Once the huts were well on their way to completion, Brumbaugh was instructed to begin another phase of the "complete restoration"—the rediscovery of Fort John Moore, popularly known as the "lost redoubt." Fort John Moore would have been an earthwork fort or redoubt surrounded by an open ditch or moat created by the digging of soil to form its walls. Tradition had it that Fort John Moore was one of two redoubts that would have been located at the northern extremity of the outer line, completing the chain of forts surrounding the encampment. Antiquarians referred to it and its sister earthwork as Fort John Moore and Fort Mordecai Moore, after the owners of the farms on which these structures had been built. Fort John Moore had been built shortly before Washington marched his men out of Valley Forge, and it had long since been plowed under by local farmers. Author Henry Woodman mentioned Fort John Moore in his 1850 history of Valley Forge, and the fort appeared to be indicated on some early maps of the encampment. The park commission found the documentary evidence "contradictory and confusing" and in the instructions to Brumbaugh emphasized the importance of accuracy and authenticity. [51]

Attempts were made to locate the fort using mine detectors to find metal relics that might be hidden beneath the soil. The Pennsylvania National Guard Military Engineers looked for the fort in February 1948, and the Second Army looked again in May. Neither came up with positive results. [52] Brumbaugh had heard that air photography was being used to locate Roman ruins in Britain and was eager to apply this new wrinkle in scientific archaeology at Valley Forge. Consequently, the Eleventh Air Force flew a mission for Valley Forge the week of May 11, 1948, and provided the architect with photographs of the suspected terrain. One picture showed a short and faint line in the grass indicating subsoil disturbances. Brumbaugh and Park Commissioner Norman Randolph, a former brigadier general, carefully located the spot in the fields at Valley Forge by relating it to trees visible in the pictures. [53]

The next step was conventional field archaeology. Military archaeologist J. Duncan Campbell was engaged to dig an exploratory trench. Right away he found "disturbed earth," which everyone interpreted as the ditch that would have been outside the fort. Dr. J. Alden Mason of the University of Pennsylvania became interested and cooperated with Campbell. The actual diggers were high school students on vacation, and when it became clear that they would have to report for classes before the excavation was done, Brumbaugh rounded up construction workers from another Valley Forge site and put them under the supervision of Mason and Campbell. Within a few more exciting days, the walls and moat of Fort John Moore were either uncovered or clearly indicated. [54] In all the digging, only one disappointing eighteenth-century relic was found: a small hand sickle. No other remains were unearthed except for some layers of charcoal and more suspected human bones. Fort John Moore was reconstructed entirely on the evidence of "feature archaeology," or the subsurface remains of a long-buried structure. [55]

No one knew what eighteenth-century soldiers would have called this particular earthwork, but Brumbaugh suggested that the fort's name be changed to Fort Muhlenberg because Muhlenberg's troops had been en camped closest to it. The Pennsylvania Historical and Museum Commission named it Fort Greene after Nathanael Greene, one of Washington's major generals. It is now known by the unsentimental name "Redoubt #2."

Having reconstructed the log huts and located the lost redoubt, the commission would take one more step toward re-creating the winter encampment with the construction of the Knox Artillery Shop. According to tradition, there had been such a shop somewhere near the old road running through the camp's artillery park (now Baptist Road Trace), a central location where the Continental Army's artillery was thought to have been collected so that cannon could be quickly dispatched in any direction in case of attack. The shop replica would be built near the artillery park's replica cannon and would be equipped with period tools. Demonstrations could be staged to show visitors how camp horses had been shod and gun carriages repaired.

Construction of the Knox Artillery Shop went fairly smoothly. The Second Army again used mine detectors to try to identify the shop's original location, but found nothing where this shop had supposedly once been. Although elsewhere they identified possible submerged metal within a 50-foot radius of a deposit of charcoal, their results were inconclusive, so Brumbaugh was instructed to erect the shop at its traditional site. Based on his knowledge of colonial blacksmith and wheelwright shops, Brumbaugh designed a building with wide doors and a wheelwright's bench appropriate for the operation of these crafts. The shop's equipment was donated by the director of the Fort Ticonderoga Museum in New York. [56]

The fourth jewel in the "complete restoration" crown was to be the drastic renovation of the Washington Inn, where hotel and restaurant operations had always been such a thorn in the side of the park commission. With its tall, round cupola and fancy balconies and ironwork, the Victorian Italianate Washington Inn also visually overpowered nearby Washington's Headquarters and dominated the park's major intersection. To tourists it had always been a landmark; to the park commission it was an eyesore. Park Commissioner Norman Randolph used that very word, calling it "an eyesore in grotesque contrast with one of the principal restorations designed to enhance the beauty and dignity of the Headquarters area." [57] Brumbaugh agreed. Like most disciples of the Colonial Revival Movement, he was devoted to the merits of seventeenth-, eighteenth-, and early nineteenth-century architecture. Brumbaugh referred to the renovation that had given the Washington Inn its Victorian Italianate features as "vandalism." [58]

In 1947, Brumbaugh began carefully peeling away the layers at the Washington Inn. When the cupola was dismantled, he encountered one interesting message from the past. On the back of an arched spandrel over one of the cupola's windows, a nineteenth-century worker had penciled the words "May 5, 1854—The son of a bitch that takes this down will remember Garrett Snyder, the son of Harry Snyder, Roxborough Town Ship, Philadelphia —G. Snyder foreman for Sam Rau, Wages $2 per Day." [59] Undaunted, Brumbaugh proceeded with his careful demolition and eventually identified six or seven stages in the structure's history. He concluded that an original log cabin had been rebuilt as a stone dwelling, which was subsequently expanded into a long, two-story house around 1758. Later a rear wing had been added, which Brumbaugh interpreted as an adaptation that converted the house for the simultaneous use of two different families, A major reconstruction had occurred sometime after the encampment around 1790. This had been followed by the "vandalism" of Garrett Snyder in 1854, which had been further adapted about twenty years later to transform the house into a hotel. [60]

The history of the house made its restoration particularly problematic. So great were the changes made after the encampment that exactly what the house had looked like to Washington's men could not be determined. Brumbaugh found better clues to its appearance in its postencampment incarnation, including some attractive wall paneling and markings on the plaster walls indicating the position of chair rails and baseboards. Other marks, on the floorboards, clearly showed where fireplaces had been. [61] It was decided to make the Washington Inn look like it had shortly after the encampment, when renovations had transformed it into what is known as the Federal style.

As with Brumbaugh's other projects, work went slowly. In January 1948, Brumbaugh wrote Norman Randolph: "This house has been one of the most difficult I have ever worked upon. Changes were made so often and so well in its history, that traces and clues have been obscure and contradictory. But the mystery is yielding to patient research, and I am growing more enthusiastic as we proceed." [62] By the end of the year, Brumbaugh was again dealing with labor problems and late deliveries, while Randolph expressed his concern at the slow progress on this project. In the interest of speed, Randolph suggested that Brumbaugh eliminate the flues leading from the restored fireplaces to the house's central chimney, because these fireplaces would no longer be used. Brumbaugh reacted strongly, citing his "reluctance to build shams in as important a building as this." His letter promised more speed without sacrifice to authenticity. [63]

Brumbaugh was apparently willing to compromise on the porch. The Washington Inn once had large porches, but such porches were very uncommon in Federal-style houses, Brumbaugh found evidence on the exterior walls that the porch foundations of the house had been added sometime after the Federal renovation but before the 1854 renovation. It is now believed that the porches were added between 1825 and 1850, so they should have been omitted from Brumbaugh's Federal-style restoration. However, the park commission wanted porches for the convenience of visitors, and Brumbaugh included them. [64]

There had long been a tradition that somewhere in this structure bread had been baked for Washington's army. The legend had originally been inspired by a letter dated August 30, 1777, from Richard Peters, secretary of the Board of War, to Pennsylvania's then chief executive officer, Thomas Wharton. This letter recommended that six militia bakers be sent to Valley Forge, where supplies were hidden, to take care of "a large quantity of Flour spoiling for want of baking." In her history of the Potts family written in 1874, Mrs. (Isabella) Thomas Potts James claimed that this baking had been done in ovens located in the cellar of the Washington Inn. She was probably just guessing based on the wording of some of Washington's orders, in which the general referred to the "Bake-House by Headquarters."

For weeks, Brumbaugh eagerly sought evidence of these ovens. He found it entirely conceivable that ovens might have been constructed in the cellar just before the encampment so that large-scale baking could be done for the army in secret. However, he found nothing in the cellar except a "curious curved wall, one brick in thickness, surrounding a depressed, brick-paved pit, 27 inches below the earth floor of the basement." The pit had an equally curious terra-cotta pipe leading out of house through the west wall. At the time, this reminded Brumbaugh of a kind of oven used by the Pueblo Indians, but the architect had no way of telling whether the unusual structure had been used for baking bread in 1777—1778. There was no evidence that it had ever been subjected to heat. [65]

Brumbaugh recommended that no oven replicas be constructed in the cellar until more evidence was uncovered. Instead, Brumbaugh suggested that a more typical oven be built outside the house. Brumbaugh was not consulted in 1963 when the park commission decided to go ahead and build oven replicas in the cellar anyway, even moving the cellar entrance to a new location to accommodate their construction.

The current version of the bread-baking story holds that the house was the site of production-scale baking as early as the 1760s. It is believed that the owners of the house wanted to enter the flour trade, as well as provide supplies for the workers at the old forge. During the encampment itself, bread was probably still baked at this general location for the use of those quartered at or near Washington's Headquarters. But the house had in no way served as a central bakery for the whole Continental Army, nor were there ever any ovens in the cellar. Any old-timers who remembered "ovens" in the cellar may have been thinking of the brick structures incorporated into some houses during the 1830s and 1840s, which retained hot air and acted as a kind of central heating system. As Brumbaugh suspected, the ovens the house had possessed had probably been conventionally located outside the house somewhere near the kitchen. Brumbaugh's clay pipe leading out of the basement pit is currently believed to have allowed the pit to fill with cool water so the basement could function as a spring room for the preservation of food. Unfortunately, this unusual eighteenth-century architectural feature was obliterated by the inaccurate oven reproductions built in the cellar in 1963. [66]

While Brumbaugh was being goaded to hurry along with work at the Washington Inn, the park commission was also deciding what to do with the building once Brumbaugh was done. This restoration had not been undertaken to add anything to the headquarters area, but rather to rid the park of the objectionable architecture of the Washington Inn. In 1949, it was decided to use the building as an administrative headquarters because at that time it seemed unlikely that the park would ever get the funds for a separate administrative building.

The old Washington Inn was first renamed the Bake House. Once it was no longer being interpreted as the bakery for the Continental Army, its name was changed to "Colonel Dewees Mansion," after William Dewees— the name Brumbaugh had also used for it. It is now called the "David Potts House," after Dewees's cousin David Potts, who is believed to have owned and lived in the house at the time it was renovated in the Federal style after the death of Dewees in 1782.