Prologue Magazine

American POWs on Japanese Ships Take a Voyage into Hell, Part 2

Winter 2003, Vol. 35, No. 4

By Lee A. Gladwin

Brazil Maru
The Brazil Maru passes through the Panama Canal on March 26, 1940.
Nearly five years later, it carried Allied POWs. (185-CZ-Vol. 51-Brazil Maru)

The Brazil Maru and the Enoura Maru:
Finishing the Journey into Hell

On December 27, the prisoners at San Fernando boarded the Brazil Maru and Enoura Maru and sailed for Takao, Formosa, part of the TAMA #36 convoy, bound for the POW camps near Moji, Japan. Landing craft ferried them from the pier to the ships. To board the landing boats, the men ha to leap twenty to twenty-five feet straight down into the boats. Any reluctance was swiftly punished by a bayonet prodding. Many were injured in their jumps.

Both Brazil Maru and Enoura Maru had been hauling livestock, and no effort had been made to clean out the manure before placing the prisoners in the holds. There were no attacks or food during the voyage to Takao. The ships docked there on New Year's Day 1945, and the prisoners received their first food since leaving San Fernando, "five moldy hardtack type of biscuits" and some rice. Each day guards came down to count the number of survivors. Twenty-one prisoners were buried at sea during the voyage to Takao, five from Brazil Maru and sixteen from Enoura Maru.

Both Brazil Maru and Enoura Maru were listed as undergoing repairs on January 8, 1945, accounting for their delayed departure for Moji, Japan. On January 9, MacArthur's forces invaded Luzon. A simultaneous attack was made on Takao. John Jacobs, aboard Enoura Maru, recalled that they were just finishing their only meal of the day when the first air raid sounded. Panic spread in the hold as men tried to remove the hatch covers.

"A young captain stood up and told the prisoners to sit where they were, that they were just as safe in one place as they were in another." Moments later, "there was a blinding orange flash and a deafening explosion followed by blindness." Planks, hatches and other debris flew through the air. Some men were pinned by the hatches. One floor gave way, dropping prisoners thirty or forty feet below. When it was over, only three doctors remained to care for seventy-five injured. Thirty-five dead were placed in a pile. With the hatches gone, the temperature plummeted.

On January 10, Jacobs was asked to look into the forward hold where a bomb fragment left a gaping hole. "There were mangled Americans, some 300 of them, piled three deep and pinned down with large steel girders and hatch covers," he remembered.

On January 12, forty-five coffins were removed to the beach and burned. One-hundred and fifty more were taken to a cemetery the following day. This gruesome task done, the survivors of this second attack were transferred from Enoura Maru to Brazil Maru. Finally, on January 14, they sailed for Moji, Japan, arriving there on January 29. Of the 1,619 POWs who boarded Oryoku Maru on December 14, 1944, 497 arrived in Moji. An estimated 500 died aboard the Brazil Maru during the voyage from Takao to Moji. Soon after Brazil Maru's arrival, wrote E. R. Haase later in his diary, "senior medical officers were asked to sign 1001 death certificates representing our loss from 13 December to Jan 31, a rather heavy toll."

Japanese radio traffic records the unfolding tragedy from the enemy's side. A message of January 10, 1945, from First Shipping Headquarters, Takao Branch. to Tokyo stated that "ENOURA MARU (JOVS) suffered some damage to the outer plates caused by near misses. Personnel killed, wounded, and missing—approximately 100." It was soon realized that the damage was much more serious, and a message to Taihoku stated that 925 prisoners were being transferred from Enoura Maru to the Brazil Maru. This figure would contrast sharply with the one later provided by the Japanese.

Counting the Costs:
A Search for the Truth

A complete, timely, and accurate accounting of American POWs was of paramount importance to the Chiefs of Staff, who needed the information for logistical planning, as well as to the friends and families awaiting word of missing loved ones at home.

Standard procedure during the war was for Axis and neutral countries where POWs were interned to inform the U.S. State Department or Office of the Provost Marshal General of prisoners taken. This was usually done via the International Red Cross. In the case of the "hell ship" victims, the Japanese War Prisoners Information Bureau compiled lists and sent them to their consulate in Bern for transmission to the Red Cross. Getting information from the Japanese often took a lot of time and repeated attempts.

As early as September 19, 1944, the State Department made inquiry concerning the numbers of those who died, escaped, or were recaptured following the sinking of Shinyo Maru. Apparently, there was no adequate response, as another inquiry was made January 11, 1945. Similarly, the Japanese were slow to furnish lists of casualties following the sinking of Arisan Maru. State made further inquiries on January 10 and February 22. Initial inquiry concerning the sinking of Oryoku Maru was not made until March 5.

Though some in Tokyo thought the American sinking of Enoura Maru would make great propaganda, the Japanese had a problem. How could they tout the number of dead without revealing their own responsibility for the ghastly figures that arose after Oryoku Maru was at the bottom of the harbor? They could not resolve this dilemma, and no propaganda broadcast was ever made.

Their dilemma was rendered moot when MacArthur's forces liberated Bilibid Prison. A roster of "Prisoners of War Believed to be Missing (Japanese Records)" or dead was submitted to the general on February 5, 1945. It contained 296 names. A copy of the list was sent to the Adjutant General's Office on March 13. Its contents, however, may have been known to the State Department and the Office of the Provost Marshal General before then.

As late as April 20, there was still no definitive word from the Japanese concerning Oryoku Maru. Maj. Gen. Archer L. Lerch stated what was known at the time by the Office of the Provost Marshal General's Office:

On 15 December 1944, a Japanese transport carrying 1600 prisoners from Cabanatuan was sunk close to shore in Subic Bay. Of this number, 2 were rescued and subsequently returned to United States military control. Approximately 1200 were retaken by the Japanese, and the remainder are believed to have been lost.

He added that the State Department had "made strong representations to the Japanese Government demanding lists showing the new locations of our prisoners who have survived, and lists of those who may have been lost due to the sinking of these transports." In addition, the War Department was "making similar requests through the International Committee of the Red Cross." So far, only a list of prisoners aboard Shinyo Maru had been received.

Although there were other sources of intelligence about the POWs and the ship sinking, some of the communications intelligence could not be shared outside a small circle without betraying their source. Figures varied depending upon who collected the information and from what sources, i.e., intercepts, captured Japanese documents, survivor accounts, etc. One, perhaps the earliest of these, bore the dual stamp "TOP SECRET ULTRA," the highest of Allied security classifications. Internal evidence in an accompanying text indicates that this table was compiled some time after April 20, 1945.

Another summary table, "Extracts of P/W Camp Information," was prepared by the U.S. War Department, Military Intelligence Division, Military Intelligence Service, Captured Personnel and Material Branch "from various sources." The distribution list was a who's who of U.S. and British intelligence gathering agencies. This report set the number of casualties from the Oryoku Maru sinking at 365. A "Magic Far East Summary" of March 29, 1945, similarly put the number at 363, and both were in virtual agreement as to casualties from the Shinyo Maru and Arisan Maru sinkings. ("Magic" was the code name for intercepted Japanese consular traffic.)

In July 1945 the Japanese War Prisoners Information Bureau sent the following cable to Bern:

ST/9 Investigations concerning a ship transporting prisoners of war from the Philippines homeward revealed that the ship in question met gun and bomb attacks by enemy planes on several occasions and that out of the 1619 prisoners on board 942 were killed instantly or died subsequently of wounds caused by bomb explosions while 59 died of illness.

Names of a part of the survivors have already been given in lists FM/56 and FM/57. As for others and those dead a notification will be sent later on.

These figures were included in the State Department's "Summary of Prisoner of War Ship Sinkings (Far East) 1944" and repeated in a letter dated September 5, 1945: "Lost 942, Died Later 59, Survivors 618."

The figures reported by the Japanese just did not add up! They did not square with the intelligence reports or what was learned from interviewing former American POWs. An accurate accounting required a reexamination of the Japanese records prepared by the Japanese Government Prisoner of War Information Bureau, Tokyo, and the use of information provided by former POWs who maintained lists of casualties. It was a daunting job to collect and manually compile and code the POW "death lists" and punch the IBM cards.

The POW accounting effort ended November 15, 1947, when the "death lists of Prisoners of War of Allied nationalities who died on route during the transfer from the Philippines to Japan Proper aboard the Japanese Prisoner of War Transport ships, the Oryoku Maru, the Enoura Maru, and the Brazil Maru" were approved. A supplemental list contained twenty-one names, including the sixteen names of those who died "During Transportation From Olongapo to San Fernando, P.I., December 1944."

In his foreword to the "death lists," Henry T. Omachi wrote that, following an examination of the records held by the Japanese "and a conference with the wartime employees of the Bureau," it was determined "that the information reported by the Bureau during World War II through the International Red Cross Committee to the various Allied governments concerned, pertaining to the casualties incurred during this transfer, was incomplete and misleading."

This was "evidenced by the affidavits executed by surviving prisoners of war and by Japanese records themselves, indicating that many of the deaths actually took place aboard the Enoura Maru and the Brazil Maru as well and that the dates of death ranged from December 1944 through January 1945."

To obtain a final accounting, the Adjutant General's Office, Machine Records Branch, meticulously went through the "death lists," writing "B" and the date of death in the margin beside the names of those who died aboard the Brazil Maru. "E" indicated death aboard Enoura Maru, and "XDS" followed by "23 Dec" marked the deaths of those who died "During Transportation From Olongapo to San Fernando, P.I., December 1944."

Scattered throughout the collection are penciled notations, "Dec Sinking." "DS" was probably combined with the first letters of the transport names, or "X" for crossing, to create the codes BDS, EDS, XDS, and ODS (Oryoku Maru). These codes were punched into the IBM cards of those who died aboard these ships or during the crossing. They were then fed through the compilers and tabulators of the time to yield the final totals.

Number of Ship Sinking Casualties Reported in Different Sources
Name of Ship and Date Sunk
Intercepted Japanese Radio Messages Japanese War Prisoners Infor-
mation Bureau

American POW Information Bureau,
Nov. 15, 1947

US War Dept. Military Intel-
ligence Division (MID),
Mar. 19, 1945
Magic Far East Summary, Mar. 29, 1945 US State Dept. Summary of POW Ship Sinkings, 1944
Shinyo Maru
Sept. 7, 1944
750 aboard 750 654* 667 667 667
Arisan Maru
Oct. 24, 1944
n/a 1,775 1,640* 1,795 1,778 1,753
Oryoku Maru
Dec. 14 - 15, 1944
250 942 284
365 363 942
59 of illness
Enoura Maru
Jan. 9, 1945
444 n/a 411
n/a n/a n/a
Brazil Maru 925 n/a 264
n/a n/a n/a
POWs who died during transportation from Olongapo to San Fernando, PI, Dec. 1944     16

n/a=not available
* from World War II Prisoner of War Data File, retrieved from AAD


The story of the men who entered the holds of hell began in the battle of the Philippines and ended in a struggle for the truth. From the ships' first appearances in the wireless communications of the Japanese, their physical descriptions, destinations, convoys, and sometimes, their cargoes were known. There is little evidence of advanced information that POWs were aboard.

Individual messages were frequently garbled. Reception was sometimes poor, and decryption, translation, and interpretation flawed. What was interpreted by one intelligence unit as "troops" was interpreted by another as "prisoners of war." In the context of the many other convoy and transport messages, as well as the cargo histories of the "death" ships, there was ample evidence suggesting these ships typically carried munitions and reinforcements. Given, too, that the "hell ships" were unmarked, there was no way U.S. Navy aircraft and submarines could know Allied prisoners of war were aboard.

"Magic" and Military Intelligence Division summaries of information taken from intercepts, captured material, and prisoner interviews confirmed casualty figures from Shinyo Maru and Arisan Maru but raised great doubts about those submitted by the Japanese for Oryoku Maru. Since the source of this information was Top Secret ULTRA and an exact, name-by-name accounting was required for servicemen's families, an intensive effort was made over several years to obtain complete lists and identify those who died on board Oryoku Maru, during transport from Olongapo to San Fernando, and aboard the Brazil Maru and Enoura Maru.

In the end, early IBM punch cards and machines were required to unravel the mystery of the deaths of those who stepped into hell that distant December day.

Lee A. Gladwin is an archivist with the Archival Services Branch of the Center for Electronic Records, National Archives and Records Administration. His particular areas of study are Alan Turing, Bletchley Park and World War II codebreaking. His introductions to two of Alan Turing's Enigma-related documents discovered at NARA were published in recent issues of Cryptologia. His paper "Alan M. Turing's Contributions to Co-operation Between the UK and US" will appear in Alan Turing: Life and Legacy of a Great Thinker (Holland: Springer-Verlag, 2003).

American POWs on Japanese Ships Take a Voyage into Hell, Part 1

Note on Sources
Selected Primary Sources

"Death Lists," POW diaries, and correspondence of the Director, American Prisoner of War Information Bureau, Office of the Provost Marshal General, may be found in Record Group 389, Records of the Office of the Provost Marshal General, General Subject File, 1942–1946, American POW Information Bureau Records Branch, Entry 460A, Sinkings, Box 2276.

Decrypted Japanese ("Orange") intercepts, indexed by ship's name, may be found in Record Group 38, Records of the Chief of Naval Operations, Translations of Intercepted Enemy Radio Traffic and Miscellaneous World War II Documentation, 1940–1945, Entry 344, Box 1299.

Joint Intelligence Center Pacific Ocean Area Summaries of ULTRA Traffic, September 11–December 31, 1944, are located in Record Group 457, Records of the National Security Agency/Central Security Office, Intelligence Reports from U.S. Joint Services and Other Government Agencies, December 1941–October 1948, SRMD 5-7 (Part IV), Entry 9024, Box 2.

All of the above are located at the National Archives at College Park, Maryland.

Useful secondary sources include Frank Cain, "Signals Intelligence in Australia during the Pacific War," Intelligence and National Security 14 (Spring 1999); Lee A. Gladwin, "Top Secret: Recovering and Breaking the U.S. Army and Army Air Order of Battle Codes," 1941–1945, Prologue: Quarterly of the National Archives and Records Administration 32 (Fall 2000): 174–182; E. Bartlett Kerr, Surrender & Survival: The Experience of American POWs in the Pacific, 1941–1945 (New York: William Morrow & Co., 1985); Gregory Michno, Death on the Hellships: Prisoners at Sea in the Pacific War (Annapolis, Md.: Naval Institute Press, 2001).

For much greater detail concerning the encoding and enciphering of messages, see Edward J. Drea, MacArthur's Ultra: Codebreaking and the War against Japan, 1942–1945 (Lawrence: University Press of Kansas, 1992), and Maurice Wiles, "Japanese Military Codes" in F. H. Hinsley and Alan Stripp, eds., Codebreakers: The Inside Story of Bletchley Park (Oxford: Oxford University Press, 1993).

For extensive treatment of Japanese codes, see Michael Smith, The Emperor's Codes: The Breaking of Japan's Secret Ciphers (New York: Arcade Publishing, 2000) and W. J. Holmes, Double-Edged Secrets: U.S. Naval Intelligence Operations in the Pacific during World War II (Annapolis, Md.: Naval Institute Press, 1979).

The web site contains the story of the Oryoku Maru and names of the POWs who sailed on that "hell ship."

Articles published in Prologue do not necessarily represent the views of NARA or of any other agency of the United States Government.