Geographic and Temporal Variations in the East Asian Round Collar Robe, 500-900 CE

            Fayez Tepe, probably representing Kushan seigniors.            

In researching my Tang dynasty impression, I've looked at a lot of robes. Specifically, round collar robes, or yuanlingpao, although as I understand, this terminology is modern (much like referring to the Tang dynasty single edged sword as a dao- more aimed at modern audiences and consumers). Although the garment was introduced during the Northern & Southern Dynasties period, it is very strongly culturally and aesthetically associated with the Tang. Not unreasonably, as far more survive from this period than earlier- I believe only one from the Northern Dynasties and one from the Sui survive. There are also far, far more artistic depictions of the garment, most frequently in tomb murals but also, notably, at Dunhuang, where mid to late Tang iconography is well preserved alongside early and pre-Dynastic artwork. 

For the purposes of this blog post, I want to keep strictly to the 500-900 CE time frame, both because it is immediately relevant to my current project and because after this point, local garment variations 'speciate' even more, The Heian noshi being a good example of this trend, although it also occured on the mainland. The early start date allows us to begin with the initial transmission of the garment too. 

The round collar robe has an origin, in fact, as a Central and West Asian garment known to us as the lapel coat. Betty Hensellek's paper on the subject does far more justice to the initial development of this garment type in Central and West Asia than I could, but the gist is that it emerged in parallel across a pretty wide geographic area- the earliest visual depictions of the mature garment type that I can recall are Kushan, and tenuously date to the 2nd or 3rd century (Fayez Tepe), but a longer tradition can be discerned in the Persian and Median Kandys- essentially a woolen riding coat worn by nomadic pastoralists. By the 5th century, the mature garment form is definitely developed (although it tends to feature a waist seam and initially, only one lapel until the 7th century at the earliest, with these earlier traits persisting in more isolated regions like the Caucuses) and can be seen in some very early surviving extant garments and in Hepthalite and early Turkic murals- the Kizil Caves are a good example of this. Sassanid and Umayyad rock carvings also attest to the garments presence in Iran and West Asia. 

        The introduction of the garment type to East Asia, however, can be traced to perhaps the 5th or 6th centuries, and came about via the Xianbei led Northern Dynasties, who maintained connections into Central Asia. Northern Qi murals, dated to the 6th century, clearly depict men wearing round collared coats with the lapels flung open, paired with trousers, riding boots and sword belts and turbans. With the official cultural sinicisation that occured under the Qi and Sui, the garment was in turn 'made more Chinese', as the clear differences in patterning and construction between this quilted Northern Dynasties coat from the China Silk Museum and this Tang example from the Murong Zhi tomb show:

The Northern Dynasties robe is composed of many pattern pieces. Gores are included on the sides for fullness, a technique not commonly seen in East Asia. They are extended into the sleeves, a practice seen with the Manazan Caves tunic, dated much later to the 11th-13th centuries and excavated from a Byzantine context. Interestingly, some of the surviving original fabric appears to be quilted with straight vertical seams. The Tang Dynasty robe, on the other hand, has a body cut with two pieces. By this point, the use of ribbon ties also seems to have been replaced with loops and buttons, made with a "Turks Head" knot out of probably bias cut fabric matching the body of the coat. The cut is both straighter and fuller, with the body widening to the hem and then continuing straight down from the extension piece- later robes seem to continue the flare into the hem extension, as one particularly splendid 8th-9th century example of very large floral roundels in fiery red silk damask from the China Silk Museum illustrates.

Murong Zhi's funerary robe.

 

 

The robe of red damask, stylistically associated with the middle-late Tang period by its enormous floral medallions, featuring a centred flying crane motif.

This transformation was certainly not instantaneous. The Murong Zhi robe has an unflared hem extension, and certain white damask robes dated to the Tang in the China Silk Museum's collection seem to fasten using ties- interestingly, apparently looping round the back of the neck to tie on the opposite side. Because they are undyed, they're probably intermediate layer robes. The cut also incorporates piecing, with the sides of the skirt being pieced and a gore inserted at the centre front seam presumably to flare the skirt. A similar approach is taken with a coat from Francesca Galloway's collection which probably dates to the northern dynasties, and this is probably a holdover feature. How did the changes in pattern affect the lapel coat in its eastward transmission? The garment seems to have become longer and fuller- the former, at least, among nobles and officials- and lost its tight fitting bodice and flared skirt. These changes seem to have then been communicated outward into the sinosphere, and can be seen in the round collar garments preserved in the Shoso-in repository. The bodice is similarly unfitted, and the garments are constructed following the same principles. However, notable differences still exist.

Among the Shoso-in relics, no robes with hem extensions are preserved. This can be explained as an accidental skew (after all, the collection is hardly representative), but other distinctions are also present. The biggest one is that the extension panels at the front are not symmetrical, with the seam actually being offset significantly from the centre towards the outside of their respective panels. On surviving Tang period garments, the extension seam is always centred. In a number of Shoso-in robes, the body has no flare, and is entirely rectangular in shape, although others do include a flare from the bottom of the sleeve. Differences in textile choices can also be discerned, and a greater proportion of survivals appear to be tie-dyed (although this is probably because some of the garments were made for actors, dancers and servants) and a good number are garment dyed to produce irregular patterns, a technique entirely absent from Chinese garments. Thus, although it has been commonly argued in the past that the Shoso-in pieces are Chinese imports, this seems to cast doubt on the argument. Even the round collar robe made of green pattern-woven silk in Chinese Lingyang roundels seems to have been made following Nara Japanese cut principles- even though the garment is recorded as being imported from the mainland in 756. An alternate interpretation could be that the fabric for the garment was gifted, and then made up in a local style, which may have, in fact, been the principle behind tributary robes more broadly, which were distributed as far as Sri Lanka and probably among other allied South Asian principalities such as the Turk and Patola Shahis (Schafer, Golden Peaches of Samarkand)- this confusion between fabric and finished garment is common in premodern accounts, probably because the majority of labour was bound up in weaving, and not in sewing or finishing.

Displaced centre front seams also seem to show up in Sogdian art at Panjikent. This isn't visible in all depictions of lapel coats there however, and probably has a distinct origin. The coat here was also probably cut more closely and with a skirt flaring from the waist, judging by extant garments excavated from a Sogdian context, ones with a nearby provenance and by art produced by the Sogdians themselves, and so shouldn't be compared too closely with Chinese equivalents, outside of the obvious importation of Lingyang roundels as a persistent fashion statement from the beginning of the 8th century.  

 

Based on the Sinitic extants I've looked at above, here is how I think the robe evolved from the 6th through 10th centuries. A general trend towards increasing fullness and simplification can be observed, and the collar also progressively narrows. By the mid 7th century, ties seem to have been replaced with loop-and-button closures constructed using strips of silk and turks head knots, and skirt gores are no longer seen. Between the Northern Dynasties and the Tang, the side of opening flips too.

Tang Dynasty Banpi

     To begin my blog, I thought I'd document my most recent project It ran quite quickly- it only took 3 or 4 days to pull together, from research and patterning to sewing on the closures- and I was pleasantly surprised by how closely the finished product resembled the originals. 

    The item I set out to reproduce is a Banpi, a short-sleeved (often erroneously identified as sleeveless) vest, worn in Tang China and transmitted to Japan during the Nara period as the Hanpi, where it remained in common use until modernity (while, obviously, still undergoing development, like everything else). The Banpi is worn as an intermediate layer between the innermost skin layer (typically of white ramie cloth) and the outermost garment, the round collar robe, although intermediate round collar robes could also be worn in really cold weather under the banpi (something important to remember is that the capital Chang'an was in the northwest, and winters could be bitterly cold. Fur lined garments, from statuary and accounts (see Schafer's Golden Peaches of Samarkand) seem to have been popular, and extant hood garments also survive, and can be seen in the tomb murals of Prince Zhanghuai, dated to approximately 683-4 CE). In less extreme conditions, the Banpi would suffice, and probably was absent most of the year in the south. I, personally, can't really imagine wearing one along with everything else in an Australian summer, and silk gauze garments that would be far more appropriate also survive. 

    Extant examples of the Banpi often originate outside of central China. Many have been found with a provenance in Tarim and Turfan, and probably entered the market criminally, as tomb robbing is extremely prevalent in Xinjiang Province and elsewhere in Central Asia. Some have also been preserved in the Shoso-in Repository in Nara, Japan, but to my eye these have slight differences in the cut, like the other garments of Shoso-in, and were local products, possibly made using Chinese or Korean pattern woven silk but definitely with distinct cuts. The fact that so many have been found in Xinjiang probably suggests that sinicisation of local's dress occurred during the Tang, although this could also be explained by the presence of Tang garrisons in the region. Finally, one was found during the Murong Zhi tomb excavations, which occured from 2019-2022, alongside an outer round collar robe of purple silk damask woven with paired pheonixes in roundels and trimmed with a bright yellow floral silk roundel on the collar and cuffs, several trouser layers including a pair of open-crotch trousers and silk damask and brocade hose, and a cross-collar ramie skin layer robe (interestingly, completely different from the typical round collar interpretation, but also attested to by the Prince Zhanghuai murals). 

   

     Many Xinjiang-origin Banpi are sold on the private market as "Sogdian silk shirts", despite this kind of garment being totally unattested to in Sogdian art or archaeological finds. This is probably to impart a Silk Road cachet to the sale, in an attempt to bump up the sales price, and also because animal roundels are associated with the Sogdians in popular imagination (despite being used and worn by basically every other cultural group in the region too). I decided to reproduce one based on this original sold by Sotheby's to a private buyer. I liked the shape and the approach to the sleeves, and the colour scheme was fairly close to the materials I had saved for the project in my stash.

    My cloth was 30cm wide (a deconstructed vintage obi featuring a pattern of roundels seen in Shoso-in and Astana Cemetery), so I had to take that into account as a limitation in the patterning, and the side extension pieces made it possible to cut it in two halves, with centre front cross collar extensions. Out of 4 metres, I had just enough to make the bodice. The skirt was from a remnant of plain weave navy silk, and the collar trim and ribbon closures were from another Shoso-in patterned obi, in brown and blue with white, yellow and green accents. I used a machine to sew the seams, as the main fabric was really, extraordinarily thick and tough to pierce. I broke two needles in the process. This worked well for the silhouette however, as depictions of the Hanpi often feature a very strong shoulder, visible through the outer robe.

    After sewing the bodice, I cut out the skirt, pinned it in place to check the measurements and proportions, and then went around the front openings and hem on the machine. I pleated the centre back with an inverted box pleat (sometimes seen on originals, probably to allow for movement). I had to wait until the collar piece had been attached to sew on the skirt, however, as I wanted the trim to disappear into the skirt for a seamless look. The images below show the process for that:

1. Pinning the band of trim into place, wrong side visible, before machine sewing.

2. Folding and ironing the trim round, turning the edge under and pressing, and whipstitching it into place with some silk twist I had lying around.

3. The effect of the collar band. I think it looks really pleasing.

4. Halfway through the process. One side turned down, the other waiting to be sewn into place. Also, gives a good impression of the front of the garment.

 

 

 

I didn't capture the process of sewing the skirt on, but I can promise you it was very boring, and consisted entirely of straight seams. Then there was just the task of making up and attaching the ribbon closures at the waist, and the garment was complete. I'd sewn the sleeve pieces with the selvedge facing out, and as it was barely visible I elected to just leave it as it was. Below are some images of the final product, accompanied by sword and sword belt.