A Confederate Bombproof Interior

by Dr. Lawrence E. Babits and Rick Leech

Originally published in the Journal of The Company of Military Historians, Vol. XLII, No. 4


In April, 1862 the Union Army successfully employed rifled artillery against the brick walls of Fort Pulaski. The Confederate response, based on limited time and resources, was to erect earth fortifications. The Union Navy bombarded Fort McAllister five times between 19 November 1862 and 3 March 1863. The defenders demonstrated that, even with parapets shot away and defenders firing artillery in the open,(1) earthworks could successfully resist naval bombardment.

A side effect of the naval bombardments was construction of shelters for the defenders. A bombproof was a chamber designed to give protection from shelling. With the increased penetrating power of rifled shot already demonstrated at Fort Pulaski and Fort McAllister (see Table 1), large amounts of earth were required to provide the needed protection.

In 1985, Causton 's Bluff, Chatham County, Georgia, was subjected to clearing and road building which affected the remains of Fort Bartow.(2) The bombproof which was inspected during concurrent archaeological investigations is a large earth mound approximately 75 feet long and over 15 feet high. Its base width is approximately 75 feet which narrows to 21 feet at the top. The total amount of material contained in this mound is over 5000 cubic yards.

Table 1 - Shot Penetration(3)

Musket shot

1 foot six inches
6 pound shot 3 1/2 to 4 1/2 feet
9 pound do. 6 1/2 to 7 do.
12 pound do. 8 1/2 to 10 do.
18 and 24 pound do. 11 1/2 to 13 do.
15 inch 17+

The mound was cut through its central point to provide fill and create an access road. The bulldozing under archaeological supervision allowed inspection of the mound interior and revealed a chamber nearly 20 feet wide composed of heavy timbers. The overburden was removed to expose the roof timbers which were mapped and photographed.

Narrow trenches were then cut through the roof timbers to allow mapping of interior structures, the floor and walls. Additional spot checking was necessary to answer specific questions as drawings were being completed.

The bombproof was not a simple earth mound. Three contemporary maps show the initial construction stage involved digging a trench. A rectangular structure shown on an 1863 Confederate Map(4) does not appear on 1864-1865 Union maps.(5) National Park Service maps(6) of the 1930's show a series of parallel depressions indicating work was underway on an additional bombproof when the fort was abandoned in December 1864.

The chamber was built entirely below ground surface producing low oxygen conditions conducive to preservation. Wood decomposition did not occur because the sand overburden collapsed and compacted the timbers. When exposed to the air, wood was damp and deteriorated rapidly. Rotten wood was located along the floor and could have decayed when the structure was still open.

When the structure collapsed, roof timbers sank into the chamber. Some planks still rested on supporting pillars but sagged as much as three feet. Support pillars were found leaning toward the walls because overburden weight forced them outward.

The excavated section of bombproof was about twenty feet wide and thirty feet long. The total chamber length was at least 150 feet based on a depression running the length of the mound's top. Settling and warping, slanting and compacted timbers prevented more accurate measurement. The western entrance to the bombproof probably started about 10 to 13 feet south of the fireplace.

A massive brick fireplace was situated at the midpoint of the chamber's west wall. The fireplace is the centerpiece of the excavated portion of the bombproof due to its size. It was constructed of large reddish-brown bricks averaging 8 3/4" x 4" x 3". From base to top, the fireplace measures 8 feet 9 and 3/4 inches. The width across the front is identical to the height. From front to rear, the fireplace measures 4 feet 7 and 1/2 inches, the distance from the hearth floor to the bottom of the arch. Hearth width is 5 feet 11 inches. Smoke was passed through a wooden flue built on top of the fireplace.

To compensate for interrupting the western roof sill, supporting pillars 6 inches square were placed in front of the fireplace on both sides. A beam rested on top of these pillars supporting the roof. Whitewash residue on the bricks indicates the fireplace pillars slanted inwards. The whitewash "shadow" also shows interior clearance was 5' 10".

Two rows of pillars ran down the chamber center. The 6 3/4 x 7 inch pillars were placed on 2 feet 10 inch centers, mortised into floor sills laid four feet apart. The pillars also helped support the chamber walls. Braces were attached to studs which kept the plank walls against the dirt face of the chamber excavation. The braces were mortised into alternate pillars. Wall planks were butt-jointed and held in place by the force of the dirt mass behind the wall.

The three bottom planks were 9, 10, and 8 inches wide and about 3/4 of an inch thick. This last measurement seems unusually thin and may be due to compaction. The walls slanted out and up at an angle of about 75 degrees. These planks were probably whitewashed.

Sleepers parallel to the floor sills rested on the pillars to support the roof over the room's center. The roof timbers were laid directly on, and perpendicular to, the sleepers and overlapped the sills. The roofing sills were laid in the dirt, a yard outside the chamber excavation.

No nails were found fastening any of the roofing planks to the sills. Roofing timbers were in widths ranging from 12 to 20 inches and at least 4 inches thick. Bark may have been placed over the roof planks to keep dirt from falling through cracks into the chamber. At Fort Fisher, tarred canvas and possibly oak leaves covered the roof.(7)

The mound was constructed by piling soil over the timbers covering the chamber. Fill came from the chamber excavation and from the fort interior. Much of this work was accomplished with shovels and wheelbarrows judging from documents relating to other Savannah defenses.(8) The mound was finally covered with sod.(9)

Archaeological evidence provided enough information for a graphic reconstruction (Figure 1). The slanting walls were noted in profiles cut through the chamber. A similar slant was noted in the pillars on the fireplace front as shown by whitewash "shadow".

Figure 1 - Conjectural reconstruction of the Fort Bartow bombproof and probable placement inside mound.

The entrance is conjectural. It is suggested by a few recovered features. Considerable damage was done to the entrance way by looters who attempted to dig into the chamber through a depression in the wall at this point prior to the archaeological work. Portions of the depression which the looters followed were still extant. The depression suggests that the entrance had been in this vicinity and had collapsed.

Once the interior chamber was constructed at Fort Bartow, it was buried under 5000 cubic yards of earth. The mound was probably sodded. No direct evidence for sodding was found at the site.

The fireplace not only provided a source for heat and light but created a draft to draw fresh air through the chamber. A depression on the south end of the mound suggests that a fireplace may have been located there as well, ensuring fresh
air throughout the chamber.

It is possible that a covered area existed on the west, or interior, side of the bombproof. Evidence for the existence of a roofed area is tenuous but has a good fit with information recovered during the archaeological project. A post mold found west of the bombproof was a large one, capable of supporting large timbers. Soil around the post mold was highly compacted in a manner suggestive of a living floor.

Further evidence for a covered area is provided by a dip inside a low mound ten feet west of the bombproof. The dip and low mound have been interpreted as the border of the roofed area. That is, the mound built up outside the roofed area as materials fell from the roof or was thrown/swept from the living floor area. The dip was the result of human traffic compacting the floor and some excavation. Admittedly, evidence for a roof is slim but it does have a very good fit with Confederate earthworks at Fort Fisher, Wilmington, NC.(10) If the roof existed, it was built up with planks resting on a timber framework and then covered with earth. Since the bombproof was in the direct line of fire, the roof did not have to be extremely thick and it could remain open to the fort interior.



  1. Robert H. Anderson, "Report on Engagement" 1 Feb. 1863 cited in Civil War Naval Chronology (GPO, Washington, D.C., 1971), p. 111-20.; Letter from P. G. T. Beauregard, cited in Civil War Naval Chronology (GPO, Washington, D.C., 1971), p. ll l-20.
  2. L.E. Babits, "A Confederate Earthwork's Internal Structure", MC&H, XLI, 194-198.
  3. Letter from Henry Bryan to General Thomas Jordan, 1 Feb. 1862 in The War of the Rebellion. Official Records of the Union and Confederate Armies. (GPO, Washington,D.C., 1882),pp.XIV: 212-13; Dennis H. Mahan, A Complete Treatise on Field Fortification. (1836, reprinted Morningside Books, Dayton, 1968), p. 30.
  4. Anonymous Plan of Fort Bartow, National Archives Record Group 109, Box 36, 1-136.
  5. Charles 0. Boutelle, Reconnaissance of Wilmington River and St. Augustine Creek. Copy of Printed Map on file, Coastal Heritage Society, Old Fort Jackson, Savannah. Map Collections, Georgia Historical Society, Savannah, GA
  6. Fort Bartow on Causton's Bluff. Undated, Unsigned Map Set produced by the National Park Service, on file at Fort Pulaski National Monument, Tybee Island, GA.
  7. Gordon P. Watts, Jr. et al, Excavation of a Fort Fisher Bombproof (North Carolina Department of Cultural Resources, Raleigh, 1981), p. 26.
  8. Edward C. Anderson, Diary, dated 2 November 1863 - 13 November 1864. Typescript on file, Coastal Heritage Society, Old Fort Jackson, Savannah, GA. Entries dated 25 January and 18 March 1864.
  9. Anderson, Diary, entry dated 15 December 1863; Report of Captain John McCrady, dated 8 March 1863, cited in Edward K. Rawson, G. P. Colvocoresses, C. W. Stewart, Official Records of the Union and Confederate Navies in the War of the Rebellion. (GPO, Washington, D.C., 1901), XIII 730-4.
  10. William C. Davis, ed., The Image of War (Doubleday and Company, New York, 1982), III, pp. 176-7.