Mail from the ‘Garrison’ of Birka

A Review of Recent Research

Peter Beatson


Mail is the best attested form of body armour from early medieval Europe. Construction of mail is considered quite uniform, however it does vary in details which further research may eventually link to particular cultural origins or periods. The purpose of this article is to describe the properties of tenth century mail as known from some archaeological discoveries, in particular those at the major Viking age settlement/trading town of Birka [1] in Sweden.

The ‘Garrison’ at Birka

Organised excavations have occurred on Björkö for more than 130 years [2]. The site of Birka near the lake shore is marked by an occupation layer rich in organic material and artefacts called the ‘Black Earth’, which was the major focus of Hjalmar Stolpe’s early excavations in the nineteenth century. The town was defended by earthworks, including a hilltop fort - ‘the Borg’ which was more closely examined by Holger Arbman in the 1930’s. He was the first to determine that an area just outside the fort had a notable concentration of remains of military equipment. That area, now colloquially called the ‘Birka Garrison’, was the subject of intensive investigations by Lena Holmquist Olausson in the years 1998-2002. A barracks and workshop was uncovered, including a forge where mail was repaired but probably not manufactured on any significant scale.

Mail from the ‘Garrison’

About 100 finds of mail fragments have been recovered from the Garrison by the three excavations mentioned above, which are described by Fredrik Ehlton in his report [3]. They are mostly single rings but there are thirty-eight examples where two or more rings are interlinked, including some small patches of less than a dozen links each. Of these twenty-four finds are well preserved enough to reveal their construction, and all of these are made of alternating rows of riveted and solid rings, like almost every example of mail known from the early medieval period [4, 5].

Size of the Birka mail links vary considerably, the data are summarised in the graph below (Figure 1). Two peaks are apparent, at around 9mm and 11.5mm external diameter. It was noted that the larger type rings, though thicker also tended to be heavily corroded while the smaller, thinner ones were better preserved and had a smooth hard crust, which might indicate a different quality of iron was used, or result from a deliberate or accidental treatment (e.g. in the fire that destroyed the workshop).

Figure 1 - Size distribution of mail rings discovered at ‘Birka Garrison’ by different excavators. Data from Ehlton [6].

Thickness measurements of corroded wire are uncertain, however the results from twenty-four specimens which I judged to be sufficiently well preserved are shown in Figure 2 below, plotted with the external diameter of each ring. Data from two well-studied comparable finds, the aventail of the late 8th century Coppergate helmet (York, England) [7] and the fragmentary mail garment from the 10th century cremation grave at Gjermundbu (Buskerud, Norway) [8] have been included:

Figure 2 - Plot of wire thickness and external diameter of some mail ring finds discovered at ‘Birka Garrison’[9].
Average figures for the Coppergate [10] and Gjermundbu mail (from two studies [11]) are also shown.  Modern Brown and Sharpe standard wire gauges 12 to 18 are indicated at right for comparison.
Click here to see the data presented as wire thickness versus internal diameter of the same rings.

Naturally larger rings tend to be made of thicker wire - the typical ratio was about 6:1 though, as the plot makes obvious, there is a lot of variation [12].

Riveted links

To produce riveted links from wire the cut ends are overlapped and flattened, a hole is punched through the ‘lap’ and it is fastened together with a small rivet. All known examples of medieval mail including those from Birka have an anticlockwise lap - if held with the lap uppermost the right-hand end would be in front of the left [13]. At Birka the wire expands in width at the lap only to a slight to moderate degree [14], and there are examples of rings expanded mainly toward the outside, to the inside, and evenly (the most common). Such rivet heads as can be observed (Figure 3) are large, hemispherical and prominent [15]. They are smoothly rounded not facetted, showing they were set with a swaging tool, not hammered directly.

Figure 3 - Mail rings from ‘Birka Garrison’ [16]. Left - three riveted links showing moderate lap expansion and the variation in size of rivet heads. Right - a solid link of round cross-section, the thickened region may be a welded scarf joint. Finds (from left) G-01A8a (11.4mm); G-981173 (11.1mm); G-981085 (11.8mm); 993777 (9.0mm).

Solid links

In the literature two methods have been proposed for producing solid rings - by punching them from solid iron sheet [17], or producing them from wire in the same manner as the riveted ones and welding them closed [18]. The former method should produce a squarish cross-section like a washer, and the latter a round section. The great majority of the Birka solid rings are of round section (including three that are oval), while there are only three examples that definitely have ‘flat’ sections [19], and on this basis alone it might appear that most of them were created by welding (Figure 3, right). In practice however the distinction is not so simple, as it appears mail makers made some effort in smoothing corners off punched links [20] - while it would take just a simple hammer blow to the face of a round wire link to flatten it to resemble a square section.

Microscopic metallographic analysis is required to reliably distinguish the two constructions, but only a single sample of mail from the Birka finds was subjected to such testing - the results indicate its solid links were produced from drawn wire and welded, like the late 8th century mail aventail of the Coppergate helmet [21]. However a similar examination of the more contemporary mail find from Gjermundbu, Norway conclusively showed the punching method was also in use at the time [22].


Several characteristics should be considered when trying to authentically reproduce the appearance of mail armour based on the Birka Garrison. It should be constructed from alternating rows of riveted and solid rings, ideally both made from round section wire. Though a range of dimensions are possible, the most typical rings would be around 9 to 11mm diameter and made of 14 gauge wire. The laps would be anticlockwise and somewhat expanded to accommodate a rivet with a prominent hemispherical head.


[1] On Björkö in Lake Malaren, Uppland. Birka seems to have been founded in the second half of the eighth century and abandoned around 970AD, superseded by the nearby (and still extant) town of Sigtuna. H. Clarke & B. Ambrosiani, Towns in the Viking Age. Leicester University: London 1995 (revised), p75.

[2] E. Hyenstrand, ‘Early discoveries in the Black Earth’, in: B. Ambrosiani & H. Clarke (eds.) Birka Studies, vol. 1, 23-51. Birka Project (RAÄ-SHM): Stockholm 1992.

[3] F. Ehlton, Ringväv från Birkas garnison - dokumentation, preparering och analys. Report, 39 pages, Stockholm University 2002-03. Available (6th July 2008) at

[4] All-riveted mail was sometimes used in this period, and become standard in the late medieval era. Full riveting allows the use of smaller, relatively thick rings to produce a denser mesh, as the ring being riveted has to be linked with only two rather than four other rings during the riveting operation - S. O’Connor, ‘Technology and dating of the mail’, in: D. Tweddle, The Anglian Helmet from Coppergate, Archaeology of York 17/8, York Archaeological Trust 1992, p.1057-1081.

[5] Mail made from butted rings was not used in warfare, however the aventail of a toy helmet (Cologne, Germany 6th cent.) was made of butted links - O. Doppelfeld, ‘Das Fränkische Knabengrab unter dem Chor des Kölner Domes’. Schriftenreihe der Archäologischen Gesellschaft Köln, vol. 2. (after 1964), p24. Though the Sutton Hoo mail (early 7th cent.) is described as butted in earlier reports, subsequent analysis has shown that it is at least partly riveted - S. O’Connor, ‘Technology and dating..’, op cit p.1076.

[6] Ehlton, see footnote [3].

[7] S. O’Connor, ‘The mail curtain’, in: D. Tweddle, The Anglian Helmet from Coppergate, Archaeology of York 17/8, York Archaeological Trust 1992, p.999-1011.

[8] V. Vike, Ring weave - a metallographic analysis of ring mail material at the Oldsaksamlingen in Oslo (= Brynjevev - metallografisk analyse av brynjemateriale ved Oldsaksamlingen i Oslo). Translated report, 42 pages, Oslo University 2000. Available (6th July 2008) at

[9] Ehlton, see footnote [3].

[10] O’Connor, op cit p.1005 (number of rings measured =10).

[11] Riveted ring data only, solid links were significantly thicker - O’;Connor, op cit p.1185 (n=5); Vike, op cit p.18. (n=7).

[12] The ratio of ring O.D. to wire thickness varies between 4.7 to 1 and 8.5 to 1 in the sample chosen for Figure 2. On geometrical considerations, the ratio must always be greater than 4.8 to 1, so some of the thickest rings may not have been used for mail (or had to be mixed with thinner ones). PB.

[13] Vike, op cit p.32.

[14] Greatly expanded laps are more characteristic of late medieval mail, probably a technical refinement to increase its resistance to penetration. O’Connor, op cit p.1080-81.

[15] Whereas rivets of the Gjermundbu mail are small and less prominent, Vike, op cit p.18; and those of the Coppergate aventail are flush with the wire, O’Connor, op cit p.1004.

[16] Images copyright, Fredrik Ehlton and Stockholms Universitet. Permission for reproduction granted. PB.

[17] For a good description see Vike, op cit p.34.

[18] For a good description see O’Connor, op cit p.1064-68.

[19] Flat link mail is another story - it was apparently used in the contemporary Khazar khaganate. There is one example of a flat riveted link (Stolpe59649) from Birka, possibly a second (G-004474) but unfortunately its join is missing. Examples of flat link mail known from Europe (and Russia) seem somewhat later, Vike p.31; A.N. Kirpitchnikov, Drenvnerusskoe Oruzhie III: Series Arkheologiia SSSR, Leningrad 1971. PB.

[20] Vike, op cit p.35.

[21] O’Connor, op cit p.1024.

[22] Vike, op cit p.14. Note these solid links are clearly of square cross section.

‘Mail from the Garrison of Birka’, written and webbed by Peter Beatson.

(c) Birka Traders 2008. Not to be copied without permission.

All rights to materials in any linked pages belong to their respective owners.