Many U.S. submariners of World War IT are convinced that the official postwar assessment of credit (Japanese Naval and Merchant Shipping Losses During World War IT by All Causes, Joint Army-Navy Assessment Committee, Washington, D.C.: U.S. Government Printing Office, 1947 -Referred to as JANAC) failed to account for many ships that they believed, at the time, to have sunk. More recent analysis of Japanese sources (Alden, John D., U.S. Submarine Attacks Durin& World War ll, Annapolis, MD, Naval Institute Press, 1989) has shown that many such cases can be explained by the self-imposed rules that limited the JANAC assessment to sinkings only of regular Japanese warships and merchant vessels of 500 or more gross tons. This excluded all cases of damage short of sinking and also many smaller warships converted from merchant types, such as submarine chasers, minesweepers, patrol craft, picket boats, and various auxiliary types. When such targets are accounted for, there still remain several hundred unexplained cases where sinkings or damage had been claimed during the war.
About two years ago CAPT Roger Pineau directed my attention to some recently declassified records of so-called Ultra intercepts, available at the National Archives and the Naval Historical Center (Allied Claims and Enemy Confirmation of Damage to Japanese Ships, CNO Pacific Strategic Intelligence Section, SRH 184, Record Group 457, National Archives). The records in question, although sketchy and incomplete, nevertheless proved very useful from several viewpoints. First, they confirmed most of the sinkings officially credited to specific submarines as well as most of the cases of damage listed in the Japanese postwar summaries. (It appears that JANAC had access to much, if not all, of the Ultra data). Second, they shed light on many attacks for which hits were claimed but the target was not identified up to this time. Third, they raise some serious questions about the use, or non-use, of the Ultra information during the war.
In the first instance, the close correspondence between the Ultra intercepts and postwar records provides confidence in the accuracy of both. The fact that the Japanese, unaware that they were being intercepted, had no hesitation in reporting ship casualties, further supports the reliability of the Ultra information. Although I have identified some apparent errors in every source examined, JANAC and the Ultra records are remarkably clean. On the other hand, the intercept data cannot be accepted at face value without further checking and verification. This is because of the inherent uncertainties in the interception, decryption, and translation of the Japanese messages. Without going into detail, the records examined contain many gaps due to incomplete reception of the message; in many cases the ships involved cannot be identified. Similarly, ships referred to by their radio call signs may have been misidentified because of garbles and the lack of accurate information during the war. Also, many of the messages were decrypted and translated only in part because of the press of more urgent work. Finally, there were problems in translating and interpreting the Japanese text. For example, many Japanese ship names can have more than one meaning, and cases of incorrect interpretation are well known. Also, there are many cases where different ships had the same name, leading to confusion in the identification of the ones actually attacked.
The above caveats aside, I have found about 650 cases where the Ultra messages throw some light on hitherto unconfirmed attacks, about 55 percent by torpedo and the rest by gunfire or other means. (Gunfire attacks were made almost entirely on very small craft; while they are of considerable interest, they will be excluded from further analysis in this article.) A few of the torpedo attacks in question were made by British or Dutch submarines, but the overwhelming majority involved only U.S. subs. In about 240 cases the Ultra messages identify, to some extent, the Japanese ship attacked, either by name, radio call sign, or the convoy of which it was a part. Although the remaining cases are less definitive, all provide information that should be useful in tracking down further material in the Japanese archives.
A particularly significant finding, although one that will be a major disappointment to the submarine skippers concerned, is that over 200 or the above intercepts, or about 57 percent, confirmed that the torpedoes missed or were evaded. This finding was so intriguing that I went back to the original patrol reports and reviewed every case where hits had been claimed, but not verified postwar, to see what evidence had been cited to support the claim of sinking or damage. In the great majority of such cases I found that the claims were based on inconclusive evidence such as timed explosions heard, distant flashes seen at night, pips fading from radar screens, screws no longer heard on sonar, breaking-up noises, etc. In some cases, the evidence was visible and apparently dramatic — columns of smoke or spray, ships appearing to stop, list or settle by the bow or stem, or even blow up violently. Yet even in many of the most convincing cases, Ultra messages unequivocally reported that little or no damage had been incurred, and no contrary evidence has ever come to light.
The following examples will illustrate the quality of the Ultra material. On 1 Nov 1943, Davenport in the HADDOCK (SS-231) reported a 7,000-ton transport burned and sunk in 15 minutes, leaving many survivors in the water, and a 4,100-ton AK sunk immediately after being hit by a single torpedo. Ultra identified these ships as the TATEISHI (a cable layer) and the KIT AGAMI MARU, neither damaged. (Because of Davenport’s vivid and detailed description, I consider this case open to the possibility that some damage was caused; however, the TATEISHI is known to have been sunk in 1945 and the K.ITAGAMI MARU presumably survived the war.)
On 9 Nov 1943, Gross in the SEA WOLF (SS-197) saw and heard explosions on a 5,000-ton AK, and was credited with damage. Ultra intercepted a message from the HOKURIKU MARU reporting torpedo tracks sighted, no damage.
The next day Waterman in the BARB (SS-220) claimed an 8,000-ton AK sunk and another of 5,600-ton damaged by torpedoes that were seen to hit. Ultra identified these ships as the YAMAHAGI MARU and NISHI MARU, plus three other ships in the same convoy, and said only the NISHI MARU might have been damaged. (It was sunk on 13 November 1944 by carrier -based aircraft.)
On 18 Nov 1943, Munson in the CREVALLE (SS-291) reported four hits with a big explosion at the bow of an escort carrier and claimed a sinking. Ultra reported that the converted aircraft transport AKITSU MARU received no damage from three torpedoes.
Harral in the RAY (SS-271) on 26 Nov 1943, claimed a 4,500-ton AK sunk on the basis of hits seen and heard, the target’s pip disappearing from the radar scope, and the absence of the ship from its convoy the next morning. An Ultra intercept from the SUMIYOSHI MARU reported no damage from two torpedoes.
On 27 April 1944, Harlfinger in the TRIGGER (SS-237) fired four torpedoes at three 7,500-ton passenger/cargo ships and reported that one exploded and sank, another settled by the stern and sank, and the third was believed sunk after last being seen stopped and down by the stern. Ultra confirmed that the MilKE MARU had sunk, but reported that the NOTO MARU and TOSAN MARU were undamaged.
MacMillan in the THRESHER (SS-200) fired six torpedoes at a convoy on 16 July 1944, saw three pips disappear from the radar, and claimed a destroyer and two AKs. Ultra confirmed the attack, identified the convoy as TAMA 21C, and reported YURIN MARU and SHOSAN MARU as probably sunk.
The Ultra messages also provided valuable details about the extent of damage and the survivability of Japanese ships. For example, the fleet oiler SHIRETOKO was reported sunk on 13 September 1943 by Bennett in the PERMIT (SS-178), who last saw it low in the water. Ultra reported that it was being towed to Japan. On 13 November, Schmidt in the SCORPION (SS278) damaged the tanker, which Ultra said was being towed to Sasebo, where it remained under repair until May 1944. Then on 7 October 1944, it received two torpedoes from the COD {SS-224) under Adkins and three from the RAY {SS-271) under Kinsella, who saw the oiler limp into port. Ultra confirmed the COD’s hits but said the RAY’s were avoided and the ship ultimately reached Singapore. (It was finally sunk by Army aircraft on 1 Feb 1945.)
Based on the combination of Ultra and patrol report evidence, I have made my independent assessment of the approximately 350 Ultra messages reporting torpedo attacks, as shown in Table 1.
View full article for table data
Further, having re-evaluated almost 2,900 reported torpedo attacks, I believe the majority of the remaining unverified attacks by U.S. and allied submarines can now be explained by three factors: the fog or war, over-optimistic reliance on ambiguous obsenrations, and defective torpedoes. (I emphatically reject any implication that any submarine commander deliberately made a false report; some were obviously less skeptical than others, but I am convinced all reported what they believed they saw or heard.) There remain about 150 cases, five percent of the total, where patrol report or other evidence appears strong enough to support the possibility of a successful attack in the absence of any confirmation from the Japanese side. Japanese records may be irretrievably lost or contain errors that make it impossible to identify a matching submarine attack. However, I believe at least 95 percent or all torpedo attack claims have been verified or can now be satisfactorily explained.
The fog of war factor is easily understood, since most submerged torpedo attacks were made either by periscope with very limited visibility or by sonar with none, and most surfaced attacks were made at night, often in heavy weather or by radar alone. Add the confusion of wolf-pack attacks on milling convoys and counter-attacking Japanese escorts, and it is readily understandable how observations could be misinterpreted.
Submarine skippers can be forgiven for over-enthusiasm in basing claims of success on ambiguous observations. Records of attacks by German, Italian, and Japanese submarines (Rohwer, Jurgen, Axjs Submarine Successes 1939-1945, Annapolis, MD; Naval Institute Press, 1983) show that they too made many unverified claims. What is surprising is the uncritical acceptance of almost all such claims by the higher commanders who reviewed and endorsed the patrol reports. Even the Submarine Operations Research Group (SORG) accepted the “task force commanders assessments” and apparently did not have access to the Ultra intercepts. Consequently, its analyses must have been distorted significantly toward the over-optimistic assessment of damage. Whether this had any ill effects on submarine operations might be worth further study.
The most significant observation to be drawn from the Ultra intercepts is the Navy’s apparent failure to make the most effective use or them. It is now well known that the decryption of Japanese fleet operational messages (the JN-25 cipher system) enabled our carrier forces to win the Battle of Midway, led to the shooting down of Admiral Yamamoto, and contributed to many other successes. However, the late Rear Admiral Edwin T. Layton has written (Layton, Edwin T. with Roger Pineau and John Costello, And I Was There. New York: William Morrow and Co., Inc. 1985 [p. 470 et seg.]) that not until the so-called mam code was broken in early 1943 were our submarines routinely ordered into position to intercept convoys and sink many enemy ships, thanks to the Japanese practice of requiring daily position reports from the ships at sea.
According to Layton, the mam intercepts also supported the submariners’ charges that their torpedoes were defective and persuaded Admiral Nimitz to authorize the inactivation of the magnetic exploder (Note: this is referred to by CAPT Wilford J. Holmes in his Undersea Victmy, published by U.S. Naval Institute in 1979). In retrospect, the Ultra evidence of defective torpedoes appears so overwhelming that the continuing delays in implementing corrective measures are incomprehensible unless the detailed evidence was never passed along to other responsible commands. The intercepts showed that defects continued to exist for many months in spite of the changes that were made. Accounts of tbe torpedo fiasco have focused laraely on problems In the depth control mechanism, the failure of the magnetic exploder to detonate, and jamming or the firing pin in the Impact exploder. The Ultra records support the conclusion that premature detonations were equally serious. In particular, the only logical explanation for the many cases where ships appeared to blow up, yet remained undamaged, is that our torpedoes were going off prematurely directly in line with the target or were being countermined by other explosions.
Layton also says that the decrypted messages often enabled submarine headquarters to learn of Japanese ship sinkings even before the boats reported making attacks. However, there is little evidence in the patrol reports that the operating forces were made aware that their claims of sinkings were seriously inflated. This may have supported morale among the submariners, but it also tended to induce complacency and downplay the importance of follow-up attacks to ensure that targets were really sunk. If submarine commanders had known how frequently their apparently successful attacks were failing, and how effective Japanese damage control and salvage efforts were, they would undoubtedly have done their utmost to improve their procedures and train their crews more thoroughly.
In summary, the Ultra records constitute a rich source of new information on the success or failure of submarine attacks against Japanese ships during World War II. Further study of material such as the original intercepts, if these are still available, might enable some of the remaining questionable cases to be resolved. Similarly, research in the wartime records of higher commands might clarify the extent to which Ultra information was disclosed to or withheld from the different command levels of the submarine force and its supporting elements. Input and comments from knowledgeable members of the Naval Submarine League are invited.
