Felting between East and West

Murray Lee Eiland III


Archaeological Evidence
Using intuition alone - as the archaeological record is unlikely to provide such fragile information from the distant past - one can assume the craft of felting predates that of weaving. While the earliest humans no doubt used animal skins to cover their bodies, it is no great leap to think that hair from these skins would be matted together to make a different kind of covering, perhaps the first step in an industry that would lead to woven structures. At the same time there is no evidence that animal hairs were felted in Africa (barring the north in historic periods, although it was apparently unknown in ancient Egypt), or America, despite the extensive evidence from ancient textile cultures with the South American domestication of the llama and alpaca. While it may be argued that there is little evidence for the antiquity of sub-Saharan African fabrics, the evidence from South America is unusually rich. Part of the reason felting is not attested from these regions may be cultural. Felts were not esteemed in ancient China, where they were associated with barbaric nomads. Only during the Mongol period was the Imperial harem decorated with felts, which from documentary sources were ornamented with floral patterns and applique work (Bidder 1964: 87-91). Surprisingly, ceramic vessels from the Mongol period (Chi-chou wares) can be decorated with swirl patterns that may reflect felt designs (fig 3).

Another reason felts were not adopted in all areas was the lack of wooly animals. In the South Pacific, with no wooly animals from which to obtain hair it is not surprising that a felting tradition did not emerge. While the archaeological record may never totally disprove that felting of animal hairs of some kind did not exist at some point in these regions, it is clear that felting is not so universal that it is a common trait shared by numerous groups spanning the continents. Instead, felting may be, like other technologies, particularly suited to distinctive environments. However, it is interesting to note that in some of the areas that did not have a tradition of felting, a similar technique, although not using animal fibres, emerged.

Bark cloth has been made particularly in Africa and South America, and it is very well known from Indonesia. ‘Tapa’ cloth, first seen by Captain James Cook (1728-79) and his crew in the Pacific Islands, is the name used for the product there, and has become a label for the industry as a whole. This cloth was primarily used for clothing, and although there is something of a tourist market for it, it has largely disappeared (Harris 1993: 50-51).

There is, however, evidence to suggest that the technique of felting existed long before it was first mentioned in Chinese sources in c. 2300 B.C. Mellaart recovered a wall painting from Çatal Hüyük (Shrine E VIII 14) that was divided into panels by five vertical lines. Each panel contained whirling curvilinear motifs in a buff-brown, outlined with black lines, against a white ground (Fig 4). He speculated that in both pattern and edging they resemble felt appliquè (Mellaart 1966: 180). Burkett (1977: 113) notes that while there was no felt found in this level, level VI (above) did yield pressed animal hair (but bore no trace of a pattern). Yet it is the designs from the shrine that are the most striking, and Burkett notes that the designs in this painting are not consistent with woven structures. Black bordered edges are not like woven kilims, while much early weaving produced rectilinear patterns so that a curve would be represented by a stepped design. Felts naturally portray curved lines, as wool shrinks irregularly, making any irregularity less noticed in a curvilinear design. The most striking feature of the designs is that they find ready parallels with felts made in modern Turkmenistan, while the painting also suggests that even at this early date felts could be made using coloured designs stitched onto a base. While there will continue to be debate about this evidence, it is unlikely that ornamental felts from this remote period will be recovered.

Other sites deeper in Asia offer good evidence for felts from an early date. “Cherchen man” from Western China (c. 1000 B.C.) has leggings with bright horizontal stripes of bright red, yellow and blue (Barber 1999: plate 1), before woven socks were invented, felt must have played an important role protecting the feet. Woven structures can also be felted to make them more compact - wind and water proof - while at the same time causing shrinkage of the fabric. As the latter method is not a “primary” use of felting (in that a woven structure is then felted) it will not be considered here. The archaeological record for felt is not extensive, as among all perishable textiles it may be the most delicate. As a rule felts do not survive long when used heavily, and their structure is such that they are perhaps an ideal medium for insects.

The felts from the Pazyrk burials are notable in that they are some of the earliest and best preserved examples of a purely decorative nature. They are assumed to have been buried by an early group of Iranian speaking people. While their date has been the subject of ongoing speculation, the materials recovered from these kurgans date from the 5th-4th centuries B.C. (Barkova 1999). While the woven materials have received detailed attention, the felts have been relatively neglected. One of the most interesting observations was that of V.N. Kononov, who noted that the wool used for the cloth and the thin kinds of felt was of fine fibers (11-26µ), while the thick felts could be made of coarser hairs. White was the predominant colour, but dark grey sheep which yielded black wool were also cultivated (Rudenko 1970: 57). From the detailed designs one can surmise that there was already a long tradition of felts with figural designs (Figs 5-6). Besides decorative felts, and felt used for clothing, felts were also used for shelter, although it is likely that felt tents played a more important role in other steppe regions. From what can be deduced from the archaeological evidence, three kinds of dwellings existed in the Altai: birch bark tents, felt tents, and log cabins. There were also small conical hexapod felt covers found in all the large barrows used for smoking hemp. The top of the burial chambers bore huge sheets of birch and larch bark and felt (Rudenko 1970: 62-3). Rudenko considered it unlikely that portable felt dwellings were well represented in the Altai region, as there is plenty of timber and bark for dwellings, and there is no evidence that the Altai folk were on a migratory path. Finally, felt of such quantity as to cover a large dwelling would only be available to families with large numbers of sheep at their disposal.

About 500 years later than the Pazyrk burials, and representing a distinctive tradition, are the felts excavated from Noin Ula. This site, thought to represent a distinctive ethnic group (the Hsiung-nu: thought to be early Turks) has a distinctive felting tradition. Many of the Pazyrk felts used couched threads to separate coloured wool (Fig 7), and often depicted animal scenes. In contrast, the felts from Noin Ula relied upon abstract all-over designs and quilting with undyed materials (Rudenko 1969: pls XLI-XLV). In both technique and more generally design, these felts are similar to the quilted “Mongol” felts made in the same area today (Chambros 1988, 35). Further evidence of felt can be found in the arts of recently settled nomadic peoples. Reliefs (particularly Persepolis in Iran) suggest that the Achaemenids (550-330 B.C.), in keeping with their nomadic origins, wore tall soft felt hats (probably called kyrbasia), while the upright tiara was reserved for the king. The Parthians (150 B.C. - A.D. 250), a later dynasty who ruled Iran, used a similar type headgear. Parthian hats, particularly after Mithradates II (123-88 B.C.), differed from the earlier examples in that the portion covering the head also bore a neckflap an side pieces that covered the ears and cheeks. These additional features were not simply ornamental, but necessary in cold climates, and modern nomads have a similar kind of headdress (Olbrycht 1997: 40-41).

Perhaps some of the most difficult to interpret information about textiles comes from examining depictions of fabrics for clues as to their structure. While one can be certain that particular items of clothing were felt, floor covering are not so straightforward. Gulácsi (1994) presents material from Manichaean illumiated books, dating from the 8th-9th centuries A.D., and suggests they are felts and not woven textiles. Despite the rather fantastic colours on many of the rugs, the designs may faithfully reproduce contemporary patterns. The features that led Gulácsi to identify them as felts was the style of fine lines forming a dense, rather monotonous, decoration that covers the entire rug. The style is very similar to Central Asian felts that are composed of identical design elements with curvilinear and at times vaguely vegetal shapes. Gordon (1980) notes that while some suggest that the art of feltmaking may have died out in Western Europe only to be re-introduced from the Near East during the Crusades of the eleventh century, it is clear that during this time the tradition did not die out in the Mediterranean or in Scandinavia.

Felt was commonly used in Europe during the Middle Ages for hats, while its use for other bits of clothing was restricted to low-status items. Excavations dating to the Medieval period (14th century A.D.) in the city of London recovered eight small fragments of felt, as when aged this fabric becomes brittle and breaks apart more readily than woven fabric. Fibres were from medium to coarse thickness, mostly undyed, many natural brown. The majority of materials were probably hat fragments - as felt hats are well attested in manuscript illumination - or footwear linings (Crowfoot et al. 1992: 75-6). Unlike the case for sedentary societies, where the need for felt may be less acute, in nomadic cultures felt had, and continues to have, a very different status.