EDITED: The Arnegunde Kaftan Project: Conjectural Merovingian clothing construction of the mid 6th century


The Arnegunde costume

The Arnegunde costume…

In 1959, archaeologists excavating under the Cathedral of Saint Denis in Paris, resting place of the Kings and Queens of France, found a sarcophagus containing the body of a woman. The woman bore a ring inscribed with the name “Arnegundis.” She is thought to be Arnegunde, wife to Clotaire I (511–561) and mother of King Chilperic († 584) (Perin et al. 2007, 182).

This project brings together current research to gain a better understanding of the context of her burial and the textiles that comprised her burial garments. This project focuses on the garment called the Robe or Kaftan  in current publications. The gold thread embroidery on the sleeves stands out as a unique textile apart from the Germanic tradition suggesting an extra-Merovingian origin. The author does not contend that the garments are exact replicas of the garments in which Arnegunde was buried, but every attempt was made to make sure that they would be familiar to Arnegunde and acceptable to one of her station.

Background of Merovingian Period

The period from the late 5th century to the late 8th century northern Gaul is known as the Merovingian, after the semi-mythical ruler, Merovech.  We know very little about their daily lives, having to rely on the few written texts that survive and emerging archaeological evidence. The Franks were a group of Germanic tribes — the Chatti, the Ripuarians, and the Salians — who shared similar laws and customs. In the 4th and 5th centuries AD, they began settling in the Roman region of Gaul (which included what is now Belgium, France, Luxembourg, and some of Germany and Italy). At first they lived in Belgium; eventually most of Gaul would belong to the Frankish empire. (Geary, 1988)

Very little is known about the early rulers of the Franks. A Salian king named Merovech (or Merovee) founded the royal Merovingian dynasty, whose kings were notable for their long hair. At that time the Franks were allies of Rome, and Merovech is said to have fought with the Romans against Attila the Hun.

The time period of this study was the mid-to-late sixth century and is based on a set of grave goods associated with Arnegunde, queen and wife to Clothar. The Merovingian period was one of warfare, on the field and in the palace. Clothar came to power through violence and married the royal widow, Guntheuca, taking the widow as he took the kingdom. At this time, women were a temporary accessory in the king’s bed, and only through the bearing of an heir found any hold on position and power. As Stafford (1983) says, “his wives were taken up and put aside as political shifts demanded rather than passions prevailed.” (p. 51-2)

Between the 530s and the 550s, Clothar had a succession of wives and concubines: Radegund, daughter of King Bertacharius; Ingunde was in favor in the 520s, and again mid-530s; by 537 she was replaced by her own sister Arnegunde; before 540 he had another mistress, Chunsina. Very little is known about the origins of the sisters Ingunde and Arnegunde, and the concubine Chunsina. (Stafford, 1983)

Arnegunde, as the mother of the future king, Chilperic, was afforded a burial place in the Basilica of Saint Denis, to the north of Paris.

Archeological Evidence

The Basilica of Saint-Denis was the burial location of many of the Merovingian dynasty starting in the east end of the Basilica, near the grave of Saint-Denis, around 475. (Fleury & France-Lanord, 1998) The Merovingian graves were excavated over a period of about 30 years, starting in the 1950s by Eduard Salin and Michel Fleury. The archeological finds were published in French and German publications over the years and only recently have articles been published in English.  (Rast-Eicher, 2010)

At Saint-Denis, a principal burial place for Frankish kings, a large stone sarcophagus was found under the choir.  The contents of this tomb consisted in great part of a mass of precious textiles, silks, and other organic materials.  The large limestone coffin was found to contain the remains of a woman’s body which had been embalmed before burial, so part of the lungs survived.  She was about five feet tall, of slight build, and blonde hair. (France-Lanord, 1979; Fleury & France-Lanord, 1998; Perin, 2007) A monogrammed gold finger-ring, ARNEGUNDIS REGINE, identified the female buried inside as Queen Arnegunde. (Rast-Eicher, 2010) She is believed to have died around 580. (Périn & Calligaro, 2005)  New studies have been completed both on the skeletal remains and the conserved textiles. Rast-Eicher (2010) explains that Arnegunde was older at the time of her death than originally thought and probably died from a dysentery outbreak. She had childhood poliomyelitis leaving her with a damaged right leg.

Burial customs of the middle to late sixth century provide a wealth of metal artifacts, but textiles are far more rare. For this reason, the extraordinary preservation of the Arnegunde artifacts provides a rare glimpse into the textiles, and provides clues to a conjectural clothing construction. By the seventh centuries, burial custom were changing and grave goods become much more rare. (Owen – Crocker, 1986; Effros, 2002) The stone sarcophagus provided a protected environment, which allowed the preservation of some of the textiles, both as fragments and as pseudomorphs (mineralized fibers) on the metallic objects. (Marzinzik, 2008)

The following sections will discuss selected garment fragments found in Arnegunde’s tomb. Additional textiles, possibly shrouds, were also found but will not be discussed in this article. Please see Rast-Eicher (2010) for a longer description of the most recent investigations into the textiles.
Following this section on the extant textiles, I will discuss the project Kaftan.

Arnegunde’s Burial Clothing


Working from the outermost garment, the kaftan, we will talk about some of the most recent theories of fiber, textile and construction. Sadly, the garment is incomplete so there is much left to conjecture. (Perin & Calligro; Rast-Eicher, 2010) The garment was a front closing kaftan or robe, probably floor length. Rast-Eicher (2010) calls this garment a “mantle”.  For the purposes of this research project we will use “kaftan.” The front opening was edged with tablet woven bands and the sleeves had a gold embroidered band. The hose and garters found with Arnegunde have led to a great deal of specualation about the length of the outer layers of clothing. Owen-Crocker (1986) comments that the decorative nature of the garter and shoe beuckles indicate the clothing would have had a lower length of just below the knee. We now know that the outer layers were nearly floor length. (Rast-Eicher, 2010)

The textile of this garment is described by Rast-Eicher as “the so-called ‘violet’ piece, a textile with one system made of animal fiber, the other system is of plant fiber, but mostly not preserved.” (2010, p. 209) She does not define the textile structure or colors of the fibers. However, past publications have described this garment as reddish or purple. (France-Lanord, 1979; Marzinzik, 2008)

The lower portion of the sleeve was different fabric than the body. Rast-Eicher describes it as “ samite 2/1 Z  … with a warp proportion of 2:1.” (2010, p. 210)  she goes on to further describe the embroidered gold band as being 7 cm. below the cuff seam. The textile underlying the gold embroidery has not survived, although a few red threads are visible.

Although Rast-Eicher (2010) describes the fiber content of the textiles as animal or plant, other researchers have described the fiber content as silk, linen or wool. (France-Lanord, 1979; Marzinzik, 2008) Just how these various layers of textiles actually were worn is still highly conjectural.

When it comes to colors used in the textiles, again much is not known. France-Lanord calls this textile “violet”. (1979) An under layer has been called “reddish”. (France-Lanord, 1979; Marzinzik, 2008) Both the silk textile and the purple to red dyes were imports to the Merovingian territories. These would have been imported along the Rhone or Rhine trade routes from the Byzantine empire. The presence of these luxury items as grave goods marks the high status of the individual buried in sarcophagus 49. (Marzinzik, 2008)

The front edge of the kaftan was decorated with a brocaded tablet woven band made of at least 100 tablets and is approximately 6.5 cm wide. Rast-Eicher describes the band as “ brocaded with a triple silk thread (z-spun) and displays a pattern of diagonals and lozenges (Fig. 33.3).” (2010, p. 210) A second band in a simple tabby/repp, 16 warp threads wide is sewn to the “violet” textile. The weft has not survived, except for one brocading weft stitch. The wider band was key in identifying the layers as it lay under the belt and was found along the skeletal remains down to the lower leg. (Rast-Eicher, 2010)

Walton Rogers (2007) suggests similarities between the Arnegunde kaftan and the Woman’s garment found at Sutton Hoo. Both had ornamented cuffs, although in Sutton Hoo the ornamentation was tablet woven bands. The style of a front opening kaftan had correlations in other garments of the period from cultures to the east and in artwork from the period.

Two round brooches were found on the body in a position suggesting their use as a closure along the center axis above the waist. The period of the later sixth century is one of fashion change for Merovingian women. The earlier Continental tradition of four brooches was giving way to the Byzantine-influenced style of a single brooch at the neckline holding a mantle over a brooch-less tunic. (Rogers, 2007; Perin, 2000)

The Project Kaftan Reconstructed

The kaftan is the outermost layer of the ensemble I created for this project. It also took the most time because of the various time-consuming pieces such as the gold embroidery and embroidery edging the front opening. It actually assembled pretty quickly once the components had been finished. It was mostly machine sewn where the seam would be hidden in order to save time with a lot of hand finishing on visible portions of the garment.

The fashion textile is a tabby with a dark red linen warp and a black silk weft. The lining is silk twill with a gold warp and a purple weft. The garment is bag lined, with a couple areas that were pieced together because I didn’t have quite enough fabric. The sleeves are lined in the turquoise taffeta that is also on the lower cuff of the sleeves.

The pattern was drafted using the rectangular construction method. I made a compromise on design between a flattering fit and a more period style represented by the Grande Robe of Bathilde. This front-closing garment is the most similar extant garment available.

Figure 1. Pattern and cutting diagram for the Grande Robe of Bathilde (Laporte & Boyer, 1991)

Figure 1. Pattern and cutting diagram for the Grande Robe of Bathilde (Laporte & Boyer, 1991)

There is some question about whether this garment could be attributed to Bathilde, due to its size being much larger than would fit Bathilde (personal communication, Wamers, 2013). And there is some discussion on whether it is actually from a couple centuries later, but until new information is published, we are using the current designation at Bathilde’s Grande Robe. If the academic community makes a definitive decision otherwise, I will be happy to make the change to this publication.

The project kaftan has wedges on each side starting at the sleeve and ending at the hem. Bathilde’s Robe has side gores that start at or below the waist, so there is some difference between the two garments in silhouette. I chose this style because it was more flattering on my Rubenesque figure.

Running along the front opening edge of the project kaftan, there is a band of silk embroidery on a silk/linen tabby woven fabric. The silk embroidery floss is recycled yarn from silk sweaters purchased at thrift stores.

The design of the embroidery is inspired by the card woven bands in the Chelles museum. In the original Arnegunde kaftan, this was card woven, but the technique is beyond my current skill level. So, taking a cue from Bathilde who used embroidery on a garment as a representation of her royal jewels, I replaced the card weaving with embroidery.


Around the neckline, is a second embroidered band that I had to create to make up for the fact that I didn’t have enough of the first embroidered band to go all the way around the neckline. I used diagonals which are common in Migration period card weaving  as a main design element. The outside edge of the embroidered band is bound with the same turquoise taffeta that lines the sleeves.

Figure 2. Design for embroidery edging the front opening. Based on the card weaving in the Chelles Museum.

Figure 2. Design for embroidery edging the front opening. Based on the card weaving in the Chelles Museum.

Embroidery on Arnegunde kaftan sleeve

The gold embroidery stands out as separate in look and technique from what is currently known about western Migration era textile traditions. It is possible that the embroidery was imported as a complete piece, either as trade or as a gift, from Imperial Byzantium. (Crowfoot and Chadwick Hawkes, 1967, p. 55). The gold embroidery on the cuffs was made of a fine spiral wrapped thread. The gold foil was wrapped around a core of silk, .25 mm thick and .8 mm wide.  There were about 13 to 14 turns to an inch. The diameter of the threads of gold reached about 0.45 mm with a length of up to 150 mm. The gold was couched down using a very thin silk thread in very small stitches that were more or less close to one another, depending on the type of pattern. (France-Lanord, 1962)

Figure 3. The extant gold embroidery. (France-Lanord, 1998)

Figure 3. The extant gold embroidery. (France-Lanord, 1998)

The rosettes show three variants:

Heart flower with round petals and eight trapezoidal-shapes;

Heart flower with round and six pointed oval leaves, which are separated by spherical triangles inscribed spiral from one another;

Circle whose inner edge is decorated with small triangles, which are designed with a spiral; inside the circle are (from left to right): a small, vertical almond, a large, well vertical half almond (with the straight side left) and three small seeds that are spread like a fan in the right half of the circle. (France-Lanord, 1962)

Only 17 of the rosettes survived, although there were originally 18 or 19 in total. Parallel gold threads run along the edge of the band. The band was 30 mm wide and about 37-38 cm long. (France-Lanord, 1962)

In the project kaftan sleeve embroidery, the design of the roundels was altered slightly to make one of the designs look more like bees, which are a part of my heraldic device. The base fabric of the embroidery was a natural linen warp and a red silk weft in a tabby weave. This linen/silk fabric is conjecturally very similar to the textile found in Arnegunde’s grave. I used a synthetic metal thread for cost savings and ease of use. The threads were couched onto the textile using a silk sewing thread.

The design of the gold embroidery on the cuff.

The design of the gold embroidery on the cuff.

The same turquoise taffeta used to line the sleeves was used as an edging on the embroidered band. Narrow bands were cut and hand sewn to the embroidered band to give it a clean finish. We don’t know that the bands would have been bound, but I liked the clean finished edges.



The completed bands before they were put onto the sleeves.

The completed bands before they were put onto the sleeve


This project was the culmination of a multi-year process. The kaftan is one part of the ensemble, and the first one completed. Future projects include learning the card woven brocade technique for the front opening edge, the leather belt with gold leaf and embroidery, shoes and garters, gold brocaded vitta, and if I feel particularly daring, the metal buckles for the garter and shoes.

One of the constraining elements of this project is that very little of the published resources is in English or accessible to non-academic researchers. It took time to track down and translate many of the resources used for this project. And there was always the dangerous distraction of the most recent publication find.


Audollent, A. (1921) Les Tombes des Martres-de-Veyre. Man, 21. 161-164.

Bachrach, B.S. (1973) Liber Historiae Francorum. Coronado Press: Lawrence, Kansas.

Effros, B. (2002). Caring for body and soul. University Park, PA: Pennsylvania State  University Press

Fleury M., and France-Lanord A. (1998) Les trésors mérovingiens de la basilique de Saint-Denis, Woippy, Klopp.

France-Lanord, A. (1979) La fouille en laboratoire. Dossiers de l’Archéologie 32, 67–91.

Geary , P. J. (1988) Before France and Germany, Oxford, England: Oxford University Press.

Laporte, J. P., & Boyer, R. (1991). Trésors de Chelles: Sépultures et reliques de la Reine Bathilde (+ vers 680) et de l’Abbesse Bertille (+ vers 704). Société Archéologique et Historique de Chelles.

Marzinzik, Sonja, (2008). “Expressions of Power – Luxury textiles from early medieval northern Europe” Textile Society of America Symposium Proceedings. Paper 113.

Owen-Crocker, G.R. (2004). Dress in Anglo Saxon England. Boydell Press.

Perin, P. (2000). Aspects of Late Merovingian Costume in the Morgan Collection. In From Attila to Charlemagne: arts of the early medieval period in the Metropolitan Museum of Art. 242-267.

Périn P., et al. (2007) La tombe d’Arégonde, Nouvelles analyses en laboratoire du mobilier métallique et des restes organiques de la défunte du sarcophage 49 de la basilique de St. Denis. Antiquités nationales 36/2005, 181–206.

Périn P. and Calligaro T., (2005) “La tombe d’Arégonde: Nouvelles analyses en laboratoire du mobilier métallique et des restes organiques de la défunte du sarcophage 49 de la basilique  de Saint-Denis”, Antiquités nationales 37, 181-206.

Rast-Eicher, A. (2008) Textiles et costume du Haut Moyen Âge. Histoires et Images Médiévales 20, 50–56.

Rast-Eicher, A. (2010) Garments for a Queen. North European Symposium for Archaeological Textiles X. 208-210

Rogers, P.W. (2007). Cloth and clothing in early Anglo-Saxon England: AD 450-700. (No. 145). Council for British Archeaology.

Stafford, Pauline, (1983). Queens, Concubines, and Dowagers: The King’s Wife in the Early Middle ages. Athens, GA: The University of Georgia Press.






  1. Reply
    Alena February 24, 2013

    Lovely post, and even lovelier clothes! You research is marvelous, and the photo looks just like all the illuminations!

  2. Reply
    Siglinde February 25, 2013

    Lovely! I look forward to seeing how you complete this outfit. It’s on my wish list for my own costuming.

  3. Reply
    Erin July 30, 2013

    Hi! This is fascinating; thanks for sharing it.

    A quick question – in the paper, you cite Rast-Eicher 2010, but in the bibliography there is only a 2008 publication. It is a typo or is there a source missing from the bibliography? I would love to follow up on it.

    Also, a minor note – in the first paragraph about the kaftan project, there are a couple of minor typos. Owens instead of Owen and bueckles instead of buckles. I know that I never see my own typos, so I thought I’d mention it in case you find yourself editing the piece some time.

    I love the overall look of this kaftan on you. It hangs beautifully and looks lushly rich. What inspired your colour choices? The turquoise is so visually dramatic! I can be really conservative about colour, but I love how this looks.

Leave a Reply

Your email address will not be published. Required fields are marked *