The Twelfth Census

In 1890, the twelfth census used Hollerith's machine. The Census Bureau employed almost 50,000 people to collect and tabulate data.² Two hundred and thirty-five points of inquiry, including sex, marriage status, no of children, and native tongue, were investigated; a forty-fold increase from the six counted in 1870 census.²

Pantograph

Hollerith's machines were amazingly reliable. Most of the malfunctions resulted from the removal of mecury from the system by clerks who were seeking a break.2 Using the Pantograph, which Hollerith had designed around the start of the census, clerks were able to punch one of corresponding 240 holes into the cards with the touch of the keyboard.

Earlier, Hollerith was using a small hand-held conductor's punch to place data on the stiff cards. The punch was useful in puncturing regular paper, such as railroad tickets, but in collecting massive amounts of data on a heavy paper, the hand-held punch was inadequate. Hollerith's mother-in-law describes:

Mr. Hollerith came over from Baltimore Friday. He is completely tired out. He has been punching cards at a rate of 1000 per day -- and each card has at least a dozen holes. He has done it all with a hand punch and his arm is aching and paining dreadfully. He really looked quite badly.²

With the new tool, each clerk would process 700 cards a day on average. After cards were processed by the clerks, tabulators took over. At the peak of processing, tabulators counted between 2,000 and 3,000 families a day. In a single day, over six million people were counted by Hollerith's machines.²