Illustration 1. A Saryq main carpet, Lot 123.

A review of the Turkoman literature shows a sparceness of information on Saryq weavings. In Moshkova's monumental Carpets of the Peoples of Central Asia, one of her weakest chapters is on the Saryq. Tsareva's Rugs and Carpets from Central Asia has no Saryq main carpets. The Russian collections are the basis for much of the research on Turkoman rugs; if Saryq carpets are weak in those holdings, this may explain why they are not better represented in published material.

The purpose of this article is to present the results of a survey of published examples of Saryq main carpets, to sharpen the distinction between two groups of Saryq main carpets, to present characteristics of the earliest known examples, and to reconsider the present tribal attributions of some of these.

Thirty-five Saryq main carpets were identified in a recent literature search. Visual data from color and black/white photographs and published data were recorded. Cited publications are listed below. The main focus is Lots 123 and 133 from Sotheby's May 30, 1987, sale.

The characteristics of Saryq weavings as described by Jon Thompson are: ivory wool warp; grey, brown or ivory wefts, two shots and slight warp depression; Turkish knot; wool pile with use of cotton and silk highlights; flat selvedge normally in a two-colored check; and long plain kelim end finishes.

Reds vary from clear through red-brown, purple-red, to dark brown or deep purple. Dark blue, mid-blue and bluegreen are standard. Small amounts of yellow are found. Brown undyed natural wool is used for outlining. A varying secondary color is an apricot to orange-tan. White is from natural wool or cotton or both.

Age distinctions in Saryq main carpets are based on a combination of several characteristics:

Color. Purple, purple-brown and chocolate brown are considered late, e. g. last quarter of the 19th and early 20th centuries. Generally, the browner the color the later the rug. Rugs from mid-19th century and earlier have a definite red field with a brownish cast and sometimes a second bright red.

Cotton and Silk. The oldest rugs are believed to have neither. Silk and, occasionally, cotton are found in mid-century weavings. Their use increases dramatlcally late in the century and in early 20th century rugs. Neither are common in mid-20th century Saryq rugs from Afghanistan.

Orange. An orange shade is used in the quarters of the guls. In early rugs the oldest have a color near apricot and in later ones the orange has a browner tone. In many late rugs it is a synthetic red.

Sariq Border Patterns: A generally accepted principle of Turkoman rugs is that the oldest weavings have the simplest borders, and youth is indicated by increased number of borders and design complexity. One sees this clearly in Saryq carpets. The oldest have one main border and two flanking guard borders. In the late purple-brown rugs, there may be two main borders and as many as seven minor ones.

Saryq border patterns are fairly standard and limited in variety. The most common main border is the "naldag" pattern (Illustration 1). It is the main border design on late rugs. A second early design is a cartouche with a cruciform element in the center (Illustration 4). The secondary borders consist mainly of checkered triangles, diamonds, and tegbent patterns. There are a few exceptions.

Illustration 3. Photo courtesy of Sotheby's

Illustration 4. Photo courtesy of Sotheby's

1. Gul-i-Gul (Illustration 3)
2. Temirchin Gul (Illustration 4)
3. Saryq Gul
a. Square in quarter of gul (young) (Figure 1)
b. Zoomorphic trefoil in quarter of gul (oldest)
(Figure 2).

Left Figure 1 Right Figure 2

Illustration 2. A Sariq main carpet, lot 133.

This small group of rugs is best known by the McMullan fragment, Islamic Carpets, Plate 127, and in Thompson, Plate 17. Three other examples were identified. The technical data are fragmentary. Sizes are published on only two of the pieces. Two -- McMullan and Sotheby's Lot 75, May 18, 1985 -- may have white cotton. A Jones rug and Sotheby's Lot 69, October 30, 1982, do not. In two the cluster of three trefoils emanate from one point in the S-group manner (Illustration 5) and in two they emanate in a two/one Ersari manner (Illustration 3). Lot 69 has a Saryq type Juval secondary gul, but the Juval secondary gul of the Jones rug is of the S-group type. The McMullan and Lot 75 rugs have a kurbaghe secondary gul variant.

Illustration 5. A Sariq main carpet, courtesy of Sotheby's.

Eight Temirchin gul carpets were identified. Although only twice the number of Gul-i-Gul carpets, they show greater consistency of characteristics. Only one piece was fragmentary, Thompson, Plate 16. One is late, Edelmann, Lot 121, March 10, 1979. It has a large quantity of silk, is the second largest of the 37 surveyed carpets (8'11"x7'11"), has the greatest number of guls (5x12), and the most number of borders (7). It was correctly cataloged as early 20th century.

Of the remaining seven pieces, all are probably first half of the l9th century. Six are very squarish (8'1"x7'10" to 7'8"x7'6"). Guls vary from 4x8 to 4x12. Secondary guls are the Juval gul or the memling gul.

Sariq Gul Carpets

Sotheby's Lots 123 and 133 are now presented in detail.

Technical Data

Lot 133

Warp: Ivory wool, Z2S
Weft: Brown wool, Z2S, two shots
Knots: 15 vertical, 9 horizontal, 135 per square inch, pile thread of three strands
Colors: BRICK RED, dark blue, blue-green, orange, ivory, dark brown, lilac (silk)
Size: 8'9"x7'10"
Field: Guls 5x11; 8 1/2"x11"; minor gul, Juval type Border: 8 l/2" wide; major border, Naldag; minor border, Triangle (2), Tegbent (4)
Selvedge: Red and blue checkered, six bundles of two warps, grouped in threes by color; inner group of two warps wrapped in blue-green
Kelim: 2" remaining, originally 10"

Lot 123

Warp: Ivory wool, Z2S
Weft: Brown wool, Z2S, two shots
Knots: 12 vertical, 9 horizontal, 108 per square inch, pile thread of two strands
Colors: RUBY RED, dark blue, green, orange, yellow, white, brown, lilac and blue silk
Size: 9'1"x7'11"
Field: Guls 5x11; 9"x11 1/2"; minor gul, Chemche gul
Border: 11" wide; major border, Naldag; minor border, Triangle (2)
Selvedge: red, blue and blue-green checkered, four bundles of five warps, grouped in twos by color
Kelim: Missing

They are similar in several respects: size, 5x11 guls, some silk, no cotton, checkered selvedges, and brown wefts.

They differ in the following respects. Lot 123 is ruby red, while Lot l33 is rust red. Lot l33 has the Juval secondary gul, but Lot 123 has the chemche gul. Lot 133 has six minor borders, while Lot 123 has two. Lot 123 has a clear yellow while Lot 133 does not. The orange of Lot 123 is pinkish, while that in Lot 133 is brownish. Lot 133 has a blue-green; the comparable color in Lot 123 is decidedly green. Lot 133 is similar to the McCoy Jones rug published in Schürmann, Plate 10; Lot 123 is closer in color to the Meyer-Muller Ersari in Schürmann, Plate 57, one of the earliest known Turkoman rugs.

In the survey, 23 carpets with the Saryq gul were identified. Of these, 15 were considered to belong to the third quarter of the l9th century or earlier, e.g. without the purplish coloration and square in the gul. Including Lots 123 and 133, there are a total of 25 Saryq gul carpets. Their characteristics are as follows, with exceptions noted:

Left Figure 3 Right Figure 4

Minor Gul. In all instances but one, the minor gul is the juval gul. The exception is Lot 123 with the chemche gul.

Major Border. The major border is the naldag. An exception is the famous Wher carpet, where the border is of the cartouche type found on S-group and Saryq Temirchin gul carpets. Several late carpets had two naldag borders flanked by numerous minor borders.

Minor Borders. There are two arrangements of minor borders on early carpets. A few rugs have the main border flanked by two. The most common pattern is multicolored triangles. Less frequent is a diamond pattern. Singular instances of small florets, stars and squares were noted.

When more than two minor borders are used, four narrower borders with the tegbent pattern are used (Illustration 3).

Edges. Most edges were a blue and red checked selvedge. Three were a solid brown wool selvedge. One late piece had a blue wool overcast. The selvedge of Lot 123 is unique.

Cotton and Silk. Of the early pieces, four had silk only (Thacher, Sotheby's Lot 81, Lot 123 and 133) and one cotton only (Franses). Of the late pieces, three had both cotton and silk and one cotton only. These data were recorded on only 11 pieces and are not conclusive but do conform to characteristics cited by most authors.

Size. Of the early carpets, the smallest was 7'x7' (Thacher), which coordinates with one of the smallest number of guls, 4x9. The largest is Lot 123 at 9'1"x711". In general the oldest were squarish in dimension. Later pieces were more rectangular, 10'3"x7' 11", 7'3"x4'8".

Knots Per Square Inch. Densities recorded were 108 (Lot 123), 112 (Grote-Hasenbalg, Plate 85, est.), 162 (Jones/Schürmann), 200-250 (Thacher but probably 125), 134 (Herrmann VI), 135 (Lot 133), and 144 (Franses).

In evaluating Lots 123 and 133 against this background, Lot 123 may be one of the oldest Saryq main carpets, circa 1800, and not the last quarter of the 19th century, as cataloged. The reasons for this conclusion are that it features:

1. a major border with only two minor borders;
2. a strong green color coupled with a clear yellow;
3. a clear ruby red similar to the reds of the earliest Turkoman rugs;
4. large, spacious guls;
5. a low knot density for this group of rugs;
6. a random placement of white in the border patterns; and,
7. several random, unselfconscious elements of design, color, and use of silk.
The last relate to aspects which are not immediately obvious and affect the aesthetic presence of the rug.

On the beginning end of the carpet, the guls and borders have idiosyncratic pattern changes. The center motifs of the first three rows of guls are different. These differences carry through the border. The chemche gul is not spacially balanced until the third repetition.

Silk detailing is present mainly in the vertical and horizontal arms of the chemche gul. The arms have every color in the rug and are a kaleidoscope of prismatic colors. It is randomly used in the guls and borders. Lilac and bluegreen silk are used initially but in the upper half only lilac occurs.

A unique feature is seen in the main border. The "X" motif of the simply drawn naldag is cut horizontally by narrow stripes with varying patterns. This horizontal patterning in the main borders is common in Tekke carpets but occurs nowhere else in Saryq main carpets. These break the verticality of the naldag pattern and create square segments comparable to those in the cartouche borders.

In the center of the gul are two features which animate the static geometry of Turkoman patterns. These features affect one's subconscious response to the carpet.

The first feature is in the tendrils surrounding and defining the center of the gul. The usual drawing and coloration of the tendrils are illustrated in Figure 4. In Lot 123 the middle, angled segment of the tendril is wider than usual giving it an angle or articulation like an elbow (Figure 2). This articulation compliments the adjacent segment of color in the quarter of the gul and moves the eye in an arc around the center of the gul. In the bottom half of the rug, the color in these two areas is the same and the movement of the eye is dynamic. This effect is lessened when they are different colors in the upper half.

The second feature also results from the widening of the tendril segment. It redefines the center space of the gul (Figure 2). In all other carpets the shape is as in Figure 4. Figure 2 is a more refined outline than Figure 4. It is usually these minor changes in pattern which separate the great from the ordinary.

A final consideration supporting an early dating of the rug is the use of the chemche gul. Antique Oriental Carpets from Austrian Collections, Plate 101, illustrates a Salor (S-group) rug with the chemche gul. This rug "probably belongs to the earliest surviving Salor weavings," according to the authors. Having seen the carpet, I believe that conclusion is justified. A Saryq main carpet fragment published by Hoffmeister, Plate 60, is the only other known Saryq example with the chemche gul.

Although the chemche gul is thought of primarily as the property of the Tekkes, it was also used by the Kizil Ayaks, Yomuds, Salors, Ersaris, and apparently Saryqs on their oldest carpets and, in earlier times, was a more widely distributed tribal main carpet pattern.

Lot 133 belongs to the main body of early Saryq carpets of which the Jones and Herrmann pieces are prime examples.

When one contrasts Lot 133 and others with Lot 123, they seem selfconscious in their regularity and absence of idiosyncracies. There are no stripes across the "X" of the naldag. The silk is used exclusively in one or two pattern elements, such as the small tegbent motifs in the center of the juval guls. There are rarely pattern adjustments in the beginning end. The use of white in the main and secondary borders is unvarying. The whole impression is one of regularity and control, not individuality.

The Wher carpet presents interesting evolutionary questions from a pattern and aesthetic standpoint. The first row of guls has the standard cruciform motif and bird trefoil. The latter is replaced in the remaining guls by a rhombus, which is also found in the principal segment of the tendril motif. It seems not to be a zoomorphic element like the trefoil which it replaces. A widening of the middle segment of the tendril in the second and third rows of guls is noted. The chuval guls are crowded in the first row and are accompanied by two unusual filler motifs.

The "star within a box" in the center of the gul is unique as is the main cartouche border. The use of six minor borders qualifies the use of the word earliest for this carpet, despite these other features.

Most of these factors support including it within the earliest group but, comparing color and originality, not as early as Lot 123.

Conclusions

The early carpets have a redder coloration, larger guls, squarer dimensions, and some use of silk. The absence of cotton indicates an earlier date. Silk would appear to have been used early by Saryq weavers. Relying solely on number of guls or minor border stripes to validate age is risky; e.g. Lot 123 has 55, the Wher carpet 32, Lot 38 in Sotheby's London sale of June 15, 1983, had only two minor borders but had 65 guls coupled with the regularity of pattern elements of Lot 133.

Eleven vertical guls is the number most frequently used in the early carpets. There was one rug with eight, one with nine, two with 10, one with 12, and three with 13 guls vertically. Horizontally, there were seven each of four and five guls and two with six guls.

The relationship of Saryq carpets to the other Turkish knotted Turkoman carpets is not clear from this study. It does seem that the naldag pattern is uniquely Saryq. The fact that half of the Temirchin gul carpets have the naldag pattern indicates a strong relationship between the two groups.

The relationship to the Gul-i-Gul carpets is less clear. The four examples included in this survey point in a variety of directions.

The McMullan fragment and Sotheby's Lot 75, May 18, 1985, with kurbaghe guls point towards Ersari weaving based on the shape of the guls, interior motifs, placement of the trefoil patterns and filler design elements. The cartouche main border design is found not only on Saryq carpets but on the Saltiq Ersaris and others as well. See Plate 120 in Austrian Collections. The two other rugs pose questions in other directions. Lot 69, Sotheby's October 30, 1982, would appear to be Saryq. It has the naldag border and secondary border patterns common to Saryq carpets. The secondary gul and its interior drawing are Saryq. Except for the cruciform elements in the center of the gul -- a Saryq pattern -- the major gul is of the S-group type. The color is more S-group than Saryq.

The Jones carpet was published in color in HALI V, 3. This carpet is even closer to the S-group in style. The secondary gul is like that found on S-group rugs. The borders are unlike Saryq patterns. The motif in the center of the major guls is an "X", not a cruciform. And the red is more akin to S-group color than Saryq.

There are two implied assumptions which present problems in assessing these rugs. The first is: if they are Turkish knotted, they are Saryq. We know from past and current production that the Yomud Turkoman tribal groups used both the Turkish and Persian knots. The Saryq Turkomans in Afghanistan today weave rugs with both types of knots. They probably did so in the past. Because the Ersaris of today do not use the Turkish knot does not mean that some group in the past did not use it. Because the Salor Turkomans no longer exist, we know nothing for certain about what variety of knots they used.

The second assumption is: if the rugs have cotton in the pile, they must be Saryq. It is well documented that cotton appears commonly in Yomud tentbands, Tekke ak chuvals, and in the wefts of Chodor and Arabachi weavings. It seems obvious that cotton was available to many Turkoman tribes to be used in many ways. There is reason to assume that cotton may have been used in the pile by several different groups.

The Gul-i-Gul and Temirchin gul carpets may or may not be Saryq. More examples will have to be studied before a firm attribution can be made.

The two rugs presented here help to define some characteristics of the earliest Saryq carpets (Lot 123) and those that follow it in the l9th century (Lot 133). These were succeeded by the more numerous purple-brown rugs from the last quarter of the century. They show the use of a common gul with minor changes in drawing and color over time.

Edelmann, Lot 121, March 10, 1979

Franses, The World of Rugs, Plate 38

Grote-Hasenbalg, Masterpieces of Oriental Rugs, I925, Plates 85 & 94

HALI V, 3; II, 4

Hoffmeister, Turkoman Carpets in Franconia

Mackie and Thompson, Türkmen, 1980, Plates 16 & 18

McMullan, Islamic Carpets, 1965, Piate 127

Moshkova, Carpets of the Peoples of Central Asia, 1970, English translation, Oriental Rug Review, Vol. III, No. 1 - Vol. IV, No. 9.

Schürmann, Central Asian Rugs, 1969, Plate 10

Sotheby's, Lot 11, March 19, 1980; Lot 69, October 30, 1982; Lot 75, May 18, 1985; Lot 40, May 30, 1987

Sotheby's/London, Lot 38, January 15, 1983

Thacher, Turkoman Rugs, 1940, Plate 6

Despite this complex, age-dependent dyeing diversity, little is known specifically about the underlying dye and chemical technology for Saryq color use. It is assumed that such dye use shares much with the rest of the Turkoman dyeing tradition.1

As part of this overall study, a selective examination of the dyes in certain areas of these two old Saryq main carpets has been carried out by the author. These studies are the first systematic look we are aware of at Saryq pieces in terms of dye analysis.

General

The time frame for these studies did not permit a detailed look at either elemental composition or metal dye mordants in the wool nor were the sample sizes sufficient.

The methods used for the identification of the Saryk dye palette are already detailed in earlier issues of this journal.3, 5 (or see footnotes 6-8, p. 14, this issue).

Dye Testing Results

For both carpets, the main red consists of a madder which has a proportionately sizable fraction of purpurin as well as the typical alizarin as madder components. A trace amount of pseudopurpurin is also present. No other madder components were observed as minor constituents.

The dark blue-black color areas consist of indigo applied to grey wool, while the green area of both main carpets has indigo as the blue component.

The yellow color in the one sample provided is a dull yellowish gold and consists of a number of components, two major and three minor. In technical terms, the two main constituents are highly "polar" substances. Overall, the components are consistent with isparyk or the Turkoman species of larkspur, Delphinium sulphurium. The yellow fraction of the green fibers in both Saryq main carpets appears to be this same rather flat golden yellow.

The orange color in samples from both main carpets is derived from the madder giving rise to the reds indicated and a yellow component which appears to be the same as that used for yellow as such, isparyk.

The dark brown in the one sample provided is dyed rather than natural brown. It consists mainly of a tannin-based dye and a small amount of purpurin -- and pseudopurpurin -- rich madder.

The sole silk sample provided, of a rich crimson hue, was found to contain carminic acid as a major component, indicative of a cochineal.

Commentary

The red in the two pieces contains madder with basically the same composition, that is, madder with a rather high amount of purpurin as well as alizarin. It can be surmised that with some adjusting of alum mordant, use of such madder accounts for the typical "look" of older Saryq work, i.e., belonging to the middle period. Madder with, mainly alizarin, purpurin and pseudopurpurin are commonly encountered in Turkoman pieces and this author has noted this in various reports.

Blue and the blue component of green derive from indigo, and this is not surprising, given its universal use in Turkoman work. An important observation is the use of darker grey wool for dyeing with indigo. This would account for the blue black appearance.

The yellow color posed some challenges in analyses, as such colors often do. The multi-component nature of the yellow dye, its appearance and comparison with reference dyes indicate that it is a flavenol, the commonly used isparyk.

The green colors from these main carpets contain no surprises, as they consist of the isparyk and indigo.

The orange shades from both main carpets consist of a mixing of madder and a yellow, rather than the more often employed selective madder and mordant mixing. Furthermore, the madder is of the same composition as that used for red, rather than madder being used for orange shades. Secondly, the choice of the yellow as isparyk, an unstable flavenol, yields a gradual deterioration in the purity of the orange shade with time as the yellow component fades. This is analogous to what happens with green over time, when the yellow portion fades.

Use of madder with little variety and isparyk, rather than later and stable weld (luteolin), suggests that the Saryq dyers had limited access to dye sources or could not readily afford them. The use of isparyk suggests some age to the pieces but does not prove great age.

The dark brown is a tannin-dye and madder. The use of madder high in purpurin but small in proportion to tannin has been seen by the author in analyses of Yomud and other work. Since the level of iron as mordant was not tested, the presence of low mordanting to give a more lively brown than seen with tannin dyeing was not examined.

Are these characteristics of Saryq dye use, in the admittedly few cases tested by the author, different from those of other Turkoman tribal groups? The use of madder with little variation in composition, the use of purpurin-rich madder plus an unstable yellow for orange shades has not been commonly encountered by this writer. The producing of orange shades, per se, from red plus yellow in Turkoman work, on the other hand, is not rare. The producing of blue-black shades by indigo on dark wool (brown or grey) have been encountered by this author in Kurdish work6 and in a recently described rare Yomud asmalyk.7

Overall, these main carpets appear to be attributable to the middle period of Saryk production and the overall red effect springs from the dye composition described above. The mellowed oranges clearly relate to the use of a fading yellow and a madder inclining more to deep reds to begin with. The blue-black effect is indigo on dark wool. The dark brown color has enough added madder to the tannin-based underpinning that a better effect is probably gotten than if one used the combinations seen in Yomud work.

Further conclusions about Saryq dyeing traditions and how they fit in the mainstream of Central Asian rug and trapping production must await further research into dyes and mordants in Saryq examples. The genesis of the "clear reds" of the rarest, oldest Saryq work is of great interest. Of interest, also, are the purple and purple-brown shades. Are they overlaps of the Yomud dyeing practices?

Notes

2. Whiting, M.C., "The Dyes in Turkmen Carpets" In: Türkmen, (L. W. Mackie and J. Thompson, eds.), The Textile Museum, Washington, D.C., pp. 217-224.

3. Mushak, P. and O'Bannon, G.W.,"Chemical Analysis of a Turkoman Chuval", Oriental Rug Review, Vol.Il, (10): 6-8, 1982.

4. Mushak, P., "A Technical Dimension to Color Esthetics in Old Armenian Weavings: Chemical Analysis of an Antique Karabagh Long Rug,", Oriental Rug Review, Vol. III , (4): 3-5, I983.

5. Mushak, P., "An Early Resht Embroidery", Oriental Rug Review', Vol. IV (7):8-11, I984.

6. Opie, J., and Mushak, P., "Stylistic and Chemical Dye Analysis of Some Rugs and Trappings of the Luri. Part III", Oriental Rug Review, Vol VI, (7): 7-9, I986.

7. Mushak, P., "A Rare Yomud Asmalyk in `Jewelry' Design: Stylistic and Chemical Analysis",Oriental Rug Review, See this issue.