Aug. 15, 2011
by Mark Ollig
There was something very special about the 1890 US census.
For the first time, census results were not hand-counted using simple tallying devices like the ones Charles Seaton invented for use during the 1870 and 1880 census.
The 1890 census results were counted using an electric tabulating machine.
Herman Hollerith of New York worked for the US Census office during the 1880 census as a statistician, when they were still tabulating census results by hand counting.
It was during this time when Hollerith decided on creating an improved method of counting the census results.
I might have, too; according to the US Census webpage, the 1880 census ended up taking seven years to complete.
Hollerith consulted with his mentor, Dr. John Shaw Billings, a statistics supervisor for the US Census. Billings suggested somehow mechanically tabulating the census results using coded cards with punched holes similar to the cards used on a device called a Jacquard handloom, which is used in textile processing.
Hollerith decided to go with a punched card system. The presence or absence of a hole in the card would indicate a specific type of data characteristic. This idea came to him while observing how railroad officials would identify seated passenger characteristics using “punch photograph cards.”
Hollerith invented an electric machine using circuit-controlling indexing points and made the location holes on each punch card indexed for collecting specific individual statistical information, cross-tabulations, and number totals.
Hollerith applied for a US Patent Sept. 23, 1884.
Jan. 8, 1889, Herman Hollerith was awarded US Patent number 395,782 titled “Art Of Compiling Statistics.”
He named his tabulating device, the Hollerith Census Machine.
Hollerith also invented devices which punched, read, and sorted card tallying data.
The Hollerith electric tabulating machine sorted census returns by completing an electrical circuit wherever a hole was located on a punched card.
The US Census office, after testing other tabulating methods, awarded Hollerith a contract for tabulating the upcoming 1890 census and paid $750,000 for the lease of his machines.
The tabulating machine could process almost 10 times the number of census data than a human census clerk could via hand counting. This greatly reduced processing time and saved millions of dollars.
The operator of the tabulating machine would place each card in the reader, pull down a lever, and remove the card after each punched hole was counted.
The keyboard punch template was of a pantographic design, which quickly transferred data from the census taker’s sheet to a punched card.
The tabulation results were displayed on clock-like dials located above the tabulating machine where a sitting clerk would be working. This tabulating desk system looked much like a vintage operator telephone switchboard (minus the cords).
The 1890 census included more detailed information than the 1880 census; also, the 1890 census population count was 25 percent higher.
Hollerith’s electric tabulating census machines finished the 1890 census much sooner than the 1880 census had been completed, and saved an estimated $5 million.
According to the US Census Bureau, the US population in 1890 was 62,622,250.
After the 1890 census, Hollerith was approached with contracts by foreign governments and railroad companies wanting to use his new electrical tabulating machines.
In 1896, Hollerith founded the Tabulating Machine Company, in Washington, DC.
His new company provided the tabulating machines used for the 1900 US census, and being Hollerith had a monopoly on the electric tabulating machine business, he was able to ask for (and receive), a great sum of money for leasing his machines to the US Census Bureau.
The good times almost came to an end for Hollerith’s company when employees within the US Census Bureau created their own electric tabulating machine – processing results faster and at a lower cost than Hollerith’s machine.
This new electric tabulating machine was used during the 1910 US Census.
In 1911, James L. Powers, a US Census Bureau technician, obtained the patent for this new tabulating device and started his own tabulating machine business.
During the same year, Hollerith merged his company with four other companies and renamed it the Computer Tabulating Recording Company; however, it almost went out of business.
In 1914, Thomas J. Watson, Sr. came to work for Hollerith’s new company and became an executive and its general manager.
Watson revolutionized how the Computing Tabulating Recording Company was operated, reestablishing the business and turning it into a very successful operation.
Herman Hollerith continued working as a consulting engineer for the company until he retired in 1921.
He went on to raise Guernsey cattle on his farm in the countryside of Maryland until his death at the age of 69, Nov. 17, 1929.
You can view Herman Hollerith’s US Patent at http://tinyurl.com/3vn7ydv.
To see how the tallying clerks used his census machine, go to http://tinyurl.com/4xvkodh.
A color photograph of the Hollerith Census Machine can be seen at http://tinyurl.com/3vng6vv.
Three years after Hollerith’s retirement, the Computing Tabulating Recording Company changed its name to the International Business Machines Corporation, or what is commonly known today as IBM.