In September 1999, undersea explorers located some intriguing wreckage on the bottom of the Pacific near Midway Atoll. The question was, exactly what had they found? To answer that, a team of web-enabled historians was asked to help, and its members confirmed the identity of one of the more exciting underwater archeological finds of recent years.
In October 1999, I received an email from Jeff Morris, who was working for the undersea exploration firm Nauticos. Jeff was looking for pictures of the Japanese aircraft career Kaga. I have run a web site devoted to the imperial Japanese Navy (www.combinedfleet.com) since 1994, and people write to me all the time asking for pictures of this or that vessel. As it happened, I already was involved in preparing an illustration of the Kaga for an upcoming book. I wrote down what I knew of her demise, sent along the illustration and a scanned photo, and thought nothing more of it. Little did I realize that I was about to stumble onto a naval researcher's dream come true.
Unlike most of my correspondents, Jeff kept coming back for more information. Over the course of the next few weeks he gradually let on that Nauticos, working under a joint research agreement with the U.S. Navy, recently had found something on the seabed near Midway.1 Jeff's questions became increasingly detailed, and I became increasingly excited. If Nauticos had found the remains of a Japanese vessel near Midway, it clearly was an important discovery. The Battle of Midway remains one of the most crucial and widely studied battles of World War II. Any truly new find there would be of considerable importance to the historical record and would increase our understanding of the battle.
I soon enlisted the aid of Tony Tully—an online friend, Japanese naval researcher, and longtime contributor to my site. Tony and I then approached David Dickson, another friend and expert in Japanese carriers. After signing nondisclosure agreements with Nauticos, our little historical team gradually was allowed to view more detailed information on the organization's find. The process was a slow one. None of us ever had met face to face, and the sensitive nature of the information made Nauticos understandably cautious. Nevertheless, through a series of email correspondences and preliminary analyses, the relationship between the team and Nauticos gradually solidified as the identification process went forward.
This culminated in February 2000, with Nauticos providing us with both video and still photography of a large piece of wreckage lvin A 17,000 feet below the surface of the Pacific. To the trained eye, it immediately was apparent that the wreckage was from a Japanese carrier, and not from a smaller combatant or merchant ship. The characteristic features of Japanese carrier architecture—cantilevered support braces, hemispherical Run tubs with radial supports, and gallery walkways—all were there. The challenge for us was to identify the artifact.
An annotated photomosaic of the wreckage is presented above. The picture at right, showing the underside of a typical Japanese carrier machine-gun gallery, is useful for orientation purposes when looking at the photomosaic. The artifact has ended up in an inverted position on the sea floor, so that the viewer is looking at the underside of the structures.
The wreckage is a portion of a ship's side plating. The piece once had been a relatively flat section of the ship's side structure. This was confirmed in the video footage, which also showed a porthole in the piece. Attached to what would have been the top of the structure are two semicircular platforms of equal diameter. The platforms, as was typical of Japanese carriers, each were supported by a large cantilevered brace that met the ship's side some 15 feet below the platform.
The artifact has been severely distorted, the piece having been bent near the center of the bulkhead and rotated around its vertical axis by about 75 deg, thus forming a concave shape. This resulted in the two platforms nearly touching each other, whereas on the original ship they would have been side by side and separated by some ten feet. "The chunk" (as we began to refer to it) had comprised nearly 60 linear feet of ship's plating before it was detached.
A second mosaic (not shown) partially showed the backside of the artifact, which at one time would have been on the interior of the ship. It showed a stairway, presumably coming down from the level of the machine-gun galleries, leading into the interior of the vessel. These two mosaics, and the video footage, comprised the basic evidence we had available to pinpoint the ship's identity. At first glance, it was not much with which to work. But the wreckage contained enough unique elements to allow us to make a positive identification.
The first problem was to resolve what type of platforms at which we were looking. The telemetry information from the photo sled soon cleared that up, giving the diameter of each of the two semicircular tubs as approximately 14 feet. This was too large a tub for a fire-control director mount, and too small for a 5-inch mount (which would have been some 22 feet across), but matched the diameter of the tubs used for twin 25-mm antiaircraft mounts. Both tubs had a section of walkway appended to them of some eight to ten feet in length, one leading forward, the other leading aft. This meant that the structure once had comprised the middle part of a machine-gun gallery, with walkways fore and aft leading to other gun or fire-control mounts.
Even more intriguing was the wreckage attached to one of the gun tubs. On closer examination, this turned out to be a landing light-array (chakkan shidoto-landing-guidance lights). These arrays (usually four to a ship) were used to help Japanese pilots establish the proper glide slope during a landing approach, very similar to modern carrier landing aids. The design of such arrays varied from ship to ship, but generally comprised a triangular supporting arm structure atop which were either two or four rectangular colored lights (either green or red). The arm itself was hinged at the base, and was provided with cables at the apex, so it could be stowed flush with the ship's side when not in use or cranked out perpendicular to the flight deck during flight operations. The structure attached to the chunk fit this description nicely. Though it had crumpled on impact, the array was well defined structurally. One of the landing lights even had survived the descent to the seafloor, and the rigging cables still trailed from the end of the arm.
We now turned to identifying the ship. Based on the evidence at hand, we needed to find a portion from one of the ships that had all of the following six qualities:
- A similar 25-mm machine-gun gallery
- A similar flat portion of the ship's structure
- Two 25-mm tubs adjacent to each other, without fire-control equipment or other platforms between them
- A porthole located about five feet from one of the gun-platform support struts
- A gallery structure with walkways continuing on either side of the tubs for at least seven to eight feet in each direction
- A gun tub with a landing-light array attached to the outermost point of the tub
The Ships
The Japanese lost four aircraft carriers at the Battle of Midway: the Akagi, Kaga, Soryu, and Hiryu. Of these, the Hiryu was not a candidate for our investigation because she was sunk far to the north of Nauticos's search area. The other three vessels, however, were lost in relative proximity to each other. Though we knew that Nauticos had been searching deliberately for the Kaga, we needed to examine each of the three carriers dispassionately, for two reasons. First, warship sinking locations often are inaccurate. In the heat of battle, it is difficult to fix a position precisely. In fact, it is not uncommon for ships of the same formation to record their positions differently.2 Second, the exact Japanese formation at the time of the U.S. dive-bomber attack at 1020-1025 on 4 June 1942 still is a hotly debated issue.3 Exact course and speed information for the three victims after the attack also is very sketchy. Very little is known of the tracks these ships took during the late morning as they staggered and drifted, all three ablaze, toward their final resting places that evening. Consequently, we had to be open to the idea that the wreckage could have come from any of the three ships.
Akagi
The Akagi, proud flagship of the Japanese Mobile Strike Force (Kido Butai), could be ruled out at once. Very high quality 1:400 scale plans of her show clearly that none of her four landing-light arrays was mounted on her 25-mm gun tubs. The forward pair of lights was attached to catwalks near the bow. The aft pair, while attached to the machine-gun galleries, was not mounted on the gun tubs themselves. This was in direct opposition to the photographic evidence of the chunk. Consequently, the Akagi could be dismissed from consideration.
Soryu
In contrast to the very good plans of the Akagi that were available, the photographic record and plans for the Soryu are very rudimentary. None of the available plans or photographs showed the location of her landing-light arrays, for instance. Nevertheless, a convincing set of circumstantial arguments could be made against her. First, the Soryu was a smaller ship than either the Akagi or Kaga, both of which were built on relatively spacious capital ship hulls. By contrast, the Soryu was designed around a heavy cruiser-type hull and powerplant. Consequently, she is a much more cramped design, and this is reflected in the layout of her machine-gun galleries. Her fire-control directors and gun tubs are jammed together nearly shoulder to shoulder, in contrast to the relatively expansive gallery seen in the wreckage. The only potential structure on the Soryu that even remotely matched the wreckage was her port bow gallery. An examination of the hull below this gallery, however, revealed that it was the site of the intakes for the Soryu's engine room ventilators. As a result, the hull in this area was not a smooth expanse of metal, but a hodge-podge of ductwork. Furthermore, the plans indicated that there were no portholes in the area. This collection of factors ruled out the Soryu.
Kaga
This left the Kaga as the sole candidate. The team agreed, however, that it was imperative to find positive evidence to recommend the Kaga's case, rather than simply settling on her by eliminating her sunken compatriots. Unfortunately, the photographic and plans records for the Kaga are nearly as miserable as those for the Soryu. None of our plans showed the location of the landing-light arrays, and the photo record was equally scanty.
We knew that the Kaga's aft machine-gun galleries were the only likely candidates for matching the wreckage, since the forward 25-mm mounts were individually sponsoned, and not part of gallery structures. Fortunately, the aft galleries contained some idiosyncratic features of their own. The twin 25-mm mounts in the aft galleries (five to starboard, four to port) were controlled by a pair apiece of Type-95 fire-control directors port and starboard. These Type-95 directors were integrated into the 25-mm tubs in the gallery, thus forming four hybrid gun/director tubs. These tubs were not semicircular in shape, having extra bulges on the forward parts of the tubs. Yet, the two tubs on the chunk clearly were semicircular, meaning they could not be hybrid gun/director tubs.
This eliminated the Kaga's port gallery, since the two hybrid tubs were the middle two of the four tubs in that gallery (i.e., the port gallery did not have two semicircular tubs sitting next to each other). The starboard gallery was composed of five tubs; nos. 3 and 4 (counting toward the stern) were the hybrids. This meant that the two forward-most tubs were the only two semicircular tubs sitting next to each other on the starboard gallery. These two tubs were our last hope. But how could we prove they were part of the chunk?
Limited as the photographic record of the Kaga is, it turned out that there was one superb overhead shot of her. Taken in April 1941 by a Japanese aircraft, it shows the Kaga's flight deck in great detail. Indeed, the resolution of the photograph was good enough that we even could make out the support spars for the safety nets hung along the side of the flight deck. And almost immediately, as we studied the starboard gallery, we saw a thick white line emanating from the outermost point on one of the two forward-most tubs—the landing-light array. Even from a distance overhead, it was clearly a more substantial structure than the net supports nearby. And it was in the correct position to match the array found on the artifact. Finally, here was positive evidence to support all the negative evidence that had eliminated the rest of the Kido Butai. Nauticos indeed had found the Kaga.
In retrospect, we were incredibly lucky. The chunk was big enough, and rich enough in detail, to give some hope of making an identification. Had Nauticos found a piece of her hangar wall with no tubs attached, or a piece of the flight deck or lower hull, we most likely would have been stymied. Not only that, but the one decent overhead shot of the Kaga available in the photographic record (out of a grand total of perhaps a dozen shots of her in her post-modernized state) actually showed her landing-light array on the starboard side. Had the Japanese plane that took that photograph been flying down the Kaga's starboard side instead, the landing-light array would not have been silhouetted nicely against the dark water.
Historical Significance
Nauticos's discovery is a significant addition to the forensic record of an important modern battlefield, for three reasons. First, it indicates that the Kaga's final resting place lies somewhere close at hand. Second, the size and nature of this artifact, and its not being in direct proximity to the main wreck, also gives some indication of the violence of the Kaga's final hours, as her hangar was wracked by induced explosions from her own ordnance. Finding her main wreck will give us an even more detailed view of her demise. Third, finding the Kaga's hull will provide a firm starting point for finding the Akagi and Soryu as well.
The main question to be answered now is how and when did the chunk get blasted from the Kaga's side? The sheer size of the artifact suggests a rather powerful explosion tore it from the ship. But was this explosion one of the initial 500- and 1,000-lb bomb hits (one of which landed no more than 40 feet away), a result of induced explosions during her day-long fire, or a relic of her sinking? If the wreck is somewhat distant, it would indicate that the chunk was blown off sometime during the initial U.S. bombing attack or during the fire afterward. If the main wreck is found nearby, however, it would indicate that the debris might have come from an explosion near the time of the Kaga's sinking. Only finding the Kaga's main wreck will shed light on that mystery.
Next Steps
Nauticos hopes to return to the Midway battlefield to find the Kaga's main wreck sometime in the not-too-- distant future. Hopefully, finding the main wreck will not be overly difficult, given that there now is a positive data point from which to work. Finding the Kaga doubtless will increase our knowledge of how the Japanese carrier force fought and died that day, and help complete the record of one of the most important naval engagements in history.
[Footnote]
1Nauticos had been working on finding the Midway carriers for more than a year. In May 1999, a preliminary sonar survey of an area identified by Nauticos's proprietary renavigation techniques had detected several promising targets. In September 1999, Nauticos returned on board the Navy oceanographic survey vessel USNS Sumner (T-AGS-61) and made the discovery of the artifact.
2This is true of the U.S. formation at least. Japanese navigation and position tracking may have been better on average than the Americans' at this stage of the war, at least according to our conversations with Chuck Haberlein of the Naval Historical Center, who was intimately involved with Dr. Robert Ballard's Midway expedition that located the USS Yorktown (CV-5).
3This question has been argued since the battle, as U.S. carrier pilots debated which air groups should get credit for which ships. Japanese records of the ships' movements, at least those that have been translated into English thus far, also have proved inconclusive for reconstructing the exact formation.
Mr. Parshall is a senior business analyst for a web-development firm in Minneapolis. Mr. Tully is information systems manager for a professional technical recruiting firm in Dallas, Texas. Mr. Dickson is corporate counsel for an insurance company in Memphis, Tennessee. The authors wish to thank Spencer King, Jeff Morris, and Jeffrey Palshook of Nauticos for their assistance and collaboration during the project. They are also grateful to John Lundstrom, Eric Bergeraud, Chuck Haberlein, and Nathan Okun for their various insights regarding these matters.