are noted here; others can be noted by reference to the various census volumes. The 1954 and later censuses were conducted by mail canvasses of firms included in the active records of the Internal Revenue Service as subject to the payment of Federal Insurance Contributions Act (FICA) taxes and which were in appropriate kind-of-business classifications. Such data cover only firms with paid employees. The nonemployer segment was derived from a 50-percent sample of 1954, 1958, and 1963 tax returns. In the 1967 census, data for all nonemployers were compiled from tax records. The 1948 and earlier censuses were conducted by field enumeration. The differences in enumeration methods affect particularly the coverage of establishments without easily recognizable places of business and those leaving business prior to the end of the year. The 19541967 data are thus more complete in those areas. The 1933 and 1935 censuses were not taken under mandatory reporting requirements and may therefore be subject to some underenumeration. There are important gaps in enumerators' reports for 1933 so that substantial underenumeration, particularly of the smaller establishments, exists for 1933. Underenumerations have more effect on the number of establishments than on receipts.

In the 1963 and 1967 censuses, nonemployer establishments which did not operate the entire year have been included if, during the period they operated, their receipts were at a rate which would have reached an annual total of $1,000 or more had they operated the entire year. Establishments without paid employment and with less than $1,000 receipts were excluded in 1954 and 1958 tabulations. The data for 1948 (comparable with later years) show 1948 figures adjusted to this cutoff point. The data for 1948 (comparable with earlier years) exclude establishments which operated the entire year but had receipts less than $500. For 1939 and earlier years establishments having receipts of $100 or more are included (except as noted). Where two estimates are shown for 1939, the figures for 1939 (comparable with later years) represent a revision to conform to 1948 kind-of-business definitions.

Receipts for 1954 and later years include sales and excise taxes; receipts for 1948 and 1939 exclude them.

Establishments are classified according to their principal kinds of business. Changes in relative importance may thus serve to shift particular establishments among service categories or between service and retailing classifications from one census to another. Many service establishments derive some receipts from sales of merchandise; conversely, many establishments primarily engaged in the sale of goods, and hence included in retail trade, obtain some income from services. Receipts reported in each case represent total receipts of establishments comprising the classification, not receipts for the particular service indicated.

T 402-403, total personal services. Data for 1933 and 1935 represent groupings that correspond most closely to the 1939 scope.

T 410-411, photographic studios. Since the 1954-1967 data were obtained by mail canvass, they are believed to be substantially more complete than data for earlier years. For this industry, nonrecognizable establishments are likely to result in substantial underenumeration in a field canvass.

T 412-417, laundry, cleaning, and garment services. Included in series T 412 are power laundries, cleaning plants, press shops, linen supply, diaper service, industrial launderers, garment repair, and hand laundries. For 1933 and 1935, power laundries and dry cleaning plants with receipts of less than $5,000 were omitted. While series T 414 does not include the count of outlets owned and operated by dry cleaning plants, series T 415 does include the receipts of such outlets.

T 426-427, automobile repair shops. Data for 1935 include specialized shops as enumerated in the census of service establishments, and general repair garages as enumerated in the 1935 Census of Business, Retail Distribution, table 1A. Data for 1933 cover only general repair garages, as enumerated in the 1933 Census of American Business, Retail Distribution, table 1A, and the following types of

specialized shops as reported in 1933 Census of American Business, Service Industries: Paint shops, radiator shops, top and body repair shops, tire repair shops, and brake repair shops.

T 434-437, miscellaneous repair services. Separate data are available for some or all of the indicated years for several of the repair services in this group including shops engaged in armature rewinding, bicycle repair, blacksmithing, harness and leather goods repair, musical instrument repair, saw and tool repair, typewriter repair, upholstering and furniture repair, watch, clock, and jewelry repair, etc. Since the 1954–1967 data were obtained by mail canvass, they are believed to be substantially more complete than data for earlier years. In these industries, nonrecognizable establishments are likely to result in substantial underenumeration in a field canvass.

T 438-439, hotels, tourist courts, motels, trailer parks, and camps. Data for 1954-1967 are for establishments with payrolls only.

T 444-471. Volume of advertising, by medium, 1867–1970.

Source: Printers' Ink Publications, New York, N.Y., 1867-1934, Printers' Ink Advertisers' Annual, 1955 edition; 1935-1968, Printers' Ink Advertisers' Guide to Marketing, annual issues; 1969 and 1970, Marketing/Communications, July 1971.

The data were prepared by Robert J. Coen of McCann-Erickson, Inc., from information furnished by the American Newspaper Publishers Association, A. C. Nielsen Company, Publishers' Information Bureau, Farm Publication Reports, Inc., the Direct Mail Advertising Association, A. R. Venezian, Outdoor Advertising, Inc., and the Federal Communications Commission.

The data include the cost of preparation, and the cost of talent in the case of radio and television as well as the charges for space and time.

T 472-484. Indexes of national advertising expenditures, by medium, 1935-1970.

Source: Compiled by Robert J. Coen of McCann-Erickson, Inc., from annual dollar figures prepared and published in Advertising Age, Crain Communications, Inc., Chicago.

The general index, series T 472, is designed to reflect the changes in advertising dollar expenditures by national advertisers in major media. The national advertiser is usually a manufacturer of a product or service who does business in many markets across the country. National advertisers usually place their advertising through the facilities of an advertising agency and national media sales representative.

The index numbers for all media are based on estimates of the total expenditures made by advertisers to cover space and time charges as well as charges for programming, art and mechanical production expenses, and all commercial production items.

Estimates for the individual media are developed according to several general sources. Broadcast media estimates are derived from the financial reports of the Federal Communications Commission (FCC). The FCC provides comprehensive data on time sales revenue and also program revenue for network television, network radio, spot television, and spot radio. Adjustments are made to these figures to include additional allowances for commercial production expenses and some programming and talent charges not channeled through the stations or networks.

Magazine estimates are primarily based on the reports of the Publishers Information Bureau. Newspaper estimates are primarily based on the reports provided by the Bureau of Advertising of the American Newspaper Publishers Association. Business publications estimates are developed from data collected by the American Business Press. Outdoor estimates are developed from data provided by the Institute of Outdoor Advertising.

The above sources relate to the figures for the most recent years.

In some cases older historical data were derived from sources that are no longer in existence. All data are adjusted and linked in order to make the historical series comparable from year to year.

Data for all media include Alaska and Hawaii as well as expenditures for media in outlying areas of the United States.

T 485-491. Newspaper advertising-linage for 52 cities, 1928-1970. Source: U.S. Bureau of Economic Analysis (formerly Office of Business Economics). 1928-1938, unpublished data; 1939-1946, 1969 Business Statistics, p. 57; 1947-1970, 1971 Business Statistics, p. 57.

T 485-491, T 1-14

Data represent newspaper linage in all newspapers, daily and Sunday, in the following 52 cities: Akron, Albany, Albuquerque, Atlanta, Baltimore, Birmingham, Boston, Buffalo, Chicago, Cincinnati, Cleveland, Columbus, Dallas, Dayton, Denver, Detroit, El Paso, Fort Worth, Hartford, Houston, Indianapolis, Jacksonville, Knoxville, Los Angeles, Memphis, Milwaukee, Minneapolis, Nashville, New Orleans, Oakland, Oklahoma City, Omaha, Pittsburgh, Portland (Oreg.), Reading, Richmond, Rochester, Salt Lake City, San Antonio, San Diego, San Francisco, Seattle, South Bend, Spokane, St. Louis, Syracuse, Tacoma, Toledo, Tulsa, Washington, Worcester, and Youngstown.