A Byzantine Shirt from Manazan Caves, Turkey

Photography from other angles [12] has revealed an error in my first reconstruction. It shows that the tunic has been cut open from top to bottom, presumably this was done during or post-excavation. I misinterpreted the end of this cut at the left side of the neck as a sloping edge for the neck opening - the front panel is probably rectangular as Dawson has described.
It is also clear the corpse has not been redressed correctly for museum display. It was replaced backwards and it appears the sleeves are empty (even though the body is said to be in good condition, except for the missing skull).
The tunic is displayed with the cut unrepaired and the edges overlapping. This adds to the impression that the garment is far too large around the torso (even with the arms inside), and the hem is also much shorter than would be expected for a woman's dress - it may be a man's tunic reused as a shroud. PB.


A Byzantine community lived in dwellings carved from the rock at Manazan near modern Karaman in southern central Turkey, leaving a necropolis containing 100-150 well-preserved corpses. Most of the tomb contents had unfortunately been looted by the time the shirt and some other textiles were recovered during a rescue excavation in 1991.

The shirt is now on display in the Karaman Museum, still upon the body [1] of its owner (Figure 1):

Figure 1 - The Manazan mummy in Karaman Museum, Turkey.

A full technical examination has been performed (Cardon, 1993) but has not been published to date. The following is therefore based on the observations and photographs of Dawson (2003):

DESCRIPTION: Knee-length shirt of ‘medium weight’ linen, presumably plain weave (1/1), probably undyed and now a shade of brown. Like most early medieval tunics the front and back of the body is formed by a single long rectangular piece with no seam at the shoulders. Full length arms tapering toward the wrist are attached directly to this central panel. At each side a single large triangular godet has been inserted, which run up past the armpits and into the seams constructing the sleeves [2]. The neck hole has a square internal facing [3]. The neck area is complex and unfortunately rather damaged, the presumed front opening is covered by a narrow applied quadrilateral or subtriangular panel, the top half of the sloping side seam of which is left unsewn to form a left-opening flap [4]. A fastening near the top of this flap is suggested by ‘pulling’ of the fabric, but it is now missing. The neck also has a small standing collar [5].

DATING: By the excavators, 8th century AD. By Dawson on costume historical grounds, not before the tenth century. An intended radiocarbon dating [6], if performed, has not yet been released.


Dominique Cardon (1993). A mysterious group of textiles, probably medieval, found in the caves of Manazan, near Tazkale, Karaman province, Turkey. Bulletin du CIETA vol. 71, p183-184.

Timothy Dawson (2003). Concerning an unrecognised tunic from Eastern Anatolia. Byzantion vol.73, p201-210.


FABRIC AND LAYOUT: Use sturdy dress-weight linen (not linen shirting, too light) with a tight weave, undyed - raw (ecru) or bleached.
About 3.5 metres of 90cm wide fabric should suffice for a medium-sized person, or 2.5 m of 120 cm.
The garment pieces should be laid out ‘square’ on the fabric, like in Figure 2 at right - never at an angle to the weave.
Generally the long sides are laid parallel to the selvedge (side edges) of the fabric.
Although no side seams are reported, it would make sense to join the offcuts of the first side panel to make the second - fabric was hard to make and never wasted.
If you have proper woven selvedges don’t cut them off, you should use them just as medieval tailors did.

Figure 2 - Dimensions of the main parts of the shirt.
The five main segments (body, two side panels, and two arms) are based on these simple measurements.
NOTE - add allowances of 1.5 cm for each seam and hem, unless otherwise instructed.

A = length of body from shoulder to knee.
B = chest measurement, divided by two.
C = length of arm from point of shoulder to base of thumb.
D = circumference of hand.

x,y,z: alignment of the (right) sleeve and side panel to the body.

GENERAL SEAMS: Stitch type is whip stitch unless otherwise noted, use doubled linen or cotton thread. The main seams are a form of run-and-fell construction, quite narrow, with all raw edges hidden, sewn with whip stitches (‘saw-tooth fell’) [7].
Internal details of the seam construction are unknown. A reasonable guess would be tatbeet type [8], similar to the modern ‘French seam’ - except that one side of the seam is not cut back before before folding in the raw edges and sewing them down. The method is shown below, Figure 3:

Figure 3 - The ‘tatbeet’ seam recommended for major construction:
1 - Fold 1.5 cm wide edges inward. Whip stitch together from the outside of the garment (shown above in purple). Use closely spaced, shallow stitches.
2 - Turn the seam over and open out the cloth.
3 - Take the two edges together and fold them in half, down to the same side. In photos it looks like all the seams are folded toward the centre line of the
4 - Press the seam flat to the inside of the garment - a light ironing would help. A second row of whip stitches (shown in green) completes the seam.


I usually do the neck first as its easier to do it before the garment is put together.

NECK [9]: The shape of the neck hole on the original shirt is hard to determine. It looks like it is cut straightish over the shoulders and curved at the front (and back?). Note that the neck opening should be a bit deeper at front than back.

1. Cut a square panel for the facing to be sewn to the inside of the shoulder and neck area (Figure 4). This should be 30 to 35 cm on a side.

2. Cut the neck hole at the centre of the facing panel. Keep the neck hole as small as possible. The edges should sit right up against the sides of the neck and over the collarbones. Start off with a small hole, and enlarge cautiously, with repeated fittings(!), and don't forget to leave 1 cm allowance for hemming. Cut a short slit in the front edge, just long enough to allow you pass your head through.

3. Once you are satisfied with the size and shape of the hole, centre the facing on the inside of the body panel. Turn the edges about 1cm in and sew it down flat. Whatever stitch was used, these seams aren’t obvious on the outside of the shirt - maybe a fine running stitch would be suitable.

Figure 4 - Inside facing with neck hole and slit cut. Seam allowances shown as darkened areas.

4. Cut a matching hole and slit in the main body panel. Turn the edges of the slit inward and sew them together, as shown in Figure 5,1. Again I would use running stitch for this. Run a coarse running stitch around the neck hole to keep it tidy - no need to hem it as the edges will be covered by the collar later.

Figure 5 - Construction of the neck closure: 1 - finishing the neck slit. 2 - The external flap and where to attach the collar (A,B,C,D).

5. Cut the outside flap. Start with a rectangle with the width the same as your neck hole, and long enough to cover the whole front slit and about half as much again. Then cut along a straight line from about the mid point on the lower side to the upper left corner, as shown in Figure 5,2 above.

6. Match the upper edge of the flap to the front edge of the neck hole. Cut the edge of the flap to match the curve of the neck opening. Give the left (sloping) edge a narrow double hem. Turn the other edges in about 1cm in and sew the flap down on the front of the shirt, again using a fine running stitch. Leave the upper half of the left side unsewn to make the opening.

7. Fix the right hand edge of the inside front slit to the flap using a line of running stitch - from point C on Figure 5,2 downwards to the bottom of the slit.

Figure 6 - Forming and attaching the collar:
1 - Match one edge of the collar strip to the outside edge of the neck hole. Sew together with fine running stitches about 0.6 cm from the edge (shown in purple). Press the collar upward over the seam.
2 -Fold the top of the collar over and tuck in the raw edge on the inside of the neck. The collar should be about 2.5 cm tall. Sew it down inside with whip stitches (shown here in green).

8. The collar is made from a strip of linen, about 8 cm wide and 60 cm long. Attach one edge to the outside of the neckline with a fine running stitch, as shown in Figure 6,1.

Work from the left side of the front slit (at point A) around the neck to point B, where the outside flap starts - now sewing the collar to both layers until you get to point C, the other side of the front slit.

From here continue on the top edge of the flap only, to point D.
Press the collar upward over the seam you have just sewn, then fold it down on the inside, trim and tuck in the raw edge until the collar stands about 2.5 cm high.

Whip stitch the fold down on the inside as shown in Figure 6,2. Again pressing may help.

BODY: NOTE - the shirt can be quickly assembled and checked for fit by using only the first step of the tatbeet seam shown in Figure 3,1.

1. Attach the arms to the body, matching the centre line of the sleeves to the shoulder line of the body panel.

2. Attach the triangular side panels, refer to Figure 2. Start from the bottom, sew them to the body panel. At the armpit (x & z) continue along the underside of the sleeves until you reach the point (y). Slight recutting might be necessary, depending on how even your sewing is.

3. Sew together the arm seams to the wrists.

4. Check for fit. The bottom and wrist hems can be cut to correct length at this stage as well.

5. Finish all tatbeet seams as in Figure 3,2-4. This can get fiddly where pieces join so pin and repin until you are happy. Remember, the body seams are folded in the direction of the centre line of the garment.

6. Give the bottom edge and the wrists a narrow double hem (1 cm or less) - see Figure 7.

7. Add the fastening for the neck at point D on Figure 5,2. It was probably a small button and loop. One common form of Byzantine button was a hollow ball of sheet bronze, usually less than 1 cm diameter, with a wire loop shank [10] - small pellet bells easily available from Indian clothing shops work admirably. Round textile-covered buttons with an organic core are also used in this period [11]. Finished!

Figure 7 - Double hem constructed with whip stitch.


Dominique Cardon (1993). A mysterious group of textiles, probably medieval, found in the caves of Manazan, near Tazkale, Karaman province, Turkey. Bulletin du CIETA vol. 71, p183-184.

Timothy Dawson (2003). Concerning an unrecognised tunic from Eastern Anatolia. Byzantion vol.73, p201-210.


[1] I have seen no scientific report on the sex of the mummy, although in popular press as well as tourist websites it is always said to be female: “Found in the Manazan area of town Karaman, the mummy of a 17-year-old girl bears surprising features. Unlike other skeletons found in Manazan graveyards, the body of the young girl did not decay and surprisingly did not contain traces of any chemicals.” - Turkish Daily News, Dec 24th 2007. PB.   See also my new web page - ‘Costume of Byzantine Mummies of Anatolia’

[2] A somewhat similar extension of side panels into the sleeves has been noted on a 15th century Italian coat now kept in Bucharest. PB.

[3] A feature shared by some Egyptian shirts of later date: Gillian M. Vogelsang-Eastwood (1987). Two children’s galabiyehs from Quseir al-Qadim, Egypt. Textile History vol. 18, p133-142.


(Left) A similar neck construction seems to appear in the tunics of these two onlookers at Christ's baptism, though here the square panel becomes more of a feature by placing it on the outside of the garment and making it in a contrasting colour.
Baptism of Christ (detail), mosaic in monastery church of Nea Moni on Chios, Greece mid-11th cent. Photographed in situ June 2009. PB.

[5] This is a unique survival of a collared shirt of such early date, but the fashion was likely quite widespread in the Middle Byzantine period, as is fully discussed by Dawson. Two discoveries from Norway show this innovation has also appeared quite early elsewhere, though in a probably unconnected context. A fully-dressed bog body discovered at Skjoldehamn in 1936 had a wool undershirt with a 5cm high standing collar. The Skjoldehamn clothing is radiocarbon dated to c.1000AD and is currently held in Tromsö Museum. A bundle of clothing excavated from a churchyard at Guddal included another woolen shirt with an 8.5cm high collar, radiocarbon dated to c.1100AD. Their find locations are seperated by the length of the country, and their collar constructions are quite individual: Skjoldehamn - see
Dan Løvlid's masters thesis; Guddal - see University of Bergen Museum photo archive, particularly images 030994 & -995.

[6] Cardon, p184.

[7] Dawson, p203.

[8] As used on 12th century Egyptian linen shirts - Vogelsang-Eastwood, op cit. note 3, p136.

[9] Note this neck construction is complete conjecture based on Tim Dawson’s observations (cheers, Tim!), a blurry photograph and some knowledge of similar garments - compare the alb of St. Hugo (12th cent.), monastery of La Valsainte, Switzerland. Brigitta Schmedding, Mittelalterliche Textilien in Kirchen und Klöstern der Schweiz, Katalog. Bern 1978, cat. no. 289; and also Alanic female costume of the 7-9th century from Moshchevaja Balka - Anna Ierusalimskaja, Die Gräber der Moshchevaja Balka. Frühmittelalterliche Funde an der nordkaukasischen Seidenstrasse. Munich 1996, p159.

[10] For example, from Sardis - Jane Waldbaum, Metalwork from Sardis (Archaeological Exploration of Sardis, vol. 8). Harvard 1983, cat. nos. 784-796 (incorrectly described as beads); from Constantinople - R.M. Harrison, Excavations at Saraçhane in Istanbul. Vol. 1. The Excavations, Structures, Architectural Decoration, Small Finds, Coins, Bones and Molluscs. Princeston University 1986, cat. nos. 546-552.

[11] For example, on a Byzantine or early Islamic Egyptian child’s tunic (5-6th cent.) held in the Allard Pierson Museum, Amsterdam: KH Aben (1979). Een koptisch kinderjurke. Medelingenblad Vereniging van Vrienden van het Allard Pierson Museum vol. 17, p12-13.

[12] Refer to - Timothy Dawson, Byzantine Infantryman: Eastern Roman Empire c.900-1204 (Osprey Warrior 118). Osprey: Oxford 2007, p.17.

‘A Byzantine Shirt from Manazan Caves’, written and webbed by Peter Beatson.

(c) Birka Traders 2008-13. Not to be copied without permission.

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