change in the government's Asian posture. Japan's continued expansion did not warrant leniency. More important, reopening trade in oil with Japan would have adverse consequences in China and Japan. It would, Hull feared, be a blow to Chinese morale and might lead to a collapse of Chinese resistance. Within Japan, it would serve to reward those unregenerate groups who argued that continued expansion, not moderation, was the wisest policy. Reopening of trade could do no good in Japan and only harm in China. Though dangerous and unplanned, the de facto embargo would have to remain. Consequently, Hull ordered that no new restrictive measures be introduced and that the present attitude not be relaxed. "Whether or not we had a policy," Acheson wrote, "we had a state of affairs that would continue." "37

37 Yost memo, Sept. 5, 1941, Hornbeck Papers; Acheson, Present at the Creation: My Years in the State Department (New York, 1969), p. 26. Hull's reluctance to alter a course of action is revealed in numerous instances of U. S.-Japanese relations

That state of affairs was a result of the FFCC's refusal to release frozen Japanese funds in August. Limited funds for oil exports to Japan might have been released if Acheson had used his influence as the State Department's representative to insist that the plan of July 31 be precisely followed. Acheson, however, was not inclined to use his influence in that way. In part, his blocking of oil exports may have been due to the fact that he was not fully aware of the impact an oil embargo would have upon

during the Sino-Japanese war. One is the U. S. refusal to enter into negotiations for a new trade treaty in February 1940. To have done so would have misled the Japanese into thinking that their behavior had been acceptable. Foreign Relations of the United States, Japan, 1931-1941 (Washington, 1943), vol. 2, p. 193. Another instance involved the aviation gasoline embargo. The British approached the State Department in August 1940 seeking a quota of aviation gasoline for Japan to reduce tension the embargo was causing in Japanese-Dutch relations. Welles responded that having taken its stand the United States could not depart from it. Telegram, Lothian to Foreign Office, Aug. 6, 1940, file F.O. 371/ 25212/W9358/9160/49. See also Acheson, Present at the Creation, p. 36.

Japanese-American relations. Acheson was not particularly familiar with Asian affairs and was never privy to the intercepted Japanese messages which revealed the strain Japan was under. "I knew little... of the tension [the embargo] was producing there," he later wrote. Even if Acheson had been aware of the tension, it is not likely that he would have changed his mind. Though the term had not yet been applied, Acheson was a "hawk." He believed in a firm stand against Japan. In June 1941 he had advocated a total oil embargo against Japan and had promised Ickes he would speak to Welles about such an embargo. Morgenthau, second only to Ickes in his desire for an oil embargo against Japan, considered Acheson the only person in the State Department with whom he could speak candidly. Acheson himself later lamented that only his own corner of the State Department supported Morgenthau's "campaign to apply freezing controls to Axis countries."

" 38

Acheson found himself in a position to create a de facto embargo on oil more by accident than by design. Hired as an assistant secretary of state in January 1941, he had little influence on Asian affairs. His job consisted of working on the economic aspects of commercial treaties and agreements. Although necessary work, it fell far short of influencing Asian policy which rested in the hands of Secretary Hull and a sizeable staff of career Asian specialists. Acheson languished in the commercial back room of the State Department for only a month before Assistant Secretary Adolf A. Berle, Jr., nominated him to coordinate State and Treasury Department problems concerning freezing of foreign assets. After this first step it was logical that Acheson would be appointed the State Department's representative to the FFCC in June 1941 when German and Italian funds were frozen.39 A month later Japanese funds were frozen, and Acheson found himself controlling the very lifeblood of Japan.

Memories and memoirs are often unreliable when they try to reconstruct events a quarter century or more in the past. Perhaps that is why the reader of Acheson's memoirs will not

38 Ickes, Secret Diary, vol. 3, p. 551; Acheson, Present at the Creation, pp. 20, 23, 36. "Hawk" is the term Acheson used to describe his views in 1941.

39 Acheson, Present at the Creation, pp. 16-17, 22-23; Adolph A. Berle, Jr., Navigating the Rapids (New York, 1973), p. 358.

find this story told there. In fact, Acheson was rather humble in recounting his own efforts. In Present at the Creation, Acheson recalls that Welles told him, in late July, that the happiest solution would be to take no action on Japanese applications for funds. Treasury Department documents show just the opposite: "Mr. Acheson has suggested to Mr. Welles that pressure be applied to Japan through the medium of freezing control licenses." 40 It was on the last day of July that the president approved a policy which specified that the FFCC would "continue to hold without action applications relating to petroleum exports from the United States and subsequently grant license under the freezing order only in accordance with the policy to be initiated by Export Control."41 Acheson made no reference to the last portion of the presidential plan or to the fact that Export Control initiated a policy which granted petroleum export licenses to Japan. He also neglected to mention that the State Department's Far Eastern Division was upset with him for blocking Japanese oil exports. In fact, Acheson's memoirs ignore the events of August 1941. If it had not been for the men of the Treasury Department's Foreign Funds Control Division who wrote down what was said to them, Acheson's part in those events might have gone unrecorded.

Dean Acheson and the Foreign Funds Control Committee are important because they illustrate how the proliferation of independent organizations decreased the control of the policymakers over the conduct of foreign relations. In 1940-41 the Petroleum Coordinator for National Defense, the Administrator for Export Control, the Advisory Commission to the Council of National Defense, and the Foreign Funds Control Committee influenced oil sales to Japan. Consequently, it was no longer possible to assume that the actions of the nation reflected the will of the policy formulators. As the events of July and August 1941 indicate, what Roosevelt and Hull considered a cautious policy became bold action in the hands of those who were supposed to follow orders but preferred to lead the nation.

CHARLES M. DOLLAR

The National Archives stands between the proliferating computer-generated information within federal agencies and the increasing number of researchers whose work requires access to such information. The National Archives, therefore, is the link between this rapidly growing computer-information base and the growing number of scholars engaged in computer-aided research. The dimensions of this activity can best be determined by looking at the growth of computer-generated information in the government, the changing interests and skills of researchers, and the evolution of the National Archives.

A discussion of computer-generated information in the federal government must begin with the year 1886 when Herman Hollerith, a former employee of the Bureau of the Census, completed a system that sorted cards in which holes had been punched to represent census data.1 Hollerith developed this system because tabulation of the 1880 census still was not com

An earlier version of this article was presented at the Second International Conference on Computers and the Humanities, April 3-6, 1975, University of Southern California at Los Angeles. Several of my colleagues at the National Archives offered constructive comments, especially James E. O'Neill, deputy archivist of the United States.

'Robert H. Gregory and Richard L. Van Horn, Automatic Data Processing Systems, Principles and Procedures (San Fran

plete (and indeed was not completed until the middle of 1887), and he anticipated total chaos when it came to sorting the data of the census of 1890 unless machine tabulation was adopted. Subsequently, the use of Hollerith's system so speeded the tabulation of the census of 1890 that it was completed in two and one-half years and later censuses relied heavily on punched cards.

During World War I when the army conducted large-scale psychological testing of soldiers, punched cards were used to record the information and to tabulate it. Later, the passage of the Social Security Act of 1935 required the federal government to keep the employment records of some 30 million people current. This entailed punching, sorting, and checking as many as one-half million cards a day. By the late 1930s the federal government was the major user and processor of such cards.

This function was extended still further by the events of World War II. In 1944 a crude electromechanical computer, Mark I, was developed to generate firing tables for the U.S. Navy. A year later the first electronic computer,

cisco, 1960), pp. 627-28. For a full account of the development of punched cards, see George Jordan, "A Survey of Punched Card Development" (Master's thesis, Massachusetts Institute of Technology, 1956).

Electronic Numerical Integrator and Computer (ENIAC), went into operation.2

After World War II the U. S. Army and the Manhattan Engineer District became deeply involved in advancing the development of computers. The Office of Naval Research formed a computer section in 1947 to examine the impact of computers on science and technology. At this time the National Bureau of Standards also began research in computers and soon developed a computer called SEAC for Standard Eastem Computer. This activity was shortlived, however, as private industry rapidly took up the development of computers. The result was the development of the first modern generalpurpose computer called UNIVAC for Universal Automatic Computer. This model was installed at the Census Bureau in 1951 to process the census returns of 1950. Compared with prevailing standards UNIVAC could store and manipulate enormous amounts of information with great speed, and its usefulness was quickly recognized. In 1952 there were five computers in the federal government. During the next decade this number grew to 1,006. By 1974 the number of computers being used by the federal government exceeded 7,800.3

More important than the number of computers is the amount of information created and processed by them. Computer-generated information now stored on tape represents about 20 percent of all the records of the federal government. Currently, agencies of the federal government store and process information on the equivalent of 11 million computer tapes.4 Furthermore, it is estimated that approximately 50 percent of the information processed by federal agencies today is computer-based and is in

2 Herman H. Goldstine, The Computer From Pascal to von Neumann (Princeton, 1972), p. 226.

3 This number excludes analog computers and digital computers designed to function as part of a weapons or space system. General Services Administration, Federal Supply Service, Inventory of Automatic Data Processing Equipment in the United States Government for Fiscal Year 1974 (Washington, 1974), p. 2.

4 The actual number of reels is about eight million. Computer tape technology in the early 1960s was such that a twenty-four-hundred-foot reel of magnetic tape would store the equivalent of fifteen 500-page books. By the end of the 1960s such a reel could hold twice that amount. Current tape technology soon will permit storing the equivalent of one hundred and twenty 500-page books on a single twentyfour-hundred-foot reel of magnetic tape. L. G. Sebestyen, Digital Magnetic Tape Records for Computer Applications (London, 1973), p. 10.

creasing every year. Within the next decade we can anticipate that 80 percent of the information created by the federal government will be processed, stored, or retrieved by computers.

Almost every federal agency now uses computers to generate and store information. The largest of these is the Department of Defense. While much of its information is routine bookkeeping, there are many magnetic tape files that contain valuable military information.5 One such file is that of the Hamlet Evaluation Program, which was designed to assess the effectiveness of the army's pacification program in South Vietnam.6 The results of this program were used in formulating the Vietnamese strategy and goals of the Johnson administration. Other federal agencies that create substantial computer-based information include the Internal Revenue Service, the Department of State, the Bureau of the Census, the Bureau of Labor Statistics, the Environmental Protection Agency, the Department of Agriculture, and the Department of Health, Education, and Welfare, to list a few.

The quantity of computer-based information handled in these agencies and others is overwhelming. For example, the State Department now has a system whereby annually about onehalf million incoming telegrams from embassies and consulates are received by computer and stored automatically. The messages are also indexed by computer and assigned a reference number. Since they are arranged in the order they are received in, easy access to a particular telegram or telegrams on a specific subject is possible through a machine-readable index.7

Another example is that of the Department of Health, Education, and Welfare, which is made up of a number of separate offices, each of which carries out programs that create enormous amounts of computer-based information. Its Office of Economic Opportunity sponsored studies of poverty, using census data from 1960 and 1970. The 1970 study of poverty contained

5 For a useful review of machine-readable records relating to South Vietnam, see Rand Corporation, "Southeast Asia Combat Data Project: Symposium Proceedings, Program Review and Status Report, March 1975."

6 A file is a collection of information containing logically related records pertaining to a specific subject. Such a collection may be on a single reel or on several reels of tape.

7 For an informed discussion of this system, see Milton O. Gustafson, "Archival Implications of State Department Recordkeeping," Prologue, vol. 7, no. 1, pp. 36-38.

summary statistics on individual households for every county in the United States. Included was information about race, income, work experience, occupation, age, sex, and education. It has been estimated that a complete printout for every state and county would require approximately four million pages. This list could be greatly expanded.

Researchers who wish to study the 1960s and the 1970s must accept the fact that the bulk of the research data is stored on computer tapes and is accessible only by using a computer. Already, their changing interests and skills are evidenced in a number of new research journals.8

While letters, diaries, governmental reports, speeches, and comparable sources continue to be the media of historical research, many historians recognize that these materials do not always tell the story of the ordinary citizen. Consequently, more and more historians are interested in detailed records that describe the activities of ordinary people and which capture the transactions of business and government. This means information that has not been aggregated into a summary table. For example, health statistics of individuals that contain no identifying characteristics such as name, address, or social security number will be of greater value to historians three decades from now than a printed table that summarizes that information. The abundance of computerbased information at this level- the level of the lowest recorded unit, will permit historians to write about the 1970s in terms of the many rather than the few. Similarly, the availability of the computer-based information that was used to formulate policies in business and government will permit historians of the future to write about the 1970s with the advantage of having access to the same information that the policymakers had. The availability of such information with all of its ramifications for his

The professional journals of these disciplines clearly indicate this trend. The establishment of the journal Computer and the Humanities and the Historical Methods Newsletter suggests the same interest in the humanities and history.

9 It should be noted that most of the machine-readable files in the division's holdings have had identifying characteristics such as name, social security number, and place of residence deleted by the agency of origin. Even files declared "disclosure free" are carefully checked to ensure that identifying characteristics have indeed been erased by the agency. The National Archives prevents invasion of privacy by not granting access to a restricted file until all conditions imposed by the agency of origin have been satisfied.

torical research will lead historians to acquire skill in using computers. As such skill becomes commonplace, historians will turn increasingly to the enormous mass of computer-based information created by the federal government.

Having addressed the growth of computergenerated information in the federal government and the changing interests and skill of researchers, it is appropriate to focus upon the response of the National Archives to the challenge this material presents. Since its establishment in 1934 the Archives has sought to preserve and to make available basic records that relate to the necessary processes of the federal government, the protection of public and private rights, and the interests of scholars, students, and the general public. It is fair to say that for the first twenty-five years of its existence, the main thrust of preservation at the National Archives was directed toward textual or printed records. For example, in July 1936 the chief clerk of the Bureau of the Census consulted the Archives to ask whether he could dispose of some eight million punched cards containing information from the 1930 Census. Permission was given to destroy them because at that time no one considered them to be records. 10 Furthermore, it was argued, the cards were redundant, since tabulations derived from them existed in printed form. As late as 1961, a preliminary study at the National Archives recommended that magnetic tape be designated only an interim storage medium and that paper or microfilm continue to be the medium of permanent storage.1

11

Within a few years, however, it became evident that this view was undergoing change. In 1963 the Social Science Research Council on the Preservation and Use of Economic Data registered its concern that economic data stored on punched cards and magnetic tape in various federal agencies be preserved. 12 This concern was shared by senior officials at the National Archives who had begun to grasp the implications of the expansion of computer-based information in federal agencies. A study conducted

10 "Punched Cards in the Census Bureau," chief, Division of Department Archives to director of archival service, July 29, 1936, Central Files, 1935-44, Records of the National Archives and Records Service, Record Group 64, National Archives.

Richard Jacobs, "Records Preparation and Magnetic Tape," Training Lecture no. 56, National Archives and Records Service, (Washington, 1961), p. 2.

12 "Data Archives Staff Report, August 1970," National Archives and Records Service, (Washington, 1970), p. 1.