Newtonpedia ES
Home › Computing › Arithmometer
Arithmometer
Image: Wikimedia Commons
Computing · 1820

Arithmometer

A desktop mechanical calculator, the first of its kind to succeed commercially and to be used every day in offices around the world.

Year
1820
People involved
Charles Xavier Thomas de Colmar
Section
Computing

History

The arithmometer is one of the first mechanical calculators reliable enough to work every day in an office, able to add, subtract, multiply and divide. It matters because it was the first such machine to achieve sustained commercial success, and it launched an industry that lasted more than a century.

Before it, calculating machines were mostly limited to addition. Gottfried Leibniz had built a machine in 1694 around a stepped drum, now called the Leibniz wheel, whose teeth engaged in a way that allowed multiplication through repeated addition. Pascal had designed an earlier adding machine. These ideas formed the basis of the arithmometer, but Thomas's first designs were complicated and impractical.

Charles Xavier Thomas de Colmar, a French army officer, began work on his machine in 1818, while he had to perform a great many calculations for the military. He patented the design on 18 November 1820. That first version multiplied by pulling on a ribbon and subtracted through an indirect method; both features were later dropped. The first working example was built by Devrine, a Parisian clockmaker, who took a year and had to alter the patented design considerably. The Société d'encouragement pour l'industrie nationale reviewed it and issued a very positive report on 26 December 1821. The prototype now held by the Smithsonian Institution in Washington comes from this early stage.

Between 1822 and 1844 hardly any machines were built. That gap of 22 years almost exactly matched the period when the British government was funding Charles Babbage's difference engine, which on paper was far more sophisticated but was still unfinished. In 1844 Thomas showed his machine at a French industrial exhibition and received only an honourable mention. Around 1850, as part of a marketing push, he made a few examples in elaborate Boulle marquetry cases and gave them to the crowned heads of Europe.

The decisive change came in 1851. The multiplier was removed, leaving an adding machine with a sliding carriage that let the operator multiply and divide. That year it was shown at the Great Exhibition in London, and industrial production began. Each machine received a serial number; in 1863 a single numbering sequence starting at 500 replaced the older practice, which had given the same number to machines of different capacity. Small faults that appeared with heavy use were also corrected, including a weak carry mechanism and the over-rotation of the wheels when the crank was turned too fast, fixed with a Maltese cross. Banks, insurance companies, observatories and government offices adopted it for its sturdiness. Around 1873 the total passed one thousand machines, a first in the history of calculating machines.

Under Louis Payen and later his widow came improvements such as a removable cover, easier-to-read result windows and a faster re-zeroing mechanism. Clone makers multiplied, mostly in Germany and the United Kingdom, and by the First World War about twenty European companies were making copies. By 1900, however, cheaper and simpler machines such as the comptometer and Burroughs' adding machine had overtaken it in volume. Production stopped in 1915, and Alphonse Darras, who had bought the business that year, could not restart it because of shortages and a lack of skilled workers.

Its operation is simple: a wooden case holds the metal mechanism, with sliders for entering numbers, a lever to choose the operation and a crank that turns the Leibniz wheels. Each turn of the crank adds or subtracts the chosen digit, and the result appears in a window on top. To multiply, the user turns the crank as many times as the units digit of the multiplier, then shifts the carriage to the tens column and repeats.

Its keyboard and accumulator layout was kept in its clones and in the desktop calculators of the twentieth century. The American company Burroughs began in 1886 as the American Arithmometer Company. Though electronics later replaced it, the arithmometer marked the shift from manual to reliable mechanical calculation in offices and research centres.

Timeline

  • 1694Leibniz builds a calculating machine based on the wheel that bears his name.
  • 1820Thomas de Colmar patents the design on 18 November.
  • 1821The French industrial society issues a very favourable report on the first working example.
  • 1851It is shown at the Great Exhibition in London and industrial production begins.
  • 1863A single serial numbering sequence starting at 500 is adopted.
  • 1873More than one thousand machines have been built, a first for this kind of device.
  • 1900Cheaper mechanical calculators overtake it in production volume.
  • 1915Production ends, and Alphonse Darras cannot restart it after the war.

Trivia and anecdotes

A piano-sized calculator

A 30-digit model built for the 1855 Paris Exposition is now part of IBM's collection of mechanical calculators. Jules Verne seems to have been impressed by it, since his novel Paris in the Twentieth Century imagines calculators shaped like huge pianos.

Gifts for European courts

Around 1850 Thomas made a few machines in Boulle marquetry cases and gave them to the rulers of Europe as part of his promotion.

A near-exact gap

The long pause in production between 1822 and 1844 coincided with British funding for Babbage's difference engine, which was also unfinished at that time.

Other milestones in Computing