Larger-than-life figures like Frank Furness tend to accumulate lively anecdotes, and some of them are even true. Furness did indeed amuse himself by firing his Civil War revolver at a stuffed moose's head in the drafting room, and he actually did fantasize, as Frank Lloyd Wright recounted, about gathering together all his clients so he could tell them to go to hell.1 Other such stories are questionable: Did he really have a mistress who “ran a profitable bawdyhouse,” as Lewis Mumford reported, so profitable that she “kept his office going through the depression of 1873”?2 The most outrageous story of all concerns Furness's mighty library for the University of Pennsylvania (Figure 1). Supposedly the university, desperately needing a library building, approached the Pennsylvania Railroad and was given the discarded plans for an unbuilt railroad station. Without hesitation, Furness converted the depot into a library, barely changing the design.
No one has ever paid the slightest attention to that story, for the documentary history is absolutely clear. In April 1887, Furness was appointed to design a new library for the university, which had recently built a new campus in West Philadelphia. He studied the plans for a year, consulting experts and visiting the latest library buildings.3 Plans were approved in July 1888, and construction proceeded for the next two and a half years; in February 1891 the library was formally opened to the public. The building still stands, a stirring performance by Furness at his imaginative best. All this is recounted, admirably and intelligently, in James F. O'Gorman's 1973 monograph on Furness, and augmented by Edward R. Bosley's 1996 monograph on the building.4
Yet the library has always baffled its users. Its components have a quirky restlessness—the weirdly brawny stair tower hulking over a squat, rounded apse—that is more mechanical than architectural. The reading room is unapologetically industrial, with naked brick walls and a mighty iron truss that flaunts its rivets as if they were sequins on a cocktail dress (Figure 2). Those visitors who arrived from Furness's nearby Baltimore & Ohio Railroad Station (1886–87), which had a similar interior, would have found it plausible that the library was a thinly disguised railroad station. And so like all urban legends, this one endured because it seemed to explain satisfactorily what was otherwise inexplicable. The Pennsylvania Railroad anecdote is nonsense, of course. But the other part of the legend, that the library was based on a train station, happens to be true.
Frank Furness (1839–1912) was appointed to design the library by his brother, Horace Howard Furness (1833–1912), the celebrated Shakespearean scholar who headed the building committee. For once nepotism was justified: Furness had already designed two of Philadelphia's three library buildings as well as several private libraries. Early in his career he inserted a panopticon into an abandoned market hall, giving the Mercantile Library of Philadelphia its furiously progressive plan (Figure 3).5 His subsequent building for the Library Company of Philadelphia (1878–79) was history-minded and understandably so, the company having been founded by Benjamin Franklin. The company requested a reverent facsimile of its eighteenth-century building, which Furness provided but handsomely updated with modern materials and ventilation.6 When Horace Howard Furness gave the university commission to his brother, he was awarding it to the most experienced library architect in Philadelphia.
It would take a while for the university to raise the $300,000 it needed to build and endow the library, giving Furness seventeen months to study his plans at leisure. It was his wont, when tackling specialized buildings, to consult expert advisers. For the Pennsylvania Academy of the Fine Arts, for example, he conferred with the sculptor Howard Roberts over the drainage of the modeling room, and with Thomas Eakins for the lighting in the large life-class room.7 Similarly, Dr. Samuel Gross advised on the layout of the surgical amphitheater in the Jefferson Hospital.8 And so for the design of the university library it was natural that he would seek out the nation's foremost university librarians: Justin Winsor (1831–97), librarian of Harvard, and Melvil Dewey (1851–1931), librarian of Columbia.9 Both men were pioneers in developing the modern public library. Winsor introduced the modern book stack (and coined the term), the first example of which was his addition to Harvard's Gore Hall. Dewey, for his part, devised his famous decimal system, America's first modern system for cataloguing books. Each of these innovations would be employed at the University of Pennsylvania Library.
Furness's gave his advisers two chances to criticize his plans. First, he “prepared a preliminary plan, and a conference of eminent librarians from various cities was held at the University.”10 We do not know if any librarians besides Winsor and Dewey attended; they were the only ones who were paid for their trouble.11 Talcott Williams, the newspaper editor and founding director of the Columbia School of Journalism, was also present at their “prolonged consultation,” which lasted into the night.12 Dewey wrote that “the plans I sketched with Mr. Furness late that evening seem to me better than any college library has yet adopted.”13
As a result of that first consultation, “the plans were widely altered in many respects by Mr. Furness.” A second round of consultation followed, this time remotely: Furness's “revised plans were then sent to the expert consultants for final study. A few suggestions only were made, some of which were adopted and others rejected.”14 The final revision came with a switch from the intended site on Thirty-Sixth Street to the ultimate site at Thirty-Fourth and Locust Streets, which required rotating the plan around its axis. Furness submitted his full set of plans and elevations to the university on 2 July 1888.
Neither of Furness's previous libraries had expressed its function on the exterior; the Mercantile Library looked like the former market hall it was, and the Library Company like a Georgian clubhouse. The university commission imposed no conditions, so Furness was now free to design a library that looked like a library; he let himself go and produced a jumble of volumes that was unusually feisty even for him. He chose to work with the “French-Gothic style [which] lends itself to the irregular forms such an edifice must have.”15 Yet if the forms were irregular, it was a principled irregularity that clearly articulated every part of the plan (Figure 4). To the south was the book stack wing, expressed as a glass-and-iron basilica. At the opposite end was the reading room, modeled on the semicircular apse of a French cathedral but facing north, not east, so as to guarantee indirect light. Along its base, from which the side chapels of a cathedral would have radiated, was the ring of private study alcoves—each dedicated not to a particular saint but to a particular subject of scholarship. The hulking stair tower was set at a right angle to the axis of the basilica and apse, like a transept, but one that asserted itself as a detached and independent object. The independence was purposeful: stairs are noisy, and these stairs gave access to the important upstairs spaces, including a lecture hall and a large room for the university's unique Assyrian collection. The sound of footsteps should not penetrate the main reading room. The theme of detached volumes asserting themselves was replicated yet again in the entrance porch, which pried itself loose from the main mass of the building as a freestanding pavilion.
Just as the library distinguished its functional components on the exterior, the interior drew distinctions between different types of reading. Furness's cardinal insight, as stated in the Library Journal, was simple yet profound: “Readers differ.”
There is the casual reader who comes in for five minutes to see a magazine; students—in a college library—who come often by groups and with an instructor to consult the works on a special subject; the reader who is hastily looking up on a topic at a moment's notice or refreshing his memory by running over a pile of related works, a job of two or three hours; and the investigator who has come for a single day's work and wants quiet and space for that day.16
Instead of following “the customary plan [which] is to put all these readers in the same place to the loss of all,” Furness provided each of these categories of readers with a different type of space. The loudest part of the library was the delivery room, where books retrieved from the stacks were delivered to readers, and where a certain amount of whispered conversation was necessary. Here Furness placed the first reading room, outfitted with newspapers and journals, “a reading-room open to all the casual world, 40 × 48 [feet].” The next reading room was screened by two colossal arches, coming nearly to the ground so as to shut out any noise from the first reading room; this was the “reading room for the student and book-worker, 40 × 54 [feet].” Farthest to the north, and farthest from the entrance, were the six study alcoves with their sliding doors, “each a separate room with a strong light in which a professor can gather eight or ten students without disturbing the rest of the room.” Each class of reader, which Furness neatly summarized as “casual, topic-searcher, and student,” could find the right balance between convenience and solitude.17
One refinement of the library shows the hand of one of Furness's consultants, Melvil Dewey, whose great cause was the rational cataloguing of books. He would have had a role in the ingenious arrangement of the card catalogue, which was installed in the arcade between the cataloguing room and the southern reading room. The cases were designed so “that they may be pulled out in either direction, and can thus be consulted as conveniently by the clerks in the cataloguing room as by the readers in the reading room.”18
Because the stacks were closed and could not be browsed, the users of the library would not see its most innovative space. The stacks were entirely of iron and glass—the iron frame carrying both the books and the translucent glass floors, which were clear enough to admit diffused daylight through three stories.19 The effect, contemporary visitors noted, was staggering: “You almost fancy yourself in a glass palace, for, wherever one looks, the ceiling, the floor, and the upper part of the walls, one sees nothing but glass.”20
The modular construction of the book stack wing meant that it could be as long or short as the university required. Furness seized on this principle to propose that the south façade should be movable, so as to expand like an accordion as books were acquired. Here at last was an answer to that chronic problem of library design, how to provide for expansion.
The plan of the stack admits of indefinite expansion to the south by extending the stack a bay at a time, the end wall being moved out on jack-screws. The cost of adding a single bay, when only roof, sides, and shelving have to be estimated for, will not be over 30 cents a volume for the additional space. I need not enlarge upon the priceless value in a library of what I might call a high coefficient of expansion.21
Thirty cents per volume—this was library planning of the most utilitarian sort, expressed in terms of unit cost. As built, the library had only three bays of stacks, although the published perspective optimistically showed eight, marching stalwartly southward. The initial capacity was 500,000 volumes; Furness calculated that an extension would easily raise it to a million.
In the end, Furness's expandable wall was never put into use. Shortly after his death, the university built the Duhring Wing across the south end of the library, blocking any future extension. By this time the idea of extending the stack seemed an example of Furness's charming but willful eccentricity, and no one took it seriously. Nor have subsequent scholars. Vincent Scully, an early champion of Furness, had little to say about the building's plan and instead concentrated on its tectonic character, which “echoes the work of H. H. Richardson and marks the beginning of the submergence of Furness's own bodily active style in favor of the spatial volumes and continuous surfaces of Richardson's work.”22 O'Gorman mentions the expandable stack, but without comment. More recently, Bosley calls the stack “ingenious” but says nothing about its source. None have attributed the form to Furness's design consultants, whose roles have been, if anything, downplayed. Bosley proposes that “Winsor and Dewey contributed less to the final plan of the library than has been previously estimated.”23 In this account, the library had already been planned in all its essentials before Furness met with his consultants; their contributions would have been limited to minor technical suggestions. In actuality they contributed more than trifles; in fact, Winsor contributed the central idea of the building.
In 1879, the American Library Association held its second annual meeting, which took place in Cambridge, Massachusetts. After the customary program of presentations (e.g., “Sensational Fiction in Public Libraries,” “Insect Pests in Libraries”), the discussion turned on the third day to the topic of library architecture.24 Attending the meeting with Winsor was Henry Van Brunt, who had designed his book stack wing addition to Gore Hall just two years earlier. Raising the question of effective library planning, Winsor made a claim that must have startled his listeners: the most functional library plan in America was not a library plan at all—rather, it was the plan for the Boston & Providence Railroad Station, built a few years earlier by Peabody & Stearns. Only a few minor modifications would be necessary to turn it into an efficient and attractive library:
The main body or head house, over 200 feet long, I would convert into the public and official working quarters of the library; this is two stories high, the lower one lofty. What is now the main waiting hall, lighted by a lanterned roof, would serve for the grand delivery room. On either side are reading rooms, large and small, to be fitted up in detail as required. Over these, the official quarters and work-rooms, all opening on a balcony, indicated by the dotted line, and which overlooks the grand hall, and communicates with the stack room or train-house. The last extension, of which you see the section, and which may be continued indefinitely, is, as built, about 600 feet long and 130 feet wide, external dimensions.25
Winsor was not teasing his listeners; he was quite earnest. In the next issue of the Library Journal he published his drawing for the library station (Figure 5).
Here, then, were all the elements of Furness's plan in schematic form: the neat division between a head house (with reading rooms, cataloguing, and offices) and a storage shed to the rear, with the delivery desk at their point of contact; the provision of separate reading rooms of different sizes; the roofing of the book stacks entirely in glass; and finally, and most crucially, the idea that the stack room could be “extended indefinitely.” This was the phrase that electrified Furness and his client, and it showed up repeatedly when they spoke of their building. William Pepper, the university provost, boasted in his annual report that “the book stack admits of indefinite extension”; Talcott Williams, writing in 1888 in the Library Journal, repeated that “the plan of the stack admits of indefinite expansion.”26 Far from treating expansion as a quirk of Furness's library, these authors saw it as a fundamental element.
Setting Furness's plan beside Winsor's, one sees the essential continuity. The only significant departures are motivated by the somewhat different program, particularly the upstairs lecture room that necessitated the stair tower. Further, unconstrained by the shell of an existing building and free to orient his building however he chose, Furness was able to exploit the potential of indirect northern light with his semicircular apse. Otherwise Furness's plan is the realization of Winsor's ideal library, as adapted to a specific site and a modified program. Winsor had certainly earned his forty-dollar honorarium.
It was a peculiarly modern idea to base the book stacks of a library on a train shed. Because of the nature of locomotives, which produced both sparks and smoke, these sheds needed to be absolutely fireproof and also brightly lit. These requirements called forth the most advanced engineering and pushed iron and glass to their physical limits. Furness was deeply engaged with railroad architecture and engineering, having worked for three railroad companies during his career and having designed well over a hundred depots and stations.27 His own Broad Street Station (Philadelphia, 1892–94) would itself contain the world's largest train shed, with an uninterrupted span of 301 feet.28
The train shed was modern in one more respect. If the front of the station projected an image of reassuring stability in the form of a temple or palazzo, the shed to the rear had no façade, and its active part was an open portal through which objects moved, its outbound tracks implying limitless spatial freedom. Applied to library architecture, this had the greatest of symbolic ramifications. A library with moving parts, a library that actually grew, was not a serene temple of knowledge—it was an engine of active thought. Furness would have found the idea congenial. Books were a familiar and natural component of his life; his father wrote them, his brother edited them, and his sister translated them.29 Furness himself, who disappointed his family by not going to Harvard as his father and brother did, was the least bookish member of the family. One of his draftsman noted that Furness owned a copy of Eugène-Emmanuel Viollet-le-Duc's Dictionnaire raisonné de l'architecture and that “it was one of the few books on architecture he admired.”30 It was characteristically perverse of Furness, who never learned French, to praise a book that he could not read.
Yet precisely because books were such a familiar and natural component of Furness's life, he did not sentimentalize them. When his contemporaries set about designing libraries, they tended to treat them reverently, as if they were reliquaries for sacred objects. But Furness knew that books were active, useful, even incendiary things; after all, the abolitionist tracts that his father, the Reverend William Henry Furness, wrote before the Civil War had brought him repeated death threats.
And so, Furness's library was no temple, palace, or warehouse for books; it treated the act of reading as an active and dynamic enterprise. Perhaps that accounts for the unusual prominence of the looming stair tower, which is the pendant to the expandable stack and serves as the formal face of the building. And like a face, it wears an expression, precisely the one befitting a library—the deeply scored voussoirs above its oversized arch suggesting a stupendous brow, furrowed in thought.31
There is a coda to the story. Where did Winsor get the idea in 1879 for turning a train station into a library? No American public library of the era bore the slightest resemblance to a train station—with one exception. This was the Mercantile Library of Philadelphia, the curious panopticon structure that was Furness's first major work, inspired by Benjamin Delessert's celebrated 1835 plan for a panopticon library (Figure 6).32 Its open book shelves represented Furness's first foray into rational library planning, based in this case on a hardheaded calculation of the relative costs of book loss and staffing a full-time delivery desk. The result was that members and visitors were given free access to the books: “The amount of loss to the property of the Company is found to be very inconsiderable under this system—far less than would be represented by the expense of hiring clerks to wait upon the public.”33
The Mercantile Library was an alteration of an existing building that began life in 1859 as the Franklin Market House. The architect was John McArthur Jr. (1823–90), who designed it as a great barn of a space, measuring 187 by 74 feet. To give it its clear span, he used an arched truss of wood with iron ties. This was a particular type of truss that had been developed by the Philadelphia, Wilmington and Baltimore Railroad, which used it most spectacularly in the 150-foot span of its Prime Street Station in Philadelphia (Figure 7).34 Anyone entering the Franklin Market would have been instantly reminded of the station. The same thought occurred to the directors of the Pennsylvania Railroad, who, after the market house failed during the Civil War, saw the building as a ready-made passenger depot and bought it in 1864.35
From market house to would-be railroad depot to library: all were public functions that required light and space, and Furness had no reason to change the building's system of boldly exposed trusses when he converted it into his modern library. This is what Justin Winsor would have seen when he came to Philadelphia in early 1871 “looking into Libraries, library matters, etc.” Winsor was then director of the Boston Public Library, and he came to inspect the prospective designs for Philadelphia's new Rush Library. He was underwhelmed; he saw “three plans by three of the best architects in that city for the Rush Library, not one of which was worth a cent, not one of the architects having any practical knowledge of what was needed.”36
It would be poetic justice if Furness were one of those three architects whose ideas were not worth a cent, for Winsor took back with him to Boston a tantalizing idea that Furness eventually made a reality: a library that looked like a railroad depot. Eight years later, looking at the Boston & Providence Station, Winsor saw a railroad depot that looked like a library. Another eight years later, in 1887, he could persuade Furness of the rightness of the idea—an idea that may well have come from Furness's own Mercantile Library.
It is no great archival coup to unearth the drawing by Winsor that inspired the University of Pennsylvania Library; Carroll L. V. Meeks reprinted it in 1956 in The Railroad Station: An Architectural History, although no scholar thought at the time to connect it to Furness's library.37 The real mystery is why something that was common knowledge in the 1880s could have so quickly been forgotten, to survive only as a garbled urban legend. The answer may have to do with the library's swift collapse into unfashionability. Within a month of the laying of its cornerstone, the cornerstone for the Boston Public Library was also laid. This was McKim, Mead & White's first great public building, and it would help set the standard for civic grandeur and library design nationwide. It did not matter that it cost ten times as much as Furness's library (while storing just twice as many books), or that with its single reading room, as Talcott Williams explained, “its plan fails to discriminate between different classes of readers.”38 By comparison, all that now mattered about Furness's library was its vulgarity, “the determination to be noticed at all costs and all risks,” as Montgomery Schuyler, the dean of architectural criticism, put it in 1910.39 Already by 1908, the prestigious Architectural Record would write off Furness's library as a “‘fortified greenhouse,’ than which nothing more grotesque could be imagined.”40
Today we can once again understand the rational and systematic thought that went into the making of Furness's great library. But a century ago the notion that such a garish performance was rational defied belief. No wonder that Charles McKim, when asked about Furness's library, responded with comments that were “unfavorable and unquotable.”41