Alicia Meyer, Curator of Research Services, Kislak Center for Special Collections

aliciajmeyer

In 1970, researchers at the Society for Data Educators set out to find the first computer at an educational institution. The society’s founder, Enoch Haga, wrote to the University of Pennsylvania archivist claiming that the “first American difference engine,” and therefore, “the first ‘computer’ to be owned by an American university,” was at Penn (UPP 10 no. 4).

A difference engine was an early calculating machine invented in the 1820s by the English mathematicians Charles Babbage and Ada Lovelace. It used thousands of metal pieces and finite differences to solve polynomial equations, eliminating the need for people to calculate them. In his letter, Haga claimed that in the 1870s, the American inventor, George B. Grant, designed one such machine at the behest of Fairman Rogers, a Professor of Engineering at Penn. Haga’s letter enclosed a photograph of the machine elevated on iron A-frame legs in an exhibit hall with the added caption Grant’s Calculating Engine at the 1876 Centennial Exhibit in Philadelphia. He said that the machine was first displayed at the Machinery Hall at the World’s Fair in Fairmount Park before coming to Penn’s campus.

“Grant’s Calculating Engine” photograph enclosed with letter, Enoch Haga to Leonidas Dodson, Archivist. 1970, University of Pennsylvania Archives and Records Center, UPP 10 no. 4, “Instruments and Apparatus of U of P”.

Penn, however, lacked any clear record that Grant’s machine had ever crossed the Schuylkill. The archivist searched the University’s early account books looking for signs of the machine. While they found a mention of an “indicator” in 1883, which sounded a bit like a similar machine, there was no conclusive proof that Grant’s calculator had made it to campus (UPP 10 no. 4). The omission of Grant’s engine in the University record was surprising. Not only was the machine historically significant and therefore worthy of preservation, but it was also so large that it was hard to believe that it could have been misplaced, forgotten, or disposed of without a trace. Grant’s Difference Engine was comprised of 15,000 pieces, stood 5ft by 8ft, and weighed 2,000 pounds. It also cost Professor Rogers $10,000, an extraordinary amount of money to expend in the 1870s on such a device.

Hopeful that the machine had survived, Haga asked the archivist to search again, wanting “to preclude the possibility of the engine being unnoticed in some dusty storeroom or basement, long forgotten.” The Charles Babbage Society, a Palo Alto group working with the Society for Data Educators echoed Haga’s call when they wrote in their bulletin “Let us hope that this historically valuable machine can be located; now that we know it was actually built; its discovery would add tangible evidence to its importance as a transitional link in the transformation of mechanical calculators into modern computers” (UPP 10 no. 4). An internal note at the University suggested that if it had been on campus, “it may have been stored in the basement of Randal Morgan Laboratory,” not, as others had suspected, the Moore Building, which had housed the construction of the ENIAC years later (UPP 10 no. 4). However, the machine was never found, and the mystery of the 2,000-pound computer was soon forgotten – forgotten until a recently rediscovered set of archival materials offers new insight into this moment in the history of technology.

A smallresearch collection by Professor Rogers suggests that Haga was correct; a difference engine was acquired and installed at the University of Pennsylvania after it was exhibited at the World’s Fair. The collection sheds light on Rogers’ efforts to procure the machine and, eventually, to dispose of it.

Fairman Rogers (1833-1900), in Horace Howard Furness’s F. R.: 1833-1900. Privately printed, 1903, Kislak Center for Special Collections, Furness Collection.

Fairman Rogers was the son of a Philadelphia iron merchant who had a lifelong penchant for machinery. His family had made a fortune selling hardware at the advent of the Industrial Revolution. When Fairman came into his fortune, he acted as a patron to several projects that enmeshed his shared interest in the arts and engineering. For instance, at the University of Pennsylvania, he was a supporter of Eadweard Muybridge’s photographic studies of horse motion, which were featured in his work, Animal Locomotion. Rogers was even credited with the idea of using the zootrope to animate Muybridge’s pictures into motion. Rogers also served as patron for Thomas Eakins and was depicted in Eakins’ painting “Sunday Morning in the Park, or Fairman Rogers Four-in-hand,” an oil painting that, like Muybridge’s work, sought to trick the eye into seeing a horse in motion. At the end of his life, Rogers was memorialized by his brother-in-law, Horace Howard Furness, for his love of machines. Furness writes that “One of the first typewriters, if not the very first, was set up by its inventor in Professor Rogers’s library” (Furness).

In the 1870s, Rogers was a board member of Philadelphia’s committee to host the World’s Fair. Designed to showcase American industrial innovation, culture, and progress on the 100th anniversary of the nation’s founding, Rogers took a particular interest in the machinery building. His records indicate that he purchased two machines designed by Grant for exhibition. The first was the 2,000lb “Grant’s Difference Engine.” The second was the much smaller “Grant’s Calculating Machine,” which stood 1ft by 1ft, weighed around twenty pounds, and consisted of approximately 400 pieces.

“Grant’s Difference Engine”, Kislak Center, Ms. Coll. 1678. “Grant’s Analytical Engine”, Kislak Center, Ms. Coll. 1678.

Like Babbage and Lovelace’s earlier models, Grant’s Difference Engine could solve an equation and then print out a mathematical table using a built-in printer. A placard saved by Rogers explains that the machine “computes the terms of any such table,” and then it “prepares a mould of them stamped in wax, from which an electrotyped plate is made ready for the press” (Ms. Coll. 1678 f.3). Before screens, a built-in printer was how mathematical results were viewed and shared.

Placard for Grant’s Difference Engine and Calculating Machine, Fairman Rogers collection of George B. Grant Difference Engine material, Kislak Center, Ms. Coll. 1678 folder 3.

The correspondence between Grant and Rogers reveals their collaborative efforts not only to build and use the difference engine but also to find ways to reduce its cost. In 1881, a letter from astronomer Hubert A. Newton to Rogers indicates that the cost of creating such machines could never be recouped. He writes: “As a money venture, of course, no one can hope to make an issue of such a table repay expenses. But I think it would be a contribution to science” (Ms. Coll. 1678 f.2). However, by 1881, Rogers and Grant were increasingly finding ways to create a more affordable model, not only by reducing the size and number of pieces in the machine, but also by reimagining the built-in printing device.

Just two years after the World’s Fair, Grant wrote to Rogers explaining that he had completed a “new difference engine” that “is nothing whatever but a series of calculating machines placed side by side. The printing apparatus would be the same as on the old machine with small changes.” A year later, he writes that “I have studied the matter so thoroughly that I am quite certain that the element for the machine as thus reduced can be made for $500.00 A good framing with gearing and all apparatus necessary for manipulation by hand could be made for $300.” He was also beginning to experiment with the idea of replacing the wax mold and electroplate device with an ink ribbon. Indeed, by 1881, that was the decision: “The cheapest plan” Grant states, “seem to be to print directly from the calculating wheels by means of inked ribbons, the wheels being rounded with raised type.” In retiring the wax-mold-to-electroplate device, Grant argued that is was also time to retire the “old machine.” Grant writes to Rogers, “I think the old machine should not be preserved any longer. It is in the way and can in no way confer with the new one: With some permission, I will get what I can for the old metal in it and dispose of it finally.” By November of that year, Rogers received payment for the scrap metal (Ms. Coll. 1678 f.2).

It has been over fifty years since Haga went looking for Grant’s Difference Engine on Penn’s Campus. While the machine was here for a while, within a decade, innovation had driven its creator to consider it obsolete and ready for the scrap yard. However, among Rogers’s papers is another scrap, a tiny strip of paper pressed with a column of four-digit numbers pressed by an electrotype plate. The first three numbers of the slip are “1111,” “2222,” and “3333” (Ms. Coll. 1678 f.2). The following numbers have no discernible connection to one another, suggesting this slip may have been used as a test of the plate-making process. It may be that this scrap is all that remains of the “old machine.”

Scrap, Fairman Rogers collection of George B. Grant Difference Engine material, Kislak Center for Special Collections, Ms. Coll. 1678 folder 2.