Below you will find a description of some basic characteristics of Salor weaving, based on design, weave structure, and colors and dyes. We have also included some information about dating pieces.
We note at the outset that some controversy exists over the criteria for attributing weaving to the Salor tribe. The most widely accepted approach is to use weave structure criteria (e.g., asymmetrical knot, open to left; significant warp depression) as the exclusive basis for a Salor attribution. In his general book, however, Murray Eiland maintains that we do not currently have enough evidence either to designate such pieces (at one point called "S-group" weavings) as Salors or to assume that the only true Salors are those having "S-group" structural features. Murray L. Eiland, Oriental Rugs 190-95 (1980). He is inclined to accept as Salors many weavings that do not have S-group structural features, including some that are quite contemporary (e.g. some Afghan Mauri weavings).
(For Eiland's current views, click here for his recent exchange with R. John Howe. In sum, Eiland remains skeptical of equating S-group structural features with a Salor tribal attribution, but concedes that such an equation cannot be ruled out definitively.)
George O'Bannon also has sounded a cautionary note about the equation of S-group structural characteristics with a Salor tribal attribution. In his article Turkoman People and Turkoman Weavings, published in Vanishing Jewels: Central Asian Tribal Weavings 18-19 (1990), he suggests that the Salor tribal attribution has become too dependent on structural characteristics. "[W]e know from Tekke rugs," he observes, "that there can be considerable variety in the weaving corpus of a major tribe. This is seen not only in weaving techniques but wool quality, dyes, border pattern variations, and sizes."
Some support for Eiland's views can be drawn from General Andrei Andreevich Bogolyubov's book, first published in 1908. General Boglyubov was Head of the Transcaspian Province at the turn of the century and amassed a collection of 139 Turkoman pieces. Of the four pieces Boglyubov attributed to the Salor, one lacks S-group structural features. And he characterized four pieces having S-group structural features as "Pende" rather than Salor. Bogolyubov's attributions deserve some deference, given his geographic and temporal proximity to the source of the weavings.
As with other Turkoman weavings, the standard design features in Salor work vary according to the nature of the weaving.
Main Carpets. The gul used on Salor main carpets, a variant of the gul-I-gul is a nearly perfectly round lobed gul with four animal forms in its center. In contrast with other versions of the gul-I-gul, the three "clover-leaves" in each quarter originate from a single point. The gul historically most frequently described as the Salor gul is a quite different turreted device and appears in Salor weavings on bags but apparently not on older main carpets. This latter ornament has been used by the Saryks, Tekkes, and Ersaris and is also a favorite of Baluch weavers. "The number of rows of guls varies between four and seven, with between ten and thirteen guls per row." Turkmen: Tribal Carpets and Traditions 67 (Mackie & Thompson, eds. 1980). The minor ornament on Salor main carpets are often a small, but detailed, octagonal version of the archetypal chuval gul. Main carpet designs often have spacious placement of the ornaments on the field. Borders on Salor main carpets are simple and few. Three is usual. The major border in published examples of Salor main carpets uses the same repetitive motif.
Chuvals, torbas. Salor bags are made in both the chuval and torba formats. Some pieces have the rectangular shape of bags but were apparently woven for other, perhaps ceremonial purposes. Major ornaments on bags and such trappings include: the turreted "Salor" gul (see above), the so-called chuval gul, and the shemle or chemle, ak-su and kejebe designs. Trappings sometimes exhibit one to three large oval medallions (darvaza design) together with the kejebe device. See, e.g., Uwe Jourdan, Oriental Rugs: Volume 5 -- Turkoman (1989), plates 2-4; Turkmen: Tribal Carpets and Traditions (L. Mackie & J. Thompson, eds. 1980, plates 9 & 14. Pieces employing the kejebe design often have a T-shaped field, as in plate 14, Turkmen: Tribal Carpets and Traditions. Minor ornaments on bags include the octogonal version of the chuval gul and, on bags containing the Salor gul as the major motif, the charch-palak gul. Again spacious graphic design is frequent. Bag borders include the chamtos design, kochanak design, the "flag" design and versions of kochak and S-borders. The elems of Salor chuvals are decorated with stately, archaic-looking plant forms. Torba elems tend to show a larger three-sided band of the chamtos design.
Ensis. Salor ensis are extremely rare. When Robert Pinner wrote his superb article, Salor Ensis, 60 HALI 86, only seven were known. At the time of the 8th ICOC in Philadelphia, in November 1996, two more had been added to the list. Salor ensis share some features of other Turkoman ensis. The most basic shared feature is the four-panel hatchli design. Like almost all Turkoman ensis, Salor ensis have two elem panels and, like most Turkoman ensis, three borders. Beyond this, "the Salor ensi is to some degree a compendium of Turkoman designs which recur, as if at random, in the weavings of several Turkoman tribes." Id. at 86. As in Tekke and some other ensis, the rows of motifs in each of the four quadrants of the field are separated by lines. Also as in Tekke and some other ensis, the field motifs consist of bird's head devices. Like Tekke ensis and some Ersari, Yomut, and Arabatchi ensis, Salor ensis employ an outer border consisting of the sainak design. In contrast with other Turkoman ensis, "the lower part of the Salor ensi field is longer than the upper." Id. at 95.
There are two types of Salor ensis, distinguished principally by the border scheme and by the drawing of the bird devices in the upper elem panel. In "type A" ensis, the inner and middle borders both consist of curled leaf motifs that are coordinated "so that they extrapolate into a hexagonal lattice closely related to that which occurs . . . on . . kapunuks . . . and some other trappings." Id. at 88. In type B ensis, the middle border is clearly the main border and features octagonal cartouches containing four stars analogous to the motif often found in the main border of older Tekke main carpets. In addition, the bird devices found in the upper elem differ in type A and type B Salor ensis. Both type A and type B ensis contain the same bird devices in the bottom elem and curled leaves enclosed within rectangles in the horizontal panel transecting the field. See Jon Thompson, Oriental Carpets (1993)(first published as "Carpet Magic"), p. 31 (type B ensi).
Salor pile weavings generally have wool warps, 2ZS, ivory, browns or a brown- gray combination. Alternative warps are significantly depressed and the backs are consequently ribbed and the handle less than floppy. Wefts consist of two shoots of wool, Z2S, and are brown, red, brown-gray. Two shoots. One weft is taut, the other sinuous. The most usual knot is asymmetric open to the left, although some latter pieces apparently have knots that are asymmetric open to the right. Knot density varies from 2,000 to 5,000 knots per square decimeter.
Pile is mostly of wool but silk decoration is frequent, sometimes lavish in latter pieces. Pile wool is usually 2ZS but 1Z, 3Z and 4Z have been noted. Silk 2Z, 3Z, 4Z, and 5Z have been reported.
Side finishes are often two two warp selvages, wrapped in dark blue or green wool or in two color checkerboard pattern. Multicolor braids occur. End finishes include ends cut and wrapped in dark blue or green wool; also red, white, or blue plain weave folded back and sewn; sometimes brown embroidered tape; dark blue fringes are attached.
Except for the coloring of the quarters of some guls, Salor weavings do not exhibit a diagonal use of color. According to O'Bannon, "Although the absence of diagonal color is not exclusively Salor -- the Saryks did not always use it and Choudor weavings frequently have a stronger vertical and horizontal color usage than diagonal -- the absence of diagonal color usage is something to keep in mind when distinguishing Salor pieces from Tekke." V. G. Moshkova, Carpets of the People of Central Asia 186 (G. O'Bannon & O. K. Amanova-Olsen, eds. 1996)(O'Bannon's commentary).
Some analysts count as many as 14 colors in Salor weavings. Ten may be more typical. Three and four shades of red, including an occasional corrosive red, and of blue are often claimed. Other colors include: shades of green, brown, orange, yellow and ivory. Most of the pieces designated as Salor are estimated to have been woven before synthetic dyes were in widespread use. Both lac and cochineal dyes are reported, especially in silk decorations.
Most weavings attributed to the Salors are estimated to have been woven in the first half of the 19th century or in the 18th century. Indicators of age include the "roundest" of ornaments (said to result from knots that are nearly as tall as they are wide), the spaciousness of the drawing, larger but fewer guls, simpler, seemingly archaic forms of some ornaments, and a color palette that includes some distinctive clear reds. Thompson suggests that the earliest Turkoman rugs may have had large, closely packed ornaments, followed by a period of spacious drawing, after which the space is filled up with ornaments and borders. Turkmen: Tribal Carpets & Traditions 98, 67, 69 (L. Mackie & J. Thompson, eds. 1980).
Some writers believe that the Salors largely stopped weaving after their military defeat in the mid-19th century. If one accepts this conclusion, then the the presence of a synthetic dye would exclude a Salor attribution. As indicated above, Murray Eiland remains skeptical of the conclusion that Salor weaving stopped with the tribe's military defeat in the mid-nineteenth century. See his recent exchange with R. John Howe. He believes that there is evidence that at least some Salor weaving continued long after that date.