Archives Library Information Center (ALIC)

Archives and Records Management Resources

by T. R. Schellenberg

Bulletins of the National Archives
Number 8 (October 1956)


Informational values derive, as is evident from the very term, from the information that is in public records on the matters with which public agencies deal; not from the information that is in such records on the public agencies themselves. The greater proportion of modern public records preserved in an archival institution are valued less for the evidence they contain of Government action than for the information they contain about particular persons, situations, events, conditions, problems, materials, and properties in relation to which the question of action comes up. Most of the larger series of records in the National Archives, for example, were accessioned primarily for the information they contain relating to other matters than the action of the Government itself. Among such series are the voluminous census schedules, military service records, pension files, passenger lists, land-entry papers, and various kinds of case files. In most instances such series shed light on the activity of Government agencies, but so little in proportion to their bulk that this is not an important factor in their selection for preservation; it is presumed that other records show the activity of the agencies more effectively.


In appraising the value of information in public records, the archivist is not greatly concerned with the source of the records -- what agency created them, or what activities resulted in their creation. The concern here is with the information that is in them. There are a number of tests by which informational values of public records may be judged. These are (1) uniqueness, (2) form, and (3) importance.


The test of uniqueness must be carefully defined if it is to be meaningful. In applying the test the archivist must consider both (1) the uniqueness of the information, and (2) the uniqueness of the records that contain the information.

The term "uniqueness," as applied to information, means that the information contained in particular public records is not to be found in other documentary sources in as complete and as usable a form. Information is obviously unique if it cannot be found elsewhere. But information in public records is seldom completely unique, for generally such records relate to matters that are also dealt with in other documentary sources, and the information they contain may be similar or approximately similar to that contained in the other sources. To be regarded as unique for appraisal purposes the information need not be completely dissimilar from all other information. But it should pertain to matters on which other documentary information does not exist as fully or as conveniently as in public records.

In applying the test of uniqueness to information in records, an archivist must bring into review all other sources of information on the matter under consideration. These sources encompass materials produced outside as well as within the Government. The materials produced outside may be published or unpublished: they may consist of private manuscripts, newspapers, books, nearprint materials, or any other form of documentation. The Government materials are the various record series relating to the matter under consideration. The archivist must understand the relation of such series to each other and must be able to identify the particular series that should be preserved. To determine if a body of records is the sole adequate source of information on a given matter, he needs to be a real expert in the subject -- acquainted with all outside resources and the products of research as well as with the other records of the Government dealing with the subject in question. The Federal archivist should know of all the significant documentation that relates to his field of specialization; the State archivist should ordinarily know of all the significant documentation relating to the history of his State.

In applying the test of uniqueness to the form of the records rather than to the information contained in them, the matter to be considered by the archivist is the physical duplication of the public records. In the Federal Government of the United States, as is well known, there is a great and perhaps an unnecessary proliferation of records. Not only are records duplicated from one administrative level to another, but within a given Government office several copies of a particular record may exist. While records having informational values are not likely to be found in as many forms or as many series as are records having evidential values, it is nonetheless necessary to carefully compare records containing information on any particular matter to avoid retaining more than one copy of them. To illustrate: records containing economic data filed by various business firms with the Office of Price Administration to obtain price adjustments were physically duplicated, to a certain extent at least, in the regional and national offices, and within the national office in the Enforcement and Price Departments of the agency. A collation of the price adjustment records was necessary to avoid keeping duplicate copies.

Because of the greater technical difficulties our ancestors faced in publishing or duplicating information and because of the inevitable loss of many records through the centuries before archival care became general, records of the remote past are likely to be the only remaining source of information on many matters with which they deal. This fact led the German archivist Meissner to formulate a maxim that "old age is to be respected" (Footnote 7) in records. Archivists of various countries have set chronological date lines before which they propose that all records shall be kept. In Germany the date is 1700, in England 1750, in France 1830, and in Italy 1861. The Italian date corresponds fairly closely, by historical coincidence, to that adopted by the National Archives of the United States, where almost all surviving records created before the Civil War, which began in 1861, are being preserved.

While public records are likely to be more valuable as a source of information when other kinds of documentary materials are scanty, the converse of this statement is also true. The proportion of public records requiring permanent retention diminishes as other kinds of documentary materials increase in quantity. It is doubtful if governments are justified, in the face of other forms of recent documentation, in keeping more than a small proportion of the voluminous contemporary public records. But an archivist's job of appraisal increases in difficulty as the documentation of society increases in quantity. He must apply standards of selection with constantly greater discrimination as he deals with more recent records; in particular, he must apply the test of uniqueness to them with great severity. For "of the making of many books" -- and of many other types of documentary materials -- "there is no end," to paraphrase the Preacher.


In applying the test of form the archivist, again, must consider both (1) the form of the information in records and (2) the form of the records.

As applied to information, the term "form" relates mainly to the degree to which the information is concentrated. Information may be concentrated in records in the sense that (1) a few facts are presented in a given record about many persons, things, or phenomena, or (2) many facts are presented about a few persons, things, or phenomena, or (3) many facts are presented about diverse matters -- persons, things, and phenomena. In the first case, the information may be said to be extensive, in the second intensive, and in the third diversified. Census schedules and passenger lists, for example, provide extensive information in the sense that each schedule or list pertains to many persons. Case files of various labor boards and other adjudicative, investigative, or regulatory bodies serve as examples of records containing intensive information about a limited number of particular matters. Reports of county agents of the Agricultural Extension Service and of the consular and diplomatic agents of the State Department serve as examples of records containing information about diverse matters. In their pamphlet the British archivists expressed their ideas about the concentration of record information in their criterion that business records should be preserved which "affect, name, or touch by inference a large number of persons and/or things or topics," and particularly "if both persons and things are involved in quantities." In general, records that represent concentrations of information are the most suitable for archival preservation, for archival institutions are almost always pressed for space to house records.

The term "form" as applied to the records rather than to the information contained in them relates to the physical condition of the public records. Physical condition is important, for if records are to be preserved in an archival institution, they should be in a form that will enable others than those who created them to use them without difficulty and without resort to expensive mechanical or electronic equipment. Chemistry notebooks, for example, are not likely to be intelligible to others than the chemists who recorded the results of their experiments in them; while punchcards and tape recordings are commonly unusable without resort to expensive equipment.

Arrangement is also important. Certain record series may be preserved by the archivist simply because they are arranged in a particularly usable manner. If he has a choice among several series relating to a given matter, he will choose for preservation the series whose arrangement most facilitates the extracting of information. For example, reports of American agricultural agents and attache's, though duplicated in the files of the State Department, are being preserved as a separate series accumulated by the Foreign Agricultural Service of the Department of Agriculture because their arrangement makes it easier to use them than the copies of the reports embodied in the classified filing system of the State Department.


In applying the test of importance, the archivist is in the realm of the imponderable, for who can say definitely if a given body of records is important, and for what purpose, and to whom? An archivist assumes that his first obligation is to preserve records containing information that will satisfy the needs of the Government itself, and after that, however undefinable these needs may be, private scholars and the public generally.

He should take into account the actual research methods of various classes of persons and the likelihood that they would under ordinary circumstances make effective use of archival materials. He will normally give priority to the needs of the historian and the other social scientists, but he obviously must also preserve records of vital interest to the genealogist, the student of local history, and the antiquarian. He should not, however, preserve records for very unlikely users, such as persons in highly specialized technical and scientific fields, who do not use records extensively in the normal exercise of their professions and are not likely to use archival materials relating to them.

Public records may have a collective, as well as an individual significance. Research values are usually derived from the importance of information in aggregates of records, not from information in single items. Records are collectively significant if the information they contain is useful for studies of social, economic, political, or other phenomena, as distinct from the phenomena relating to individual persons or things. Records of the General Land Office, for example, collectively show how the public domain passed into private hands and how the West was settled; individually, the land-entry papers also have value for biographical studies and for studies of family history. In his article on "The Selection of Records for Preservation" in The American Archivist for October 1940, Dr. Philip C. Brooks has correctly observed that ". . . most records having historical value possess it not as individual documents but as groups which, considered together, reflect the activities of some organization or person or portray everyday, rather than unique, events and conditions."

Records relating to persons and things may, of course, have an individual research value in relation to particular persons or things. Normally, the more important the person or thing, the more important is the record relating to it. Such records may also have sentimental values because of their association with heroes, dramatic episodes, or places where significant events took place. Usually such values are attached to single record items, such as the Emancipation Proclamation, though extreme sentimentalists sometimes attach them to all records relating to the objects of their reverence, no matter how voluminous or trifling they may be. Utility for determining significant facts is with such persons only a secondary consideration. But archivists must exercise their sense of proportion in judging sentimental value.

Before applying the test of importance, an archivist should be sure that records meet the tests of uniqueness and form. The test of importance relates, as has been noted, to imponderable matters -- to matters that cannot be appraised with real certainty. The tests of uniqueness and form, in contrast, relate to ponderables -- to matters that are capable of being appraised on the basis of ascertainable facts.

An archivist normally brings to his task a general knowledge of the resources and products of research, which he acquired during his academic training. In the discharge of his duties he normally acquires a specialized knowledge of subject-matter fields pertinent to the records with which he works. And while performing reference service he learns to know of genuine research needs. He will also acquire a knowledge of the documentation produced by the agencies with which he deals so that he can reduce to manageable proportions the quantity of records that must be used for research. But if he does not have such knowledge, he should deliberately seek it by searching out and comparing the documentation available on various matters; and if his investigation fails to yield an answer he should not hesitate to consult subject-matter specialists.


Let us now see how the tests of uniqueness, form, and importance have been applied to groups of records in the National Archives containing information on (1) persons, (2) things, or (3) phenomena. In discussing information relating to these three matters it is not assumed that records relate exclusively to one or the other of them; they may, and often do, relate to more than one of them.

Records Relating to Persons

The term "persons," it will be recalled, was defined to include both individuals and corporate bodies. The values of records relating to persons will be discussed with reference to the information they contain on the persons themselves, not with reference to their information on the conditions, problems, situations, and the like, that affect the persons.

Records relating to persons are produced in great quantities by modern governments. Certain types of records, like census schedules, are intended to cover all human beings in a country; others, relating to specific classes, often represent large segments of the population, such as laborers, farmers, soldiers, and recipients of social welfare services; still others relate to even more specialized classes, such as transient Mexican or Puerto Rican laborers, Indians, and other nationality groups. As the controls of the Government over its individual citizens are extended, more records are created in relation to them. With universal military service, for example, records are created on the entire male population of a certain age group which may duplicate, in part at least, the information contained in census schedules. Records on a given soldier, again, may be created in relation to various phases of his military life -- his service in the armed forces, his medical history, his retirement and pensioning. And these records, in turn, may be supplemented by records on his life as a civilian, such as on the taxation of his property, his relation to governmental welfare programs, the control of his business if it is of the type that is subject to Government regulation, and various other of his activities that may bring him in touch with his Government. Social welfare activities, in particular, result in the production of voluminous records pertaining to poverty and dependency, crime and delinquency, disease and sanitary problems, and the like.

The problem of deciding which records on human beings to keep is a particularly difficult one. The records obviously are very great in quantity and duplicative in content. The information in them about persons is largely impersonal in character, particularly in recent years as the relations of the Government with its citizens have become more formal and impersonal. The information about any particular person, moreover, is not extensive, and often consists of nothing more than the bareboned facts necessary to establish his identity. The records contain few of the intimate details that are found in diaries or personal correspondence.

If considered singly and solely with reference to the personal information they contain, most records pertaining to persons have relatively little research value. From the point of view of their significance for demographic, sociological, or economic studies, they are usually important only in the aggregate. For such studies they have value only if used collectively and because of their information on phenomena that concern a number of persons, and not because of their information on single persons. And summarizations of the data they contain are usually available in statistical enumerations and tabulations, either in published or unpublished form. From the point of view of their historical or biographical significance, they are important individually only to the degree that the persons to whom they pertain are important. An archivist obviously will preserve all records, whatever their character, for notable persons who lived in the past; but how is he to know who will become notable among the millions about whom records are now being created?

Among the large series of records in the National Archives that pertain to persons are records of the censuses of population. These serve to illustrate most of the problems that arise in the appraisal of personal records.

Since the original schedules are very voluminous, an archivist is justified in questioning, momentarily at least, if printed statistical summarizations of their contents will not adequately meet scholarly needs. A goodly proportion of the general information in the schedules is available in published final reports. The Statistical Abstract of the United States, which is issued periodically by the Bureau of the Census, also contains a wealth of statistical data. It is supplemented by Historical Statistics of the United States, 1789-1945, published by the Bureau in 1949; and a supplement to this was issued in 1954. The statistical publications of the Bureau are so numerous that a sizable book is required to list them -- the Catalog of United States Census Publications, 1790-1945, published by the Bureau in 1950.

The unpublished social data produced by the Bureau, however, are also quite valuable, according to C. Luther Fry, who some years ago wrote an article on "Making Use of Census Data," published in the Journal of the American Statistical Association for June 1930. Fry points out that the data on population contained in unpublished enumerations and tabulations of recent years are classified according to sex, color or race, nativity, and parentage; that they show various social phenomena, such as the rural population by counties, the marital status of classes of population, and facts concerning the tenure of homes; and that generally they are more adaptable to the purposes of research than the published statistics because they are broken down by smaller localities.

The original schedules of the censuses of population from 1790 to 1880 are, nevertheless, preserved in the National Archives, which is one of the few large archival institutions in the world that preserves this type of record. While the published and unpublished statistical summarizations are likely to contain most of the information needed by the scholar, such summarizations are occasionally found to be inadequate. This is the case, for example, for studies of the settlement or the movement of nationality groups that can be identified only by noting the names of individuals. Here the original schedules must be used. The schedules are used occasionally, also, by scholars wishing to obtain or verify basic facts about persons in historical or biographical studies. They are used most extensively, however, for genealogical searches and for establishing facts about persons that are ordinarily derived from vital records when such records are available. To a degree, then, population census schedules meet the tests of uniqueness and importance.

The information on the population of the United States that is provided in the census schedules is comprehensive both as to time and place. The schedules provide almost complete coverage of the population of the country at intervals of ten years, and they represent concentrations of information about individuals, containing many facts about persons in relatively small compass (though the later census schedules are so voluminous that they have been reduced to microfilm). Beginning with the census of 1850 the schedules usually show the name and age and the State, territory, or country of birth of every free inhabitant in the United States. While census schedules vary in their content from one country to another, and within a country from one census to the next, they usually contain information in regard to the personal characteristics (family, sex, marital status, age), the political status (birthplace, nationality, language, race), the social status (literacy, religion), and the economic status (occupation, earnings) of individual citizens. Because of their arrangement and the concentrated form of their information, census schedules also meet the test of form.

There are numerous other groups of records in the National Archives that contain personal data on individuals. Some of these are valuable because of the individual information they contain. Examples are homestead applications, passport applications, pension applications, passenger lists, old personnel records (both military and civilian), and immigration and naturalization records. Others, although again containing information on individuals, are not valued for that reason, but rather because they deal with a class of persons. Examples are the case files on farmers participating in the Agricultural Adjustment Administration and the rural rehabilitation programs during the economic depression of the 1930's, southern sharecroppers, migrant workers, transient Mexican and Puerto Rican laborers, Indians, criminals, and others. Here records have value because of their information on a class of persons, not because of their information on specific persons. On such classes the information is not exclusively or even primarily of a personal nature; it may relate to economic, geographic, or other phenomena. This brings us away from personal values to values for studies of phenomena; these will be discussed in later paragraphs in which records on various types of phenomena will be considered.

In selecting records for the information they contain about persons, two alternative courses are possible. The first is to select those that represent concentrations of information, such as census schedules, in which single documents provide extensive, intensive, or diversified information in a concentrated form. The second is to select a limited number of documents or case folders that are representative or illustrative of the whole, or that are adequate to throw light on the phenomena under investigation.

In the latter of the alternatives, namely that of selecting for retention a limited number of case files on individuals, two principles may be followed: (1) that of special selection, and (2) that of statistical sampling. The principle of special selection may be illustrated by the retention in the National Archives of a limited number of personnel folders for civilian Federal employees. For the early years such folders are replete with documents of an informative nature and are being retained. For the later years only those for key employees who served the Government in an administrative, executive, or supervisory capacity are selected for retention. Here the persons are individually important so that a selection is made in relation to individuals rather than to matters of a social nature. The principle of special selection may also be applied to obtain a documentation of social or other phenomena. The principle of statistical sampling is applied only when records are being selected for studies of collective, not individual phenomena. Such applications will be discussed in later paragraphs in which records on various types of phenomena will be considered.

Before closing the discussion of records containing information about persons, a little more attention should be given to the matter of the personal uses of such records. The uses that are to be considered here are those that relate to financial, legal, or civil rights of individuals. To what extent, in a word, is an archival institution obligated to preserve records for purely personal uses?

Public records are the ultimate proof of all permanent civic rights and privileges; and the immediate proof of all temporary property and financial rights that are derived from or are connected with the citizen's relations to the Government. Certain of the property and financial rights are of long duration; others are of a passing nature.

Among the most important records relating to persons are those that establish the facts of their existence, identity, and marital status. These facts are essential in establishing a whole host of collateral rights, such as the rights to property, to the privileges of citizenship, and to social benefits of various kinds. The National Archives is preserving a number of large groups of records containing the vital facts about persons, and has compiled a list of them for those who seek information concerning age and citizenship. The census schedules, which have been considered at some length, are the most important group of such records, and are used extensively to establish facts about persons that are ordinarily derived from vital records when such are available. The Census Bureau itself has established an organizational unit devoted exclusively to providing such vital data from the schedules of the censuses of 1880, 1900, and 1920. Its services are similar to those performed by registries of vital records.

In every advanced society, the state has provided for the maintenance of vital records of births, marriages, and deaths. The history of their maintenance is a long one. The formal registration of information about births, marriages, and deaths in the English-speaking world began in 1538 when Henry VIII required that the incumbents of parish churches throughout England enter in books a record of each baptism, marriage, and burial that occurred in the parish. This practice spread to other Christian countries so that by the 18th century legal registration of vital data by Protestant and Catholic officials was widespread. In 1789, during the French Revolution, the responsibility for handling French registrations was transferred from church officials to town halls throughout the country. In the next century other European countries followed the practice of France in making such registrations a state rather than a church responsibility. In England a registration law was enacted in 1836 that created a central register officer with responsibility for the records and statistics of births, marriages, and deaths -- by cause -- for all England and Wales. This act of 1836 was the prototype of registration laws for the British colonies, including those of Australia, and for certain of the American States, notably Massachusetts, which enacted the first registration law in America in 1842. In the middle of the 19th century a number of American States passed laws requiring that public records of births, marriages, and deaths be made and that copies of such records be sent to a central bureau of vital statistics in the capital city of the State. New Jersey began the practice in 1848, and Rhode Island and Virginia in 1853. Largely through the agitation of the American Public Health Association, founded in 1872, various other States adopted registration systems, so that by 1919 every State had a central registry of vital statistics. ( Footnote 8).

Vital records pertaining to births, deaths, and marriages should be and are being permanently maintained by the respective States. The Federal Government in the future will therefore be relieved of any necessity to keep large bodies of records because they contain incidental information on births, marriages, and deaths, as it now does for earlier periods in our history.

Another important class of records relating to persons is that which establishes facts regarding property. Most such records relate to property rights of a purely temporary nature, such as arise out of contracts with the Government, loan agreements, and the like. These have a value only for the duration of the commitments between the Government and the persons involved. There are, however, certain property rights which, as the German archivist Meissner has pointed out, relate to substantial matters, such as titles to real property that was once owned by the state. In the National Archives this class of record is best exemplified by the records of the General Land Office that relate to the transfer to private persons of title to land on the public domain.

Another important class of records relating to persons is that which establishes facts regarding their service to the Government either in a military or a civilian capacity. These facts are also essential in establishing a number of collateral rights, such as rights to pensions and other benefits. The personnel records of Federal civil servants have been found to be very sketchy in recent years. They contain only the information necessary to establish the employee's rights, and they are therefore being retained (in the Federal Records Center in St. Louis, Mo.) only for the duration of such rights.

There are innumerable other classes of records that are important to persons in support of their "rights." The list is endless. They arise every time an individual has any sort of dealing with his Government. The extent to which, the duration for which, and the place at which the Government should preserve such records are matters of public policy. For records that relate to purely temporary relations between the citizen and his Government, the conclusions of Bauer in his paper are valid, viz. : first, that "an agency established to protect or regulate certain private interests ought, of course, to maintain appropriate records and preserve them as long as the interests primarily affected by them subsist," and secondly, that "a fair working principle for fixing the retention period of such records would be to consider them only in relation to those interests that fall within the jurisdiction of the agency creating or accumulating them and not in relation to all the limitless rights and interests that could be defended by their collateral use."

Besides the records that deal with persons individually there are numerous groups of records in the National Archives that contain data on corporate bodies. Such records are usually in the form of case files that pertain to the Government s relations with particular corporate bodies or in the form of returns (or reports) that are furnished to the Government by corporate bodies of a particular kind. Among the case files are those pertaining to cases of bankruptcy, equity, and law before district and claims courts; to labor disputes before various labor boards to the manufacture and marketing of foods and drugs; and to the regulation of interstate commerce, trade, transportation, and communications. Among the returns or reports are those submitted to the Bureau of Mines on sales, production, employment, and the like, by the mineral industries to the Commodity Exchange Administration on trading at the Chicago Board of Trade; to the Securities and Exchange Commission on corporate bodies issuing securities; to the Federal Trade Commission; and to other regulatory and investigative agencies of the Federal Government. Usually such returns or reports are submitted to the Government in compliance with regulations or under subpoena powers provided by statute, and their use is restricted for relatively long periods of time. They are preserved for the information they contain about business and financial conditions generally, not for their information on particular firms.

There are exceptional instances, however, in which they are kept for a study of particular firms. This is the case with the records of the Chesapeake and Ohio Canal Company covering the years 1785 to 1939, which were acquired by the Government and which are obviously valuable for a study of business history, as well as for a study of internal improvements. If records are to be preserved on particular firms, the criteria for their selection suggested by the British archivists are as good as any. In their pamphlet on the principles of elimination the British archivists suggested, it will be recalled, that records should be preserved for firms that belong to a category for which records "have rarely been preserved," or that are of "outstanding importance" in comparison with others in the same category, or that belong to a category the general history and development of which "can only be traced by the use of collective evidence." Like most records on individuals, however, records on corporate bodies are preserved mainly for their collective significance; not for their value in studying the history of individual firms. In this respect they have value for studies of various economic and social phenomena, and will therefore be discussed in later paragraphs.

Records Relating to Things

The term "things," it will be recalled, was defined to include places, buildings, and other material objects. In discussing records on things the values to be considered are those that derive from the information they contain on the things themselves, not from the information on what happens to things.

Among the most fundamental things with which human beings are concerned is the land on which they live. The National Archives preserves many series of records relating to land: records on its mineral resources, produced by the Bureau of Mines; on the classification of its soils, produced by the Bureau of Soils; on its survey and exploration, produced by the Geological Survey; on the ownership of lands that were once part of the public domain, produced by the General Land Office; and on various other of its topographical, geological, and geographical features.

The records pertaining to lands that were once part of the public domain will serve to illustrate most of the problems of appraisal that arise in relation to records on things. These are the land-entry papers of the General Land Office, of which there are almost 19,000 cubic feet in the National Archives, and among which are many applications for homestead lands during the years 1862 to 1950. The land-entry papers contain descriptions of the land by subdivision, section, township, and range. Since the title to the land is based upon the documents transferring it from the public domain into private hands, the records are primarily retained for the evidence they contain of the legal or property rights of individuals who now have possession of the land. The records will have to be retained for this purpose so long as the present system of recording title to real property exists. If the Torrens system of land title registration were in use, the retention of the whole chain of records on conveyancing transactions going back to the original land-entry papers would be unnecessary. (Footnote 9). The land-entry papers of each applicant for a homestead on the public domain, as has been noted before, also contain personal information, such as his age, place of birth, and, where appropriate, information about his naturalization. They are, therefore, used quite extensively for genealogical purposes. While the land-entry papers, as a whole, can be used for studies of the settlement of the West, and for study of the alienation of public lands, they are seldom so used. Their information on the character of the lands themselves is insignificant. The papers, considering their volume and arrangement, hardly meet the test of form and barely meet the test of importance, but the primary values that still inhere in them are such that under the present conveyancing system of land titles no Government officer would venture to recommend their disposal.

Records in the National Archives relating more broadly to the land of this country include military and nonmilitary geographical explorations and surveys such as those of Lewis and Clark in 1803-6; geological surveys such as those of Hayden (1867-79), King (1867-80), Powell (1869-79), and Wheeler (1869-79); surveys of the public domain by the General Land Office; and boundary and railroad surveys. These records contain geographical, topographical, geological, botanical, and ethnographical information. This information is important; the sources in which it is presented are unique; and while the sources are widely scattered and, on occasion, have been removed from public custody, they nonetheless are quite workable and would be more so if they could be brought together.

Among other things on which records are being preserved in the National Archives are manmade things -- things that, by and large, are impermanent and the records of which, therefore, are less likely to have enduring value. Among such records are those relating to the internal improvements of the nation, such as the records of the Chesapeake and Ohio Canal Company which have already been mentioned, of railroads on which records are found in a number of record groups, and of buildings. The records on buildings may serve to illustrate problems of appraising records relating to artificial things. It is obvious that records need not be kept on most buildings, whether private or public, that they need not even be kept, for instance, on their architectural or structural details, for printed information is available on such matters. Records on buildings are archivally important only if the buildings themselves are important; and buildings acquire an importance because of the associations with them, because they are identified with important historical personages or important historical events, or because they are outstanding examples of period buildings. The homes of our Presidents -- Mount Vernon, Monticello, The Hermitage and the buildings in which important historical events occurred -- Independence Hall, the White House, the Capitol -- these are important for their associations and practically all records pertaining to them are therefore important. In evaluating records on such structures the dictum for records of the last century should be to "keep everything." For records of quite recent origin, however, everything obviously cannot be kept even for the most important places; for many of the records are likely to relate to very trifling housekeeping matters.

The interest of the National Park Service of the Interior Department in records on truly historical places is an important interest. Every scrap of information on such places may be important to Park Service historians and should, therefore, be preserved for them. In general, the observation of the German archivist Meissner in regard to records on buildings is valid; namely, that you should keep files relating to real property if they establish the rights of the state to such property or if they relate to the administration of property that is of special or historical interest.

Another class of records relating to manmade objects, of which numerous examples are found in the holdings of the National Archives, consists of records relating to ships. Very large series of such records exist, such as plans, including tracings, drawings, blueprints, and the like, of naval vessels among the records of the Navy Department, and of commercial vessels among the records of the Commerce Department. There are both antiquarian and scholarly interests in the early vessels constructed in this country that are reflected quite well in The American Neptune, a quarterly periodical devoted to various aspects of marine research. It is obvious from a perusal of its pages that records containing information on the design, construction, and operation of various types of vessels at different periods are of real interest to a large group of persons. But it is doubtful if the same research interest attaches to most records of new vessels as attaches to practically all records of old ones, and if, therefore, an archivist is justified in keeping more than a few selected classes of records on relatively new vessels.

Another type of record on manmade objects found in the National Archives is that relating to the granting of patents by the United States Government. This group of records illustrates why certain records have value for the information they contain on objects rather than because they reflect the administrative processes of government, though, of course, patent files may also serve to do the latter. The patent granted to Galileo by the Doge of Venice in 1594 for inventing "a machine for raising water and irrigating land," ( Footnote 10) for example, throws interesting light on the patent system existing at that time as well as on the technological developments of the period. The monopolies and patents granted by the American Colonies before the establishment of the Federal Government shed light on the industrial life of the colonial period. The first patent granted in America for machinery pertained naturally to agricultural equipment. It was given by Massachusetts in 1646 to Joseph Jenks and related to a mill for making scythes and "diverse sorts of edge tools." (Footnote 11). Other rights of manufacturing during the colonial period pertain to making salt, potash, pitch, molasses, sperm candles, linseed oil, duck canvas, paper, and nails. The patents granted by Thomas Jefferson while Secretary of State include those to John Fitch for his invention of the steamboat and to Eli Whitney for his invention of the cotton gin; they also throw incidental light on Jefferson's part in establishing the patent system. The early patent files, with their applications and related plans, drawings, and sketches, are important for the information they contain on the patents themselves and for the information they contain on the technological development of the country.

But while the earlier patent "case" files have an undoubted research interest, this is less certain for the more recent ones. As the country has developed technologically, the patents relating to its industrial and mechanical processes and devices have progressively become more specialized. The recent files, particularly those since 1900, usually relate to small parts of highly complicated processes or machines and seldom to an entirely new mechanical device that has had, or may have, a major impact on the economic life of the country. Thus they are less significant individually than the older files. And the information they contain is also available, to a greater extent in recent years, in published documentary sources. The printed patents themselves contain most of the information that the Government is justified in keeping to show the development of technology, from a patent point of view. Only a very limited number of individual patent files relating to the most significant technological developments appear to be worth preserving for the period after 1900.

Records Relating to Phenomena

The term "phenomena," it will be recalled, refers in the present context to what happens to either persons or things -- to conditions, activities, programs, events, episodes, and the like. The phenomena recorded in public records are of interest chiefly to social scientists, but some of them may be of interest to natural scientists. If the phenomena are old, they are of chief concern to the historian; if new, to the sociologist, the economist, or the student of government.

Since most records that come into the care of an archivist are relatively old, the interests of historical research are most important to him. An archivist, no matter what his training, will ordinarily appraise records primarily on the basis of their historical value or interest. This is the basis on which Armand Gaston Camus (1740-1804) and Pierre Claude François Daunou (1761-1840), the first heads of the Archives Nationales, appraised the prerevolutionary records of France.

Modern archivists are generally trained as historians, and it may therefore be assumed that they are competent to appraise the value of public records for historical research. Most archivists are likely to preserve all records that relate significantly to important personages, episodes, or events. No American archivist, for instance, would knowingly destroy anything of value relating to an episode like the Whisky Rebellion, an event like the Louisiana Purchase, or a personage like Abraham Lincoln. And if an archivist's knowledge of history is extensive, he is likely to preserve records relating to personages and episodes whose influence on the course of events, though less widely known, was considerable. Most archivists are likely to keep the basic source materials for studies in diplomatic, political, and military history, which were once the chief concern of historians. The National Archives, for example, keeps the official despatches, reports, and instructions of the State Department that are needed for a study of foreign affairs; the committee files, reports, and journals of the House and Senate needed for a study of political affairs; and the various series needed for a study of the conduct of war, produced by the War and Navy Departments. If a full picture is to be obtained of diplomatic, political, and military affairs, however, these basic sources must be supplemented by many other record series of a specialized nature. The series on diplomatic matters, for example, must be supplemented by records pertaining to economic matters, particularly records produced by Government agencies that are concerned with international trade, as well as by records pertaining to public opinion, such as press releases, broadcast scripts, and films and recordings.

The appraisal of records from the point of view of their historical interest becomes difficult when the records relate to broad historical movements, historical causation, and the like. Here a discriminating choice may have to be made among the records that are available. A movement like the westward expansion of the United States, for example, can be traced in a number of record groups in the National Archives, including those for the Bureau of Indian Affairs, the Bureau of Land Management, and various other Government bureaus.

When records relate to recent social or economic matters, a greater degree of specialized knowledge is required for their appraisal than is ordinarily possessed by historians. Here the knowledge of economists, sociologists, and scholars in other disciplines comes into play. Recent public records that are of interest to such scholars arise especially from the regulatory and social welfare activities of modern governments. They may be of real significance for studies of various aspects of modern society. They may be used, for example, to study the consequences of public welfare activities -- what happened to private economic organizations under Government regulation or the rural and urban patterns that are developing in the country, social trends, and the like.

As one goes backward in time information on social and economic matters becomes less complete. Records on business are almost as scanty for the 19th century as they are full for the 20th century; and almost all of them that are still extant for the earlier period should therefore be preserved. Generally public records relating to social and economic matters that are earlier than the First World War should be carefully compared with other documentary sources to determine if they contain unique information.

Present-day documentation of social and economic matters, however, is very voluminous. The publications of the Federal Government alone (which has in fact become the world's largest publisher) provide a wealth of information on such matters. On the economy of the country, in particular, an ever-increasing range and volume of information is available in published form. This relates, among other things, to the Nation's agricultural and industrial production, trade, consumption, unemployment, financial condition, prices, income, and living costs. Among the agencies of the Government that issue in print statistical data and information on economic conditions are the Bureau of the Census, the Bureau of Foreign and Domestic Commerce, the Bureau of Labor Statistics, the Office of Business Economics, the Bureau of Agricultural Economics, the United States Tariff Commission, the Federal Trade Commission, and the Board of Governors of the Federal Reserve System.

In appraising records on social and economic matters, the archivist must rigidly apply the test of uniqueness. In the National Archives and Records Service this test was recently applied to a large series of tax returns submitted by certain kinds of corporations to the Internal Revenue Service. ( Footnote 12). The returns, comprising about 100,000 cubic feet for the period from 1909 to date, admittedly contain information which, if not to be found elsewhere, would be useful for certain kinds of research. The analysis showed, however, that substantially similar information was available elsewhere on many, though not on all the corporations submitting returns. The analyst concluded that commercially published sources, and other official sources, were better than the series in question for a study of the business economy generally, and also for a study of any important corporation particularly.

By and large, the scholar can usually rely on the overwhelming mass of published literature for information on recent day-to-day social and economic developments in this country. Published sources generally provide adequate information on them. The original public records on them are far too voluminous to be preserved in extenso; and it is mainly in regard to the abnormal or the unusual that an archivist should preserve such records. If he preserves records on normal contemporary social and economic matters at all, he should preserve them in summary form or in exemplary selection.

Several large record series are being preserved in the National Archives because they contain information on unusual or abnormal economic or social conditions. These are illustrated by the transcripts of hearings of the National Recovery Administration which reflect the condition of industry during the 1930 economic depression, and by the price and accounting records of the Office of Price Administration which reflect the condition of industry under the controlled economy of the Second World War.

The principle of special selection should be applied to more recent records on social and economic matters. This principle simply means that a few records are selected for preservation because they contain data that are representative or illustrative of the whole, because they deal with an important or significant event or action, or because they contain data that are considered adequate for a study of particular social or economic conditions. It is well to distinguish this principle at once from the principle of statistical sampling. The latter, which was developed early in the present century, requires a knowledge of method that is not ordinarily possessed by the archivist. The techniques of collecting, classifying, and analyzing statistics, of correlating data, computing averages and probabilities, making forecasts, plotting curves, and compiling index numbers are highly specialized techniques that are part of a distinct discipline. And statistical sampling techniques, even if known to the archivist, cannot ordinarily be applied to the selection of records.

The archivist preserves records for unknown uses; the statistician must know in advance the particular ways in which his samples are to be used. The archivist selects records that have characteristics illustrative of the whole; the statistician, in accordance with well-defined mathematical formulae, selects a sample that presents information of measurable reliability on particular characteristics of the universe from which it is taken. A statistical sample is more exact than the representative or illustrative body of records preserved by the archivist.

Criteria based on and closely resembling statistical methods were applied in selecting records on the rehabilitation loan program of the Farm Security Administration, an agricultural agency of the last economic depression in the United States. This procedure has been described by Dr. Carl J. Kulsrud in an article in The American Archivist for October 1947, entitled "Sampling Rural Rehabilitation Records." In granting rehabilitation loans to relief clients, the agency developed for each such client case files containing reports, correspondence, and other papers. These case files are rich in information on the social, economic, and human factors that led to the rehabilitation loan program. They are useful, therefore, for social studies, and studies of the economic conditions in the depression period as well as for a study and evaluation of the procedures, ideologies, and techniques followed in the program. Since the files were very voluminous, a sampling was made of them that saved only 3 percent of the total. The sample consisted of all case files for typical counties in 134 distinct farming areas as classified by the Bureau of Agricultural Economics of the Department of Agriculture.

The principle of special selection is illustrated more typically in the action taken by the National Archives in preserving various kinds of labor board case files. In selecting files of the National Labor Relations Board for retention, for example, the importance of the individual cases was established in reaction to the following standards (1) the issues involved in the case; (2) the influence of the case in the development of principles, precedents, or standards of judgment; (3) the contribution of the case to the development of methods and procedure; (4) the intensity of public interest in the case (5) the effect of the case on the national or local economy or on the industry and (6) the strikes, lockouts, etc., attendant upon the case.

Records that contain concentrations of social and economic data which may be statistically exploited are similar in character to those that contain summaries of personal data. The schedules produced by censuses of industry and agriculture, outwardly at least, are similar to the population census schedules. Schedules of business and agricultural censuses, however, do not possess the same value as schedules of population censuses, mainly because the information they contain is almost always used in the aggregate and not in relation to individual business or agricultural units, and because the aggregates have been tabulated and enumerated satisfactorily.

In appraising records the contents of which can be statistically summarized, such as administrative forms and statistical questionnaires and schedules, the archivist is well-advised to proceed cautiously. If the Government agency that created the records for statistical purposes did not fully exploit them, it is hardly likely that anyone else will; for scholars outside the Government do not ordinarily have the resources for the costly exploitation of such records. If the records were not created for statistical purposes, it is hardly likely that they will yield accurate or meaningful statistics. During the Second World War (on January 7, 1944), a conference was held with a group of business and technical experts on the possible uses of the Office of Price Administration rationing applications for gasoline, tires, and automobiles. After exhaustive discussion, the experts agreed that the applications need not be saved for the purpose of compiling any national statistics from them. "The arguments used," according to Dr. W. J. Wilson, "would seem to have covered the major principles governing the evaluation of records for statistical purposes." (Footnote 13). They were:

1. Masses of raw statistical data need not be preserved after the statistical information has been satisfactorily extracted.

2. Masses of unusable data need not be retained longer than is necessary to determine their irremediably faulty character.

3. Masses of usable data will seldom be used at all if not used fairly promptly.

4. Masses of usable data should not be retained for indefinite periods of time on the mere chance that they may one day be employed.

5. All these considerations apply still more cogently to data assembled on applications, registrations, and other administrative forms than they do to data assembled on regular questionnaires.

Generally, then, the archival institution should preserve only summary information -- not the great mass of schedules and questionnaires on which the summaries are based.

While records of interest to the social scientist relate primarily to phenomena involving persons, those of interest to the natural scientist relate largely to phenomena involving material things.

Scientific records present special problems of evaluation to the archivist. These arise mainly with respect to records needed for further scientific research, not with respect to records pertaining to the history of scientific activities in the Federal Government, which are clearly suitable for archival preservation.

Scientific records may be in the form of raw data resulting from observing and measuring various phenomena or in the form of tabulations and summaries of such data. The archivist normally prefers to keep only tabulations and summaries. In the case of scientific investigations, however, the raw, or original, records may also have value; for much of the essential detail in such records may be lost in the course of their tabulation and summarization. Tabulations usually present only averages, and summarizations only the most important characteristics of each type of measurement.

The virtue of the raw original data depends on the nature of the phenomena that were observed and measured and on the degree to which the observations and measurements can be exploited by others than those that made them. An archivist is perhaps justified in keeping data that are derived from measurements of basic phenomena, such as those of the earth, the oceans, and the atmosphere. In the National Archives records containing observations of the earth are best exemplified by the reports of the Coast and Geodetic Survey, 1832-1942. These reports contain data derived from astronomical, magnetic, seismographic, gravity, and other kinds of observations. In many cases they constitute the only authentic source from which can be deduced natural or artificial changes in the physical condition of the area surveyed. Records containing observations of the oceans include the reports of various surveys made by the Hydrographic Office, as well as the logbooks collected by Matthew Fontaine Maury (1806-73) to compile his wind and ocean current charts. Records containing observations on the atmosphere are represented by the climatological and meteorological records received from the Weather Bureau, which were returned by the National Archives to the Bureau after it had established proper facilities for their care and exploitation, though the National Archives retained a film copy of the records created before 1890.

In general, scientific records, just as any other type of records, should have values beyond the temporary ones that resulted in their production if they are to be preserved in an archival institution. This is usually not the case when raw scientific data relate to the measurements and observations made in controlled laboratory experiments, which can be repeated. Records of chemical and biological laboratory experiments are thus not likely to be worth keeping in an archival institution.

Scientific records in their raw form may also present difficulties to the archivist because of their form. They are usually quite voluminous. Often they have attributes that make their further use impracticable. They may be intelligible only to the persons who recorded the data. Like punchcards produced in statistical work, they may be in a form that is difficult to interpret without resort to mechanical or electronic devices.

They may be in the form of recordings made by instruments on tape or film or photographic plates, or charts, or cards. And these forms present special problems of storage as well as of use.

See Also: