My greatest love is the late Migration Era, but for this project I wanted to create a full set of soft kit that would be appropriate for a “Roman Iron Age” presentation from Scandinavia in the late 3rd or 4th century CE. The goal was to put together a set of everyday clothing that was generic enough to be plausible for a variety of locations (Norway, Sweden, Denmark) and did not represent a high-status individual, but could be supplemented with weapons or accessories to raise the perceived status if necessary.
For this project, I wanted to make sure I created an ensemble rooted in evidence but also wanted to avoid creating a pastiche of very distinctive elements that would likely never have been combined in a single outfit. Each garment is based on an extant piece from Norway or the Jutland Peninsula that is carbon-dated to within my target timeframe, but each of those garments, in turn, has at least one other extant parallel from somewhere at least a day’s journey away, so I can be reasonably confident that I am not recreating a unique or distinctive item. Although not perfect replicas of the specific garments, my pieces use materials and construction techniques that are consistent with textile finds from this period. Those of us who reenact the Migration Era and early Medieval period often rely on these much earlier finds to bridge the gaps in the archaeological record, but for this project, I was able to use them in their own timeframe, which was a nice change of pace.
Here is the finished ensemble. It consists of a tunic, trousers, legwraps, shoes, and a belt. The sections below summarize the basis for each element, explanations for material choices and construction methods, and things I learned in the process of making the garments. There is an image of a disembodied foot of a bog body in the “legwraps” section, so be forewarned if that is going to bother you.
Tunic
The tunic is modeled on the Lendbreen tunic, a garment recovered from a Norwegian glacier and C14 dated between 230 and 390 CE. It is generally similar to the contemporary Thorsberg tunic, consisting of a fairly fitted rectangular body, narrow sleeves, and a boat neckline, with sleeve seems positioned at the back of the arm rather than along the underside of the arm. This article provides a detailed description of the Lendbreen find, and this one discusses the process of recreating it, starting with the raw fleece. Between the two articles, I had access to extremely specific information including the exact dimensions, seam treatments and stitch types for each of the various seams, and clear enough photographs to see the fabric grain and understand the cutting pattern.
The original tunic body was woven from a diamond twill that used two different colors of natural wool – a dark and a light. The light colored wool was used for the weft, and the warp uses alternating light and dark threads. While I was not able to source an exact replica of the original fabric, I was able to find a wool twill that used natural light weft and and warp, and every fourth warp thread is a dark brown. This produces a fairly similar visual effect and felt like an acceptable compromise. The original garment had sleeves made from a different diamond twill in a dark brown, but since I was not attempting to create an exact replica of the garment, I made the complete tunic from the same fabric, which I overdyed with a light brown dye made from walnut hulls. All seams were sewn with a two-ply wool thread. The original sleeves are torn and their full length is unknown; I used the Thorsberg tunic as the basis for making mine full length and for leaving the last few inches of the sleeve seam unsewn to create a slit at the cuff.
Detail of fabric at left (after dyeing); at right, in-process image of armscye finishing, using the same 1cm double-rolled hem found on the original.
In other respects, the cut and construction follow the original as faithfully as I was able to, using the measurements and seam/stitch descriptions provided in the two articles linked above. The tunic body consists of a single piece of fabric folded into a tube and sewn up most of the way on the long side using the Thorsberg stitch with the side seam on the wearer’s left. To create a slight flare, the maker cut one edge of the tube on a slight diagonal; this adds a few inches of extra circumference at the bottom hem and twists the side seam slightly, but is otherwise irrelevant to the construction process. The right armhole is cut into the tube at the opposite shoulder, but there is no seam continuing down the right side. The bottom of the tube is left open, while the top edge is sewn at each shoulder, leaving the middle open in a boatneck that is raised 4cm at the back (the height difference is accomplished by gathering extra fabric at the neck back – see following paragraph – and by removing a narrow strip of material from the top edge of the front panel). My tunic is constructed in the same way, using the dimensions from the original, which fits me perfectly (it seems as though I am historically sized!). Along the way, I encountered a few interesting elements of this construction that were not explicitly discussed in either of the two articles linked above – the way the neckline shaping in turn affects the shoulder construction and armholes and the method for attaching the set-in sleeves.
In the published photograph, it is clear that the neck back is wider than the neck front. To accomplish this, the maker could not simply lay the tube flat and sew up the shoulder seams working inward from each corner. Some extra fabric must be gathered by bringing the corners of the back side several centimeters closer the center of the body before beginning to stitch the shoulder seams. This must be done prior to cutting the armscye and since it creates a garment where the back is slightly wider than the front, it places the armholes slightly closer to the front of the body when worn. This aligns well with natural human range of motion (since the arms are mostly waving around at the side or front of the body rather than reaching behind the back).
The raised neck-back also creates a slight slope at the shoulders that accommodates the trapezius muscles and allows for a narrower neck opening than if a simple slit was left that split the head circumference equally between front and back. I was able to fit my 57cm head through a neck opening that is only 23cm wide across the front instead of 28.5 cm. Overall, the tunic construction is intuitive and follows the human form better than a strictly rectilinear “pillowcase” pattern.
The sleeve construction was also interesting. A set-in sleeve requires more sophisticated shaping of the sleeve cap, and while the photographs of the original showed that the sleeves may have had some slight puckering, they do not appear crudely attached. A photograph of the back of the garment shows the sleeve seam at the upper rear portion of the sleeve – not aligned with the neck opening or the shoulder seam – and this made it much simpler to do without having to draft the typical bell-shaped set-in sleeve cap ahead of time.
To create the sleeve, I started with a rectangle large enough to form the rough tube of the sleeve. Beginning at the back of the shoulder where I wanted the sleeve seam to begin, I sewed the edge of the tube to the armscye of the tunic, working my way up and over the shoulder, and as I approached the armpit, I increased the allowance on the sleeve side of the seam by about two fingers’ width, gradually tapering it back as I exited the armpit on the back of the armscye. This successfully avoids bunching in the underarm area and is not especially scientific. When I reached my starting point, I snipped off any excess sleeve material beyond what I would need for a seam allowance and sewed the sleeve seam down toward the wrist.
I did a test sleeve in scrap linen just to see how it would work. When I removed the test sleeve and flattened it out, the result was a slight S-curve along the top edge of the sleeve – less exaggerated but very similar to the edge of sleeves on much later medieval garments that use a back seam (although these are more fitted and have a triangular insert as well). Since the armpit portion of both the original Lendbreen sleeves are missing, I cannot say for sure whether this was the technique used, but it is an easy and intuitive way to attach a set in sleeve without the need for any real planning or drafting and is almost identical to the process for attaching a sleeve to a similar tunic without armhole shaping.
In the images of the finished garment above, you can see the nice anatomical fit of the neck back, as well as the position of the sleeve seam relative to the elbow and the cuff opening at the outside of the wrist.
Trousers
The trousers are based closely on the pair found in a bog in Damendorf, Germany, located on the Jutland Peninsula about 50 km south of the present Danish border. The trousers were found along with a leather belt, shoes, and a pair of legwraps, and are dated between 135 and 335 CE. The originals are a wool diamond twill and the garment pieces are well preserved, although the thread has deteriorated completely, thus it may have been linen or some other vegetable fiber. I sewed mine using linen thread on a 2/2 wool twill in a slightly mottled gray color.
I had made other pairs of trousers based on the Thorsberg and Damendorf examples before, and for this project I wanted to see if I could get a well-fitting pair of pants using the Damendorf seat pattern without using a tape measure, paper pattern, math, or any other modern convenience. I describe the whole process for this in a separate blog post here, so I will spare the details here, but these were essentially draped on my body, using my handspan to check some proportions, and used almost exactly one yard of fabric with very little waste. They’re sized so that in a full crouch they are skin tight, are form fitting while seated, and have just a touch of ease when standing. I found that when I size them to hit my natural waist and then add a very wide waistband like the originals, I can fold the waistband over and they actually stay up on their own. With a simple cloth belt or tie, they are definitely not going anywhere and it adds almost no bulk under a tunic.
The original trousers were rather ratty at the bottom, so we don’t know whether or not they had attached feet like the Thorsberg trousers. Based on published scale drawings of the Damendorf trousers in Karl Schlabow’s Textilfunde der Eisenzeit in Norddeutschland, the outseam (exclusive of the attached waistband) is about 100cm and the inseam is a little over 7o cm. The waist diameter (assuming a narrow strip for the crotch gusset) is about 100 cm. These dimensions suggest that if the trousers originally ended just below the damaged area, they would have fit an adult who is roughly my size. I am about 175 cm tall and weigh around 63 kg (5′ 9″ and 140 lbs), and my trousers ended up very close to the original dimensions as scaled from the drawings. I would say that they would likely have fit an individual who (like me) would wear a 30 waist, 32 inseam in modern (US) trousers, although if the wearer were slightly shorter, the ragged flaps at the ankle may have been part of an integral foot as well. I opted to reconstruct mine as ankle-length trousers without feet.
Legwraps
The legwraps are based on a pair found in Søgård Mose II at Viborg, Denmark, dated between 130 and 320 CE, now on display at the Skive Museum. Unlike the pair found with the Damendorf trousers (which are longer, narrower strips), these consist of two lengths of cloth sufficient to go around the entire lower leg with plenty of overlap. They were found still wrapped around the lower legs of a corpse, leaving no ambiguity as to their purpose or exactly how they were worn. Based on the photographs, the cloth appears to be a 2/2 wool twill that was woven to width so that the selvedge edges form the long sides of the rectangular pieces while the short ends have a rolled hem. Cords are attached to one short end and are used to secure the wraps at the ankle and just below the knee.
While I could not obtain wool at the exact width necessary, I used a 2/2 twill cut so that one long edge used the selvedge and the other was secured with a rolled hem like the short ends. Although I realize that the color of the originals may be the result of tannins in the bog rather than an intentional color, I wanted to experiment with natural dyes, so I dyed the fabric using onion skins. I attached braided cords to the edges to use as ties; I’m not certain what the original cords were made from, but since they survived in the bog, it is likely that they were some sort of animal (protein-based) fiber, so I have used wool for mine.
Shoes
Intact shoes from any part of Scandinavia during this time period are as scarce as hens’ teeth. A Norwegian glacier near Oppland (Runningsbreen) has recently yielded a reconstructable shoe dated to the 4th century CE. At the time I originally put this together, information on that shoe (discovered by hikers in 2019) had not yet become available. A conservator tweeted images of the find in April of 2022, so I updated this outfit with a pair of shoes based on that find. The shoe is similar to examples from the same period found in southern Jutland/Schleswig-Holstein and slightly further west in the Netherlands, shown below (a one-piece leather shoe with slits on the sides).
These include the Damendorf shoes (see Hald 1972) from the Jutland Peninsula, carbon dated to between 135 and 335 CE (van der Plicht et al. 2004), another shoe from Wijster, a settlement in the Netherlands beyond the Roman frontier that was active between 150 and 425 CE (van Es 1967), and a fourth example from a 4th-century frontier settlement at Cuijk, Netherlands (van Driel-Murray 2007). All four shoes are slightly different but rely on the same basic principle: starting with a roughly oval shape, rows of staggered slits are cut into the instep. When soaked in water, this area expands to form a net-like panel that conforms to the wearer’s foot, although the extent of the net varies a bit across the examples noted, as does the size of the slits themselves.
Wijster shoe at left (image from van Es, Plate 17); Damendorf shoes at right.
Although I was not intending to make a set of clothes specific to 4th century Norway, I could not resist the temptation to pair the Lendbreen tunic with a contemporary pair of shoes found less than 10km away, so I decided to base mine on the Runningsbreen shoes. I derived my pattern from scale drawings and photographs prepared by the conservation team at the Kulturhistorisk Museum in Oslo, Norway. The toe box includes four small tabs that are still gathered together with a fragment of leather thong, and each side has four large slits that stretch to form lacing loops while four long rows of smaller slits run across at a slight diagonal. Although the additional rows of slits on the Runningsbreen shoes may at first appear to be more decorative than structural as compared to the other three examples, I found that they dramatically increased the elasticity of the leather as it stretches to form to the foot.
Unlike Roman shoes of this period, examples from the Germanic sphere were typically fastened with a single thong laced in a spiral starting at the gathered toe. The Damendorf shoes were tied with a single integral lace that was cut as part of the shoe pattern and runs through the toe tip and then zig-zags up the instep to tie at the ankle. The Wijster shoe does not appear to have this feature, nor does the Cuijk shoe (Volken 2016), although the toe is too damaged to be certain. On the other hand, the Runningsbreen shoes are unusual in that they are a more symmetrical cutting pattern with the expansion slits on both the lateral and medial instep, whereas the other three examples only have slits on the lateral side, a more typical configuration in the Germanic-language sphere at the time, where asymmetrical cutting patterns were the norm (Volken 2016).
Interestingly, the Damendorf pair lacks the long tab across the instep, but the Wijster and Cuijk examples (within or much closer to Roman-occupied territory) share this detail with the Runningsbreen shoe. This may indicate a higher degree of Roman influence, further emphasized by the symmetrical cutting pattern of the Norwegian example, and made me wonder if perhaps I should use a dual-ended lacing pattern comparable to Roman one-piece shoes, but the shoe toe has an odd number of slits, which means I would have to fudge things a little bit. Lacing in a spiral eliminates this problem, but the conserved shoe looks as if the lacing points are directly opposite one another rather than staggered, as would be expected with a spiral lacing. When I laced the shoes, however, I found that there was very little stagger, so it really could go either way (see below).
According to conservators, the original was made of untanned hide (possibly oil- or smoke- cured, but currently unknown) 2-3 mm thick, although other examples including the Cuijk shoes are of vegetable-tanned leather (van Driel-Murray 2007). I used a 2.5mm thick vegetable tanned leather, which stretches well when soaked, and formed the shoes over my own feet. They are sewn with waxed linen thread using a butted whipstitch at the heel, as shown in photographs and drawings of the original shoe; I could not obtain information as to the original stitch material (which could have been sinew, thin thong, or vegetable fiber) but the whip stitch is still visible at the upper end of the heel seam. I also tried to emulate the larger stitch pitch visible in the photographs of the shoe, despite my preference for smaller and tidier stitches. In retrospect, the linen thread stretched out a bit more than I would have liked, and another option might be to use sinew or very thin rawhide thong that will shrink as it dries and tighten the seam.
I have omitted the short bit of thong inserted into slits partway up the heel and knotted at either end. It is not integral to the shoe closure, and may have had some other function, perhaps as part of a ski binding system. Similarly, I did not include the thong fragment beneath the ball of the foot, which could have anchored a repair patch or been part of a binding. The strong similarities to the other shoe examples of this period leave me with no doubt that it did originally function as a shoe, but it is possible that when the soles wore through and were no longer useful as footwear, the remaining portion was repurposed as a ski, snowshoe, or crampon binding that could essentially be worn as an overshoe.
Belt
The only missing element at this point was the belt. The Damendorf Man’s body was found with a leather belt, but it has been 122 years since he was discovered and while the belt is often mentioned in passing, I have yet to find a clear photograph of it or any additional detail. His garments were deposited in the bog with him, but not on him, and apparently the shoes, trousers, leg wraps, and belt were bundled together, so it is unclear whether the belt held up the trousers or cinched a tunic (perhaps a linen, hemp, or nettle fiber garment that did not survive in the bog). I struggled to find a simple garment belt buckle that would be appropriate to the time and place; everything I was running across came from more ornate belt sets in weapon graves or from ritual deposition contexts. I have tried to be careful and evidence-based in the rest of this project, but for the belt I decided to push the boat out just a little bit, since this will be the least visible element of the ensemble.
If we look at Medieval examples beginning the late 8th or 9th centuries, we find bone buckles that closely resemble their metal counterparts in size, shape, and decoration; some are quite fine, others are crude imitations. I located two examples of bronze buckles dated to the Roman Iron Age – one is from Hafslo, Norway, not too far from Lendbreen, and the other is from Vallstena, in central Gotland. In Gotlandic contexts, Almgren and Nerman assign this style to the second half of the 4th century (see p.135). Compared to the Gotlandic buckle, the Norwegian one is a fairly crude design, but the stylistic details are the same: single-piece construction, a trapezoidal frame, an integral plate with three rivet points and a notched “foot” at the short end, and a few incised lines as decoration.
Buckle from Gotland at left (Almgren & Nerman), buckle from Norway at right (Universitetsmuseet, Bergen, Item no. B11340; image from Unimus database).
I carved this one from cow bone, keeping to the general shape and proportions of the exemplars. The Norwegian one is about 36mm long and 33mm wide, and I scaled this one down a touch to fit the size of the bone that was on hand – it is about 32 mm long and 26mm wide, and the frame needed to be much chunkier than the bronze one in order for it to be as sturdy. I formed the tongue from some bronze wire and riveted it to a strip of dark brown veg-tan leather that I found while rummaging in a box. I think it was part of a vinegaroon dye test at one point.
I’m not entirely satisfied with this solution, but it will be mostly hidden by my tunic and is no more speculative than using a textile belt or a leather one with no buckle at all. Of course, it is certainly possible that no belt was worn aside from whatever was used to hold up the trousers; however, this creates a number of small inconveniences for the wearer that makes me think this is unlikely. It would require that utility knives or other small, needful items be suspended from a trouser belt up underneath the tunic, and it also makes it impossible to adjust the tunic hem height by blousing. In addition, the Lendbreen tunic also showed signs of wear around the hip area, which is at least suggestive of a belt worn over the tunic.
Sources
Almgren, O., Nerman, B., Kungl. Vitterhets. 1923. Die ältere Eisenzeit Gotlands: Nach den in Statens Historiska Museum, Stockholm, aufbewahrten Funden und Ausgrabungsberichten. Stockholm: Druck von I. Haeggströms boktryckeri a.b.
de Rigo, Daniele & Enescu, Cristian & Durrant, Tracy & Tinner, Willy & Caudullo, Giovanni. (2016). Juglans regia in Europe: distribution, habitat, usage and threats. European Atlas of Forest Tree Species.
Es, W.A. van, 1967. Wijster. A native village beyond the Imperial frontier 150-425 A. D. (Palaeohistoria XI) Groningen.
Hald, M. 1972. Primitive shoes: An archaeological-ethnological study based upon shoe finds from the Jutland peninsula. Copenhagen: The National museum of Denmark.
Kwaspen, Anne & Zink, Gabriele. (2015). “The Datgen Trousers.” in Karina Grömer and Frances Pritchard (eds.), Aspects of the Design, Production and Use of Textiles and Clothing from the Bronze Age to the Early Modern Era, NESAT XII. 175-186.
Schlabow, K. 1976 Textilfunde der Eisenzeit in Norddeutschland, Göttinger Schriften zur Vor- und Frühgeschichte, Band 15, Neumünster, Karl Wachholtz Verlag.
van der Plicht, J., van der Sanden, W. A. B., Aerts, A. T., & Streurman, H. J. (2004). Dating bog bodies by means of 14C-AMS. Journal of Archaeological Science, 31(4), 471–491. https://doi.org/10.1016/j.jas.2003.09.012
van Driel-Murray, Carol. (2007). Mode in de nadagen van het Keizerrrijk: de schoenen van Cuijk. Westerheem 3, 133-141.
van Es, W.A. (1967). Wijster. A native village beyond the Imperial frontier 150-425 A. D. Palaeohistoria XI, Groningen.
Vedeler, Marianne. “Reconstructing the Tunic from Lendbreen in Norway.” Archaeological Textiles Review no. 59 2017, p. 24-33, SFA, Center of Textile Research, Copenhagen.
Vedeler, Marianne & Bender Jørgensen, Lise. “Out of the Norwegian glaciers: Lendbreen—a tunic from the early first millennium AD.” Antiquity Vol. 87, No. 337. Pp. 788–801.
farolfvs me fecit 