Below is text from one section of Douglas Galbi’s work, “Sense in Communication.”  This work includes text and images.  Some images may be missing (due to use restrictions) or improperly formatted below.  The full work in pdf format, as well as other text sections, are available at www.galbithink.org

 

Table Notes

Sources, Definitions, and Calculations

 

 

Table 2

 

Ranks for England are for names from forty dispersed English parishes.  See Smith-Bannister (1997), Appendix C, pp. 196-203.  This source does not provide the popularity percent.  It is estimated as the average across the corresponding rank for the Warwick sample (1538-1659) and the North/Cumbria sample (see Table A3; est. birth years included here, 1538-1659).

      The figures for Warwick for 1381-1405 and 1465-1509 are from persons entered into the Guild of the Holy Cross, the Blessed Mary, and St. John the Baptist for the years 1406-1430 and 1490-1534, respectively.[1]  This guild was located in Stratford-upon-Avon.  It provided social, educational, and spiritual services to its members, and owned a school, a chapel and other real estate in Stratford.  Early in the fifteenth century, it had perhaps 250 members, while the population of Stratford-upon-Avon may have been around 1000.[2]  Thus membership in the guild amounted to perhaps half the adult population of Stratford.  In the early fifteenth century, the cost for a living person to enroll in the guild was 20s, while a soul could be enrolled for 6s 8d.  Payments in installments, or with goods and services, or at a discount, were occasionally made.[3]  By earlier in the sixteenth century, a (living) husband and wife could enroll for 6s 8d, while a (dead) soul could be enrolled for 20d (1s 8d).[4]   Wage rates for building craftsmen and laborers in southern England, 1412-1532, were 6d and 4d per day respectively.[5] Thus a laborer could join the guild for about a month’s worth of work.  Given that this was a once-in-a-lifetime fee and that persons attached high value to having a good funeral, which the guild arranged, the entrance fee was probably not considered high to most persons in Stratford in the early sixteenth century.  These facts provide evidence that the names of guild entrants were representative of the names of all persons in Stratford.

      The figures for Warwick for years 1513 and later are from the parish registers for Stratford-upon-Avon and Solihull in the county of Warwick.[6]  Solihull is about 15 miles north of Stratford.  All years are for baptism records, except for 1513-1525.  Names for those years are from marriages in Solihull, 1538-1550.   Frequencies for Mary do not differ significantly between Stratford-upon-Avon and Solihull.   Sample sizes are provided in Table A2.  The name database is available at http://www.galbithink.org/names/engb1800.htm

 

Table 3

 

Data for activities other than telephone conversations are from Galbi (2001a) pp. 11-20 and are reported to the nearest hour.  The figure for 1925 is for circa 1925.

      Unfortunately, time budget studies have not typically separated telephone conversations from voice conversations among persons in physical proximity.  In the context of reporting time use, the amount of telephone time is small and tends not to be accurately reported (see Brandon (1981) Figures 5.B.13 and 5.B.14, and Sorokin and Berger (1939) pp. 52) . 

      To get per person telephone use time during discretionary (non-work) time, total minutes of telephone conversation time (see sources and estimates for Table 7) has been multiplied by two to get minutes of telephone use for the population [ver. 1.0 of this paper incorrectly used conversation time].  Telephony use time has been divided by the total US residential population ages 10 and older, scaled to minutes per week, and multiplied by .5 to separate home use from business use.  The figure has been rounded to the nearest half-hour.  Thus zero for 1925 indicates less than a quarter hour of discretionary telephone use per person per week.

 

Table 4

 

Year 1890:  The figures for photographers, artists, and dentists are the figures given for “photographers,” “artists and teachers of art,” and “dentists,” respectively, in Census (1900) Table III, pp. xlix, xxxiv.  Figures for accountants and authors are the figures for “accountants and auditors” and “authors” given in Historical Statistics (1975) D 235, 244 (p. 140) for 1900, deflated by the per decade growth rate indicated by the figures for 1880 and 1900 for these categories in the IMPUMS OCC1950 coding frequency count, available at http://www.ipums.org/usa/pwork/occ1950b.html

 

Year 2000: Figures from Statistical Abstract: 2001, No. 593, p. 380.  Category descriptions are:

Table 4 Heading

Statistical Abstract Category

Accountants

Accountants and auditors

Artists

Painters, sculptors, craft-artists, and                                          artist printmakers

Authors

Authors

 

 

Table 5

 

End-user supplies includes camera, film, photofinishing services, etc. sold to produce custom photographs to be viewed by individuals.  It does not include photography equipment and supplies used in industrial production and health services, such as photoduplicating, x-ray photographs, and dental film.  Professional services are the receipts of professional photographers, including both portrait and commercial studio categories.

 

Year 1890:  Statistics are from the Census (1890), Table 4, p. 104.  End-user supplies are the sum of “value of products…” in categories “photographic apparatus” ($624,432) and “photographic materials” ($2,121,387), both inflated by 20% to account for retailing margins.  Professional services are “value of products, including custom work and repairing” in the category “photography,” which was considered a “manufacture” through the Census of 1900.  Subsequently, professional photography was considered a service establishment, and included only in the first business census covering services in 1939.  In 1890, 3,105 establishments were reported with value of products, $15,488,324.  In 1900, 7,553 establishments were reported with value of products, $23,238,719.   The large increase in the number of establishments reporting in 1900 relative to 1890, and the small share of establishments relative to the number of photographers reported in the population census of 1890, indicates underreporting of establishments in 1890.  The number of establishments in 1890 was estimated as 5,618, which is establishments per photographer in 1900 (0.28) times the number of photographers in 1890 (20,040).  The value of products of the additional imputed establishments was estimated based on the revenue per establishment for the increase in establishments and revenue from 1890 to 1900.

 

Year 1939:  Professional services are “receipts” in category “photographic studios” in Census (1939), Service Establishments, v. 3, Table 1A, p. 16. 

      End-user supplies for 1939 are estimated from Kodak’s 1939 US sales ($104,470,778; see Kodak Annual Report, 1939, p. 5), multiplied by an estimate of Kodak’s non-industrial/health sales (.5; see id.), and divided by Kodak’s market share in film about this time (0.85; see Kadiyali (1998) p. 91-2), and marked up by a distribution margin roughly estimated at 20%.  The resulting estimate is $74 million. 

      A second estimate of end-user supplies for 1939 can be built from census service and retail categories.   The Census (1939) Service Establishments, v. 3, Table 1A, p. 16, lists, under “business services,” receipts of photofinishing labs ($16.1 million).  Retail photofinishing services include some mark-up, estimated at 20%, over this total.  The Census (1939), Retail Trade, Table 18, pp. 163-9, provides retail sales of cameras, projectors, photo equipment and supplies in six kinds of businesses:

 

1939 Census of Business

Retail Sales of End-User Photo Supplies

 

Retail Store Class

Sales of Photo Supplies

Photographic supply-camera stores

20,144

Jewelry stores

1,858

Opticians

998

Sporting goods stores

899

Radio stores

373

Radio-musical instrument stores

104

 

Combining photofinishing, marked up, and the above retail sales gives an end-user supplies estimate of $43 million.

      A third estimate can be constructed from the Census (1939), Distribution of Manufacturers’ Sales, Table 1, p. 197.  In the category “optical instruments and lenses,” $2.1 million in sales were distributed to non-industrial entities.  In the category “photographic apparatus and materials and projection equipment (except lenses),” distributed sales were $141.2 million.  Subtracting from those sales $12.1 million exported and $26.5 million that went directly to industrial users, adding sales to non-industrial entities from “optical instruments and lenses,” and marking up by 20% gives $125 million.

      The first estimate, $74 million, which is close to the average of the other two, is used in the table.

 

Year 1967:  Professional services are from Census (1967), Table 2, “Photographic Studios” (SIC 722) [which includes portrait and commercial], all establishments.  End-user supplies are total “retail sales of photographic merchandise and photofinishing” for 1967 from Wolfman Report for 1968.

 

Year 1997:  Professional services are from Census (1997), “Photographic services” [NAICS code 54192], all taxable firms, receipts.  The last Wolfman Report was published in 1994, and covered the 1993.  End-user supplies for 1997 are extrapolated from Wolfman Report’s retail sales for 1983 and 1993.

 

Table 6

 

Year 1890:  Historical Statistics (1975), Series R 2, indicates 3.7 telephones per 1000 population in 1890.  There were roughly 80,000 cameras in the US in 1890, or 1.3 cameras per 1000 population (see below Table 7, notes concerning amateur photo count in 1890).  The ratio of population to households in 1890 was about 5 (Historical Statistics (1975) Series A 257).  These figures imply that about 2% and 1% of households had telephones and cameras, respectively.

 

Telephone Household Percentage, 1938, 1958, and 1995:  These figures are from the Statistical Abstract: 1999, No. 1440, p. 885, interpolated where necessary.

 

Camera Household Percentage for 1938:  Taft (1938) p. 404 estimates 15 million amateur photographers.  One amateur photographer is assumed to be per household.  Household figure from Historical Statistics (1975) Series A 242. 

 

Camera Household Percentage for 1958: Wolfman Report for 1960, p. 17.

 

Camera Household Percentage for 1995:  PMA (2000) Chart1a-1, p. 14.  The prevalence of one-time-use cameras has reduced camera ownership in 1999 to 87% of households.

 

Table 7

 

Professional photos are added to amateur photos to get total photos.  Since the issue of interest is photo use, the number of professional photos is considered to be the number of photos persons purchased from professional photographers, not the number taken.  Identification card photos are considered product inputs, not end-use photos, and hence are not considered.  Photos replicated in magazines, newspapers, and posters are not counted because persons do not make explicit and separable choices to acquire these photos.  The number of amateur photos has grown much faster than the number of professional photos, and hence in the estimate of total photos the number of professional photos matters most for the photo counts in earlier years.  

 

Professional Photographs

 

In a knowledgeable article describing the development of photography, Tribune (1853) reported the estimate that “there cannot be less than 3,000,000 daguerreotypes taken annually in the United States.”   The Census of 1850 reported 938 daguerreotypists, while the Census of 1860 reported 3,154 daguerreotypists and photographers.  Interpolating these figures implies 1350 photographers in 1853.  Bear (1873) describes an industrious daguerreotypist working by himself who sold 10 daguerreotypes a day, at $1.50 per daguerreotype, for several months in 1846.  This operation was highly profitable, but the daguerreotypist moved regularly from placed to place.  This suggests that his business was not a steady-state operation.  Photographer (1896) describes a large establishment with an intricate division of labor that produced 400-500 daguerreotypes a day at $3 to $8 per daguerreotype in 1841.  Werge (1890), pp. 196-202, describes a “factory” portrait operation c. 1854 that made portraits for 25 cents per photo.  The estimate of 3,000,000 daguerreotypes annually in 1853 implies about 2200 daguerreotypes per photographer in that year.  The cost of photography and its technical difficulty surely fell from 1853 to 1890, but the demand benefits of offering a new service also declined.   Scaling the 1853 daguerreotype estimate by number of photographers indicates about 44 million photos in 1890. 

      Most professional photographs are personal portraits.  West (2000), p. 2, makes an informed estimate that, in 1890, a typical family probably owned about 10 photographs.  In 1965 and 1983, about one out of every three households purchased professional portraits (Wolfman Report for 1968, p. 36; Report for 1983-84, p. 84).  In 1999, about 26% of households purchased professional photographic services (PMA(2000) Chart 13-3, p. 157).  The number of professional photographs purchased per year can be estimated as the number of households, times the share of household purchasing, times the number of photos purchased per households purchasing.  Combining Census data on households with rough guesses for shares purchasing and photos purchased gives the following estimates of professional photos:

Estimate of Total Professional Photographs Purchased

 

Year

Households

Share

Purchasing

Photos

Purchased

Total

Photos

1890

12,690

0.33

5

20,939

1939

34,408

0.33

15

170,322

1973

68,179

0.33

25

562,479

1995

98,984

0.33

25

816,616

These figures are used, except that, in light of the alternative estimate of 44 million professional photos in 1890, a middle estimate of 30 million is used for 1890.

 

Amateur Photographs

 

Year 1890:  Ford (1989) p. 62 states, “By September of 1889, over 5,000 Kodak cameras had been sold in the USA, and the company was daily printing 6-7,000 negatives.”  Taking the work week as 5.5 days, 6,500 negatives per day implies 1.86 million photos per year.  Jenkins (1975), p. 18, notes Kodak sales (U.S.) as $0.45 million in 1889 [cited to U.S. v Eastman Kodak. pp. 2565-69] and value of US photography materials and apparatus [rom 1890 Census of Manufactures] at $2.75 million.  This implies a 0.16 Kodak share of the photography industry.  Kodak camera and apparatus sales in 1892 [$207,212, from United States v. Eastman Kodak (1915), 226 F. 62, 67] amount to 0.33 of total US “value of products” of photographic apparatus manufacturers [$624,342, from Census (1890)].  Since Kodak focused played a lead role in promoting amateur photography, its share of amateur photos was probably larger than its share of other aspects of the photography industry.  Taking 0.5 as the share of Kodak photos in total amateur photos in 1890 implies 4 million amateur photos in 1890.

      United States v. Eastman Kodak (1915) [226 F. 62, 67] states that sales of “Kodak and apparatus” amounted to $207,212 and $74,594 in 1892 and 1894, respectively, while sale of [Kodak] film amounted to $102,404 and $81,319 in 1892 and 1894, respectively.  The decrease was attributed to competition from the recently introduced Bullseye camera.  In 1890 Kodak was selling the original Kodak, renamed Kodak No. 1, and the Kodak No. 2, which was introduced in 1889.  Kodak’s 1895 price list [Kodak (1895) p. 68] indicates the prices for “reloading, developing, and printing” for the Kodak No. 1 and No. 2 were both $10, while the numbers of exposures included were 100 and 60, respectively.  The film itself sold for $2.50.  Assuming an average exposure number of 80 [equal numbers of Kodak No. 1 and No. 2 in use] and an average sale price for film and any associated developing and printing of $5 implies 1.64 million Kodak photos.  Taking Kodak’s photo share to be 0.5 implies about 3 million total amateur photos in 1890.

      A figure of about 4 million amateur photos in 1890 is reasonable in light of other evidence.  Figures in West (2000) p. 24, Ford (1989) pp. 62-3, and Kodak (2003) suggest that Kodak sold about 20,000 cameras by the end of 1890.  Taking Kodak’s share of camera sales in 1890, estimated as 0.25, as also a rough estimate of Kodak’s share of total camera’s sold by that time implies total cameras extant were 80,000.  These figures seem consistent with Kodak’s emergence as a camera seller in 1888, following on from a successful business selling dry plates and film.  The 1890 Census found 20,040 persons who identified themselves as photographers.  Assuming one camera per photographer, professional or amateur, implies about 60,000 amateur cameras.  George Eastman estimated that there were about 50,000 amateur photographers in the U.S. about 1889 (Brayer (1996) p. 68).  Along with the total photo figure, the figure for total photographers implies about 80 photos per photographer per year.  This is about the average number of photos included in an early Kodak film cartridge (see previous paragraph).

 

Year 1939:  The Wolfman Report  for 1960, p. 11, gives 2.2 billion and 1.8 billion for amateur pictures made in 1959 and 1953, respectively.  The Eastman Kodak Annual Report for 1954, p. 20, indicates that sales in 1954 were 4.1 times greater than sales in 1939, holding prices constant (1939 dollars).  The sales category “amateur photographic” fell from 33% of Kodak sales in 1939 (Annual Report, 1939, p. 5) to 29% of Kodak sales in 1954 (Annual Report, 1954, p. 19).  These figures are consistent with 3.5 and 4.35 times growth from 1939 to 1953 in amateur and non-amateur Kodak sales.  Scaling 1.8 billion amateur photos in 1953 down by a factor of 3.5 gives 514 million amateur photos in 1939.

      Taft (1938) p. 404 states, “there are at the present time at least fifteen million amateur photographers who are actually engaged in their hobby in this country.”  A footnote adds:

The New York Times (May 16, 1937, section 4, p. 9) estimated that even during the recent depression there were one and one-half million cameras owned by New York’s metropolitan population of seven million and that one million three hundred thousand photographs were being taken weekly.

Scaling the number of photos taken by New York’s population relative to total US population and a year’s duration implies 1.3 billion photos per year.  A weakness of this figure is that persons in a large city probably take on average more persons than persons who live in areas with lower population density.  The data also imply that about 45 photographs were being taken per camera per year in New York.  Multiplying this figure by 15 million amateur photographers implies 676 million amateur photographs per year.  This figure is taken as the best estimate of amateur photos in 1939, and is used in the calculation of total photographs.

 

Years 1973 and 1995:  The amateur photo count for 1973 is from the Wolfman Report for 1993, the last report.  The amateur photo count for 1995 was extrapolated from the Wolfman Report counts for 1983 and 1993.

 

Telephone Conversation Minutes

 

Year 1973:  FCC (1999) Table 21-1 provides total telephone minutes of use data yearly from 1980.  The figures are called “dial equipment minutes-of-use” (DEMs) and are used in FCC analysis of cost support for telephony company rates.  Because of the way DEMs are measured, about two DEMs are measured for each minute of telephone conversation between two persons. The minute estimates in Table 7 are conversation minutes. 

      The growth rate of telephone minutes increased sharply toward the end of the 1990s.  This change is associated with dial-up access to the Internet.  The year 1995 was chosen as the final date of comparison because that is the most recent year in which the average telephone minute volumes probably was not significantly affected by Internet use.  Wireless telephony minutes were not significant relative to wireline minutes in 1995.

      Historical Statistics (1975), series R 9-12, and FCC SOCC 1979, Table 9, p. 18, together provide total telephone calls yearly from 1880 to 1979.  Average call duration is needed to estimate call minutes from these statistics.  Average call duration depends significantly on whether uncompleted calls (busy signals and no answers) are included in call totals.  Mayer (1973) p. 228 describes his sample as having almost 30% of calls lasting less than thirty seconds (see also id. Fig.3, p. 229).  In contrast, Garfinkel and Linhart (1980) Fig. 5, p. 18, seems to include almost no calls less than thirty seconds.  The presence or absence of short calls, plausible associated with uncompleted calls, can significantly affect average call duration.  Average call duration also depends on the mix between business and residential users.  Compared to a residential user, a business user tends to make more calls, but of shorter duration (Garfinkel and Linhart (1980) p. 18).

      What is included in available call volume time series is not clear.  FCC SOCC 1979 Table 9, p. 18, ft. 3 indicates that local and toll figures are partially estimated and include both “completed and uncompleted calls.”  In an earlier version of the table, FCC SOCC 1968 Table 9, p. 21, ft. 3 states, “Local calls after 1950 include only completed calls.  Toll calls include both completed and uncompleted calls for all years.”  The call volume data in FCC SOCC 1979 and FCC SOCC 1968 appear to be based on the same data used in the call volume data in Historical Statistics (1975) Series R 9-12.  The notes to those series indicate that local calls include completed and uncompleted calls, and does not mention any change in figure definition in 1950.  Finally, Census (1912), p. 30, suggests that the call volume statistics generally excluded uncompleted calls,.  It notes that while some uncompleted calls may have been included, “the number is relatively small and does not seriously affect the comparative value of the statistics.” 

      Studies made presenting traffic data collected about 1973 provide the best additional evidence for estimating total telephone minutes.   Data from the studies give estimates of residential household telephone minutes.  In each estimate, residential household minutes are scaled by the number of households with phones in 1973 (90% of households, as indicated from Historical Statistics (1975) R 3 and FCC (1999) Table 17.1, where total households is 61,361, interpolated from Statistical Abstract: 1999, No. 1419 p. 873).  Bell company filings to the FCC in 1973 indicate that business lines were 38.3% of residential lines.  Total minutes of calling on a business line is assumed to equal to minutes of calling on a residential line.  Thus total telephone minutes are estimated as residential minutes times 1.383.  The table below summarizes estimates based on these studies.

 

Telephone Traffic Studies and Minutes Estimates, c. 1973

 

Source No.

Source Name

Residential

Calls/Month

Ave Res.

Call Duration

Total

Res. Min

Total Res. &

Business Min.

1

Brandon (1981)

106

5.8

451

624

2

Garfinkel &

 Linhart (1980)

125

4.5

414

572

3

Mayer (1977)-1

128

4.3

404

558

4

Mayer (1977)-2

149

5.1

562

777

 

Average

127

4.9

457

633

      Notes for no. 1:  Data from Brandon (1981), Table 8.B.1, 8.B.3, Figure 10.A.3, and Figure 10.A.12.  Averages for blacks and whites are weighted by respective 1973 population shares (0.11/0.89).   Calculating average local-suburban call duration and comparing to Table 8.B.2 indicates that the call total contains about 10% zero-length (uncompleted) calls. 

     Notes for no. 2: Data from Garfinkel and Linhart (1980) Fig. 3, p. 17; and Fig. 5, p. 18.  These data are for local calls.  In 1973, toll calls were 6.7% of local calls by call numbers (FCC (1999) Table 9).  The figure for local calls per month is inflated by this factor to get total residential calls/month.

     Notes for no. 3: Data from Mayer (1977) p. 227-9.  The local call figure has been inflated as for estimate no. 2 above.

     Notes for no. 4: Data from Mayer (1977) p. Figs. 5a, 5b.  Average figures have been constructed using Census data on income classes.  Calls per household per day have been estimated from calls per user per day by multiplying by the average number of persons 10 years old and older per household in 1977 (2.55).

 

      Total calls can also be estimated two other ways.  FCC SOCC 1984, Table 8, p. 13, indicates 213 billion calls in 1973.  Multiplying this figure by 5 minutes per call gives 1065 billion minutes.  This figure is significantly higher than all the estimates in the table above.

      Alternatively, the yearly average growth rate of DEMs from 1980 to 1995 (3.8%, calculated from DEMs given in FCC (1999) Table 17-1) can be used to back-project total conversation minutes (one-half total DEMs) for 1973.  Doing so gives 668 billion minutes.  This figure, which is close to the average total minute figure from the table above, is taken as the best estimate for total telephone conversation minutes in 1973. 

     

Year 1995:  The minute figure for 1995 is calculated as one-half of total DEMs for 1995. 

 

Year 1939:  Using Bell company ratios of phones to main lines for 1939 and 1973, total phones for 1939 and 1973 are used to estimate total main lines for 1939 and 1973 of 14.7 million and 70.4 million, respectively.  These figures imply a main line growth rate of 3.6% per year.  Mayer (1977) p.232 notes “the steady growth in telephone usage – at the rate of abut 2 percent per household for some years….”  Assuming minute grew 2% per main line per year from 1939 to 1973 implies 5.6% increase in total minutes per year across this period.  Back-projecting from 668 billion minutes in 1973 gives 58 billion minutes in 1939.

      Alternatively, Willey and Rice (1933) p. 144 report AT&T data indicating that the average time spent in each telephone conversation over “long lines” was about 5 minutes for 1924 to 1929, with little variation per year.  Id. also cites another AT&T source from 1930 indicating “the average time spent in transcontinental conversations is about six minutes.”  Average call length generally increases with call distance.  Brandon (1977) Table 8.B.2 and Fig. 10.A.11 indicate that local-suburban call duration is about 25% shorter than toll call duration.  Reducing the AT&T long-lines duration figure by this amount implies average call duration of 3.7 minutes per call in 1939 (toll calls were only 3% of total calls in 1939, and hence difference between toll and local call duration have a negligible effect on average call duration). 

      From above, the best estimate of average call duration in 1973 is 5 minutes per call.  An increase in average call duration from 3.7 to 5 minutes per call from 1939 to 1973 is consistent with the increase in share of residential lines from 62% to 72% across that period.  Business lines tend to have higher call volume but shorter call duration (Garfinkel and Linhart (1980) p. 18). 

      An average call duration of 7.5 minutes reported in 1909 (see Judson (1909) p. 646) seems, in light of the above data, to be not at all representative.  Judson, Chief Engineer of the Independent Telephone Co. in Seattle, was concerned about the large amount of “unprofitable traffic generated by “idle gossip” under flat-rate calling plans.  That he was concerned about traffic volume indicates the potential for selection bias: high average calling duration is correlated with his motivation to report his statistics.   Moreover, early in the twentieth century non-Bell telephone companies pursued residential subscribers much more aggressively than Bell companies.  Thus Judson’s sample probably included a much higher fraction of residential subscribers than was representative for the industry as a whole.  The median call duration in Judson’s sample was between three and six minutes, while the median in a call sample from 1973 was 4.05 minutes (Garfinkel and Linhart (1980) Fig. 5).  Judson’s sample had 8% of calls longer than 8 minutes, and 16% longer than 15 minutes.  The 1973 sample, by contrast, had only about 3% of calls longer than 10 minutes.  Thus Judson’s high average duration seems to have been created by a small fraction of his subscribers who were making long-duration calls.

      The best estimates of total minutes (668 million) and average call duration (5 minutes per call) for 1973 imply that total reported calls in 1973 should be scaled down by a factor of 0.62.  Applying this same scaling factor to total calls in 1939 and taking average call duration to be 3.7 minutes implies 78 billion minutes in 1939.  Averaging this estimate with the above estimate of 58 billion gives the final estimate of 68 billion minutes in 1939.

 

Year 1890:  The estimate for 1890 involves the most uncertainty.  In 1890, service quality was undoubtedly significantly worse than in 1939.  Until its patents expired in 1894, Bell Telephone effectively had a monopoly on telephone service.  In the late nineteenth century, Bell Telephone primarily sold its service to businesses, and thus most telephone users in 1890, perhaps 90% (Mueller (1997) p. 4), were business users.  Awkwardness associated with lack of familiarity with telephone service, worse quality service, a much higher share of business users, most of whom were probably men, all suggest a significantly lower average call duration.  A plausible estimate is 2 minutes per call.  That figure implies 1,057 million minutes of telephone calls in 1890.

 

 

Table 8

 

Photo industry revenue:  See sources and calculations for Table 5.

 

Telephone industry revenue:  Figures for 1890, 1939, and 1967 are the sum of operating revenue for Bell and Independent companies, from Historical Statistics (1975), Series R-20 and R-35.  For 1967, the sum is 9.5% higher than operating revenue of telephone carriers filing annual reports to the FCC, as given in FCC SOCC 1984, Table 8, p. 13.  Although the latter figure may better account for inter-company duplications in operating revenue among the Bell companies (see id., Table 5, notes), the former figure has been used for comparability with earlier figures.   The 1997 figure is end-user revenue reported in FCC (1998), Table 1.

 

Table A1

 

Data from sources referenced in Galbi (2001b) Table 5, except “1350, Yorkshire”, which is from Gwynek (n.d.).  Years given in the table are estimated average date of birth for the corresponding sample.

 

Table A2

 

See notes for Table 2 above.  This table adds the sample sizes for the Warwick sample in Table 2.  Smith-Bannister (1997) does not provide sample sizes for the England rank data.

 

Table A3

 

Data from Galbi (2001c), available online.  Names are from marriage records; average age at marriage estimated at 25.

 

Table A4

 

See Galbi (2001b), Table 3 and Appendix D.

 

Tables C1 and C2

 

The first U.S. census that included detailed occupational data was in 1850.  Occupation categories and counts 1850 to 1980 were taken from corresponding census report on detailed occupations.  Data for 1990 are from Census (1990), Table 1.  Data for 2000 are from Hecker (2001), Table 2.   For discussion of revisions of the Standard Occupational Classification System, see Bureau of Labor Statistics (1999).

 



[1] Bloom (1907) provides a transcript of entrants from 1406 to 1534.

[2] Attendance at the guild’s annual feast in 1409, 1411, and 1413 was 133, 108, and 160 persons, respectively.  Rosser (1994) p. 439, which also estimates total membership as 250.  On Stratford’s population, see Dyer (1997) pp. 44-5.  Note that the population would have fallen after the great plague of the mid-fourteenth century.  Wrigley and Schofield (1981) p. 528 estimates that in 1541 48% of the English population was 25 years old or older.

[3] E.g. Bloom (1907) p. 12, “William Reve, & Agnes, his wife…40s…payable by installment.”  John Prynce, “Master Cook of the Hospice of Sir Richard, Earl of Warrewyke, entered in 1416 thus: “And made his fine, giving nothing save that he will be always glad (assiduous) to give advice & aid if he be forewarned, annually during his life, and that he will come & labour at the Common (Feast).” Id. p. 26. Johane, wife of John Phippus, entered in 1415 for 13s 4d.

[4] Persons were also enrolled individually for 6s 8d, while some husband-wife combinations were enrolled for twice that, i.e.13s 4d. 

[5] Phelps Brown and Hopkins (1981) p. 11.  Dyer (1997) p. 53 provides a similar estimate for Stratford.

[6] See Galbi (2003), compiled from Savage (1905) and Savage (1904).