This volume is the third in the Historical Statistics series issued by the Bureau of the Census as a supplement to the annual Statistical Abstract of the United States.

Statistics are a valuable adjunct to historical analysis. They often clarify and enrich qualitative history and on occasion become important parts of a historical record on their own. However, users of historical data are faced with the paradox of over-abundance and scarcity. A burdensome multiplicity of sources has frequently to be consulted in order to reconstruct one quantitative aspect of a particular subject. Just as often, users are confronted by a discouraging barrenness of data, discoverable only after much costly work and delay.

The objective of the Historical Statistics volumes is to provide a convenient reference source which has two functions, collecting and referring. The collecting function consists of assembling, selecting, and arranging data from hundreds of sources and making them available within a single source. The referring function consists of text annotations to the data which act as a guide to sources of greater detail. The annotations also define terms used in the tables and include essential qualifying statements.

The first volume in this series, Historical Statistics of the United States, 1789-1945, was published in 1949. It provided a wide range of series quantifying various aspects of the development of the Nation. An interim Continuation to 1952 was issued in 1954 to provide data for 1946 to 1952 for the still-active series shown in the first volume. Limited resources confined the scope of the first volume to data most readily available, usually from governmental agency sources. theless, some 3,000 statistical time series were presented.

Never

Historical Statistics of the United States, Colonial Times to 1957, issued in 1960, represented a substantial expansion of the data shown in the original volume. It presented more than 8,000 time series, mostly annual, on a greater variety of subjects and for longer time periods. The statistics were also more fully annotated and more precise references to original sources were provided. For a greater number of series, in addition, there were more detailed descriptions of the development and reliability of the data. A Continuation to 1962 and Revisions was issued in 1965, presenting revisions of data in the basic volume and extensions to 1962 of the more than 6,000 series still current at that time.

Each of the first two volumes was prepared with the cooperation of the Social Science Research Council, the guidance of a distinguished Advisory Committee, and the assistance of numerous scholars, research analysts, and particular subject specialists. A description tracing the development of the first two editions appears below under "Origin of Historical Statistics of the United States."

During the latter 1960's, the supply of copies of Historical Statistics... to 1957 available for sale from the U.S. Superintendent of Documents was exhausted. The edition had already been through a cycle of five printings and a question was raised concerning the advisability of further printings in the light of a possible new edition. The question was timely. Experience with the first two editions and their Continuation supplements had shown that a new edition was desirable at 10 to 12 year intervals. The Continuation supplements were at best handy stopgaps for researchers, a serviceable minimum seriously lacking in documentation. As each year lengthened the interval between editions, the "convenience" value of both the Continuation to 1962 and its parent Historical Statistics... to 1957 diminished. More and more time series were revised in part or entirely replaced. Further, the task for the user of updating the still active, unrevised, series became more burdensome despite the special efforts of the an

nual Statistical Abstract to maintain a direct linkage to as many historical series as possible in its current tables. As a result, a decision was made in 1969 to begin preparation of a new edition.

The plans for the new edition immediately encountered the problem of funding and resources. It was clearly impractical at that time, given the available resources, to consider undertaking a fullfledged new edition of Historical Statistics. The determination to make a start, however, was very strong and more modest objectives were adopted. In effect, the early plans for the present edition proposed that it comprise little more than: (1) An extension to 1970 of those series for which current data were available; (2) revisions of data which had occurred since issuance of the Continuation to 1962; and (3) a reprinting of those series in Historical Statistics... to 1957 which had not been affected by either updating or revisions. No time span was specifically set down to complete the work because there was a clear understanding that it was a part-time staff project. Two other aspects of this plan differed considerably from the procedures followed for the last edition. For that edition, a large number of consultants were enlisted for their expertise in assembling and developing new time series, reviewing and adjusting old time series, and providing explanatory and bibliographic notes for both. Although most of these consultants, especially those in Federal agencies, contributed their own and their agencies services without compensation, many were compensated from funds provided by the Ford Foundation (by arrangement through the Social Science Research Council). For the new edition, given the limited resources, consultants' contributions were recruited on a public service basis entirely. Partly for the same reason, it was decided not to revive the collaboration of the Census Bureau with the Social Science Research Council which had proved so highly effective for the first two Historical Statistics editions. Even more convincing for the Bureau decision to undertake the project alone was the solidity of the base which those editions now provided for the next edition. Seeking such collaboration again seemed unwarranted in the light of the modest objectives outlined above.

As the work slowly progressed and as the many consultants and contributors gave generously of their knowledge and talent, it became clear that our objectives were too restrictive; that our contemplated mere updating would, if adhered to, have to ignore a large accumulation of new time series which were either ineligible for the last edition (at that time they covered a period of less than 20 years) or had not been discovered or properly developed prior to that edition. The gradual accretion of new material plus the additions to old material substantially changed the planned scope of the present edition. What follows are some measures of the changes in content introduced in the present edition.

All of the broad subject fields shown as separate chapters in the last edition are included in this edition and follow the same sequence. Within some of the chapters, however, chapter segments have been regrouped into new subchapters (as in chapters K and X) and in others, the sequence of the subchapters has been changed (as in chapters H, Q, and U) to achieve minor improvements in the juxtaposition of subjects.

In two chapters, two entirely new subchapters have been added: "Input-Output Structure of the U.S. Economy" to chapter F and "Flow of Funds" to chapter X.

The present edition presents more than 12,500 time series, a 50percent increase over the last edition. Every chapter has undergone some expansion with respect to new time series. Chapter F, national income and wealth, and chapter H, social statistics, doubled in num

ber of series; the former from 345 to 723 and the latter from 543 to 1,170. The increase in chapter F was largely due to newly-added data for economic growth rates, greater detail than was previously shown for national and personal income, and data showing valuation of capital stocks. Unsurprisingly, the largest increase in series occurred in chapter H where the data for social insurance and welfare, education, and crime and correction reflect the great public attention given to these subjects in recent decades. Almost equally large increases took place for chapter K, agriculture, and chapter X, financial markets and institutions (formerly banking and finance); chapter K from 328 series to 623; chapter X from 480 to 962. Partly to accommodate the increase in series, chapter K has been subdivided into 4 parts. Most of the new series in chapter K relate to farm population and farm-operator characteristics, farm marketings, government payments and price supports, and a number of new measures of farm productivity. For chapter X, the bulk of the increase in series is in the new flow-of-funds subchapter.

Several chapters now include for the first time a number of data series below the national level. In all, there are 13 new tables (comprising 484 series) in this category, 9 of which present data for the individual States and 4 for either regions (e.g. the South or the West) or the smaller geographic divisions (e.g. New England, South Atlantic). Perhaps of special interest among these tables are the series on population characteristics and land area for each State (A 195–263), those on selected items for farms and farm population by State (K 17-81), those on voter participation in presidential elections by State (Y 27-78), and those on population censuses taken in the colonies and States during the colonial and pre-Federal period (Z 24–132).

In addition, each of 4 chapters (D, G, Q, and Z) includes at least 200 or more new series and each of 10 chapters (A, B, L, N, P, S, T, U, V, and Y) includes 100 or more. A summary of selected new series included in each chapter is shown on p. XV.

One other important change is the reinstatement of a time period index (see p. A-4) which first appeared in Historical Statistics..., 1789 to 1945. The index enables users to identify quickly which time series (or statistics for particular subjects) begin in the specified 10- or 20-year time segment (e.g. 1800-1819, 1820-1839).

As a result of the complete review and updating of the contents of the last edition of Historical Statistics, many changes, apart from the entirely new series, have occurred in both the tables, the descriptive text, and the bibliographic notes. Most of the changes are due to revisions and corrections made during the interval between the last and present editions by the sources of the data affected. Where users of both editions become aware of discrepancies in what purport to be identical sets of data, it is safe to assume that the figures, descriptive text, and notes in the present edition supersede those in the last edition.

With rare exception, all of the series shown in the last edition are also included here. 76 series were omitted. They were primarily discontinued series replaced on recommendations of consultants by other series of a similar kind or were considered of marginal importance or relatively weak in other respects. In one or two instances, space was also a factor.

Origin of Historical Statistics of the United States

The first edition, Historical Statistics of the United States, 17891945, was formally initiated by a recommendation in 1945 by the Social Science Research Council that the Secretary of Commerce consider compilation and publication by the Bureau of the Census of a source book of economic statistics.

Earlier the same year, J. Frederic Dewhurst urged the development of an historical source book in a proposal to the American Statistical Association and the American Economic Association. A joint committee was named by these associations, joined by the Economic History Association, to explore the practical problems of preparing such a volume. Dr. Dewhurst's proposal coincided closely with

Bureau of the Census plans to prepare an historical supplement to the Statistical Abstract of the United States. The formal decision in 1945 by the Bureau of the Census to compile and publish such a volume led to the reconstitution of the joint committee, which then became the Social Science Research Council Committee on the Source Book of Historical Statistics, Advisory to the Bureau of the Census.

After the first edition was issued in June 1949, the Economic History Association, in response to a request from the Bureau of the Census, appointed an advisory committee in September 1950 to evaluate the volume and to make specific recommendations affecting the question of its revision. This committee, formally designated as the Committee of the Economic History Association on the Revision of Historical Statistics of the United States, 1789-1945, was under the chairmanship of G. Heberton Evans, Jr., The Johns Hopkins University, and included the following as members: Arthur H. Cole, Harvard University; Shepard Clough, Columbia University; T. C. Cochran, University of Pennsylvania, and Solomon Fabricant, National Bureau of Economic Research, Inc. In April 1952 the committee submitted a report to the Bureau of the Census entitled "On the Revision of Historical Statistics of the United States, 1789–1945." The conclusions and comments presented in this report were subsequently influential in getting underway the project for a revised volume.

For the second edition, Historical Statistics of the United States, Colonial Times to 1957, the Bureau designated a project director who also acted as secretary of the Committee on Historical Statistics appointed by the Social Science Research Council to serve as an advisory group similar to the committee which participated in the preparation of the first edition. The Census Bureau again assumed the responsibility for publishing the volume as a part of its Statistical Abstract program. The Social Science Research Council, in turn,

obtained a grant from the Ford Foundation which provided funds for the procurement of services of experts in each field. More than 125 such specialists were engaged to serve as consultants. The Council also made arrangements with some of the consultants for the preparation of bibliographic essays on statistics in selected fields, five of which were subsequently published in the Journal of the American Statistical Association.

The Problem of Historical Statistics

The scattered sources of historical statistics of the United States include the annual reports of the executive heads of the agencies of the Federal Government, reports of special Federal commissions, the U.S. census volumes, printed debates of the Congress, published reports of committees of the Congress and transcripts of hearings on important legislative measures, published reports and documents of the State governments, statistical publications of private research organizations and of the universities and colleges of the Nation, together with the great mass of statistical volumes printed by other private organizations and individuals.

It has been noted that on occasion compilers, desiring to save the time and effort required to obtain data directly from the original sources, make use of successive issues of the annual Statistical Abstract of the United States to construct long-term time series. The results of such a procedure are not always sound, since the space available in the Statistical Abstract for describing major revisions in time series may not permit adequate clarification. Of the many revised figures appearing in each issue, most revisions apply to the immediate preceding years, but revisions of much earlier years are not uncommon. Moreover, the revisions shown have followed no systematic pattern and may be scattered irregularly over many issues.

Impediments to the use of historical statistics, then, include the initial difficulty of determining whether the data in fact exist, of identifying the document in which the data may be found, of constructing time series where the data may not be arranged in suitable form, and of identifying and interpreting changes in concept and

coverage. Definitions employed in published historical tables, moreover, may have to be sought in separate publications if, indeed, they have been published at all.

Technical Notes and Explanations

Arrangement of the data. Data are arranged for broad subjects in lettered chapters and for more specific and detailed subjects in numbered series within each chapter. To facilitate reference, subject groups are organized in summary form under chapter and subchapter titles in the table of contents (p. IV). In addition, there is a detailed alphabetical subject index (p. A-10). The data are presented in conventional tabular form, each table comprising a group of subjectrelated series. Each series or tabular column is assigned a unique letter and number. The letter prefix identifies the chapter and the number represents the order of the series in the chapter. Thus the 44th series in the chapter on agriculture is designated K 44 to distinguish it from the 44th series in the chapter on transportation, Q 44. Because of possible confusion with numerals, the capital letters I and O have been omitted in identifying chapters. Source citations and descriptive text material (see below) are linked to the data series by use of the assigned series numbers.

All series begin with the most recent year for which data have been obtained and run backward in time. This arrangement was selected because it lent itself to more compact, less space-consuming presentation than the alternative of beginning with the earliest year. Insofar as possible, there are uniformly placed spaces above every year ending in 0 or 5. No data are shown for years subsequent to 1970. Figures for later years for most of the current series are presented in the Statistical Abstract of the United States beginning with the 1973 edition.

Basic guidelines. The guidelines adopted for this edition to aid Census Bureau staff members, subject matter consultants, and other participants with respect to selection and presentation of the data are quite similar to those of the last edition. As was the case then, however, the guidelines were not followed with complete rigidity. At times, the scope, variety, and complexity of the data involved made it necessary to modify the rules for the sake of clarity or internal consistency. The guidelines applied and the elements subject to application are discussed below.

Area coverage. Except as otherwise specified, data generally represent conterminous United States or the 48 States (including the District of Columbia) prior to the admission of Alaska and Hawaii to statehood and the 50 States thereafter. Asterisks on individual tables or series indicate the first year for which the figures include Alaska and Hawaii to the extent that their inclusion could be ascertained. For some series, especially in chapter K, the notes specifying inclusion or exclusion of Alaska and Hawaii appear in the text. In some instances, the sources used for data failed to specify the area covered. Where practicable, the data were examined and the appropriate qualifications were added.

Because of limitations of space, data are not generally shown for regions, States, or localities. Some exceptions were permitted, however, as noted above with respect to data for regions and States. Other exceptions were of a more specialized nature as in the following instances: Where regional statistics are helpful for correct interpretation of data, such as presentation of merchant marine statistics separately for each coast and for inland waters; where data in the subject field cannot (by definition) be summarized effectively for the United States, such as internal migration data; where summary data for a given subarea or market are indicative of general trend or level, such as prices on the New York Stock Exchange or in specified cities; where data for a given area effectively represent the national picture because of concentration of production, etc., as Pennsylvania anthracite; where data are available for only a given area as in the case of many series concerned with early American history and limited to the Atlantic seaboard.

Time coverage. In general, only annual or census-period data which cover at least 20 years are presented. A major exception was made for series covering the colonial or pre-Federal period. Other exceptions were permitted where newly developed series of recent origin were the only data available to represent an important subject field or where a short series was an important extension of other longer series.

The general requirements as to time coverage were specifically designed to permit inclusion of "lapsed" series, particularly those falling within the nineteenth century or extending into the early twentieth century. The lapsed series, which begin and terminate in the past, represent major fields of interest during various phases of American historical development; frequently they must be sought in out-of-print documents which are available in few libraries.

The identification of time-periods was complicated by failure of some sources to state whether the data were prepared on a calendaryear or on a fiscal-year basis; by shifts in time coverage from calendar to fiscal year during the period of the series, and, in some instances, by the lack of identification of the beginning or ending date of the fiscal year. In all such cases, particularly where time shifts seemed likely to have occurred, an effort was made to identify the correct basis. Frequency of data. Annual data are given preference but certain series are presented only for years in which a national census was conducted, and, in some instances (for example, telephone and telegraph rates), only for the scattered dates for which the data are available. Where both annual figures and decennial or quinquennial benchmark or census data exist, both series are frequently shown. Series linkage. No formal attempt was made to extend a single series back through time by linking it to another series which terminated at or near the date on which the first began. In a number of instances, however, such series are presented in adjoining columns, with an overlap for a period of years, when available.

Selection of data. The criteria of selection varied broadly, depending on a number of factors applicable to the subject matter involved. Generally, summary measures or one-dimensional aggregates at gross levels and immediately below were given highest priority for inclusion. Below such levels, selection was governed by the interplay of: The amount of space already devoted to a particular subject; the attempt to achieve a relatively balanced presentation among subject fields; the "uniqueness" (in the sense that other data did or did not fairly cover a particular subject) of the data; the quantity of data available; the quality of the data available; and the extent to which data might be related to and enhance the value of other data.

Among less discretionary factors, both area detail (see above under area coverage) and subject detail, such as cross-classifications or data for specific commodities, were held to a minimum because of space limitations. Inevitably, there were exceptions where synthesis or summarization did serious damage to the value of a series or where it was clearly more meaningful to show series for specific commodities than a group aggregate.

Presentation of absolute rather than derived data. Primary emphasis was placed on the presentation of absolute figures rather than on derived data since the absolute figures offer somewhat greater flexibility to the user. The major exception was the presentation of index numbers. In general, percentage distributions of absolute data already shown are not presented. Other percentage data, and averages, medians, ratios, and rates were used only where they resulted in a significant economy in space or where they significantly facilitated interpretation. No attempt was made to convert various series of index numbers to a base year or period other than that shown in the source. Large numbers (8 digits or more, for example) shown in the source documents have been rounded to thousands, millions, or billions for ease of use and reference only as staff resources allowed. Omissions of data, “blank” cells. The significance of dashes in tabular cells varies from series to series. In general, the presence of cell "leaders" or "dashes" indicates merely that no information was provided. Dash entries may mean that no information exists for the given year; the entry, if shown, would be zero; the information

was not available; or the information is believed to exist in published form but it was not practicable to do the research necessary to locate the appropriate source. The user will have to judge from the context which meaning is appropriate in each particular instance.

The practices of the various sources of information differ as to the meaning of dashes in cells, the extent to which they label data as "not available," the meaning of the term "not available," the use of the zero entry, etc. In general, the policy adopted in preparing this volume was to retain "not available" notations where they appeared for intermediate years in the series; to change them to dashes where they appeared at the beginning or end of the series. Where cells were left blank in the sources, they were filled with dashes unless there was evidence that "not available" was a more appropriate entry. Since series of varying length taken from different sources are frequently found in adjoining columns in a table, the stub listings for years necessarily encompass the earliest and latest date for which any of the series in the table are shown. In itself, this tends to create many additional blank cells since missing entries have been replaced by dashes in order to make it easier for the user's eye to trace the entries for a given year across the entire table.

Source citations and text. For every series shown, the text notes present the source or sources of the data. In most cases, precise publication dates and page or table numbers are given. However, where numerous issues of a certain publication were used, the source citations are usually limited to "annual issues," "various monthly issues," or similar notations. The term "unpublished data" means that the data were not in published form at the time they were obtained for use in this edition. In many cases, such data were scheduled for inclusion in forthcoming publications.

Where possible the descriptive text includes definitions of concepts and terms used, and sufficient methodological and historical information to permit intelligent use of the data. For many series the text also includes reference to where more detail can be found. Unusual values in a series are explained and major changes which affect comparability are noted. Methods used for adjusted or derived figures are described, often with reference to a more complete description. Copyright material. Copyright restrictions, where applicable, are noted in source citations. Permission to quote or reprint copyright material should be obtained directly from the copyright owner.

Statistical Reliability and Responsibility

The contents of this volume were obtained from a large number of sources. All data from either censuses and surveys or based on estimates or administrative records are subject to error arising from a number of sources: Sampling variability (for statistics based on samples), reporting errors in the data for individual units, incomplete coverage, nonresponse, imputation, and processing error. The Bureau of the Census cannot accept responsibility for the accuracy or the limitations of data presented here, other than for those which it collects. Every attempt has been made, within the limits of time and available personnel, to verify and correctly identify the material. Final responsibility for selection of the material, and for its accurate and proper presentation, rests with the Bureau of the Census, even though carried out with the cooperation of many individuals and agencies who devoted much time and energy in providing data and descriptions of series for this publication.

Summary of selected new series

Urban population characteristics (A 82-90); foreign born, by sex and race (A 105-118); population characteristics and land area, by States (A 195-263); households, by number of persons (A 335-349); inmates of institutions (A 359-371)

Fertility and birth rates, by age of mother (B 11-19); illegitimate births and rates (B 28-35); health expenditures (B 221-261); hospital use rates, expenses, personnel, and insurance coverage (B 381-422)

Aliens naturalized (C 162-167); passenger arrivals and departures (C 296-331)

Working women (D 49-74); unemployment rates (D 87-115); employee output (D 683-704); employee earnings (D 705-786); union membership and work stoppages (D 927-939, D 986-1021)

GNP implicit price deflators (E 1-22)

GNP growth rates and per capita (F 10-31); GNP summary and by type of industry (F 32-46), F 130162); national and personal income (F 163-209, F 250-348); national tangible assets (F 349-376); value and age of capital stocks (F 470534)

Distribution of families and persons, by money income levels (G 1-268); nutritive value of city diets (G 857865); food used at home (G 866-880)

Employee benefit plans (H 70–114); OASDHI (social security) coverage, benefits, and trust fund (H 186-259); Civil Service retirement (H 260-270); private philanthropy (H 398-411); schools and school enrollment (H 418-432, H 442-522); school retention rates (H 587-597); income, by years of school completed (H 648663); illiteracy (H 669-688); doctorates, by field (H 766-787); foreign travel, passports issued, and foreign visitors to U.S. (H 894-951); crimes and crime rates (H 952-961); homicides and suicides (H 971-986); criminal justice expenditures (H 1012-1027); lawyers (H 1028-1062); courts (H 1063-1124)

Indian lands (J 16-19); tornadoes, floods, and cyclones (J 268-278)

Farm population, land in farms, farm property, farm products sold (K 181); farm operator characteristics (K 82-108); balance sheet of farming and value of gross farm product (K 204-239); farm income and farm marketings (K 256-343); farm productivity (K 407-495)

Summary of selected new series

Forest products production, imports, exports, and consumption (L 56–71, L 87-97, L 138-165, L 178-198); fishery imports and exports (L 224235); value of landed fish catches (L 244-253); prices received by fishermen (L 321-337)

Employment and man-hours in mineral industries (M 287-306)

Wholesale price indexes for construction materials (N 140-155); new publicly-owned housing starts (N 171-179); value and mean age of stocks of residential structures (N 200-231); housing vacancy rates (N 246–258)

General statistics (P 58-67); horsepower of power equipment (P 68-73); shipments, inventories, and orders (P 74-92); corporation sales, profits, and equity (P 93-106); shares of largest companies (P 177-204); consumption of selected commodities (P 216230)

Intercity passenger traffic (Q 1-11); revenues and employment, by type of transport (Q 23-46); intercity motor carriers (Q 69-81); long-term public highway debt (Q 136-147); motor vehicle insurance, ownership, and financing (Q 163-186); motor vehicle speed, miles of travel, and accidents (Q 187-232); merchant vessels built, repaired, launched, and active (Q 438-480, Q 487-502)

Cable TV (R 98-101); radio and TV finances (R 113-120, R 130-137); new books published (R 195-217); newsprint consumption (R 218-223) Consumption of raw materials and fuel resources (S 15-31); privatelyowned electric utility generating plants and balance sheet and income items (S 58-73, S 133-146); REA summary of operations (S 147-159); waterpower (S 160-175); natural gas consumption and gas utility industry (S 178-218)

Annual earnings of full-time employees (T 29-42); legal form of organization of industries (T 43–57); merchant wholesalers sales and stocks (T 375-383); index of national advertising expenditures (T 472-484)

Value of direct foreign investment in U.S. (U 47-74); U.S. Government foreign grants and credits (U 75–186); exports and imports, by broad enduse class (U 249–263)

Number of firms, by type of ownership (V 1-12); income of unincorporated firms (V 66-77); manufacturing and trade sales and inventories (V 78-107); gross product and unit costs of nonfinancial corporations

XV