TOP SECRET

Recovering and Breaking the U.S. Army and Army Air Force Order of Battle Codes, 1941–1945

Fall 2000, Vol. 32, No. 3

By Lee A. Gladwin

A keypunch operator of the Thirty-sixth Machine Records Unit in Belgium enters data from morning reports onto cards. (NARA, 111-SC-196472)

What was the current strength of the U.S. Army and Army Air Force each month during World War II? What were the numbers of casualties and prisoners of war? How many more men would have to be drafted, trained, and sent overseas? The answers were vital to the war effort. The authorized and actual numbers of officers and enlisted men for specific units was classified as "top secret" and were reported monthly in the Analysis of the Present Status of the War Department Troop Basis, large, thick reports circulated only among those with a right and need to know. Only eighty-six copies were printed for distribution. Not even President Franklin D. Roosevelt received a copy.

Tracking the men and material needs of the U.S. Army ground and air forces was a daunting task. In order to quickly and accurately tally actual versus authorized strength, the War Department's Adjutant General's Office (AGO) relied upon IBM data-processing equipment. Members of the AGO's Machine Records Units keypunched raw data onto eighty-column IBM punch cards, which were then fed into various machines that tallied the figures and printed reports. For the purpose of identifying specific units, three fields had to be punched: Type of Organization, Parent Unit Number, and Parent Unit Type.

The Type of Organization identified the kind of unit to which the serviceman belonged: e.g., "Heavy Bomber (B17 or B24)" or "Field Artillery 105MM Howitzer." The Parent Unit Number was the actual unit's number. Parent Unit Type indicated whether the unit was, for example, a group, squadron, regiment, or battalion. Most of the punch cards containing troop basis-related codes were destroyed after the war, but one collection remains: the World War II Prisoners of War Punch Cards, Record Group 389, Records of the Office of the Provost Marshal General, 1941–. It contains the records of 143,360 POWs. But for the actions of War Department personnel, later army records managers, and some perceptive army archivists, these cards might have been branded "nonarchival" and relegated to the lost treasures of the world. Thanks to them, the evidence of the War Department's pioneering effort in encoding the troop basis and estimating man and material needs now resides in the National Archives and Records Administration's electronic records.

Researchers who need copies of their personal records or those of family members who were prisoners of Germany or Japan during World War II usually write to NARA. They may use the information to support a claim for benefits or to complete a genealogical history. Until a couple of years ago, reference staff conducted manual searches of the printouts of some of the records that they had created for reference purposes and, if a record were found, composed a letter to the researcher that incorporated the field names and their values from the record. The following data fields may appear in any record: Serial Number, Name, Grade alpha, Grade Code, Service Code, Arm or Service Code, Date Reported (DDMMY), Race, State of Residence, Type of Organization, Parent Unit Number, Parent Unit Type, Area, Latest Report (DDMMY), Source of Report, Status, Detaining Power, Camp, Repatriation Status, and, in a small subset of the records, data on Ship Sinkings.

Many of these fields use alphanumeric codes to record the data. To interpret the coded data, NARA staff had to turn to the codebooks included as part of the documentation for the electronic datafile. After years of manually searching printouts and codebooks in order to write responses, the entire collection of punch cards was migrated to a digital format, and the thirteen data files composing this POW collection were recently loaded into an Access 2.0/97 database along with separate tables containing all of the codes and all available meanings. At the start of the project, the Type of Organization, Parent Unit Number, and Parent Unit Type code values had not been found. The Access database includes not only the raw data from the electronic records but also additional fields to hold the code meanings. Happily, I recently found many, though not all, of the meanings of the Type of Organization codes, and they were added to the documentation package for the electronic records file and also to the Access database.

This is the saga of how IBM punch cards and the War Department's Machine Records Units helped win World War II. It is also the tale of how, nearly sixty years later, the Type of Organization codes were partly discovered or broken through the application of quantitative methods and recently discovered code documentation.

Office Automation Goes to War

Before 1939, neither the War Department nor any of the twenty-four European war ministries served by the International Business Machines European Headquarters in Geneva showed much interest in machine-readable records. Administration, military or otherwise, was viewed as a "'necessary evil' of little importance for the defence of the country," wrote J. W. Schotte, IBM's general manager for Europe. This attitude changed abruptly after the German occupation of Austria and Czechoslovakia. Germany's preparation for war included "internal organization," or the "organization of the second front." Newspapers were filled with stories of the "necessity of having in all phases of life behind the front an organization, which would remain intact and would function with 'Blitzkrieg' efficiency in times of general mobilization and war."¹ Confronted by German expansion, the war ministries of Yugoslavia, Romania, Hungary, Poland, Sweden, Holland, and France ordered "punch card equipment" in an effort to overtake the German lead. Late in 1939, Germany seized most of IBM's equipment within its area of influence for use by its own installations.²

Punch cards, keypunch machines, and mechanical sorters had been introduced by Herman Hollerith in time for the tabulation of the 1890 U.S. population census. Having proved their effectiveness in recording, storing, and manipulating large amounts of records, they were soon adopted by business and some federal agencies. Their acceptance and use gained ground during the New Deal era as a means of implementing its social policies and tracking payroll deductions and unemployment insurance.³

Incredibly, the War Department had not fully embraced current business machine technology, with the exception of the typewriter. Nevertheless, before World War II, the army's Adjutant General's Office used punch cards to maintain officer and enlisted strength statistics. In fact, the Surgeon General's Office used punch cards as early as World War I to maintain records of the sick and wounded and annually tabulate medical statistics. A very few machines were allowed the Army Signal Corps and the Office of Naval Intelligence for cryptological work. Two ordnance arsenals and two engineer districts employed business machines to process payroll and accountingrecords.⁴

Following the outbreak of war in Europe in September 1939, the Adjutant General's Office of the War Department was confronted with the eventual need to account for men and materials on a massive scale. There may have been some who remembered events during World War I, when the effort to "maintain locator cards and keep track of changes in status of units and individuals broke down completely under the deluge of so-called 'snow flakes' or change slips which showered the Adjutant General's Office from overseas."⁵ Despite the Roosevelt administration's encouragement of the use of IBM basic punch-card machines for civilian purposes, that characteristic enthusiam and urgency was lacking with regard to military preparedness. An isolationist Congress may have thought that buying fewer business machines would keep America out of war. As late as April 23, 1940, the Adjutant General's Office could boast of only one "Hollerith punch card system" being used in its Enlisted and Reserve Divisions.⁶ When it was announced that some money might be available from the President's Contingency Fund, the Adjutant General's Office requested $200,000 for the "complete modernization and mechanization of the administrative system of the army." It was proposed to use such funds for establishment of a central machine records section in the office of the Adjutant General, in each corps area and department headquarters, at air corps general headquarters, larger army posts, and at least "one experimental unit at an army headquarters in the field, to develop procedures for tactical purposes." The AGO proposed to staff these units with civilians so as to free military personnel from administrative duties.⁷ To buttress their argument for funding, they pointed out that by placing the "administrative system of Army on a mechanical basis," the office could more quickly handle a mass of personnel records while reducing the number of personnel. The AGO predicted that one set of machines could replace sixty men.⁸

The announced purpose of the system was to "shift the burden of personnel administration from tactical units (company, regiment, division, corps and army) to installations of corps area service commands through the establishment of machine administration service units, as part of corps area service commands." While "normal channels of communication" would continue to be used for general administrative purposes, machine records units (MRUs) would convert personnel accounting records into "punched card records."⁹ An MRU was to be established at each corps area and departmental headquarters; auxiliary units were set up at air corps general headquarters and at locations with the heaviest personnel administrative load. Additionally, an "experimental" mobile unit was to serve an army headquarters, and funding was found for nineteen such units.¹⁰ The army hoped to put mobile MRUs as close to the action as possible.¹¹

A series of punch cards was even then being printed. The statistical card would record such personal information as the individual's height, age, and date and place of birth. Other cards included the status card, report of change (personal and unit), a medical card to record the serviceman's medical history based upon morning and sickness reports, locator cards, rosters and returns, and casualty reports.¹²

Despite the October conference and initial funding, financial support remained a problem into 1941. As Europe and the Far East plunged ever deeper into war, in January 1941 the Adjutant General's Office pleaded for parity of funding with other services as it defended "the necessity for the Reorganization of the Army Personnel System and the installation of Business Machines for the handling of personnel matters for the Army."¹³

The necessary note of urgency was struck at Pearl Harbor on December 7, 1941. At the Office of the Assistant Chief of Staff, G-1, the staff worked through the night. They had to quickly "compute the defense reserves of equipment then in the overseas departments, particularly in Hawaii and the Philippines, which had been established on punched cards."¹⁴ Following the attack upon Pearl Harbor, thousands of patriotic young men descended upon army recruiting centers across the country. Area commands were confronted with more servicemen than their current MRUs could handle and cabled the Adjutant General's Office for assistance.¹⁵

Demands for punch-card machines and men to operate them far exceeded initial supplies. By July 31, 1942, there were forty-four machine records units serving the U.S. Army, and more keypunch machine operators, tabulating machine operators, and card-punch-sorting machine operators were desperately needed. The shortage of experienced keypunch operators was described as "acute." MRUs were ordered to search their collections of qualification and status cards for qualified enlistees, and training programs were conducted on a massive scale that summer.¹⁶

Construction of the mobile trailers apparently began during the winter of 1942–1943. Each mobile unit consisted of two truck-trailers mounted on rubber shock absorbers. The machines contained within them were powered by two generators producing eighty horsepower. Following construction, the trailers were sent to the International Business Machine Corporation's Engineering Laboratory at Endicott, New York, for installation of the punch-card machines. Huge as they were, these trailers were mobile enough to follow the troop landings in every theater from Sicily and Normandy in the European theater to the islands of the Pacific. At least one mobile machine records unit was attached to every army corps. Of the more than one hundred MRUs, thirty-three received unit citations.¹⁷

Punch cards recorded personnel and unit information. These cards and their related reports allowed the army chief of staff to monitor troop strength as printed out in the troop basis or its monthly analysis. The reports compared authorized and actual strength for the purposes of drafting, training, and replacing service personnel.

Preserving the Punch Cards on Electronic Media

In 1959 the army transferred boxes of punch cards to the National Archives and Records Service as part of the wholesale transfer of the army's World War II archives. They were assigned to Record Group 389. No effort was made to convert them to a digital format until the late 1970s, when the Veterans Administration borrowed some of the punch cards and migrated these records to digital format as two files copied to nine-track, reel-to-reel tape. These two files were known as the U.S. Military Prisoners of War (POW)—World War II, Military Personnel Returned Alive from the European and Pacific Theaters files in the Records of the Office of the Provost Marshal General. In 1998 the Center for Electronic Records, in collaboration with the Textual Archives Services Division and the Information Technology Branch of the Office of Human Resources, arranged for the conversion of eleven additional files to magnetic media. As a result, all of the World War II POW punch cards were migrated to a digital format. There are thirteen files containing 143,000 records.

In Quest of Meaning: Finding the Documentation, 1970–1999

Without documentation for coded data fields, individual data records often have little or no meaning. In coded data, much of the information exists only in code; e.g., a number "1" in the Race field stands for "White." No documentation for the punch card records was identified directly by the army when it transferred the cards to the National Archives. However, we now know that the code values and the card formats are defined in textual records from the Machine Records Units and these, too, are preserved in the National Archives. Documentation is absolutely essential for using the data in any way. The documentation identifies the names of the fields, their sizes, and type (whether alphabetical, numeric, or alphanumeric). For those seeking to understand the codes in the records, codebooks are a vital part of this documentation.

The World War II Prisoner of War Codes

IBM punch card for Lt. Gen. Jonathan Wainwright, captured commander of the Philippine Division. (Records of the Office of the Provost Marshal General, 1941–, RG 389)

Throughout the war, the International Red Cross Committee in Bern, Switzerland, routinely sent lists of POWs to the Office of the Provost Marshal General. The office then sent a letter to the next of kin and a copy of the casualty report to the Office of the Adjutant General, Machine Records Branch. A telegram was also sent to the next of kin by the Office of the Adjutant General while the Machine Records Branch punched the POW card.

We owe a great deal to those who preserved these cards and transferred them to the National Archives in 1959. Much is owed also to William H. Cunliffe, assistant chief for reference, National Archives Modern Military Branch, 1975–1980. He found the first set of documentation that matched most of the observed fields in the War Department Adjutant General's Office Form No. 0326, Prisoner of War punch card.

In the 1970s, when former POWs requested documentation of their incarceration, National Archives staff copied the POW card and Red Cross reports for the camps in question. According to Cunliffe, "We were all aware that the punch holes in the cards represented information beyond what was printed out along the top of the card." Trying to find the key to this information, Cunliffe, "worked out allotment of the columns all across the card (assigned to name-rank serial #, etc.). This activity identified the columns where codes were being used." He then searched a variety of records and reference materials for files, issuances, and regulations that covered codes for military units, personnel, stations, and geographic locations.

Cunliffe discovered eleven-by-fourteen-inch goldenrod coding sheets that included a sheet titled "Wounds, Amputations and Diseases." The codes, both alpha and numeric, could be combined to form three-character combinations, e.g., S54. Since the sheet of codes was included with the other code sheets and the codes matched some of what we later discovered were the Type of Organization codes found in the cards, the sheet was believed to contain codes for a possible field, Wounds, Amputations and Diseases. It was not until 1989, when one of the cards was copied onto a transparency for a presentation, that the previously obscured field names could be read. In mid-sentence, Margaret Adams, Center for Electronic Records, turned to the enlarged image on the screen and realized that the field layout for the cards had been preprinted on each of the punch cards. At that point, she realized that rather than information on "Wounds, Amputations and Diseases," columns 56–58 in each card recorded "Type of Organization," "Parent Unit Number," and "Parent Unit Type."

In 1994, when I was looking for codebooks in connection with documenting another series of records, the World War II army serial enlistment file, the first codebooks for Type of Organization, Parent Unit Number, and Parent Unit Type came to light. These codebooks contained alphanumeric codes that consisted of three numbers followed by a letter (e.g., 522P for armored infantry regiment). Though codebooks for 1944 and 1945 were now in hand, no effort was made at that time to decode the Type of Organization and Parent Unit Number and Type fields in the POW records. It was not until several years later, after the Center for Electronic Records began creating search, retrieval, and reporting systems for its Korean Conflict and Vietnam Conflict casualty and POW records, that interest turned to creating a similar system for the World War II POW records.¹⁸

Following a question from another archivist concerning the meanings of the uninterpreted Types of Organization codes in a POW's record, I began a search in June 1999 for additional technical manuals, codebooks, or correspondence pertaining to the creation and use of these codes and the three coding systems.

Discovering Code Books and Breaking Type of Organization Codes

The earliest known systems for encoding Type of Organization are numeric and alphanumeric. Both were in use between 1942 and 1945. A mimeographed copy of the numeric system codes was found, but no printed version was discovered. In the "Army Ground Force Units" version (March 8, 1943), Type of Organization and Parent Unit Type are given in parallel columns. Spaces were left for adding new organization types. A defect of this system was that there was no way to identify higher-level parent organizations, such as division. This system was somewhat limited, therefore, in its descriptive power.

The creation of these alphanumeric codes begins in mystery and ends in enigmatic uncertainty. No codebooks were ever found for this system, though some of the earliest POW records, including that of Lt. Gen. Jonathan Mayhew Wainwright, commander of the Philippine Division, who surrendered his forces to the Japanese after the fall of Bataan and Corregidor, were encoded using this system (J32 in his case). The system was used simultaneously with the numeric system and, in some ways, supplements it. Only by analyzing the records themselves and trying to identify unit numbers with those given in the Analysis of the Present Status of the War Department Troop Basis and cross-checking with online biographies at various private POW web sites can this coding system be broken.

Analysis revealed large percentages of Type of Organization codes beginning with certain letters clustered around specific arm or service codes. For example, code C associated with the Chemical Warfare Service, code D with the Quartermaster Corps, and code E with the Corps of Engineers. The next step was to link unit numbers with specific Type of Organization codes. For this purpose, great reliance was placed on two sources: Analysis of the Present Status of the War Department Troop Basis and Shelby L. Stanton's World War II Order of Battle. Breaking the codes begins with finding the arm or service associated with the given unit number. Consider the example of William E. Harman's World War II POW record. His record gives the Type of Organization as K15, the Parent Unit Number as 0092, and the Parent Unit Type as 06. In the Analysis of the Present Status we find 92 BOMB HV [92nd Bombardment, Heavy]. The parent unit type 06 means Group, giving us the complete title, 92nd Heavy Bombardment Group. To confirm the interpretations of these alphanumeric codes, I consulted various private World War II POW and army air force–related web sites.¹⁹ I then compared hundreds of records in these online collections and cross-referenced them with the information in the World War II POW punch card records to confirm the interpretations of Type of Organization and Parent Unit Type codes (see sidebar: "Meanings of Type of Organization Codes"). By this method, I am confident that I have broken 25 of 235 alphanumeric codes. In time, hopefully, the same codebreaking methods may be successfully applied to the remaining unbroken codes.

In order to interpret the Type of Organization codes found in the World War II POW punch cards, it was necessary to discover additional codebooks or find some way to break the codes. Three coding systems were identified: a three-digit numeric system (e.g., 215) and two alphanumeric systems (e.g., S53 and 401A). The first two coding systems were used as early as May 1942. After the third coding system was introduced in 1944, all three systems were used simultaneously.

All three systems were based on two common sources: the table of organization and equipment (TOE) and troop basis reports. The TOE provided a numeric system of hierarchical organization for the U.S. Army and Army Air Forces together with brief "titles" or organization identifiers.²⁰ TOE numbers and units were printed in small pamphlets issued on a regular basis as new numbers and organizations were added or old ones changed. They did not contain all numbers and organization names. Much more extensive and inclusive were the monthly troop basis reports, which specified organization strength in officers and enlisted men. They were based on the numbers and titles found in the TOE but were briefer in their "type unit" descriptions.²¹

More voluminous still, and more useful for identifying specific organizations and their component units, is the Analysis of the Present Status of the War Department Troop Basis (December 1, 1944). This monthly statistical report was based upon the troop basis, which specified authorized unit strength, but it also identifies all organizations and their units in the hierarchical order laid down in the TOE. These analyses revealed both authorized and actual strength and were the basis for making decisions about where to send troops and how many more men would need to be drafted, trained, and supplied; i.e., it was the staffing plan for the U.S. armed forces. In format, the Analysis of the Present Status of the War Department Troop Basis is a step closer to the surviving codes so far discovered. The TOE number remains constant, and the "type unit" descriptions are similar to those found in the troop basis. What is different is that the Analysis identifies specific units and their positions in the TOE hierarchy (e.g., 43 TROOP CARR designates the Forty-third Troop Carrier).²²

To track changes in unit strength and compute differences between authorized and actual numbers, the AGO created codebooks that set up a one-to-one correspondence between the TOE numbers and troop basis type unit descriptions and the codes punched into the Type of Organization and Parent Unit Type fields.

As early as 1941, an effort was made to encode unit type information:

In this early version of subunit codes, "09" means "company," and "A" is its designation. In the parent unit pair, "06" means "regiment," and "12" designates this as the Twelfth Infantry Regiment. Even at this early stage, Parent Unit Number was the actual number of the unit, not a code, and the unit type was already encoded.²³

Changing the Type of Organization and Parent Unit Type Codes (1944)

Late in 1944, the Type of Organization field was expanded from three to four characters consisting of three numbers and a letter. The Parent Unit Number henceforth was the "actual division number." This change appears to have been made in order to identify units within the larger divisional hierarchy, not just as types of units, as was previously the case. The first character indicates the type of division. The second and third characters specified the sequence for sorting and printing the units in the same order in which they appeared in the Troop Basis. The fourth, alphabetical, character denotes the arm or service of the organization.

The two-character Parent Unit Type field identified all units except divisions with regard to size and importance in the chain of command. It no longer was used to specify whether a unit was a regiment, battalion, or some other type of unit.

How to Punch a Four-Character Code into a Three-Character Field

The new set of coding instructions posed a dilemma. Keypunch operators now had to squeeze a four-character Type of Organization code into a three-character field. The record for POW Staff Sgt. William T. Reyenga illustrates how operators solved this problem. Reyenga was attached to the infantry, which was coded 722P. The last three characters, 22P, also match two other infantry organization types:

522P ARMORED INFANTRY REGIMENT
722P INFANTRY
822P INFANTRY

Since the first character, 7, indicated "type of division," or infantry, and this was already indicated by the arm or service code, the first digit was apparently stripped away to create the three-character code 22P. In order to compute the actual versus authorized strength of specific units, the Type of Organization, Parent Unit Number, and Parent Unit Type fields had to be completed. Such information was not provided by the Office of the Provost Marshal General. The information may have been taken from the status or report of change cards that identified the individual's organization. Once the card was found, the keypunch operator punched the codes using one of the available codebooks.

Conclusion

The three coding systems used by the World War II POW punch cards for the Type of Organization and Parent Unit Type fields can now be broken with a great deal of effort and a little luck. More of both will be necessary if we are to complete the task. Still, we are far ahead of where we were in understanding these coding systems and can make the resultant code interpretations available to researchers, while explaining which code values come directly from extant War Department documentation and which are interpretations. At the same time, however, enigmas remain. Why were three systems developed and used simultaneously? What guided the selection of any coding system? Hopefully, additional codebooks or correspondence will be found that will answer these questions.

For whatever reasons, these early coding systems, and the data fields themselves, were abandoned by the time of the outbreak of the Korean conflict. The concept, however, remained in the form of the troop sequence number (TPSN) codes found in the Records of the Office of the [Army] Adjutant General, RG 407, Korean War Casualty File, 1950–53.²⁴ It uses an eleven-character number, providing the flexibility and power sought by the codemakers of World War II.

The value of IBM machines to the war effort was clearly proven each and every time the troop basis or analysis was printed. These machines allowed the Office of the Chiefs of Staff to plan and to direct men and material where they were needed, when they were needed. The Machine Records Units and their machines won the logistical battles that helped win World War II.

Meanings of Type of Organization Codes

Notes

¹ Confidential Report to Mr. L. H. La Motte, "Our Dealing with War Ministries in Europe," May 16, 1940, p.1; Modernization of Army Administrative Methods Through Use of Business Machines Reorganization of Army Personnel System Initiation of Machine Records System, box 1629, 310 Business Methods and Procedures From (5-21-40) (1) Section 1 To Section 1A, Army Adjutant General's Decimal File 1940–1945, Records of the Adjutant General's Office 1917–, Record Group (RG) 407, National Archives at College Park, MD (NACP). Hereinafter referred to as Modernization of Army Administrative Methods.

² Ibid.

³ Margaret O. Adams, "Punch Card Records: Precursors of Electronic Records," American Archivist 58 (Spring 1995): 182–201.

⁴ Lt. Col. Carl G. Allen, "The Use of Machine Records by the Army," p. 7, IBM War History Files, copy provided from collection of Adjutant General Corps Museum, Fort Jackson, SC.

⁵ Ibid., p. 10.

⁶ Memorandum for the Administrative Assistant, AGO, Subject: Contingent Expenses, War Department, 1942, from Maj. H. G. Holdridge, AGD War Plans Officer, Apr. 23, 1940, Modernization of Army Administrative Methods, RG 407, NACP.

⁷ Memorandum for the Chief of Staff, through the Assistant Chief of Staff, G-1, Subject: Modernization of Army Administrative Methods through use of Business Machines, from Maj. Gen. E. S. Adams, Adjutant General, May 25, 1940, Modernization of Army Administrative Methods, RG 407, NACP.

⁸ "Possible Uses For President's Contingency Fund (Personnel)," May 18, 1940, Modernization of Army Administrative Methods, RG 407, NACP.

⁹ Memorandum for the Assistant Chief of Staff, G-1, Subject: Reestablishment of Post and Regimental Personnel Sections, ca. October 1940, pp. 2–4, Modernization of Army Administrative Methods, RG 407, NACP.

¹⁰ Ibid., p. 5.

¹¹ Ibid., pp. 11–12.

¹² Ibid., pp. 6–9.

¹³ Memorandum to Chief of the Budget and Legislative Planning Branch, War Department, General Staff, Attention Colonel Brown, from Lt. Col. R. E. Cummings, Adjutant General, Jan. 24, 1941, p. 2, Modernization of Army Administrative Methods, RG 407, NACP.

¹⁴ Allen, "The Use of Machine Records by the Army," p. 8.

¹⁵ Cable from Presidio of San Francisco 455P, Dec. 15, 1941, to the Adjutant General, Army Adjutant General Decimal File 1940–1945, box 1626, 310 Business Methods and Procedures (10-6-41) to (12-31-41), RG 407, NACP.

¹⁶ Memorandum from Brig. Gen. H. B. Lewis, Acting Adjutant General, to Commanding Generals, Subject: Machine records unit personnel, July 31, 1942, box 1628, 310 Business Methods and Procedures (10-6-41) to (12-31-41), RG 407, NACP.

¹⁷ Allen, "The Use of Machine Records by the Army," p. 11.

¹⁸ See Ted Hull "Electronic Records of Korean and Vietnam Conflict Casualties," Prologue: Quarterly of the National Archives and Records Administration 32 (Spring 2000): 54–61.

¹⁹ See www.axpow.org/biographies.htm (formerly combatvets.net/asp/biopage.asp) for POW biographies; www.armyairforces.com/ for lists of bomber groups based in the European Theater and photographs of aircraft; www.littlefriends.co.uk/ for an extensive lists of the 20th, 55th, 78th, and 364th Fighter Groups' aircraft, pilots, casualties, and POWs; and http://www.fourthfightergroup.com/.

²⁰ Table of Organization and Equipment (1943–1945), Table of Organization and Equipment (TAGO Series TOE) Beginning TOE Index, Ending: TOE 1-167T, box TOE1, Publications of the U.S. Government, U.S. Army, 1941–, RG 287, NACP.

²¹ The War Department Troop Basis (Apr. 1, 1945), box 63, DRB Reference Collection, Document No. 214, War Department Troop Basis (December to May 1945), RG 407, NACP.

²² Analysis of the Present Status of the War Department Troop Basis (Dec. 1, 1944), box 117, Office of the Comptroller of the Army, Program Review and Analysis Division Statistics Branch, Records of the Army Staff, RG 319, NACP.

²³ Army Regulations No. 330-63, Machine Records Codes, Code 63—Unit and Installation Code—Types of Units and Installations, Army Adjutant General Decimal File 1940 - 45, box 1626, 310 Business Methods and Procedures 10-6-41 to 12-31-41, RG 407, NACP.

²⁴ [U.S. Army] Korean War Casualty File (TAGOKOR), RG 407, NACP.

Lee A. Gladwin is an archivist with the Electronic and Special Media Records Services Division of the National Archives and Records Administration.