I am indebted to John Lefgren who supplied the data that got me started on this research project.
Brad Lepper, Curator of Archaeology for
Ohio History Connection, believes there were once 10,000 discrete mounds, 600 elaborate earthworks such as
Serpent Mound, and 14 hilltop enclosures such as
Fort Ancient in Ohio. Modern agricultural plowing and tilling have destroyed most of them. A notable Ohio earthwork no longer obvious to the untrained eye goes by the name "East Fork Works" (EFW) referring to its location on the East Fork of the Little Miami River. The East Fork of the Little Miami flows through Clinton, Highland, Brown and Clermont Counties in SW Ohio. As with all images on this blog, click to enlarge.
![]() |
| East Fork of Little Miami in Yellow |
The EFW was originally mapped in 1803 by William Lytle (1770 - 1831), a prominent early Ohio surveyor. Lytle founded Cincinnati College, now the
University of Cincinnati. The Lytle map was among the papers Thomas Jefferson perused before finalizing the Louisiana Purchase.
A map of the EFW first appeared in print in Hugh Williamson,
Observations on the Climate in Different Parts of America (New York: T. & J. Swords, 1811, p. 195). Hugh Williamson (1735 - 1819) was a signer of the US Constitution and a scholar of international reputation.
![]() |
| East Fork Works Williamson 1811 |
Williamson says the wall labelled "a" was 2,100 feet long. The letter "b" represents a valley. The letter "c" is a spring with multiple seeps. The letter "d" is low ground. "B" indicates a branch of the Miami River which is a mistake on Williamson's part. The caption should have said "B" indicates a branch of the Little Miami River.
Another EFW map was created in 1823 by Major Isaac Roberdeau (1763 - 1829) of the US Army Corps of Engineers. Roberdeau assisted Pierre Charles L'Enfant (1754 - 1825) as a surveyor and engineer laying out the city of Washington, D.C.
![]() |
| Milford and East Fork Works Roberdeau 1823 |
This is a closeup of Roberdeau's EFW:
![]() |
| Closeup East Fork Works Roberdeau 1823 |
This map is in the
National Archives and
Science Views has copies at various resolutions available for download.
A fourth version of the map, obviously redrawn after Roberdeau, was printed in David Baillie Warden, Recherches sur les Antiquités de L'Amérique Septentrionale (Paris: Everat, 1827, Plate XI). Warden (1772 - 1845) was an Irish-American scholar.
![]() |
| Milford and East Fork Works Warden 1827 |
This is a closeup of Warden's EFW:
![]() |
| Closeup East Fork Works Warden 1827 |
A widely-distributed map of the EFW, embellishing Roberdeau with additional topographical details, appeared in the Smithsonian Institution's very first publication: E.G. Squier and E.H. Davis,
Ancient Monuments of the Mississippi Valley (Washington, D.C.: Smithsonian Institution Contributions to Knowledge Vol. 1, 1848, plate XXXIV). The authors aptly called this site "singular," an "extraordinary outline."
![]() |
| East Fork Works Squier & Davis 1848 |
Ephraim George Squier (1821 - 1888) was an American archaeologist and newspaper editor. Edwin Hamilton Davis (1811 - 1888) was an American archaeologist and physician. For their landmark 1848 publication, they ground-truthed many of the earthworks they described, but not the EFW (they said they could not vouch "for the entire accuracy of the plan" and "its thorough investigation is an object greatly to be desired.") So, it is likely that the five EFW maps created through 1848 (Lytle 1803, Williamson 1811, Roberdeau 1823, Warden 1827, and Squier & Davis 1848) were all based on Lytle's original survey.
In 1894 the Smithsonian Institution published an encyclopedic work on the mounds and earthworks in many parts of the US: Cyrus Thomas, Report on the Mound Explorations of the Bureau of Ethnology (Washington, D.C.: Smithsonian Institution Twelfth Annual Report of the Bureau of Ethnology 1890 - '91, 1894). On p. 566 Thomas said ground-truthing by assistants demonstrated the Squier & Davis 1848 map was "to a large extent imaginary." Thomas did not say the EFW did not exist, just that published maps were inaccurate. Unfortunately, the Bureau did not create an improved map, probably because by 1890 decades of agriculture had obliterated many site features.
Gerard Fowke, Archaeological History of Ohio (Columbus: Ohio State Archaeological and Historical Society, 1902) reproduced a great deal of content from the 1848 and 1894 Smithsonian publications. On p. 218 he rotated the Squier & Davis 1848 map 90 degrees, called the parallel walls a "gridiron," located it in Clermont County, and expressed skepticism about the map's accuracy.
![]() |
| East Fork Works Fowke 1902 |
William C. Mills,
Archaeological Atlas of Ohio (Columbus: Ohio State Archaeological and Historical Society, 1914) employed assistants to verify many sites before including them on his maps by township and county. When he came to Clermont County, he simply reproduced Fowke's rotated "gridiron" image without elaboration.
The EFW, one of the most unique archaeological sites ever documented, had versions of its map published 6 times over 103 years in two languages before it faded into relative obscurity for decades.
All that changed in 1988 when a local newspaper in Marietta, Ohio ran an article suggesting the EFW image in Squier & Davis 1848 looked like a Jewish menorah or 9-branched candelabrum enclosed within an effigy of an ancient Near Eastern oil lamp.
![]() |
| Jewish Menorah for Celebrating Hanukkah |
The EFW maps with 9 parallel walls did resemble a stylized menorah and the top part of the enclosure walls did resemble an ancient oil lamp with a finger ring.
![]() |
| Ancient Oil Lamp, The National Maritime Museum, Haifa, Israel |
Jews in ancient Ohio? Suddenly people who believed in prehistoric transoceanic contacts had something new to talk about besides the dubious Bat Creek Stone (currently on display at the Museum of the Cherokee Indian in North Carolina) and the clearly fraudulent Newark Holy Stones (currently on display at the Johnson-Humrickhouse Museum in Coshocton, Ohio). Before long, articles about the "Menorah Mound" were circulating among members of the
Church of Jesus Chirst of Latter-day Saints already pre-disposed by the
Book of Mormon toward a belief in cultural diffusion. (Full disclosure: I am an active member of the Church of Jesus Christ of Latter-day Saints and believe the Book of Mormon is the earliest ancient American codex currently known to science.) When the ultra-diffusionist magazine
Ancient American published a
1996 article by J. Huston McCulloch, a PhD economist with a strong
interest in archaeological outliers, mythology around the "Menorah Mound" increased. That mythology today often includes an element of conspiracy theory: a paradigm-changing earthwork built by Jews who once lived in ancient Ohio was known to Thomas Jefferson, but contemporary knowledge of it has been systematically suppressed by mainstream archaeologists who deny cultural diffusion.
The much more likely scenario: a sophisticated earthwork built by ancestors of modern Native Americans was surveyed in 1803 and visited occasionally throughout the nineteenth century as it gradually disappeared under the plow. Partially accurate maps based on the original survey were published through 1914, but today the location of this extraordinary site is generally unknown.
So where is this so-called "Menorah Mound?" It could be completely plowed under by now and only detectable by a geophysical survey using magnetometry, ground penetrating radar, or electrical resistivity. On the other hand, archaeological remains can sometimes be detected via high resolution satellite imagery using tools such as Google Earth. I invite you along as we apply standard reconnaissance techniques to locate this most interesting ancient earthwork.
Historical sources describe the EFW as being on the East Fork about 20 miles east of Milford, Ohio, and about 20 miles from the confluence of the East Fork of the Little Miami with the Little Miami. On the map below, the East Fork is in yellow and the red circle has a radius of 20 miles from the confluence of the East Fork with the Little Miami. The teal circle has a radius of 20 miles from Milford center.
![]() |
| 20 Mile Radius Circles around Base Points |
The point where these two circles intersect the East Fork is probably in the ballpark of the area we seek. This points up the first discrepancy with published sources. Fowke 1902 followed slavishly by Mills 1914 said the EFW was in Clermont County, but the mileage estimates in earlier published sources would suggest the EFW is actually in the northern panhandle of Brown County.
Early published sources give us another way to triangulate location. There was a major ancient earthwork located on the East Fork about 1 mile from Milford. That would put it about where Greenlawn Cemetery is today. Archaeologists refer to this site, long since obliterated by urbanization, as the Milford Works.
![]() |
| Milford Works Site near Modern Greenlawn Cemetery |
4 miles above (upstream from) Milford Works was another ancient site with no name attached.
![]() |
| Unnamed Site 4 Miles Upstream from Milford Works |
According to the early sources, the EFW was another 20 miles above (upstream from) this unnamed site. That would put it where the village of Marathon is today on the Clermont County/Brown County line.
![]() |
| Approximate Location of EFW 20 Miles Upstream from Unnamed Site |
How accurate are these mileage estimates (20 miles from Milford, 20 miles from the confluence, 24 miles upstream from Milford Works) likely to be? Fairly accurate. The sources say "about" but these were surveyors who made their living laying out plats and maps. One generic standard deviation around 20 miles would be 34% or 6.8 miles. One half of a standard deviation would be 3.4 miles. We plot a 3.4 mile radius circle around the two 20 mile points of interest.
![]() |
| Half Standard Deviation Circles around 20 Mile Distance Points |
The intersection of these two 3.4 mile radius circles is the sweet spot where we will begin our search for the EFW. And what should we expect to find? J. Huston McCulloch's experience pinning down the
better documented Milford Works shows that we may have to rotate the map as Fowke 1902 and Mills 1914 did. We may even have to turn the map over and use the mirror image. Williamson 1811 and Roberdeau 1823 show the spring flowing in different directions offset about 90 degrees from each other. We know the scale on the published maps cannot be right. Roberdeau says the parallel walls were on 66 foot centers. Surveyors were unlikely to mistake that distance. A surveyor's chain in the early 1800's was exactly 66 feet (4 rods) long. 66 X 8 = 528 feet so we would expect the parallel wall structure to be near that width. But, on all published maps the parallel walls dominate the total enclosure which is described as 2,000 or 2,100 feet on a side and about 100 acres in total area. 2,100 X 2,100 = 4,410,000 square feet. One acre = 43,560 square feet. 4,410,000/43,560 = 101.23 so 100 acres is a reasonably accurate number. Surveyors were unlikely to be too far off on that measurement. But, a parallel wall structure 528 feet wide could not have had more than about 4.5 acres in total land area. That is less than 5% of the enclosed space, while the maps make the walls look like they take up at least 33% of the enclosure. The mapmakers were obviously illustrating what they considered the most notable feature of the site without regard to accurate scale.
Based on published maps and descriptions, these are EFW characteristics we expect to find:
1. A site about 100 acres in area.
2. At least one lateral measurement close to 2,100 feet.
3. A spring flowing away from the site toward the East Fork of the Little Miami.
4. The spring will have at least 3 sources (seeps) coming together to form a single stream.
5. Some marshy bottom lands (low ground).
6. A valley with a stream flowing through it.
7. A noteworthy structure with 9 parallel walls.
8. The 9 parallel walls will be perpendicular to a base line.
9. The base line on one end will extend beyond the 9th wall.
10. The 9 parallel walls will be different lengths.
11. Some of the 9 varying length parallel walls will taper evenly to a point.
12. Some of the 9 parallel walls will flare out on their ends.
13. The 9 parallel wall structure will occupy less than 5% of the total site area.
14. The 9 parallel wall structure will be on the order of 528 feet wide.
15. Some parallel walls will be 66 feet apart.
16. Some walls will be at 90 degree right angles to each other.
17. Some walls will be at 45 degree or other angles to each other.
18. There will be at least 3 sets of parallel walls at the site.
19. Two shorter walls will be diverging rather than parallel to each other.
20. One part of the site will have curved walls in a knob or tongue shape.
21. The curved knob or tongue-shaped walls will be on the same side of the site fronting the river as the stream flowing from the spring.
22. The valley with a stream flowing through it will be on the opposite side of the site as the 9 parallel wall structure.
23. The low ground will be on a different side of the site than the valley with the stream flowing through it and the curved knob or tongue-shaped walls.
24. The curved knob or tongue-shaped walls will be on the opposite side and corner of the site as the 9 parallel wall structure.
25. The curved knob or tongue-shaped walls be be oriented generally in the same direction as the East Fork stream flow.
26. Hilly topography will exist between the site and the East Fork of the Little Miami.
Remains of ancient structures on this triangle-shaped piece of ground along US 50 2 miles SW of the village of Fayetteville in Perry Township, Brown County, Ohio, do show up on Google Earth. They are in the area we identified as the sweet spot based on our location analysis and they meet all 26 of our site correlation criteria with flying colors.
![]() |
| The East Fork Works Have Almost Certainly Been Identified |
We will now visually show this correlation's high degree of fit. First, the Perry Township site is in the area we identified as the most likely location for the EFW.
![]() |
| Sweet Spot Intersection of Two 3.4 Mile Radius Circles |
1. The Perry Township site is about 100 acres in size. The Brown County Land Map shows 3 adjacent tracts: tract 0341, is 96.12 acres; tract 0449 is 1.69 acres; and the Veracruz Cemetery, tract 1051, is 4.76 acres.
![]() |
| Brown County, Ohio Land Map |
These 3 adjacent tracts total 102.57 acres. Criterion #1 satisfied.
2. The proposed site has two lateral lengths that measure precisely 2,100 feet.
![]() |
2,100 Feet Measured along Western Edge
The Northern Edge also Measures 2,100 Feet in Length |
Criterion #2 satisfied.
3. A spring feeds a small stream that flows out of the site, into Glady Run, which joins the East Fork of the Little Miami.
![]() |
| Thin Yellow Line Represents the Stream Flowing from the Spring |
Criterion #3 satisfied.
4. The spring has 3 sources (seeps) that come together to form the stream leaving the site. Two of these source streamlets have tiny tributaries of their own.
![]() |
| Thin Yellow Lines Represent 3 Spring Sources Coming Together |
Criterion #4 satisfied.
5. The site has an area of marshy bottoms or low ground that drains into the spring-fed stream leaving the site.
![]() |
| White Overlay Shows Marshy Bottoms or Low Ground |
Criterion #5 satisfied.
6. Glady Run is a tributary of the East Fork of the Little Miami flowing through a valley bordering the site on the west. 1,000 feet north of the site Glady Run has been dammed to form Lake Lorelei.
![]() |
| Glady Run Flowing Southward through its Valley |
Criterion #6 satisfied.
7. And here is the unusual structure with 9 parallel walls that has had people excited since 1803.
![]() |
| Ancient Structure with 9 Parallel Walls |
Criterion #7 satisfied. These parallel walls also show up in Google Maps Street View. This is the view from Highway 131 looking south at the NE corner of the 9 parallel wall structure:
![]() |
| 3 of 9 Parallel Walls Photographed by a Google Car in May, 2018 |
In the image above, the 3 down arrows point to parallel walls. The up arrow points to the base line.
8. The 9 parallel walls are perpendicular to a base line.
![]() |
| 9 Parallel Walls at 90 Degree Right Angles to a Base Line |
Criterion #8 satisfied.
9. The base line on its west end does extend beyond the 9th wall.
![]() |
| Base Line Extending Past 9th Wall |
Criterion #9 satisfied.
10, 11. The 9 parallel walls are in fact different lengths and some of them do taper evenly to a point.
![]() |
| Variable Length, Tapering Parallel Walls |
Criteria #10, 11 satisfied.
12. Two of the 9 parallel walls do flare out on their ends.
![]() |
| Parallel Walls with Flared Out Ends Identified in Red |
Criterion #12 satisfied.
13. The parallel wall structure is 501 feet long X 399 feet wide which equals 199,899 square feet or 4.59 acres. This is less than 5% of the surface area of the site (approximately 100 acres).
![]() |
| Parallel Wall Structure 501 Feet Long |
Criterion #13 satisfied.
14. The parallel wall structure base line is 451 feet long which is reasonably close to the the 528 feet we would expect based on the Roberdeau 1823 scale.
![]() |
| Wall Structure Base Line 451 Feet Long |
Criterion #14 satisfied.
15. The 9 parallel walls are only 50 feet apart, but two other sets of parallel walls at the site are spaced precisely 66 feet apart. Roberdeau probably applied this well-known 66 foot (4 rod) measurement to the 9 wall structure.
![]() |
| 3 Parallel Walls Laid Out on 66 Foot Center Lines |
Criterion #15 satisfied.
16, 17. Many walls at the Fayetteville Township site are joined at 90 degree right angles. Others are at 45 degree and lower angles.
![]() |
| Ancient Walls Visible in Google Earth Traced in White |
Criteria #16, 17 satisfied.
18. This site is so complex it has 5 different sets of ancient parallel walls.
![]() |
| Parallel Wall Sets Identified in Red |
Criterion #18 satisfied.
19. Two walls are diverging rather than parallel.
![]() |
| Diverging Walls Traced in White |
Criterion #19 satisfied.
20. One part of the site does show walls curved in a knob or tongue shape.
![]() |
| Curvature |
Criterion #20 satisfied.
21. The knob or tongue-shaped curvature is on the southwestern side of the site. The stream flowing from the spring runs southwesterly.
![]() |
| Southwest Side of the Site has Curvature, Spring-fed Stream |
Criterion #21 satisfied.
22. Glady Run flowing in its valley is on the western edge of the site. The 9 parallel wall structure is on the eastern edge of the site.
![]() |
| Valley to the West, Parallel Wall Structure to the East |
Criterion #22 satisfied.
23. The valley with the stream flowing through it and the curved knob or tongue feature are on the western side of the site. The marshy low ground is on the south.
![]() |
| Glady Run, Curvature to the West, Marshy Low Ground to the South |
Criterion #23 satisfied.
24. The curved feature is on the SW corner of the site, the 9 parallel wall structure on the NE corner.
![]() |
| Opposite Corners |
Criterion #24 satisfied.
25. The curved knob or tongue-shaped walls are oriented generally southwesterly. The East Fork of the Little Miami in this area flows generally southwesterly.
![]() |
| Curved Feature Orientation, East Fork Flow Both Southwesterly |
Criterion #25 satisfied.
26. Turning on the Google Maps Terrain Layer clearly shows the hilly topography Squier & Davis 1848 mapped between the site and the East Fork of the Little Miami. The East Fork and its unnamed tributary east of the site have carved a small canyon with 50-foot-high walls.
![]() |
| Hilly Topography Between the Site and the East Fork |
Criterion #26 satisfied.
So how did the site we see today on Google Earth end up as the image that got published 6 times between 1811 and 1914? Flip the various images and it is not hard to see how this:
![]() |
| Selected EFW Features in Context |
Came to be represented like this:
![]() |
| Williamson 1811 Flipped |
And like this:
![]() |
| Roberdeau 1823 Flipped |
And eventually like this:
![]() |
| Squier & Davis 1848 Flipped |
If the direction of the East Fork stream flow seems incongruous, keep in mind that the site adjoins two small meanders. Within 1/2 mile of the site, the East Fork of the Little Miami flows N, S, E, and W.
East Fork Works (EFW) is one of the most complex North American earthworks ever built and there is neither a menorah nor an oil lamp in sight.
How accurate are my tracings overlaid on Google Earth? The lines in the earth I think I see via Google Earth do exist. You are welcome to
download the KMZ file and verify them. Ground truthing via magnetometry or ground penetrating radar would reveal many more features and fill in a number of gaps. But, the site is a full 100 acres and either of those geophysical survey techniques would require a solid week of work with good equipment which would be quite expensive.
In Ohio archaeology, it is well known that mounds and more complex enclosures often cluster in groups, and that is exactly what we see with the EFW. Less than a mile away more than a dozen other ancient parallel walls show an obvious cultural connection.
![]() |
| Similar Ancient Structures Nearby |
I scanned upriver and downriver for several miles via Google Earth and found no other sites of interest, so I believe these are the significant ancient structures riverside in this region.
Others deserve the credit for first correlating this Perry Township site with the EFW. In the material John Lefgren sent me, he said James Hill identified this site. All I did was bring up Hill's proposal on Google Earth and the pieces quickly fell into place with assistance from
Prof. McCulloch and author
Jason Colavito. Ryan Dahle shared this
embarrassingly naive video which also makes the association.
Even though he mis-identified the well-known EFW image as being in Clermont County following Fowke 1902, William C. Mills in his 1914
Archaeological Atlas of Ohio shows an enclosure site in Perry Township at the confluence of Glady Run with the East Fork of the Little Miami.
![]() |
| 1914 Map Showing the East Fork Works |
The mound has been found.
Kirk Magleby volunteers as Executive Director of
Book of Mormon Central, world's leading source for reliable Book of Mormon enrichment and contextual content in English and
Spanish.
