Pretend Swordplay: Hercules Swords and the Hornets’ Nest of Reality

When I was an adolescent, among my favorite movies were Zorro, The Three Musketeers, and Conan the Barbarian. My friends and I spent countless hours pretending to be swashbucklers with thin sticks that we lunged and parried with in various places like my best friend’s back yard where the grass grew tall. One year, unknown to us both, the grass wasn’t the only thing that grew with fruition. As I parried and traversed my friend’s thrust, I stepped right through the home of Dolichovespula maculata. The bald-faced hornet!

At the first sting or two on my arm and leg, the sudden realization of my mistake was clear. And like the cartoon protagonist with an arrow of flying insects after him, I ran faster than I’ve probably ever run before, arriving at the doorstep of my friends house with his mother ushering me quickly inside. Three hornets were able to keep up. And they were out for blood. But my friend’s mom was a… mom. Battle tested and a quick aim with a can of wasp spray. She nailed all three. I think I suffered six stings all together.

That event is one of the many bookmarks in the memories of my childhood that remind me of how far back my fascination with swords goes. I suspect I’m no different from many other boys and probably even girls who grew up watching many of the same movies, reading many of the same books, and having many of the same sorts of day dreams of swashbuckling pirates, masked defenders of the common man, and the loyalty and dedication to great causes that come with being one for all and all for one.

And so when a grandiose story about a sword makes the news in the way the alleged Roman Hercules sword did in association with the television series “The Curse of Oak Island” (History Channel), people pay attention. Skeptics and believers alike.

Quick Background
In case you are a reader unaware of what has been dubbed by some as “Swordgate,” or of the “Oak Island” TV series, here’s the quick rundown.

For decades, there has been a legend about great treasure to be found on an island in Nova Scotia and, in recent times, two brothers with a little cash to spend, bought much of the island and embarked on a multi-year effort to recover the treasure which the History Channel was all-to-eager to televise for at least four seasons. The show rates 6.9 out of 10 on IMDB and has a huge following through several Facebook groups where every detail is discussed after each episode.

Nova Scotia Hercules sword, courtesy of The History Channel.

Around episode 10 of the third season, in January, the brothers are presented with a sword that was alleged to have been discovered by a scalloper and his son decades before, caught in their rakes then held as a treasured keepsake for years. The sword is alleged to be Roman in origin, bronze, over 2,000 years old, and evidence that there was a pre-Columbus visit to Oak Island by Romans.

News of the sword and its alleged importance made its way around the internet about one month prior as J. Hutton Pulitzer, a self-described “researcher” and “Treasureforce Commander,” pronounced it a “smoking gun” that supports the claim that Europeans were coming to the Americas long before Columbus made the trip.

Looking over Pulitzer’s posts and comments on Facebook, Twitter, and various blogs of a year ago (and even today), it’s clear that he consistently insists that the sword that turned up in Nova Scotia and was featured on the Oak Island television show, was 100% of Roman origin, dating to about the first century AD. Indeed, Pulitzer insisted that he was to publish a white paper with the AAPS (Ancient Artifact Preservation Society) which would reveal the data and evidence which “100% confirms” the date and origin of the sword itself.

As of January 2017, no such white paper has materialized, though Pulitzer did release a colorful, if disjointed PDF that was largely a copy and paste compendium of semi-related material. This was titled “Oak Island Roman Sword Vs. St. Mary’s University, Halifax, Nova Scotia, Canada as Authored by Dr. Christa Brosseau (a.k.a. Brosseau Report),” and dated to February of 2016. This was Pulitzer’s response to the testing the brothers on the television show paid for at the renowned St. Mary’s University.

Some Added Detail

You see, just a little over a month after Pulitzer made his assertion that the sword was “100% confirmed” to be of 1st century Roman origin, the sword was featured on episodes 10 and 11 in season 3. The very public test, using a scanning electron microscope (SEM) and energy dispersive x-ray spectroscopy (EDS) as two of the primary instruments concluded that the sword was of modern manufacture. This was something that Pulitzer could not, apparently, accept–perhaps since it effectively left him out of the story of Oak Island; perhaps because he truly believes the claims he’s generated. Either way, Pulitzer has continued to claim that he and his team conducted an elemental analysis using a portable x-ray florescence (pXRF) device. This is an instrument that is very similar to the EDS that Brosseau used in her analysis.

Nova Scotia sword being analyzed by St. Mary’s University, courtesy The History Channel.

The primary goal of this “200 page report” that Pulitzer released was clearly an effort to discredit Brosseau, but one that was only successful with those unfamiliar with real scientific reports and assessments. Whereas Brosseau was careful to provide a concise introduction, a detailed and step-by-step walk-through of her methods, and an informative section that discussed the results of her efforts, Pulitzer produced a flashy, rhetoric-ridden, full of photographs and text, much of it liberally copy and pasted from sources often without attribution. Ironically, he was careful to make himself the copyright bearer for all those who provided him with graphics and text by plastering a copyright notice on each page.

From Peter Geuzen’s Swordgate Poster hosted by Andy White. Note the finer detail with the California sword’s hilt when compared to the Nova Scotia version.

It’s also fascinating to compare how the two provided references and notes. Whereas Brosseau did so the expected fashion that any scientist would: with a half-dozen or so relevant citations linked to specific statements within the text that were ordered and consistent, and a brief set of bulleted notes that offered information relevant to the entire process; Pulitzer provided hundreds of bibliographic entries and entire pages from journal papers (some of which seemed to obviously contradict his position!). And not just those that were relevant, but over 80 pages of journal copy/pastes and 22 pages of seemingly random references. I say “seemingly,” but he did, at least, categorize them as being references for “SEM” or “EDS” or “variance in alloys,” etc.

It was as if Pulitzer was attempting to show that he’s smarter than Brosseau because his bibliography was bigger!

Pulitzer’s argument in his “200 page report” amounts to the kind of Neuro-linguistic programming and rhetoric that bad business leaders use to stay ahead of their competition or to gain temporary public approval. They do things like re-enforce how great things are by saying “things are great!” over and over. They tell you the bad news but tell you how it doesn’t matter because, “by golly, we’re doing good things and that’s what matters!” In the end, that’s the kind of business leader that leaves employees, investors, and consumers holding the bag. The common person, not understanding the behind-the-scenes details only remembers those buzz words: “great,” “fantastic,” “wonderful,” “the good we’re doing”…

And so it goes with Pulitzer’s “200 page report.” He repeats over and over how Brosseau doesn’t follow the scientific method and how she’s just wrong. She’s wrong.

But was she? Pulitzer never really outlines how this might be. His “report” doesn’t include his alleged pXRF data. He’ll tell you it does if you ask him. But he won’t tell you what page. Instead, he’ll grill you on “What version of the report do you have? What’s the date? They’re all different, so I can’t answer your question until I know….” But, Hut… it’s a “200-page report.” Wouldn’t the pagination be 1 to 200?

As far as this sword is concerned, it appears to be one of about 21 variations of a set of decorative swords, probably originally sold as souvenirs but now available even on Tell-tale indicators of a common mold are present among many of the variations, and Andy White hosts an excellent poster (created by Peter Geuzen) and a series of blog articles that describe the variations from sword to sword. With careful examination, it’s easy to see that an inexpensive sword, originally purchased as a souvenir in 1975, has finer details and resolution, implying that its mold was of better quality than the mold used in the Nova Scotia sword. In fact, if you look closely at the Nova Scotia sword, you can see filing marks on the sides where the flash was filed down. Were this a genuine sword, greater care would have been taken in the molding process.

Molding Process of copper alloys

Bronze sword in its ceramic mold. Courtesy

When used to create swords or even objects like sculptures, ewers, candlesticks, and even architectural items, copper alloys like bronze and brass are poured into molds, casting their final form into a rough product that is further refined into its final form.

Well before brass alloys came along, bronze was used as a medium for swords. The swordsmith created a mold from one of several methods. A model of the sword is first created, perhaps out of wood. Then a casting in ceramic is made of opposite sides. The casting is fired to create a mold. For a simple sword with little detail, the casting itself can be used as a mold for the liquid bronze. The two halves are placed together, a pouring cup is at one end of the cast, probably the tip of the sword since this would be an easy place to cut it off and shape the blade afterwards. The crucible, with molten copper and a bit of added tin, is then poured into the mold.

However, the details on the Nova Scotia Hercules sword would have required a bit more care. In order to allow the bronze to flow more freely and gases to escape, some vents, called sprues or gates, must be added. In a mostly linear form of a normal sword, this is not usually a problem. The figurine of Hercules that forms the handle, however, has many nooks and crannies that can trap gases as the molten metal (probably at a temperature of over 2100 degrees F) enters the mold. In this case, the method of choice would be the cire perdue, also known as the “lost wax” method.

This wasn’t called “lost” because the method was ancient and secret, rather because of the method itself. The original model would be cast in ceramic (or, today, in a silicone or alginate base) to create a mold for wax. The melted wax is poured into the mold, and the the mold removed once cooled. Additional wax sprues are added at various points around the object (heat the spot, press in a wax stick). Think of how the plastic model airplane kits for kids are shipped with parts in a framework of plastic that is broken or cut off and you have the idea. These sprues, or gates, act as vents for the gases in the final mold.

Bronze cast after ceramic mold removed but spruing still attached. Photo by José-Manuel Benito Álvarez

Once the wax model is modified with the sprues, it is then dipped in a ceramic slurry, coated with fine dry clay, allowed to dry, then dipped, coated, dried 2 or 3 more times. The ceramic is then fired, and the wax is “lost” in the process because it drips out during firing. What’s left is a hollow, ceramic mold that the molten bronze can be poured in.

In all likelihood, the molds for the Hercules style swords (including the one alleged to have been found in Nova Scotia) were probably two-part ceramic or soapstone pieces that fit together like clamshells. This is based on the seams apparent in the close up photos of the swords (including the Nova Scotia version). These seams are very roughly ground with obvious striations visible on the lateral edges.

Close examination of the known versions of the Hercules sword (to date there are at least 21) shows varied degrees of detail. Some swords have a Hercules figure that is well-defined. Others, not so much. The Nova Scotia sword is one that is not so well defined. The obvious conclusion for anyone familiar with copying techniques used for both fraud and commerce (legit copies not claimed to be originals are just good commerce), is that a wax mold was made of a good copy; then a wax mold made of that copy; and so on. If you’ve ever tried to photocopy a high-resolution photograph, then copy the copy a few times, you understand what I mean. There’s a loss of fidelity.

The fact is, the Nova Scotia Hercules sword has less fidelity than others of the 21.

As mentioned above, the Nova Scotia Hercules sword was tested by a scientist at St. Mary’s University in Halifax, Canada. The composition was determined to be 56% copper and 35% zinc. Moreover, she determined that the copper itself was very pure and must have been refined. These are important facts which are indicative of the period in which this sword could have been manufactured.

Bronze, as an alloy of copper, begins to show up routinely in the archaeological record by about 3500 BCE in Europe. Typically, bronze is about 80-90% copper and 10-20% tin. Some bronzes have traces of other metals or non-metals ranging from lead to arsenic. Bronze continues to be the metal of choice for centuries to come, but over time, many additional copper alloys are introduced. Brass being one such alloy. Like Bronze, brass is mostly copper offset by varied percentages of zinc instead of tin. While it is definitely possible to have zinc in some bronzes, it is rare that there is tin in a brass. Among the rare exceptions is red brass, which contains as much as 5% tin along with 5% each of lead and zinc offsetting the remainder in copper at 85%. Interestingly, red brass is considered both a brass and a bronze since it effectively straddles the qualities of both alloys.

As Paul Craddock mentions in his Scientific Investigation of Copies, Fakes, and Forgeries (New York: 2001), “It is surprising how many forgeries are given away by the presence of zinc in the alloy” (p. 70). He goes on to say, “the presence of zinc in a copper alloy in quantities above trace amounts is such a widely used and potent indicator of recent date…”

Zinc occurs naturally in and around copper ores and finds its way into bronzes (there’s still tin in the alloy). But these are trace amounts and the majority of this sort of zinc would be expected to be vaporized in the furnace, escaping as a gas rather than mixing with the copper. Lowering the temperature during smelting allowed the zinc to stick around a little longer and mingle. The earliest brasses emerged around the first millennium BCE, mostly as coins and rarely with zinc percentages over 28% due to the problem with vapor pressure allowing the zinc to vaporize and escape through the ventilation. Generally speaking, the maximum zinc content until the late 18th Century was 28% due to cementation limits of zinc to copper in a sealed crucible even with added charcoal.

From Craddock (2009). Frequency of zinc in copper alloys prior to 19th century.

This is because zinc boils at 1742ºF which is well below the temperature needed to melt Copper at 1984ºF. Zinc was not isolated as a metal until 1746, but was still used to create brass by smelting compounds that contained zinc with copper. In order to get the zinc to mix with the copper, the vapors had to be condensed into the copper. So the entire mix was heated to above the boiling point of zinc, but below the melting point of copper, the vapors then infused the copper, which was then melted to form a brass alloy. Only about 28% of zinc will permeate the copper through zinc vapor though. In 1832, however, G.F. Muntz patented a process of creating brass that was 60% copper and 40% zinc by smelting both metals together with a little added zinc to compensate for vaporization. He was able to do this since, by then, zinc was in production as a metal of its own, so he wasn’t using zinc-sulfides, -carbonates, or -silicates.

The 19th century was also one for an advance in copper ore refinement. Copper is often found in ores that include other elements such as the same sulfides, carbonates, and silicates as zinc, making it difficult to obtain consistent results when alloying with tin, zinc, lead, etc. During the late 19th century, advances in the understanding of magnetism and electricity gave rise to the process of electrowinning.

Electrowinning is a process by which copper ore is first placed in a solution and then chemically leached from the ore solution by use of electricity passed through the solution between a cathode and anode, thus winning ions of copper that collect on the anode. It was first done by Duke Maximilian Herzog von Leuchtenberg in 1847 then patented as a process in Whales in 1865 by James Elkington. By 1883, Balbach and Sons refining and Smelting Company was using the process in the United States. This process provides pure, refined copper, free of sulfides, carbonates, and silicates for alloying and is still in use today.

How is all this relevant to the Nova Scotia Hercules sword?
If you’ll remember, Brosseau made a couple key observations:

1) The copper had very little in the way of impurities like silicates, sulfides, or carbonates. In other words, it was refined. Pure. As mentioned above, the refining process of electrowinning wasn’t invented until 1847.

2) The zinc content of the sword was 35%. Brasses prior to around 1832 very rarely contained more than 28% and when they did it was never more than 33% for very small objects (like coins or tiny ornaments and jewelry).

This is why Brosseau concluded that the sword was a fairly modern production. It could not have been produced earlier than the mid-19th century because it is cast in brass with a zinc content of over 28% and the copper is refined.

Nor can this be a bronze sword as Pulitzer has claimed (he even wrote “high-zinc bronze” in his 200-page report. If there were such a thing as a high-zinc bronze, it would be called “brass.”) since it is not made of bronze (there’s no tin but there is zinc; bronze is copper-tin and brass is copper-zinc).

Pultizer has stated many times in many venues of social and even traditional media that he has portable X-ray fluorescence (pXRF) data which refutes Brosseau’s SEM-EDS data, but he never shows the data. It’s just not available for review directly, but we can take his word for it. So he says.

He was once asked in a Facebook group what methods he used to prepare the sample and what parts of the sword were sampled. It was pointed out to Pulitzer that the pXRF will sample everything in the viewing window and would essentially be limited to the surface considering there’s a patina. So, essentially, if the sample area isn’t first prepared by removing the patinated layer, he’d be sampling the patina and potentially not the underlying metal. This is something he scoffed at and attempted to provide a blast of “references” before simply deleting the entire message thread. One was left to think that the conversation bothered him.

Energy dispersive x-ray spectroscopy (EDS) was an excellent choice for investigating the Nova Scotia Hercules sword because it gives both the relative compositions (like a pie chart) and the distribution of elements within the sample–which is why Brosseau was able to see that the copper was refined. When using EDS, the sample is loaded into a chamber within the instrument so it needs to be fairly small, usually less than the size of a quarter. Brosseau removed filings from two polished sections of the sword, the tip and the hilt, and these were independently loaded into the instrument’s sampling chamber. Using the SEM’s image, the EDS can target specific regions of the sample and the result is extremely precise sampling.

Using the pXRF to sample an artifact is also extremely useful and has the added benefit of being non-destructive since you don’t need to put a sample in a chamber, thus no need to file off metal shavings which damages the artifact even if a little. But it has to be understood that everything in the handheld detector’s sampling window will be part of the results. Also, some sample preparation for an object like the Nova Scotia Hercules sword would certainly be necessary unless there was a bare section of metal, free of patina and corrosion.

The XRF will penetrate the surface of a sample, but only the first 100 ?m, so analyzing an unprepared sample would likely provide seriously skewed information of the patina, corrosion, weathering, plating, etc. in addition to the underlying material. This is known among experienced XRF analysts as the “matrix effect.” It’s more likely to adversely affect an unprepared sample than one that was prepared by removing the patina, corrosion, or outer layer so that an underlying section the size of the sample window is available for analysis.

On the other hand, and in the case of this sword, if you investigate the patina first, understand it’s composition, then find a section with bare metal, test it with the pXRF, then subtract the patinated elements from the primary target, what’s left is probably a good representation of the underlying metal. In this way, you could potentially scrape back just a small section of the patina. The results would hardly be as accurate as the SEM-EDS, which can target specific zones of metal at the microscopic level, but it should be sufficient to tell the difference between bronze and brass. Either significant quantities of zinc would be present (brass) or significant quantities of tin (bronze). The pXRF would not, however, give any clue as to the purity of the copper, which might indicate post 18th century electrowinning.

Given the data and facts, it seems clear that the Brosseau analysis adhered to the strictest of scientific protocols and was well reported. She and her team took two samples of the underlying metal and two of the patinated area. She not only used SEM-EDS, but also Confocal Raman Spectroscopy and Surface Enhanced Raman Spectroscopy (SERS) both to evaluate the patina. Her samples were specific, precise, and targeted as opposed to general and potentially random with the pXRF.

But even when you cast aside the scientific analyses, what’s always bothered me about this silly sword is this: if we assume it is a genuine Roman artifact of bronze dating over 2000 years ago (which it isn’t), then why would it be evidence of Roman visitation to the New World before Columbus? At best, all we’d be able to say is that a Roman sword was lost in the waters near Nova Scotia at some period between the Roman and the modern. Occam’s razor makes it far more likely that the deposition be closer to the modern period since we know that 1) ships traveled back and forth with regularity in the last few hundred years; and 2) the fascination with collecting Roman antiquities also goes back hundreds of years.

And the sword has limited provenance (its documented chain of custody) and zero provenience (the documented context of its original find). These two facts make it useless in either an archaeological or historical sense.


I returned to that field behind my friend’s house a few days later both humbled and cautious. The hornet’s nest appeared deserted then, torn in half by my backward blunder. At least a foot tall, it must have been the home of hundreds of angry-to-be-evicted hornets. I was lucky to receive only a half-dozen or so stings.

In the pretend swordplay surrounding the Nova Scotia Hercules sword, Pulitzer has been parrying and thrusting with those that doubt his claims, thinking that he’s fighting those archaeologists and historians before him. He seems completely unaware that the stings he’s getting are actually from the data and facts he so carelessly blunders through.

notes and references

Bayley, Justine; Crossley, David; and Ponting, Matthew (2008)”Part Three: Knowledge and Understanding” in The Historical Metallurgy Society Occasional Publication No. 6. London: The Historical Metallurgy Society Ltd., pp. 39-65.

Craddock, Paul (2009). Scientific Investigation of Copies, Fakes and Forgeries. Oxford: Butte

Geuzen, Peter  (2016). Swordgate poster link:

IMDB (2017). “The Curse of Oak Island.”

La Niece, Susan; and Craddock, Paul (eds) (1993). Metal Plating and Patination: Cultural, Technical and Historical Developments. Oxford: Butterworth-Heinemann Ltd.

Malainey, Mary E. (2011). A Consumer’s Guide to Archaeological Science: Analytical Techniques. New York: Springer.

Shackley, M. Steven () An Introduction to X-Ray Fluorescence (XRF) Analysis in Archaeology. In X-Ray Fluorescence Spectrometry (XRF) in Geoarchaeology, edited by M. S. Shackley, pp. 7–44. New York: Springer.

White, Andy (2016). Summary of the Analysis of the Nova Scotia Sword Performed by Christa Brosseau. Andy White Anthropology.

About Carl Feagans 321 Articles
Professional archaeologist that currently works for the United States Forest Service at the Land Between the Lakes Recreation Area in Kentucky and Tennessee. I'm also a 12-year veteran of the U.S. Army and spent another 10 years doing adventure programming with at-risk teens before earning my master's degree at the University of Texas at Arlington.

Be the first to comment

Leave a Reply