Washington and Lee University

The Building of Arlington House
Murray Nelligan

Note: The following is taken from the May 1951 issue of the Journal of the Society of Architectural Historians (volume 10), pp. 11–15.


MURRAY NELLIGAN is the National Park Service Historian for the Lee Mansion in Arlington, Virginia. His investigations clarify the successive stages of the construction of the mansion.

THE BUILDING OF ARLINGTON HOUSE

For well over a century Arlington House, now the Lee Mansion National Memorial, has been a conspicuous feature of the scene across the Potomac from Washington. Yet for all its architectural excellence and historic interest, surprisingly little has been definitely known about its origin and construction.

Unfortunately, neither George Hadfield, who designed it, nor George Washington Parke Custis, who built it for his home, left any records concerning it. Its history must be pieced together from the existing fragmentary evidence. Custis was a man of considerable culture, but evidently he had little interest in architecture, for he is known to have discussed the subject on only one occasion, when in the early 1820's a young Army officer named John Farley was at Arlington. Farley was at Arlington again in the summer of 1861, a few months after the outbreak of the Civil War, when General Irvin McDowell had his headquarters in the mansion. With him was his son Joseph, aide to General Mansfield. Years later, young Farley had this to say of their visit: “My father called me to a particular spot on the lawn in front of the house, remarking as he did so: ‘I want to explain to you the architecture of this building, just as Mr. Custis did forty years ago.’ ” Whereupon, young Farley went on to something else, thus depriving architectural historians of what might have been the best, if third-hand, account of the building of Arlington House.1

Fortunately, there are more enlightening sources than Joseph Farley. For one thing, it is now known that Custis tried to buy Mount Vernon from Bushrod Washington in 1802, after the death of Martha Washington, Custis' grandmother, and that he was unable to do so because of the opposition of Mrs. Bushrod Washington to the projected sale.2 Custis thereupon moved to Arlington, which he had inherited from his father, John Parke Custis, Martha Washington's son by her first marriage. Young Custis was twenty-one and unmarried when he moved to Arlington and set up bachelor quarters in the four-room cottage built by a former owner near the river. In it he stored the many heirlooms of his family and the Washingtons he had inherited or bought at the various sales held by the executors of the Washington estate: silver and china, paintings and furniture, and a vast number of relics of the first President, including his death-bed. His first concern was to get the Arlington estate into production, since his fortune consisted largely of land and slaves. But when he found that his cherished mementoes were being irreparably damaged by dampness and rats, he decided to begin the construction of a house where they could be safely preserved and suitably exhibited.3

If he did not already have them, Custis must have obtained plans for his house and its outbuildings at this time, for there are no indications that any significant changes in design were made in the mansion after it was begun, and the two servants' quarters and the stable (no longer existing) were in the same architectural style.4 These plans have been attributed at one time or another to numerous architects and builders, including Thomas Jefferson, tavern-keeper John Gadsby, a nameless Turkish ambassador, and Custis himself, though ineither in his letters nor in the many accounts written about him by his contemporaries, are such plans mentioned.

In recent years the house has been correctly attributed to George Hadfield, the gifted young English architect who stayed on in Washington after losing his job supervising the construction of the Capitol. Hadfield died in 1826 and his obituary lists “Mr. Custis's house” as one of his creations.5 William P. Elliot, architect and surveyor, who was for five years a student of Hadfield and a great admirer of what he termed Hadfield's “beautiful proportions” and “chaste architecture,” says the same thing in a letter to the editor of a local newspaper seven years later.6 Elliot published the first known picture of the house in the 1826 edition of his Washington Guide.7 William Dunlap, artist, dramatist and diarist, likewise attributes “the plan of . . . Custis's Mansion” to Hadfield in a book published eight years after the latter's death.8 These men were in a position to know, and their attribution must be accepted in the absence of better evidence to the contrary.

Influenced by the contemporary interest in classical architecture, Custis wanted his house in that style. Accordingly, his architect planned one with an impressive Greek Doric portico in front of the two-story main section, balanced by one-story extended wings, the whole composition having proportions strong enough to make the building show to best advantage when seen from across the river. Since ornamentation would be useless at such a distance, Hadfield relied almost entirely on good proportions to give beauty to his creation. Inside, the rooms would be large and have high ceilings, and their walls would be severely plain except for a molding near the ceiling. Having them open into one another would give an illusion of great space, while semi-circular arches and beautifully proportioned windows would relieve any impression of austerity.

In building his house, Custis followed the common method of constructing the wings first, leaving the center section and portico to be put up later. As originally built, the entire north wing was one large banquet room with a kitchen and a laundry in the basement below, as is shown by the distinct difference found between the finish of the wood trim on the perimeter and that on the present partition walls. Furthermore, the old chimney breast, still to be seen in the middle of the west wall, is in precisely the same location as the present one in the south wing. It is likely, too, that the beautiful wooden mantel, now in one of the small inside rooms, was originally intended for this fireplace, for it is disproportionately large for its present situation. The location of the pantry and the indoor kitchen in this wing and the outdoor one just west of it, supports this belief.9

The corresponding wing to the south was outwardly identical in appearance, but was divided inside into a large parlor for entertaining and a smaller room for use as an office and study.10 Each wing was a complete building in itself, indicating that Custis anticipated some delay in building the main part of the house.11 Even at this stage “Arlington House” attracted favorable attention; one young lady visiting there in 1804 reported that “The House will be a very handsome building when compleated.”12 Even more enthlusiastic was the comment of a traveller a few years later: “This seat is on a superb mount, and his buildings are begun in a stile of superior taste and elegance.”13

There is some evidence that both wings were built by the summer of 1804 when Custis married Mary Lee Fitzhugh, of “Chatham,” near Fredericksburg.14 To provide living quarters, the north wing was partitioned into three small rooms, the present arrangement. No doubt financial reasons account for the delay in building the center section and portico until after the end of the War of 1812. Years later, Mrs. Lee wrote that she could just barely remember when the main section was built; while one of her childhood playmates, born in 1812, could recall chasing chickens about its foundations when a little girl.15 Apparently, the main section of the house was built in 1817, for in December of that year a waterproof stucco plaster devised by David Meade Randolph, whose wife, Mary, lies buried just below the house, was applied to its northwest corner.16 It may be inferred, too, that the portico was built about this time, else the term, “Custis' Folly,” applied to the house by a traveller from the North early the next year, would hardly have been justified.17 Miscellaneous bills for plastering and the services of a builder, together with a letter by Custis putting off the payment of an overdue bill because of “the heavy expense of my building,” further support this belief.18 Not until 1824, when General Lafayette visited Arlington, is the portico specifically mentioned.19

The building was most solidly constructed. All the walls and most of the foundations were of brick, as were the portico columns, but there is no evidence that these were made of clay dug on the estate, as is the tradition, though this was a common practice about Washington at the time.20 All the brickwork exposed to the weather was covered by a hard stucco plaster, scored to imitate cut stone, as was the pediment, a balustrade on top of the wings and some of the interior walls. Slight horizontal grooves around the portico columns indicated drum lines.21 Originally the cement on the columns seems to have been tinted to look like marble, but later was painted white, providing more contrast to the warm buff or ochre color of the rest of the exterior. Joists, studs and rafters were of hewn timber, neatly mortised together or pinned with woodeln pegs, the main roof being supported by great trusses spanning the entire width of the center section, and every piece of timber being numbered, however small. Doors, cornices, floors and woodwork were of native pine.22


Arlington House in 1853,
a sketch by Benson J. Lossing.


The portico and south wing in 1864.
Photo by Mathew Brady,
courtesy of the National Archives.


The family parlor and dining room at the time the mansion
was being restored by the War Department. Originally there
were doors and a fanlight in the center arch, those on the
outside being filled in with lath and plaster, perhaps to
make the room easier to heat. Mary Custis and Robert E. Lee
stood under the center arch during their marriage ceremony in 1831.

For all its imposing appearance when seen at a distance, the real size of the mansion is not apparent until seen close at hand, for magnolias now completely hide the one-story wings, each 40 feet long and 25 feet wide. All that is presently seen is the portico and two-storied center section of the building, the latter being 60 feet wide and 40 feet deep. Although the wings, unobscured, would be quite pleasing because of their tall recessed windows, the magnificent Doric portico is the salient architectural feature. This extends out 25 feet from the house and has eight columns, each 23 feet high and a little over 5 feet thick at the base. Early writers differ as to whether the portico was derived from the smaller, well-proportioned Thesium at Athens, or the larger, more imposing Temple of Neptune at Paestum, Italy.23 Actually, the portico of Arlington House seems to be a compromise between the two, being larger than the first named and smaller than the second.24 Regardless of the prototype, there is no doubt as to the effectiveness of the architectural style chosen, for no other would have had the strength and massiveness necessary to make the building impressive from the distance at which it was normally viewed.25 Yet for all its simplicity and solidity, the proportions of the mansion are so well balanced and refined as to make it one of the most appealing of its type.

Actually the building was conceived on too grand a scale for Custis to complete it before his death in 1857. The large drawing-room south of the main hall was not plastered until 1855, at which time marble mantels of Victorian design were installed by the Lees, together with a hot-air furnace in the basement. During this “renovation,” two dressing rooms at the east end of the second floor hall were made into one small guest room and a doorway opened from it into the hall.26 The partition walls in the north wing were never removed; and it was not until after the Civil War, when the house was serving as headquarters for the Arlington National Cemetery, that the west side of the house was covered with stucco cement to correspond with the rest of the exterior.

Though he was a man of considerable ability and one of the best-known personalities of his time, George Washington Parke Custis of course never approached the stature of George Washington or Robert E. Lee, between whom he was a living link. Yet he had his own merits and achievements. His home was one of the latter, a monument to his taste and perseverance. In building it, he gave to future generations not only an outstanding example of a Greek Revival building of the early 19th century, but one whose architectural character, dignity and strength combined with simplicity and beauty—make it most appropriate for the purpose to which it is now dedicated, a memorial to General Robert E. Lee.

NATIONAL PARK SERVICE