Roman Mail Construction

[Kineas]

Salwe,
I'm thinking of putting together a Roman Auxiliary Cavalry impression (let's call it 100 AD). I'm looking at chain mail. I'm confused.

More than one reenacting-related website has told me that I want round wire with round rivets and a washer in a 4-1 pattern. None of them offered me any documentation, but it does seem a popular view.

On the other hand, David Sim in his article "Roman Chain-Mail: Experiments to Reproduce the Techniques of Manufacture" states

To gain accurate dimensions for the reproduction of chain-mail, ancient rings were measured to obtain the size and shape of the wire and the diameter, and the method of assembly of the rings was examined. Ferrous chain-mail rings from Caerleon,l3l Wales,and also some from Thorsberg Germany, were examined.The rings were made from wire which was square or rectangular in section, rather than circular.

His article is quite well researched and he includes examples of period rings. His photo of Thorsberg rings seems to show flat iron wire that has rounded edges from touching up on a mandrel.

Before I spend money, I'd like some guidance. It looks to me as if the popular view is incorrect, but experience tells me that there's often more to research than one article, however good.

[jvrjenivs]

There is written about this on RAT some times before. The best advice you can get is by Erik D. Schmidt.

Then my own two cents, but I'm no expert in this, but doing a auxilia cavalry impression myself, I've looking into this a bit, so for auxilia cavalry I would go with a 4 in 1 shirt with alternated solid and riveted rings, with rounded doubler. See Adrian's marketplace on here for this item. A normal riveted/solid hamata will also do. The 6mm shirts are good and have a nice price.

[L_CORNELIUS_CETHEGUS]

Flat vs Round cross-section: One thing worth considering is that when creating the overlap, it can be difficult limiting the flattening force to small area where the ends overlap. If you place the overlapped rings into a die (while still round in cross section) then hammer the flat punch onto them, there will be can some small degree of flattening over the non-overlapped portion as the overlapped portion gets smashed together. This could explain the flat top/bottom yet slightly rounded outside/inside of the wire after the overlapping is done. You can achieve a nice profile that I think would still fall into the 'round cross section' if you don't flatten it any more at that point. Keep flattening of course and you'll be going toward more late medieval mail which is completely flattened.

As for the solid links in roman mail, I should expect these would be flat in cross-section, since it's probable that they were punched from sheet metal, rather than formed from wire. You might see online that the slag patterns in the iron differ between the solid and riveted rings, lending some physical evidence for this.

--Kelsey

[Matthew Amt]

I know I've seen an article in JRMES (probably, though it might have been an old ARMA issue) which very deliberately showed cross-sections of the rings, and the riveted ones were round wire. Sorry, it was years ago, and I don't have a citation on hand. But Erik Schmid wasn't the only person examining surviving bits of mail and finding round-section wire rings. I also remember a photo he sent me of a very well-preserved piece, and the punched rings still had visible burrs around the edge from the punching! Very cool. Darned if I can find that photo NOW, of course...

Matthew

[Matt Lukes]

Oh there's no doubt the solid rings that were punched would have a square cross-section and even those purported to be welded would at least near the joint (since hammering is necessary for fire welding), although it still seems dubious this was a usual practice given the technical difficulties involved. The riveted rings are another matter- I'd be wary Christian of even expecting there's a right and wrong; since neither is technically impossible for any given period, it may be a matter of prevalence more than anything- thus neither is 'wrong'. Indeed Erik is the man to talk to- mail is his expertise, mail of all periods, so I'm sure he could advise you with more certainty than probably anyone. I have images of what's supposed to be Roman mail that show round wire riveted rings and flat punched ones, but they're collected from the Internet, so don't have site information often enough, etc. I can say the mail of the various examples of Lorica hamata squamataque (L. plumata, some call it) do have round riveted rings and flat punched rings so I see no reason to expect regular mail was any different on the whole.

[Matt Lukes]

These are from Kalkriese so dated pretty securely to 9CE- they look mostly rounded although it's hard to say for sure if there are slightly flattened bits there; they're cetainly not square in section though...

Curle describes a bit of iron mail from Newstead as alternating riveted wire rings and solid, welded ones but it's hard to say just how accurate the description is since flattened or not may not have seemed a vital detail to mention.

[Kineas]

Thanks to everyone. All aspects of recreation are always awesomely complicated. I thin it is one of the reasons I enjoy it so much.

My own research offline coincides with what I think I'm hearing here. I'll summarize, and you folks tell me if I'm wrong.

There is evidence--strong evidence--for round wire riveted links with round-head rivets.

There is evidence--strong evidence--for Roman use of washers, whether punched or otherwise.

There is other evidence--both pictographic and archaeological--for flat stock or square wire rings riveted with round head rivets, BUT by the time such rings were shaped on a mandril and/or finished with an interior punch, they often had merely a "flattish" roundness to them.

Conclusion--if purchasing a mail shirt for a 2nd C Cav impression, it should be round wire, riveted with round head rivets, with washers, 4/1. Even the "square stock" wire would, buy the time it was serving in the field, have a "rounded" edge or look to it. Ideally, I would punch 120K rings by hand and finish them on mandrils. that's not happening this lifetime.

Is that a fair summary?

[L_CORNELIUS_CETHEGUS]

Good on your assumptions I think. My own assumption is that the default wire working material may have been round, although I may be projecting based on the materials I/we have available, which are mostly round. (Somebody more knowledgeable on wire pulling technology/tools could probably give a better answer).

I think the (slight) flattening of the riveted rings happens mostly as a byproduct of the overlapping process.

I think round rivet heads are generally assumed to be the norm for the Roman period - certainly not wedge as seen in the 15th century. Technically, round makes a lot of sense since (I assume) the round wire for the rivet is readily available, punching/piercing the overlapped rings is more straightforward (i.e. no need to orient a piercing drift as done with wedges), and the clinching tools necessary to close the rings would probably have a dome-shaped indent where the peening takes place.

BTW, here is an early attempt at some rings (pre-punching) that I did some time ago. My technique actually got a lot better after what you see here but maybe it gives a better idea of the round/flat part I was talking about before.

[attachment=0:fkgyr317]<!-- ia0 DSC_0049.jpg<!-- ia0 [/attachment:fkgyr317]

Of course, I realize that you're not looking into building your own hamata from scratch (not a bad decision and after all the work and research I've put into it, I would probably buy what you're describing. I'd probably save my money and go with 6mm since it's simply amazing to see, but I'm sure the 8mm would be awesome in the riveted/solid 4-1 setup you're describing.

Hope to see a photo someday

--Kelsey

[L_CORNELIUS_CETHEGUS]

...to be welded would at least near the joint (since hammering is necessary for fire welding), although it still seems dubious this was a usual practice given the technical difficulties involved."

I completely agree about welding and I'm surprised when I hear that theory brought up out there...

I think forge welding requires a couple things which are terribly difficult when you are working with fine wire: one can get sufficiently high temperature (I think white hot is necessary for welding) but keeping it from melting the iron on such a tiny piece of metal would be just one challenge. The second challenge that immediately comes to mind is keeping that ring hot enough as you take it from the fire to the tool/anvil where it's being hit. Items with a certain mass (a sword or an iron bar for example) will stay hot enough for welding just until they touch that cold anvil, so even with those you have to be on-the-ball and go to work quickly before the heat gets too low for welding. A 6mm ring would lose heat extremely quickly. Finally, contaminants that ruin the weld would be another obstacle. In order to work with wire (today at least) we anneal the metal first. I'm not sure since I've never done forge welding, but I wouldn't be very confident about forge weld something that has already accumulated some scale from the previous annealing. In case you're imagining the logistics of getting around these technical challenges, I would also add that you would be not welding a single ring, but adding a ring to some pounds of a pre-existing matrix of rings that have already been assembled. I can't even figure out how one would try to add a white hot ring to a shirt under construction and weld it shut :!: when it's hard enough to do that with the metal cold.

--Kelsey

[L_CORNELIUS_CETHEGUS]

Darned if I can find that photo NOW, of course...

Matthew - The image I always think of on the subject of solid/riveted slag patterns is Fig. 3 in this publication by Arne Jouttijärvi, posted on Erik's website: [url:1stq4bz3]http://www.themailresearchsociety.erikds.com/pdf_files/tmrs_pdf_10.pdf[/url].

--Kelsey

[Kineas]

Kelsey, if you look at the Greek reeanctment forum you'll see that I just finished a 1300 scale shirt where I made all the scales one at a time. So I did THINK about doing it myself...

Then I thought again. LOL.

[L_CORNELIUS_CETHEGUS]

That's an absolutely beautiful scale thorax and I get a sense of the time involved - not only producing it but the research as well.

I dropped out of riveted mail after assembling about four square inches of it

Seeing your work perks my interest up to tackle a lorica squamata...

--Kelsey

[John Conyard]

Over breakfast I was reading James S. (2004) Excavations at Dura-Europos 1928-1937 Final report VII The Arms and Armour and Other Military Equipment (London 2004). A page 110-111 contains a summary on mail shirts.

I would have said it was relatively safe to go for flat square section punch rings, alternating with riveted rings of round cross section with round rivet heads.

Putting together a 25,000 ring shirt would be a mad project, even over a Canadian winter.

A few years ago for the business I bought two of the 6mm shirts from DSC via Armamentaria, one with shoulder doubling, one without. I’m very pleased with them, and they have flat square section punch rings, alternating with riveted rings of round cross section with round rivet heads. They have survived two years of hard wear unscathed and I haven’t lost a rivet.

For 100 AD and after I would opt for a simple shirt without doubling, http://www.armamentaria.com/store/index ... cts_id=118

Mine is fine over padding. But I have a bigger build than your self. Such a shirt is very versatile and can be used for many different impressions.

A shirt with doubling would date the shirt, and the “s” hooks on the front would date it even more.

[PhilusEstilius]

John.
In the case of a Ist century cavalry outfit might it not be alright to have a shoulder cap which of course is more appropriate, but then there is also the point just where and when did this style of mail go out we wonder.

[John Conyard]

Hello Brian,

I don't want to repeat what has been said already. <!-- l <a class="postlink-local" href="http://www.romanarmytalk.com/rat/viewtopic.php?f=20&t=27939">viewtopic.php?f=20&t=27939<!-- l

As you say for the 1st century by all means use a shoulder doubling.

For after 100 AD mail without shoulder doubling would be fine, and useful for many other impressons.

Cheers

John