On a cold March 11, 1809, Thomas Jefferson paid the ferryman $1 to take him and his carriage across the Potomac River at Georgetown and headed south toward retirement. What he left behind at the President’s House were unfulfilled dreams of remodeling the still-unfinished mansion and completing its partly built domestic service wings, which were entirely his idea. It is ironic, in retrospect, that these wings, with their zigzag roofs and flat terrace platforms, would become his physical heritage there, because they have been mostly forgotten.

If the White House is, as the historian William Seale has written, “an American Idea,”1 it includes one of Jefferson’s most tenacious architectural ideas— domestic service wings. Jefferson did not invent the concept but borrowed it from those seen attached to Renaissance villas in Andrea Palladio’s Four Books of Architecture (1570). Domestic service wings appear in Jefferson’s earliest drawings for his bookish Palladian- style home, Monticello, built in the early 1770s. Placing domestic outbuilding functions in wings gave order to what would have been the typical scattered backyard arrangement of necessary buildings; more important for Jefferson, it saved the space for ornamental landscaped gardens. He adapted the same ideal for the White House.

Urban houses of pretense needed the same domes- tic services as large country houses. In America those functions were squeezed into backyard spaces as connected or detached buildings, while in Europe they filled the lowest floor and continued into connected wings or were separated and grouped as service courts. The overall service requirements of the White House amplified those of even large urban American houses on small lots like the ones George Washington and Jefferson had used in Philadelphia. Pierre Charles L’Enfant, in his famous plan for the Federal City, had, in fact, indicated a palace-derived solution for the White House with greatly extended wings and terraced ensembles. Washington, however, especially liked James Hoban’s Irish Georgian house design and declared it the winner of the architectural competition for the President’s House in 1792.2 Historians have indicated that Washington and Hoban actually discussed wings on the house, but details are not known and Hoban confined the immediate service needs to the basement story of the White House.3 He probably rationalized that typical food-related functions could be supplemented through daily trips to the local meat and produce market, as indeed happened.

When President John Adams arrived in Washington to move into the White House in november 1800, there were no separate service structures except for simple brick stables two blocks away; the grounds held only workers’ sheds. It was up to him and Mrs. Adams to make the unfinished brick and stone interior shell habitable. The basement, containing work and chamber spaces for cooks, housekeepers, and servants, probably seemed more cozy and finished than the principal rooms above. Adams’s addition to the house consisted of rickety wooden stairs to a scabbed-on wooden balcony that gave the public unintended access to the south entrance of the house. When Jefferson replaced John Adams as president in 1801, he quietly slipped into a shell of a house still reeking of fresh plaster. His creative design propensities must have been stoked as never before.

White House Remodeling

Thomas Jefferson never resided in any house very long without altering it to suit his ever-developing taste for comfort and convenience. The incomplete interior of the White House presented him with a remodeling opportunity he had already practiced in Williamsburg, Paris, new York, and Philadelphia.4 This practice on landlords’ houses later paid dividends when he remodeled Monticello, constructed his idealistic villa retreat Poplar Forest in Bedford Country, Virginia, and eventually created his largest work, the University of Virginia in Charlottesville.5 While his remodeling skill bore mostly unrealized fruit in his interior plans for the White House, the two service wings he added established an important link to his three most personal projects and represent an enduring experiment in architecture.

Probably even before Jefferson took up residence in the White House, his creative mind whirred with ideas of how to tweak the large Georgian pile. What he had been given as a starting point was a house of generous size with domestic services confined to the basement story and flanking government buildings about 500 feet to the east and west. The mostly barren shell vacated by John and Abigail Adams was a tabula rasa, as were the grounds around the house, facilitating the amateur architect’s fertile imagination to produce remodeling solutions to the former and initial plans for the latter.

To Jefferson’s dismay, domestic necessities such as shelters for fowl, goats, and cows had begun to supplant the workers’ shanties that had sprung up around the house during construction. Jefferson found other needs wanting. The stables were two blocks distant; guests and servants shared an exterior privy; keeping ice was difficult without an ice house; the basement offered inadequate storage for wood and coal; and other functions such as a hen house and smokehouse were needed and best placed outside the basement. Jefferson’s solution to put these functions in wings developed from his use of attached Palladian-style service wings at Monticello. To accomplish this would take time, congressional budget approval, and a capable construction superintendent. The latter job fell to Benjamin Henry Latrobe as surveyor of the public buildings and his assistant John Lenthall.6

Jefferson had corresponded with Latrobe since 1798, most likely meeting him in Philadelphia about that time, and had first hired him in 1802 to study the Tiber Creek Canal between the White House and the Capitol and then to design a dry dock for the fledgling U.S. navy on the Anacostia River at the base of Capitol Hill. Having confirmed Latrobe’s skills as the most accomplished, and arguably the only, professional architect and engineer in America, Jefferson hired him in March 1803 as surveyor of the public buildings, a position he held until 1811.

Latrobe’s initial task at the White House was to replace a leaking roof of slate embedded in mortar and gutters that were letting buckets of water into the house, threatening to destroy the few newly installed architectural finishes. Jefferson also called upon Latrobe to fix the two new water closets on the Second Floor that were fed from rain-collecting cisterns in the attic. Jefferson redesigned the grand, but unbuilt, west stair- case that Latrobe would construct off-axis in order to open the vista and public access to the terrace promenade once there was a wing built to support it. A similar glass door was retrofitted in the similar eastern Serliana or venetian window for east wing terrace access. Jefferson’s other interior plans to fashion a French three- room chamber and study suite on the First Floor were never accomplished but are shown on plans by Latrobe.7 What Jefferson focused on and pursued to at least par- tial completion were the domestic service or “office” wings and their intended connection to the flanking federal buildings.

Jefferson began remodeling with the easiest, yet important, projects related to convenience and service: replacing the outdoor privy with two indoor water closets from Philadelphia; establishing a cooled wine cellar; hanging service bells throughout the house; upgrading the kitchen for his French chef by installing stew stoves, boilers, and ranges; and having the unsightly and dangerous south stairs removed in favor of a bridge-like entrance on the north side of the house, which had been intended for public access. Jefferson did keep Adams’s scabbed-on wooden balcony as a temporary outdoor room from which to view the distant Potomac basin’s wilderness, quickly disappearing beyond the already denuded White House grounds. Later he could enjoy
the outdoors on his own terraces, but the view would remain less than scenic.

The Domestic Service Wings

Jefferson’s solution to the deficiencies of the White House came from his own house rather than from the sophisticated Paris town houses where he resided in the 1780s. His plan drawing of c. 1804 shows what he wanted (illustrations 16, 17). These drawings were first published by the architect and architectural historian Fiske Kimball in Thomas Jefferson, Architect (1916), a monumental work that established Jefferson as an accomplished self-trained architect in addition to his myriad other capabilities. onto each side of the White House he proposed attached service wings, commonly called “offices.” These wings, partially set into the grade on their north or public side, would expand east and west as needed, or as funded, until they joined the Treasury Department building on the east and the War Department building on the west (illustration 15). The common thread visibly connecting these segments on the south would be a Tuscan order colonnade that provided a covered walkway. Jefferson’s drawing also shows a 100 foot section of parallel wings to the south of the main block at the east and west ends, intended for government clerks’ offices for the Treasury and War Departments, to which they connected. Latrobe’s advocacy of fireproof construction influenced Jefferson to designate the extremities of the wings, opposite the span of proposed clerks’ offices, for fireproof storage rooms for each department. This lateral expansion to each side of the White House served to visibly connect the three existing federal buildings in an American way, stretched out horizontally in the wide-open space.8 The grouping of services in these wings left the large, ungraded expanse south of the house for Jefferson’s private landscape mixture of formal and picturesque features (illustrations 11, 12).

Kimball’s book also depicted Monticello’s prototypical north and south colonnaded wings. The difference in function between these and the White House wings had to do with what was already housed in the basements and the scope and function of each house. The common denominator was service. Monticello’s daily routine and its economy depended upon enslaved servants. Jefferson’s preference at the White House was for paid servants.9 Monticello’s wings captured most of the economic and domestic functions of a large Virginia plantation that ordinarily existed as separate buildings arranged hierarchically in the surrounding and distant landscape. Jefferson’s imitation of Palladio’s domestic service wings reserved the surrounding grounds for picturesque pleasure gardens as well as making the linked and covered buildings convenient.

The White House wing plans can be scaled for size because of a drafting convention Jefferson learned in Paris. The famous antiquarian architect Charles-Louis Clérisseau taught Jefferson the architectural drafting advantages of using pencil on carpetmakers’ point paper. Jefferson thereafter used this ruled graph paper (which became standard for architects in the twentieth century) as a visible scale in architectural drawing. The paper consisted of a grid of large red lines containing ten small squares between them. Typically, for Jefferson, each small square equaled 1 foot. Thus this “decimal” paper required no scale rule or written dimensions for understanding scaled size. By scaling Jefferson’s White House wing plan with this method, one can see that he is placing the intended columns on 10 foot centers with door and window openings centered between them. The interior room wall divisions are also centered on the columns, with some larger rooms necessarily being a double or a multiple of the 10 foot module. The first sections of the wings, separated 20 feet from the main house wall, were meant to be 24 feet wide and 150 feet long. The plan shows partially penciled walls, indicating the subsequent extension to the adjacent federal buildings and the parallel row of clerks’ offices. Jefferson’s precision on the graph paper can be seen in the thickness of the brick walls dividing each room, which are shown to be about 6 inches less than two squares, or about 18 inches, being the standard size of a 2 wythe thick wall using standard 81⁄2 inch long handmade brick with a mortar joint of about 1⁄2 inch.10 In this scaled method, the room sizes are as follows: on the west the wine cellar was 18 feet in diameter, the wood room with coal cellar, 8 feet wide; the necessary, 9 feet; the saddle room, 81⁄2 feet; the servant’s room 18 feet; and the coach house 58 feet. on the east the meat house with vault below was 131⁄2 feet; the cellar stairs, 31⁄2 feet; the necessary, 8 feet; the servant’s room, 18 feet; the hen house, 18 feet; and the stables, 58 feet.

Jefferson’s first addition was neither in the house nor in the wings but an exterior ice house constructed just west of the house in 1801. Some functions might wait, but those related to the quality of food and drink could not! This 18 foot round by 16 foot deep structure was not unlike the ice house constructed as part of the north wing at Monticello in 1802. Jefferson’s memorandum book notes that he paid for filling the White House ice house in 1802 and for carpentry work by John Lennox the same year.11 The work by Lennox could be for any or all of the tasks of constructing the ice house roof, for an internal platform, or for an enclosure that appeared by 1803. The ice house was then joined to the west side of the basement level by what must have been a simple frame structure that also served to shelter the preexisting well just west of the house.

This was more than simply an ice house, however. Having no deep cellars in the basement posed a problem for keeping expensive drink at cool temperatures.

Apparently the subject became one of anecdote, as Sir Augustus John Foster mentioned in his travel memoir that the ice house–wine cellar had been occasioned after President Jefferson experienced “great losses in wine” from inadequate storage in the basement.12 Latrobe also gossiped of Jefferson’s losses when, in a letter to William Lee, he shared his opinion that, due to the absence of an original underground cellar, “Mr. Jefferson lost 800 bottles of Crab cider for want of one” and that “a good provision [for a] cider and beer cellar never existed in the house.”13 To prevent further loss of pre- cious liquors, Jefferson had a platform fitted out inside the ice house as a sort of wine cellar room that had impressed Foster enough to remark on its temperature relative to the heat outside. Latrobe lamented that “the wine cellar in the West wing is fit for nothing but wine” but that the “present kitchen will admirably supply the deficiency [for beer and cider].”14

Digging for the west wing began in the summer
of 1804 under the direction of architect and engineer Latrobe and his assistant and construction supervisor Lenthall, the same team working on the Capitol construction under Jefferson’s supervision. Jefferson had chosen the right man to undertake the most complex of American building projects. While the completion of the Capitol and the President’s House was his primary responsibility, Latrobe found that he first needed to supervise emergency rebuilding on both buildings.15 He had promptly engaged builder John Lenthall as his assistant, beginning an interesting correspondence among the three regarding both building projects, with Lenthall serving as an outlet for Latrobe’s frequent frustrations with the architect-president.

Latrobe and Jefferson found themselves intellectual and architectural soul mates of a sort, while Latrobe and Lenthall bonded over details of building construction. In the beginning, at least, Latrobe was taken with Jefferson, writing home to his wife after a White House dinner in 1802: “It is a long time since I have been present at so elegant a mental treat. Literature, wit, and a little business, with a great deal of miscellaneous remarks on agriculture and building, filled every minute. There is a degree of ease in Mr. Jefferson’s company that every one seems to feel and to enjoy.”16 The honeymoon would soon be over, however, when Jefferson’s architectural ideas and taste clashed with Latrobe’s over certain building concepts and construction details.

Construction of the White House wings in 1805 marked the beginning of a restrained and testy relation- ship between Latrobe and Jefferson. Responding to Jefferson’s early concern for a seamless height and connection of the wings from the White House to the Treasury and War Department buildings, Latrobe wrote, “I find many difficulties in the arrangement of the connecting porticos of the public offices with the President’s houses which however I do not despair of conquering.”17 Still, Latrobe found that this collab- oration of his own architectural taste and reason with Jefferson’s as “damned hard work” and confessed, “I have bestowed much labor upon them [designs for the wings] already, and find myself exceedingly puzzled how to determine the exact mode of accommodating the two ends of the wing between the President’s house and the Treasury to each other so as to answer the object of each in the best manner.”18

In a letter to Lenthall, Latrobe complained of the difficulty of aligning the east wing deck level with the flanking Treasury fireproof wing at a higher grade level. Latrobe lashed out in frustration and famously remarked: “I am sorry that I am cramped in this design by his [Jefferson’s] prejudices in favor of the old French books, out of which he fishes everything, but it is a small sacrifice to my personal attachment to him to humor him, and the less so, because the style of the Colonnade he proposes is exactly coincident with Hoban’s Pile,—a litter of pigs worthy of the great Sow it surrounds, and of the Irish boar, the father of her.”19 When Latrobe accidentally addressed the letter to Jefferson, Jefferson returned it stating he had not read it. Latrobe then wrote, somewhat embarrassed, to Lenthall: “The president, might have very safely read the whole of my last letter to you, even to the litter of pigs. He is certainly one of the best hearted men that ever came out of the hand of nature and has one of the best heads too.”20 An indication that Jefferson actually did read Latrobe’s reference to “old French books” is a letter Jefferson wrote in 1807 asking Latrobe for the return of his Kraft and Ransonnette book on the mansions of Paris, saying: “Being about to build some little temples in my grounds at Monticello, I must pray the return of the Plans des Maisons des Paris . . . as I expect to find some good designs in that.”21 In other words, Jefferson planned a fishing expedition in new (1801) French books.

The first portion of wings, east and west, appeared by the end of 1805 under Lenthall’s supervision. When Latrobe submitted the surveyor’s annual report on the public buildings to the president and Congress, part 3 of that report addressed the White House: “At the President’s House two small buildings have been erected, containing some of those domestic offices without which that building could not conveniently be inhabited. They contain a meathouse, cellars for liquors, coal and wood, and privies, and are intended to be faced to the South by a covered passage, or colonnade. Further menial offices, and some of them of the first necessity, are still wanted, before the dwelling of the President of the United States will be provided with all those domes- tic accommodations which are required by most private citizens.”22

Evidence for the Design of the Wings

Were these first sections of the White House wings, with later “menial” and subsequent sections, constructed according to Jefferson’s original plan? Did Latrobe influence changes to what Jefferson designed and intended? Did President James Monroe and the architect James Hoban alter the original design after the 1814 fire? This article is the first to seriously address these questions. To decipher the answers, a rather detailed account must be given to support an interpretation that stands on the variety of sources. The best answer to all of these questions requires some historical sleuthing based on the analysis of different sets of relevant evidence: Jefferson’s c. 1804 drawing of the intended wings; drawings and written references to the wings during construction; Hoban’s account of rebuild- ing the wings in 1817; the earliest known post-Jefferson wing plan drawing, by the architect Thomas U. Walter from 1853; and illustrations and photographs from the nineteenth and twentieth centuries, especially photo- graphs of White House renovation projects from 1969. The myriad sources and types of evidence are not always in perfect corroboration, but they have been assembled and reassembled to test all the possible hypotheses and present the case that is best supported.23

Correspondence among Jefferson, Latrobe, and Lenthall implies that Jefferson’s initial design was stubbornly adhered to despite Latrobe’s personal preferences. one of many lengthy letters indicates that Latrobe worked from Jefferson’s design but was producing working drawings, more or less in accord with Jefferson’s drawing, for Lenthall’s use. Latrobe wrote Lenthall during the initial season of building:

As to the president’s calculations of his coal cellars, pray don’t plague yourself about them,
nor about the necessary. You have my ground plan; let that be your guide & of what consequence is it, whether there be a foot or two more or less for coals or dung provided there be the room enough. now the president’s 20 feet, are two spaces of 10 feet, whereas my two spaces make only 19'6"—& my are produced by the actual division of the spaces allotted for the Colonnade into equal parts & his are assumption near the truth and as to the door & windows, it is also only of consequence whether they fall centrically into the intercolumni- ations or not, but of none at all whether they fall symmetrically internally.24

Latrobe considered his own drawings the more accurate ones for Lenthall to follow while reconciling Jefferson’s ideal design regularity to the reality of wall thicknesses and the practical space in each room. Walter’s plan, subsequent plans, and what is still stand- ing of the west wing confirm that exterior appearance and style did matter. Door and window openings were centered between each Tuscan column on the south (the columns did not come until 1808) and between the evenly spaced lunette windows on the north. That this design trumped internal room division walls is at the heart of the matter for understanding changes. Latrobe’s mention of rooms 10 feet wide does not match the sur- viving Jefferson plan and indicates either some missing updated drawings by Jefferson or some tweaking by Latrobe.

The 1805 report to Congress describes the first wing segments as “two small buildings” containing a “meathouse, cellars for liquors, coal and wood, and privies” (illustrations 27, 28). Taken at face value, according to Jefferson’s plan, this description would include three spaces on each side: ice house, coal and wood room, and privy on the west; and meat house with vault, stairway, and privy on the east. Latrobe’s working drawings for the wings are not known to survive, and the later physical evolution of the wings makes it is hard to confirm this arrangement and size except for the first room on the west. Key to distinguishing between Jefferson’s intentions and Latrobe and Lenthall’s con- struction is Walter’s 1853 wing plan, the second oldest known after Jefferson’s (illustrations 16, 17).

While it offers a distant comparative view in time, we know that President James Madison was determined to rebuild the house and wings as they were before the fire, a symbolic gesture that also implied the already iconic nature of the house. Hoban left a pretty good description of the extent of the main house’s standing walls but not of the wings. The wing’s masonry walls, partially in-ground, might have withstood substantial destruction just like those of the basement. The Walter plan thus serves as one verifiable measure in addition to other evidence when determining the original room usage and size for the east and west wings.

on the west the subterranean ice house became re-enclosed in brick as part of the west wing in 1805. It constitutes the “room” and contained the wine cellar space referred to in the report to Congress as “cellar for liquors.”25 The proposed Walter plan substantiates the ice house use and location in 1853, since its survival was structurally assured and its function easily reactivated after the fire.

In Jefferson’s plan next came the wood room with a coal cellar below. The period use of the term “cellar” did not necessarily mean a below-grade room but could denote any type of storage room or space. In this case, however, the Jefferson drawing labels this space “coal below, wood above.” What it does not indicate is how servants accessed the piles of Virginia Midlothian coal.26 What seems odd on the Jefferson plan is the narrowness of the wood room and coal cellar, scaled to be about
8 feet wide. It is here that a reference in the above mentioned letter from Latrobe to Lenthall makes sense. Latrobe complained about the president’s “calculations of his coal cellars” and that a few inches difference in size for the cellars, be they for “coals or dung,” did not matter. This reference seems to be linked to the one that immediately followed, referring to two spaces of 10 feet each. Jefferson apparently increased the size of the sec-ond room to 10 feet wide, a width that still worked with the exterior openings and made the adjacent necessary 10 feet wide. The lingering question is how anyone reached the below-grade cellar. The Walter plan shows
a much larger second room of about 20 feet wide, the combination of Jefferson’s two rooms, that in later plans is still labeled as wood and coal storage.

Jefferson’s third space, the necessary or privy, Seale indicates was for servants on this side whereas the east privy would be reserved for the family and guests.27 At 10 feet wide it might have been divided into two stalls but, one or two, was probably still “unisex,” as per the custom. Latrobe’s reference to cellars for “coal or dung,” mentioned in relationship to rooms 10 feet wide, seems to confirm a combined cellar space for the coal cellar and privy waste removal. This combination works only for the west wing, but it does not resolve the question of access. A fourth room is unidentified and was probably used by servants.

A surviving drawing helps establish the size of the initial west wing (illustration 9). This drawing shows someone taking measurements of the existing house and wings. If it is related to working out the dimensions of Latrobe’s design of a pilaster and pier, it would date to 1807 or 1808. The west wing length given in this draw- ing is 50 feet.28 This size accommodates five window or door bays and accounts for all three of the Jefferson designated spaces as well as an extra space on the end. on the Jefferson plan that fourth space is labeled “saddle room,” but at this period there were no stables in the west wing. Walter’s plan shows a necessary of about 12 feet wide, divided into two stalls with seats, but placed to the west of Jefferson’s necessary.

What complicates a reconciliation of the Jefferson and Walter plans is the odd fenestration shown in 1853. While we can expect the room division walls to vary from the regularity of the exterior features, four blind windows shown on Walter’s west wing plan for the first five bays produce an awkward collision with internal walls. It is here that other sources help sort out the contradictory evidence. These include a period drawing, personal and official correspondence, and physical evidence shown in photographs.

Latrobe drew a beautiful south elevation of the White House showing a proposed portico (illustration 18). It also shows four bays of the east and west wings at that time. on the west is pictured: a lunette window on the ice house wall; a doorway with typical lunette window above, another doorway, and a window. This fenestration would seem to support the Jefferson room plan and sequence of an ice house, a wood room, and a necessary.

Another piece of evidence is an unlikely photo- graph taken on november 28, 1969, by a White House photographer documenting three men digging out the west end of the west wing for Richard nixon’s new Press Room (illustration 22). The men are working below grade in a Piranesi-like view amid piles of dirt and fragments of masonry structures. While far from a suitable documentary recording of architectural evidence, this photograph sadly provides our known universe of physical evidence from which to interpret and test hypotheses of the initial west wing room plan.

The photograph clearly shows a round brick out- line of the ice house, the wing room most substantiated. By using the window and door bays as markers, it is possible to place the “ghosts” on the upper walls and the below-grade remains in a proper context despite the photograph’s warped perspective. on the right, the south wall, can be seen the first lunette window bay adjacent to the ice house, with a later doorway below it. To the right of that bay is the “ghost” of a missing wall that separated the ice house from room two. The second bay shows the sill of a doorway accessing the wood room that was later filled with brick. The third bay is shown as a doorway that would have accessed the necessary. These three bays confirm Latrobe’s elevation and the Jefferson plan. The evidence would seem to confirm the Walter plan’s depiction showing a window in bay two.

on the north side (left side of photograph) the ice house wall “ghost” is visible to the right, and continuing below window bay two. This defines the wall between the ice house and the wood room, whose walls are shown on the Walter plan as recessed for the original window and door to open into the room. on the wall to the right of window bay three (upper left foreground), and below to the right of the mason, is evidence of the missing wall between the wood room and the privy. This same wall also appears on the right side of the photograph in the foreground just below and east of the doorway (due to the perspective of the photograph things do not seem to align if the opening bays are not used as reference). This wall also establishes the eastern wall of room three, the necessary, whose full extent can- not be seen in the photograph. These walls for the original rooms two and three prove that the larger room shown in the Walter plan was created out of two earlier rooms. To the left of the mason is a lower brick wall that seems to be directly under the third window bay, with an arch springing from it.
In addition to showing five later period doorways,

there are three things of interest in the photograph. First, at the far wall, the eastern end of the wing, a large masonry arch has been filled in. This arch presumably carried the weight of the wing’s eastern brick end wall above the ice house that protruded beyond it at a lower level. Constructed first, the ice house was covered by a wooden roof structure that was demolished when the wing was constructed, leaving the protruding wall that needed to be captured in the squared brick wing walls. The ice house might have been identical to that Jefferson constructed at Monticello at the same time, with a wooden roof structure below the level of the wing roof.29 The original arch extended just to the south (right in the photograph) of the later doorway that was inserted in the filled wall when the ice house ceased to function.

Because the ice house is off-center to the north in the wing, there is room on its south side under the arch for space containing a stairway to be squeezed between the curved ice house wall and the south exterior wall. This is the most likely place for a narrow stairway accessing the cellar space below rooms two and three, and making sense of a cellar for “coals or dung.” Jefferson’s drawing hints at this arrangement by show- ing a south doorway in bay one, although with no indication of stairs.30 The Walter plan does not show any access or stairs but simply a solid mass of masonry in this location.31 The only reasonable access to cellars below would be a narrow stair space defining a third room, like that shown on the east wing plan, but nothing supports this scenario.

Finally, one perplexing bit of evidence in the photograph is the remains of two arches on the north side shown on either side of the mason. It would be reason- able to think that Jefferson might have created vaulted spaces for the coal cellar and/or a necessary clean-out. However, a letter from Latrobe seems to rule out this possibility. Latrobe remarked to Jefferson: “I regret that your Coal cellars were not arched. I have seen so much rotten timber in every building erected in Washington, that my passion to exclude it altogether grows upon me daily.”32 The arches could have been added in the 1818 rebuilding to help support the combined spaces of rooms two and three for wood and coal storage. There is no means of a lower-level access in Walter’s plan, so the arches would not be for the necessary, which must have been cleaned from the space itself.33 The other possibility is that the arches are relieving arches for the north wall, allowing the “Coal cellars” to extend beyond the wall into the higher north grade. As such, they would not have fit Latrobe’s definition of an arched floor like that designated for the meat house’s vaulted floor.

For the east wing a comparison of the Jefferson and Walter plans can be made, but the photographic and physical evidence is lacking due to that wing’s demolition in 1866. Jefferson’s plan calls for room one to
be “Meat house above, vault below,” room two as “Descent into cellar” stairway, and room three as “necessary.” All three functions are mentioned in the 1805 report. What is seen on Walter’s plan is consistent with Jefferson’s room use and sequence but shows the rearranging of internal room walls to achieve a larger meat house and necessary.

“Meat house” was a term contemporary with “smokehouse,” a structure where meats were smoked and salted for curing and then usually suspended from the rafters until needed. The difference with this meat house, due to Jefferson’s flat terrace roof above, was that it did not have the typical tall pyramidal peaked roof wherein the hams cured in the upper reaches of smoke. Did this meat house contain only hams smoked elsewhere? Two clues suggest that fires were actually made in the space. The “vault below” refers to a mason- ry vault of brick. This would not be necessary for constructing the floor above a cellar or for creating a ceiling above a cellar space, but it would be necessary if the floor structure needed to be fireproof for a fireplace or firepit in the room. The other possible clue to smoke being present is that the Walter plan shows a possible opening for ventilation on the west wall of this space facing the exterior passage between wing and house.

In two other applications where Jefferson created a low ceiling smokehouse under a terrace deck roof— Monticello and Poplar Forest—there was no apparent exit for smoke other than what seeped through the roof or deck.34 The vault was not necessary for a fireproof floor since typical smokehouses had floors of dirt or brick. The vault not only made for a fireproof floor but created a cellar space below, as the stairway indicates. What the cellar was used for is unknown, although the proximity might indicate that it could have housed wood for the fires above. The possibility that the fireplace was located below the space seems to be negated by Jefferson’s use of the word “cellar,” which implies storage. The stairway, constituting the second room space, in addition to accessing the “cellar below,” could also have accessed the underside of the necessary where waste could be removed. Its window most certainly would have been used originally for light, as opposed to the blind window shown on the Walter plan.

The third space on Jefferson’s east plan was the “necessary” that, enlarged to 10 feet wide, might have accommodated two separate stalls for bench seats as shown in the Walter plan. Typically, Jefferson put louvers in his privy window openings for ventilation, but unfortunately his plan does not show windows and the Walter plan does not distinguish anything different in this opening.

In summary, it seems Jefferson stuck to his original plan but allowed for larger spaces (illustrations 27, 28). The first section of the west wing ended up as a five bay section, 50 feet long. The east was first a four bay section, 50 feet long. This arrangement seems to be con- firmed by the fact that in 1805 Jefferson mentioned that he wanted to extend the east wing by 60 feet and in 1807 mentioned extending the west wing 50 feet, giving each by 1809 their 100 foot lengths.35 on the east this extension was accomplished in two sections: a stable space was added in 1806, and a carriage house was added to the end in 1809. The west wing did not get its matched extension until 1818.

The Roof Structure

The innovative significance of Jefferson’s service wings lies in their roof structure. To create his flat terrace—or what he called “terras”—roof deck or walking platform, Jefferson had experimented at Monticello with several ways to compress the roof structure by altering the ceiling joists of the rooms below. The roof needed to act in a typical manner to keep out water while giving support to the deck and rising to a minimal height, typically behind the entablature trim. It was important that the deck level be at the same level as the floor of the house from which one stepped.36

Jefferson might have seen versions of flat roofs and terrace decks in Europe, where masonry “platform” roofs were more common, but his documented evolution of experimenting with different methods leads to the conclusion that he did invent his ultimate “terras” roof-deck construction.37 Latrobe, who certainly knew European construction methods, called it the “President’s zig zag roof of sheet iron uniting all the good qualities of the pantile, without its bad ones.”38 Latrobe had more than a passing interest in sheet iron for roofs, because he was partners with Samuel Mifflin in the first iron rolling mill in the United States and claimed to have been the first to use sheet iron for roof coverings.39 Latrobe and Mifflin were providing sheet iron for Monticello in 1803 at the same time that they were busily producing it for both the White House and the Capitol. At the White House the iron was needed for a hasty retrofit to replace the leaking slate roof and gutters of Hoban’s roof.

Jefferson had ultimately settled on a compressed series of high and low ceiling joists for his wings that he called the ridge and gutter joists (illustrations 4, 5, 6). By spanning the closely spaced high and low joists with two layers of shingles, he created a miniature shingle roof that was hidden underneath the deck and at the same time acted as its support. If sheet iron was used, wide horizontal boards would be used to span the joists. Rain water fell through the cracks between deck boards to the sloped shingles or tin-covered boards below, where it was directed into the scooped-out wooden gutter joists that were pitched outward to carry it, by gravity, to one or both sides. In some instances the water was directed into cisterns, and at other times it fell out to the ground through scuppers in the entablature.40 At the White House, hidden gutters directed water to tin pipes that ran down the outside of the room walls and from there it probably fed into cisterns in the basement, just as the main roof gutters delivered water into attic cis- terns that fed the interior water closets. As Latrobe put it to Lenthall, “[There are] cross gutters inside the [wing] wall, into which all the others piss, as you say.”41

The Achilles’ heel of this system was the wooden gutter under the deck that stayed dark and damp and rotted. Jefferson solved this problem by using some sheet iron intended for the Capitol. Folded sheet iron was placed over the wooden ridges that overlapped, like a pantile roof, with inverted folded sheets in the gutters. Jefferson’s “terras” roof drawing for the White House shows a v-shaped gutter to accommodate the sheet metal.42 The terrace deck above required a wooden cap on top of the ridge in order to hold and secure the perpendicular sleepers to which the deck boards would rest and fasten. Even with sheet iron covering the wood, the same longevity problem existed: the thin sheets of rolled iron could not be tinned like smaller sizes and had to be continually painted to be preserved.

The fixed deck not only kept air and sun out but also prevented maintenance, causing the iron sheet metal to rust or uncovered wood gutter joists to rot. Jefferson apparently first saw a similar use of sheet iron at General Samuel Smith’s modern house, Montebello, outside of Baltimore. In a letter to Latrobe in 1803, Jefferson asked him to stop to see the house with its flat roof that used iron sheet metal in gutters, being “the first and only example yet executed.” Jefferson asked Latrobe to examine it so that “it may furnish us, by the manner of its execution, information both as to what succeeds, and as to what may not succeed, and therefore is to be avoided, if anything about it does not succeed.”43

There is no known further correspondence on the matter, but Jefferson proceeded with the sheet iron, and in fact had already ordered the material for Monticello’s roof. While he used sheet metal for gutters at the White House and later at the University of Virginia, he came to acknowledge its failure. Concerning a similar “terras” deck and zigzag roof on the 1815 service wing at Poplar Forest, Jefferson remarked that sheet iron was expensive (when not at federal expense) and lasted no longer than the thick wooden gutter joists, which at Poplar Forest was only ten years.44 At Poplar Forest wood shingles spanned the high and low joists and directed rain water into the wooden gutters and out through scuppers in the entablature, where it fell to the ground.45 Using sheath- ing boards covering with sheet iron rather than wooden shingles saved a great amount of labor. A surviving wing roof shingle found at Poplar Forest shows that they were not typical shingles but were specially made. Each pine shingle—and there were about nine thousand for a 100 foot long roof—had to be planed on four sides, cut with beveled tops and bottoms, and incised with two parallel 1⁄4 inch grooves on their faces to facilitate downward movement of water. The labor, it must be remembered, on Jefferson’s own projects at Monticello and Poplar Forest, was slave labor. At the University of Virginia in the 1820s, with state money, Jefferson experimented with even more versions of this system over the student rooms on The Lawn, sometimes with wood shingles and at other times with sheet metal in the gutter joists.46

The sheet metal used in the White House roof gutters probably lasted until the British torched the place in 1814, but another typical type of failure before that pre- vented President James Madison from strolling or sitting on the deck as Jefferson had. Latrobe wrote Madison in 1812 that the “platforms covering the gutters were rot- ten and must be replaced.”47 This message must have chagrined Madison, who had adopted one of Jefferson’s zigzag roof systems to create decks over the 1809 wings at his Montpelier. Whether they were replaced at the White House, or not, is hardly relevant because they would be burned in two years. When the wings were rebuilt by James Hoban in 1818, the roofs were copper. Whether Hoban constructed the roof structure with Jefferson’s joist system or simply constructed a low pitch rafter roof is unknown.