Sunday, November 15, 2026

Late Roman Craticula (Gridiron)



Finished Piece


This piece is a reproduction of a cooking gridiron (craticula) excavated in Icklingham, Suffolk,

England, and currently in the collection of the British Museum1

.

While a precise date is not known, the museum estimates it to be 4th century, shortly before or

concurrent with the withdrawal of Roman control over the area. The piece shares clear elements with

grid irons or stoves found across the Roman Empire, and dating as early as the destruction of Pompeii.

The unique feature of this particular piece are the “omega” shaped central bars.


© The Trustees of the British Museum (CC BY-NC-SA 4.0)

Period Techniques:

A blacksmith in Roman England would have been working with wrought iron. The only tools required

for this piece were a hammer, tongs, a forge, a bellows, and an anvil. The fuel would have been coal or

charcoal, as both were available in Roman England.2 The piece is largely forge welded from pre-made

bars, with bent and tapered elements. There is a central band that is consistent with bent over staple

banding that is common in assembled ironwork. It appears to be relatively solidly attached to the

piece, but I believe that is the result of use and aging, as heating, even just to cooking temperatures can

create welds over time, and rusting can create a conglomeration affect among disparate iron pieces.

Finally there is a welded ring with a clinch nail type clip attaching it to the piece.

Figure 2: "Funeray Stele of a Blacksmith" 1st century CE CC BY-NC-SA 4.0 Carole Raddato


Modern Compromises:

Unfortunately, wrought iron is no longer widely available, so I used mild steel instead. The silica slag

in wrought iron makes it largely self fluxing and relatively easy to weld. In addition, a period

blacksmith may have used silica (sand) as an additional flux, but there is no way to tell for a given

piece. Since mild is more difficult to weld, I used borax as a flux, offsetting some of the difficulty. I

also used a hand crank blower in place of bellows for convenience, and a modern (London pattern)

anvil. Otherwise the tools I used are largely consistent with those found in images of Roman forge

work. I also used industrially produced steel bar. This was purely a labor saving decision, as the

construction of this piece clearly started with similar bars, even if the smith had to produce them

personally.


My Tools


Methods:

The central bars were bent into a matched pair and tapered. Each bar was then welded to the third and

fourth bars. A square hole was punched and drifted into the fifth bar to provide a mounting point for

the hanging ring, but the ring was not attached at this time. These bars were welded one at a time to the

existing assembly, then this was repeated with the last two bars. The feet were then hammered and

bent up, and then the legs bent down to achieve the final shape.

After the core was complete, the ring was welded separately, and the clip was formed as a flat piece,

prior to being folded around the ring. The ring and clip were inserted, heated, bent around, and cut

with a chisel to fit the piece. Finally, a central band was made, hammered into a square staple, and bent

over in a traditional blacksmithing technique.




Work in Progress


Difficulties and Learning Experiences:

I significantly improved my welding techniques, particularly an approach of closing the weld with

tongs while the piece was still in the fire. This potentially allows for more precision in placement, and

multiple attempts without significantly deforming the metal. While I am happy with the end result, I

did end up with the final piece slightly larger than the original (16” long dimension versus 12”). In

addition, my “omega” bends are slightly smaller relative to the finished piece, and the two side bars are

not parallel. Finally, the pre-punched ring mounting hole ended up off center during the final assembly.

I still believe this is the correct approach, as trying to punch that hole on the finished piece would be

extremely difficult, but in the future I would take more care in making sure it’s centered during

assembly.


Other Thoughts:

This piece was made with bars of uniform dimensions, and based on my work, I believe equal starting

lengths. This makes me think it is likely that the smith doing the work was purchasing bar stock from

another source, as that level of uniformity makes more sense if it’s specified to ensure a standardized

purchase. It also results in a piece that takes more metal than a more standard style using smaller

central bars (see figure 4).

While the omega shaped bars are decorative, it seems likely that they were also intended to serve as

receptacles for conical or round bottom pots common in Roman kitchens. A ring that appears to be for

this purpose is found in figure 5, and similar rings are found in a stove from Pompeii, and heated food

counters from the same city.

That being said, the choice to band the central bars rather then weld them is surprising because it

interferes with both placing round bottom pots into the omega shapes, and with putting a flat pot on the

gridiron. I am inclined to think, admittedly without evidence, that it might be a later addition to the

piece.


Similar Extant Pieces:

Excavated at Trimontium (Newstead) England


This piece represents the most common design found in multiple locations, and over a time period extending beyond the (western) Roman Empire, with tenoned cross bars.


Excavated at Castra Vetoniana, Pfunz Germany


Uses the same sort of forge welding assembly found in the recreated piece. This piece was identified

as a window grid by the original researchers, but I believe it is a gridiron.


Excavated at Calleva Atrebatum, Silchester, England

This piece has a center ring reminiscent of the omegas found in the recreated piece, and

similar hanging rings. However the pieces are tenoned like the common simple grid.


These pieces were all found along the Roman frontiers, frequently in military posts, making it possible

these were Legion equipment.


Most references were pulled directly from a curators note on the item in question, with additional web

searches on key terms. In order to track down reference 4, which had an incorrect name in the curators

note, I reached out to the Trimontium Museum, who pointed me to an online copy of the work. The

error has now been corrected on the British Museum website.