CARPETS (qālī; Ar. and Pers. farš, heavy textiles used as coverings for floors, walls, and other large surfaces, as well as for various kinds of furnishing.

(See also under individual tribes and locations)

The history of Persian carpet manufacture is fraught with unproved hypotheses, rash assumptions, and confident assertions that flout historical evidence. Indeed, scholars hold widely differing opinions on almost every aspect of the production of Persian carpets. As Edward Gibbon said in another connection, “the melancholy duty is imposed upon the historian” of trying to determine what facts are beyond dispute (p. 505). They are few.

The discovery of the Pazyryk carpet and an even earlier fragment in tombs in Siberia revealed that the manufacture of knotted-pile carpets is of far greater antiquity than had previously been supposed (see vi, below). The oldest previously known pile carpets were manufactured in Anatolia in the 7-8th/13-14th centuries (Denny, p. 23). With the exception of several fragments from Sasanian levels at Shahr-e Qumis (see vi, below) and possibly of a fragmentary lion carpet recently purchased by the De Young Museum in San Francisco (see vii, below) no Persian pile carpet can definitely be said to survive from before the 10th/16th century (Hubel, p. 160). There is literary evidence for the existence of carpets in western Asia from pre-Islamic times (cf. Edwards, pp. 2-3; cf. vii, below), but it is impossible to say whether or not they were knotted. The famous jeweled garden carpet of Ḵosrow I at Ctesiphon was almost certainly not a pile carpet but a flat weave or an embroidery (see bahār-e kesrā); a pile carpet of the size described (ca. 27 m square) would probably have weighed more than two tons (cf. Edwards, p. 2).

According to Kurt Erdmann, Oriental carpets were not regularly imported into Europe before the end of the 7th/13th century (Survey of Persian Art, p. 3160). At about that time European artists began to depict such carpets in their paintings; Giotto (ca. 1266-1337) seems to have been the first (Sylvester, p. 4; cf. Erdmann, 1970, pp. 17-20, 47-51). Most of the carpets depicted are thought to be Turkish, however, not Persian; it was generally quicker, easier, and cheaper to import carpets into western Europe from Anatolia than from Persia as, especially during the Safavid period, the Ottoman empire, which controlled the traditional overland routes from Persia to the west, was periodically at war with the Persian state. Carpets of presumed Persian design are nevertheless to be found in the paintings of Mantegna (1435-1506), Van Dyck (1599-1641), and Rubens (1577-1640); in fact, in the 17th century some seventy Dutch artists included representations of Persian carpets in their paintings, mainly types originating in northwestern Persia (Mills, pp. 32-34; see ix, below). Oriental carpets and rugs purchased by Europeans were, of course, often too valuable to be laid on the floor, as had been the practice in the Orient since earliest times. Terms used in Venetian inventories indicate that Oriental rugs were placed on tables (tapedi da desco; tapedi da tavola) and on chests that also served as seats (tapedi da cassa; Erdmann, in Survey or Persian Art, p. 3161 ), and European paintings confirm such uses (e.g., L. Bassano, “Portrait of a Senator,” late 16th century; King and Sylvester, p. 18, fig. 13).

Beginning with the Safavid period large quantities of pile carpets survive, but the dating and attribution of these and later examples to particular weaving centers are pitfalls for the unwary (see ix, x, below). Some general observations are relevant, however. First, of an estimated total of between 1,500 and 2,000 surviving Safavid carpets only five are dated (Edwards, p. 7; Spuhler, p. 701). The number is greater from the Qajar and Pahlavi periods (see xi, xiii, below), yet, even when a carpet is dated, careful examination is necessary to ensure that the date has not been tampered with (Ittig, 1981, pp. 125-27). Carpet inscriptions can nevertheless provide invaluable information on craftsmen, places of manufacture, patrons, and so on; furthermore, once a particular carpet has been firmly fixed in time and place, it can be used to identify other structurally related pieces (Ittig, 1981, pp. 125-27; cf. ix, below). Arthur Upham Pope pointed out that parallel designs can frequently also be found in architectural decoration, which can be useful in dating and localizing carpets (Survey of Persian Art, p. 2268).

The attribution of a carpet made before the late 13th/19th century to a particular place of manufacture is even more hazardous than dating. Friedrich Spuhler declares bluntly that assumptions about the origins of Safavid carpets “which have gained credence over the years are purely hypothetical” (p. 704), and E. Gans- Ruedin concurs: “Too few written sources are at our disposal today to enable light to be thrown upon the making of any given carpet. For this reason, the method of attribution to a specific production center, used for more than half a century, with a few exceptions, is unreliable. Perhaps the time will come when documents will come to light which will allow us to ascertain the exact provenance of some rugs from the 16th and 17th centuries. Meanwhile, it is possible to formulate hypotheses only” (p. 28). In the few instances in which a master craftsman has woven his name into the carpet, it would be rash to assume that his nesba (attributive name) indicates the place where the carpet was manufactured. Yet only recently have scholars become more conservative about attributing specific carpets to specific manufacturing centers (see ix, x, below).

On the other hand, the location of the principal centers of carpet production during the Safavid period and later (see ix-xi, below) is generally known from contemporary texts and travelers’ reports, even though it is usually not possible to associate particular carpets with them. It seems generally accepted that it was the Safavids who raised the status of carpet weaving in Persia from a nomadic and rural craft to a national industry, the products of which constituted a significant item of export to India, the East Indies, the Ottoman empire, and Europe. Nor is there much doubt that it was the Safavids who established the first independent carpet factories in some of Persia’s major urban centers (Varzī, p. 59).

It is not reasonable, however, to postulate that the Safavid carpet masterpieces came into being ex nihilo. Spuhler raises the pertinent question of why no definitely Timurid carpets, or even carpet fragments, are extant (p. 700; see viii, below). He points out that the Milan hunting carpet, possibly dated only a few years after the Uzbeks had brought Timurid rule to an end in Transoxania (906/1500-01) and Khorasan (913/1507; Savory, 1987, p. 78; see ix, below), is in excellent condition and speculates that perhaps some “early Safavid” carpets are really “late Timurid” carpets. Although Spuhler’s hypothesis cannot be proved, the current consensus is that Erdmann went too far in claiming that Safavid carpets represent a complete break from previous knotted pile carpets (1977, pp. 53-54; cf. Helfgott, p. 113). Beattie has found design elements on floor coverings depicted in miniature paintings of the 9th/15th century and earlier that portend those of Safavid carpets, including hints of Chinese influence (Beattie, p. 8; cf. viii, below). Leonard Helfgott’s verdict is “Whether or not a transitional period in carpet production preceded this fusion [of carpet-weaving skills with designs from the arts of the book, ceramics, and textiles], or whether this fusion began in a primitive state as a result of contacts with urban culture, or existed as a tendency or undercurrent as a result of Chinese influence, is impossible to determine at this time” (p. 114).

It is possible that Tabrīz was the principal center of carpet production during the early 10th/16th century, for Shah Esmāʿīl I (907-30/1501-24) was initially master only of Azerbaijan; it required ten years of hard fighting to bring the rest of Persia and Mesopotamia under Safavid control. The process was completed only with the capture of Herat in 916/1510 (Savory, 1987, pp. 17, 79-80). There is no agreement among scholars, however, as to whether or not Esmāʿīl or his successor, Shah Ṭahmāsb (930-84/1524-76), established a carpet factory at Tabrīz, the first capital of the Safavid state. Esmāʿīl is known to have assembled there an array of men skilled in the arts of the book: calligraphers, illuminators, and miniature painters. A major category of Safavid carpets is that known as “medallion” carpets (naqš-e toronj, toronj-dār; see iv, below), for which, according to Pope, the “most immediate and exact models . . . are to be found in book covers . . .” Both bookbindings (q.v.) and carpets were frequently designed by the illuminators of manuscripts (Pope, p. 2312). But did Esmāʿīl establish a carpet factory at Tabrīz? Edwards thought that the shah’s military preoccupations precluded such a move, though he admitted the possibility that Ṭahmāsb might have done so (p. 4). Maurice Dimand, on the other hand, categorically states, on stylistic grounds, that “All these early medallion rugs may be assigned to northwest Persia, particularly to Tabrīz, the Safavids’ main artistic center” (Dimand and Mailey, p. 42). He also assigned the Milan hunting carpet (see Plate CVIII in ix, below) and the “Emperor’s carpet” in Vienna (and its companion piece, now in The Metropolitan Museum of Art, New York; see Plate CX in ix, below) to Tabrīz in the period of Shah Esmāʿīl (Dimand and Mailey, pp. 42, 46; cf. ix, below). Ṭahmāsb, is known to have been personally interested in the technicalities of carpet production, for example, details of dyeing, design, and weaving, and is personally credited with having created some carpet designs (Varzī, p. 58). Furthermore, Dimand claimed that the Ardabīl carpets, which were woven during his reign, must have come from a Tabrīz workshop (Dimand and Mailey, p.46; cf. Gans-Ruedin, p. 38; see ardabīl carpet). There is thus a strong possibility that there was a carpet factory at Tabrīz, but, if so, production must have been interrupted by the Ottoman invasions of Azerbaijan and brief occupations of Tabrīz in 940/1533-34, 941/1534-35, and 955/1548 .

In 955/1548 Ṭahmāsb relocated the Safavid capital from Tabrīz to Qazvīn, and presumably the production of carpets under royal patronage was also moved. The Safavid court remained at Qazvīn for fifty years until it was transferred to Isfahan by Shah ʿAbbās I in 1006-07/1598. N. Aram-Zanganeh (p. 64) claims that the royal workshops were moved from Tabrīz to Qazvīn to Isfahan as each city successively became the capital, which is what would be expected, but Erdmann noted that “the manufacture of carpets at Qazvīṇ . . . cannot be authenticated” (1977, p. 36). The Hungarian ambassador to the court of the Ottoman sultan Solaymān reported the arrival in 975/1567 of gifts from Shah Ṭahmāsb, which included carpets made in Hamadān and Dargazīn; no mention was made of carpets from Qazvīn (J. von Hammer, Geschichte des osmanischen Reiches III, pp. 520-21; cf. Pope, p. 2335). The Spanish traveler Father Florencio del Niño Jesús reported that in the early 11th/17th century there was an abundance of silk and brocade carpets at Qazvīn (p. 100; cf. Spuhler, p. 702).

On the basis of textual sources and travelers’ reports most authorities seem to agree that there were royal Safavid carpet workshops at Isfahan, Kāšān, and Kermān (Florencio, p. 102; Tavernier, I, pp. 397, 589; Chardin, III, p. 120; cf. Beattie, p. 9; Erdmann, 1976, pp. 37, 41, 42). On the basis of Jean Chardin’s report of thirty-two royal workshops (boyūtāt-e ḵāṣṣa-ye šarīfa, boyūtāt-e salṭanatī, q.v.) employing an average of 150 craftsmen each, the total number employed in such workshops has been estimated at approximately 5,000 (VII, p. 329; according to Taḏkerat al-molūk, tr. Minorsky, p. 50/fol. 20a, in ca. 1139/1726 there were thirty-three royal workshops; cf. Helfgott, p. 113). Of course, members of many different crafts and trades were employed in these workshops, and the number of weavers may have been quite small. The workshops not only produced luxury carpets for the Safavid shahs’ mosques and palaces, for gifts to neighboring rulers like the Ottoman sultan, and for the export trade; they also fulfilled commissions from members of the nobility and other private citizens. In at least one instance toward the end of the 11th/17th century the court seems to have contracted out the production of carpets and other textiles to specialized craft guilds, a process known as taḥwīl-e aṣnāf (see Taḏkerat al-molūk, tr. Minorsky, p. 30). The shah or another client put up the capital in the form of raw materials and paid weavers regular wages while work was in progress. Sometimes the malek al-tojjār (state supervisor of textile guilds) played this role (Floor, p. 23).

At Isfahan the royal carpet workshops were located between the Čehel Sotūn (q.v.) and the Meydān-e Šāh (Tavernier, I, p. 444; cf. Varzī, p. 60). Father Tadeusz Krusiński, who was in Persia in 1116-42/1704-29, reported that rugs were made there for the royal Court (apud Mańkowski, p. 2431 and n. 5; cf. Beattie, p. 9); Tavernier had observed the same in the late 11th/17th century (I, p. 654). According to Florencio, Shah ʿAbbās I, like his grandfather Ṭahmāsb, was personally interested in the carpet industry and was himself skilled in carpet weaving (see Beattie, p. 9), and it would therefore be expected that the volume of carpets produced under royal patronage in Shah ʿAbbās’ reign would at least equal that produced under royal patronage during the reign of Shah Ṭahmāsb. Yet few major authorities are prepared to attribute to Isfahan the manufacture of even such a major category as the “vase carpets” (ṭarḥ-e goldānī, goldān-dār; see ix, below).

Father Krusiński also mentioned Kāšān as one of the cities in which royal carpet workshops had been established by Shah ʿAbbās. In 1601 Sigismund III Vasa, king of Poland, sent his purveyor Sefer Muratowicz there with instructions to order several rugs for the king and to superintend their weaving personally (for extracts from Muratowicz’s own report of his journey, see Mańkowski, pp. 2433-34; see also ix, below). At about the same time several Spanish Carmelites described Kāšān as a center of fine weaving of carpets in silk and gold (Florencio, p. 102; cf. Mańkowski, p. 2432 and n. 5, p. 2433 and nn. 1-2).

Kermān was another city of royal workshops, according to Father Krusiński (Mańkowski, p. 2431 n. 5). Engelbert Kaempfer saw woolen carpets with animal designs from there in the royal palace in 1095/1684 (Kaempfer, p. 225).

There was a flourishing export trade in carpets to Europe (sometimes via the Portuguese colony of Goa; see Gans-Ruedin, p. 11) and to the Mughal empire, where Safavid exports stimulated the local carpet industry; in effect, some Mughal products were almost “indistinguishable from Safavid carpets” (cf. Walker; Helfgott, p. 114). Some Safavid carpets were regularly shipped by the Dutch East India Company to Batavia, Ceylon, Malaysia, the Coromandel coast, Bengal, and Cochin, as well as to the Netherlands itself (Floor, p. 23). Apart from luxury carpets, which in London might cost as much as 100 pounds sterling for a pair, smaller carpets, in the 7-15 pounds sterling range, were sold there in 1685 and 1688 (Floor, p. 23). In general, however, trade between Persia and the Indies was more reliable and more profitable than trade between Persia and Europe; Persian exports could be shipped to Europe via the overland route through Ottoman territory but only when the Ottomans and the Safavids were not at war. Safavid foreign trade was almost exclusively in the hands of Armenians, who had established businesses in Livorno, Venice, and Amsterdam. The commercial and financial expertise of the Armenians and their European contacts was essential to Shah ʿAbbās’s policy of expanding overseas trade (Tavernier, II, p. 604; cf. Savory, 1980, p. 198).

Although the Safavids thus transformed a simple rural craft into a courtly art (Aram-Zanganeh, p. 12), carpets did continue to be produced by nomadic tribes and in smaller urban workshops. This production continued after the overthrow of the Safavids by the Afghans in 1135/1722 and the consequent decline of the luxury carpet industry, which had depended in large part on royal patronage. Nevertheless, there is some evidence that both Nāder Shah and Karīm Khan Zand commissioned carpets in southern Persia (see x, below). The breakdown of social and political institutions resulting from a half century of civil war between the Zand and the Qajars (ca. 1148-1209/1736-94) also affected the functioning of the economy. Law and order, especially security on the roads, which had played such an important part in encouraging the movement of commercial caravans in the time of Shah ʿAbbās I, could no longer be maintained. With the establishment of the Qajar dynasty in 1212/1797, the carpet industry began to flourish once more, largely owing to local financing and domestic market demand (see x, xi, below). By the middle of the 13th/19th century carpets of fine quality were being manufactured in a number of centers, as well as by several nomadic groups (Housego, p. 170; cf. x, below). The number of Persian carpets exported, however, remained small; it may have been the fortuitous conjunction of two factors, Persia’s need for an export commodity to substitute for silk and strong European demand, that produced a boom in Persian carpets at the end of the 13th/19th century.

Between 1281/1864 and 1284/1867 the silk-worm disease pébrine reached Persia from Europe and caused a drastic slump in Persian exports of raw silk (Housego, p. 169). According to the British consul in Tabrīz in 1289/1872 “the trade of Persia fell to the lowest point since commerce first seriously attracted the attention of foreign merchants.” Many Persian firms were bankrupted (Housego, p. 170). The subsequent sudden rise in European demand was stimulated particularly by the Vienna Exhibition of 1873, where carpets from Khorasan, Farāhān, Senna (Sanandaj), and other cities were exhibited (Housego, p. 170). Furthermore, the rise of the middle class in Britain and elsewhere increased the size of the potential market (Ittig, 1983, p. 68; idem, 1985, pp. 116-18 n. 22; see xi, below).

In 1294/1877 the Manchester firm Messrs. Ziegler and Co., which had been importing silk through its agents in Tabrīz, Isfahan, and Tehran, opened a new branch in Solṭānābād (Arāk) for the manufacture of carpets, which soon became “the centre of the carpet trade in Western Persia” (Curzon, II, p. 524). A rival British firm, Hotz and Co., also established a carpet factory there, and by 1296/1879 the carpet boom was well under way (see xi, below). At the same time an agent of the Paris department store Maison du Louvre was making annual visits to Persia for the purpose of buying carpets, and Persia was also beginning to export to the United States (Housego, p. 170).

By 1310/1892, according to Lord Curzon (II, p. 523), Persian carpets had “attained so wide a celebrity that the upper-class householder in England or America is rare who does not think the acquisition of such an article, whether genuine or spurious, an indispensable testimony both to culture and to civilisation.” Carpets of good quality were manufactured in the districts of Qāʾen and Bīrjand and Qohestān; at Saraḵs in Khorasan; in the districts of Qarādāḡ and Farāhān in Azerbaijan; in Kurdistan; and in Fārs, where carpets were made at Shiraz and tribal rugs woven by the Qašqāʾī (Curzon, II, p. 524). The boom in carpet production was particularly noticeable at Kermān (Ittig, 1985, passim). Whereas “in 1871 there were only six carpet-weaving establishments in the city, and fewer than thirty looms”—there were probably no more than one hundred looms operative in the whole country at that time (Ittig, 1985, pp. 121-22 and n. 2)—“by 1900 Sykes found over one thousand looms in Kirman City alone, each supervised by a master weaver” (English, p. 28 and n. 64; cf. xiii, below). Unfortunately, in an effort to keep up with European demand, Persian carpet manufacturers made the nearly fatal mistake of introducing the use of chemical dyes (see ii, xi-xiii, below), and, though the Persian government banned their use in 1304/1877, they continued to be smuggled in, to the detriment of the industry (Housego, p. 171). The late Qajar period was thus characterized, on one hand, by general deterioration in the quality of carpets, resulting from the use of aniline dyes and from hasty production methods introduced by manufacturers eager to meet the demand of foreign markets, and, on the other, by production of sumptuous silk carpets, the like of which had not been seen since the 11th/17th century (Tattersall, p. 99). High-quality carpets were frequently commissioned by members of the Qajar nobility and of the upper classes generally (Ittig, 1985, passim; Gans-Ruedin, pp. 249, 285, 440). Even in the last quarter of the 13th/19th century, however, Safavid carpets still had an overwhelming impact on those who saw them for the first time. William Morris, writing to a friend in 1877, reported: “I saw yesterday a piece of ancient Persian, time of Shah ʿAbbās . . . that fairly threw me on my back; I had no idea that such wonders could be done with carpets” (Sylvester, p. 7).

The Persian carpet industry has had a checkered history since its resurgence in the last quarter of the 13th/19th century (see xii, below). Not surprisingly, the two world wars represented periods of decline. Persia, caught up in both conflicts by reason of its geopolitical situation, twice experienced loss of foreign markets and resulting severe impairment to domestic prosperity. Furthermore, prophets of doom confidently predicted that the introduction of machinery would put an end to the carpet industry (Edwards, p. 362). Nevertheless, by the mid-1920s Persians were again producing “sound and attractive carpets,” and, until the occupation by Allied forces in 1320 Š./1941, “the standards of production were, on the whole, well maintained” (Edwards, p. 362). With the accession of Reżā Shah (1303 Š./1924), royal patronage was resumed, and the industry once again produced carpets that could be compared, in beauty and delicacy of design and finesse of weave, to the finest productions of the 10-11th/16-17th centuries. Some of the carpets woven at the famous Amoḡlī factory in Mašhad actually surpassed the finest extant Safavid carpets in density of knotting. For example, two Mašhad carpets of 1309 Š./1930, with 906 knots to the square inch, and another of comparable quality made in 1314 Š./1935 can be compared to a fine Safavid carpet, for years considered to have a silk pile because its knotting is so dense, which has only 822 knots to the square inch (Gans-Ruedin, p. 66).

By 1327 Š./1948, however, the Persian carpet industry was once more in a “hapless state” (Edwards, p. 362), and in the following year the government convened a conference in Tehran to draw up recommendations for dealing with the three principal evils that had beset the industry intermittently for the previous sixty years and more: the use of aniline and chrome dyes, a decline in the standard of design, and the use of the joftī knot, which produced a carpet inferior both in beauty and in durability (Edwards, pp. 363-66; see iii, below). The remedial measures instituted by the government, together with growing consumer resistance to inferior carpets, once again led to a revival of the industry. Foreign demand for Persian carpets, especially in the United States, continued strong. After 1328 Š./1949 there were periods of “intense” carpet production in Kermān, owing primarily to demand from the United States for the very thick rugs known in the market as “American Kirmans” (Gans-Ruedin, p. 423; see xii, xiii, below). Royal patronage ensured that some carpets of exceptional technical quality were also produced. Gans-Ruedin (pp. 375, 38, 58) mentions, for example, a 1349 Š./1970 Isfahan carpet with 710 knots to the square inch, which may be compared with three famous 10th/16th-century carpets, those known as the Ardabīl carpets (400 knots per square inch; Plate CIX in ix, below) and the Milan hunting carpet (367 knots per square inch; Plate CVII in ix, below). The establishment of the Carpet Museum in Tehran in the mid-1350s Š./1970s contributed to the revival of public interest in this aspect of Persian culture.

Bibliography : N. Aram-Zanganeh, Le tapis persan aux XVI^e et XVII^e siècles. Contribution à une sociologie de l’art persan, Lausanne, 1984. M. H. Beattie, Carpets of Central Persia (with Special Reference to Kirman), Sheffield, 1976. C. Bier, “Court and Commerce. Carpets of Safavid Iran,” in C. Bier, ed., Woven from the Soul, Spun from the Heart. Textile Arts of Safavid and Qajar Iran, 16th-19th Centuries, Washington, D.C., 1987, pp. 97-106. A. Briggs, “Timurid Carpets,” Ars Islamica 7, 1940, pp. 20-41; 11-12, 1943-46, pp. 146-58. J. Chardin, Voyages du chevalier Chardin en Perse . . ., 4 vols., Amsterdam, 1735. G. N. Curzon, Persia and the Persian Question, 2 vols., London, 1892. W. B. Denny, Oriental Rugs. Cooper-Hewitt Museum, Washington, D.C., 1979. M. S. Dimand and J. Mailey, Oriental Rugs in the Metropolitan Museum of Art, New York, 1973. A. C. Edwards, The Persian Carpet. A Survey of the Carpet-Weaving Industry of Persia, London, 1953. P. W. English, City and Village in Iran. Settlement and Economy in the Kirman Basin, Madison, Wis., 1966. K. Erdmann, Seven Hundred Years of Oriental Carpets, tr. M. H. Beattie and H. Herzog, Berkeley and Los Angeles, 1970. Idem, Oriental Carpets and Their History, Fishguard, Wales, 1976. Idem, The History of the Early Turkish Carpet, London, 1977. Idem, “The Pattern Structure of the Arabesque Carpets,” in Survey of Persian Art XIV, pp. 3160-62. W. Floor, “Economy and Society. Fibers, Fabrics, Factories,” in C. Bier, ed., Woven from the Soul, Spun from the Heart. Textile Arts of Safavid and Qajar Iran, 16th-19th Centuries, Washington, D.C., 1987, pp. 20-32. Florencio del Niño Jesús, “A Persia (1604-9). Peripecias de una embajada pontificia que fué a Persia a principios del siglo XVII,” in Biblioteca Carmelitana-Teresiana de Misiones II, Pamplona, 1929. E. Gans-Ruedin, Iranian Carpets. Art, Craft and History, London, 1978. E. Gibbon, The Decline and Fall of the Roman Empire I, New York, 1900. L. M. Helfgott, “Production and Trade. The Persian Carpet Industry,” in C. Bier, ed., Woven from the Soul, Spun from the Heart. Textile Arts of Safavid and Qajar Iran, 16th-19th Centuries, Washington, D.C., 1987, pp. 107-20. J. Housego, “The 19th Century Carpet. Persian Carpet Boom,” Oriental Art 19, 1973, pp. 169-71. R. G. Hubel, The Book of Carpets, tr. K. Watson, New York and Washington, D.C., 1970. A. Ittig, “A Group of Inscribed Carpets from Persian Kurdistan,” Halı 4/2, 1981, pp. 24-27. Idem, A Technical and Historical Study of the Qajar Carpet Industry. The Case of Garrus, Sultanabad and Kirman, Ph.D. dissertation, University of Oxford, 1983. Idem, “The Kirmani Carpet Boom,” Oriental Carpet and Textile Studies 1, 1985, pp. 111-23. E. Kaempfer, Amoenitatum Exoticarum Politico-Physico-Medicarum fasciculi V, quibus continentuṛ . . ., Lemgo, 1712. T. Kawami, “Archaeological Evidence for Textiles in Pre-Islamic Iran,” Iranian Studies (forthcoming). J. Mills, Carpets in Pictures, Themes and Painters in the National Gallery, ser. 2, no. 1, London, n.d. T. Mańkowski, “The Art of Carpet Making B. Some Documents from Polish Sources Relating to Carpet Making in the Time of Shah ʿAbbās I,” in Survey of Persian Art VI, pp. 2431-36. A. U. Pope, “The Art of Carpet Making. A. History,” in Survey of Persian Art VI, pp. 2257-2430. R. Savory, Iran under the Safavids, Cambridge, 1980. Idem, Studies on the History of Ṣafawid Iran, London, 1987. R. B. Serjeant, “Material for a History of Islamic Textiles up to the Mongol Conquest,” Ars Orientalis 9-15/16, 1942-50; repr. Beirut, 1972. F. Spuhler, “Carpets and Textiles,” in Camb. Hist. Iran VI, pp. 698-727. D. Sylvester, “On Western Attitudes to Eastern Carpets,” in Islamic Carpets from the Collection of Joseph V. McMullan, Arts Council of Great Britain, 1972. C. E. C. Tattersall, “Carpets,” in E. D. Ross, ed., Persian Art, London, 1930, pp. 94-100. J. B. Tavernier, Les six voyages de J. B. Tavernier en Turquie, en Perse et aux Indes, 2 vols., Paris, 1676. M. Varzī, Honar wa ṣaṇʿat-e qālī dar Īrān, Tehran, 1350 Š./1971. D. Walker, “Classical Indian Rugs,” Halı 4/3, 1982, pp. 252-57.

(Roger Savory)

For centuries Persian carpet weaving has depended primarily on local materials processed by traditional techniques. The introduction of merino wool from Australia was short-lived (Report of the Iran Carpet Company), and, even after synthetic chemical dyes were introduced around the turn of the 14th/20th century, they were often processed according to traditional techniques (see, e.g., A. C. Edwards, p. 167). Until the 1320s Š./1940s vegetable dyes continued in common use (Iran Carpet Company, 1949).

Fibers and their preparation. The basic raw materials for carpets are natural fibers, of which sheep wool (pašm-e gūsfand) is important. In different parts of Persia different kinds of sheep are raised, depending on the prevailing climate and available pasturage. In colder areas, for example, the sheep produce a finer, long-staple wool fiber (Formenton, p. 50); that of Khorasan is particularly fine (A. C. Edwards, p. 166). In warmer Kermān, however, a shorter, springier fiber is produced (Dhamija, 1971a); it is particularly durable and has a sheen. In Fārs local wool is used by the Qašqāʾī, Ḵamsa, and settled rural weavers, often from their own sheep. It is soft, lustrous, and absorbs dyes very well (Dhamija, 1971b). The wool of Kermānšāh is lighter and finer. In northwestern Persia the wool is relatively coarse (A. C. Edwards, p. 61). That from around Ardabīl is strong and lustrous and absorbs dyes well. Weavers in the vicinity of Herīs (Harīs/Harīz), near Tabrīz, formerly used a fine, lustrous wool obtained from the nomadic Šāhsevan, but recently they have been buying it from Ardabīl; only in remote villages are the Šāhsevan still the main suppliers. At Dargazīn (or Darra-ye Gazīn), near Hamadān, wool is also still purchased from the Šāhsevan and blended with the fine local wool (A. C. Edwards, p. 91). In the northwestern region of Mākū, after the spring shearing (pašm borīdan/čīdan, qeyčī zadan/kardan), wool was customarily blended with lamb’s wool (quzem) to produce a yarn of very high quality. Such a blend was used in the weaving of fine carpets (A. C. Edwards, p. 58).

In Persia wool was traditionally either sheared with scissors (qeyčī) or plucked (kandan), but in recent decades only the wool of dead animals (ṭabāḵī) has been plucked (Wulff, Crafts, p. 177). In most areas the sheep are prepared for shearing in the following way: Their wool is first combed (šāna zadan), then washed (šostan) with soap, potash (qālī-āb), or a solution of soap nuts (Sapindus saponaria) and potash (Wulff, Crafts, p. 179). Sometimes the sheared wool is sorted according to whether it comes from the underbelly (šekam), shoulder (šāna), back (pošt), or neck (gardan). The fibers are then straightened by hand and separated by means of an upright comb fastened to a base (šāna-ye mīḵ; see Wulff, Crafts, p. 182, fig. 260). This method is preferred to bowing (kaman zadan) because it loosens and separates the fibers and arranges them in parallel fashion, which facilitates spinning the yarn. The combed wool is then gathered into a coil (kalāfa), wound around (pīčīdan) a rod or the arm of the spinner (rīsanda), or simply put into a bag (kīsa). A spindle (dūk, pīl, parra) is most commonly used; the spinning wheel (čarḵ-e rīsandagī), though it is found among the settled peasantry (raʿīyatī), is more often used for plying or twisting the yarn. The spinner feeds the raw wool (ḵāma) onto the spindle, which is set spinning rapidly by means of a sharp twist of the wrist; the spinner continues to pay out the wool as the weighted spindle rotates toward the ground, spinning the fibers into yarn. The spindle is then lifted and the yarn (naḵ, rīsmān, qātma) wound round it before the entire process is repeated. In Persia the spindle is generally twisted outward to the left, and the resulting yarn thus has a Z twist (i.e., the yarn spirals in the same direction as the diagonal stroke in the letter Z), though an S twist does occasionally occur (Wulff, Crafts, pp. 185-86; see also early Islamic carpets, below). Lamb’s wool is processed in the same way.

Camel’s hair (pašm-e šotor) is used for less expensive carpets in the Čahār Maḥāl(l) area (see baḵtǰārǰ tribe iii) and in Baluchistan (see A. C. Edwards, p. 186; baluchistan v), as well as to provide tan and brown colors in the weaving of kilims (gelīms, kelīms) in Khorasan and Baluchistan (cf. Jacoby, p. 2457). It has been particularly preferred for woven flour wrappers (sofra-ye ārd) in all parts of Persia, as it generates heat and thus ensures the rising of dough (information collected in Birjand, 1970). Camel’s hair is usually plucked from the animals in the spring, often by the camel drivers as they walk along, and then processed in much the same way as wool (Wulff, Crafts, p. 177).

Goat hair (mū-ye boz), mixed with coarse wool from sheep, is often used for the warps of tribal carpets, especially those woven for domestic use (see also boz; field studies in Bīrjand, Sīstān, Baluchistan, Fārs). It is also in Baluch flat-woven carpets and is favored for the selvedges of the coarser Baluch carpets (A. C. Edwards, p. 25; baluchistan v). In contrast to wool, goat hair is clipped, washed, combed and then spun or combined with sheep wool and spun.

In the past silk (abrīšam, q.v.) was used in the warps and pile of finer carpets (see Safavid carpets, below) and sometimes also to provide areas of white in deep maroon or crimson and black wool pile. In central Persia, particularly Kāšān, Nāʾīn, and Qom, carpets were woven entirely of silk; particularly noteworthy are the silk barjasta (relief) carpets, with only the patterns knotted in pile on a plain-weave foundation, which were made in Kāšān (personal communications from G. Anavian, Tehran, and traditional carpet weavers in Kāšān). Silkworms (pīla-ye abrīšam) are cultivated and processed by the peasants of Gīlān, Māzandarān, and the Isfahan area. The silk of the first two regions is reeled and marketed for carpet weaving in three grades: dāna, the finest, for knotted pile; haštī, used in the warps of very fine wool carpets; and pūdī, the coarsest grade, used for the wefts of silk carpets (Wulff, Crafts, pp. 182-83). The Turkmen of Māzandarān, especially around Marv Tappa, and in the Jargarān and Zāvīn areas of Khorasan also raise silkworms in a very primitive fashion and use part of the silk in their carpets (observed during field studies, 1972 and 1978; see also Dhamija, 1979, color pl. 4).

Gold-wrapped thread (golābetūn), which contributed to the luxury of court carpets in the past, when they were woven into the pile or non-tufted fine-quality silk (see Dimand, pp. 2, 59; see Safavid carpets, below), is only rarely used today, though it does occur in the form of brocading (zarī) on the flat-woven portions of the barjasta carpets (seen in private collections in Persia, India) and in the very fine kilims of Ḵᵛoy in Azerbaijan (private collection of A. de Franchis; also seen with dealers in Mākū and Ḵᵛoy).

Short-staple cotton (panba, katān) is grown in many parts of Persia and was an important source of fiber for the carpet industry (see cotton). Generally commercially prepared undyed cotton is used in the warps of finer carpets made in commercial workshops because it is less elastic and holds its shape better (Jacoby, p. 2458; A. C. Edwards, pp. 25-26). It is also occasionally used to add white highlights in the kilims of the Lor and Baḵtīārī and to the so-called palās of the Turkmen (Persian Handicraft, p. 308; for medieval references to rugs called palās, see Ḥodūd al-ʿālam, fols. 20b, 23a, 32b; tr. Minorsky, pp. 106, 114, 142). Although as recently as the 1330s Š./1950s handspun cotton was used for less expensive commercial carpets (A. C. Edwards, p. 25), weavers generally buy cotton in skeins, then wind it into balls and ply or rewind it as required.

For further discussion of fibers used in tribal weaving, see Tribal carpets, below.

Dyeing. Persia has been renowned for its dyes for many centuries; in 987/1579, for example, an Englishman was sent to learn the secrets of dyeing wool and silk in the Persian manner (Hakluyt, p. 202; cf. Jacoby, p. 2459 n. 1). Traditionally Persian dyers have used vegetable, animal, and mineral products to produce both dyes and mordants, the agents used to enhance the fibers’ capacity to absorb dyes and to fix the colors. Unfortunately, this aspect of carpet making has been little explored by scholars. Because reports in medieval texts are seldom detailed and often somewhat confused and because the colors of antique carpets have rarely been identified scientifically, it is difficult to trace the history of the dyestuffs and dyeing practices observed in the 13th/19th and 14th/20th centuries.

In the dyeing process as observed by ethnologists and other field workers wool yarn is first scoured (šostan) with a mixture of carbonate of soda and soap, then repeatedly rinsed in potash (qālī-āb) and hot water until the dirt and most of the grease have been removed (Qashqāʾi, p. 54; A. C. Edwards, pp. 31-32). Care must be taken not to remove the grease so completely that the fibers become brittle. The clean wool is then immersed in a mordant. The most commonly used mordant is alum (zāj), of which two varieties have been commonly available in Iran: aluminum sulphate (zāj-e safīd) and potassium alum (zāj-e qālīya). In this connection it should be noted that “pure alum” was exported from Persia to China as early as the T’ang period (Schafer, pp. 217, 330 n. 16). In the late 10th/16th century alum from Gīlān was being sold at Shemakha in the eastern Caucasus (A. Edwards, pp. 378-79). Sometimes a variety of astringents with high concentrations of tannic acid are combined with the mordants in order to enhance the fixative properties of the latter and also to darken the colors of the dyes. Very similar procedures were observed by A. C. Edwards in the late 1320s Š./1940s (pp. 31-34).

The range of dyes available in Persia seems to have been rather limited before the introduction of modern aniline dyes in the late 13th/19th century (Jacoby, p. 2459; cf. A. C. Edwards, pp. 29-34). The most popular color for carpets has always been red (qermez, sorḵ), of which a wide range of shades and tones can be obtained from various materials. Historically the most common was madder (rūnās/rūnīās), extracted from the root of Rubia tinctorum. Today the plant grows wild in Māzandarān, Kermān, and around Tehran and Yazd (cf. A. C. Edwards, p. 31). The dye yields different hues, shades, and tints, depending on the mordant used (e.g., whether or not astringents or acids are included), the quality of the fibers, the age of the plant from which the dye is extracted, the season in which it is harvested, and the like. The purity of the water and the addition of substances like dried yogurt (dūḡ, q.v.) and the juice of sour grapes, āb-e deraḵš-e torš also affect the results (Wulff, Crafts, p. 190; Jacoby, pp. 2460-61; Edwards, pp. 31-32; Handbook, pp. 34-35). In the 4th/10th century madder was produced throughout Azerbaijan and Arrān, especially at Mūqān and in Barḏaʿa (q.v.), in Warthān (Vartan, modern Altan, on the Araks river north of Ardabīl), and at Bāb al-Abwāb (Darband) and a neighboring island in the Caspian Sea; from these sources it was shipped by sea to Gorgān and thence overland to India (Ḥodūd al-ʿālam, fol. 32b; tr. Minorsky, pp. 142-43; Ebn Ḥawqal, pp. 347, 388; tr. Kramers, II, pp. 340-41, 378). At the same period it was exported to India from Qowāḏīān in Transoxiana (Eṣṭaḵrī, p. 298; Ebn Ḥawqal, p. 477; tr. Kramers, II, p. 459). According to Qazvīnī, in the Mongol period madder was produced in abundance at Ḵᵛāf in Khorasan (Nozhat al-qolūb, p. 154), and it was also grown at Yazd and Nāʾīn (Camb. Hist. Iran V, p. 502 n. 2, citing the anonymous 8th/14th-century author of Ketāb-e ʿelm-e felāḥat wa zerāʿat, ed. ʿA. Najm-al-Dawla, Tehran, 1323/1905, p. 94). In the reign of Shah Esmāʿīl (907-30/1501-24) the dye made at Ḵᵛoy in northwestern Azerbaijan was being shipped to India via Hormoz (Grey, pp. 165-66; cf. Jacoby, p. 2460 n. 1). At about the same time madder was also cultivated around Ḡazna in Afghanistan, and the entire crop was exported to India (Bābor-nāma, p. 218). Because of growing world demand for Indian textiles, the Indian market for madder remained important; throughout the 11-12th/17-18th centuries it continued to be shipped from northwestern Persia via Bandar-e ʿAbbās (q.v.), which had supplanted nearby Hormoz, except for a brief period between 1025/1616 and 1031/1622, when hostilities between England and Portugal forced a shift from the Gulf to an overland route from Ardabīl via Isfahan and Qandahār (Tavernier, I, p. 53; Ferrier, p. 202 and n. 99; Camb. Hist. Iran VI, pp. 447, 481; Fukasawa, pp. 49, 65 n. 90). Centers for production of the dye in that period included, beside Ardabīl, Tiflis (Tbilisi) in Georgia and Armenian Astabat (Ordūbād?), on the Araxes below Jolfā (Ferrier, p. 202; Tournefort, III, p. 170; Tavernier, I, p. 52; cf. Fukasawa, p. 64 n. 89). In the early 12th/18th century madder was also shipped in quantity from Tiflis to Erzurum, whence it reached other Anatolian cities (Tournefort, III, pp. 111, 170; Fukasawa, p. 65 n. 91). A few years later an Armenian from a village near Naḵčevān introduced the cultivation of madder into southern France (see althen). Today the dye is sold in prepared form at Shiraz and Isfahan (Qashqāʾī, p. 55), but it had already become scarce and expensive by mid-century, and it has been largely replaced commercially by aniline reds (A. C. Edwards, p. 30). Another category of red dyes includes those made from the dried bodies of female insects (qermez- dāna), found mostly in northern Persia and Armenia; it includes kermes, from Kermococcus vermilio, and “Armenian red” (Armenian vortan), from Porphyrophora hamelii, which is chemically similar to New World cochineal (cf. Forbes, IV, pp. 102-03; on cochineal see A. C. Edwards, pp. 33-34). Lac (lāk), a red dye from female insects (Coccus lacca), was produced in India. These dyes are frequently confused in medieval sources, and modern scholars also sometimes use the name of one type as a generic term for all (Lombard, pp. 118-22; Wulff, Crafts, pp. 189-90; Jacoby, p. 2460; Kurdian; Serjeant, XV, p. 35; A. C. Edwards, pp. 33-34). Red-dyed cloth from Armenia (Urartu) was mentioned in an inscription of Sargon II in 714 b.c. (Forbes, IV, p. 102). The dye used was very probably “Armenian red,” for the insects from which it comes feed on grasses that are particularly common in the territory at the base of Mount Ararat (Kurdian, pp. 105-06). In the 3rd/9th century Balāḏorī noted that the town of Ardesāṭ (Artaxata, q.v.) in Armenia was also known as qarya qermez (the village of kermes; Fotūḥ, p. 200; cf. Kurdian, p. 105; Levi della Vida, p. 288). His contemporary Jāḥeẓ mentions a red dye from insects that feed on grass (Tabaṣṣor, p. 24, apud Kramers, in Ebn Ḥawqal, tr., II, p. 335 n. 631) in Armenia. In the next century such dyes were important products of Azerbaijan and of Dabīl (Dwin) in Armenia (Ḥodūd al-ʿālām, fol. 32b; tr. Minorsky, pp. 142-43; Ebn Ḥawqal, II, pp. 342-43; tr. II, p. 335; Eṣṭaḵrī, p. 188). In the 7-8th/14-15th centuries kermes was still produced at Marand and elsewhere in Azerbaijan (Nozhat al-qolūb, ed. Le Strange, I, p. 88, II, p. 89 and n. 2). Other red dyes were imported. In the early 11th/17th century sappanwood (baqqam; Caesalpinia sappan) from the Far East was being imported to Persia (Ferrier, p. 210), as was lac from India (Ferrier, p. 207 and n. 128; Wulff, Crafts, p. 191; A. C. Edwards, p. 33; cf. Heyd, II, pp. 612-13). Both were among the commodities brought from the east to Bandar-e ʿAbbās by the Dutch in 1135/1721 (Camb. Hist. Iran VI, p. 483). By the late 10th/16th century the English were importing cochineal from the New World into Baku (Burroughs, p. 446), and both cochineal and vermilion (a mineral dye) were among English imports to Persia in the early 11th/17th century as well (Ferrier, pp. 212-13; cf. Camb. Hist. Iran VI, p. 448). Later in the century Armenians from Isfahan were trading for cochineal in Aleppo (Camb. Hist. Iran VI, p. 460).

Blue (nīl, rang-e kermānī, rang-e wasma) is produced from the fermented leaves of the indigo plant (Indigofera tinctoria). According to one Chinese source, as early as the 98-99/717 indigo was sent as a gift from Samarkand to China (Schafer, pp. 212, 329 n. 47). In the Middle Ages the plant was grown in Kermān: at Sīrjān and in the districts between Maḡūn and Valāšgerd, on one hand, and Hormoz, on the other (Ebn Ḥawqal, p. 312, tr. Kramers, p. 307; Moqaddasī, pp. 465, 467; Eṣṭaḵrī, p. 167; Ḥodūd al-ʿālam, fol. 26b; tr. Minorsky, pp. 123-24; cf. Heyd, II, p. 598 n. 6). It was also grown in the region around Kabul (Ebn Ḥawqal, pp. 184-85, 450; tr. Kramers, p. 183; cf. Marco Polo, pp. 302, 304, 306). Maḥmūd of Ḡazna and his successors received indigo from their conquered territory of Multan; they kept part and sent some as diplomatic gifts to Baghdad and elsewhere (Bayhaqī, p. 293; ʿAwfī, Jawāmeʿ al-ḥekāyāt, in Elliot and Dowson, History of India II, p. 189; S. H. Hodivala, Studies in Indo-Muslim History, p. 176; cf. Bosworth, Ghaznavids, p. 76). By the 8th/14th century cultivation of indigo had died out in Persia, though Ḡāzān Khan made an unsuccessful attempt to revive it (Camb. Hist. Iran V, p. 502, citing the anonymous 8th/14th-century author of Ketāb-e ʿelm-e felāḥat). Instead dyers relied on imports from India, which seems to have been the main source of the dye through most of the subsequent centuries (see, e.g., Ferrier, pp. 207-08; Camb. Hist. Iran VI, p. 475). In the 14th/20th century, however, small quantities of the plant have been grown near Šūštar and Dezfūl in Ḵūzestān, though synthetic indigo seems largely to have replaced the natural dye (Wulff, Crafts, p. 192; Jacoby, p. 2460; A. C. Edwards, p. 33). Recently the Persian Carpet Company attempted to revive the cultivation and use of indigo in Persia.

Yellow dyes also come from a variety of sources. The flowers of esparak (weld, Reseda luteola; A. C. Edwards, p. 32; cf. Handbook, p. 35), which grows wild throughout Persia and is also cultivated in Khorasan, yield a dye that is particularly effective with silk, though it can be used on wool if mixed with other dyes. Another important source of yellow dye is safflowers (gol-e rang; Carthamus tinctorius), also widely cultivated in Persia; the extract from the petals, when combined in solution with fuller’s earth and potash, yields a range of colors from clear yellow to orange. Yellow is also obtained from autumn vine leaves, pomegranate rinds, turmeric (zard-čūba), somāq, and saffron (Crocus sativus; Wulff, Crafts, p. 191; Jacoby, p. 2461). Both saffron (kurkum) and turmeric were mentioned as dyestuffs in the Bundahišn (TD₂, p. 118.4; tr. Anklesaria, chap. 26.20, pp. 148-49). Saffron was particularly costly, and, though it is mentioned frequently in medieval sources, it is not clear that it continued to be used as a dye. All these yellow dyes are also used in combination with other dyes to achieve special shades. Cloth already dyed with indigo will turn green when immersed in esparak or gol-e rang. Esparak mordanted with copper sulphate produces a soft blue-green color (āb-e sangar); copper sulphate tends to destroy woolen fibers, however, so that it has not often been used for carpets (Jacoby, p. 2462 and n. 1). Green can also be obtained by mixing willow leaves with indigo before dyeing (Wulff, Crafts, p. 191).

Shades of brown, beige, and tan are produced from green walnut hulls (pūst-e gerdū; Juglans regia) and oak bark (jaft; Edwards, p. 32). Dyeing dark or natural black wool with iron oxide (ferruginous litharge) produces a deep, glowing black, which, however, fades quickly (A. C. Edwards, p. 186); it is most often found in Baluchistan. In other areas black was obtained by using oak galls (Formenton, p.53). Compounds of indigo and henna also produce a luminous black. The Lor and Baḵtīārī obtain black from the hulls of wild acorns (balūṭ, q.v.) from which they occasionally make bread (q.v.; information from the late Nāder Afšār Nāderī).

In most Persian towns there are professional dyers (rang-raz), whose secret recipes have often been handed down from father to son for generations. The dyeing trade was highly regarded in earlier times, and, as demonstrated above, dyers traveled to many parts of the world in order to train others to prepare special dyes for large workshops (kār-ḵānas). In remote rural areas and among tribal people dyeing is done by women, who gather flowers, bark, fruit seeds and peelings, nuts, and the urine of cows and horses (for fermentation and to achieve special tints). They also buy dried roots and other essentials. Their skills and knowledge are passed on to their daughters and daughters-in-law.

Bibliography : Ẓahīr-al-Dīn Moḥammad Bābor, Bābor-nāma, tr. A. S. Beveridge, Bābur-nāma (Memoirs of Bābur) I, London, 1922; repr. New Delhi, 1970. C. Burroughs, “Aduertisements and Reports of the 6. Voyage into the Parts of Persia and Media . . .,” in Hakluyt, The Principal Navigations I, pp. 440-53. J. Dhamija, Survey of Carpet Manufacturing in Sirjan, Kerman Ostan, Report ILO/UNDP/Rural Non-Farm Employment, Tehran, 1971a. Idem, Survey of Income Generating Activities Merv Dasht, Ostan Fars, ILO/UNDP/Rural Non-Farm Employment, Tehran, 1971b. Idem, Living Tradition in Iran’s Crafts, New Delhi, 1979. M. S. Dimand and J. Mailey, Oriental Rugs in the Metropolitan Museum of Art, New York, 1973. A. Edwards, “An Other Letter of the Said M. Arthur Edwards, Written the 26. of April 1566. in Shamakie in Media . . .” in Hakluyt, The Principal Navigations I, pp. 377-79. A. C. Edwards, The Persian Carpet. A Survey of the Carpet-Weaving Industry of Persia, London, 1953. R. Ferrier, “An English View of Persian Trade in 1618. Reports from the Merchants Edward Pettus and Thomas Barker,” JESHO 19, 1976, pp. 182-214. R. J. Forbes, Studies in Ancient Technology, 2nd ed., Leiden, 1964. F. Formenton, Oriental Rugs and Carpets, London, 1972. C. Grey, ed. and tr., “The Travels of a Merchant in Persia,” in A Narrative of Italian Travels in Persia, in the Fifteenth and Sixteenth Centuries, Hakluyt Society 1/49, London, 1873; repr. New York, n.d., pp. 139-207. K. Fukasawa, Toileries et commerce du Levant d’Alep à Marseille, Paris, 1987. R. Hakluyt, The Principal Navigations, Voyages, Traffiques and Discoveries of the English Nation, London, 1589; facsim. ed., 2 vols., Cambridge, 1965. R. Hakluyt, “Certaine Directions Given by M. Richard Hackluit of the Middle Temple, to M. Morgan Hubblethorne, Dier, Sent into Persia, 1579,” in Hakluyt, The Principal Navigations I, pp. 454-55. Handbook of Dye Plants and Dyeing, Plants and Gardens 20/3, 1964. W. Heyd, Geschichte des Levantehandels im Mittelalter, 2 vols., Stuttgart, 1879; repr. Hildesheim, 1971. D. Hooper and H. Field, “Useful Plants and Drugs of Iran and Iraq,” Field Museum of Natural History, Fieldiana, Botanical Series 9/3, pp. 73-241. Iran Carpet Company, Report, Tehran, 1339 Š./1960. H. Jacoby, “Materials Used in the Making of Carpets,” in Survey of Persian Art, pp. 2456-65. H. Kurdian, “Kirmiz,” JAOS 61, 1941, pp. 105-07. G. Levi della Vida, “On Kirmiz,” JAOS 61, 1941, pp. 287-88. M. Lombard, Les textiles dans le monde musulman du VII^e au XII^e siècle, Civilisations et sociétés 61, Paris and The Hague, 1978. Persian Handicrafts. A Survey, Tehran, 1977. The Qashqāʾi of Iran [exhibition catalogue], London, 1976. G. Schaefer, “Der Anbau und die Veredelung der Krappwurzel,” CIBA-Rundschau 47, 1940, p. 1715. E. H. Schafer, The Golden Peaches of Samarkand. A Study of T’ang Exotics, Berkeley and Los Angeles, 1963. J. Takasuku, “Aomi-no Mabito Genkai (779). Le voyage de Kanshin en Orient (742-754),” Bulletin de l’Ēcole française d’Extrême-Orient 28, 1928, pp. 1-41, 441-72; 29, 1929, pp. 47-62. J.-B. Tavernier, Les six voyages de Jean-Baptiste Tavernieṛ . . ., 2 vols., Paris, 1679. P. de Tournefort, Relation d’un voyage du Levant fait par ordre du roy, 3 vols., Lyons, 1727.

(Jasleen Dhamija)

The techniques of carpet making are the processes of weaving, knotting, and finishing; structure is the complex of interrelations among the elements of the finished carpet. One of the major problems in carpet studies is the lack of a standard terminology to describe specific techniques, structures, and designs. The nomenclature most widely used in scholarly publications has been adopted here. (For an extensive listing of colloquial terms, see Varzī, pp. 289-99).

Technique. Whether the carpet is flat-woven (see v, below) or piled, the basic components are the warps (tār, tān, tūn, čella; Figure 64a), which run from end to end of the rug, and the wefts (pūd, pūt; Figure 64b), which are passed from side to side alternately over and under the warps, thus binding them together to produce the foundation of the carpet. Each drawing of a weft across the warps is called a “shoot,” or “pass.” In pile rugs supplementary weft yarns, or “knots” (ḵeft, gereh, ḡond), are wrapped around the warps between the foundation wefts, then cut; the cut ends project on the surface to form the pile (porz; Figure 65a-c) and create the design. Knots are generally tied to pairs of warps. The two basic types of knots found in Persian carpets are the symmetrical (Figure 65a) and the asymmetrical both of which may be open either to the right (Figure 65b) or, more commonly, to the left (Figure 65c). In earlier carpet literature the symmetrical knot was generally called the Turkish or Ghiordes knot: among Persians in the trade the technique of knotting carpets in this fashion is commonly known as torkībāf. The corresponding terms for the asymmetrical knot are Persian or Senna and fārsībāf. This older terminology is problematic, however, as both types of knot are used throughout the Near and Middle East, without respect to geopolitical boundaries or the linguistic or ethnic affiliations of the weavers. Knots of either type tied over more than two warps are known colloquially as joftī “doubled, paired.” Although the joftī technique requires less time and material than tying knots onto every two warps and thus reduces production costs, the resulting decrease in knot density leads to less detailed designs and sometimes to carpets of lower quality.

Despite official adoption of the metric system in Persia (1314 Š./1935), pre-existing systems of carpet measurement, which vary from region to region, are still in use. The standard measure of length is the ḏaṛʿ (cubit), subdivided into 16 units, also called gerehs, or knots; the ḏaṛʿ varies in length from 100 to 111.25 cm (40-44.5 inches). Knot density is measured by raj, that is, the number of rows of knots per longitudinal gereh (6.25-6.95 cm/2.5-2.78 inches). In the West, by contrast, density is determined by the number of knots per square decimeter or square inch.

The loom (dastgāh, lit. “equipment,” dār, lit. “pole”; Figure 66 is the frame upon which carpets are woven). The warps (Figure 66a) are stretched between two parallel beams (navard), which may be fixed or rotating. When the beams are fixed the length of the carpet is limited to the distance between them. With rotating beams, on the other hand, the carpet can be wound around the lower beam as weaving progresses, so that the finished carpet can be longer than the loom. To facilitate weaving, alternate warps are usually separated and raised to make a shed through which weft yarns can be passed. The shed is created by means of a flat shed stick (pošt-e gūla, pošt-e kūjī, bača-gord; Figure 66b), which is inserted under alternate warps; the remaining warps are attached by means of loops, called leashes, to a heddle rod (šemša, vard, gūla, gord, gurt, kūjī; Figure 66c), a bar supported or suspended at each end and extending the full width of the loom. During weaving the shed stick is first turned on its edge, separating and forcing up alternate warps, and the weft is passed through the resulting shed (pūd-e ru@; Figure 66d). For the return pass the warps must be reversed: The shed stick is turned flat, closing the shed, and the heddle rod is manipulated to raise the second set of warps, creating the countershed (pūd-e zīr), through which the weft is returned.

In Persia looms may be set up either horizontally or vertically. The ends of horizontal looms (rū-zamīnī) are usually pegged to the ground, and sometimes the sides are also supported. They can be quickly dismantled and easily transported and are thus favored by nomadic peoples. More commonly used is the vertical (dīvārī) loom, the upper and lower beams of which are either linked by two upright poles or posts or are fitted into holes in the side walls of the workroom (kārgāh).

The complete weaving process consists of several stages, each demanding different skills. In the commercial setting a highly specialized hierarchy of craftsmen is linked contractually to the manufacturer, or loom owner. Generally weavers in supervised workshops receive wages according to the hours worked and the carpet units woven; cottage weavers usually receive an advance against the price of the completed carpet (Bonine, pp. 167-68; Costello, pp. 125-26; Ehlers, pp. 291-92; English, pp. 92-94, pl. 8a and Appendix B; Gleadowe-Newcomen, pp. 93-94; Ittig, 1985, pp. 118-21 and figs. 1, 5, 8; Parhām, n.p.). The first procedure is to lay the warps on the loom, which may be accomplished either by a professional warp winder (tārbanda, čellakaš) or by the weavers (qālī-bāfān) themselves. When a specific knotting density is required, the upper and lower beams are divided into gerehs to ensure uniform placement and number of warps in each unit across the loom. After preparation of the warp several shoots of weft are inserted to create a protective strip of plain weave (gelīm). Balls (gorūk) of yarn (ḵāma) in the colors required for the pile are suspended from the upper beam. The first row of knots is then tied to the warps across the width of the carpet. In some regions a knife with a hooked point (gollāb) is used for tying knots. After each row of knots one or more shoots of weft are passed to hold them in place. The successive rows of knots and wefts are compacted with a comb beater (šāna, dastak, daftī, daftīn). In the commercial workshop a head weaver (ostād), who may also be the loom owner, is assisted in the actual weaving by one or more apprentices (šāgerd).

The sides (kenār), the edges of a carpet parallel to the warps, may be finished either by overcasting (sar-dūzī) a single cord of warp yarn (Figure 67a) or by wrapping two or more cords or warps in a circular fashion to produce a selvedge (kenārapīč; see Figure 67b). The yarns used in overcasting or selvedges may be of wool or another fiber. The upper and lower ends of a rug (sar-e farš; tah-e farš), that is, the edges running parallel to the weft, are usually finished in plain weave (Figure 67); they may have a simple warp fringe, or the warp ends may be braided, knotted, netted, or otherwise embellished. Fringes (rīša; ḥāšīa) may also be added.

Many carpets, notably those of village or nomadic origin with traditional designs, are woven without a particular pattern guide. The weaver may also work from a cartoon (naqša), in which the design is executed on a paper grid, each square (ḵāna) representing one knot (see iv, below). Alternatively, a knotted “sampler” (vāgīra), in which a section or sections of the field and border are depicted, may be used as a model (see iv, below).

Although preliminary trimming (čīdan) of the pile occurs during the weaving process, the final clipping (pardāḵt) is undertaken by an experienced finisher (pardāḵṭčī). Both offset scissors (qeyčī) and a sharp trimming knife (kārdak) are used to reduce the pile. After the completed carpet is removed from the loom it is washed. In addition to detergent, a variety of chemicals may be added to the bath in order to alter the rug’s colors according to market demands (Edwards, p. 30, 140 n. 2, 142).

Structure. Structure must be taken into account, along with stylistic features, documented inscriptions, and archival and historical sources, in classifying Persian carpets. The type, spin, ply, and diameter of the fibers; the number of weft shoots and degree of weft tension; and the knot type and knotting density are all factors that influence carpet structure. For example, quite different structures can be produced by varying weft tension because the degree of weft tension determines the extent to which alternate warps will be displaced. If none of the wefts is pulled taut, the warps will lie on the level (taḵt), and both nodes of each knot will be visible on the back of the rug. If the warps are closely laid and the tension of one weft is greater than that of the next, alternate warps will be displaced, or depressed, creating two different levels of warps; on the back of the rug one node of each knot will be at least partially concealed. In the former structure the warps are commonly referred to as level or flat and in the latter as depressed.

The structures of Persian carpets are extremely varied; in general each is associated with a specific carpet-producing region or group. One distinctive structure that has been described in the literature was originally associated with various designs in which vases figure and is therefore often called the “vase technique.” It consists of a particular type of triple-wefted weave with asymmetrical knots open to the left and is found in certain pieces attributed to Safavid production (Beattie, pp. 14-15; Martin, 1908, p. 76; see ix, below). Today triple-wefted carpets are produced in both the Kermān and Bījār areas, but, because of variations in other components of their weaves, quite different structures result (Azadi, nos. 29-31; Edwards, p. 205; Ittig, 1981, table I; 1985, table I).

Bibliography : S. Āzādī, Farš-e Īrān/Persian Carpets, Hamburg, 1978. M. H. Beattie, The Carpets of Central Persia, London, 1976. M. E. Bonine, Yazd and Its Hinterland, Marburg, 1980. D. K. Burnham, Warp and Weft. A Textile Terminology, Toronto, 1980. S. Churchill, “The Present State of the Carpet Industry in Persia,” in C. Purdon Clarke, ed., Oriental Carpets I, Vienna, 1892, pp. 1-6. V. F. Costello, Kashan. A City and Region of Iran, London and New York, 1976. A. C. Edwards, The Persian Carpet. A Survey of the Carpet-Weaving Industry of Persia, London, 1953. E. Ehlers, “City and Hinterland in Iran. The Example of Tabas/Khorassan,” Tijdschrift voor economische en sociale geografie 68, 1977, pp. 284-96. I. Emery, The Primary Structures of Fabrics, Washington, D.C., 1966. P. W. English, City and Village in Iran. Settlement and Economy in the Kirman Basin, Madison, Wisc., 1966. A. H. Gleadowe-Newcomen, Report on the Commercial Mission to South Eastern Persia for 1904-05, Calcutta, 1905 (United Kingdom, Foreign Office 368.38, file 3938). A. Ittig, “A Group of Inscribed Carpets from Persian Kurdistan,” Halı 4/2, 1981, pp. 124­-27. Idem, “The Kirmani Boom. A Study in Carpet Entrepreneurship,” Oriental Carpet and Textile Studies 1, 1985, pp. 111-23. F. R. Martin, A History of Oriental Carpets before 1800, Vienna, 1908. S. Parhām, Nāmāyešgāh-e naqša-ye qālī-e Kermān, Tehran, 1356 Š./1977. A. U. Pope, “The Art of Carpet Making. C. The Technique of Persian Carpet Weaving,” in Survey of Persian Art, pp. 2437-55. M. Varzī, Honar o ṣaṇʿat-e qālī dar Īrān, Tehran, 1350 Š./1971.

(Annette Ittig)

In this discussion “design” refers to the overall composition of decorative elements on a carpet; the simplest elements in designs are single motifs, which are most frequently combined in more complex units; these units in turn may be arranged in various combinations and sequences to form patterns. The terminology for carpet designs, like that for technique (see iii, above; v, below), is quite controversial, and several sets of nomenclature are currently in use. In academic studies rugs are traditionally classified on the basis of motifs and patterns, with names usually derived from the broader repertory of Islamic art, for example “arabesque” (see below). On the other hand, in trade publications carpets are commonly identified according to a place or period of manufacture, and a variety of terms, generally colloquial, are used to label designs. In the absence of a standard vocabulary, the terminology adopted here is that most widely used in scholarly publications (for an extensive list of colloquial terms, see Varzī, pp. 289-99).

Designs. The most common design format in Persian carpets is a central field (matn; Figure 68a) enclosed by a border (ḥāšīa; Figure 68b-e) of patterned stripes alternating with narrow bands of solid colors. The borders of Persian carpets generally have three or more patterned bands; the widest is called the “main stripe” (Figure 68b) and those flanking it “minor” stripes (Figure 68c). The narrow bands of solid color that usually separate the main and minor stripes are known as “guards” (Figure 68e). Carpet formats may be categorized further as either directional or nondirectional. The former include pictorial, prayer-niche (meḥrāb), and naturalistic vegetal designs intended to be viewed from a single vantage point. Nondirectional designs, on the other hand, are coherent when viewed from any angle; examples include radial designs, as well as abstract repeat patterns associated with tribal weaving (see Āzādī, 1978, no. 40; Edwards, pls. 156-57; Gans-Ruedin, pp. 474, 477, 481).

Some Persian carpet designs may be either directional or nondirectional, depending upon the nature and arrangement of their specific components. The following are particularly common types:

1. Centralized designs are balanced longitudinally and transversely around a central element, usually a medallion (toranj; Figure 68a). In medallion-and-corner (lačak toranj) designs there are a medallion in the center of the field and a quarter-medallion in each corner (e.g., Āzādī, 1978, no. 39; Gans-Ruedin, p. 452). Arguably the best-known examples of Persian carpets with centralized designs are the so-called Ardabīl carpets (Plate CIX in ix, below; see ardabīl carpet).

2. In lattice designs linear motifs intersect to form an overall grid, or lattice. Two, three, or more lattice systems may be superimposed so that their components cross at regular intervals. Although this type of design is most familiar from certain carpets in the so-called “vase technique” attributed to the Safavid period (Beattie, 1976, pp. 40-58, nos. 33-37; see ix, below), a variety of lattice designs are still woven today (e.g., Āzādī, 1978, no. 37; Edwards, pl. 218).

3. In multiple-medallion designs the field is dominated by rows of medallions, either linked or overlapping; they may be arranged either directionally or nondirectionally (e.g., Beattie, 1976, p. 43, no. 7).

Motifs. The following motifs are common in both the fields and borders of Persian carpets.

1. The medallion (toranj) is an ornament, usually curvilinear, that occurs frequently as the central motif(s) in the field design. The most common variants are ogival, circular, shield-shaped, and rhomboid. The edges are generally lobed or stepped.

2. The cartouche is an elongated ogival compartment (Figure 69). When used in a border pattern (as, e.g., in the Ardabīl carpet, see Plate CIX), a single cartouche form may be repeated, or two variant forms may alternate. Inscriptions are generally framed in cartouches.

3. The arabesque (šāḵa-ye eslīmī, lit. “patterned” or “arabesque branch”; Figure 70; see also Survey of Persian Art, figs. 755, 760, 762, 785, 787) is a stylized vegetal scroll with bifurcated leaves. In border patterns arabesques can occur in continuous scrolls, contiguous pairs, or alternately reversed contiguous pairs.

4. The bota (lit. “bush, shrub”; Figure 71a-b) resembles a leaf with a curved tip; it is familiar in the West as the “pear” or “teardrop” motif of paisley shawls (Edwards, pls. 21-22; Varzī, pp. 263-64).

5. The palmette (gol-e eslīmī, lit. “patterned” or “arabesque flower”; Figure 72); see also Survey of Persian Art, figs. 774-80) is a stylized fan-shaped motif somewhat resembling a lotus flower in section. In the most frequent type of palmette border the motif faces alternately inward and outward and is flanked by arabesques.

6. The rosette resembles a fully opened rose seen from above (Survey of Persian Art, fig. 783a-d).

7. The sickle leaf is a long, curved leaf with a serrated concave edge (Survey of Persian Art, fig. 784a-d).

8. The cloud band is an extended ribbon-like motif ultimately derived from Chinese art (Survey of Persian Art, figs. 789a-o, 791a-n).

Patterns. Among the most frequent patterns on Persian carpets are the following.

1 . Herātī (lit. “from Herāt”; Figure 73) is named for the city in which it was formerly thought to have been developed (see ix, below). The defining unit consists of a rosette framed in a rhombus with palmettes at the corners. Four sickle leaves are arranged around the whole. In Persian the pattern is also known as māhī “fish” from the resemblance of the sickle leaves to small fish. The pattern is woven throughout Persia, with local variations (Edwards, pls. 17-19).

2. Mīnāḵānī (literal meaning unknown, possibly derived from a proper name) is based on a two-plane lattice system of scrolling stems, with rosettes at the points of contact. This field pattern is particularly associated with carpets from the Varāmīn area (Edwards, pp. 42-43 and pl. 25).

3. The reciprocal-trefoil (madāḵel lit. “doorways”) pattern, common on borders, consists of two interlocking systems of stylized trefoils (Dimand and Mailey, p. 63, fig. 89; Gans-Ruedin, p. 534).

4. The S-stem (šāḵa-ye mawwāj lit. “wavy branch,” tāk lit. “vine”; Figure 74) pattern, consisting of repeated short stems in an S configuration, is usually a border pattern. The curving stem generally terminates in small vegetal motifs, with larger forms like palmettes or rosettes centered on the diagonals. S stems may be arranged in either single-, double-, or triple-plane systems. The most common single-plane systems are (i) continuous, in which S-stems are sequentially placed (Figure 74b); (ii) contiguous continuous, a sequential arrangement of linked S-stems (Figure 74c); (iii) counterpoised, with pairs of confronting S-stems (Figure 74d); (iv) continuous counterpoised, a sequential arrangement of pairs of confronting S-stems (Figure 74e; for terminology, see Beattie, 1976, pp. 20, 32). In the most common double-plane system contiguous counterpoised S-stems are interlaced with a confronting system. The so-called “crab” (ḵaṛčang) border consists of a double-plane system of interlacing continuous S-stems with stylized flame palmettes, or “crabs,” at their points of contact, and leaves, or “claws,” on their terminals (Edwards, p. 49 and pl. 33). The S-leaf is a variant of the S-stem, in which a leaf covers the entire diagonal of the stem (Varzī, fig. 14).

Šāh-ʿAbbāsī (lit. “of Shah ʿAbbās”), a repeat pattern of various types of blossoms, cloud bands, and palmettes associated with Safavid carpets, e.g., the so-called “Polonaise carpets” (Plate CXI in ix, below; Āzādī, 1978, p. 19; Edwards, p. 43). More recently this pattern has been associated with workshop production in such urban centers as Tabrīz, Kermān, Nāʾīn, and Isfahan (Edwards, pp. 36, 336, and pls. IV, 213; Parhām, pls. 6, 7a-s).

Certain patterns and designs, as well as distinctive renditions of common designs and motifs, have customarily been associated with specific weaving centers: for instance, the “brick” (ḵešt)-patterned rugs of Čahār Mahāl-e Baḵtīārī (Housego, 1978, pl. 89; Edwards, pl. 364; baḵtǰarǰ iii. carpets, p. 562) and the angular medallion-and-corner carpets of Herīs (Edwards, pls. 42, 68-69). Today, especially in weavings for export, such distinctions have been blurred, as pile-carpet manufacturers in Iran, Turkey, Egypt, India, China, eastern Europe, and other areas compete to satisfy the market for particular designs and palettes. For example, the label “Indo-Kirman” is applied to carpets woven in India in a floral design derived from a provincial Persian prototype (see xi, below).

Cartoons (naqša) and samplers (vā-gīra). As noted in iii, above, many carpets are woven without specific models, especially those with traditional designs that have not been produced in manufactories. The weaver may also use a cartoon in which the design (ṭarḥ) has been drawn on graph paper, usually to full scale, and colored; each square (ḵāna lit. “house”) represents one knot. The spacing of the squares serves as a guide to knotting density. Before printed graph paper was available designs were drawn, then painted, and finally divided by superimposed grid lines. Cartoons are prepared by specialized designers (naqqāš; Edwards, p. 206; Parhām, n.p.). Once completed, a cartoon is cut into horizontal sections that can be conveniently placed on the loom in front of the weaver. The knotting sequence on the cartoon may also be called out to the weaver either by a professional design caller (naqšaḵᵛān) or the master (ostād) of the workshop.

Alternatively, the weaver may use a knotted sampler (vā-gīra) in which a section or sections of the field and border designs have been woven according to a cartoon. Although initial preparation of the vā-gīra requires more time and labor than does that of a paper pattern, it provides a more effective representation of a particular choice of colors and design before weaving actually begins, so that costly and time-consuming errors can be avoided. It is also more durable than a paper cartoon. The vā-gīra is noted in descriptions of late 13th/19th- and early 14th/20th-century carpet production, and it has been suggested that it was introduced by Western exporters at that time (Edwards, p. 125). It is quite possible, however, that such carpet samplers were in use earlier.

Today the patterns woven in urban carpet workshops in Persia are usually copied directly from cartoons. The cartoon also plays a vital role in cottage carpet weaving, as it allows an entrepreneur to place orders for a specific pattern with any number of individual loom owners. As a single loom owner may be simultaneously under contract to more than one manufacturing firm, however, it is virtually impossible to copyright designs at the cottage level, and many incidents of design piracy have been recorded in correspondence from foreign manufacturing firms in Iran to their respective consuls (e.g., U.K. Foreign Office 248.1072, 11/21/13).

Designers (naqqāš). The importance of the designer in the hierarchy of artisans involved in carpet production is reflected in the occasional inclusion of designers’ names in custom-woven carpets (Ittig, 1985, no. 5); generally, carpet inscriptions mention only patrons and/or loom owners (ostād), rather than actual craftsmen. Design training has traditionally been conducted according to the apprenticeship system (Parhām, n.p.; Edwards, p. 207); it was not unusual for the skill to be passed down from father to son (Parhām, n.p.; Edwards, p. 207). A designer might work either for a specific firm or on commission, supplying designs to more than one manufacturer (Edwards, pp. 308, 337). During the late 13th/19th and early 14th/20th centuries, a “boom” period for Persian carpets, competition for designers was keen; for example, in Kermān one Āqā Moḥammad, a designer under contract to the Eastern Rug and Trading Company, was lured by higher wages into the employ of the German firm Persische Teppichgesellschaft AG (PETAG; U.K Foreign Office 248.1072, 9/8/13).

Design and provenience. No extant example of a knotted-pile carpet that can be firmly attributed to Persian manufacture before the 10th/16th century is known (see vi-viii, below). Early geographers, historians, and travelers provided only cursory descriptions of carpet designs and manufacture (for an annotated compilation of historical references to trade and production of textiles, including carpets, into the 8th/14th century, see Serjeant). Perhaps the most famous pre-Safavid carpets are the legendary Bahārestān carpet (see bahār-e kesrā), or “Spring of Kosrow,” and the Pazyryk carpet (see vi, below). The former, if in fact it was a woven carpet, was probably of flat-woven, rather than pile, construction (Edwards, p. 2). Recent archeological evidence suggests that the Pazyryk carpet, on the other hand, may not be of Persian origin but a product of Central Asia or Siberia (Kawami, pp. 16-17 and nn. 64-66).

Although the chronology of Anatolian carpets in the 9th-12th/15th-18th centuries is based largely upon comparison with representations in European painting (Mills, pp. 10-33), similar methods are less useful in dating carpets attributed to Persian manufacture (see i, above). Persian carpets, usually of the “Herat” or “Indo-Persian” group, appear in the 11th/17th century, in the paintings of van Dyck, Rubens, and Vermeer, and other European artists, but for earlier periods scholars have traditionally relied for dating on stylistic comparisons with architectural ornament, bookbindings (q.v.), manuscript illuminations, and depictions of floor coverings and textiles in miniature paintings (see vii, viii, below).

Among the earliest visual evidence for Persian carpets are the depictions of floor coverings in Timurid miniature paintings, where the predominant field designs are small repeat patterns of squares, hexagons, octagons, circles, and crosses; borders often contain repetitious “pseudo-Kufic” designs (Briggs, 1940; for a critique of Briggs’ method, see viii, below). Unfortunately, these paintings give no indication of technique and may represent flat weaves, pile carpets, or even, in some instances, felts. Furthermore, extant Anatolian, Spanish, and Mamluk rugs, as well as carpets depicted in contemporary European paintings, attest that an “octagon” style was in vogue internationally at that time (King, pp. 14-21, pls. 7-12, 14-16, 22-25; Mills, pp. 10-26, pls. 3, 5-7, 11-12). The floor covering that most closely resembles in design those in Timurid paintings is the zīlū, a sturdy cotton flat weave with a double-cloth structure (Wulff, Crafts, pp. 210-11; Beattie, 1981, p. 171, figs. 2, 4, 6-7; see v, below). The date on a fragmentary zīlū with a waqf inscription found in the congregational mosque at Maybod, near Yazd, has been interpreted as 808/1405 (Afšār, pls. 44-45); if that is correct, this fragment would be of Timurid or Turkman manufacture, the oldest carpet firmly attributable to Persian looms.

At present there are only five published pile carpets that can be assigned a Safavid provenience on the basis of inwoven inscriptions and dates (Ittig, 1984, no. 26); the two Ardabīl carpets of 942/1535-36 (see Plate CIX), the Milan hunting carpet (see Plate CVIII) of 929/1522-23 or 949/1542-43, a fragment with a multiple medallion field design in Sarajevo dated 1066/1655-56 (Survey of Persian Art, pl. 1238), and a shaped silk carpet with a directional design of cypresses and flowering plants in Qom dated 1082/1662 (Survey of Persian Art, pl. 1258B). In both their field designs and their predominantly curvilinear motifs, these pieces differ significantly from the repeat patterns of small geometric motifs that scholars have traditionally associated with Timurid carpets, though it is now recognized that “Safavid” features also occurred on floor coverings in Timurid paintings (see viii, below). Interestingly, most of the floor coverings depicted in miniature paintings ascribed to the Safavid period have centralized designs and curvilinear motifs, but none of those illustrated exactly correspond to any extant carpets considered to be of Safavid manufacture.

Carpets attributed to this period are characteristically filled with a variety of stylized floral and foliate motifs, many classifiable as chinoiserie, including lotus and leaf palmettes, split-leaf arabesques, scrolling vines, and blossoms, as well as real and fantastic animals, cloud bands, and cloud knots. Human and animal figures and tableaux from Persian literature also occur occasionally. Despite the wide range of decorative devices employed, these carpets are normally strongly symmetrical on the longitudinal axis and often transversely as well. The terms used to distinguish groups of Safavid carpets, for example, Polonaise, Sanguszko, and Herat, were devised during the early days of carpet studies. Today, a century later, they remain in use as labels of convenience, for no satisfactory classification based on place of manufacture has yet been developed (see ix, below). It is generally thought that the designs for Safavid carpets were produced by court craftsmen, who were also responsible for designs for the arts of the book and for painted architectural decoration. Yet such assumptions must be treated with caution, especially as the development of miniature painting itself is not yet entirely clear; and miniatures formerly thought to have been produced under royal patronage in court workshops have been reattributed to other settings (e.g., Simpson, p. 97). Furthermore, there are no extant carpets known to have been ordered specifically for the Safavid court, though there is contemporary textual evidence that carpets of various other kinds—nomadic, village, and urban, including both commercial weaving and weaving to order—were manufactured in various regions, as they are today (Keyvānī, p. 237; Chardin, VII, pp. 329-34; Du Mans, p. 187; Mańkowski, p. 2433; Thevenot, p. 77). Finally, the assertion that “the golden age of the (classical) class doubtlessly falls within the first half of the sixteenth century” (Bode and Kuhnel, p. 87) is unproven: at present, only three dated pieces (see above) support this statement. The traditional chronology and classifications of Safavid carpets therefore require careful reexamination. More recent studies have stressed structural as well as design features in attributing provenience (e.g., Beattie, 1976, passim); closer attention to primary sources may also provide clearer understanding of the historical context of carpet manufacture.

As few securely dated 12th/18th-century carpets are presently known, only the most general hypotheses about their designs are possible. A preliminary survey suggests that Safavid designs and patterns continued to be woven, albeit with less fluidity. New motifs, like the bota, were introduced later in the century (Housego, 1986, pp. 40-51, figs. 3, 9; see x, below).

The volume of objects and documentary sources relating to carpet production and trade in the 19th century permit a number of general observations on Qajar carpet designs. First, various designs and motifs known from the Safavid and succeeding periods continued in use; in addition, such small repeat patterns as the herātī, mīnāḵānī, and offset bota became established in the repertory, with distinctive regional variations. Furthermore, as part of a contemporary fashion for European imagery in the Persian arts generally (art in iran x. qajar art, p. 631), Western pictorial themes were also included in carpets. As the economic importance of the export market increased in the late 13th/19th and early 14th/20th century, widespread changes in traditional sizes, palettes, and designs were introduced in order to meet the requirements of Western consumers (see xi, below). Finally, it is only from this period that significant quantities of so-called “nomadic,” or “tribal,” carpets survive, with their (usually) bolder geometric patterning (e.g., Housego, 1978, passim), in marked contrast to the types of designs described here. Similarities between certain tribal ornaments, especially the Turkman göl (Pers. gol “flower”), and motifs in floor coverings in Timurid miniature paintings suggest that at least some nomadic carpet patterns represent survivals from the pre-Safavid repertory (Denny, pp. 331-34; Pinner and Franses, p. 88; see viii, below).

Since the Qajar period the designs of Persian carpets have continued to be modified to meet demands of the international market, for example, production of the so-called “American Kirmans” (Edwards, pp. 207-09; see xii, below). As traditional Persian motifs and patterns are increasingly copied in centers outside Iran, design must now be considered more indicative of the targeted market than of the place of origin.

Bibliography : Ī. Afšār, Yādgārhā-ye Yazd I, Tehran, 1348 Š./1969. S. Āzādī, Fārš-e Īrān/Persian Carpets, Hamburg, 1978. Idem, “Baluchistan v. Baluch Carpets, Rugs and Other Products,” in EIr. III/6, pp. 645-46. M. H. Beattie, Carpets of Central Persia with Special Reference to Rugs of Kirman, Sheffield, 1976. Idem, “A Note on Zilu,” in C. Cootner, ed., Flatwoven Textiles, Washington, 1981, pp. 168-73. Idem, “Ardabīl Carpet,” in EIr. II/4, pp. 365-68. W. von Bode and E. Kühnel, Antique Rugs from the Near East, 4th ed., tr. C. G. Ellis, Berlin, 1958. A. Briggs, “Timurid Carpets,” Ars Islamica 7, 1940, pp. 20-54; 11-12, 1946, pp. 146-58. I. Chardin, Yoyages du chevalier Chardin en Perse VII, Paris, 1811. W. B. Denny, “Turkmen Rugs and Early Rug Weaving in the Western Islamic World,” Halı 4/4, 1982, pp. 329-37. M. S. Dimand and J. Mailey, Oriental Rugs in The Metropolitan Museum of Art, New York, 1973. R. Du Mans, Estat de la Perse en 1610, ed. Ch. Schefer, Paris, 1890. A. C. Edwards, The Persian Carpet, 1953. K. Erdmann, Oriental Carpets, tr. C. G. Ellis, New York, 1960. E. Gans-Ruedin, Iranian Carpets. Art, Craft, and History, London, 1978. J. Housego, Tribal Rugs, London, 1978. Idem, “18th Century Persian Carpets. Continuity and Change,” in Oriental Carpet and Textile Studies 1, 1986, pp. 40-51. A. Ittig, “A Group of Inscribed Carpets from Persian Kurdistan,” Halı 4/2, 1981, pp. 124-27. Idem, A Technical and Historical Study of the Qajar Carpet Industry, Ph.D. dissertation, Oxford University, 1983. Idem, “The Persian Carpet. Previous Studies and Future Directions,” Oriental Rug Review 4/7, 1984, pp. 1-4. Idem, “The Kirmani Carpet Boom—A Study in Carpet Entrepreneurship,” in Oriental Carpet and Textile Studies 1, 1985, pp. 111-23. Idem, “Baḵtīārī iii. Carpets,” in EIr. III/6, pp. 561-62. T. Kawami, “Archaeological Evidence for Textiles in Pre-Islamic Iran,” Iranian Studies 22/2-3, 1989, pp. 1-22. M. Keyvānī, Artisans and Guild Life in the Later Safavid Period, Berlin, 1982. D. King, ed., The Eastern Carpet in the Western World from the 15th to the 17th Century, London, 1983. J. Lessing, Ancient Carpet Patterns after Originals of the 15th and 16th Centuries, London, 1879. T. Mańkowski, “Some Documents from Polish Sources Relating to Carpet Making in the Time of Shah ʿAbbās I,” in Survey of Persian Art, pp. 2431-36. F. R. Martin, A History of Oriental Carpets before 1800, Vienna, 1908. J. Mills, Carpets in Pictures, London, 1975. S. Parhām, Namāyešgāh-e naqša-ye qālī-e Kermān, Tehran, 1356 Š./1978. R. Pinner and M. Franses, “Two "Turkoman" Carpets of the 15th Century,” in R. Pinner and M. Franses, eds., Turkoman Studies I, London, 1980, pp. 83-89. A. U. Pope, “The Art of Carpet-Making. A. History,” in Survey of Persian Art, pp. 2257-2430. B. W. Robinson, “Art in Iran x. Qajar 2. Painting,” in EIr. II/6, pp. 637-40. J. Scarce, “Art in Iran x. Qajar 1. General” in EIr. II/6, pp. 627-37. R. B. Serjeant, Islamic Textiles, Material for a History up to the Mongol Conquest, Beirut, 1972. M. S. Simpson, “The Production and Patronage of the Haft Aurang by Jami in the Freer Gallery,” Ars Orientalis 13, 1982, pp. 92-119. F. Spuhler, “Carpets and Textiles,” in Camb. Hist. Iran VI, pp. 698-727. J. Thévenot, Travels of Monsieur de Thevenot into the Levant, pt. 2, London, 1686. U.K. Foreign Office Records, Public Records Office, London, no. 248.1072, Sept. 8, 1913, Nov. 21, 1913. M. Varzī, Honar o ṣaṇʿat-e qālī dar Īrān, Tehran, 1350 Š./1971.

(Annette Ittig)

Most of the structures in Persian flat-woven carpets belong to the category called “interlacing” by textile specialists; the term designates the most straightforward way in which each thread of a fabric passes under or over threads that cross its path (Emery, pp. 62, 74-178). Interlacing can involve one or more sets of warps (longitudinal elements; čella, tār) and welts (transverse elements; pūd). Persian flat-weave structures range from simple interlacing—one set of warps and one set of wefts in various forms of plain weave (see below)—to a variety of compound weaves involving additional sets of warps or wefts or both. Patterning and decoration may also be achieved through various flat-weave techniques, including the introduction of supplementary warps, wefts, or both. Occasionally small motifs are worked in a supplementary-weft technique known as knotted and cut pile (see iii, above). The other two basic categories of woven textiles are represented in Persian flat-woven carpets by weft twining (Emery, pp. 196, 200-01) and weft wrapping (Emery, pp. 214-21, 225). A flat-woven carpet is made in either one piece or, if the loom is narrow, in two lengths, the long sides of which are then stitched together, leaving a seam down the middle of the carpet.

The looms used for flat-woven carpets are of the same basic types as those used for pile carpets (see iii, above). Nomads have traditionally used horizontal ground looms (rū-zamīnī), which are easy to dismantle for the biannual migrations. Certain nomadic groups do use vertical looms (dārī, dīvārī), however; for example, weavers of the Šāhsevan, whose summer quarters are on high mountain slopes where the climate is cold, work inside tents, where space is limited and the vertical loom thus more practical. Villagers tend to use vertical looms fixed permanently in place, but recently settled nomads, including some Baḵtīārī, favor their traditional horizontal looms, even in village houses.

Vertical looms have also been used by urban weavers of flat-woven rugs. From the Iranian workshops that produced mainly pile carpets for the court and commerce the only flat-woven carpets known to date from the “classical” period (10-11th/16-17th centuries) are silk tapestry-woven examples with brocaded (supplementary; see below) decoration in silk threads wrapped with thin strips of silver or gold. These carpets, together with certain silk pile carpets brocaded in a similar way, were products of Kāšān, mainly during the reign of Shah ʿAbbās I (996-1038/1588-1629), and are known as “Polonaise” (Beattie, 1972, pp. 31-37 and pl. III; Dimand and Mailey, pp. 59-67; for illustrations of two tapestry-woven examples, see Bier, nos. 40, 49; see also ix, below).

A wide variety of flat-weave structures have been used by nomads and villagers throughout Persia, both in carpets and in such other items as storage bags, horse trappings, and tent decorations (see xvi, below; for discussion of techniques, diagrams of structures, and illustrations of flat weaves, see Black and Loveless; Cootner; Enderlein; Ford and Pohl-Schillings; Wertime, 1979; idem, 1981; Housego; Petsopoulos; Tanavoli).

Plain weaves. The most basic type of interlacing of warp and weft is called plain weave (for structural terminology, definitions, and diagrams, see Emery; Burnham). It is created by passing successive rows of wefts alternately under and over single warps. If the warps and wefts are equally spaced and of virtually the same thickness and flexibility, the structure is called balanced plain weave (Figure 75). If the wefts are tightly packed so that the warps are hidden or almost hidden, the structure is called weft-faced plain weave (Figure 76), or tapestry weave. In Persia this structure is called gelīm (Turk. kilim). Conversely, if the warps are packed so closely that they completely cover the wefts, the weave structure is called warp-faced plain weave (jājīm; Figure 77).

In tapestry weave the wefts may be continuous (carried from edge to edge) or discontinuous (turned back around adjacent warps part way across the work as required by the pattern). Discontinuous wefts may be manipulated in various ways in order to produce different effects, sometimes in a single weaving; certain of these effects can be said to characterize particular regions or tribal groups (see xvi, below). Discontinuous wefts may either interact or not. In the latter instance, where discontinuous wefts are turned back at the boundary between two colors, the adjacent warps are not bound together and a slit is left between them; this weave structure is thus called slit-tapestry weave (Figure 78; in Figures 78-82 the weft-faced tapestry weaves have been schematized in order to clarify the structures, whereas in actual weaving the wefts would be tightly packed together). The slits may be sewn up, but in Persian weaving they usually are not. The boundary between two colors can also be dovetailed (Figure 79). Dovetailing is a shared-warp technique, in which two discontinuous wefts from adjacent color areas in the pattern are wrapped around a common warp before being returned to their own color areas. A third way of handling the meeting of discontinuous wefts of different colors is single or double interlocking, in which the wefts from adjoining color areas are wrapped around each other between two adjacent warps; it produces a stronger and neater join than does slit weave or dovetailing. In single interlocking (Figure 80) each discontinuous weft is wrapped around one corresponding weft from the adjacent color area. In double interlocking the discontinuous wefts are wrapped around each other in pairs, forming an even closer join; where the two color areas meet, the face of the piece is smooth (Figure 81), but a ridge is formed on the back (Figure 82). Double interlocking is particularly characteristic of Baḵtīārī tapestry-woven pieces (see baḵtǰārǰ iii. carpets). Tapestry-woven carpets characterized by slit weave, dovetailing, or single interlocking are reversible and will look the same on both faces, unless supplementary decoration has been woven in.

The techniques discussed so far were also used in European tapestries; the Polish Carmelite father Paul Simon thus used the words “tissues of the Arras kind” (referring to the famous Franco-Flemish tapestry-weaving center the name of which had become virtually a generic term for tapestries in Europe) to describe tapestry-woven carpets that he saw when he visited Kāšān in 1016/1608 (Dimand and Mailey, p. 59).

Float weaves. Another group of structures involving only one set of warps and one set of wefts is known as float weaves. They are characterized by warps or wefts that, instead of simply being passed alternately under one and over one of the opposite element, as in plain weave, are passed over and under more than one element at a time. The longer spans are known as floats. One type of float weave is twill weave (Emery, pp. 75, 92-107), which occurs in some Persian tribal and village pieces, sometimes in combination with other structures used for decoration; it is characterized by a diagonal alignment of floats (see Figure 83). The diagonals—forming herringbone, chevron, or lozenge patterns—result when the weft floats, instead of floating in alternate passes over the same warps (alternation), are offset one warp beyond the float in the previous row (progression). Figure 83 illustrates a plain twill weave found in some nomadic carpets, in which each weft is passed over two and under two warps (expressed as a ratio, 2:2). Different effects are achieved by different ratios of float spans. The direction of the diagonal in twill weave is characterized as either “S” or “Z” (according to the direction in which the middle section of each letter is slanted). The direction of the diagonal in Figure 83, for example, is that of S, and the twill ratio and diagonal direction can be described as 2S2 or, alternatively, 2$ 2 (2Z2 may also be written 2/2). On the reverse face of twill weave the diagonal is in the opposite direction.

A type of warp-float weave derived from plain weave (Emery, pp. 113-21) and favored by the Šāhsevan (Tanavoli, p. 72) is based on both alternation and progression, rather than simply the latter, as in twill weave. In the example illustrated in Figure 84 the wefts are passed under three warps and over one warp (3:1) in alternate rows and under one and over one (1:1) in the intervening rows (a form of alternation); in addition, the pass of the weft under three warps is offset one warp in successive rows of 3:1 (progression). The warps in this variant float on the front of the fabric (the face seen in Figure 84); the reverse side shows weft floats.

Compound weaves. In addition to weft wrapping with a ground weft and to extra-weft contour wrapping (see below), nomadic and village weavers use a number of other compound weaves, which may be either complementary or supplementary (Burnham, pp. 29, 138, 174, 184; Emery, pp.74, 140-54; Wertime, 1981, pp.195-96; Tanavoli, pp. 73-77). In contrast to plain weave (whether balanced, weft faced, or warp faced), in which there is only one set of warps and one set of wefts in the foundation, compound weaves consist of more than one set of either warps or wefts or both. When the additional sets are integral to the foundation structure they are said to be complementary; the resulting fabric may or may not be reversible, and the areas of color or floats forming the pattern may be identical or reversed on the two faces. Supplementary warps and wefts are not integral to the foundation structure but are added only for purposes of design. Either supplementary or complementary warps and wefts may be used to create a pattern or enrich a ground weave. The resulting fabric is described as either warp patterned or weft patterned.

The compound weaves used in nomadic and village carpets include supplementary weft-float brocading, or patterning (Figure 85, Figure 86), which may be used on either a weft-faced plain-weave ground (gelīm) or a warp-faced plain-weave ground (jājīm). The weft floats may be continuous (from selvage to selvage) or discontinuous (used in more widely spaced areas of the piece), and they may span variable numbers of warps.

In complementary compound weaves the second set of warps or wefts can be used over the entire face of the fabric or limited to decorative stripes (vertical, in the direction of the warp) or bands (horizontal, in the direction of the weft). Threads of contrasting colors may either be substituted in pattern areas on the face (Figure 87) while the unused elements float on the back, or they may be used more equally throughout the structure. In either instance the front and back of the fabric are similar, but the colors are reciprocally reversed on each face.

Weft twining. Another type of compound structure, used to decorate ground weaves, is weft twining (Emery, pp. 196-205; Tanavoli, p. 78), in which pairs of wefts in contrasting colors are twined around each other, enclosing successive warps, to produce a row of diagonal “strokes.” Weft twining may be worked in a single direction (Figure 88) or countered (Figure 89).

Weft wrapping. Nomadic and village weavers employ a variety of weft-wrapping techniques, in which the weft encircles the warps, usually two or more at a time. Weft wrapping may cover the entire surface of a woven piece (Emery, p. 225), or it may be an added decorative feature on tapestry-woven gelīms. In some pieces, when weft wrapping covers the entire surface, a simple ground weft (1:1) may be passed between two rows of weft wrapping, which is then considered supplementary to the ground weft (Figure 90); such supplementary weft wrapping is also known as “extra-weft wrapping.” Persian flat weaves composed entirely of weft wrapping, with or without ground wefts, are often called “sumak,” a word of obscure origin. Weft wrapping can be handled in such a way that it creates a horizontal, diagonal, or vertical line (Emery, pp. 214-21; Tanavoli, pp. 80-88; Wertime, 1981, pp. 175-91, 197-99; Acar, pp. 69-75). The precise wrapping sequence can be designated by the numbers of warps or paired warps the weft spans as it encircles the warps: forward over (Fo), forward under (Fu), back over (Bo), or back under (Bu). The diagram in Figure 90 shows weft wrapping in a sequence Fo4 Bu2. As wrapping progresses, a row of slightly oblique spans is produced on the front surface of the piece. They may all incline in one direction, which is known as “plain sumak” (Figure 90), or they may be countered (the direction of slant reversed) in alternate rows to produce a herringbone effect, known as “countered sumak” (Figure 91). A form of weft wrapping known as “reverse sumak,” in which the shorter spans of the weft wrapping are seen on the front of the work and the longer spans on the back (Figure 92, showing a sequence Bo2 Fu4 Bo2), is also found both as a ground weave and as a decorative feature in some Persian flat-woven pieces, particularly those of the Šāhsevan of northwestern Iran. The shorter spans tend to be less oblique than the longer spans, which lends reverse sumak a characteristic appearance distinct from that of plain or countered sumak.

Diagonal extra-weft wrapping can be used as decoration on a plain-woven ground, producing a kind of compound-weave structure. It is used particularly to outline motifs worked in a color that contrasts with that of the adjacent background. In Figure 93 this structure is illustrated in a sequence Fo4 Bu2, Fo4 Bu3, Fo4 Bu2, which creates a diagonal line on the face of a plain-woven ground.

Zīlū carpets. An entirely different type of flat-woven carpet made in Persia is an inexpensive sturdy floor covering known as zīlū, which is usually made of cotton, though occasionally in the past wool was also used. It is woven on a special loom called dastgāh-e zīlū-bāfī and is produced almost entirely for local use, rather than for export.

The structure of zīlū carpets is a compound weave, a particular kind of weft-faced plain weave with two sets of warps and pairs of wefts that are complementary to each other (see Figure 94). This structure (known by the commonly used French term taqueté) characterizes the entire carpet, rather than being limited to specific areas. Each of the two sets of warps has a different and essential function. One set of warps is known as the main, or binding, warps; it interlaces with the wefts to create the foundation of the fabric. The other warps, called inner warps (the dark vertical threads in Figure 94), control the pattern, by determining on which side of the fabric the wefts will appear; the inner warps themselves cannot be seen on either surface. This set of inner warps distinguishes zīlū carpets from other Persian flat-woven carpets. The two sets of wefts, which are continuous and worked in pairs to form the pattern, are complementary to each other; they are usually of different colors. When a weft of one color appears on one surface, a weft of the other color appears in the same area on the opposite surface. Zīlū carpets are thus reversible (or, more precisely, counterfaced) when the wefts are of only two colors; that is, the colors are reversed in the same pattern area on each surface, though a woven inscription will appear in mirror image on the opposite surface. In the plain-weave version of this structure (see Figure 94) the wefts (on both sides of the fabric) float over three warps (one inner, one binding, one inner) and then go under (are bound by) one binding warp before the next float span over three warps.

According to literary sources, carpets known as zīlū were woven in Persia at least as early as the 4th/10th century (see vii, below), but it is not certain that they were woven in the same manner and materials as modern zīlū carpets. The earliest dated example of the latter type was woven in 807/1405 (Beattie, 1981, pp. 169-74; Wilber; Mackie). Today zīlū carpets are commonly found in mosques, homes, hotel rooms, offices, and shops. Some pieces, particularly those given to mosques, bear inscriptions and dates. The region around Yazd, notably the villages of Maybod and Ardakān, has long been known for the production of these carpets.

Finishing. The edges of flat-woven carpets parallel to the warp, like those of pile carpets, may be finished either by overcasting with the ground weft or by wrapping an extra weft over two or more pairs of warps either parallel or in a figure eight. If the figure eight is wrapped over more than two pairs of warps, an additional extra weft in a contrasting color may be wrapped vertically over part of the selvage as further decoration. The variety of techniques used to finish the upper and lower ends of flat-woven carpets (Tanavoli, pp. 97-102) fall into two main categories. Either a hem may be created by turning under a few centimeters of the fabric and sewing it down, or the loose warp ends may be formed into a fringe. Fringes may be formed in a variety of ways:

1. Groups of warp ends may be knotted together (Figure 95).

2. Alternate groups of warps may he knotted together in successive rows to create a net or web (Figure 96).

3. Warp threads may be interlooped singly or in groups in a series of three to five chain-like rows (Figure 97).

4. Groups of warps may be braided in flat vertical plaits, each about 2 cm wide (Figure 98).

5. The ends of the warp threads only may be diagonally plaited in a structure like plain weave to create a band about 2 cm wide extending across the width of the carpet; the ends may then be knotted in a simple fringe (Figure 99).

6. Warp threads may be twisted together in pairs, often with the ends left as uncut loops (Figure 100).

7. Groups of warps may be twisted together (Figure 101).

Bibliography : B. B. Acar, Kilim-Cicim, Zili-Sumak. Turkish Flatweaves, Istanbul, 1983. M. H. Beattie, The Thyssen-Bornemisza Collection of Oriental Rugs, Castagnola, Switz., 1972, pp. 31-37, pl. 111. Idem, “A Note on Zilu,” in C. Cootner, The Arthur D. Jenkins Collection I: Flat-Woven Textiles, Washington, D.C., 1981, pp. 169-74. C. Bier, ed., Woven from the Soul, Spun from the Heart. Textile Arts From Safavid and Qajar Iran, 16th-19th Centuries, Washington, D.C., 1987, nos. 40, 49, 76, 100-02, 104, 105. D. Black and C. Loveless, eds., The Undiscovered Kilim, London, 1977. D. K. Burnham, Warp and Weft. A Textile Terminology, Toronto, 1980. C. Cootner, The Arthur D. Jenkins Collection I: Flat-Woven Textiles, Washington, D.C., 1981. M. S. Dimand and J. Mailey, Oriental Rugs in the Metropolitan Museum of Art, New York, 1973, pp. 59-67. I. Emery, The Primary Structures of Fabrics. An Illustrated Classification, Washington, D.C., 1966. V. Enderlein, Orientalische Kelims, Flachgewebe aus Anatolien, dem Iran und dem Kaukasus, Berlin, 1986. P. R. J. Ford and H. E. Pohl-Schillings, Persische Flachgewebe, Bilder einer Ausstellung, Cologne, 1987. A. de Franchis and J. T. Wertime, Lori and Bakhtiyari Flat-weaves, Tehran, 1976. J. Housego, Tribal Rugs, New York and London, 1978. L. W. Mackie, “Covered with Flowers. Medieval Floor Coverings Excavated at Fustat in 1980,” Oriental Carpet and Textile Studies 1, 1985, pp. 23-35. S. Parhām and S. Āzādī, Dastbāfhā-ye ʿašāyerī wa rūstāʾī-e Fārs I, Tehran, 1364 Š./1986. Y. Petsopoulos, Kilims. Flat-Woven Tapestry Rugs, New York and Fribourg, 1979. P. Tanavoli, Shahsavan. Iranian Rugs and Textiles, New York and Fribourg, 1985. C. E. C. Tattersall, Notes on Carpet Weaving, London, 1969. J. T. Wertime, “Flat-Woven Structures Found in Nomadic and Village Weavings from the Near East and Central Asia,” Textile Museum Journal, 1979, pp. 33-54. Idem, “Weft-Wrapping in Nomadic and Village Flat-Woven Textiles from the Near East and Central Asia. A Guide to Flat-Woven Structures,” in C. Cootner, The Arthur D. Jenkins Collection I: Flat-Woven Textiles, Washington, D.C., 1981, pp. 175-203. Idem et al., Tribal Animal Covers from Iran, Tehran, 1975. D. N. Wilber, “A Very Old Flat-Weave?” Halı 3/4, 1981, p. 309.

The figures are reproduced with the kind permissions of the authors and publishers:

Figures 75-77, 83-89, and 95-101 from P. Tanavoli, Shahsevan. Iranian Rugs and Textiles, New York: Rizzoli, 1985, with the author’s permission;

Figures 79 and 8l-82 from J. Housego, Tribal Rugs, London: Scorpion Publishing Ltd., 1978;

Figure 80 from D. K. Burnham, Warp and Weft. A Textile Terminology, Toronto: Royal Ontario Museum, 1980, with the author’s permission;

Figures 78 and 90-93 from B. B. Acar, Kilim-Cicim, Zili-Sumak, Turkish Flatweaves, Istanbul: Eren Yayınçılık ve Kitapçılık Ltd. Şti., 1983.

(Sarah B. Sherrill)

Since 1966, when Hans Wulff wrote “There is no field in the industrial arts in Persia that is as important as carpet weaving, and yet little is known about its early development” (Crafts, p. 212), little new information has become available. Indeed, the evidence for textiles of all kinds in pre-Islamic Iran is very sparse. It is necessary to supplement the few remains of actual textiles with examination of representations in art and other kinds of indirect evidence of production, for example preserved impressions and pseudomorphs from excavations (for a recent summary of all this evidence, see Kawami). In this discussion the focus will be on preserved textiles in general, in order to demonstrate the range of fibers and techniques in use in Iran over time and thus to provide a context for the development of carpet making.

Interworking of fibers to produce cloth was known in Iran as early as the 5th millennium b.c.; two ceramic vessels of the Dalma period (see ceramics, chalcolithic, zagros) containing infant burials, from mound B at Seh Gābī in Kurdistan, yielded impressions of textiles and a preserved fragment as well (Levine p. 180). The corroded surface of a proto-Elamite (Period IV) copper mirror found in Tomb 2 at Sialk (Sīalk) in north central Iran, dating to about 3000 b.c., preserves what appear to be shreds of S-spun linen (Pfister, pp. 201-02 and pl. XXXII/13; Ghirshman 1938, p. 64 and pl. XXIX/3). Other textiles are preserved from later 3rd-millennium contexts. At Tepe Hissar (Ḥeṣār), in northeastern Iran, two copper daggers with fragments of charred cloth in the area of the grip were found in the Burned Building of level IIIB, and another, in a IIIC grave, preserved traces of a grip of cloth and wood (Schmidt, p. 201 and pls. I, L). A copper ax from Susa had been wrapped in two pieces of cloth, of which traces were preserved. Analysis indicates that both were of linen, one coarse, the other finely woven (Lecaisne, p. 162 and suppl. pl. XLIII).

By the late 2nd millennium b.c. cording was known; the technique was later used in the selvages of carpets (see above, iii, v). In the tombs at Marlik (Mārlīk) on the southwest Caspian coast, which date from this period or perhaps the beginning of the 1st millennium b.c. (Muscarella, pp. 416-17), were found fragments of unidentified fabric, including cords and woven bands (Negahban, 1964, fig. 88 and p. 50). In addition, adhering to the surface of a bronze beaker from tomb 39 were traces of a fabric in which it may have been wrapped before burial (Negahban, 1983, p. 77, no. 50). Tomb C-I at the approximately contemporary site of Ghalekuti (Qaḷʿa-ye Kotī) in Deylamān contained a mirror with remains of hemp or flax cloth adhering to it, as well as fragmentary strings of hemp or flax that appear to have provided suspension for jewelry (Egami et al., 1965, p. 31 and pl. LXXV/88-90).

Hasanlu (Ḥasanlū), near Lake Urmia in Azerbaijan, was destroyed in an attack at the end of the 9th century b.c. (level IVB), and a relatively large sample of textiles survived the resulting fire (Dyson and Muscarella; Maude de Schauensee, personal communication). The pieces include one- and two-ply yarns, exhibiting both S and Z spin, preserved as woven fabric, fringe, tassels, and balls of thread. One example appears to have a kind of pile, with small loops occurring at intervals; its function is, however, unknown. Although the fibers have not been completely analyzed, the textiles from both Bronze and Iron Age levels include a large preponderance of wool, as well as bast; a second, unspecified vegetal fiber; and goat hair (de Schauensee, personal communication). Impressions of textiles were also preserved on a silver-and-electrum beaker; presumably it had been wrapped in fabric for storage (Porada, p. 2971, pl. 1488). Excavations in the grave of a young girl (grave 3, area P_l) at Haftavan (Haftavān) Tepe, north of Hasanlu, also of the 9th century b.c., produced the remains of a headdress, which had been made of plain-woven cloth, as well as bits of the thread that had secured decorative bronze disks to it (Burney, pp. 134-35, fig. 8, and pl. IVa).

Another piece of evidence that pile carpets may have been produced in the ancient Near East is to be found on stone slabs from the entrances of residences of the Neo-Assyrian kings, beginning with Tiglath-pileser III (745-27 b.c.), which are decorated with geometric designs that may represent such carpets (Albenda). Although Xenophon mentions that the Persian kings made use of “Median carpets” of a yielding softness, no example is preserved from Achaemenid Persia itself (Cyropaedia 5.5.7, 8.8.15-16; cf. Wulff, p. 212). The earliest well-preserved pile carpet that has been excavated was found in kurgan 5 at the site of Pazyryk in the Altai mountains of southern Siberia; it has often erroneously been identified as “Persian,” even by the excavator (Rudenko, 1970, p. 304). As kurgan 5 dates from the end of the 4th or the beginning of the 3rd century b.c., however, the carpet probably does not belong to the Achaemenid period, though it may well reflect the influence of Achaemenid carpet design. The carpet, which is preserved to nearly full size, originally measured 1.83 by 2 m (Plate CV). The field is divided into 24 framed squares, each containing a stylized floral element that may ultimately have been derived from the quatrefoils of Assyrian prototypes (Albenda, p. 3). The border consists of five stripes separated by narrow guards of alternating black, tan, and red squares outlined in black, flanked by narrow tan bands edged in black (Rudenko, 1968, pp. 40ff.; idem, 1970, pp. 298ff.). The outer and inner stripes contain schematic lion-griffins with heads turned back; the heads in the outer stripe face to the left, those of the inner stripe to the right. In the second stripe from the outside is a row of horses with crenellated manes, some ridden and some accompanied by walking male figures wearing caps; the ground is rust-colored. The walking figures resemble the tribute bearers on reliefs at Persepolis (e.g., Ghirshman, 1964, p. 184, pl. 231, p. 176 pl. 222), and it is this motif that originally suggested a Persian origin for the carpet. Near one corner of this stripe there are two circles with rosettes in the centers, which have led one scholar to infer that the carpet may have been a game board (Jettmar, p. 121). The fact that the horsemen are both mounted and on foot, in contrast to the figures on the Persepolis friezes, suggests that this border pattern was a rather distant adaptation of the Achaemenid original. (This observation was made by Judith Lerner.) The central border stripe contains stylized floral elements similar to those in the squares of the field, in rust, white, and black on a tan ground. The remaining stripe contains a frieze of reddish, spotted animals with lowered heads on a buff ground; they were identified by the excavator as fallow deer (Cervus dama; Rudenko, 1968, p. 42). The haunches are patterned with the Achaemenid “dot and comma” motif (cf. Ghirshman, 1964, p. 239, pl. 286), and decorative polychrome bands run along the backs and necks, but the relatively naturalistic body contours and pose with lowered head suggest that the carpet was made in Central Asia or Siberia, where this animal and the naturalistic representation of animals in general were indigenous (Rubinson; Farkas, pp. 24-25). The pile consists of symmetrical knots, 3,600 to the dm² (Rudenko, 1968, pp. 48-49; idem, 1970, p. 302). It should be noted that kurgan 2 at Bashadar in the Altai, dating about 100 years earlier than Pazyryk kurgan 5, yielded fragments of a much finer knotted carpet, with 7,000 asymmetrical knots to the dm² (Rudenko, 1968, p. 49; idem, 1970, p. 302). It is, however, too fragmentary to permit determination of its design or function.

No pile carpet remains are extant from Persia before the Sasanian period, though fragments of Parthian wool textiles are common. Some pieces from tombs in the Germī region of eastern Azerbaijan, near the Soviet border, are reported to be colored blue, orange, brown, and cream and to include among their design repertory swastika motifs and checkered patterns, as well as Greek-key borders (Matheson, p. 75). At Shahr-i Qumis (Šahr-e Qūmes) in Gorgān part of a black/brown felt garment, with the edge of the neckline and part of the left shoulder preserved, has been excavated. It is dated to the first half of the 1st century b.c. The silk cord that presumably secured the neck opening is the oldest excavated piece of silk known from Iran at present. Felt fragments, colored red and black, were also found in the Parthian levels at Shahr-i Qumis (Hansman and Stronach, 1970a, pp. 49-52; Nunome, pp. 341-43). Hemp cord and other textile remains from the Parthian period have been found in tombs at Naruzmahale (Nowrūz-maḥalla) on the Caspian coast (Egami et al., 1966, Japanese text pp. 14, fig. 2/1; 17, fig. 3; English tr., p. 18).

From the Sasanian period a rich legacy of textile imagery and literary references has survived (see, e.g., bahār-e kesrā), but it is only from the Sasanian levels at Shahr-i Qumis that actual Sasanian textiles are preserved (Ackerman, pp. 69lff.; Bier, pp. 119-25). They include several yellowed fragments with pile, identified as rugs, and fabrics of wool or wool and cotton, some with patterns; a fragment striped in red/orange, white, and blue; another with brown and white stripes; a white shroud ornamented with a single green thread; a piece decorated with white palmettes; and another with a blue ground and white “pearls.” Fragments of linen were also found at the site (Hansman and Stronach, 1970a, p. 40; idem, 1970b, pp. 154-55, pls. I-IV and p. 154 fig. 7; Bier, pp. 119, 123; Kawami, appendix). One of the Sasanian fragments from Shahr-i Qumis has been identified as a zīlū (Vogelsang-Eastwood), but the identification is controversial (see v, above).

Bibliography : P. Ackerman, “Textiles through the Sasanian Period,” in Survey of Persian Art I, pp. 681-715. P. Albenda, “Assyrian Carpets in Stone,” The Journal of the Ancient Near Eastern Society of Columbia University 10, 1978, pp. 1-34. C. M. Bier, “Textiles” in P. O. Harper, The Royal Hunter. Art of the Sasanian Empire, New York, 1978, pp. 119-40. C. Burney, “Excavations at Haftavan Tepe 1969. Second Preliminary Report,” Iran 10, 1972, pp. 127-42. R. H. Dyson, Jr., and O. W. Muscarella, “Constructing the Chronology and Historical Implications of Hasanlu IV,” Iran 27, 1989, pp. 1-27. N. Egami, S. Fukai, and S. Masuda, Dailaman I: The Excavations at Ghalekuti and Lasulkan, The Tokyo University Iraq-Iran Archaeological Expedition Report 6, Tokyo, 1965; II: The Excavations at Noruzmahale and Khoramrud 1960, The Tokyo University Iraq-Iran Archaeological Expedition Report 7, Tokyo, 1966. A. R. Farkas, “The Near East,” in E. C. Bunker, C. B. Chatwin, and A. R. Farkas, “Animal Style” Art from East to West, New York, 1970, pp. 19-59. R. Ghirshman, Fouilles de Sialk, près de Kashan, 2 vols., Paris, 1938-39. Idem, The Art of Ancient Iran, New York, 1964. J. Hansman and D. Stronach, “Excavations at Shahr-i Qumis, 1967,” JRAS, 1970a/1, pp. 29-62. Idem, “A Sasanian Repository at Qumis,” JRAS 1970b/2, pp. 142-55. K. Jettmar, Art of the Steppes, New York, 1967. T. Kawami, “Archaeological Evidence for Textiles in Pre-Islamic Iran,” Iranian Studies 23, 1989 (forthcoming). M. L. Lecaisne, “Note sur les tissus recouvrant des haches en cuivre,” MDFP 13, 1912, pp. 163-64 and suppl. pl. XLIII. L. D. Levine, “Seh Gabi,” Iran 10, 1972, pp. 179-80. S. Matheson, Persia. An Archaeological Guide, Park Ridge, N. J., 1973. O. W. Muscarella, “Fibulae and Chronology, Marlik and Assur,” Journal of Field Archaeology, 1984, pp. 413-19. E. O. Negahban, A Preliminary Report on Marlik Excavation. Gohar Rud Expedition, Rudbar, 1961-62, Tehran, 1964. Idem, Metal Vessels From Marlik, Prähistorische Bronzefunde, Abt. II, Bd. 3, Munich, 1983. J. Nunome, The Origins of Sericulture and Ancient Silks (in Japanese), Tokyo, 1979; Eng. tr. courtesy Department of Near Eastern Art, The Metropolitan Museum of Art, New York. C. Parham, “How Altaic/Nomadic Is the Pazyryk Carpet?” Oriental Rug Review 13/5, June-July, 1993, pp. 34-39. R. Pfister, “Traces de tissu sur un miroir en cuivre provenant d’une tombe proto-élamite de Sialk (période VI),” in R. Ghirshman, Fouilles de Sialk, près de Kashan II, Paris, 1939, pp. 201-02 and pl. XXXII. E. Porada, “Notes on the Gold Bowl and Silver Beaker from Hasanlu,” in Survey of Persian Art XIV, pp. 2971-78. K. S. Rubinson, “A Reconsideration of Pazyryk,” in G. Seaman, ed., Ecology and Empire. A Symposium on Nomads in the Cultural Evolution of the Old World, forthcoming. S. I. Rudenko, Drevneĭshie v mire khudozhestvennye kovry i tkani, Moscow, 1968. Idem, Frozen Tombs of Siberia, tr. M. W. Thompson, Berkeley, 1970. E. F. Schmidt, Excavations at Tepe Hissar, Damghan, Philadelphia, 1937. G. M. Vogelsang-Eastwood, “Zilu Carpets from Iran,” Studia Iranica 17, 1988, pp. 225-40. [E. J. Laing, “Evidence for Two Possible Sasanian Rugs Depicted in Tun-Huang Murals of A.D. 642,” Ars Orientalis 12, 1981, pp. 69-71.] Dr. Trudy Kawami kindly supplied some of the information reported here.

(Karen S. Rubinson)

Because of the scarcity of surviving materials it is difficult to separate the history of carpet making in Iran from that of the rest of the Islamic world before the Mongol invasion (656/1258). Furthermore, the kind of rigid distinction between carpet and other textile designs that characterizes later production probably did not exist in the early Islamic period. On surviving fragments and contemporary depictions of both carpets and textiles simple patterns of lozenges, circles, stripes, and the like are the main designs; animal designs on early carpets seem to have been modeled on those of deluxe figured textiles, again suggesting a more general approach to the design of floor coverings at this period.

Until recently the entire history of carpets in the first centuries of Islam was known only through written accounts (for compilations and extracts, see Wiet, pp. 93-129; Pope, pp. 2276-80; Serjeant; Lombard, pp. 181-86). Excavation of early carpets and fragments in Egypt, including some that may have been made in greater Iran, has permitted the first step toward establishing a corpus of scientifically excavated and analyzed early pieces that may soon lead to a better understanding of early floor coverings. The extant material includes not only tapestry-woven and pile carpets but also fiber mats, textiles, and felt rugs.

Fiber mats (ḥaṣīr, ḥoṣr). Mats woven of rushes, straw, and undyed hemp were made throughout the warmer parts of the Islamic world. Among Persian centers of manufacture were Dārābjerd in Fārs, ʿAbbādān (Ābādān), and Sāmān in Ḵūzestān, and the province of Gīlān (Moqaddasī [Maqdesī], p. 442; Ḥodūd al-ʿālam, tr. Minorsky, p. 137, cf. Serjeant, 10, pp. 89, 102; 15-16, pp. 76-77; Dimand, 1942, p. 79; Lombard, p. 185). The mats of Sāmān and ʿAbbādān must have been especially fine, for they were described in the work of the 5th/11th-century writer Abu’l-Moṭahhar al-Azdī as lovelier than carpets and softer than silk, so that they could be folded in two as if they were made of cloth (p. 36; cf. Serjeant, 15-16, p. 76). Several examples and fragments survive from this period, though the provenience of most remains to be established. A large straw mat in the Benaki Museum, Athens, was made, according to its inscription, in Tiberias, in Palestine (Dimand, 1942, p. 78), a city mentioned by several medieval writers as a center for production of fine reed mats, including prayer mats (Dimand, 1973, p. 13; Ferber, no. 10). In The Metropolitan Museum of Art, New York, there is a similar, though smaller, example (Acc. no. 39.111) with two bands of stylized writing expressing blessings and good wishes to an unnamed owner. Both these mats have panels and borders of purple and black, and the weave is similar to a basic cloth weave (Dimand, 1942, p. 78). Other designs are also known: confronted lions rendered in a geometric style (on a fragment in the Benaki Museum; Delivorrias, p. 96), panels of script with decorative borders (al-Sabah collection, Kuwait; Jenkins, p. 43), and Kufic writing on a speckled ground (several fragments formerly in the H. P. Kraus collection, exhibited in Binghamton, New York, in 1975; Ferber, figs. 10a-b). All these examples have been attributed to the 4th-5th/10th-11th centuries. In 1980 a considerable number of fragments woven in various fibers and of various weights and structures were excavated in Fosṭāṭ (old Cairo) by the American Research Center in Egypt, under the direction of George Scanlon; they are presumed to date from the 5th/11th century, though they have not yet been systematically examined (Mackie, p. 26). Early textile inscriptions were an instrument of government economic control, usually indicating the city or workshop where the pieces were made (Bierman, p. 19); further study of the paleography on surviving mats may thus provide additional evidence of their origins (see calligraphy).

Textiles as floor coverings. On early Islamic metalwork and ceramics from Iran male figures are commonly depicted seated directly on the ground, with no cushions or textiles beneath them; kings and other important figures, on the other hand, sometimes sit on thrones covered with cloth. A fine throne rug of this type is represented on a silver bowl in the Hermitage Museum, Leningrad, generally dated to the 2nd-3rd/8th-9th century (Survey of Persian Art, pl. 208). In the field is an elaborate asymmetrical floral design, perhaps related to Chinese designs of the Tang period (Dimand, 1973, p. 13). At Pyandzhikent, near Samarqand, in murals dated to the 2nd/8th century figures are shown sitting on textiles with wide borders; borders and fields are sometimes patterned in pearled roundels, rosettes, and other repeat motifs that are also found on silk-brocade clothing of the period (Azarpay, figs. 52, 53, pls. 18). Identical or similar textiles appear as saddle cloths in these paintings (Azarpay, figs. 7, 50, pl. 20, 21). A reconstruction of another Pyandzhikent painting shows a god seated on a throne supported by horses. A cloth panel with a curving edge hangs down in front of the throne (Azarpay, fig. 35); it may be either one end of a small oval rug or possibly a swag attached to the edge of the seat. Thrones with similar pendants are represented on Persian metalwork. For example, on a shallow silver dish in the Islamic section of the Staatliche Museum, Berlin (West), a lute player rendered in relief is seated on a sturdy bench; an oval cloth with asymmetrical folds is draped beneath him (Survey of Persian Art, pl. 1354[3]B; 5th-6th/11th-12th century). Several inlaid medallions in the upper register on the bronze “Bobrinsky bucket” (made in Herat in 559/1163; now in the Hermitage, Leningrad) frame figures seated on simple bench thrones with small round textile swags visible (Survey of Persian Art, pl. 1308). Approximately seventy-five years later several personifications of zodiac constellations were represented in very much the same fashion on the lid of the famed “Vaso Vescovali” in the British Museum (1950, 7-25.1), seated on benches with legs in the shape of horse protomes, partially hidden by semicircular drapes (Baer, p. 255, fig. 207). The bunched folds of some of these drapes suggest that most of them were lightweight textiles, rather than heavier carpets. Ṭabarī (II, p. 952) mentions a military commander’s sitting on a ṭenfesa, a heavy carpet or cloak, during battle, and Ebn Jawzī (510-97/1126-1200) has recorded that the ascetic ʿOmar b. ʿAbd-al-ʿAzīz, while he was governor of Medina (87-93/706-12), used to sit on a cotton shirt spread over a Qaṭawānī cloak (ʿabāʾa) made of wool flock (Ebn Jawzī, Manāqeb, p. 31 ). Qaṭawān near Kūfa in southern Iraq was the probable origin of such rough woolen cloaks, but it is also possible that they were the products of Qaṭawān (Qaṭvān) near Samarqand (cf. Balāḏorī, Fotūḥ, p. 246; Yāqūt, Boldān IV, pp. 139-40; Serjeant, 9, p.63). These stories suggest that some “rugs” were no more than outer garments, which were perhaps used for bedding as well.

Felts (named; namaṭ, plur. anmāṭ). In colder parts of the Islamic world it is likely that carpets and felts were more widely used than lighter floor coverings; the surviving remains of such materials are, however, very fragmentary (see vi, above). In the 4th/10th century the term al-anmāṭ al-moḥkama (“sturdy anmāṭ”) is applied by Moqaddasī (Maqdesī) to products of Jahram (Jahrom) in Fārs (p. 442; cf. Pope, p. 2279; Serjeant, 10, F 86). Other towns that produced anmāṭ were Fasā (ibid.) in Fārs, Tostar (Šūštar) and Qorqūb in Ḵūzestān, and Dabīl (Dwin) and Barḏaʿa (q.v.) in Armenia (Moqaddasī, pp. 377, 380, 416; cf. Serjeant, 10, pp. 74, 78, 94, 96). It is not clear that at this early period the word necessarily meant felt, however, for Yāqūt (Boldān I p. 656) mentions that the men and women of Baṣenna (Boṣanna) wove anmāṭ and other items. At Pyandzhikent there are paintings of paladins and deities seated on rectangular and oval floor coverings that probably represent felts (Azarpay, pls. 13, 27); the fields are unornamented, though the narrow border, which are in contrasting colors, sometimes may be decorated with zigzags or other simple designs. In the excavations a Fosṭāṭ several small fragments of felt were uncovered in a 5th/11th-century context; the colors include yellow, blue, green, tan, and red, but there is no patterning (Mackie, p. 33).

Flat-woven carpets (gelīms). These carpets are woven of spun wool (see v, above), and most of the fragments known from the period have been found in Egypt. In Egypt wool was generally spun toward the right (S-spun), following the practice with linen, the fibers of which curl naturally toward the right when they are drying. As some of the pieces discovered in Egypt are made of wool spun to the left (Z-spun), they are believed to be imports, or possibly made locally of imported yarns, for example, a flat-woven piece with slit joins of an unusual type and a design of polychrome squares arranged in a stepped-lozenge design (Mackie, p. 332, figs. 8-9). Several groups have been distinguished within these finds. One large group of tapestry-woven bands, distinguished by the use of three-ply wool warps and distinctive lattice patterns containing birds and geometric motifs in the compartments (Thompson, no. 35, pl. XV, pp. 80-81; Kühnel, pp. 79-89, pl. I, fig. 2), has been dated by the paleography of some in-woven inscriptions to the 3rd-4th/9th-10th centuries. It is possible that these bands were sewn together in shrouds or coffin covers, and they were probably imported from Armenia, Turkey, or northern Syria (L. Bellinger, in Thompson, p. 80); similar designs occur on some 13th/19th-ceutury Anatolian gelīms (Michael Francis, personal communication). There are many other early Islamic woolen tapestries found in Egypt that have not yet been analyzed or classified according to type; they are believed to come from various sites in Syria, Mesopotamia, and Persia, as well as Egypt (Dimand, 1973, pp. 10-11). Gelīms are mentioned as products of Wāseṭ (in southern Iraq) and Fārs in Ḥodūd al-ʿālam (written 372/982; tr. Minorsky, pp. 126, 138, cf. Serjeant, 9, p. 89; 10, p. 80). The term does not occur frequently in early texts, which suggests that gelīms were made largely for private use, most probably by nomads, as they still are today.

The term zīlū also appears in early sources; these flat weaves were sometimes called Jahramī after the city in Fārs that was renowned for their production (Ḥodūd al-ʿālam, pp. 126, 129; Eṣṭaḵrī, p. 152; Ebn Ḥawqal, p. 212; cf. Serjeant, 10, pp. 80-81 ). They were also noted among the products of Sīstān (Ḥodūd al-ʿālam, p. 110); in the 5th/11th century Tūn in Kūhestān was an especially active production center (Nāṣer-e Ḵosrow, p. 141 ). Fine zalālī were produced at Dabīl, as well as at Badlīs (modern Bitlis; see bedlǰs), Ḵelāṭ (Ahlat), Arjīš (Erciş), Barkarī (or Bargīrī, modern Muradiye), Ḵᵛoy, Qālīqalā (Erzurum), and Naḵjavān (Nakhichevan) in Armenia (Moqaddasī, p. 374; Ṯaʿālebī, Laṭāʾef al-maʿāref, pp. 184, 237; Ḥodūd al-ʿālam, p. 143; cf. Serjeant, 10, p. 94; Vogelsang-Eastwood, p. 233) and at Warthān (or Vartān) on the southern bank of the Aras river in Azerbaijan (Ḥodūd al-ʿālam, p. 142). Thick, heavy fragments of cloth made of both S- and Z-spun dyed wool and Z-spun undyed cotton were found in the Fosṭāṭ excavations; the weave is related to modern pieces, also called zīlū, made in Iran and Iraq (Mackie, p. 33, figs. 14-15; see v, above). One piece has a stepped crenellated pattern, a second flattened “pearls” arranged in tangential rows; both designs may well be derived from traditional Persian stone and textile designs, for example, those represented on clothing on the reliefs at Ṭāq-e Bostān (see clothing iii). Another, technically similar fragment has been excavated at the Red Sea port of Quseir al-Qadim (al-Qoṣayr al-Qadīm) in a context probably of the 7-8th/13-14th centuries. It has been suggested that this entire group was influenced by a type of zīlū produced in Iran in the Sasanian period (Vogelsang-Eastwood, p. 233; see vi, above), but there is no conclusive evidence on the point.

Pile carpets. There is some evidence in Chinese sources suggesting that pile carpets from the Iranian world were reaching China in the 2nd/8th century (Schafer, pp. 198-99, 325 n. 25). Three fragments of knotted pile carpets were excavated in Egypt in 1980 and are believed to be of the 5th/11th century (Mackie, figs. 3-6, p. 28). Two of them have pile of symmetrical knots (see iii, above); the wool (Z2S) is not Egyptian, but the place of origin remains unknown. Each fragment has only two colors, crimson or tan and bluish-green, and both show an unusual treatment of the wefts, with multiple shoots (up to eight) interlaced with pairs of warp threads. Symmetrically knotted rugs with multiple weft shots of a.d. the 3rd century have been found at Dura Europos in Syria (1933.541; Pfister and Bellinger, no. 224, pl. IV) and in the al-Tār caves in Iraq (3rd century b.c.-a.d. 3rd century; Sakamoto, p. 10, fig. 3; Eiland, pp. 111, 230-33). This type of wefting is not known after the 7th/13th century (Mackie, p. 28). In 1986 the San Francisco Museum of Art purchased an entire small carpet (168 x 89 cm) executed in symmetrical knots. The design in the field is a large, geometrically stylized lion; the border is filled with running vines with palmettes, and there is a single outer guard stripe with a jewel pattern (see Plate CVI). The dating suggested on the basis of radiocarbon-14 analysis is 2nd-3rd/8th-9th centuries. Although no consensus has yet been reached on the origin of this piece, a number of features point to northeastern Iran. The stylization of the lion, with triple-clawed feet and broad bands partly filled with interlocking Ys outlining its body, has close parallels on the well-known “St. Josse silk” in the Louvre (Survey of Persian Art, pl. 981), which is inscribed with the name Abu’l-Manṣūr Baḵtegīn, an officer of the Samanid ruler ʿAbd-al-Malek b. Nūḥ (d. 350/961; Survey of Persian Art, p. 2002). The patterned treatment of the foreleg and tail has parallels in the rendering of lions on silk textiles made at Zandana, near Bukhara, in the 2nd-3rd/8th-9th centuries (Shepherd and Henning, pp. 21-29, figs. 5, 7, 8, 10, 12; Ierusalimskaya, pp. 5-56, 178-79; see also abrǰšam, pl. XII/2, where the legend has been reversed with that of pl. XII/1).

A special group of pile carpets consists of types in which the forms of the individual supplementary weft loops differ from those of the familiar symmetrical and asymmetrical knots; these types, though extremely varied, are often lumped together under the misleading term “cut-loop pile.” Yaʿqūbī (Boldān, p. 332), writing in 278/891, referred to “cut rugs” (al-farš al-qoṭūʿ) made in Aḵmīm, in upper Egypt, a term that has been interpreted as referring to carpets with a pile of cut loops (Dimand, 1973, p. 12), though it is not possible from such texts alone to specify the precise forms of the knots. Fragments of such carpets were known in northern Iran in the Sasanian period (see vi, above), and the region apparently remained a center of their manufacture for some centuries afterward. Six hundred pieces of qeṭʿ forš (also called Ṭabarī carpets) were sent as part of the tax payment of Ṭabarestān, Rūyān, and Nehāvand to the ʿAbbasid court in the late 2nd/8th century (Ebn Ḵaldūn, I, 323; cf. Serjeant, 9, p. 63; for the date see von Kremer, pp. 265-67). Two small fragments of carpets identified as of a cut-loop type, one in the Victoria and Albert Museum, London (T.33-1942), the other formerly in a private collection in Washington, D.C., have designs of stylized birds in polygonal frames decorated with circles and radiating trefoils, designs considered to be Persian in character; they are believed to date from the 3rd-4th/9th-10th century and to be among the antecedents of the better-known knotted animal carpets of the 7th/14th century (Ettinghausen, pp. 92-98).

Bibliography : B. B. Acar, Kelim-Cicim-Zili-Sumak. Turkish Flatweaves, Istanbul, 1983. G. Azarpay, Sogdian Painting. The Pictorial Epic in Oriental Art, Berkeley and Los Angeles, 1981. Moḥammad b. Aḥmad Abi’l-Moṭahhar al-Azdī, Ḥekāyat Abi’l-Qāsem Baḡdādī, ed. A. Mez, Heidelberg, 1902. E. Baer, Metalwork in Medieval Islamic Art, Albany, 1983. I. A. Bierman, “The Significance of Arabic Script on Carpets,” Halı 5/1, 1982, pp. 18-22. A. Delivorrias, Guide to the Benaki Museum, Athens, 1980. M. Dimand, “Two Abbasid Straw Mats Made in Palestine,” Metropolitan Museum of Art Bulletin, N.S. 1, summer 1942, pp. 76-77. Idem, Oriental Rugs in the Metropolitan Museum of Art, New York, 1973. Abu’l-Faraj ʿAbd al-Raḥmān Ebn Jawzī, Manāqeb ʿOmar b. ʿAbd-al-ʿAzīz, ed. C. H. Becker, Leipzig, 1899. R. M. Eiland, Chinese and Exotic Rugs, Boston, 1979. R. Ettinghausen, “Ḳālī,” in EI¹, Suppl., pp. 106-11. Idem, “New Light on Early Animal Carpets,” Aus der Welt der Islamischen Kunst, Festschrift für Ernst Kühnel, ed. R. Ettinghausen, Berlin, 1959, pp. 92-98. S. Ferber, ed., Islam and the Medieval West, Binghamton, N.Y., 1975. Islamic Art, The Metropolitan Museum of Art Bulletin, N.S. 40 spring 1983. A. A. Ierusalimskaya, “K slozheniyu shkoly khudozhestvennogo shelkotkatchestva v Sogdie” (On the formation of schools of artistic weaving in silk in Sogdia), Srednyaya Aziya i Iran (Central Asia and Iran), Leningrad, 1972, pp. 5-56; English abstract, pp. 178-79. M. Jenkins, ed., Islamic Art in the Kuwait National Museum. The Al-Sabah Collection, London, 1983. A. von Kremer, Kulturgeschichte des Orients unter den Chalifen I, Vienna, 1875; repr. Aalen, 1966. E. Kühnel, “La tradition copte dans les tissus musulmans,” Bulletin de la Société d’archéologie copte 4, 1938, pp. 79-89. Idem and L. Bellinger, Catalogue of Dated Tiraz Fabrics, Washington, D.C., 1950. M. Lombard, Les textiles dans le monde musulman VII^e-XII^e siècle, Paris and the Hague, 1978. L. W. Mackie, “Covered with Flowers: Medieval Floor Coverings Excavated at Fustat in 1980,” Oriental Carpet and Textile Studies 1, 1985, pp. 23-35. C. C. Mayer-Thurman and B. Williams, Ancient Textiles from Nubia, Chicago, 1979. Nāṣer-e Ḵosrow, Safar-nāma, Berlin, 1341 Š./1962. R. Pfister and L. Bellinger, The Excavations at Dura Europos. The Textiles, New Haven, 1945. A. U. Pope, “The Art of Carpet Making. A. History,” Survey of Persian Art, pp. 2257-2430. K. Sakamoto, “Ancient Pile Textiles from the At-Tar Caves in Iraq,” in Oriental Carpet and Textile Studies 1, 1985, pp. 9-17. E. H. Schafer, The Golden Peaches of Samarkand. A Study of T’ang Exotics, Berkeley and Los Angeles, 1963. R. B. Serjeant, “Material for a History of Islamic Textiles up to the Mongol Conquest,” Ars Orientalis 9-15/16, 1942-50; repr. Beirut, 1972. D. G. Shepherd and W. B. Henning, “Zandaniji Identified?” Aus der Welt der Islamischen Kunst. Festschrift für Ernst Kühnel, ed. R. Ettinghausen, Berlin, 1959, pp. 21-29. F. Spuhler, Islamic Carpets and Textiles in the Keir Collection, London, 1978. Ṯaʿālebī, Laṭāʾef al-maʿāref; ed. I. al-Abyārī and Ḥ. K. al-Ṣayrafī, Cairo, 1960. D. Thompson, Coptic Textiles in the Brooklyn Museum, New York, 1971. E. de Unger, “An Ancestor of the Mamluk Carpets,” Halı 5/1, 1982, pp. 44-46. G. M. Vogelsang-Eastwood, “Zilu Carpets from Iran,” Studia Iranica 17, 1988, pp. 225-40. G. Wiet, L’exposition persane de 1931, Cairo, 1933.

(Barbara Schimtz)

Persian carpets that can be indisputably identified a having been produced in the 8-9th/14-15th centuries are virtually nonexistent. One fragment of a pile carpet in the Benaki Museum in Athens has recently been attributed to “Iran or Central Asia” in the 9th/15th century (Lentz and Lowry, p. 220, no. 119), though the reasons for this attribution are not clear. It is made entirely of wool and measures 14 x 22" (35 x 56 cm; Mackie). Many features of the design link it (and comparable fragments reported to be in the same collection; R. Pittinger, personal communication) to Anatolian Saljuq and “Holbein” rugs dated between the 7th/13th and the early 10th/16th centuries (Yetkin, pls. 1-22, esp. 26-27; cf. Spuhler, 1978, nos. 1, 3, 5-14; for a zīlū carpet said to be dated to 807/1405, see iv and v, above).

That carpets were used and produced in Persia in the 8-9th/14-15th centuries has nonetheless been inferred from written sources, both contemporary and slightly earlier (e.g., Barbaro and Contarini, p. 119; Erdmann, 1962, p. 18; idem, 1977, p. 14). The existence of carpets and weavings from contemporary Anatolia and the Turkman tribal confederations, and possibly also from Egypt and even Spain (Spuhler, 1978, pp. 27-32; Helfgott, pp. 107-14), permits the inference that carpets were being produced in Persia as well. Finally, it has been argued that “the finest surviving knotted carpets . . . of the Safavid dynasty . . . could not have originated spontaneously” (Spuhler, 1986, p. 698).

In the absence of surviving examples from the period, hypotheses about the colors and designs of Persian carpets have been based on comparisons with 9th/15th-century illustrated Persian manuscripts, most fully by Amy Briggs. The validity of using pictorial art as a source of evidence for textiles that no longer survive is, however, questionable at best; a concomitant problem is that the surviving parallels for the depiction of rugs in 9th/15th-century Persian painting usually turn out to be Anatolian. Nonetheless, in the continuing absence of material evidence from Persia, the approach still has adherents. If representations of rugs in contemporary painting are to be used as evidence for Persian carpets in the 8-9th/14-15th centuries, the paintings must be chosen according to the strictest standards of quality, dating and localization, and purpose, for only at that level can specific illustrations of the material environment be found.

Timurid carpets. Briggs’ arguments are flawed by the underlying assumption that a rug shown in a Timurid painting exactly represents a Timurid carpet. In her first article (1940) she discussed geometrically patterned rugs; in her second (1946), on arabesque and “flower” rugs, she continued to work from the same flawed hypothesis. It is now clear that not only geometric rugs but also other types can be found in illustrated Timurid manuscripts from court workshops, dated throughout the century, which casts doubt on the argument that the classical Timurid carpet was geometrically patterned (Spuhler, 1986, pp. 698-700; Lentz and Lowry, p. 221). For example, in “Rūdakī and the Samanid Amir” from a copy of Neẓāmī ʿArūżī’s Čahār maqāla made for Bāysonḡor in 835/1431, the two figures sit on a rug with floral arabesques in the field and a contrasting border in pseudo-Kufic script with knotted ascenders (Istanbul, Türk ve Islam Eserleri Müzesi, 1954, fol. 22r; Gray, 1979, pl. L); in “Laylī and Majnūn at School” from a copy of Neẓāmī Ganjaviʾs Ḵamsa dated 866/1461 a rug with multiple niches is depicted (Plate CVII; Topkapı Sarayı , Istanbul, Hazine 761, fol. 106r; cf. Grube, pl. 33A).

Briggs’, and others’, arguments regarding Timurid carpets become less significant as other possible sources of evidence emerge. For example, scholars are now considering the possibility that rugs once considered Safavid are actually Timurid in date (e.g., a fragmentary arabesque- and floral-patterned multiple prayer rug in the Museum für Islamische Kunst, Berlin; Spuhler, 1986, p. 700 and pl. 5). The identification of Timurid woven textiles, notably those with curvilinear patterns (Lentz and Lowry, p. 218), also allows the comparison of designs on illustrated Timurid clothing and furnishing textiles with those presumed to be typical of Timurid rugs, which in turn indicates the importance of the moḏahheb “draftsman” and other personnel of the ketāb-ḵāna “library” as the source of Timurid designs on all materials, regardless of scale (Lentz and Lowry, chap. 3).

Il-khanid carpets. The carpets of the preceding, Il-khanid, period, however, remain as little known as ever. Briggs had restricted herself to Timurid painting, as carpets are only occasionally represented in Il-khanid miniatures (1946, p. 20 n. 3), and her approach, based on contemporary book illustrations, provides few clues. For example, two floor coverings with the appearance of carpets are depicted in the great Šāh-nāma, now attributed to the period of Abū Saʿīd, the last of the Il-khanid rulers of Iran (717-36/1317-35; e.g., Grabar and Blair, pp. 46-55): Unfortunately, the surviving material for comparison is from Anatolia, rather than from Persia. Part of a large octagon with a dragon-like animal in the foreground can be discerned on the floor covering depicted in “Żaḥḥāk Enthroned” (Freer Gallery of Art, Washington, D.C., 23.5; Grabar and Blair, p. 59); this design was characteristic of 8th/14th-century Anatolia (Yetkin, pp. 27-41, esp. fig. 12). In the foreground of “Mourning over the Bier of Eskandar” (Freer, 38.3; Gray, 1961, p. 32) there is a large rug with a field pattern similar to those on brick surfaces depicted in other Persian miniatures and an outer border in white pseudo-Kufic script on a red ground, a pictorial formula known from 7th/13th-century Anatolian carpet fragments (cf. Yetkin, pls. 3, 4, 7, 13).

As in the 9th/15th century, the similarity in form and color between illustrated 8th/14th-century Persian carpets and contemporary Anatolian weavings is striking. It is thus difficult to escape the conviction that before the 10th/16th century the development of the knotted carpet in Anatolia, Iran, and the tribal areas adjacent to both was characterized both aesthetically and technically more by similarities than by differences.

Bibliography : J. Barbaro and A. Contarini, I viaggi in Persia degli ambasciatori veneti Barbaro e Contarini, ed. L. Lockhart et al., Il nuovo Ramusio 7, Rome, 1973, pp. 119, 120, 124, 127, 128. B. Brend, “A Carpet and Related Pictures—A Legacy of Timur’s Samarqand?” Oriental Art N.S. 30/2, 1984, pp. 178-88. A. Briggs, “Timurid Carpets,” Ars Islamica 7, 1940, pp. 20-54; 11-12, 1946, pp. 146-58. K. Erdmann, Oriental Carpets. An Essay on Their History, tr. C. G. Ellis, New York, 1962. Idem, The History of the Early Turkish Carpet, tr. R. Pinner, London, 1977. O. Grabar and S. Blair, Epic Images and Contemporary History. The Illustrations of the Great Mongol Shahnama, Chicago, 1980. B. Gray, Persian Painting, Geneva, 1961. Idem, ed., The Arts of the Book in Central Asia, Paris and London, 1979. E. J. Grube, “Notes on the Decorative Arts of the Timurid Period,” in Gururājamañjarikā. Studi in onore di Giuseppe Tucci, Naples, 1974, pp. 233-79, pls. 24-111, esp. pp. 256-59 (rugs and textiles). Idem, La pittura dell’Islam, Bologna, 1980. L. M. Helfgott, “Production and Trade. The Persian Carpet Industry,” in C. Bier, ed., Woven from the Soul, Spun from the Heart. Textile Arts of Safavid and Qajar Iran 16th-19th Centuries, Washington, D.C., 1987, pp. 107-120. T. W. Lentz and G. D. Lowry, Timur and the Princely Vision. Persian Art and Culture in the Fifteenth Century, Los Angeles and Washington, D.C., 1989. L. Mackie, “A Piece of the Puzzle: A 14th-15th Century Persian Carpet Revealed,” Halı 47, 1989, pp. 16-23, 55-56. F. Spuhler, Islamic Carpets and Textiles in the Keir Collection, London, 1978. Idem, “Carpets and Textiles,” in Camb. Hist. Iran VI, 1986, pp. 698-727. S. Yetkin, Historical Turkish Carpets, tr. M. Quigley, Istanbul, 1981.

(Eleanor Sims)

The high point in Persian carpet design and manufacture was attained under the Safavid dynasty (907-1152/1501-1739). It was the result of a unique conjunction of historical factors—royal patronage, the influence of court designers at all levels of artistic production, the wide availability of locally produced and imported materials and dyes (see ii, above), and commercial acceptance, particularly in foreign markets. Although no certain examples of 9th/15th-century Persian carpets may survive today (see viii, above), the high level of quality achieved early in the Safavid period confirms what seems apparent from travelers’ accounts (e.g., that of the Venetian Giosafat Barbaro, q.v., who was in Tabrīz in 878-79/1474; Barbaro and Contarini, pp. 124, 126-27; Eng. tr., pp. 57, 59-60) and representations of carpets in 9th/15th-century Persian paintings (Briggs; see viii, above): that a mature carpet-weaving tradition already existed in the century before the Safavids assumed the throne. Surviving Safavid carpets represent both court production (i.e., carpets produced to court standards but not necessarily at the court itself) and commercial production, which often reflects courtly tastes. Carpets were woven in palace workshops and on urban and village looms. No example of a nomadic carpet remains from the Safavid period, probably owing to hard use, but such carpets surely existed well before courtly and commercial interests engendered an urban industry (see v, above, and xiii, below). Depictions of nomadic carpets in manuscript paintings, though rare, confirm that production of such rugs took place during the Safavid period (Grube, p. 127, right center).

Because of significant changes early in the 11th/17th century, it is convenient to discuss carpets produced before and after that dividing line separately. Although the major types are described here, there remain many Safavid carpets that are either anomalous or have not been available for direct examination and classification. The increased use of structural analysis in recent years has helped to clarify the criteria for assigning carpets to specific groups; nevertheless, several types bear traditional names that, though not technically correct, are still used for convenience. Geographical names, for example, can be taken only as general identifiers of type, not as absolute indications of the centers where the rugs were woven. Indeed, the localization of specific categories of weaving is generally still obscure, for it is rarely possible to match references in travelers’ descriptions with existing rugs. The chronology of types can also be outlined only generally because of the scarcity of datable pieces around which a more precise chronology could be constructed.

I. The 10th/16th century.

The history of carpet weaving during the first century of Safavid rule can be traced only sketchily. The most important documentary evidence is three carpets with dated inscriptions. The date on a medallion rug in the Poldi Pezzoli collection in Milan has been read by some scholars as 929/1522-23 and by others as 949/1542-43 (Spuhler, in Camb. Hist. Iran VI, p. 700; cf. Beattie, 1986, p. 366). If the earlier date is correct, it was made late in the reign of Shah Esmāʿīl (r. 907-30/1501-24), just at the time when the court atelier was beginning to produce large numbers of royal manuscripts. The pair of carpets allegedly from the shrine at Ardabīl (see Beattie, 1986) is dated 946/1539-40, the moment of supreme artistic achievement under Esmāʿīl’s son and successor, Shah Ṭahmāsb. The inscription on each of these carpets contains the weaver’s name with a nesba (attributive name), Jāmī on the Milan carpet, Kāšānī on those from Ardabīl, but the nesba does not necessarily refer to the place of manufacture. There are also references in 10th/16th-century sources to specific weaving sites (see i, above). A gift from Ṭahmāsb to Sultan Bāyazīd, the fugitive son of the Ottoman sultan Solaymān, in 961/1553-54 included carpets from Kermān and Jowšaqān (near Kāšān) woven with gold thread, and Kermān and Jowšaqān carpets were presented by Shah ʿAbbās I to the shrine of Imam Reżā in Mašhad in 1007/1598-99 (Eskandar Beg, I, pp. 112, 578; tr. Savory, I, p. 169, II, p. 764). Silk carpets from Hamadān and from Dargazīn in Khorasan were presented as gifts to the Ottomans in 974-75/1567 (Hammer, III, p. 520). Among the abundant trade goods brought to the Persian Gulf entrepôt of Hormoz were carpets from Khorasan (Linschoten, I, pp. 47, 166). At the end of the century Abu’l-Fażl ʿAllāmī (q.v.), biographer of the Mughal emperor Akbar, referred to carpets that were still being imported to India from Jowšaqān, Ḵūzestān, Kermān, and Sabzavār (Abu’l-Fażl, p. 55). Pedro Teixeira (p. 243) judged the carpets of Yazd to be the best, followed by those of Kermān and Khorasan. Carpets are not specifically mentioned in connection with the three Safavid capital cities of the 10th/16th century—Tabrīz, Qazvīn, and Isfahan—but it seems likely that they were produced there (see i, above). Although there is no direct evidence that royal weaving workshops had yet been established, the influence of court designers on carpet weaving is clear. The two key design features of rugs in this period, the medallion design and figural elements, were borrowed directly from the arts of the book as practiced in the royal atelier. Persian carpets continued to find their way abroad: Gifts of carpets to the Ottomans were recorded or depicted in 965-66/1558, 974-75/1567, and 988-89/1581 (Atil, p. 180; Hammer, III, p. 520), and India continued to import Persian rugs even after the establishment of royal workshops there (Abu’l-Fażl, p. 55). An English dyer was dispatched to Persia in 987/1579 to learn what he could of local dyeing methods (Hakluyt). Nevertheless, it does not seem, either from the quantity of extant rugs or from consistent use of certain designs, that carpet production was particularly stimulated by foreign demand.

Northwest Persian medallion carpets (cf. Survey of Persian Art, pls. 1112-14, 1116-25, 1127-28). This group consists of about thirty carpets, most with wool pile (though at least two include some cotton), cotton warps, and either cotton or wool wefts. The designs feature multipointed (often sixteen-pointed) central medallions. The most celebrated example is a medallion carpet in the Museo Poldi Pezzoli, Milan (no. DT 1), with horsemen hunting animals across the field (Plate CVIII). In the center of the medallion is an inscription bearing a date (read as either 929/1522-23 or 949/1542-43). Another field design, more characteristic of the group, consists of a repeat pattern of scrolling vines and forked leaves. These carpets appear to have been woven from about the turn of the 10th/16th century, or perhaps earlier, into the 11th/17th century, though it is uncertain whether lower quality of design and execution indicates a decline over time or simply contemporary production at an inferior level. The group is usually associated with northwestern Persia because of perceived similarities to other medallion carpets, including the “Ardabīl carpets,” traditionally assigned to northwestern Persia (see “Silk-foundation carpets,” below). Variations in coloring suggest that the carpets were made at more than one weaving center.

Silk-foundation carpets (cf. Survey of Persian Art, pls. 1130-36, 1144-47, 1152-54, 1156-62, 1165-70, 1177). A significant number of extremely fine carpets woven on silk foundations have so far eluded categorization except by design (defined by single central medallions, multiple medallions, cartouches, prayer niches, and so on). Characteristics include extremely fine weave (in some examples there are about 500 knots to the square inch), luxurious materials (beside the silk foundations silver and gold brocading also occurs occasionally), elegant drawing, and rich colors. Some, like the multiple-medallion “Chelsea carpet” (Victoria and Albert Museum, London, no. 589-1890, Survey of Persian Art, pls. 1130-32) and a pair of fine cartouche rugs (The Metropolitan Museum of Art, New York, no. 10.61.3, Survey of Persian Art, pl. 1133; Musée Historique des Tissus, Lyons, no. 25.423) include elaborate figural motifs. Others, like the renowned “Ardabīl carpets” (Victoria and Albert Museum, no. 272-1983; Los Angeles County Museum of Art, no. 53.50.2), dated by inscriptions to 946/1539-40, are exclusively nonfigural (see Plate CIX). This famous pair of medallion carpets was long thought to have been made for the ancestral shrine of the Safavids at Ardabīl, but a study of inventories and the sizes of the rooms in the shrine, as well as suggestions that the carpets actually came from the Shrine of Imam Reżā at Mašhad (Weaver, p. 48), has cast doubt on this association. The so-called “Salting group,” the nucleus of which is the “Salting rug,” named for its donor (Victoria and Albert Museum, no. T. 402-1910, Survey of Persian Art, pl. 1162), originally thought to be from 9th/16th-century Persia, has been reattributed to 13th/19th-century Turkey. Doubts about many similar pieces with medallion and prayer-niche designs and either secular or religious inscriptions have also been raised (Erdmann, 1970, pp. 76-80). Nevertheless, the earlier attribution should not be dismissed for every rug in the group without careful examination; some may indeed be early pieces. The Salting rugs (at least the possibly authentic ones) have silk foundations, and most are brocaded. Carpets of the silk-foundation class appear to span the full 10th/16th century and may have been manufactured in several places; Tabrīz, Kāšān, Qazvīn, and Herat have all been suggested as possibilities.

Kāšān silk carpets (cf. Survey of Persian Art, pls. 1191-1202). A group of twenty rugs with silk warps, wefts, and pile is characterized by designs based on a variety of lobed and ogival medallions and on figural elements found also in the arts of the book. Of the four large pieces the most famous is a hunting carpet in Vienna (Österreichisches Museum für Angewandte Kunst, no. T8336, Survey of Persian Art, pls. 1191-92), with a central medallion around which are arranged horsemen hunting animals against a background of flowering plants. The figure style is comparable to that in royal illustrated manuscripts of the court of Shah Ṭahmāsb (r. 930-84/1524-76). This carpet is embellished with gold and silver brocade. Of the sixteen smaller pieces twelve have various medallion designs, and four have animals, some in combat, arranged pictorially (i.e., oriented toward one end of the rug, rather than symmetrically disposed; see iv, above). The dating of this group ranges from about 946/1540, when artistic production under the patronage of Shah Ṭahmāsb was at its peak, to perhaps the end of the century. The traditional attribution of this group to Kāšān is based primarily on the fame of that city’s silk industry and on the 11th/17th-century production at Kāšān of carpets of similar luxury materials (see “Tapestry-woven carpets” and “Polonaise carpets,” below).

Herat carpets (cf. Survey of Persian Art, pls. 1140-41, 1143, 1148-51, 1174-76, 1178-79, 1182-84, 1186-87). Several dozen carpets and fragments belong to this group. The most prevalent field design consists of symmetrically disposed scrolling vines with palmettes and leaves. Animals and birds are sometimes added to the floral pattern, and palmettes and blossoms may take on highly complex and fantastic forms. Elaborate striped cloud bands occur. Usually the field is red, and the main border, which also contains floral motifs, is dark green. The best-known carpet in this group is the “Emperor’s carpet,” actually a pair of carpets (Österreichisches Museum für Angewandte Kunst, no. T8334; The Metropolitan Museum of Art, no. 43.121.1), supposedly a gift from Peter the Great of Russia to Leopold I, emperor of Austria, in 1698 (Plate CX). The finest examples have silk foundations; others, though closely related in style, design, and color, have mixed foundations of wool, silk, and cotton, sometimes plied together, a peculiarity rarely encountered elsewhere. Some examples are brocaded. A few fragments characterized by exquisite drawing and with pile of very fine wool, probably goat hair, seem to belong to this group. The Herat group is thus one in which a real hierarchy of weaving quality can be observed: Similar designs were woven in several easily distinguishable grades. The association with Herat is based on affinities between the rug designs and late 9th/15th-century painting and illumination in Herat, the Timurid capital. The Herat group probably dates from the second half of the 10th/16th and early 11th/17th centuries and was the forerunner of the so-called Indo-Persian group of the 10th/17th century (see below).

Sanguszko carpets (cf. Survey of Persian Art, pls. 1205-10, 1212A, 1213-14). The Sanguszko carpets, of which more than a dozen examples are known, are distinguished by figural decoration closely tied to manuscript illustration, a bright appearance owing to abundant use of white in the pile, and certain technical peculiarities. The group owes its name to the former owner of one magnificent example of the group (currently on loan to The Metropolitan Museum of Art, no. L1985.3; Survey of Persian Art, pl. 1206). Medallion designs with symmetrically arranged floral patterns predominate, but directional designs with figural motifs are also found. Animal-combat motifs are plentiful, and animal heads often inhabit palmettes and border designs. Clusters of human figures embellish discrete compartments in the field or border; some of these figures can be directly linked to manuscript illustrations, for example, scenes from Laylī o Majnūn and images of fighting camels, a popular subject first seen in an early 10th/16th-century painting by Behzād (q.v.). Technically the Sanguszko pieces resemble the so-called vase carpets (see below) with cotton warps and wefts of wool and silk, but they differ in other ways. The localization of this group is highly conjectural; Kāšān, Yazd, Kermān, and Qazvīn have all been suggested, but certain differences within the group suggest that these carpets were produced at more than one weaving center. A date late in the 10th/16th, or perhaps even early in the 11th/17th, century seems plausible on the basis of the figure style and the existence of Indian copies that probably date from the early 11th/17th century.

II. The 11th/17th century.

Several factors contributed to major alterations in the organization of rug production in the 11th/17th century. Most important, the Persian carpet became a commercial commodity. A dramatic increase in the number of extant pieces (perhaps ten times as many 11th/17th-century rugs survive as those of the 10th/16th century), most of which have been found in the West, attests to an extensive trade in carpets among English, Dutch, and Portuguese companies operating in the Persian Gulf (see i, above). Royal workshops were established by Shah ʿAbbās I (q.v.) in Isfahan, the capital, and elsewhere, but their production was not limited to materials for use at court; the surplus was exported for sale in Europe and India, and the profits were returned to the treasury (Tadeusz Krusiński, cited in Mańkowski, in Survey of Persian Art, pp. 2431-32; Keyvani, p. 171). The contemporary Turkish writer Evliya Çelebi (Awlīā Čelebī/Čalabī) tells of Istanbul dealers who sold carpets from Isfahan (Atil, p. 182). Rural carpet weavers also worked for the shah, using materials provided from the court stores (see i, above). References to carpet weaving in 11th/17th-century travel accounts are far more numerous than before, but it is difficult to determine which factories or production centers were connected with the court and to judge which extant carpets were woven where. Jean-Baptiste Tavernier gave the precise location of the workshops in the Isfahan bāzār (Tavernier, I, p. 444), and Adam Olearius (p. 303) mentioned 1,460 villages in the vicinity that were engaged in weaving carpets and textiles. Other travelers commented on carpets produced in Isfahan (Silva y Figueroa, p. 216; Chardin, VII, p. 330). Kāšān is cited for its carpets, especially gold and silk ones, in many accounts (Paul Simon, cited in Mańkowski, in Survey of Persian Art, p. 2432; Silva y Figueroa, p. 216; Chardin, III, pp. 3-4). In a scroll dating from the reign of Shah Solaymān I (1077-1105/1666-94) the twenty-one main caravansaries of Isfahan are listed, two of them noted for Kāšān carpets and one for carpets from Khorasan and Herat (British Library, Sloane 4096; cf. Keyvani, pp. 237, 238). Herat was cited for producing the best carpets (Olearius, p. 199), but so was Kermān (Chardin, IV, p. 154). Engelbert Kaempfer (p. 202) described fine woolen rugs with animal patterns from Kermān. Other 11th/17th-century travel accounts refer to the carpets of Sīstān (Chardin, IV, p. 154) and Tabrīz (Schillinger, cited by Spuhler, in Camb. Hist. Iran VI, p. 702). Richard Steel and John Crowther (p. 275) reported a number of villages scattered along the edge of the salt desert where carpets, presumably of local manufacture, were sold in abundance. The style of carpets also changed in the early 11th/17th century. Although several earlier types characterized by figural decoration may have persisted, on the whole figural decoration was superseded by strictly floral designs. The relation between carpet design and the arts of the book, so notable in the 10th/16th century, seems to have ceased. All the 11th/17th-century pieces that can be dated by inscriptions, inventory lists, or depictions in European paintings have floral decoration.

Tapestry-woven carpets (gelīms; see v, above; cf. Survey of Persian Art, pls. 1262-69). One coherent group consists of about two dozen rugs that are tapestry-woven in silk and gold and silver thread. Although they are not executed in the knotted technique of the other pieces discussed here, they have many other features in common with Safavid knotted carpets. At least nine of the tapestry-woven pieces include animals or human figures in medallion or compartmented designs; the finest example (in the Residenzmuseum, Munich, Survey of Persian Art, pl. 1264) incorporates both well-drawn figures and poetic inscriptions. The remainder are strictly floral, typically with a central medallion on a field of scrolling vines, palmettes, and leaves. Three of these rugs include medallions containing the coat of arms of the Vasa kings of Poland (Survey of Persian Art, pl. 1268B). Documents show that in 1010/1601 King Sigismund III (1587-1632) ordered carpets from Kāšān; they were delivered in 1011/1602, with an itemized account referring to the execution of the royal arms, which were not reproduced with perfect accuracy (Mańkowski, 1936, pp. 152-53). These particular rugs, then, were almost certainly woven in Kāšān in 1010-11/1601-02, and the figural piece in Munich was probably obtained there also; all of them were included in the dowry of Sigismund’s daughter Anna Katharina Constanza in 1642. Many other rugs of this class may therefore have been woven in Kāšān, but it is not correct to infer that all were; some may well have been made in Isfahan or elsewhere. Production probably continued into the second half of the 11th/17th century.

Polonaise carpets (cf. Survey of Persian Art, pls. 1242-257, 1261). The Polonaise group is an unusually large one; more than 200 examples are known. The name is somewhat inappropriate, for it is based on misidentification of a coat of arms on one example as Polish, but it has remained in use as a matter of convenience. It is also a reminder that many of these sumptuous pieces, of silk (occasionally with some cotton in the foundations) with silver and gold brocading, though woven for local consumption, were also presented or sold to Europeans. The Polonaise carpets reflect a new aesthetic, perhaps owing to the tastes of Shah ʿAbbās I alone or to commercial influences. About a dozen different field patterns were used. In the few large pieces the field pattern is complete or nearly so; in the much more numerous small pieces only a detail of one of the standard patterns is seen. The concept of a single ground color was generally abandoned in favor of design units with different ground colors. The palettes of many seem now rather “sweet,” pastel and muted, and some colors have faded; the few examples in which the colors remain fresh demonstrate that Polonaise carpets were originally as brilliant as other types. The design elements of most Polonaise carpets are purely floral and include palmettes, curving leaves, and vines. At least twenty-five pairs of such carpets exist. There is some documentary evidence regarding these pieces. Traveler’s reports (Tavernier, I, pp. 444, 654; Krusiński, cited in Mańkowski, in Survey of Persian Art, p. 2432) attest that such luxury weavings were produced in Isfahan itself, in the bāzār area near the public square (maydān), though it is likely that they were woven in other places as well. One of the rugs in Anna Katharina Constanza’s dowry, obtained and possibly woven in Kāšān in 1010-11/1601-02, is in fact a Polonaise (Survey of Persian Art, pl. 1251). Production of the type probably began early in the reign of Shah ʿAbbās, late in the 10th/16th century, and continued into the second half of the 11th/17th century. Five examples in the Treasury of Saint Mark’s, Venice (Survey of Persian Art, pls. 1244-45), can be linked to ambassadorial gifts of 1012/1603 and 1030/1621 (Spuhler, 1968, pp. 102-3). One beautifully preserved carpet in Copenhagen (Plate CXI, Plate CXII) is thought to have been presented to the Dutch queen Sophie Amalie in 1076-77/1666 (Mellbye-Hansen, p. 14). A silk carpet in the mausoleum of Shah ʿAbbās II at Qom is dated to the year 1081/1671 (Survey of Persian Art, pl. 1258B); there are uninscribed companion pieces (pls. 1258A, 1259-60). These examples are so different in style from most Polonaise carpets, however, that their inclusion in the group is uncertain.

Vase carpets (cf. Survey of Persian Art, pls. 1218-39). Vase carpets take their name from the vases incorporated into the elaborate, multileveled floral patterns of some of the best-known examples. Rugs in this distinctive design, with two or three superimposed grids, or lattices, of vines and blossoms, often oriented toward one end of the rug (see iv, above), as well as pieces in several other designs, share specific technical features that are in fact now the basis for inclusion in the group. The warps are cotton, the wefts are wool and either silk or cotton, and the pile is wool. There are three weft passes between adjacent rows of knots; the first and third are of mixed colors of wool, generally either the colors of the pile fiber or brindle (mixed fibers in natural colors). This construction results in a distinctive “feel” and an easily recognizable pattern of wear. In general, the coloring is rich and varied. In addition to floral lattice designs, “vase carpets” are known with garden designs (water canals in aerial view and vegetation in profile), sickle-leaf designs, directional designs with rows of flowering shrubs or animals amid flowering plants and trees, and symmetrically arranged arabesque designs. Even classical medallion designs are found in examples belonging to this group. “Vase carpets” have most often been attributed to Kermān. They were produced in the reign of Shah ʿAbbās I, perhaps as early as the late 10th/16th century, and continued until post-Safavid times. It is ironic that the only documentary evidence for dating this group pertains to a carpet on which no vase is depicted, a magnificent garden carpet in Jaipur, India (Central Museum, no. 2225), which bears a label showing that it was first inventoried in the Amber fort, near Jaipur, in 1042/1632 (Dimand, p. 95).

Indo-Persian carpets (cf. Survey of Persian Art, pls. 1179a, 1180-81, 1188-89). Some time in the early 11th/17th century carpets with cotton foundations and wool pile in designs emulating those of the 10th/16th-century Herat group (see above) began to be produced in large quantities. The designs are simplified versions of the Herat type, usually consisting of symmetrical arrangements of vines and palmettes on a deep-red field with dark-green borders. Animals occur rarely and only on a few carpets that may be considered “transitional”; that is, the drawing is typical of classical 10th/16th-century Herat pieces, but the cotton foundations are more typical of 17th-century pieces. Such transitional rugs in fact suggest that the Indo-Persian class represents a continuation of the Herat type and its adaptation to commercial production. The origin of the Indo-Persian carpets has long been debated, and attributions to Isfahan, Herat, and northern India have all been put forth. The issue is far from resolved, but the evidence for a Persian attribution seems stronger: Indo-Persian carpets are technically more closely related to Persian products, and inventory information in Jaipur confirms that several such pieces were of foreign manufacture (Campbell, nos. 43, 45). It is possible that the variety in design and color typical of this class reflects different centers of manufacture. Several hundred such rugs are extant, many in Portugal, and they are frequently depicted in European paintings, especially by Dutch and Flemish artists, from about 1615 into the 18th century. All the evidence indicates that they were commercial products. Although many are small, some are more than 50 feet long.

Portuguese carpets (cf. Survey of Persian Art, pls. 1216-17). A particularly enigmatic class includes the “Portuguese carpets,” so called because the sailing vessels and figures in European costumes depicted on them have been identified as Portuguese. These maritime scenes are crowded into the small corners of the field, the only space left by the main design, which consists of a diamond-shaped central zone with concentric, serrated extensions in different colors. The conventional border design consists of palmettes and a broad, strap-like arabesque. Portuguese carpets feature wool pile on cotton foundations, and the coloring tends to be bright and to include large amounts of pink, light blue, and tan. Various places in Persia and also in India have been suggested as the place of manufacture, but a definitive determination has yet to be made. Certain aspects of the maritime scenes recall conventions in 10th/16th-century and early 11th/17th-century Persian and Indian painting, but specific identification of the scenes remains elusive. It may be that no specific incident is represented and that the maritime imagery was included simply to appeal to Europeans for whom the carpets may have been made.

See also anhalt carpet.

Bibliography : Abu’l-Fażl ʿAllāmī, Āʾīn-e akbarī, tr. H. Blochmann: The Ain i Akbari, 2 vols., Calcutta, 1873. E. Atil, The Age of Suleyman the Magnificent, Washington, D.C., 1987. G. Barbaro and A. Contarini, I viaggi in Persia degli ambasciatori veneti Barbaro e Contarini, ed. L. Lockhart et al., Rome, 1973; Eng. tr. Travels to Tana and Persia. A Narrative of Italian Travels in Persia in the 15th and 16th Centuries, The Hakluyt Society, series 1, no. 49, London, 1873. M. Beattie, Carpets of Central Persia, Sheffield, 1976. Idem, “Ardabīl Carpet,” in EIr. II/4, 1986, pp. 365-68. W. von Bode and E. Kuhnel, Antique Rugs from the Near East, 4th ed., tr. C. G. Ellis, London, 1970. A. Briggs, “Timurid Carpets,” Ars Islamica 7, 1940, pp. 20-54; 11-12, 1946, pp. 146-58. S. V. R. Cammann, “Symbolic Meanings in Oriental Rug Patterns,” Textile Museum Journal 3/3, 1972, pp. 5-54. A. J. D. Campbell, Report on carpets in the collection of the Maharaja of Jaipur, 1929, unpublished. J. Chardin, Les voyages du chevalier Chardin en Perse, et autres lieux de l’Orient, ed. L. Langles, 10 vols., Paris, 1811. C. P. Clark, ed., Oriental Carpets, 2 vols., Vienna, 1892-96. M. Dimand, “A Persian Garden Carpet in the Jaipur Museum,” Ars Islamica 7, 1940, pp. 93-96. Idem and J. Mailey, Oriental Rugs in The Metropolitan Museum of Art, New York, 1973. The Eastern Carpet in the Western World from the 15th to the 17th Century, London, 1983. C. G. Ellis, “Some Compartment Designs for Carpets and Herat,” Textile Museum Journal 1/2, 1965, pp. 42-56. Idem, “The Portuguese Carpets of Gujarat,” in R. Ettinghausen, ed., Islamic Art in the Metropolitan Museum of Art, New York, 1975, pp. 267-89. Idem, Oriental Carpets in the Philadelphia Museum of Art, Philadelphia, 1988. K. Erdmann, “Persische Wirkteppiche der Safidenzeit [sic],” Pantheon 10, 1932, pp. 227-31. Idem, “"The Art of Carpet Making," in A Survey of Persian Art. Rezension,” Ars Islamica 8, 1941, pp. 121-91. Idem, Seven Hundred Years of Oriental Carpets, tr. M. Beattie and H. Herzog, Berkeley and Los Angeles, 1970. Idem, Oriental Carpets, tr. C. G. Ellis, Fishguard, Wales, 1976. E. Grube, The World of Islam, London, 1966. R. Hakluyt, “Certaine Directions Given by M. Richard Hackluit of the Middle Temple, to M. Morgan Hubblethorne, Dier, Sent into Persia, 1579,” in R. Hakluyt, The Principal Navigations Voyages Traffiques and Discoveries of the English Nation III, Glasgow, 1903, pp. 249-51. J. von Hammer, Geschichte des osmanischen Reiches, 10 vols., Budapest, 1827-35. E. Kaempfer, Am Hofe des persischen Grosskönigs (1864-85), tr. W. Hinz, Leipzig, 1940. M. Keyvani, Artisans and Guild Life in the Later Safavid Period, Berlin, 1982. J. H. van Linschoten, The Voyage of Jan Huyghen van Linschoten to the East Indies, ed. and tr. A. C. Burnell and P. A. Tiele, 2 vols., London, 1885. T. Mańkowski, “Note on the Cost of the Kashan Carpets at the Beginning of the Seventeenth Century,” Bulletin of the American Institute for Persian Art and Archaeology 4, 1936, pp. 152-53. Idem, “The Art of Carpet Making. B. Some Documents from Polish Sources Relating to Carpet Making in the Time of Shah ʿAbbās I,” in A Survey of Persian Art III, pp. 2431-36. P. Mellbye-Hansen, “The Carpets in Denmark,” in M. Bencard, ed., Denmark’s Coronation Carpets, Copenhagen, 1987, pp. 7-28. A. Olearius, The Voyages and Travels of the Ambassadors Sent by Frederick Duke of Holstein, to the Great Duke of Muscovy, and the King of Persia, tr. J. Davies, London, 1662. A. U. Pope, “The Art of Carpet Making. A. History,” in A Survey of Persian Art III, pp. 2257-2430. F. Sarre and F. R. Martin, Die Ausstellung von Meisterwerken muhammadanischer Kunst in München 1910, 3 vols., Munich, 1912. F. Sarre and H. Trenkwald, Old Oriental Carpets, tr. A. F. Kendrick, 2 vols., Vienna and Leipzig, 1926-29. G. Silva y Figueroa, L’ambassade de D. Garcias de Silva Figueroa en Perse, tr. de Wicqfort, Paris, 1667. F. Spuhler, Seidene Repräsentationsteppiche der mittleren bis späten Safawidenzeit—Die sogennanten Polenteppiche, Ph.D. dissertation, Berlin, 1968. Idem, “Bisāṭ. i. Carpets of the Central and Eastern Islamic Lands,” in EI², Suppl., 1981, pp. 136-44. Idem, “Carpets and Textiles,” in Camb. Hist. Iran VI, pp. 698-720. Idem, Oriental Carpets in the Museum of Islamic Art, Berlin, tr. R. Pinner, Washington, D.C., 1987. R. Steel and J. Crowther, “A Journall of the Journey of Richard Steel and John Crowther, from Azmere in India . . . to SpahaŋAnn. 1615. 1616.,” in S. Purchas, Hakluytus Posthumus or Purchas His Pilgrimes IV, Glasgow, 1905, pp. 266-77. J. B. Tavernier, Les six voyages de Jean Baptiste Tavernier, 2 vols., Paris, 1679. P. Teixeira, The Travels of Pedro Teixeira, ed. and tr. W. F. Sinclair, London, 1902. Travels to Tana and Persia by Josafa Barbaro and Ambrogio Contarini, tr. W. Thomas and S. A. Q. Roy, ed. Lord Stanley of Alderley, London, 1873. W. E. Weaver, “The Ardabil Puzzle,” Textile Museum Journal 23/1, 1984, pp. 43-51. “Carpet Making in the Time of Shah ʿAbbās I,” in A Survey of Persian Art III, pp. 2431-36. P. Mellbye-Hansen, “The Carpets in Denmark,” in M. Bencard, ed., Denmark’s Coronation Carpets, Copenhagen, 1987, pp. 7-28. A. Olearius, The Voyages and Travels of the Ambassadors Sent by Frederick Duke of Holstein, to the Great Duke of Muscovy, and the King of Persia, tr. J. Davies, London, 1662. A. U. Pope, “The Art of Carpet Making. A. History,” in A Survey of Persian Art III, pp. 2257-2430. F. Sarre and F. R. Martin, Die Ausstellung von Meisterwerken muhammadanischer Kunst in München 1910, 3 vols., Munich, 1912. F. Sarre and H. Trenkwald, Old Oriental Carpets, tr. A. F. Kendrick, 2 vols., Vienna and Leipzig, 1926-29. G. Silva y Figueroa, L’ambassade de D. Garcias de Silva Figueroa en Perse, tr. de Wicqfort, Paris, 1667. F. Spuhler, Seidene Repräsentationsteppiche der mittleren bis späten Safawidenzeit—Die sogennanten Polenteppiche, Ph.D. dissertation, Berlin, 1968. Idem, “Bisāṭ. i. Carpets of the Central and Eastern Islamic Lands,” in EI², Suppl., 1981, pp. 136-44. Idem, “Carpets and Textiles,” in Camb. Hist. Iran VI, pp. 698-720. Idem, Oriental Carpets in the Museum of Islamic Art, Berlin, tr. R. Pinner, Washington, D.C., 1987. R. Steel and J. Crowther, “A Journall of the Journey of Richard Steel and John Crowther, from Azmere in India . . . to Spahaṇ . . . Ann. 1615. 1616.,” in S. Purchas, Hakluytus Posthumus or Purchas His Pilgrimes IV, Glasgow, 1905, pp. 266-77. J. B. Tavernier, Les six voyages de Jean Baptiste Tavernier, 2 vols., Paris, 1679. P. Teixeira, The Travels of Pedro Teixeira, ed. and tr. W. F. Sinclair, London, 1902. Travels to Tana and Persia by Josafa Barbaro and Ambrogio Contarini, tr. W. Thomas and S. A. Q. Roy, ed. Lord Stanley of Alderley, London, 1873. W. E. Weaver, “The Ardabil Puzzle,” Textile Museum Journal 23/1, 1984, pp. 43-51.

(Daniel Walker)

The rug production of Persia in the Afsharid (q.v.; 1148-61/1736-48) and Zand periods (1163-1209/1750-79) has been difficult both to identify and to document, owing to the lack of extant examples and the absence or misuse of historical and archival evidence. The generally held view has been that, in the political and economic decline resulting from the invasion of Persia by Nāder Shah Afšār, beginning in 1141/1727, rug production came almost to an end (Edwards, p. 5). Such concomitant factors as the destruction of urban centers, forced migrations, and other large population movements have been considered the immediate causes of the apparent establishment of new production centers and the diffusion of carpet designs, especially in northwestern Persia and the Caucasus. This explanation, which has caused much confusion in carpet studies, is based on the commonly accepted notion that Nāder Shah resettled large groups of carpet weavers both in the Caucasus and in west central Persia and Fārs (Martin, pp. 84, 86; Survey of Persian Art VI, p. 2356 and n. 2; Yetkin, I, pp. 90-91).

A careful review of 12th/18th-century histories, travelers’ accounts, and trade records suggests a somewhat different picture of contemporary Persian carpet production, however. Historical evidence suggests that forced migrations during this period were instituted largely in order to remove troublesome elements from the western frontier; Nāder Shah thus sought to shift certain western tribal groups to Khorasan while at the same time securing the latter area against enemy incursions and providing himself with manpower for his campaigns. These migrations did not have serious long term effects, however, for, according to a recent study, the resettled populations returned to their places of origin after Nāder Shah’s death in 1160/1747 (Perry, pp. 208-09).

Although it is probable that magnificent silk-and-brocade rugs in the style of the Safavid court manufactories (see ix, above) were no longer produced in significant quantities, it seems reasonable to assume that production of less luxurious wool rugs continued in many traditional centers, even though on a smaller scale and mainly for domestic consumption, rather than for export. It is also likely that tribal rugs were still woven and that felts and other types of nonpile floor coverings, both local and imported, were still in demand.

According to an agent of the French Compagnie des Indes, in 1150/1738 “beautiful carpets of silk and wool” were available in the bāzār at Bandar-e ʿAbbās (Housego, p. 40). Nāder Shah himself ordered rugs and felt floor coverings from the city of Kermān to redecorate the church of Yerevan (Īravān) in Armenia and the shrine of ʿAlī at Najaf in Mesopotamia (Marvī, I, p. 410, III, p. 892). Such a sizable order seems to indicate active production in this center. He may also have ordered similar rugs for the shrine of Imam ʿAlī al-Reżā at Mašhad, which he repaired and improved (see āstān-e qods), as well as for his treasure house at Kalāt(-e Nāderī) in Khorasan and the small palace that he built in Qazvīn. The grandiose royal tent of silk installed for his coronation in 1148/1736 at Dašt-e Moḡān in Azerbaijan was strewn with silk carpets and carpets from Kermān (Marvī, II, p. 454).

The renewed local prosperity and stability that accompanied the reign of Moḥammad Karīm Khan Zand in Shiraz and southern Persia (1163-93/1750-79) must have been favorable for rug production there. Indeed, according to one of his contemporaries Karīm Khan himself claimed to be skilled in carpet weaving, as well as in several other métiers (Āṣaf, p. 309). The belief that carpet production continued is confirmed by a wool carpet in the Mūza-ye Īrān-e Bāstān, Tehran, with a design consisting of a single-plane lattice with flowering plants in the compartments; the knotting technique is the same as that associated with vase carpets (the “vase technique”; see iii, above). The date 1172/1758 and the name of the weaver, Moḥammad Šarīf Kermānī, are woven into an epigraphic cartouche (Housego, p. 44; cf. Pope, p. 2266). This dated example can thus serve as the nucleus for a larger group of related rugs that are not precisely dated. A number of extant rugs in “millefiori” and “paisley” patterns (see iv, above) resembling rugs illustrated in contemporary paintings or with color schemes and designs paralleled in other Zand decorative media, have also been suggested as possible examples of Shiraz and Kermān production. Some rugs may also have been produced in Khorasan in this period (Housego, pp. 46-50).

By the 1780s and 1790s East India Company merchants in Persia were reporting the availability of carpets in Būšehr (Bushire; Issawi, p. 88) and extolling Khorasan carpets for their “brightness of colors and elegance of workmanship” (Housego, p. 48). Other travelers and diplomats in the 12th/18th century made only brief references to rugs: In 1786 it was noted that Yazd and Kermān were exporting felts and carpets (Franklin, p. 148); in the late 1780s beautiful carpets woven in Khorasan, some even with grounds of gold thread, were noted (Ferrières de Sauvebœuf, II, p. 8). In 1801, in a confidential report on the manufactures, exports, and imports of Persia made by Sir John Malcolm to the director of the board of control of the East India Company, Yazd, Kāšān, Ṭabas, and cities in independent Khorasan are mentioned as carpet-manufacturing centers (Malcolm apud Issawi, pp. 262-63; cf. Hambly, p. 81). In 1826 an immensely useful and detailed account of Persian commerce was published; it shows that carpets were being produced in sizable numbers in both traditional and new centers: Herat, Kermān, Yazd, Borūjerd, the Turkman areas of Khorasan, Isfahan, and Azerbaijan (Fraser, p. 362).

There thus seems to be enough evidence to establish continuity of rug production throughout the Afsharid and Zand periods; indeed, there is insufficient evidence for claiming the contrary. There is, however, a clear need for much more careful correlation between historical evidence and the dating and attribution of extant rugs to the 12th/18th century. Attention should also be focused on the scale of production and types of design, as well as on attributions to specific weaving centers. It is probable that many “classical” designs of the 11th/17th century (e.g., those on garden, lattice-and-vase, and floral-directional carpets; see iv and ix, above) were continued into the later period and that some new ones were also introduced (e.g., millefiori, Plate CXIII; repeat bota, Figure 71 above in iv; and mīnāḵānī designs, Edwards, pp. 42-43, fig. 25). It may in fact be suggested that the “movement” of Safavid lattice, floral, and garden designs into northwestern Persia and the Caucasus in the 12th/18th century resulted, not from undocumented migrations of weavers, but from the “inspiration of great and colorful designs themselves” (Beattie, p. 70), like those on the carpets ordered by Nāder Shah to be sent to Erevan.

The decline in luxury rug production was as much a reflection of the overall decline of the Persian economy as of the depredations of the Afghans and Nāder Shah’s military campaigns; it seems to have been brought about by growing European control of the maritime trade (Hambly, p. 77), a shrinking European market for Oriental rugs in the late 11-12th/17-18th centuries (Housego, p. 40), and a drop in the production of raw silk in Persia.

Rugs with traditional Persian designs but woven in local techniques continued to be produced in border areas that were no longer directly under Persian political control. They included the floral and hunting rugs of the Caucasus, rugs with interwoven Armenian inscriptions and Armenian calendar dates of the 12th/18th-century, and the rugs of Herat, which was by then part of Afghanistan. Examples of the last group were even included in the list of imports from Afghanistan to Persia in the early 13th/19th century (Hambly, p. 79).

Bibliography : M. H. Āṣaf, Rostam al-tawārīḵ, ed. M. Mošīrī, 2nd ed., Tehran, 1313 Š./1935. M. H. Beattie, Carpets of Central Persia, Westerham, Kent, 1976. A. C. Edwards, The Persian Carpet. A Survey of the Carpet Weaving Industry in Persia, London, 1953. L. F. de Ferrières de Sauvebœuf, Mémoires historiques, politiques et géographiques des voyages du comte de Sauvebœuf faites en Turquie, en Perse et en Arabie, depuis 1782 à 1789, Maastricht, 1790. W. Franklin, Observations Made on a Tour from Bengal to Persia in the Years 1786 to 1787, London, 1790. J. B. Fraser, Travels and Adventures in the Persian Provinces on the Southern Banks of the Caspian Sea, London, 1826 (esp. appendix II, pp. 352-38). G. Hambly, “An Introduction to the Economic Organization of Early Qajar Iran,” Iran 2, 1964, pp. 69-81. J. Housego, “18th Century Persian Carpets. Continuity and Change,” Oriental Carpet and Textiles Studies 3/1, 1987, pp. 40-51. C. Issawi, ed., The Economic History of Iran 1800-1914, Chicago, 1971. F. R. Martin, A History of Oriental Carpets before 1800, Vienna, 1908. M.-K. Marvī, ʿĀlamārā-ye nāderī, ed. M. A. Rīāḥī, Tehran, 1345 Š./1966. J. Perry, “Forced Migrations in Iran during the 17th and 18th Centuries,” Iranian Studies 8/4, 1975, pp. 199-215. A. U. Pope, “The Art of Carpet Making. A. History,” in Survey of Persian Art VI, pp. 2257-2430. S. Yetkin, Early Caucasian Carpets in Turkey, 2 vols., London, 1978.

(Layla S. Diba)

During the Qajar period there were dramatic alterations in the traditional organization and orientation of the Persian carpet industry and, consequently, in Persian carpets themselves. Particularly significant was the substantial increase both in the number of looms and in the volume of carpet exports from the 1290s/1870s to World War I. (For a summary of visible carpet exports from Persia in the years 1885-1914, compiled from available British consular and commercial reports, see Ittig, 1983, I, table I.) Although the literature on Qajar carpet production is focused almost exclusively on external and imposed factors, the indigenous, locally funded industry, which provided the organizational model for export manufacture both before and after this “boom” period, must also be considered.

The locally financed industry. In publications on Persian carpets 13th/19th-century production before the boom is generally categorized as “an isolated, largely rural craft, practiced primarily by pastoral nomads and carried on only in a few urban locations” (Helfgott, p. 117). Yet Persian and Western primary sources, as well as inscriptions on extant examples, confirm that carpets were produced for both domestic use and sale in urban, rural, and tribal settings in most parts of Persia throughout the Qajar period. This industry consisted of a mosaic of regional units; the centers most frequently mentioned include Kermān, Borūjerd, Isfahan, Azerbaijan, Kurdistan, Kāšān, Farāhān, Hamadān, Herat, and Mašhad (Blau, p. 104; Fraser, 1826, pp. 354, 356, 362; U.K., House of Commons, Parliamentary Papers LXVII, p. 107; Ferrier, pp. 26 n. 125; Burgess, 1942, p. 4; Amanat, 1983, pp. 146, 160; Polak, II, p. 168; Armstrong, p. 152; Kinneir, p. 115; Malcolm, p. 25). Each locale produced its own distinctive weaves and patterns, for example, the herātī-patterned rugs (see iv, above) of Farāhān and the arabesque carpets of Garrūs (Brugsch, II, p. 13 and ill.; Ittig, 1983, I, appendix; idem, 1981, pp. 125-26 and figs. l, 3, 5, 6; for other regional types, see Polak, II, p. 168; Goldsmid et al., p. 99; Malcolm, p. 25; Fraser, 1826, p. 356; 1838, II, p. 333; Blau, p. 104). The importance of regionalism in Persian carpet weaving is further demonstrated by the persistence of local designs, structures, and traditional palettes even after the introduction of Western-influenced decorative themes, machine-spun yarns, and synthetic dyestuffs in the later decades of the 13th/19th century (e.g., Ittig, 1981, table I and figs. 1, 3, 5; Housego, 1978, pls. 45, 97, 125, 126; for specific local carpet structures see Edwards; Neff and Maggs). Undoubtedly the variety of both sedentary and nomadic ethnic and tribal groups in Persia, as well as the physical geography, which made interregional communication among regions difficult, contributed to and reinforced the development of local weaving styles.

Despite the variety in types of carpets woven, certain patterns of production and finance were observed throughout the country both before and after the boom. Pieces not made for the weaver’s own use were either commissioned directly or manufactured commercially for more general markets, both in Persia and abroad. Commercial production was sufficiently diversified so that in weaving centers like Mašhad and Farāhān more than one quality, or line, of carpet was manufactured to suit a range of customers (Blau, p. 106; Ittig, 1983, I, appendix). Both commercial and custom weaving was conducted in cottage, as well as workshop, settings (Armstrong, p. 152; Blau, p. 106; Churchill, p. 2; Brugsch, II, p. 13; Goldsmid, p. 280; Sayyāḥ, p. 164; Ittig, 1985, p. 119). At least some tribes also worked on commission and were involved in commercial production (Amanat, p. 116; Blau, p. 106; Ker Porter, I, p. 265).

Carpet manufacturing, whether specially commissioned or for a mass market, required inputs of material and labor (Ittig, 1985, p. 118); there was an implicit contractual arrangement between entrepreneur and weaver for supply of materials, completion of each carpet, and payment (Wazīrī, p. 189; Sayyāḥ, p. 164; Ittig, 1985, p. 118 and figs. 1, 5, 7). Until the 1290s/1870s it appears that both commercial and custom production were financed locally, either by Persian notables—court figures, provincial dignitaries (Ferrier, p. 26 n.; Blau, p. 106; Āzādī, 1978, no. 32; Gans-Ruedin, pp. 384-85; Ittig, 1983, II, no. 4), and tribal khans (Blau, p. 106); merchants (Blau, p. 105); or loom owners, ostāds (lit., “masters”; Gans-Ruedin, pp. 426-27; Ittig, 1983, II, no. 29). The economic basis for the involvement of these three groups in carpet manufacture can be summarized as follows.

Provincial notables owned both the requisite raw materials and the transport animals necessary to collect and distribute these materials and the completed carpets (Ittig, 1983, I, pp. 48-49; idem, 1985, p. 118). Not only did this group finance commercial production (Ittig, 1985, p. 112; Blau, p. 106), but sources and surviving examples of rugs with commission inscriptions from the Qajar period (e.g., Wazīrī, p. 189; Goldsmid, p. 280; Āzādī, no. 32; Gans-Ruedin, pp. 384-85; Ittig, 1983, II, no. 4; 1985, fig. 1) also suggest that it provided the principal patronage for custom weaving. Custom and commercial carpet manufacturing and distribution by merchants were funded through established commercial and credit networks (Blau, p. 105; Ittig, 1983, I, p. 49; Ittig, 1985, p. 119 n. 32 and figs. 7-9; Polak, II, p. 169; White, II, p. 121). The importance of individual loom owners in the production process (Ittig, 1983, I, p. 50) is reflected in substantial numbers of inscribed Qajar carpets on which specific ostāds are named as manufacturers, either singly (evidence of involvement in commercial production) or in combination with those who commissioned particular carpets (e.g., Āzādī, no. 29; Gans-Ruedin, pp. 426-31, 436-39; Ittig, 1985, fig. 1). Unfortunately, though references to various quantities, qualities, and prices of rugs available in local bāzārs and carpet anbārs (warehouses) suggest that a large portion of production was oriented to market demand even before the boom (see, e.g., Brugsch, II, p. 13; Blau, pp. 104, 106), there are no known statistics on the volume of either custom or commercial carpet weaving financed by local sources at this time.

Little is known about the hierarchy of craftspeople involved either in preparation of raw materials for carpet manufacture or in the actual weaving process before the 1290s/1870s, but references to the spinning and dyeing of yarns for use in carpet workshops, differentiation in wages between senior and apprentice weavers, and preparation and use of cartoons (Sayyāḥ, p. 164; Goldsmid, p. 280) demonstrate that a highly specialized personnel structure was already in existence before the boom and probably even before the Qajar period.

Nor are there quantitative data on either domestic or international trade in Persian carpets before the 1280s/1860s. Certainly carpets were transported from one province to another (U.K. Foreign Office Records, Public Record Office, 60.92, enc. no. 1; Burgess, p. 4). Indeed, it appears from the sources that during the first half of the century carpets were among the chief articles of internal commerce, along with silks, cottons, and shawls (U.K. Foreign Office Records, Public Record Office, 60.92, enc. no. 1; Malcolm, p. 24; Polak, II, pp. 168-69; Blau, pp. 104ff.). Accounts from the same period indicate that most carpets exported from Persia were initially destined for India, Turkey, or Russia (Malcolm, p. 26; Shoberl, p. 24; Jaubert, p. 288; Fraser, 1826, pp. 354, 362; Pottinger, p. 226; U. K. Foreign Office Records, Public Record Office, 60.117, no. 25 and enc. no. 2; Blau, p. 106). Although, according to contemporary observers, the volume of Persian carpets on the Constantinople market declined during the 1250-70s/1840-50s (White, II, p. 118), values cited for visible Persian carpet exports to Constantinople through Trebizond in 1266/1850 and 1272/1856 indicate that trade was still significant in that period (Blau, p. 105). While there are unfortunately no statistics, at least some of these carpets were subsequently forwarded to Europe; but traditional Persian design formats and dimensions were generally considered more suited to Eastern than to Western interiors and tastes (Blau, pp. 105-06; Polak, II, p. 169).

By the middle of the century Persian carpets faced competition in their traditional export markets due to the development of substantial pile-carpet industries in both Turkey and India. At the same time, however, European and American demand became significant, and carpets therefore became more important as exports to these new markets (Quataert, p. 474 and table I; Ittig, 1983, I, table I). The rapid rise in both numbers of looms and carpet exports from Persia beginning in the 1290s/1870s, already noted, reflects the general growth of the industry. These figures cannot be considered complete, since, as attested in numerous references widespread smuggling also occurred; nevertheless, they do demonstrate that carpets were becoming more prominent in the international trade of Persia.

The Persian carpet “boom.” One of the most common explanations in the literature for the general increase in carpet production during the third quarter of the 13th/19th century is that the “boom” resulted from a search by Western exporters of manufactured goods to Persia for profitable commodities to replace silk, the single most important visible Persian export before 1281/1865. An outbreak of the silkworm disease pébrine in that year had sharply reduced the amount of silk available, and it has been argued that consequently “income from the sale of carpets supplanted that once derived from the export of silk” (Bier, p. 254). This hypothesis does not, however, take into consideration the importance of regionalism and regional economies in Persia; a surplus of imports in one region was thus not necessarily offset by a surplus of exports in another (Gilbar, pp. 328-29). The local response to pébrine in the Caspian provinces and northeastern and central Persia was to increase cultivation of such cash crops as rice (see berenj), opium (see afyuᵛn), tobacco, and cotton (q.v.). Furthermore, it appears that even in northern Persia, where silk was the most important visible export, it may already have been equaled or surpassed in the balance of payments by specie before 1281/1865 (U.K. Foreign Office Records, Public Record Office, 60.21, 30 October 1820; idem, 60.75, 29 October 1840; Burgess, p. 85; U.K., House of Commons, Parliamentary Papers LXIV, p. 201). Furthermore, the impact of the carpet boom was not uniformly felt throughout Persia; rather, different regions were affected at different times (Ittig, 1983, I, pp. 256-57; Ittig, 1985, pp. 116-17).

It appears, in fact, that the major catalyst in the Persian carpet boom was a growing international market for these products, directly linked with the increasing strategic and commercial importance of the Middle East to the industrialized nations. Factors contributing to the popularization of both antique and contemporary Oriental carpets included the great world’s fairs and major museum exhibitions, the contemporary Arts and Crafts movement in England and comparable phenomena in Europe, and resulting changes in furnishing fashions (Ittig, 1985, p. 118; Bierman, pp. 53-59; Quataert, p. 474; Reitlinger, II, pp. 520-2l; Ittig, 1983, I, pp.70-72). The growing purchasing power of the Western middle classes fueled this demand (Ittig, 1985, p. 118; Hobsbawn, p. 160).

To satisfy the rise in demand for Persian carpets in turn required capital investment, but contemporary observers suggested that, as local Persian capital was not adequate to support expansion beyond traditional parameters (U.K. Foreign Office Records, Diplomatic and Consular Reports, A.S. no. 1376, p. 57; Ittig, 1985, p. 122 and n. 39), the gap was filled by funds from outside groups. Prominent among such investors were Persian, particularly Azerbaijani, wholesalers, import-export traders (tojjār), and merchant bankers (ṣarrāfs), as well as Western importers. It was through investment from these sources that the weaving industry was able to expand. Furthermore, the participation of such Persian and Western investors in international trade and banking made it possible to integrate the manufacture and export of Persian carpets into previously established systems of collection and distribution (Ittig, 1985, pp. 119-21 ).

Although the exact date when large-scale production specifically for a Western clientele commenced is uncertain, by at least as early as 1274/1858 Persian carpet manufacturers had begun to alter the traditional dimensions and designs of their carpets for that market (Blau, p. 105). In 1283/1866 rugs were first listed as an export commodity separate from other woven articles in British Foreign Office reports on Persia (U.K., House of Commons, Parliamentary Papers LXI), but it is not known whether or not these items differed in any way from those woven for the domestic Persian market. By 1290/1873, however, some regions were definitely producing for the Western market (U.K., House of Commons, Parliamentary Papers LXXV, p. 206). The organization of this export-oriented industry was initially undertaken by native, rather than foreign, investors, at least, in Tabrīz and Kermān (Edwards, pp. 55-56; Ittig, 1985, pp. 119-120). Furthermore, contemporary accounts and inscriptions on carpets provide evidence that Persian entrepreneurs followed extant domestic production models, arranging contracts with local ostāds and retaining the traditional hierarchy of craftsmen while expanding and reorienting these regional industries to the new markets (Ittig, 1985, p. 120 and figs. 7-8). Notable among tojjār and ṣarrāfs engaged in carpet manufacture and export during (and after) the Qajar period were the Dīlmaḡānī, Ṣadaqīānī, and Mīlānī families (Ittig, 1985, pp. 119-20 and fig. 9; Edwards, p. 56 n. 2; cf. xiii, below).

The earliest known effort to organize rug manufacture in Persia with foreign capital was the Ziegler enterprise in Solṭānābād (now Arāk). Ziegler’s was a firm of Manchester-based factors primarily involved in the import to the Near and Middle East of European printed cottons and the export of a wide variety of raw and manufactured goods for remittances. The firm’s first Persian agencies were established in Tabrīz and Rašt in 1284/1867. Although the common hypothesis of a connection between pébrine and the carpet boom has often been applied specifically to Ziegler’s organization of a cottage industry in Solṭānābād, the facts that the firm did not establish its first Persian agencies until two years after the onset of the disease and that it subsequently expanded commercial operations (including the exporting of silk) in Gīlān during the 1290s/1870s (U.K., House of Commons, Parliamentary Papers XLVII, p. 363; U.K. Foreign Office Records, Public Record Office, 60.401, no. 13) are evidence that it was not seeking alternative commodities to silk for export. Rather, the decision to reinvest capital in carpet production at Solṭānābād appears to have been influenced primarily by contemporary growth in Western demand for Oriental carpets (Ittig, 1983, pp. 133-36).

Ziegler’s initial involvement in the Persian carpet trade was limited to the purchase and export of both new and “used” rugs that were available on the market (Ittig, 1983, I, p. 138 and appendix); undoubtedly the most famous antique carpets to be handled by the firm were the so-called Ardabīls (see ardabīl carpet). The necessity for a dependable supply of suitable carpets to meet the ever-increasing demands of Western retailers was a major factor in establishment of an agency in Solṭānābād between 1294/1877 and 1299/1882 (Ittig, 1983, I, pp. 139-40 and notes 33, 37). The choice of this town was influenced not only by its role as market center for a major carpet-weaving area but also by its position on the route between Tehran and Baghdad via Hamadān and Kermānšāh, through which the company had already begun to import goods into Persia.

Ziegler’s manufacturing activities involved the already-existing locally financed cottage industry. The firm negotiated agreements with local agents (ʿāmels; Dehqān, II, pp. 165ff.), through whom it provided raw materials, including dyed wool and patterns (Plate CXIV, Plate CXV) prepared by specialized dyers and designers engaged by Ziegler’s, and cash advances toward the finished carpets (Ittig, 1983, I, pp. 165-66, 175-78; Whigham, p. 5; U.K. Foreign Office Records, Diplomatic and Consular Reports, A.S. no. 1376, pp. 57, 59). The ʿāmels then distributed the materials to the weavers, who were often their own kinswomen (Aubin, p. 315; Dehqān, II, pp. 165ff.; U.K. Foreign Office Records, Diplomatic and Consular Reports, A.S., no. 1376, p. 57). They also supervised the weaving process and collected the finished carpets for delivery to Ziegler’s compound in Solṭānābād, where the rugs were carefully inspected before being packed for export.

Ziegler’s production was oriented particularly toward European and American consumers, and the firm altered traditional palettes, dimensions, and designs to the specifications of Western retailers (see below; U.K. Foreign Office Records, Diplomatic and Consular Reports 69, p. 499; Dehqān, II, pp. 166-68). Although such changes were deplored as contributing to the commercialization and deterioration of the Persian carpet (Whigham, p. 5), Ziegler’s product sold extremely well, and by 1311/1894 the firm had expanded its operations to include 111 villages around Solṭānābād (see below; U.K. Foreign Office Records, Diplomatic and Consular Reports, A.S. no. 1376, p. 58). The firm’s success encouraged other Western entrepreneurs to invest in the carpet industry, and before World War I such foreign firms as Oriental Carpet Manufacturers, Nearco Castelli & Brothers, and Persische Teppichgesellschaft-AG were in competition in various carpet-weaving centers in Persia (see xii, below). Ziegler’s organization of a cottage industry provided the model for these “second generation” companies, but they operated with a significant difference: In addition to organizing cottage and workshop production through local agents (e.g., Ittig, 1985, pp. 120-21 ), they also established their own weaving factories (e.g., U.K. Foreign Office Records, Public Record Office, 248.1147, no. 42, p. 16), which permitted greater quality control through more immediate supervision of weavers. The factory weavers were salaried (U.K. Foreign Office Records, Public Record Office, 368.38, file 393R, pp. 93-94), which obviated the risk inherent in making advances payments when little redress was possible in instances of default (U.K. Foreign Office Records, Public Record Office, 248.1177, 24 October 1917).

Production of carpets in Persia for export was influenced by fluctuations in supply and demand on the world market, as well as by the state of international trade generally. The industry thus suffered a slump in 1322/1904, a period of worldwide recession, which was aggravated by a surplus of inventory on the British market (Sykes, 1906, p. 427; U.K. Foreign Office Records, Public Record Office, 368.38, file 3938, pp. 50, 94; U.K. Foreign Office Records, Diplomatic and Consular Reports, A.S. 3374, pp. 5-6). The industry soon recovered, however, and the value of Persian carpet exports seems to have peaked in 1324/1906-07 (Ittig, 1983, I, table I), despite considerable domestic insecurity in Persia during the period of the Constitutional movement (q.v.). Production was interrupted during World War I, and hostilities between German- and British-protected interests in Persia resulted in substantial losses for both local and foreign carpet manufacturers there (e.g., U.K. Foreign Office Records, Public Record Office, 248.1147, 42/16, 42/206, 76). With resumption of large-scale manufacturing for export after the war an even larger proportion of production was geared to the American market (Edwards, pp. 92-93, 136, 140-41, 208), though locally financed commercial and custom carpet weaving continued throughout the period.

The Qajar carpet. The volume of extant Persian rugs securely datable to the Qajar period permits a number of general observations on carpet design. Formats and motifs derivative from those of the Safavid period, particularly centralized-medallion, medallion-and-corner, and lattice designs in arabesque and Shah ʿAbbāsī patterns were woven throughout this period (see iv, above). Indeed, fine carpets woven in designs reminiscent of the classical period were produced into the late 13th/19th and early 14th/20th centuries (for example, Ittig, 1981, figs. 1-5; Ittig, 1983, II, no. 4 and pls. 28, 29; Gans-Ruedin, pp. 164-65); such pieces were generally specially commissioned and may represent local survivals of pre-Qajar workshop production. In contrast are those carpets with “classical” designs deliberately copied from older pieces or illustrations in carpet books and museum and sales catalogues, which were represented as antiques to meet a particular market demand (see above); such rugs were often treated in chemical baths and by other methods to make them appear older (Sayyāḥ, p. 192; Ittig, 1983, I, p. 245).

Depictions of carpets in early 13th/19th-century miniatures and paintings provide a terminus ante quem for the incorporation of such small repeat patterns as the herātī and mīnāḵānī into the design repertory (Falk, 1972, fig. 2 and pls. 19, 20, 25). The offset bota repeat pattern so popular in both Qajar textiles and carpets may have been introduced during the Zand period (Housego, 1986, p. 48, fig. 9; Parhām, n.p.; see iv, x, above). Descriptions of the dwellings of notables in the early 13th/19th century attest that carpets from Herat were also fashionable at the time (Fraser, 1826, p. 60; idem, 1838, II, p. 56; Ker Porter, I, p. 235), though, surprisingly, carpets with field patterns generally categorized as “Herat” (or “Indo-Persian”; see ix, above) do not seem to be illustrated in contemporary Persian paintings.

In addition to nonfigural patterns, Qajar carpets also exhibit a multitude of pictorial motifs, of both local and foreign derivation. Among those of Persian origin the most common are based on literary themes, particularly scenes from Ferdowsī’s Šāh-nāma and Neẓāmī’s Ḵamsa (Gans-Ruedin, 1978, pp. 192-93, 446-47). Also popular were depictions of historic Persian figures (Ittig, 1983, II, no. 27). The appearance of Western pictorial themes in carpets, particularly notable from the 1290s/1870s onward, was part of the contemporary fashion for European imagery in Persian art generally (see art in iran. qajar i-ii). Favorite foreign themes for carpets included scenes from classical mythology and portraits of Western historic personages, which were derived from available European newspapers, books, postcards, and paintings (Burgess, p. 93; Ittig, 1985, fig. 1). Manchester cotton goods were another source of Western designs (U.K. Foreign Office Records, Diplomatic and Consular Reports, A.S. 1376, p. 58).

As the economic importance of the export market increased in the late 13th/19th and early 14th/20th centuries, widespread adaptation of traditional designs, dimensions, and palettes was undertaken to meet the requirements of foreign consumers. Some of the earliest documentation for the introduction of foreign models into the traditional weaving repertory is provided by a series of carpet cartoons prepared for Emil Alpiger, manager of Ziegler’s Solṭānābād agency between 1883 and 1896; many of these exhibit the elaborate floral medallion-and-corner designs then fashionable in the West (Plate CXIV; cf. Ittig, 1983, I, p. 162, II, nos. 15, 17, 19). Rugs for export to the West were produced in standard dimensions suitable to European and American interiors—for example, 3 x 5, 4 x 6, 8 x 10, and 9 x 12 feet (Ittig, 1983, I, pp. 238, 258-59)—in contrast to those used in traditional formal Persian settings (Edwards, 1953, pp. 55, 125 and pl. 38).

Various types of synthetic dyes, first developed in Europe in the 1850s (Mushak, table 2), were noted in Persian carpets by the 1290s/1870s (Floyer, p. 278). Although less expensive and time-consuming to apply than natural dyestuffs (see ii, above), some of the early synthetic colorants were often unstable in light; others were not fast when washed (Mushak, 1984, p. 9). Furthermore, the vivid shades produced by these dyes were unacceptable to the American and European markets, where a softer color palette was preferred, and they were probably one of the factors in the 1322/1904 slump (F.O. 368.38, p. 94). A variety of methods was devised to soften the harsh synthetic colors, including washing in chemical baths (Ittig, 1983, I, p. 245, II, no. 37). Although efforts by the Persian government to ban the import of synthetic colorants were ineffective, as were attempts to confiscate all carpets made with artificially dyed yarns, it is not clear to what extent synthetic dyes were used, other than in lower-grade Persian rugs made for the Western market in the late 13th/19th and early 14th/20th centuries. Edwards’ observations that several carpet-weaving areas continued to use natural colorants after World War I (Edwards, 1953, passim) are supported by the variety of extant late Qajar carpets in which no artificial dyes appear to have been used.

Bibliography : A. Amanat, ed., Cities and Trade. Consul Abbott on the Economy and Society of Iran 1847-1866, London, 1983. T. B. Armstrong, Journal of Travels in the Seat of War, London, 1831. E. Aubin, La Perse d’aujourd’hui, Paris, 1908. S. Āzādī, Farš-e Īrān/Persian Carpets, Hamburg, 1978. C. Bier, “Commodity and Consumption,” in Bier, ed., Woven from the Soul, Spun from the Heart. Textile Arts of Safavid and Qajar Iran 16th-19th Centuries, Washington, D.C., 1987, pp. 253-55. I. A. Bierman, “Oriental Rugs of the Pacific Northwest 1850-1930,” The Warp and Weft of Islam, ed. J. L. Bacharach and Bierman, Seattle, 1978, pp. 53-57. O. E. Blau, Commerzielle Zustände Persiens, aus den Erfahrungen einer Reise im Sommer 1857 dargestellt, Berlin, 1858. H. Brugsch, Reise der kaiserlichen preussischen Gesandtschaft nach Persien 1860 und 1861 II, Leipzig, 1863. C. and E. Burgess, Letters from Persia, ed. B. Schwartz, New York, 1942. S. T. A. Churchill, “The Present State of the Carpet Industry in Persia,” in Oriental Carpets, ed. C. P. Clarke, Vienna, 1892, pp. 1-6. E. Dehqān, Tārīḵ-e Arāk, 2 vols., Tehran, 1330 Š./1951-52. A. C. Edwards, The Persian Carpet, London, 1953. S. J. Falk, Qajar Paintings, London, 1972. J. P. Ferrier, Caravan Journeys and Wanderings in Persia, Afghanistan, Turkistan and Beloochistan, London, 1857. E. A. Floyer, Unexplored Baluchistan, London, 1882. J. B. Fraser, Travels and Adventures in the Persian Provinces, London, 1826. Idem, A Winter’s Journey from Constantinople to Tehran, 2 vols., London, 1838. E. Gans-Ruedin, Iranian Carpets. Art, Craft and History, London, 1978. G. Gilbar, “Persian Agriculture in the Late Qajar Period, 1860-1906,” Asian and African Studies 12/3, 1978, pp. 312-65. F. J. Goldsmid, “Notes on Eastern Persia and Western Baluchistan,” JRGS 37, 1867, pp. 269-97. Idem et al., Eastern Persia. An Account of the Journey of the Persian Boundary Commission, 1870-72, London, 1876. L. Helfgott, “Production and Trade. The Persian Carpet Industry,” in C. Bier, ed., Woven from the Soul, Spun from the Heart. Textiles Arts of Safavid and Qajar Iran 16th-19th Centuries, Washington, D.C., 1987, pp. 107-20. E. J. Hobsbawn, Industry and Empire, Suffolk, 1970. J. Housego, “The Nineteenth Century Persian Carpet Boom,” Oriental Art, N.S. 19/2, 1973, pp. 169-71. Idem, Tribal Rugs, London, 1978. Idem, “18th Century Persian Carpets. Continuity and Change,” Oriental Carpet and Textile Studies 2, 1986, pp. 40-51. A. Ittig, “A Group of Inscribed Carpets from Persian Kurdistan,” Halı 4/2, 1981, pp. 124-27. Idem, A Technical and Historical Study of the Qajar Carpet Industry, 2 vols., Ph.D. dissertation, Oxford University, 1983. Idem, “The Kirmani Boom. A Study in Carpet Entrepreneurship,” Oriental Carpet and Textile Studies 2, 1985, pp. 111-23. P. A. Jaubert, Voyage en Arménie et en Perse, fait dans les années 1805 et 1806, Paris, 1821. R. Ker Porter, Travels in Georgia, Persia, Armenia . . . during the Years 1817, 1818, 1819 and 1820, 2 vols., London, 1822. J. M. Kinneir, Geographical Memoir of the Persian Empire, London, 1813. J. Malcolm, The Melville Papers, ed. A. Wilson, London, 1930. P. Mushak, “A General Technical Commentary on Synthetic Dyes in Oriental Rugs,” Oriental Rug Review 4/5, 1984, pp. 9-10. I. C. Neff and C. V. Maggs, A Dictionary of Oriental Rugs, London, 1977. S. Parhām, Namāyešgāh-e naqša-ye qālī-e Kermān, Tehran, 1356 Š./1977. J. E. Polak, Persien. Das Land und seine Bewohner. Ethnographische Schilderungen, 2 vols., Leipzig, 1865. H. Pottinger, Travels in Beloochistan and Sind, London, 1816. D. Quataert, “Machine Breaking and the Changing Carpet Industry of Western Anatolia, 1860-1908,” Journal of Social History 19, 1985-86, pp. 473-89. G. Reitlinger, The Economics of Taste, 2 vols., London, 1963. M.-A. Sayyāḥ, Ḵāṭerāt-e Ḥājj Sayyāḥ, ed. Ḥ. Sayyāḥ and Ṣ. Kāmgār, Tehran, 1346 Š./1967. E. Shoberl, Persia, London, 1822. P. M. Sykes, “A Fifth Journey in Persia,” Geographical Journal 28/5, 1906, pp. 425-53, 560-92. U.K. Foreign Office Records, Diplomatic and Consular Reports, Annual Series 69, 1882: W. J. Dickson, “The Trade of Persia.” Idem, Annual Series, 1376, 1894: J. R. Preece, “Report for the Years 1892-3 and 1893-4 on the Trade, etc., of the Consular District of Ispahan.” Idem, Annual Series, 3374: P. M. Sykes, “Report on the Trade of Kirman Consular District for the Year 1904-05.” U.K. Foreign Office Records, Public Record Office (London), F060.21, Willock to Castlereagh, 20 October 1820; F060.75, McNeill to Blackhouse, 29 October 1840; F060.92: K. E. Abbott, “Report on Trade for 1841”; F060.117, no. 2, enc. no. 2, Abbott to Aberdeen, 27 November 1845; F060.165: “Notes on the Trade, Manufactures and Productions of Various Cities and Countries Visited by Mr. Consul [K. E.] Abbott in 1849-50”; F060.401, no. 13, Churchill to Earl of Derby, 26 July 1877; F0248.1147, decipher nos. 42/16, 42/206, 76; F0368.38, file 3938: A. H. Gleadowe-Newcomen, Report on the Commercial Mission to South Eastern Persia for 1904-05, Calcutta, 1905. U.K., House of Commons, Parliamentary Papers. Accounts and Papers LXI, 1864: “Report by Consul General Abbott on the Trade of Tabreez for the Year 1862-63”; LXVIII, 1867-68: “Report by Mr. Consul General [K. E.] Abbott on the Trade and Commerce of Tabriz for the Year 1866”; LXVII, 1873: “Rasht. Report by Consul Abbott [for 1871]”; LXXV, pt. 1, 1875: “Tabreez. Report by Consul General Jones [for 1873].” M. Varzī, Honar o ṣaṇʿat-e qālī dar Īrān, Tehran, 1350 Š./1971. A. ʿA. Wazīrī, Joḡrāfīā-ye Kermān, ed. M.-E. Bāstānī Pārīzī, Tehran, 1340 Š./1961. H. J. Whigham, “Persia’s Resources. The Carpet Industry,” The Morning Post, 21 July 1902, p. 5. C. White, Three Years in Constantinople, London, 1845. E. Wirth, “Der Orrientteppich und Europa,” Mitteilungen der frankischen geographischen Gesellschaft 21/22, 1974-75, pp. 297-400.

(Annette Ittig)

Throughout the 14th/20th century carpet manufacturing has been, from the point of view of both employment and domestic and foreign market demand, by far the most important Persian industry after oil refining. Although it had been to a large extent organized by foreign capital in the 13th/19th century (see xi, above), after World War I it was gradually taken over by Persian entrepreneurs as a result of the Iranization policy instituted by Reżā Shah. The carpet market did remain largely geared to exports, however, and therefore it was affected by worldwide economic cycles. For example, exports grew after World War I but dropped sharply after the stock-market crash of 1929; the industry also suffered from technical, organizational, and commercial inadequacies. The technical problems included low productivity in spinning, preparation of warps and wefts, and knotting and beating operations, owing to the use of outmoded equipment and lack of quality control. The industry was organized in such a way that weavers were dependent upon a large number of middlemen, which led to confusion and inefficiency. On the commercial side, the use of chemical dyes and the lack of novelty in design after World War II had intensified foreign resistance to Persian carpets, and there was a need for strong market development through promotion of exports. In the 1350s Š./1970s there were improvements in all these areas, owing to installation of modern equipment and some reorganization of production.

Development of the carpet industry. Established production centers like the Solṭānābād area and Tabrīz have remained important in the present century, and, in addition, several cities like Kermān, which had been of secondary importance in the 13th/19th century, rose to prominence in carpet manufacturing and trade. Sir Percy Sykes (History of Persia II, p. 537) claimed that, while serving as consul at Kermān, he had convinced local manufacturers to export carpets via Bandar-e ʿAbbās (q.v.) instead of Tabrīz, thus reducing their costs by 50 percent. He also (1902, p. 199) persuaded them to produce only traditional Persian designs, rather than European ones, which made it possible to capture a larger share of the existing market, as well as to open new ones. In 1327/1909 Nearco Castelli & Brothers and the Eastern Rug and Trading Company, both with headquarters in New York, began to invest in the Kermān carpet industry (Edwards, p. 202). In the next decade they were followed by other large British and American firms. As a result, by 1329/1911 about 10,000 people were employed in the carpet industry in Kermān alone. Between 1326/1908 and 1329/1911 the output of the Kermān carpet industry doubled, and by 1308 Š./1929 about 5,000 looms were in operation; 90 percent of the output was destined for the American market. By that time the original contractors from Tabrīz had been almost entirely replaced by European companies (Dillon, pp. 292-94). After the stock-market crash of 1929 the Western firms left Kermān; when they returned later it was as buyers, rather than as contractors, owing to Reżā Shah’s Iranization policy. One of the largest firms sold its interest to the Iran Carpet Company (Šerkat-e Farš-e Īrān), which had been created in 1314 Š./1935 to produce quality carpets (see below). To that end it had been given a monopoly over carpet exports (Edwards, p. 202).

Starting in about 1300 Š./1921 Isfahan also became a major center of carpet production, a latecomer to the trade focused on the European and American markets (Soltani-Tirani, pp. 75-76). This new development resulted from economic pressures in the region, generated by natural disasters like drought and loss of former industries; for example, at nearby Nāʾīn the production of ʿabāʾs (traditional cloaks), a major source of employment, had been discontinued. Many villages in the Isfahan area did have carpet-weaving traditions, which could be adapted to commercial production. The new carpet industry at Isfahan suffered especially severely, however, from the loss of foreign markets during World War II, and recovery began only when entrepreneurs turned to the Tehran market, which had greatly increased (see Edwards, p. 308).

Aside from Kermān and Isfahan, the major carpet-manufacturing areas in the 1350s Š./1970s were largely the same as those in the 13th/19th century: Kāšān, Kermān, Tabrīz, Qom, Arāk, Isfahan, Nāʾīn, Malāyer, Bījār, Yazd, Ḵᵛānsār, and some parts of Khorasan, Fārs, Baluchistan, and Gorgān (Iran Almanac, 1973, p. 311). These areas still lead the industry, each with its own characteristic designs, commercial practices, and production methods.

Organization of carpet production. Weaving in factories, defined as single-story buildings housing twenty looms, twenty to sixty weavers, and administrative personnel, has always been marginal in Persia. This type of organization has the advantage of permitting better quality control, but it is also more subject to enforcement of the labor laws (see below). Most carpet production is in the hands of subcontractors, who supply labor from their own families or other local sources. Only about 20-30 percent of weavers work for themselves (Soltani-Tirani, p. 95; Bazin, n.d., p. 73); lack of market information, contacts, and capital militates against such independence. Most contract weavers have their own workshops, usually in their homes, but some may rent working space. There are also weavers who have workshops but no looms. A loom consisting of two roller beams that fit into sockets in the side walls of the workshops (see iii, above) represents a significant investment. Wooden beams cost 300 tomans the pair before 1339 Š./1960, when cheaper iron beams were introduced (Dillon, p. 318); the latter cost 100 tomans in 1350 Š./1971.

The demands of the export market impose certain standards on contract work. For example, American buyers prefer “large areas of plain, light colored background. Uniformity of color, especially the background, is important. In view of variability of dyeing, a large quantity of dyed wool needs to be bought at the outset. A weaver usually does not have that kind of money. Also, carpet making requires some 6 months of labor cost; he cannot advance this either, while credit is hardly available” (Dillon, pp. 304-10; cf. ILO, surveys; Soltani-Tirani, pp. 108-15). Because of the need for quick turnover and payment, smaller carpets predominate, a phenomenon that was already apparent at the beginning of commercial carpet production in the early 1300s/1880s (Ehlers, p. 248). Economic necessity also makes it difficult for weavers to sell carpets themselves, for they can rarely support themselves while they are waiting for buyers. One option open to the independent weaver, however, is amānat-forūšī, selling carpets through an agent at a commission of 5-20 percent of the price (Soltani-Tirani, p. 96).

In rural areas weaving is organized by middlemen (dallāl “broker,” ḥāmel “agent,” ostād(-e) kār “master of work,” maḥaḷčī “representative”) who supply all necessary materials, by heads of families who buy materials in the bāzār, or by the Iran Carpet Company (Ehlers, pp. 233, 235). Three types of middleman can be distinguished, all based primarily in urban centers: those who work only for carpet dealers; pīlavars (itinerant traders), who supply the general needs of the rural population as well; and teachers, government officials, and others who trade in carpets on the side.

The marketing structure is thus heavily dependent upon intermediaries; whether a weaver is independent or works as a subcontractor, he or she must deal with such intermediaries, particularly for the sale of medium-sized carpets (primarily 2.5 x 3.5 m), which are subject to considerable price elasticity in importing countries (Ehlers, p. 230; ILO, surveys). The extreme fragmentation of the contracting system has led, paradoxically, to concentration of the wholesale carpet trade and ancillary activities in a few large enterprises with oligopolistic market features. Wholesalers in the carpet and wool trades, operating mostly out of the bāzārs, play an essential role in the carpet industry. They buy carpets directly or indirectly from villagers and nomads and also from dealers and colleagues in the bāzārs; they also supply wool and yarn to weavers, take care of having the wool dyed, and engage in other processing activities (Ehlers, p. 231). Although there are several hundred carpet dealers in, for example, Arāk, the bulk of the trade is controlled by only twenty or thirty. They supervise the weaving through dallāls or ḥāmels. Tehran wholesalers also leave purchasing and manufacturing arrangement to such middlemen. In addition, local wholesalers may sell in Tehran or to agents of Tehran firms (Ehlers, p. 235; Dillon, p. 321; Edwards, p. 211).

Among middlemen the bigger operators combine several functions: They own flocks of sheep; have the sheared wool washed, spun, and dyed; and supply it with other necessities to “their” weavers. Usually, however, the situation is more fragmented, especially when a household hires workers or when two or more households pool their resources. Some raw material is locally available; for example, local sheep herders provide wool, which can also be bought in neighboring towns and major provincial centers. Raw materials can also be obtained from local or itinerant traders; their prices are higher, but they give credit. Weavers who do their own dyeing buy their colors from retailers (saqat-forūš), but most dyeing is done in small establishments located in large towns; the dyes are purchased from tājer-e rangs (dye merchants) or in large quantities directly from Tehran. Specialists, for example, warpers, are often required in preliminary stages of production; the cost of having the warps laid is the weaver’s responsibility. New designs may be prepared by the naqšakaš-e qālī, though they are not necessary where traditional patterns persist. Designs can be purchased from middlemen or rented at 15-20 percent of their purchase price (Soltani-Tirani, pp. 82-91). The middlemen keep records of good weavers, which means that the latter find it easier to obtain advances. Discipline is also relatively easy, as alternative work is difficult to find.

In the Isfahan area the ostād(-e) kār supplies the bulk of raw materials for cottage weavers, but the old system of pīšforūš (“presale”), in which he had a prior right to the carpets produced, no longer exists there. Weavers owe him money, but they may sell their carpets as they wish. Nevertheless, the weaver is at a great disadvantage, especially in villages where the ostād(-e) kār is simultaneously the main (or sole) source of credit, the supplier of consumer goods, and the carpet dealer (Soltani-Tirani, p. 91 ). The situation is somewhat similar in Arāk, where middlemen known as maḥaḷčīs supply all raw materials to weavers in exchange for the right to purchase the finished carpets at fixed prices. They buy the wool ready-made or arrange to have it carded, washed, and spun. They also supply credit for up to six months either at interest or with the right of piš-ḵarīd (prior purchase); the weavers thus become economically dependent on them. The middlemen, who control the weaving through their monopoly of credit, raw materials, and the carpet trade, constitute a kind of “informal union” (Ehlers, pp. 232-34).

In Qom the industry is organized by the municipality, which supplies all materials and designs. Pay is determined according to the size and complexity of the work. About 30 percent of the weavers are independent and concentrated in a few villages. They purchase additional services and raw materials from elsewhere and sell their carpets in Qom. Estimates for 1348 Š./1969 indicate that 1,000-1,500 tomans, or 20-40 percent, of annual family revenues came from carpet weaving (Bazin, 1973, pp. 73-74; idem, n.d., pp. 85f.). In fact, these revenues were more important for many rural families than their incomes from agricultural activities (Bazin, n.d., pp. 85f.; Costello; Ehlers, p. 241).

Urban capital and organization have also played a role where the carpet industry was still relatively undeveloped, as in Kurdistan in the 1350s Š./1970s. Wool was either sold in raw form in the Sanandaj market or collected from the villages by agents for outside traders; the finished carpets were sold at Sanandaj or Kermānšāh, largely by people from outside the region. In as many as twenty-five villages it was mentioned that marketing was a serious problem; lack of buyers had discouraged weavers from enlarging the scale of their operations (ILO, Sanandaj report).

Carpet weaving in tribal areas is still relatively independent, for the subcontracting system has made few inroads. City investors do not risk capital among the tribes, which are always on the move and themselves own the necessary raw materials (Dillon, p. 301). In the territory of the Baḵtīārī, for example, there is hardly an ostād(-e) kār to be found, though carpets are sold to pīlavars and other middlemen. The same is true in the Najafābād valley, which was formerly dominated by ostād(e) kārs

(Soltani-Tirani, pp. 95-97). Owing to the absence of contractors in the Kermān tribal area after 1327 Š./1948, however, the quality of carpets (determined by the foreign importers) declined; because of high prices for wool cheaper cotton warps were substituted for traditional woolen ones. Also natural dyes were increasingly replaced by chemical colors; the small tribal carpets were thus mostly destined for the domestic market (Dillon, p. 302). The opposite effect was noted in the Ardabīl area, where the quality of carpets has suffered from the commercialization of production (ILO, Ardabīl survey), part of a general decline that began in the 13th/19th century (Ehlers, p. 248).

The Iran Carpet Company was founded in 1314 Š./1935 to oversee the export of carpets and to guarantee certain standards of quality, though, because of declining foreign markets (see below), restrictions on exports were lifted after only a year. Subsequently the Company came to serve as guarantor for credit extended by the Agricultural Bank (Bānk-e Kešāvarzī) to village cooperative societies for the purchase of raw materials to be used in carpet production. The finished carpets were to be bought by the company, which would then repay the loan from the Agricultural Bank. The importation of machinery to make carpets was forbidden (Iran Almanac, 1963, p. 246). The Iran Carpet Company also established an inventory of carpet designs, in order to copyright them and to prevent machine-made carpets from usurping the market for hand-woven carpets (Iran Almanac, 1965, p. 646). To improve quality control the Company set up six factories for washing and combing wool, three dye factories, and one each for wool carding and spinning. It also announced that 80 percent of its profits would be set aside to improve the lot of rural weavers through creation of cooperatives. The company planned to help them purchase looms, raw materials, and other necessities and to insist on higher wages, elimination of child labor, better working conditions, and higher standard of output. These measures were expected to increase carpet prices 25 percent (Iran Almanac, 1971, p. 311). The experience in Zābol (see below) is evidence that the Company achieved at least some of these objectives.

Since 1354 Š./1975 the Company has operated in eight provinces (Tehran, Tabrīz, Mašhad, Kermān, Isfahan, Kermānšāh, Hamadān, and Arāk); in 1356 Š./1977 it controlled 2-3 percent of the total national output of carpets. In the Isfahan area the Company owns about 1,000 looms (Soltani-Tirani, p. 93). Before 1353 Š./1974 the Arāk office had produced all its own raw materials, but since then it has been supplied with dyes by the Company’s factory at Karaj; whereas before it used 100 dye shades, it now uses only thirty-seven standardized colors. The Company also supplies all inputs; its agents, known as ḥāmels are often village headmen (kadḵodās). The Company enforces higher quality requirements than do the maḥaḷčīs of the free market, but it employs only a limited number of designs (Ehlers, p. 236).

Although data on employment are based on estimates and are not always reliable, it is possible to discern three general trends during the 14th/20th century. First was a steady growth in overall employment in the carpet industry linked to the steady growth of the industry itself. Second, the proportion of rural workers in the labor force increased, so that by 1309 Š./1930 75 percent of all carpet workers were located in rural areas. Third, women and girls have always played a predominant role in carpet weaving and now represent almost 80 percent of the labor force.

Growth in employment. At the end of the first decade of this century about 65,000 people were working in the urban carpet industry; the number in rural areas is unknown. An estimate by Abdullaev (p. 212) for the second decade puts urban employment in the Persian carpet industry at 60,000; this figure remained fairly steady until 1319 Š./1940 (Floor, 1985, pp. 5, 27-28), when it was approximately matched by the number of people employed in rural areas (Abdullaev, p. 84). A more detailed estimate by C. A. Edwards (based on extensive personal contacts in the industry over many years) put the total (rural and urban) employment at more than 250,000 in 1319 Š./1940. His figure tallies generally with those of Abdullaev if it is assumed that each loom was operated by three or four people, the ostād, or master weaver; the naqšanevīs, or design supervisor; and one or two šāgerds, or apprentices (Edwards, p. 99).

According to the Iran Almanac, in 1335 Š./1956 there were 103,492 looms for weaving carpets and cotton druggets in eighty-eight towns in Persia. Approximately 166,000 workers operated these looms. In 1338 Š./1959 the industry employed about 231,000 workers, 5,000 of them in government-owned workshops (1963, pp. 246, 300, 302). Almost a decade later, in 1347 Š./1968, there were approximately 400,000 people engaged in carpet weaving, operating about 120,000 looms. In that year the Persian government established the Center for Cottage Industry (Markaz-e Ṣanāyeʿ-e Dastī-e Īrān) to develop handicrafts in rural areas (1971, p. 311), and by 1354 Š./1975 there were about 500,000 weavers working for 3,000 employers. The figures published by the Institute of Social Studies (Moʾassasa-ye Moṭālaʿāt-e Ejtemāʿī) in Tehran differ slightly: 748,000 weavers in 1344 Š./1965 and 800,000 in 1354 Š./1975. These estimates also include nonprofessional weavers, however (Iran Almanac, 1976, p. 209). The most recent estimates are that there were 2 million carpet weavers in Iran in 1367 Š./1988 and that the total number of people directly or indirectly involved in the carpet industry was five million (Iran Yearbook, p. 432); these figures seem rather high.

Progressive concentration in rural areas. Before 1309 Š./1930 carpet weaving appears to have been primarily an urban industrial activity, but in that year it began to change sharply. The numbers of looms in the most important production centers ten years later are shown in Table 48.

In an important carpet-weaving center like Arāk the urban industry still plays an important role. In 1354 Š./1975, for example, of 9,233 business establishments in the town 45 percent, or 4,181, were related to carpets (carpet repairmen, cutters, designers, etc.); 3,440 were weaving establishments. Most of them were household workshops (Ehlers, p. 224). Nevertheless, the production of carpets in urban Arāk, as elsewhere, has declined in comparison with the 1310s Š./1930s. A similar situation prevails in Kermān. In 1327 Š./1948 there were 4,000 looms in operation there. In 1345 Š./1966 32,000 weavers were employed in the Kermān area, operating 11,000 looms, mostly in home workshops. The number of looms in the city itself, however, had fallen by about 35 percent since 1316 Š./1937 (Dillon, pp. 279, 299).

On the other hand, after World War II thriving carpet industries developed in rural areas where they had previously been nonexistent. In 1327 Š./1948, for example, only eleven villages in the Qom area were producing carpets, probably for local and personal needs. In the 1335 Š./1956 census only 730 people were listed as engaged in carpet production, whereas in the census of 1345 Š./1966 3,671, or 36 percent, of 10,036 households were engaged in carpet production. The total number employed was about 5,000, mostly women and girls. A field survey in 1348 Š./1969 confirmed the steady growth and continued spread of carpet production in the Qom area (Bazin, n.d., pp. 70, 73). In Ardabīl, another important center, the carpet industry employed about 5,500 weavers, 60 percent of whom were men, in 1351 Š./1972. There was, however, a steady shift of the industry to the villages in order to circumvent laws related to the employment of minors (defined as under twelve years old; ILO, Ardabīl survey). Even the government-owned Iran Carpet Company circumvented the laws by signing weaving contracts with parents in full knowledge that it was mostly minor children who would do the weaving (Rist, p. 152). An additional inducement to move production to rural areas was villagers’ acceptance of lower wages.

Many other cities produced almost no carpets at all; instead the industry was concentrated in surrounding rural areas. For example, according to both the 1335 Š./1956 and 1345 Š./1966 censuses there were only twenty-three carpet-weaving establishments in the town of Kermānšāh, 2.5 percent of the enumerated household workshops, reflecting a general decline in home industry there. On the other hand, in 1347 Š./1968 fifty shops handled pile carpets and gelīms (see v, above) in the bāzār, where they were brought from the countryside to be sold to agents from the Tehran carpet market (Clarke and Clark, pp. 62-63, 75). In Shiraz, according to the 1335 Š./1956 census, only 15.9 percent of urban household workshops made carpets and related products, whereas in the surrounding area 53.4 percent of household workshops were so engaged (Clarke, p. 38). In 1344 Š./1965 the manufacture of rugs was proportionately even more important in the neighboring villages and among the tribes (especially the Qašqāʾī). In the Isfahan ostān (administrative region) at the end of the 1350s Š./1970s urban weaving occurred only in Isfahan itself, Qomša (Šahreżā), Homāyūnšahr, and Ḵᵛānsār (Soltani-Tirani, p. 82).

Not all rural carpet production was for the market, however. According to a 1350 Š./1971 survey of forty-five villages in the vicinity of Qazvīn, about 8 percent of the total, most weaving was for domestic use. The weavers used chemical dyes and produced inferior work. The once famous Qarabāḡī design was still being made but by only two professional weavers. The explanation for such backward conditions so close to the major Tehran carpet market was an absence of promotional and technical support (ILO, Qazvīn survey). There was a similar problem around Sanandaj, where weaving on small looms with throw shuttles took place in most households. Carpets of medium and coarse quality with traditional folk motifs and designs were produced: “jainajaj” of “kaškūla” (begging bowl) design, mūj (blanket of four large pieces sewn together), narrow jājīms in simple striped patterns, gelīms (see v, above), and so on. Yarn was spun from local wool by the villagers, but facilities for dyeing, cutting, and the like were not available in the seventy-three villages surveyed. Dyeing was done in Sanandaj (ILO, Sanandaj survey).

In Semnān even rural carpet weaving was of little importance in the 1350s Š./1970s (Connell, p. 66, 82). In southeastern Persia at the end of the decade more than 80 percent of carpet production was concentrated in the vicinities of Zābol (71 percent, based on a sample survey) and Bam (12 percent), yet in the former area the production of carpets was insignificant and mainly for the villagers’ own use (Ital Consult, chap. 8). In fact, before the Iran Carpet Company was established there in 1348 Š./1969, as part of a program to develop the region, there had been no carpet production at all, either public or private. Even in 1355 Š./1976 there were only forty-five looms and ninety carpet workers around Zābol. Kermāni weavers were brought in to train Sīstāni women, who naturally adopted Kermāni designs. The entire production process remained under the control of the Company (Rist, pp. 150, 153, 187).

The latest reliable figures on the size and distribution of rural employment in the carpet industry are from 1350 Š./1971, when the International Labor Organization conducted an extensive survey of Persian industries. At that time most farm families were engaged part time in nonagricultural occupations; of a total of 1.2 million engaged in manufacturing 300,000 were carpet weavers. “The bulk of rural manufacturing consists of carpet and cloth weaving. Population centers with fewer than 5,000 inhabitants account for 72% of employment in the carpet industry and 70% of the employment in cloth weaving. Both male and female workers are employed in carpet and cloth weaving, where part-time employment and low wages are frequent. Workers are highly dependent on the middlemen who provide credit by making advance payment for the product. In the carpet industry there is ample room for development of the market through export promotion, reorganization of production and improvement of equipment and marketing” (ILO, 1973, pp. 46-47).

Labor force and working conditions. Weaving is an almost exclusively female occupation, mainly for young girls; men purchase the raw materials, prepare the warps, and sell the finished products (Soltani-Tirani, p. 91). Female workers account for as much as 67.5 percent of total employment in the carpet industry and for more than 75 percent of the rural sector alone (ILO, 1973). Working conditions among women and children in the carpet workshops of Kermān led to the first attempts at labor legislation in Persia. Crouching over their work in subterranean areas like caves—badly lit, cold, humid, and unventilated—many women and children were crippled for life. Workers were often afflicted with ankylosis of the pelvis and lower spine, with resulting difficulties in childbirth (Floor, 1987, pp. 101-02). Even as recently as the 1350s Š./1970s there was evidence that disproportionate numbers of female children in rural areas of Persia suffered from rickets and other diseases caused or aggravated by the conditions under which carpet weaving took place (Rašīdīān, pp. 25-26, with illustrations).

In 1331/1913 the deputy governor of Kermān province attempted to introduce some regulation of working conditions into the carpet-weaving industry. He was believed, however, to have been motivated primarily by the desire to extract more money from employers, and under pressure from the British and Russian consuls in Kermān he was forced to withdraw the regulations. The consuls proposed that improvements be introduced where necessary, especially to ameliorate overcrowding, poor ventilation, and cruelty to children (Floor, 1987, pp. 82-84). After World War I the International Labor Organization, the British government, and Christian missionaries in Kermān pressed for voluntary improvements by employers, which were adopted in 1300 Š./1921. In Dey 1302 Š./December 1923 the governor-general of Kermān decreed a compulsory eight-hour working day, an age limit for workers, gender separation, and better working conditions; fines were established for violations. As a result it was widely acknowledged that the situation improved greatly over that of 1332/1913. By 1307 Š./1928, however, reports indicated that there had been a relapse to earlier conditions. The resulting publicity forced the Persian government to issue new regulations for the entire carpet industry, including the forty-eight-hour work week, abolition of child labor, and improved working conditions. These regulations were apparently never enforced, however. According to a report by the United States legation in 1313 Š./1934, “no such reforms have been made and probably were never a matter of serious consideration” (Floor, 1987, pp. 86-93; cf. Edwards, p. 206).

Because of the decentralized structure of the carpet industry (see above), it has been difficult to apply the various labor laws that have been promulgated by the Persian government since 1325 Š./1946. Furthermore, a flexible labor policy, for example, toward the hiring of minors, allows entrepreneurs to react more quickly to fluctuations in the market (Soltani-Tirani, p. 93). The labor laws were thus still not being enforced in rural areas as recently as the late 1350s Š./1970s. The mainly female labor force still put in long hours without holidays or overtime pay (Iran Almanac, 1976, p. 210), especially those working on piece-rate contracts. In areas where demand for carpets was low, however, underemployment was a greater problem than exploitation. Around Sanandaj, for example, where 48 percent of those employed in crafts were in the carpet industry, female weavers worked an average of only twenty-seven hours a week in 1350 Š./1971 (ILO, Sanandaj survey).

In the early part of this century European contractors tried to exercise strict control over their labor force, for example, locking up a defaulting weaver in a stable until his relatives came forward with guarantees that the carpet would be finished or the advance returned (Edwards, p. 136 n. 1). When Persian contractors came to dominate the industry in the 1310s Š./1930s such measures were replaced by more indirect controls. Contract work was based on piece rates, which meant that the problem of disciplining labor was transferred to the head of the family, whose income depended on output. It also increased competition among weavers. The latter tried to countervail this trend by cutting corners, particularly using the joftī knot (see iii, above) and “kimanshi” weaving, which reduce the amount of labor required. These methods spread in the late 1310s Š./1930s, when production for Tehran companies and foreign trade intensified. Currently such methods are universal, abetted by employers, because consumers do not easily recognize their hallmarks. Pieces of high quality are therefore produced only under direct supervision by company foremen (Dillon, pp. 313-17; Soltani-Tirani, p. 92) or as a result of manipulation by middlemen. For example, in Golpāyegān the ostād(-e) kār owns all factors of production and pays his weavers by the ḏaṛʿ (cubit; see iii, above), though he prices carpets by the meter for sale. He may also provide thicker yarns for weaving and knotting, which, because the dimensions and knotting density of the carpet are fixed, force the workers to weave more tightly and thus to produce better quality at lower cost (Soltani-Tirani, p. 92).

Under the contract system weavers are paid subsistence wages at best, often less. Particularly in Sīstān and Baluchistan, where unemployment was a serious problem in the 1350s Š./1970s, peasants were at the mercy of middlemen. Because of large-scale migration to other, more prosperous areas, fewer sheep were raised locally, and the peasants could thus not produce sufficient wool for weaving carpets. As there were no credit facilities to allow purchase of materials, middlemen from nearby towns provided them. The middlemen also usually handled the sale of the finished carpets, with half the price going to the weavers. The International Labor Organization estimated, on the basis of standard carpets and prices, that a weaver’s wages averaged about 18 rials a day, whether paid as a percentage of the sale price or as a piece rate (ILO, Sīstān report). In the Marvdašt area rural female weavers were making about 20 rials a day in 1350 Š./1971 (ILO, Marvdašt report). The same types and levels of remuneration also prevailed elsewhere in Persia (Momeni, p. 182; Šafaqī, p. 8). Urban weavers received more, an average of 90 rials a day in 1347 Š./1968, compared to 20 rials for rural weavers (Šafaqī, p. 8), but still barely at subsistence level (wages are usually paid by the piece but have been converted to daily rates here for comparison with other branches of industry). Furthermore, these low wages were eroded by rising inflation. During 1345-50 Š./1966-71 carpet prices rose by 11-20 percent a year, which was blamed partly on higher production costs, while weavers’ wages rose only about 5 percent (ILO, “The role”). According to a report in 1354 Š./1976, wages absorbed about 60 percent of the cost of carpet production. They had risen by a total of 250 percent in the preceding decade but nevertheless remained low in relation to the overall Persian wage structure. Whereas in 1343 Š./1964 it cost 8,000 rials to weave a carpet, in 1354 Š./1976 it cost 28,000 rials; in the latter year the same carpet would sell for 60,000 rials (Iran Almanac, 1976, p. 210). Wages for urban weavers in that year ranged between 120 and 240 rials a day (the higher rate for master weavers), whereas the average daily wage for unskilled urban laborers was 300 rials (Iran Almanac, 1976, p. 210). In 1357 Š./1978 beginning girl weavers in Isfahan were paid 10-50 rials a day for a ten-hour day, experienced weavers 60-150 rials, and highly qualified weavers 200-400 rials. Comparable wages were even lower in rural areas (Soltani-Tirani, p. 108). It is unlikely that the situation has changed in subsequent years.

Only the Iran Carpet Company issues formal written contracts. Whether written or oral, the contract stipulates design, colors, dimensions, and knotting density of the carpet, as well as rate of pay, date of completion, and conditions for payment of a bonus for particularly good work (Dillon, pp. 318-21; cf. p. 322 for the wording of such a contract; Rist, p. 152). Both factory and rural workers are paid by the piece. The rate depends on intricacy of design, tightness of weave, and type of knotting, as specified in the contract. Piece rates for factory and other hired workers in the Kermān area are based on the sad-nešān (100 nešāns: a nešān = 160 knots along the weft, or pūd, of the carpet) and a theoretical daily output of 1,600 knots (one tenth of a sad-nešān), though few workers tie that many knots in one day. For contract weavers the base unit is the gaz, which equals 1,280 nešān (Dillon, p. 312). In the Arāk area the čārak (čahārak, lit. “one fourth”; 4 čārak = 1 ḏaṛʿ of 104 x 104 cm) is the basic unit of measure for piece rates (Ehlers, p. 239). In the villages two to four weavers usually work at one loom. Younger, inexperienced weavers (porkon) work on the background areas, while the more experienced women execute the pattern. The density of the pile depends on the number of knots (raj) per unit of length and the number of knots (rīša) per unit of set up on the loom, in a consistent ratio of 2:1. There are four standards: 70/35, 80/40, 90/45, and 100/50. The 80/40 ratio is export quality, the 70/35 for cheaper carpets; the other two standards are for very fine carpets (for breakdown of production costs, see Dillon, pp. 312, 322, 339-48; for Arāk, Ehlers, p. 239; for Zābol, Rist, p. 152; for Isfahan, Soltani-Tirani, p. 109). The weaver often receives a month’s pay in advance, after which pro rated payments follow every thirty-five days; there is no pay during the final month of work because of the advance. Monthly inspections can lead to disputes over the rate of progress; the weavers are always at a disadvantage in such disputes.

Exports. For the early part of the 14th/20th century the data on carpet exports are not entirely reliable. The customs administration did not check the declared values of exports because duty was charged only on carpets in which yarns dyed with aniline had been used. Trade expanded considerably up to 1308 Š./1929, owing primarily to demand in Germany for cheap and medium-grade carpets after the reduction of German import duties in July 1927. Traditional American demand had begun to fall even before the Great Depression reduced it further. European demand for antique carpets also declined (Lingeman, pp. 14, 32). More important than the Depression were American tariffs on cheap carpets, which Persian exporters regarded as prohibitive. They were then faced with the choice of making carpets of higher quality or halting exports to the United States altogether. At the same time increased use of aniline dyes had harmed the reputation of Persian carpets. Finally, the carpet industry suffered from Russian competition in the Persian wool market. Although trade did improve somewhat after export controls were lifted and the Iran Carpet Company imposed higher levels of quality control (Gray, p. 20; see above) and although carpets remained the most important Persian export commodity after oil (cf. Table 49), there was a steady decline, aggravated by higher tariffs, quotas, and exchange restrictions in former markets. As a result carpet-manufacturing firms in Iran either restricted their activities or, in some instances, closed down. Before World War II the chief export markets were the United States, the United Kingdom, Turkey, and Iraq (Simmonds, p. 28).

For most of the period after World War II carpets were also Persia’s leading export commodity after oil, though in some years they fell to second place behind raw cotton. In 1327 Š./1948 production rebounded from wartime lows, though quality (as viewed by US importers) had deteriorated (Dillon, p. 302). In 1339-40 Š./1960-61 carpet exports totaled 1,930 million rials, second only to oil; the following year they reached 2,192 million rials but were surpassed by cotton.

Despite declining quality, competition from machine-made carpets, and protective tariffs abroad, woven carpets generally retained their value and market (for major markets in 1330-42 Š./1951-1963, see Dillon, p. 301). Since 133l Š./1952 West Germany has been the main customer for Persian carpets, followed by the United Kingdom, the United States, and Switzerland. The Swiss prefer expensive carpets with delicate designs. Americans prefer thick carpets; for that reason those woven for the American market have long stems. They are predominantly in pastel colors, especially light blue and pink, with less intricate patterns. Kermān is still the main supplier for this market (Iran Almanac, 1963, pp. 300-02). Toward the end of the 1340s Š./1960s, however, low standards of quality (referring to dyes, knots, etc.) began to affect exports adversely. In 1350-51 Š./1971-72 only $73 million worth of carpets were exported (Iran Almanac, 1973, p. 247). This decline led to improvements, and as a result exports and prices soared in the first half of 1352 Š./1973; in the second half they dropped sharply again, owing to the impact of the international oil crisis on the main carpet markets. Many factories lowered output, and in some carpets were even left half finished on the looms (Iran Almanac, 1974, p. 323). The slump was of short duration, however, and both quality of production and volume of exports increased in 1353 Š./1974, partly as a result of better marketing (Iran Almanac, 1975, p. 250).

Rising prices and increased competition from Indian and Pakistani carpets were nevertheless having a long-term effect on demand in both the domestic and world markets, and Persian carpets dropped to third place among exports (Iran Almanac, 1976, p. 210); the industry also suffered enormously as a result of the 1357 Š./1978-79 revolution in Persia and the ensuing war with Iraq. In 1358-59 Š./1978-79 carpet exports had amounted to only $84 million, reflecting uncertainty about the future of the Pahlavi regime and the direction of the Islamic Republic of Iran. In 1359 Š./1980-81, however, exports reached a new high of 30.2 million rials ($425 million), partly owing to the lifting of previous government restrictions. With the outbreak of the war with Iraq in 1360 Š./1981 and new government trade restrictions carpet exports dropped again to the 1357 Š./1978 level. Especially the obligation to sell all foreign-currency receipts to the government at the official exchange rate not only caused carpet exports to drop to 5.6 million rials in 1361 Š./1982-83 and 7.14 million rials in 1362 Š./1983-84 but also led to a fall in production and employment (see xiii, below). With the costs of war rising and the concomitant growing need for employment and revenues both at state and household levels, the government began to provide incentives to exporters, as a result of which the carpet industry regained strength. In 1365 Š./1986-87 there was an upsurge in carpet exports to 28,217 million rials. It is expected that, with better marketing and improved quality, the Persian carpet industry can regain the markets lost to Indian, Pakistani, and Afghan carpets and can even pass 1 billion rials in the near future (Iran Yearbook, p. 433; Iran Times, 8 Ābān 1366 Š./30 October 1987; 12 Esfand 1366 Š./2 March 1988).

Bibliography : Z. Z. Abdullaev, Promyshlennost’ i zarozhdenie rabochego klassa Irana v kontse XIX-nachale XX vv., Baku, 1963. M. Bazin, “Le travail du tapis dans la région de Qom,” Bulletin de la Société languedocienne de géographie 7, 1973, pp. 83-92. Idem, La vie rurale dans le region de Qom, Paris, n. d. J. I. Clarke, The Iranian City of Shiraz, Durham, 1963. Idem and B. D. Clark, Kermanshah an Iranian Provincial City, Durham, 1969. J. Connell, ed., Semnan. Persian City and Region, London, 1969. V. F. Costello, Kashan. A City and Region in Iran, Durham, 1976. R. Dillon, Carpet Capitalism and Craft Evolution in Kerman, Ph.D. diss., Columbia University, New York, 1976. A. C. Edwards, The Persian Carpet. A Survey of the Carpet-Weaving Industry of Persia, London, 1953. E. Ehlers, “Teppich-manufaktur und Teppichhandel in Arak/Farahan-Iran,” Der Islam 59, 1982, pp. 222-53. W. M. Floor, Industrialization in Iran, Durham, 1985. Idem, Labor Law, Unions and Conditions in Iran 1900-1941, Durham, 1987. F. A. G. Gray, Report on Economic and Commercial Conditions in Iran, London, 1937. M. C. Hadow, Report on the Trade and Industry of Persia, 2 vols., London, 1923-25. International Labor Organization (ILO), unpublished reports: “Report on Fact-Finding Mission to Tabriz, July 1972”; “. . . to Sanandaj, Saqqez, Baneh and Marivan in Kurdistan, April 1972”; “. . . to Kermanshah and Sirjan, April 1972”; “Report on Rural Industries of Sistan and Baluchistan, March 1972”; “Report on Rural Manpower Development Possibilities in Selected Districts of Qazvin, September 1972”; “The Role of the Non-Farm Sector in the Overall Strategy for Employment and Income Promotion, February 15, 1972”; Idem, Employment and Income Policies for Iran, Geneva, 1973. Iran Almanac, Tehran, 1963, 1965, 1971, 1974, 1975, 1976. Iran Yearbook 1988, Bonn, 1988. Ital Consult, Economic and Social Development Plan for the South-Eastern Region I, Rome, 1961. E. R. Lingeman, Economic Conditions in Persia, London, 1930. M. Momeni, Malayer und sein Umland, Marburg, 1976. Ḵ. Rašīdīān, “Pažūhes-ī dar bāra-ye qālī-bāfī wa masāʾel-e ān dar rūstāhā-ye manṭaqa-ye bīabānī-e šarq-e Kāšān,” Bīābān (Markaz-e Taḥqīqāt-e Manāṭeq-e Kavīrī o Bīābanī-e Īrān, Tehran University) 8, 2537 = 1357 Š./1978, pp. 1-35. B. Rist, Die Stadt Zabol, Marburg, 1981. S. Šafaqī, Moṭālaʿa-ye joḡrāfīā-ye šahr-e Esfahān, Isfahan, 1347 Š./1968. S. Simmonds, Economic Conditions in Iran, London, 1935. M.-A. Soltani-Tirani, Handwerker und Handwerk in Esfahân. Räumliche, wirtschaftliche und soziale Organisationsformen, Marburger geographische Schriften 87, Marburg, 1982. P. Sykes, Ten Thousand Miles in Persia or Eight Years in Iran, London, 1902. B. Temple, Report on Trade and Transport Conditions in Persia, London, 1921. H. E. Wirth, Der Orientteppich und Europa, Erlangen, 1976.

(Willem M. Floor)

In the period immediately following the shah’s flight from the country in 1358 Š./1979 the prices for Persian carpets reached record highs on Western markets. These prices were determined by the convergence of three main factors: the underlying trend toward higher prices as demand outstripped supply, which had been apparent since the 1340s Š./1960s; the particular circumstances of the last two years of the Pahlavi era; and the exchange-rate policy adopted by the new regime. In the period immediately following World War II a boom in carpet production occurred in Iran, fueled in particular by the German “economic miracle” after 1332 Š./1953; the art of weaving on a commercial scale thus spread to almost every corner of the country. But by the mid-1340s Š./960s the shah’s industrialization program began to have a significant effect on both urban and village carpet production. As the shepherds of the Khorasan uplands left their villages to work in the steel mills of Mašhad queues were seen outside the butchers’ shops. But the developing urbanization meant a shortage not only of mutton, but also of wool. Also, the shepherds’ wives came to find watching television more congenial than weaving, and the supply of carpets began to be insufficient to meet demand; prices rose more rapidly than many Western importers could accept, and as a result carpets with Persian designs began to be produced in large numbers in India and Pakistan and later in China. In the late 1350s Š./1970s prices of Persian carpets in Persia became still more inflated because huge quantities of goods were purchased and shipped out of the country to Hamburg and Zurich, especially by Jewish dealers who feared for the security of their capital as the Revolution of 1357 Š./1978-79 approached. Finally, under the shah the rial, which was pegged to the dollar, was one of the world’s strongest currencies, supported by massive oil exports. After the revolution the government continued to maintain a high exchange rate with the dollar, but political and economic realities in Iran made a mockery of this attempt and engendered a black market in which the world market value of the currency was reflected more clearly. As carpets could be exported only at the official exchange rate, the price levels reached in 1360-62 Š./1981-83 were rejected in the international carpet trade, and Western importers began to buy from other countries (see Table 50). Carpet exports from Iran during this period served in many instances mainly as a conduit for wealthy Iranians eager to transfer their capital abroad and willing to do so at any price.

The government soon recognized that drastic measures had become necessary if the slide in carpet exports was to be halted. A commission from the Tehran customs office was instructed to visit Hamburg, London, Zurich, and other international trading centers to recommend a scheme for revitalizing the trade. The commissioners discovered, however, that their product was overvalued by 40-60 percent. For example, in the early 1350s Š./1970s, when Bījār carpets cost DM 500 per m² in the Hamburg free port, they had seemed a good investment; when the price rose above DM 1,000 per m² those who had bought at DM 500 congratulated themselves. By 1358 Š./1979 they had reached DM 2,000, which most dealers considered overpriced, and, when DM 3,000 was quoted shortly afterward, it was clear that the bubble had burst.

The simple solution would have been to abolish the official rate of exchange with the dollar, but such a move would probably have led to an undesirable drain of capital. A system was thus devised to allow carpet exporters to obtain part of the purchase price of their goods on the black market; in this way only a proportion of the value had to be paid in foreign currency at the official exchange rate. An advantage of the scheme was that the proportion could be varied at will to increase or decrease the effective discount on the official rate. Indeed, details were altered frequently in order to provide ever greater incentives to dealers as Persia’s need for foreign exchange increased. The system did have the disadvantage of being very cumbersome and time-consuming: As Persian invoices are notoriously unreliable, the only way that the Tehran customs could arrive at a true valuation of goods being exported was to have its own officials inspect and value every carpet before it was dispatched. Despite the frustrations caused by this procedure, the desired result was achieved. Beginning in 1364 Š./1985, the average price in dollars or German marks fell (even though the price in rials soared), and exports began to rise again (see Table 50). The effect on local prices may be judged by comparing the cost of the Nehāvand rug. The villages in the area between Nehāvand and Ošvand, south of Hamadān, produced large blue-ground, single-wefted rugs of around 40 sq feet stylistically related to the products of neighboring Kurdistan. They have maintained a consistent quality of wool, weave, and dyestuffs over a long period and thus offer a more reliable yardstick for price comparisons than areas where quality has varied considerably.

Table 51 shows price movements for this type of carpet at Hamadān over the fifteen years between 1352 Š./1973 and 1367 Š./1988. From 1358 Š./1979 to the end of 1366 Š./1987 the rial price increased fivefold, while the dollar price remained more or less constant. Further explosive rises in 1367 Š./1988, however, again led importers to cancel orders. On the other hand, the new system, involving as it does manipulation of prices and exchange rates by the government, creates its own distortions. For example, the valuation scheme in operation at the end of the 1360s Š./1980s favored high-priced merchandise and penalized cheaper goods. This bias was reflected in the domestic market. In dollars a fine semi-antique Kashan (Kāšān) carpet cost about twice as much as a good example from Herīs (Herīz/Harīz), a market center in Azerbaijan, but in rials the former was four times the price of the latter. If exchange controls were lifted, a major realignment of prices would thus occur.

The pattern of production of new rugs did not differ significantly from the one that was developing before the revolution. There were, indeed, a few remarkable new creations, like superbly woven weft-wrapped flat weaves from Sīrjān and fashion-oriented Heriz-style carpets woven in Tabrīz. But elsewhere the trend toward ever harsher color combinations produced with ever more inferior dyestuffs was accelerated by the shortages of a wartime economy in the 1360s Š./1980s. Many districts in the northwest, where previously the best wools and most honest colors were used, succumbed to the vogue for overfine, spongy imported yarns and debased palettes; dark blue, for example, was largely replaced by black, which is crisp and bold when new but soon fades to a streaky gray.

The most positive feature of the market, from the point of view of the trade, was the effect of the new export-valuation system on the supply of flat weaves. Beginning in 1364 Š./1985 these carpets received exceptionally low valuations for foreign-exchange purposes, which allowed prices in rials to climb to unprecedented levels while keeping costs in foreign currencies lower than they had been ten years earlier. As a result an abundant supply of exceptional pieces appeared on the market. The nomad who was offered a few thousand rials for his treasures in 1355 Š./1975 simply hid them, but the seven-figure prices being paid a decade later tempted the proudest weavers to part with their heirlooms, and flat weaves came to light in quantities and in a diversity of types and origins that even specialized dealers had barely suspected. Despite the gloomy outlook for the Persian economy while the war with Iraq dragged on, the technical and cultural skills to produce beautiful textiles remain strong in the country. It remains to be seen whether the ending of the war can create the favorable conditions under which these skills could flourish anew.

(P. R. Ford)

The term “tribal carpets” includes both floor coverings and other objects woven by nomads and semi-nomads for their own needs (see ʿašāyer; confederations, tribal). Although the terms “Persian carpet” and “Oriental carpet” theoretically cover all carpets produced in the Iranian world, as a rule they are taken to refer only to the finely knotted floral carpets manufactured in cities like Mašhad, Kermān, Kāšān, Isfahan, Tabrīz, and so on (see iii, above). Knowledge of rural carpets, especially nomadic carpets, is still very limited because specialists in carpet studies have generally interested themselves only in court carpets from the 10th/16th and 11th/17th centuries, often going so far as to dismiss other weavings as peripheral local products of no scholarly interest. For example, the noted early authority Wilhelm von Bode stressed the necessity of distinguishing “mainstream Persian from provincial work and the typical from what is only of historical or local interest” (p. 7). Half a century later Kurt Erdmann expressed the same fundamental notion in an only slightly modified form (p. 44). The European assumption that an art form is generally developed in an ecclesiastical or courtly environment and then is imitated with varying degrees of simplification and misunderstanding in more popular settings cannot, however, be applied to study of Persian carpets. In fact, the carpet was not introduced as a courtly art but was evolved among nomadic peoples at an early date (see vi, above). During the approximately 150 years when Safavid power was at its height (907-1077/1501-1666) perhaps 2,000-3,000 court carpets were produced; connoisseurs are familiar with most of those that survive (see ix, above). In the same period, however, approximately 2 million carpets were made by nomads, peasants, and city dwellers on the Iranian plateau. These weavings should not automatically be labeled Safavid carpets, however, for they cannot be viewed as such (Āzādī, 1978d, pp. 262-64).

In Persia rural carpets have been made in nearly every possible technical variation and for a wide range of uses (see v, above). Yet there are many nomadic groups whose works are absolutely unknown, and the weavings of other groups have been only very imperfectly studied and described. For that reason there are still many objects of which the function is obscure; in this connection it must be acknowledged that at present the nomads of Iran, as elsewhere, are experiencing a sudden and irreversible change in their way of life, owing to the impact of modern communications media. Tribal weavings can thus already be said to belong to an earlier cultural phase.

Almost all nomadic weavers work on horizontal looms, in contrast to village and urban weavers, who usually work on vertical looms (see iii, v, above). For warps, wefts, and pile they normally use wool from sheep that they shear themselves. The wool used for warps is two-plied with an S twist and Z spin; usually the wefts are also of wool, two-plied and Z spun but with an S twist so slight that they cannot properly be said to be twisted. The Baḵtīārī almost always used undyed cotton for warps and wefts, though wool was the primary material for knotting. The Šāhsevan and Afšār used cotton for wefts and the former occasionally for warps as well. Both warps and wefts of Qaragözlü works (so-called “Hamadān carpets”; see below) from the turn of this century and earlier were of cotton, the main feature that distinguishes these works from carpets produced by the Šāhsevan from around Sarāb. The latter are identical or very similar in color and design and in the extensive use of undyed camel hair; usually, however, there are two weft shoots after each row of knots, whereas in Qaragözlü pieces from the Hamadān area there is only one. Among other groups camel hair is seldom used for warps, though occasionally it is used for pile, usually undyed. Goat hair is not often used in Persia, though it is more common in Afghanistan. The use of silk for wefts is attested in the finest objects produced by the Qašqāʾī, the Ḵamsa confederation, and the Göklan Turkmen (Gūklān). Whereas it was the custom in the first two tribes to introduce pairs of ruby-red silk wefts only in pieces of exceptional quality, the Göklan normally used undyed silk wefts, a practice that distinguished their work from that of all other Turkman tribes in Iran and elsewhere (Āzādī, 1989); they were able to do so because they cultivated silk. It is rare for silk to be used for knotting in nomad works; when it is used, it is limited to small areas and is a sign of the highest luxury.

Nomads traditionally obtained most of their dyes from natural sources (see ii, above). The most common red was madder (rūnās; Rubia tinctorum). Yellow was obtained from weld (esparak; Reseda luteola) and turmeric (zard-čūba; Curcuma longa), as well as from grape leaves (barg-e mow), pomegranate rinds (pūst-e anār), walnut shells (pūst-e gerdū), and many other vegetal sources. Indigo (nīl; Indigofera tinctoria) usually had to be purchased or obtained through barter. Although the plant was cultivated in Kermān and Ḵūzestān, the best kinds came from India. Most nomads dyed their wool themselves, but for certain shades, like blue, the yarn had to be sent to professional dyers.

Tribal carpets are known to have been produced in the following regions.

Azerbaijan. The Šāhsevan in particular have been making carpets, mainly gelīms and sumaks but also knotted carpets, since the 11th/17th century (see v, above; Āzādī and Andrews; Housego; Tanavoli). In contrast to the floral carpets from urban workshops in Azerbaijan, the Šāhsevan carpets have geometric designs. Such carpets were first comprehensively exhibited in 1971 at the Museum für Kunst und Gewerbe in Hamburg and the Museum für Kunsthandwerk in Frankfurt (Meister and Azadi). In addition to the Šāhsevan, the Kurdish tribes of Azerbaijan, especially in the Qarādāḡ region, weave carpets that are very similar to those of the Šāhsevan and of Kurdish groups in the neighboring districts of Iraq, Turkey, and the southern Caucasus.

Kurdistan. Aside from the urban products of this western Persian province there are also Kurdish nomadic carpets, the best known of which are those of the Sanjābī and Kolyāʾī, though until the recent publication of two monographs (Eagleton; Buckhurst-Hill) little research had been done on Kurdish weaving at all. On the other hand, there are Afšār in Kurdistan, mainly settled north of Bījār, whose works have hardly been studied; their early production in this province had red wool wefts, as did the carpets of the main body of this tribe, which is located primarily in the province of Kermān (see below). There are also several Šāhsevan groups in Kurdistan; they produce work similar to that of their fellow tribesmen in Azerbaijan, though naturally with some differences. For example, their mafrašes (see Figure 103.1 ) are ornamented only on the front sides, whereas in Azerbaijan all four sides are decorated. Until the turn of this century members of the Qaragözlü (Qaragozlū) tribe, who may already have been half-settled, wove carpets; most so-called “Hamadān carpets,” which incorporate undyed camel’s hair, are products of the Qaragözlü.

Isfahan. The following nomadic tribes in this province have traditions of rug weaving: the Lorī (Opie; Collins); the Baḵtīārī, the largest tribal group in Iran (s.v. baḵtǰārǰ carpets); the Kurds, probably Lak, whose works are unknown; and the settled Qašqāʾī, Lorī, and other groups, who presumably play a significant role as weavers of so-called Hinegun carpets.

Fārs. The Qašqāʾī, with their numerous subtribes and clans, form the largest weaving group (Black and Loveless, 1979; Parhām and Āzādī; Azadi, 1987). Other tribal weavers in the province include the Boir Aḥmad (q.v.; Löffler and Friedl); the Mamasanī; and the Ḵamsa confederation, which consists of the Nafar, ʿArab (q.v.), Bāṣerī (q.v.), Bahārlū (q.v.), and Aynallū (q.v.) tribes, on whose work no research seems to have been conducted. It is not known whether the Afšār and Kurds of this province produced carpets.

Kermān. The main tribal carpet weavers are Afšār, whose works have been treated only superficially in general works. Although the majority of the Baluch tribes live in Kermān and Baluchistan, they produce very few carpets. Whether or not the Kurds of the province of Kermān weave carpets is entirely unknown.

Khorasan. This province has the most varied population of tribal weavers in Iran. Beside the Baluch (see baluchistan v; Azadi and Besim) the Tīmūrī, the Brahui (q.v.), the ʿArab (mainly around Ferdows and Ṭabas), and the Barbarī all work in the Baluch tradition, as do the Jamšīdī, whose carpets are somewhat cruder. The Kurds produce several types of carpet; as they have mingled with neighboring tribes, some of these types show the influence of other traditions. There are a purely Kurdish type, the Qūčān carpets, which are knotted and have polychrome geometric designs; knotted carpets that are somewhat darker than the Qūčān, though still lighter than the monochromatic Baluch carpets; and quasi-monochromatic geometric carpets in the Baluch tradition. Among the Turkmen three main tribes have woven carpets in northern Khorasan for a very long time. The largest is the Yomūt (Yamūt), with numerous subtribes and clans. The range of objects produced, in both knotted and flat-woven techniques, is more varied than that of almost any other Turkman tribe (Figure 102, Figure 103). The works are generally characterized by geometric ornament and monochromatic color schemes (Azadi, 1975; idem and Vossen). The work of a second Turkman tribe, the Göklan, which has lived in the area of Gorgān and Atrak for at least 700 years, is almost always incorrectly ascribed to the Yomūt (Azadi, 1986). The third tribe is the Tekke (Takka), small groups of which live with the Awlīād tribes. The Afšār of Khorasan weave and knot carpets, but their work is largely unknown (Azadi, 1979). Although other tribal groups have long been established in Khorasan, for instance, the Lak, Qājār, Jalāyer, Bayāt, Qelīčī, and Gārīlī, it is not known whether or not they produced carpets, either in flat-woven or knotted technique.

Māzandarān and Gīlān. There are also tribes in this province whose works are unknown. The Kurdish work of Kalārdašt has been known only since the beginning of the 1340s Š./1960s. In addition, there is no information on whether the Qājār made knotted carpets.

Tehran. Knowledge of tribal weaving in this province is limited to the works of the Šāhsevan, Hadāvandī Lor, Borbor, Afšār, and other tribes from the vicinity of Varāmīn and some groups at Garmsīr (Garmsār).

As mentioned above the nomads produce many objects in both flat-woven and knotted techniques for their own use. Originally they were only simple pieces designed to fulfill daily needs, but at some point objects began to be made for functions that had been elaborated beyond daily needs. It seems, for example, that such objects as horse blankets, the trappings of the marriage camel, knee decorations, and double bags constitute a separate tradition, one that is still partly observed and followed, for example, among the Yomūt Turkman tribes in Iran, as can be seen in the drawings of specific items in Figure 102, Figure 103. In addition to the illustrated objects, the following are included in the common Yomūt repertoire: the germeč, a threshold carpet between a gap and a torba in size but ornamented in the same way as the added lower border (alam) of an ensi (see Figure 103.1, 9, 15); the ojaq bāš, a U-shaped hearth carpet, measuring 250-350 by 150-80 cm, in the opening of which the fire is placed; the at čeki, the saddle girth, which in exceptional cases can also be made in knotted technique; and the hali, or qali, measuring 170-350 by 120-230cm, a Turkman domestic carpet that has taken on cultic significance and must be understood as a highly precious object. The terms and explanations for all these objects have been drawn both from the literature and from research in the field (Āzādī, 1975; idem and Vossen).

Bibliography : P. A. Andrews, The Turcoman of Iran, Kendal, Eng., 1971. S. Azadi, “Orientalische Flachgewebe,” Die Weltkunst 6, 1971, pp. 286-87; 9, 1971, p. 510. Idem, “Rendez-vous mit Raritäten. Kaschkai,” Textile Heimkultur 3, 1973, p. 42. Idem, Turkoman Carpets, tr. R. Pinner, Fishguard, Wales, 1975. Idem, “Ein Gartenteppich der Nomaden. Bericht über einen Teppich der Bakhtiaris,” Kunst und Antiquitäten 1, 1977a, pp. 50-52. Idem, Farš-e Īrān/Persian Carpets, Eng. tr. R. Pinner, Hamburg, 1977b. Idem, “Einige Teppiche in Belutschtradition,” Die Weltkunst 48/8, 1978a, pp. 876-77. Idem, “Kaschkai-Gaben, spezielle Knüpfarbeit der südpersischen Nomaden,” Die Weltkunst 12, 1978b, pp. 1060-62. Idem, “Some Methodological Aspects and Problems of Research into Oriental Carpets,” Halı 3/3, 1978c, pp. 262-64. Idem, “Die vernachlägten persischen Bauern- und Nomaden-Teppiche — Eine neue Afschargruppe,” Halı 2/4, 1979, pp. 289-91. Idem, Mystik der Gab-beh, Hamburg, 1987. Idem, in Oriental Textile and Carpet Studies 3, 1989, forthcoming. Idem and P. A. Andrews, Mafrash. Woven Transport Packs as an Art Form among the Shahsewan and Other Nomads in Persia, Berlin and Munich, 1985. S. Azadi and A. Besim, Carpets in the Baluch Tradition, Munich, 1986. S. Azadi and R. Vossen, Turkmenische Teppiche und die ethnographische Bedeutung ihrer Ornamente, Hamburg, 1970. R. D. Biggs, ed., Discoveries from Kurdish Looms, Evanston, Ill., 1983. D. Black and C. Loveless, Rugs of the Wandering Baluchi, London, 1976. Idem, eds., Woven Gardens, London, 1979. W. von Bode, Altpersische Knüpfteppiche, Berlin, 1904. J. J. Collins, Jr., Shiraz, Newburyport, Mass., 1987. W. Eagleton, An Introduction to Kurdish Rugs and Other Weavings, London and New York, 1988. A. C. Edwards, The Persian Carpet. A Survey of the Carpet-Weaving Industry of Persia, London, 1953. M. E. Enay and S. Azadi, Einhundert Jahre Orientteppich-Literatur, Hanover, 1977. K. Erdmann, Der orientalische Knüpfteppich, Tübingen, 1960. A. de Franchis and J. T. Wertime, Lori and Bakhtiyari Flatweaves, Tehran, 1976. J. S. Gluck, A Survey of Persian Handicrafts, Tehran, 1977. J. Housego, Tribal Rugs. An Introduction to the Weaving of the Tribes of Iran, London, 1978. A. Ittig, “Baḵtīārī Carpets,” in EIr. III, pp. 561-62. H. M. Jones, R. S. Yohe, and J. W. Boucher, Persian Tribal Rugs, Washington, D. C., 1971. R. Löffler and E. Friedl, “Eine ethnographische Sammlung von den Boir Ahmad, Südiran,” Archiv für Völkerkunde 21, 1967, pp. 95-208. Idem and A. Janata, “Die materielle Kultur von Boir Ahmad, Südiran,” Archiv für Völkerkunde 28, 1974, pp. 61-142. D. W. Martin, “Die Gabbehs von Fars,” Halı 5/4, 1983, pp. 462-73. P. W. Meister and S. Azadi, Persische Teppiche, Hamburg, 1971. J. Opie, Tribal Rugs of Southern Persia, Portland, 1981. S. Parhām, Qali Bulvardi, Tehran, 1973. Idem and S. Āzādī, Dastbāfhā-ye ʿašayerī wa rūstāʾī-e Fārs, Tehran, 1365 Š./1986. The Qashqai of Iran, Manchester, 1976. H. Reinisch, The Georges D. Bornet Collection, Graz and London, 1986. K. Schlaminger and P. L. Wilson, Weaver of Tales. Persian Picture Rugs, Munich, 1980. W. Stanzer, Kordi, Vienna, 1988. P. Tanavoli, Lion Rugs of Fars, Washington, D.C., 1974. Idem, “Gabbeh,” Halı 5/4, 1983, pp. 474-76. Idem, Lion Rugs. The Lion in the Art and Culture of Iran, Basel, 1985a. Idem, Shahsavan. Iranian Rugs and Textiles, Fribourg and New York, 1985b. Idem, “The Afshar. 1. A Tribal History,” Halı 371, 1988, pp. 22-29. Idem, Kings, Heroes, and Lovers. Pictorial Rugs from the Tribes and Villages of Iran, New York and London, 1989. J. T. Wertime, “Flat-Woven Structures Found in Nomadic and Village Weavings from the Near East and Central Asia,” Textile Museum Journal 18, 1979, pp. 33-54. Idem, “Salt Bags from Iran,” Halı 2/3, 1979, pp. 198-205. Idem et al., Tribal Animal Covers from Iran, Tehran, 1975.

(Siawosch Azadi)

Figure 64. Foundation: a. warps; b. wefts

Figure 65. Knots: a. symmetrical knots; b. asymmetrical knots, open right; c. asymmetrical knots, open left

Figure 66. Loom: a. warp; b. shed stick; c. heddle rod; d. weft

Figure 67. Side finishes: a. overcasting; b: selvedges

Figure 68. Central-medallion format a. Field; b. Main border stripe; c. Minor border stripes; d. Medallion; e. Guards

Figure 69. Cartouche

Figure 70. Arabesque

Figure 71. Bota

Figure 72. Palmette

Figure 73. Herātī pattern unit

Figure 74. S-stem

Figure 75. Balanced plain weave. Drawing by Houshang Adorbehi; after Tanavoli, p. 67, fig. 51

Figure 76. Loosely packed weft-faced plain weave. Drawing by Houshang Adorbehi; after Tanavoli, p. 68, fig. 53

Figure 77. Tightly packed warp-faced plain weave. Drawing by Houshang Adorbehi; after Tanavoli, p. 68, fig. 52

Figure 78. Loosely packed weft-faced plain weave with discontinuous weft, known as slit-tapestry weave. After Acar, p. 47, design 3

Figure 79. Dovetailing (shared warp). After Housego p. 22, no. 4

Figure 80. Single-interlocking wefts. After Burnham, p. 145, center

Figure 81. Double-interlocking wefts (front of carpet). After Housego, p. 23, no. 5

Figure 82. Double-interlocking wefts (back of carpet). After Housego, p. 23, no. 6

Figure 83. Twill weave in a ratio of 2:2. Drawing by Houshang Adorbehi; after Tanavoli, p. 72, fig. 62

Figure 84. The warp-float face of a float weave with weft passes of 3:1 and 1:1 in alternating rows, with the span under three warps in each successive row of 3:1 offset one warp. Drawing by Houshang Adorbehi; after Tanavoli, p. 72, fig. 64

Figure 85. Face of extra-weft float brocading with spans of different lengths on balanced plain-weave ground. Drawing by Houshang Adorbehi; after Tanavoli, p. 73, fig. 66

Figure 86. Back of weaving in Figure 85, showing reverse pattern created by extra wefts, which float on the back when not used on the front. Drawing by Houshang Adorhehi; after Tanavoli. p. 73, fig. 67

Figure 87. Two-color complementary-weft compound weave forming lozenge motifs. Drawing by Houshang Adorbehi; after Tanavoli, p. 77, fig. 77

Figure 88. Plain two-strand, two-color weft twining on balanced plain-weave ground. Drawing by Houshang Adorbehi; after Tanavoli, p. 78, fig. 81

Figure 89. Countered two-strand, two-color weft twining on balanced plain-weave ground. Drawing by Houshang Adorbehi; after Tanavoli, p. 78, fig. 82 .

Figure 90. Plain extra-weft wrapping (plain sumak) with ground weft. After Acar, P. 70. design 20

Figure 91. Countered extra-weft wrapping (countered sumak), with ground weft. After Acar, p. 72, design 22

Figure 92. Reverse extra-weft wrapping (reverse sumak), with ground weft. Acar, p. 74, design 24

Figure 93. Diagonal supplementary (“extra”) weft wrapping on plain weave with discontinuous wefts. After Acar, p. 53, design 9

Figure 94. Structure of zīlū carpet, showing both faces. For demonstration purposes the diagram depicts the structure of a fabric with a checked pattern. Drawing by Milton Sonday (Copyright 1989)

Figure 95. Fringe produced by knotting groups of warps together. Drawing by Houshang Adorbehi; after Tanavoli, p. 99, fig. 146

Figure 96. Fringe produced by knotting alternate groups of warps together in successive rows to create a net or web. Drawing by Houshang Adorbehi; after Tanavoli, p. 99, fig. 148

Figure 97. Fringe produced by interlooping warps. Drawing by Houshang Adorbehi; after Tanavoli, p. 100, fig. 150

Figure 98. Fringe produced by braiding groups of warps in flat vertical plaits. Drawing by Houshang Adorbehi; after Tanavoli, p. 100, fig. 152

Figure 99. Fringe produced by plaiting warps diagonally. Drawing by Houshang Adorbehi; after Tanavoli, p. 101, fig. 154

Figure 100. Fringe produced by twisting pairs of warps, and leaving ends uncut. Drawing by Houshang Adorbehi; after Tanavoli, p. 101, fig. 156

Figure 101. Fringe produced by twisting groups of warps together. Drawing by Houshang Adorbehi; after Tanavoli, p. 101, fig. 158

Figure 102. Examples of Yomūt weaving. 1. Tent band; the finest examples are considered to be those with white ground (aq yüp) in warp-faced weave with knotted decoration. Tent bands range from 11 to 20 m long and from 20 to 75 cm wide. 2. Kaplyk, frame for the interior of the tent entrance, ornamented with fringes, tassels, and sequins. 3. Horse blanket (tainakša, konakča, at joli). 4. Saddle cover (čerlyk, ayerlyk). 5. Kalyk, similar to the kaplyk but much smaller, decorative frame for the opening of the camel litter (kajava) in the bridal procession.

Figure 103. Yomūt woven objects. 1. Qap, qab, mafraš, small bag that Turkman women use for storing their jewelry and other valuables. 2. Dis (duz) torba, small bag for storing salt, an extremely valuable commodity among nomads; similar bags in different shapes are used for sugar, tea, rye, and other staples. 3. Bokča, an unusual kind of bag the function of which has not been clearly established; often used for carrying the Koran. 4. Tutaš, tutač, gazan tutaš, small pockets for carrying large or small jars (dieg). 5. Another form of saddle blanket. 6. Darak baš, small bag woven on both sides and with projecting teeth, used for carding and cleaning wool. 7. Diah dizlyk, small pentagonal object covered with tassels and small sequins and used as knee decoration for the bridal camel. 8. Ok baš, hood for the stakes (ok) of the new white tent (aq oy) given to each bridal couple; the fifty to sixty stakes are tied in two bundles with an ok bas decorated with bells and sequins sheathing the sharp points at each end; the two bundles are carried by camels on the right and left sides of the procession. 9. Čowal, gaṛčīn, largest Turkman storage bag, for clothing and other household furnishings. 10. Torba, similar to the čowal but less than half as large. 11. Asmalyk, one of a pair of pentagonal or heptagonal objects, sometimes with pendants, placed on each side of the wedding camel under the litter and later hung in the tent as memorabilia (comparable to wedding pictures). 12. Čemča-torba, small bag for storing wooden spoons and other utensils. 13. Ḵorjīn, double bag. 14. Salaṯčak, used as a weighing bag or cradle, with inherent magical function of protecting babies from illness, insects, snakes, and other threats. 15. Namaslyk, ayatlyk, prayer carpet, with specific form, dimensions, and ornamentation, used especially in funerary rites. 16. Ensi, engsi, carpet for the tent entrance, with specific dimensions and ornamentation, much of which seems to have an amuletic character.

Plate CV. Drawing of Pazyryk Carpet; after K. Jettmar, Art of Steppes, Art of the World, New York: Random House, 1964, fig. 103 p. 121

Plate CVI. “Lion Carpet”—reconstruction drawing by Steven Sechovec. Museum purchase, the Rosco and Margaret Oakes Income Fund, The Fine Arts Museums of San Francisco, no. 1986.4

Plate CVII. “Laylī and Majnūn at School,” detail, from a copy of Neẓāmī, Ḵamsa, illustrated in Baghdad or western Persia in 866/1461. Topkapı Sarayı , Istanbul, Hazine 761, fol. 106r

Plate CVIII. Medallion carpet, Museo Poldi Pezzoli, Milan, no. DT 1 (partly restored)

Plate CIX. Ardabīl carpet. Victoria and Albert Museum, London, no. 272-1893

Plate CX. Emperor’s carpet (one of a pair). The Metropolitan Museum of Art, New York, no. 43.121.1: Rogers Fund, 1943

Plate CXI. Polonaise carpet, Rosenborg Castle, Copenhagen, no. 31-R1

Plate CXII. Polonaise carpet, Rosenborg Castle, Copenhagen, no. 31-R1, detail

Plate CXIII. Millefiori carpet attributed to Persia in the second half of the 12th/18th century. Victoria and Albert Museum, London, T.99-1973

Plate CXIV. Cartoon for one-quarter of a medallion-and-corner carpet commissioned by Emil Alpiger from the Farāhān, (šahrestān of) Solṭānābād, workshop, 1311/1893-94; private collection; watercolor on hand-drawn graph paper, 39 x 40 cm

Plate CXV. Cartoon for one-quarter of a medallion-and-corner carpet produced for Ziegler’s before 1313/1896; private collection; watercolor on hand-drawn graph paper, 50 x 50 cm

Table 48. Distribution of looms, 1319 Š./1940

Table 49. Persian carpet exports before World War II

Table 50. Knotted-pile carpet imports into U.S.A. and West Germany 1978-1987

Table 51. Prices of Nehāvand rugs in the Hamadān bāzār (rials per m²)

tem esse traco aqui?

tem esse traco aqui mesmo?

Backward slash, missing in the font.

que letra ı essa?

idem

q letra ı essa?

idem