by George O'Bannon
(Photo Captions by Tom Cole)

This article appeared in Oriental Rug Review, December/January1989, Vol. 9, No. 2, Baluch Focus

This is a fourth and final article in a series considering Baluch rugs which I saw and studied in Afghanistan in the 1970s. All of the other articles (see author's bibliography entries on Taimani, Aksi, and Mushwani rugs) were about groups of rugs which were woven in the mid 20th century. This article is about two distinctive design types which derive from a group of Baluch rugs called Timuri in scholarly, collector and trade circles which are generally attributed to the last half of the 19th century. These two types are Yacub Khani and Dokhtar-i-Ghazi Baluch rugs. With respect to the name Timuri, I never once heard it used by rug dealers in Afghanistan. Recent books and articles on the enthnographic make-up of the Afghanistan population do not refer to Timuris as a constituent group in either the Baluch or Aimaq peoples. To my knowledge the term appears in the enthnographic writings of the 19th century and, where it is used in 20th century sources, the word Timuri is always derived from one of these earlier tabulations of ethnic groups. Bogolubov in his Carpets of Central Asia illustrates and attributes two rugs to the Timuri. This would seem to have stimulated the use of the term in rug and auction literature in recent years.

Figure 1. A main carpet from Afghanistan, that most likely comes from the region of Chakhansur, similar in patterning and colour to the main rug that now graces the walls of the DeYoung Museum, SF Such rugs are not seen in NW Afghanistan where most of the Afghan based Taimuri related weaving groups reside -TC. See Raising the Bar

I am of the opinion that the Timuri ceased to exist as a separate group sometime around the turn of the century and that they have coalesced with units of the Aimaqs or other settled populations in northwest Afghanistan.

The names that I have used in all of these articles are those which were common in the rug trade in Afghanistan in the 1960s and 1970s. They were used in provincial rug centers, such as Herat, Qandahar, Mazar-i-Sharif, Maimana, and Kunduz, as well as in Kabul. My use of these names reflects validation of this usage in several areas for rugs with similar weave and design characteristics.

To say that some of the rugs I am about to describe are not Timuri is not to deny the existence of Timuri Baluch rugs; it is rather to affirm a belief that the weavers of the rugs we call Baluch are a dynamic and constantly changing group of weavers. These weavers existed in the most marginal of ecological niches and were constantly forming new political, social and economic units for survival. I believe it is impossible to consistently ascribe, over a 100 year period for instance, a type of Baluch rug to one specific group as we can, for example, Tekke Turkoman rugs.

Figure 2. The so-called kalam danni design, literally translated means pen holder. This rug is definitely Afghan, and the pattern is unknown among the Persian Baluch weaving groups. -TC

The Yacub Khani and Dokhtar-i-Ghazi rugs may reflect one of the longest traditions that exist in Baluch rugs. These two terms are at some time in the late 19th and early 20th centuries probably applied to specific groups of weavers who were known tribally by these names. By mid-century, I suspect that these tribes no longer survived as separate units, but the designs were adopted by other groups and have survived. Hence, when I use these terms, I am referring primarily to a group of rugs with certain designs, technical characteristics, and wool quality rather than a group of rugs woven by a specific group of Baluch weavers. The technical characteristics of both of these types are not unusual for Baluch rugs. In general, they have white wool warps, natural brown wool wefts, Persian knots open to the left with an average knot density of 50 to 90 per square inch, and a lustrous but soft wool. The oldest pieces tended to have wide kilim ends, which vary from simple stripes to patterning in both plain, slit, and dovetail tapestry and various brocade techniques. The edges may be a simple one-colored selvedge to multicolored wrapping. Depressed warps rarely occur. Sizes vary from small prayer rugs to floor carpets up to 7x10, which may be woven as one piece or two matching pieces and sewn together. The rugs come from the area north and west of Herat to the Iranian border.

Figure 3. Another Afghan Baluch rug, but with an interesting and unusual secondary border located between the two meandering vine borders. -TC

Although I do not know the Iranian market well, on the rare occasions when I encountered these rugs in the Tehran bazaar, the dealers did not consider that they were from Iran. This confirms for me that Afghanistan or adjacent areas of Central Asia or both are the origin of these rugs.One can see how these two types have evolved from the Timuri rugs, how they have changed over time, and how they are being woven today from the wealth of extant antique rugs down to new examples woven in recent years. Figure1 is a rug of the so-called Timuri type. It is probably from the third quarter of the 19th century, and I believe all the dyes are vegetal. For a Baluch rug, its palette is broad (two reds, two blues, blue-green, white, brown, and dark aubergine). The field consists of large palmettes surrounded by a veritable sampler of Baluch patterns: plant forms, small animals, geometric figures, Ss, botehs, small rosettes, and the "duck" or peacock common to a group of donkeybags and rugs. The palmettes alternate two and three per horizontal row in an offset manner.

Figure 4. A superlative example of Taimuri group weaving. Whether this piece isfrom Afghanistan or Persia is unclear to me; the design is extraordinary and the colours appear to be fantastic. O'Bannon notes the rug was offered at Skinners and did not sell. So much for the acumen of the Baluch buyers of that period, late 1989 when one would have thought they would have known better -TC

To observe the changes and alteration of patterns over time, one should note some minor details of this rug. In the center of the palmette is a footed vase-shaped design containing a tree. Note to the left the Tree of Life which occurs on many Baluch prayer rugs and the small white stick-figure animal. Above the Tree of Life is a geometric figure consisting of a kotshak topped by a triangle. This small pattern will occur in later rugs.

Starting from the rug's outermost border and working toward the field, border designs are: a gyak stripe, then a tick-tack-toe meander, gyak, the main border with a floral quincunx and
various small patterns including small botehs between major elements, then gyak, and the leaf and tegbent meander so characteristic of these rugs, a reciprocal step pattern, a tagbent, and reciprocal V. In the field note the kotshak pattern, which functions more as an edging of the field than a border.
The reason for focusing on these small designs is to show how they persist over time, how a design can change easily with slightly different drawing, and to emphasize how a simple change in color alters one's perception of a design. The altered design, however, still shows these elements relating back to its antecedents.

Figure 5. A late example of Afghan Baluch weaving with a singularly uninteresting field design and multiple border systems. -TC

In Figure 2, if one focuses on the light area of the complete pattern element on the right, the palmette we see in the Timuri rugs is obvious. If one focuses on the light area on the left, it is the footed vase which stands out as the principal pattern. Both of these elements can be seen in the Timuri rug, but it is the palmette which is the paramount and intended visual pattern. The same is not true of this rug.

This is an example of one of the two main patterns of Yacub Khani rugs. The footed vase design is referred to as Qalem Dani, or pen holder. Because of the protruding leaf forms, Westerners think of an upright holder for a quill or pen. The
weavers more likely had in mind the Persian type of long, ovate, papier-mâché or wooden pencil box, richly ornamented with lacquer painting. It graphically shows how a pattern is simplified and then a specific element is extracted to become a major design element.

This particular rug contains a rare date panel containing the number 1330, or 1911. Specific age references for Baluch rugs are difficult to come by, but this one aids in interpreting the evolution of these specific patterns. The inscription has not been deciphered. Note the kilim ends with plain and slit tapestry and the multicolor, wrapped selvedge.

Figure 6. A late and ultimately uninteresting example of late BAluch weaving from Afghanistan, reportedly made in 1970.-TC

Figure 3 is a variation of the Qalem Dani pattern. Note that the rug has a 2-3-2 usage which was noted in Illustration 1. At each end of the field are three elements labeled Gul-i-Khaf by Wegner and rare in these rugs. Bogolubov's "Turkoman Line" is used in two borders, the outer fully drawn and the inner one a scaled down version but perfectly balanced and harmonious. The middle border with white Xs shows its Saryq origins but, in later rugs, this border becomes more static and less detailed.

Figure 4 shows the earliest form of the Yacub Khani Baluch type which some people label Timuri. The field design consists of rectangles with serrated ends. This form can be seen at the center of the palmette in some Timuri rugs. It has been extracted from that pattern to become the primary "gul" in these rugs. The center of the rectangles contains some of the geometric figures noted earlier in Figure 1. The palette is more subdued; more brown is used, but a light apricot color has been added. The
rectangles are arranged in horizontal rows of three and four in an offset manner. The field is covered with varied patterns. The excellent quality. inner border is a finely drawn geometric meander. The main border is a loosely drawn cartouche in blues alternating with an X outlined in brown. The outer border is a series of bracketed diamonds which is used extensively in later rugs.

In later rugs of the serrate rectangle type, the ends of the rectangles are first blunted and later the projections at the ends and edges are lost. The alternation of odd and even numbers of rectangles per row is dropped in favor of the same number of units in a row. By the middle of the 20th century the form has evolved into the type of rug in Figure 5. Here there is an even number of rectangles per row, the serrated ends have been eliminated, and the patterning of the spaces between has become uniform. There are no random geometric and animal figures in these spaces. The rugs of this type also have a limited palette of red, blue and brown. The wool in most cases is of

Figure 7. A seemingly fine and nicely proportioned example of a Dokhtor-i-Qazi prayer rug.-TC

Although the serrate rectangles and the Qalem Dani patterns predominate, other designs were woven by these weavers. One pattern is an allover repeat of the small Tree of Life pattern. Some people see this pattern as a version of the Yomud curled leaf pattern, which I feel is erroneous.
Published color examples of rugs I believe to belong to this group but which I have not seen personally are: Azadi, Plates 9, 40 and 41; Black, Plates 5, 27, 28 and 38; and Homer, Plate 8. See the Baluch bibliography for full citations of these publications.

Although Yacub Khani did not weave many prayer rugs, they are not unknown. The Tree of Life pattern in Figure 1 was used in them more frequently than others, perhaps because of the smaller format area. In post-1950s rugs, the pattern seen in Illustration 6 was the most common on Yacub Khani prayer rugs. (See Craycraft, No. 2, and Black, Plate 11).
There are a few donkeybags which can be attributed to the Yacub Khani with some certainty. The triangular plant form visible in Illustration 1 as a filler pattern was frequently used on these weavings. See Craycraft's No. 30 for an example of this design in prayer format, and Homer's Plate 24 for a complete donkeybag of the Dokhtar-i-Ghazi type.

Concurrently with this evolution of the traditional patterns to simpler ones and a narrower color range, some new patterns emerged in the 1960s which were also called Yacub Khani. Figure 6 is one of those. The field is covered with an allover Zili Sultan design. It was not uncommon to frequently find two square panels with aou, a native antelope, frolicking in a tree-filled landscape. An inscription at the top and bottom of the field says "Made in Afghanistan 1970," another rare dating reference for these rugs. A totally new, largely synthetic, color palette of orange, green, and purple accompanied this evolution and is found in other Yacub Khani

Plate 15, David Black & Clive Loveless, Rugs of the Wandering Baluch.

The Yacub Khani were mainly weavers of floor rugs, with very few prayer rugs or other utilitarian pieces made. By contrast the second group of rugs are predominantly prayer rugs; these are called Dokhtar-i-Ghazi.

They are distinctive in the drawing of the prayer niche. In most Baluch prayer rugs, the mihrab is squared off on three sides. In the Dokhtar-i-Ghazi prayer rugs, there is usually a second mihrab within the square one of a head-and-shoulders type or with six sides. (See "Baluch at Auction," pg. 26.)

The second distinctive feature is the small allover repeat design of a small arrow-shaped "bush" on a blue field. This "bush" may be found as a filler pattern in Figure 1. The palette of most of these rugs is very restrained. In the oldest examples the pattern of the mihrab spandrels continues around the top side of the rug. See Craycraft's No. 14 for an excellent specimen and example of this use. His Nos. 13-16 show the many variations which occur in this type. Also see Black, Plate 10, and Homer, Plates 15 and 16.

Plate 15, Homer, Exclusively Baluch.

Although most of these rugs fall within the knot density mentioned above, some of the finest knotted Baluch rugs have been noted in this group with knot densities over 150 per square inch. Some silk wefts have been noted in the finest of these. Although the palette is generally quite consistent, there are pieces with a pastel palette, a result of the use of muavine or early aniline dyes, resulting in abrash fading to pale gold and gray tones.

Seemingly until the 1950s, rugs of this type continued to show a very strong relationship to the Timuri rugs in the border patterns: leaf meander, trefoil, and the inner border of terraced Vs. The
frequently used botehs in the mihrab spandrels are drawn much as they appeared in the Timuri rugs. See Black, Plate 10.

Craycraft's Nos. 16 and 13 are good examples of two changes which occur in these rugs. No. 16 shows the small drawing of the field pattern and typical borders, but it has the standard Baluch three-sided mihrab without the inner head-and-shoulders. In No. 13 the field pattern is drawn much larger and also with only the three-sided arch. These are features which I believe come later in the design evolution of these rugs.

Plate 16, Homer, Exclusively Baluch.

An examination of most of the published illustrations cited shows the continued use of many of the border patterns used in Illustration 1.

As with the Yacub Khani, this design lives on in new rugs of the post-1950s era. A typical example of this new production is found in Parsons (1983), Plate 98. The double mihrab is used and the field pattern is the same small version, but the palette is restricted to red, blue-black, and brown; the vibrant indigo blues found in the older pieces is missing. But these new rugs on the
average are much more finely woven than the old pieces and have excellent wools which should soften and develop a nice sheen with time and use.

It should perhaps be noted that the recent Yacub Khani and Dokhtar-i-Ghazi rugs frequently have a black-dyed, fine cotton weft. This appears to be used as early as the 1930s but became much more extensively generalized in many of the Baluch rugs of all types from Afghanistan during the 1960s. The source of this thread has not been determined.

Plate 16, Craycraft, Belouch Prayer Rugs.

IA system of nomenclature for Baluch rugs continues to be one of the challenges facing those of use who chose to write about them. I believe that the Yacub Khani and Dokhtar-i-Ghazi types illustrate some of the ways associations may be shown but also the problems which arise when considering Baluch rugs longitudinally. A nomenclature based on groupings by design and/or technical features is likely to prove much more workable than one based on tribal or place affiliations until on-site investigations show otherwise.

With thanks to Oriental Rug Review, and Ron O'Callaghan for permission to reproduce this article here.

<---previous article