangars was placed well within the hull for greater stability than afforded by the Ryūjō’s layout. The ship was provided with three elevators, one forward and two smaller elevators amidships and aft. In what was becoming a trademark of Japanese carriers, the boiler uptakes were trunked over the starboard side amidships in two downward-venting funnels. Just forward of these, on the starboard side, was located a small island superstructure.27
In July 1936, six months after the Sōryū’s launching, the keel was laid for her sister ship, the Hiryū. But by this time experience and research seemed to argue for some fairly basic changes in the design of the class. Taking a cue from the near disaster that had threatened the Fourth Fleet in the great storm of 1935, the Hiryū’s design was modified by strengthening the hull, raising the forecastle, and increasing the beam for greater stability. Additional armor was contributed to the ship’s main belt (though the horizontal protection was the same as the Sōryū), and the antiaircraft battery was redistributed.28 Although she normally operated sixty-four aircraft during the interwar period, the Hiryū’s maximum aircraft capacity, seventy-three, was slightly larger than that of the Sōryū. The Hiryū’s island superstructure was located, like the Akagi’s, on the port side of the ship, just forward of the starboard side funnels, an arrangement that, again as with the Akagi, was eventually found to be troublesome for flight operations.
By the end of the treaty era, therefore, the Japanese navy had completed four aircraft carriers and had two more building. From their specific characteristics and capabilities, it is possible to trace the evolution of Japanese carrier design, operating doctrine, and operating efficiency for the period from the end of World War I to about 1937. The Hōshō, as we have seen, had been an experimental ship whose design had been influenced by British experience during World War I, but whose small size restricted the number of aircraft she could embark and thus limited her offensive potential. But the fourfold jump in displacement from the Hōshō to the Akagi and Kaga had less to do with Japanese perceptions of the need to increase aircraft capacity than with the fact that the Japanese navy had available capital ship hulls on which work had been suspended as a result of the naval agreements at Washington in 1922. Indeed, as several studies have suggested, in the mid-1920s the Japanese navy, like the American navy, believed that small carriers—10,000 tons or less—in larger numbers would provide greater aircraft operating capacity. Given the vulnerability of carriers, such a dispersal of carrier air power would lessen the impact of the loss of any single carrier.29 If true, this belief may have been nearly as much a factor in the design of the small carrier Ryūjō as the desire to keep her displacement under 10,000 tons and thus exempt her from the Washington Treaty limits. Yet the Japanese and American navies later came independently to the conclusion that a carrier of 10,000 tons was ineffective as a combat unit because it could not operate enough aircraft. In the 1930s, other factors were to favor the increased construction of larger and faster carriers, including the continuing need for longer flight decks to allow the takeoffs of larger and heavier aircraft.30
CARRIER CONSTRUCTION IN THE POST-TREATY ERA
The end of the naval arms limitations treaties in December 1936 allowed Japan to build ships of all classes, including aircraft carriers of unprecedented size and performance. During the five years between the end of the treaty era and the onset of the Pacific War, the navy added the two finest carriers it ever built: the Shōkaku and Zuikaku. Indeed, some have argued that they were the two most successful warships ever built for the navy. Ordered in 1937 under the “Circle Three” construction program to counter growing U.S. carrier strength, the Shōkaku class was conceived as part of a fast carrier group capable of operating with the monster Yamato-class battleships; such missions called for large, fast carriers, capable of embarking a powerful aerial strike force and able to defend themselves effectively. The General Staff specifications called for the same aircraft complement as the remodeled Akagi and Kaga (seventy-two aircraft, with twenty-four in reserve); the same speed (34 knots) as the Sōryū; a greater radius of action (9,700 nautical miles at 18 knots) than any of their predecessors; and hull armor to withstand an 8-inch (20.3-centimeter) shell fired at 12,000–20,000 meters (13,000–22,000 yards) or an 800-kilogram (1,764-pound) bomb launched from a high-level bomber.31
As the two ships were almost identical, were laid down within six months of each other, and were completed a month apart, a discussion of the arrangements for the namesake of the class is appropriate for both carriers. Designed with the experience gained in operating the navy’s four other large fleet carriers, the Shōkaku, in a number of respects, particularly in profile, was a bigger and more heavily armed and armored version of the Sōryū. Though her flight deck was planked and unarmored, she was 10,000 tons heavier than the Sōryū because of the armor on her main deck above her machinery, magazines, and aviation gas. Nevertheless, flooding calculations made at the time of her construction ensured that she had the stability equal to that of battleships of the day. Her island, like the Sōryū’s, was well forward on the starboard side; abaft on that side were two funnels trunked downward. Three elevators brought the aircraft up from her two hangar decks to her flight deck. All three decks were equipped with fueling stations, so that aircraft could be refueled on all three decks. Her closed-in hangars presented the same fire hazards, however, as those of other Japanese carriers (see below). Her armament—eight twin 5-inch (12.7-centimeter) dual-purpose gun mounts controlled by four directors, as well as a light battery of forty-two 25-millimeter antiaircraft guns in triple mounts—was certainly more formidable than that of the Sōryū. Her engines, which delivered 160,000 horsepower, 10,000 more than even the Yamato, were the most powerful ever installed in a Japanese warship. By the time of her participation in the Hawai’i strike, the Shōkaku, though embarking slightly fewer aircraft than demanded in the specifications, provided a formidable offensive capability: twenty-seven Aichi D3A dive bombers, twenty-seven B5N torpedo bombers, and eighteen Zero fighters, seventy-two aircraft in all (not counting the twelve reserve aircraft).32
The Shōkaku and Zuikaku were critical additions to the Japanese carrier fleet. In performance and capability they exceeded all American carriers until the wartime appearance of the Essex-class carriers. So valuable were these two warships that the decision to mount the Pearl Harbor strike was based in part on their availability to the mobile task force organized for that operation.33 Certainly their service record lived up to the care that went into their design and construction; on the American side their battle honors were matched only by the Enterprise. By the time they met their ends in 1944, they had participated in every carrier battle except Midway, where their absence may have been the margin between victory and defeat.34 Indeed, their ubiquity, performance, and durability in more than two years of fierce combat in the Pacific led Admiral Nimitz to aver with some heat that the happiest day in his life would be when “those two ships” were finally sunk.35 It can be said that when that time came, their destruction spelled the end of the Japanese carrier fleet.
Before leaving this discussion of the design and construction of Japan’s carrier fleet, I should touch upon the navy’s effort to supplement its publicly announced carrier construction programs by the creation of a “shadow fleet” of auxiliary and merchant vessels that could be quickly converted into carriers in case of a wartime emergency.
Two practical considerations shaped this effort. The first of these had to do with the tonnage restrictions imposed by international agreements. After laying down the Sōryū and Hiryū, Japan had reached (and indeed exceeded) the carrier tonnage that it could build under the terms of the Washington Treaty. The second consideration anticipated the prospect of an unrestricted building program in a post-treaty era, the navy recognizing that the nation had limited shipyard capacity for keel-up naval construction.
For these reasons, early in the interwar period the Japanese navy, sometimes assisting and subsidizing private shipbuilding firms, began to design and construct both naval auxiliaries and merchant vessels suitable for easy conversion to light fleet carriers. The navy had begun this “shadow” building program even before the London Naval Treaty.36 In the late 1920s it worked with the Nihon Yūsen Kaisha on the design of three fast passenger vessels subsidized by the Ministry of Transportation and built to navy specifications for easy conversion to carrier configuration. In the event, these particular ships were launched and completed but were never actually converted into aircraft carriers.
The first ship of this “shadow fleet” that was built and actually converted to a carrier (during the Pacific War) was the 10,000-standard-ton submarine tender Taigei, on which conversion work was secretly started in 1941 and completed in November 1942, when she was renamed the carrier Ryūhō. But her small flight deck, insufficient hull strength, and inadequate diesel propulsion system (later replaced by destroyer turbines) prevented her from ever becoming an effective fleet unit. In 1934, under the “Circle Two” naval construction program, the navy ordered four auxiliary vessels that could be converted into submarine tenders, fast oil tankers, or light fleet carriers, as needed. The Tsurugisaki, the first of these ships laid down, was completed as a submarine tender in 1939 but converted into a carrier in late 1941 and renamed Shōhō. Her sister ship, the Takasaki, laid down in early 1935, was commissioned as the carrier Zuihō in 1940. The other two vessels, the Chitose and Chiyoda, were laid down as seaplane carriers and converted to flush-deck carrier configuration during the Pacific War. As a group, these conversions, though quickly completed for the most part, were largely unsuccessful because of their low speed, small aircraft capacity, and inadequate protection. The Zuihō, with her destroyer-type turbines, was probably the best of the lot and was the only one of this group that had joined the fleet when the war began, serving in five of the major carrier operations of the Pacific War.37
It remains only to mention in passing those aircraft carriers under construction as hostilities began in December 1941. None proved to be a durable asset to the navy in the coming struggle, but the most promising of the group was the 30,000-ton Taihō, laid down in July 1941 and launched in April 1943, the first Japanese carrier with an armored flight deck. To that extent, she appears to have been patterned on the British Illustrious, which she closely resembled. One of the last big fleet carriers to join the Japanese fleet, in her short life the Taihō was never to have the opportunity to test her flight deck against aerial bombs. Completed in March 1944, she was sunk by a submarine three months later.38
Conversion work on the hulls of two new 27,000-ton luxury liners of the Nihon Yūsen Kaisha began in October 1940, and they were launched in June 1941 as the Hiyō and Jun’yō, joining the fleet in the summer of 1942. But their hybrid propulsion system of destroyer-type boilers mated to turbines designed for mercantile use provided inadequate power and thus made them too slow for fully effective fleet operations and subject to frequent breakdowns. The Taiyō was another carrier converted from an NYK liner under construction in the months immediately prior to the Pacific War. Intended as a convoy escort carrier, she was to prove inadequate to the task after she joined the fleet. Her sisters, the Chūyō and Un’yō, converted after the war began from NYK liners already in service, proved similarly ineffective.39
IMPLICATIONS OF JAPANESE CARRIER DESIGN
The single greatest factor affecting the difference in the design and operation of Japanese and American carriers during this period was the method of calculating maximum aircraft capacity. In the Japanese navy, as in the Royal Navy, aircraft capacity was determined by the size of the hangar rather than by the size of the flight deck. While from the beginning American carriers normally parked most of their aircraft on the flight deck and used the hangars below only for aircraft repair and maintenance, Japanese carriers used their hangars as their main storage area, as well as for servicing, refueling, and reloading ordnance (although rearmament of dive bombers was often performed on the flight deck). Such an arrangement had important implications for both aircraft capacity and flight operations. As flight-deck storage provides considerably more aircraft space, Japanese carriers of roughly the same displacement as American carriers usually had significantly smaller air groups. Considering the numerical disadvantage that the Japanese thus imposed upon themselves, it may seem odd that Japanese aircraft designers did not place more emphasis on folding wings for their carrier aircraft. Of the three main types of Japanese carrier aircraft at the time of Pearl Harbor, only one, with a considerable wingspan, the Nakajima B5N carrier attack aircraft, had folding wings. The other two—the Aichi D3A carrier bomber and the Mitsubishi A6M2 carrier fighter—had wings that folded only at the tips, an arrangement made out of a concern for keeping the wingspan within the dimensions of the elevators, not out of a concern for saving hangar space. (Fig. 3-1.) The incorporation of folding mechanisms into the wing structure was believed to add too much weight and thus to represent an unacceptable sacrifice in performance.40
Yet if Japanese aircraft carriers built in the 1920s and early 1930s differed from American carriers in their arrangements for aircraft storage, they shared with their American counterparts the common characteristic that their design sacrificed defensive protection for the ships themselves in order to maximize the offensive potential of their aircraft. While it is true that the Akagi and Kaga retained some of their original capital ship armor, it was drastically reduced during their conversion. On the Sōryū and Hiryū, armor was limited to modest protection over machinery, magazines, and aviation gasoline tanks, while on the Ryūjō it was practically nil. As in the case of American carriers, the flight decks were superimposed upon the hull rather than constituting a strength deck that formed an integral part of the hull, as in the carriers of the Royal Navy of the period.41 Because their designers were unwilling to make the sacrifices of speed, weight, or space that substantial horizontal protection would have required, the flight decks of prewar Japanese carriers consisted of wooden planking laid lengthwise over thin steel decks (U.S. carriers had their deck planking athwartships). Nor were the hangars of prewar Japanese carriers armored, though they were enclosed by storerooms, so that their aircraft and ready crews were shielded from wind and weather. (Fig. 3-2.)
Fig. 3-1. Wing-folding arrangements for Japanese carrier aircraft
Such arrangements entailed operational consequences that ranged from inconvenient to outright dangerous. Because of ventilation problems, aircraft could not normally be warmed up in the hangars. Much worse, as evidenced by the damage suffered by Japanese carriers during the Pacific War, when enemy bombs penetrated the unarmored flight decks and exploded in the hangars, the effects of the resultant blast pressures could be disastrous: the flight deck blown apart and the hangar sides buckled. Enclosed hangars also prevented the easy disposal of fuel and ordnance over the side in the event of a hangar fire, and effectively prevented fighting such a fire with hoses from any screening ships that might be in a position to act as fireboats. Moreover, as the hangars were neither vapor-tight nor flash-tight, they were vulnerable to explosions from ordnance or aviation fuel.42
Japanese aviation fuel arrangements were also suspect. While even early-war carrier aviation gas tanks were surrounded by spaces filled with inert carbon dioxide gas, the tanks themselves were incorporated into the structure of the vessel.43 This meant that shock stresses imparted to the hull could be directly transmitted to the tanks, causing them to crack and leak, sometimes disastrously. Late-war carriers attempted to remedy this problem by filling the void spaces with reinforced concrete, but this measure was only partly successful. Furthermore, there is no indication that early-war Japanese carriers were capable of filling their fuel lines with carbon dioxide, as became American practice soon after the outbreak of hostilities.44
In addition, while Japanese carriers incorporated elaborate firefighting equipment, including fireproof curtains, banks of foam dischargers, flooding systems, and armored damage-control stations throughout the hangar decks, the very nature of enclosed hangars and the vulnerable fueling system design meant that the damage mitigation measures were largely doomed from the start.45 This fact, combined with a mediocre level of damage-control training and organization (see chap. 6), meant that Japanese carriers were unduly prone to aviation fuel fires, a defect that contributed in one degree or another to the loss of the four carriers at Midway as well as to the later sinkings of the Taihō and Shōkaku.46
From the beginning of carrier aviation, it was important in the Japanese navy, as it was in its American and British counterparts, to control the movement of aircraft, particularly landing aircraft, across carrier flight decks. The first necessity was to bring landing aircraft to a secure stop, a function they were unable to perform entirely on their own. Early in the 1920s the Japanese navy adopted the British system of longitudinal wires across the flight deck to slow incoming aircraft, but this system too often failed to stop aircraft attempting to land. By the 1930s, Japanese carriers switched to the far more effective French system of transverse wires to engage the arrester hooks of landing aircraft.47 As in the U.S. and Royal navies, Japanese carriers also made use of a crash-barrier system to prevent aircraft that missed or failed to be restrained by the arresting gear from smashing into aircraft spotted forward. This system consisted of a set of three wires (in later carriers, two sets) about 2 meters (6 feet) high, strung one above the other across the deck between supports on either side of the flight deck, each of which was attached to a hydraulically operated sliding restraint. When an errant aircraft hit the wires, they were pulled forward and were ultimately stopped by the hydraulic mechanism, taking up the shock and stopping the aircraft. If aircraft were to be parked forward of the barrier, it could be lowered for each incoming aircraft and reset for the next.48 Arresting gear was fitted both fore and aft, indicating that just as in U.S. practice, Japanese fleet carriers were equipped to conduct emergency landing operations over the bow with the ship going astern.49
Fig. 3-2. Interior of the Hiryū’s upper hangar deck, aft
As Hugh Popham has written in his fine study of British naval aviation, all carrier flight operations are governed by two determinants: space and time.50 These determinants were particularly severe in the days of cyclic flight operations before the “flex deck” operations of the post–World War II era, when carriers came to have ready decks for both landings and takeoffs. Popham’s additional point that in a carrier space is organized vertically, rather than horizontally, was truer of the British and Japanese navies than of the American navy because the British and Japanese tended to store most of their aircraft below, on the hangar deck, and to do most of the servicing of aircraft in hangars rather than on the flight deck. This arrangement thus made elevators more central to the spotting, maintenance, and storage of the carrier’s aircraft than on American carriers—which parked most of their aircraft on the flight deck—and thus had a greater impact on the duration of flight cycles. In preparation for a flight cycle, aircraft in the hangars had to be manhandled onto the elevators, off them, and then pushed or pulled down the deck, where they were properly spotted for launch. The laborious process—tricky and even dangerous in foul weather—was reversed when, at the end of a recovery cycle, landed aircraft at the forward end of the flight deck were struck below.
Flight operations on Japanese carriers were directed from a platform located on the rear of the island bridge by an air operations officer (hikōchō). A launch began with the carrier headed into the wind,51 and aircraft to be used on a particular mission properly spotted on the deck with engines run up and chocks still under the wheels. The air operations officer checked the wind-over-deck determined from a speed indicator at his side, and when the wind speed was judged sufficient—14 meters per second, or approximately 27 knots—he would have a white flag with a large black ball hoisted. (In low-light conditions, as during the launching of the first attack wave from the Akagi at the Battle of Midway, the procedure may have involved a lamp.) The air operations officer would then give a signal to a flight-deck officer, who would wave away the chocks; the deck crew would drag the chocks from the wheels and scurry to the sides of the flight deck. The air operations officer would then raise a white flag from the bridge as a signal for launch, and the first aircraft would move up the deck. The pilot judged wind direction during takeoff by means of a steam jet trailing vapor at the forward end of the flight deck. As long as the white flag flew from the bridge, aircraft took off at twenty- to thirty-second intervals.52 A minimum takeoff length of 70 meters (230 feet) was required by the Zero fighter plane; heavier aircraft required nearly twice as much takeoff distance.53 Catapults, while used to launch floatplanes from other aircraft-carrying vessels, were never installed on any Japanese carriers.54
In general, the Japanese did not use a landing signals officer (“Paddles,” in the U.S. and Royal navies) to recover aircraft, though a crewman (seibiin) under the supervision of the air operations officer was stationed aft to signal to the aircraft as required. The air operations officer would hoist a red flag in case of a fouled deck to signal a wave-off. A set of standardized signals was developed to communicate with incoming aircraft. Carrier identities themselves were indicated by a large “kana” symbol painted on their flight decks (usually on the port quarter, but sometimes on the port bow). By day, a black ball with two numerical flags under it was hoisted, indicating that the carrier was ready to receive aircraft and giving the wind velocity over the deck in meters per second. Wind direction was indicated by the steam jet at the front of the flight deck. A red flag from the seibiin signaled that the pilot should go around again; a white flag with an H meant that the aircraft’s hook was not lowered.55
At night these signals were replicated via a series of colored lights. A white light in the center of the landing array could be used to blink messages to the incoming pilot. A line of red, blue, and white lights aligned fore and aft just abaft the island were used to indicate relative wind over the deck. The outline and centerline of the flight deck were illuminated in white lights, while crash barriers and the aft edge of the deck were indicated by transverse rows of red lights. Each bank of lights could be dimmed independently by rheostats.56
Aircraft waiting to land formed a group to starboard of the ship, circling to the right. When it was time to land, the next aircraft in the cycle would break left, circle the ship across the bow and down the port side, then turn in to the left and line up approximately 700 meters (2,000 feet) astern at approximately 200 meters’ (700 feet) altitude.57 Each pilot would establish his glidepath by lining up his incoming aircraft with a series of adjustable and differently positioned red and green lights (chakkan shidōtō, “landing guidance lights”) on either side of the deck, an arrangement similar to the “call the ball” system later developed by the British and used on American carriers. (Fig. 3-3.) Each pilot flying in astern of the carrier with wheels, flaps, and arrester hook down would have to adjust his flight path until he had both pairs of lights in his sight, a perfect horizontal lineup for landing. To determine the correct height for his approach to the carrier’s deck, he adjusted his glide angle according to the correct vertical position of the colored lights in his sight. A perfect glide slope would have the green light positioned immediately above the red. If he could see only the red light, he was below the perfect slope, and if the red light was over the green, he was coming in so low that he would smash into the carrier’s stern if he didn’t pull up. If the green light was positioned far above the red, the pilot was coming in too high and risked missing the arresting gear and hitting the crash barrier, or worse.58 He was further aided in this process by the white-painted outrigger platforms near the aft end of the flight deck. These platforms helped him gauge the orientation of the flight deck even if the deck itself was obscured by the nose of the aircraft.59
Fig. 3-3. Japanese carrier landing operations
The pilot aimed to maintain an airspeed of approximately 10 knots above stall speed, which typically translated to 70–75 knots.60 Assuming he had kept the lights lined up properly, just prior to crossing the after end of the flight deck the pilot would cut his engine, and gliding in, his hook would catch the arrester wire and stop the aircraft. He would then release the hook and taxi forward.61 The crash barriers would be lowered,62 the aircraft moved to the forward end of the flight deck, and the barriers reset for the next aircraft.
This technique of “continuous stowage” (renzoku shūyō)—the recovery of landing aircraft in quick succession by respotting the just-landed planes forward of the crash barriers—was regularly practiced aboard Japanese carriers by the mid-1930s. By the Pacific War, Japanese carriers could land aircraft every twenty-five to forty-five seconds. When necessary, they were pushed back from the forward end of the flight deck to the after end and serviced there in preparation for the next launch.63 But in most cases, once a recovery cycle had been completed, the aircraft were struck below, to be serviced and armed. It was this procedure, governed by elevator cycles,64 that kept Japanese carrier flight operations at a slower pace than those of their American opposites, who in general performed most refueling and rearming functions on the flight deck. This slower pace, combined with all the hazards brought about by the proximity, in a confined space, of highly combustible fuels and ordnance, held a potential for disaster. There was always the possibility that in urgent combat situations, the delay implicit in the slower flight cycles of Japanese carriers might cause proper rearming procedures to be cast aside in favor of greater speed in rearming and refueling. Again, the often-told loss of the Akagi and Kaga at Midway is testimony to this ominous problem.65
As in the other two carrier navies, the hazards of operations on and off a Japanese carrier were considerable and crashes too frequent an occurrence. Okumiya Masatake, who learned to fly off carriers aboard the Ryūjō, recalled in later years how very small the carrier looked coming in to land, and how easily a pilot could misjudge his approach despite the best guidance systems. It was difficult but essential, for example, to compensate instantly for the slightest deviation between the carrier’s heading and the wind direction as indicated by the steam jet on the flight deck. In a side wind of 10 meters (33 feet) per second, a second of inattention by a pilot approaching the carrier’s stern would cause the aircraft to be blown 10 meters sideways. With the Ryūjō’s deck only 23 meters (76 feet) wide, this would mean that one of the aircraft’s wheels would miss the flight deck, and the plane would plummet into the ocean. Only the selection of carrier pilots with instant reflexes and great enthusiasm, near perfect teamwork between air and deck crews, and constant and rigorous training kept the occasional crackups on deck and splashes into the sea from becoming a chain of disasters.66 Of course, in flight operations there were always a few aircraft that went over or completely missed the deck and fell into the ocean. To minimize personnel losses in this event, Japanese carriers engaged in flight operations, like their American counterparts, always trailed one or two guard destroyers astern, ready to rescue downed aircrew.
Source: Fukui, Shashin Nihon kaigun zen kantei shi, 1:345.
EMERGENCE OF JAPANESE CARRIER DOCTRINE
In the early days of carrier aviation, the justifications for building carriers were speculative and the designs for constructing them were necessarily experimental. The variety of carriers produced by the Japanese navy in this period is ample testimony to these conditions. As in the U.S. and Royal navies, carrier doctrine was quite tentative. It developed sporadically on the basis of both the demands of the Navy General Staff for superior warships and the practical results of Japanese carrier operations in these years.
When the Hōshō was first added to the Japanese fleet and the Akagi and Kaga were being constructed as carriers, the principal functions of naval aviation were spotting, reconnaissance, and antisubmarine patrol. There was as yet little thought given to the use of carrier aircraft as important components in offensive operations, and thus little consideration of carrier doctrine. With the formation of the First Carrier Division in 1928 and with the availability of more than one carrier, serious consideration began of the study of the role and place of carriers in a fleet engagement.67 Yet even then, carrier aviation was seen as a force multiplier, not as a decisive element of sea power on its own. The chief function of carrier aircraft was still tactical and defensive in nature: to secure “command of the air” over the surface battle area, increasing the effectiveness of friendly battleships over their enemy counterparts. Of course, at this point there was little possibility of carriers operating in a long-range strategic role, largely because of the limited effective range of carrier aircraft, which in the late 1920s was not much more than 100 miles.68
In the early 1930s, therefore, the navy still placed emphasis on the tactical use of carrier air power, particularly in support of the main battle force. For example, a principal function of carrier attack aircraft during this period was cooperation with the surface batteries of the main force and the laying down of aerial smoke screens for supra–smoke screen fire by battleship main batteries, a role that reflected the continuing “big-gun” priorities of the navy. Moreover, in the fleet maneuvers from 1928 to 1935, which always pitted a “Red Fleet” representing the Americans against a “Blue Fleet” representing Japan, carrier fighters were employed mainly in repelling enemy spotter aircraft and in protecting friendly target observation planes. Japanese naval tacticians had come to believe that for one side or the other to prevail in such a contest, a major air engagement would have to be waged for control of the air over the opposing fleets, a duel that would precede the showdown surface battle. Because their performance was so limited at this time, carrier fighters were conceived of only as defensive aircraft over the fleet, but always in direct support of fleet units engaged with the enemy’s main surface forces. Not unreasonably, therefore, by the early 1930s naval aviators, for their part, came to see the enemy’s carriers as the main target of carrier offensive air power, whereas the navy’s capital ship gunnery staffs continued to believe that carrier strikes should have enemy battleships as their priority targets.69
Thus, early in the 1930s there was as yet no unified doctrine in the Japanese navy as to how carriers were to be used in a fleet action. The Battle Instructions (Kaisen yōmurei) of the navy,70 which had governed fleet operations since roughly 1900, were ambivalent on this issue even as late as the 1930s, though they appeared to lean toward the “battleship first” principle, as before. They declared that “attacks on the enemy main force [his battle line] should be timed at the opportunity for the decisive fleet engagement,” and they urged “attacks on enemy carriers at the outset of such an encounter so as to immobilize them in the ensuing [surface] battle.” Such equivocation was part of a larger ambiguity of the instructions concerning the role of air power in naval warfare. Because they eventually became so out-of-date in relation to air power, in 1933 the Navy General Staff set up a committee to study the problem of revision of the instructions. Its members were so divided over the issue of air power that the committee made little progress. Eventually, since the whole matter had become controversial, what emerged from the committee in 1934 was a set of vacillating and far narrower guidelines, “Draft Instructions for Air Combat” (Kōkūsen yōmu sōan), which satisfied no one and was regarded with contempt by the navy’s air officers.71
With every increase in the range and power of carrier aircraft, the most salient characteristic of carrier air power came to be seen as its ability to reach out beyond the range of existing surface weapons. It was inevitable, therefore, that Japanese gunnery staffs as well as naval airmen should begin to argue that the reconnaissance capabilities of naval air power should be combined with its growing offensive potential to perfect the concept of a preemptive aerial strike at the enemy’s aircraft carriers.72 It was only by the elimination of the enemy’s carriers that Japanese forces could achieve air superiority over the area of the surface battle. Beginning about 1932–33, therefore, the Japanese navy began to shift its initial aerial targets from the enemy’s battleships to his carriers, and by middecade, with the heightened performance of bombing aircraft, particularly the dive bomber, the destruction of the enemy carrier force became the focus of intensified research and practice by the naval air forces. As I shall demonstrate in a later chapter, the navy’s increasing tactical skill in offensive air operations in the late 1930s was to turn its confidence in naval aviation into an article of faith. But what is important to understand here is that by the mid-1930s, among navy airmen the proper exercise of air power was seen to be less in the direct support of surface forces than in seeking its own opportunities to destroy an enemy who might be over the horizon.73 In 1937 the navy worked out a scheme of air operations against the U.S. Pacific Fleet that included the following principle: air operations were to take place before the decisive battle, and such operations would have as their objectives both the command of air space over the battle zone and the destruction of the enemy’s carriers and battleships.74
The essential condition for that destruction was that Japanese naval air forces be able to strike first, before Japan’s carriers came within range of the enemy’s carriers. That in turn depended on the Japanese ability to “outrange” the enemy in the air, just as Japanese surface forces were hoping to do by shell and torpedo on and below the ocean’s surface. This meant that the Japanese navy would have to develop attack aircraft that could “outrange” those of the United States Navy. It was for this reason that throughout the 1930s the Japanese navy came to emphasize range in its specifications when calling for bids for aircraft design. This emphasis was driven home in the design of Japan’s front-line carrier attack aircraft in service as the Pacific War opened, and it continued throughout that conflict. Because of their lighter construction, Japanese carrier attack aircraft could indeed search and attack at greater ranges than their American opposite numbers. During the Pacific War, U.S. searches were generally limited to 325–360 miles, while Japanese naval aircraft often searched to ranges of 560 miles. American carrier aircraft could attack to about 200–250 miles, whereas their Japanese counterparts could strike up to about 300–350 miles. Yet as Eric Bergerud has pointed out, during the war both sides learned from experience that in flying any given aircraft in combat, it was essential to allow for a large margin of error rather than flying to its combat radius (the maximum distance the aircraft could fly, perform its mission, and return to base with sufficient fuel to land safely).75
The realization of “outranging” enemy naval air power thus gave substance to the concept of preemptive mass aerial attack. This in turn meant the freeing of Japanese carriers from having to provide direct support to the main battle force. Now they might operate in the vanguard of the fleet as a means of strengthening the fleet’s reconnaissance capabilities, acting either independently or in cooperation with cruiser forces.76 Yet as I noted earlier and shall discuss again in a later chapter, one difficulty with this concept was that Japan’s carrier forces themselves did not devote adequate training and attention to the reconnaissance mission.77
Moreover, the concept gave rise to a difficult problem that was to take a good number of years to solve: whether to concentrate or to disperse the fleet’s carriers in battle, an issue highlighted by Japan’s continuing numerical inferiority under the unfavorable ratios that had been imposed by treaty limitations. The nub of the problem lay in the fact that while carriers possessed great striking power, they were also extremely vulnerable. Unlike a battleship, a carrier did not need to be sunk in order to be eliminated as a fighting unit. Since the flight decks of both Japanese and American carriers were unarmored, destruction of these decks would essentially neutralize these warships. True, intensive navy studies in the 1930s pointed to the difficulty of judging at what point bomb damage would neutralize a carrier beyond repair. Yet it is hard to see how a carrier could remain effective if its aircraft elevators were put out of action.78
Although their offensive potential was best realized by concentrating them in order to launch devastating mass aerial attacks, by the same token, should all the carriers be surprised and struck, they could be annihilated at one blow, a possibility that was enormously increased in the second half of the decade with the advent of dive bombing. On the other hand, if carriers were dispersed in order to minimize the impact of any one enemy aerial assault, not only would the ability of individual carriers to defend themselves be lessened, but their weaker formations of attack aircraft would be more likely to fall to enemy fighters, and because of factors of time and fuel, their attacks would be more sporadic. It was difficult, moreover, to concentrate aircraft groups from carriers not within sight of one another. In the early 1930s, when the navy had only two fleet carriers, it was impossible to test the relative merits of concentration and dispersal in fleet maneuvers, so that much theorizing was tested in the form of table-top war-gaming. By 1936, however, on the basis of fleet maneuvers that year, as well as of map exercises, most Japanese tacticians had come to favor the doctrine of carrier dispersal, so that a numerically superior foe could not destroy all the Japanese carriers at once, even though they recognized the disadvantages that this dispersal might bring.79 Yet as I shall explain in chapter 6, the doctrine of carrier dispersal, as the Japanese navy conceived it in 1936, gave scant attention to the problem of fleet air defense. Given the meager offensive power of carrier aircraft in the mid-1930s, this may have been understandable, but fleet air defense continued to be a fundamental weakness for Japanese carrier formations straight into the Pacific War.
Contemporary Japanese tactical thinking concerning the optimum employment of dispersed carrier air power was set forth in a Staff College study in November 1936. The study proposed that smaller Japanese carriers were to steam in a dispersed formation, and large fleet carriers were to operate alone. Dispersed formations were to be deployed so as to encircle the enemy, enabling individual carriers to concentrate their attacks on the enemy fleet without themselves becoming a massed target. Yet with Japanese carriers so far out in front of their own main battle force, serious losses were to be expected by the most advanced units. In the words of the study, the leading Japanese carrier “must be prepared to be impaled as it impales the enemy” (gūshi suru no kakugo aru o yō su). Air attacks against the enemy fleet were to be launched at a range of 350 miles, and all attack aircraft were to be committed by the time Japanese carriers were 250 miles from the enemy. The study argued that if Japanese attack aircraft at the time of hostilities had no better range than their American opposites, their bomb loads were to be reduced, and midair-refueling (kūchū nenryō hokyū) techniques were to be developed in order to give them the edge. (There is no evidence that the Japanese navy ever actually practiced such techniques.) The key to success was to be mass attacks, delivered preemptively because of the advantages of surprise and of “outranging” the enemy. Given the Japanese navy’s limited confidence in the ability of fighter aircraft to defend fleet units against the assaults of enemy torpedo and dive bombers, it is not surprising that the study said almost nothing about the maintenance of combat air patrols over Japanese carriers. On the other hand, if the enemy’s offensive carrier power was destroyed at the outset, air defenses would be unnecessary. In any event, until the advent of radar and its contribution to efficient vectoring of defending aircraft, carrier-borne fighters could not hope to defend their carriers effectively against incoming enemy air attacks.80
The general composition of Japanese carrier air groups during these years reflected these ideas. It is difficult to be precise about the composition of such air groups, in part because individual carrier hangar configurations changed with the modernization and refitting of individual carriers, and in part because postwar reference works on Japanese naval aviation often tabulate significantly different data concerning air groups. Nevertheless, two principles in the composition of Japanese carrier air groups seem to be constant throughout the period.
First, in keeping with the offensive priorities of Japanese carrier doctrine, carrier air groups were weighted in favor of attack aircraft—dive bombers and torpedo bombers—over fighter planes for defense. Second, in the mid-1930s, between attack aircraft, priority was given to torpedo over dive bombers, unlike the composition of carrier air groups in the United States Navy, where the emphasis was reversed. For example, the composition of the air groups of the two largest carriers reflected these preferences. As modernized in 1935 and 1938 respectively, the Kaga embarked forty-five torpedo bombers, thirty dive bombers, and sixteen fighters, and the Akagi carried fifty-one torpedo bombers, twenty-four dive bombers, and sixteen fighters. In comparison, typical U.S. carrier air groups of the 1938–41 period included one fighter, one torpedo, and two dive-bomber squadrons, with each squadron consisting of twenty-one aircraft, eighteen operational and three spares.81
This Japanese preference undoubtedly reflected a traditional faith in torpedo warfare and the belief that torpedoes simply had greater explosive power than the ordnance that dive bombers were capable of carrying. With the demonstrations of the effectiveness of dive bombing late in the 1930s, the emphasis shifted, however, so that by early 1941 most Japanese carriers embarked roughly equal numbers of torpedo and dive bombers, though still far fewer fighter aircraft. Although torpedo planes and dive bombers together greatly outnumbered fighters on Japanese carriers, fighters still accounted for a higher proportion of the total (about one-third) than on U.S. carriers. The comparable figure in the United States Navy was one-fourth; only after the Pacific War began did the proportion of fighters on U.S. carriers increase.82
