Detail map of London, England, United Kingdom,Albany, New York, United States

A: London, England, United Kingdom, B: Albany, New York, United States

William Farr Publishes the First Instances of a Printing Calculator Used to Do Original Work

1857 to 1864
William Farr, circa 1850.

William Farr, circa 1850.

In 1859 English statistician and epidemiologist William Farr published "On the Construction of  Life-Tables, Illustrated by a New Life-Table of the Healthy Districts of England," Philosophical Transactions 149, pt. 2 (1859) 837-78. This was the first report describing the use of the Scheutz Engine no. 3 to prepare life tables, and it included a table calculated and typeset by the calculator. Farr, a pioneer in the quantitative study of morbidity and mortality, was chief statistician of the General Register Office, England's central statistical office. Influenced by Charles Babbage, he had long been interested in the use of a calculating machine such as Babbage's Difference Engine No. 1 to compute life tables. On page 854 of his paper Farr referred to his 1843 letter on this subject to the registrar-general. Farr had seen and tested the machine's predecessor, the Scheutz Engine no. 2, when it was on display in London. It was at Farr's recommendation that the British government authorized in 1857 the sum of £1200 for the Scheutz Engine no. 3 to be constructed by the firm of Bryan Donkin, a manufacturer of machinery, including those for the color printing of bank notes and stamps. Costs overran and Donkin delivered the machine in July 1859, several weeks past the deadline, at a loss of £615 (Lindgren 1987, 224-25). Farr's preliminary report, received by the Royal Society on March 17 of 1859, was written while the Scheutz Engine no. 3 was still "in the course of construction by the Messrs. Donkin" (p. 854). The report's table B1, "Life-Table of Healthy English Districts," made from stereotype plates produced by the calculator, represents the very first application of a difference engine to medical statistics.

Prior to their production of their Difference Engine No. 3, in 1857 the Scheutz brothers had brought the Scheutz Engine no. 2 from Sweden to London, where it was used to produce Specimens of Tables, Calculated, Stereomoulded, and Printed by Machinery. (London, 1857. These were the first mathematical tables calculated and typeset by a mechanical calculator. 

The Scheutz Difference Engine No. 2 was purchased in 1857 by the Dudley Observatory in Albany, New York. The following year the observatory used the machine in the computation of tables for the planet Mars; however, these were experimental and probably never printed on paper (Lindgren 1978, 211). The Scheutzes, Farr, and the Dudley Observatory were the first to use the Scheutz calculator in a scientific context.

In 1864 Farr published English Life Table. Tables of Lifetimes, annuities, and premiums. . . . Published by authority of the Registrar-General of births, deaths and marriages in England. The colophon leaf of this book indicated that 500 copies were printed. Farr's English Life Table contained, what was for its time, a tremendous amount of data— 6.5 million deaths sorted by age. Included in English Life Table no. 3 were the first lengthy working tables produced by the Scheutz printing calculator— the first instance of such a machine being used extensively to do original work. However, none of the hoped-for benefits of mechanizing the calculation of the tables were realized, since the Scheutz machine failed to include any of Babbage's security mechanisms to guard against mechanical error, and it required constant maintenance.

The machine did accomplish some of the typesetting which it stamped into stereotype plates; however, the process was so problematic that there was little cost savings from automation. Of the 600 pages of printed tables in the book, only 28 pages were composed entirely by the machine; a further 216 pages were partially composed by the machine, and the rest were typeset by hand. Nor was there the hoped-for savings from using the machine to prepare stereotype plates. Her Majesty's Stationery Office, printer of the volume, stated that having the machine set the entire book automatically would have saved only 10 percent over the cost of conventional typesetting (Swade, The Cogwheel Brain [2000] 203-8).

Pages cxxxix-cxliv contained Farr's appendix entitled "Scheutz's calculating machine and its use in the construction of the English life table no. 3," in which he emphasized the usefulness of the new machine, but also the delicacy and skill necessary for its operation:

The Machine required incessant attention. The differences had to be inserted at the proper terms of the various series, checking was required, and when the mechanism got out of order it had to be set right. Of the first watch nothing is known, but the first steam-engine was indisputably imperfect; and here we had to do with the second Calculating Machine as it came from the designs of its constructors and from the workshop of the engineer. The idea had been as beautifully embodied in metal by Mr. Bryan Donkin as it had been conceived by the genius of its inventors; but it was untried. So its work had to be watched with anxiety, and its arithmetical music had to be elicited by frequent tuning and skilful handling, in the quiet most congenial to such productions.

This volume is the result; and thus—if I may use the expression—the soul of the Machine is exhibited in a series of Tables which are submitted to the criticism of the consummate judges of this kind of work in England and in the world (p. cxl)

Farr also noted Babbage's contribution to the venture—it was Babbage who "explained the principles [of the Scheutz calculator] and first demonstrated the practicability of performing certain calculations, and printing the results by machinery" (p. xiii).

Having invested so much time and money in the project while realizing only token gains, the British government showed little patience with the Scheutz calculating machine. The General Register Office soon reverted to manual calculations by human computers employing logarithms, which they used until the GRO's conversion to mechanical calculation methods in 1911.  

Hook & Norman, Origins of Cyberspace (2002) Nos. 77 & 85.

(This entry was last revised on 01-14-2015.)

Timeline Themes

Related Entries