Wednesday 3 December 2014

Doors of Jannah-Turkestan West

A group of fragments from the Magaki Mosque in Bukhara woven in a pure,kilim-like style have surfaced over the last few years at various auctions.The original Saph is said to have been constructed in 1874.The most substantial piece is still held by the Mosque(actually 2 pieces sewn together)


Another fragment from presumably the same carpet was auctioned at Rippon-Boswell`s in 2014,achieving a hefty $166,110

136-RB 31.5.2014-96

A last fragment was published by Bausback in his 75 Years Anniversary Catalogue


Martin Andersen has  published a graphic collage


Two fragments from another carpet both surfaced in London.They were commented upon by Hali as being significantly older than the preceeding,an assumption supported by their condition.They are more in the style of the Beshir prayer rugs typified by a piece in the Dudin Collection


The larger fragment was sold to a solitary bidder for $94,355 in London 1998

140-Sothebys 29 April 1998(96)

A second fragment sold at Christies London on 17 October 2002(141)for $27,730


Interestingly the serrated outlines seem to imitate dovetail kilim technique.

Further reading:Hali APG 99-127;126-135;180-138.

A group of three Ersari carpets with chevron designs appear also to have been inspired by flatwoven counterparts


143-Skinners 5 December 2009(71)

144-Herrmann,SOT X-97

A number of other Turkmen manufacturers specialised in Chevron design Saph carpets with kilim designs,such as the "Eagle-Göl" group.A batch of knotted carpets in traditional prayer rug design is attributed to the Middle Amu Darya Ersari,no doubt produced in large urban centers

145-Straka Collection

The Straka Collection  piece above is generally considered the best of a poor lot.Of unmitigated ugliness was an enormous carpet sold at Christies on 9 April 1988(79) for a measly $6,050

146-Christies 1988

Another large item went at Skinners in 2002

147-Skinners 20 April 2002(54)

Two related examples were published by Schurmann and Eiland,both featuring a mihrab-within-a-mihrab scheme

148-Central Asian Rugs 47


A version with stacked mihrabs appeared at Nagels 37th Sale

150-Nagels 37(27)

Published by David Sorgato in Ghereh 39


The following recently sent in by Michael Black features an appealing Mina Khani in spontaneous style

152-Courtesy Michael Black

Not a Saph,but a small carpet with the remains of its original Shirasi,the following Lot was sold at Rippon Boswells on 19 May 2007 for 7000 €

153-RB 2007

Last but not least a solitary panel from an Ersari Saph in the boteh style


A small contingent of Turkmen trappings mirrors the saph theme,especially a group from the Ersari/Kizilayak Zone


A smaller group from Central Asia,known commonly as "Uzbek" feature a more pictorial view


Plus a late entry kindly sent in by Igo Licht

157-Igo Licht

Doors of Jannah-India

Stone floor Saphs are not uncommon in India

116-Moti Masjid,Red Fort,Delhi

117-Jami Mosque,Hyderabad

Knotted Saphs are less frequent.A Mughal era Pashmina Prayer carpet has been suggested as a possibility,although Daniel Walker in Flowers Underfoot(95) seems to think it may have been part of a Tent-hanging,or Qanat.The Aynard carpet,now Thyssen-Bornemisza Collection,is nevertheless a spectacular product.Its border plan certainly suggest a multi-niche arrangement-the carpet was also woven in a non-perpendicular fashion

118-Thyssen Collection Catalogue,Beattie,plate IX

For comparison,a velvet tent-panel(qanat) now in the MET


From the 17th-18th century,there are three basic types of Indian Saphs.
The first has been attributed to North India(Keir Collection-Spuhler,Beselin)or to Warangal in the Deccan(Steven Cohen,Daniel Walker)This type is stylistically quite different from the other two,although sharing similar construction features(z4-6 warp,perpendicular weave)and more prosaic colouring with tans and duller secondaries.

The most well-known example is in the Keir Collection and consists of a long,probably 7 niche carpet made up into two rugs


The borders of both carpets have been renewed and it is likely that a panel with pommegranate design would have been the centrepiece.This pommegranate design went on to become a popular design in the Tarim Basin.It only occurs on the four pieces in this first group

121-Lefevre 1978

122-Sandra Whitman Hali 142

Thus emphasising the close cultural ties between India and East Turkestan.

A second carpet,apparently complete,was sold at Lefevre`s in 1978 for 58,000 pounds,and may or may not be the carpet already illustrated by Jekyll in the 1920`s

123-Lefevre 6 October 1978

124-Jekyll 2 (1928-9)

A last piece is often mentioned as having been in the Dumbarton Oaks Collection,and seems slightly closer to the Jekylls rug in design(the flattened rosettes in the spandrels)and condition(similar old repairs in the first spandrel,right)

125-Textile Museum Journal II-19

The first group with its "Taj Mahal" look and more ornate borders seems decidely up-market when compared to the Warangal group,in which a strict architecture has been replaced by more window-like panels,simpler border-guards and a frequent intrusion by large trefoil outer enclosures.A looser,more playful atmosphere is afoot.There are two types:a white field  with one repeating design,and a multicoloured layout with differing mihrabs.
Six whiteground examples are known.The first appeared at the Untermeyer Auction in 1940,one of two examples on offer plus a single niche fragment

126-Park-Bernet May 11 1940(210-211)

It reappeared at Sothebys on 25 November 2008 in a truncated form,missing its left border,but still selling for $ 6,875

127-Sothebys 2008

A third time around in colour it sold for $8750 at Sothebys NY on 25 November 2008(139)

128-Sothebys 2008

Yet another four-panel fragment from the same carpet appeared at Christies on 7 October 2010.Perhaps it is the second,unillustrated Untermeyer piece(Lot 211)?

129-Christies 2010

A fragment in the Tapi Collection may well be the missing single niche item from the Untermeyer Sale,with its long vertical stem

130-Foto courtesy Markus Voigt

A similar example,but with elongated mihrab"neck" is in the Reeves Collection,Art Museum Dallas


Two further single panels are known.The first,in the V&A,is said to have come from the Treasury of the Nizam of Hyderabad.It closely resembles the Reeve example

132-S.Cohen,Sultans of the South 125

A last example in Philadelphia has been woven entirely in cotton,with the exception of red areas piled in Pashmina,yet it resembles the others in this group,even mimicking the champfered column protrusions towards the apex of the Mihrab

133-Ellis,Philadelphia,plate 66

Even in the 19th century interesting and beautiful Saph-like carpets were produced,as can be seen in an all-cotton small piece once with Bausback,which seems to have been inspired by Kalamkar textiles

134-78 x 98 cms

Tuesday 2 December 2014

Doors of Jannah-The Caucasus and Persia

Considering the business-hold that Caucasian manufacturers had on the prayer rug market in the 19th century,the lack of Saphs from the area is puzzling.Perhaps the answer lies in their concentration on small rugs: the lions-share of large carpets was held by the British and German companies working in Persia,Turkey and India.Be that as it may,very few(mostly mediocre) examples were produced,chiefly in the "Shirvan" Zone.

A carpet from the Michaelian Collection must take pride of place,having been frequently exhibited and finally auctioned at the Vojtech Blau sale in 2006 for $39,000,a price stimulated no doubt by its "Marasali" design

80-Sothebys 14 December 1996(74)

A carpet dated 1911 was auctioned some years before at Sothebys for $6520.Its panels feature a recycling of standard designs

81-Sothebys 2 May 2001(25)

A very similar example was sold earlier still at Christies London for 3,240 pounds

82-Christies 25 October 1984(26)

A carpet with Maktabi featured the classic 5 panel format mostly commonly encountered in Caucasian saphs,and a weave typical of Talish rugs

83-Hali 115-137

Similar in feeling was a piece auctioned at Christies for $4,200,as "Genje"

84-Christies 23 June 2006(81)

Two later Southern Caucasian weavings share a similar design.The first was published in "The Treasure of the Caucasus" in 1993

85-Keshishian Collection

A second example was with the German-Turkish dealer Sari in Karlsruhe

86-ebay 2008

Another Internet appearance was made by a rare example with Lesghi star


A last example with recycled designs is in a Russian collection



Little is known about traditional mosque fittings in Persia.It is claimed that zilus,flatwoven items chiefly held in blue and white,were formerly used.Few have survived,but some are apparently of great age as an example in the Hermitage museum shows,which is dated to the 14th century


Parviz Tanavoli illustrates a Zilu Saph

90-Persian Flatweaves(196)

A photograph from Jon Thompson shows their use


Yet the final balance of knotted rugs is meagre.Only four pieces are known,mostly of venerable age,and nothing from the 19th century.A salient example,divided between four collections,is one of the most beautiful of all known Saphs,and a highpoint of Safavid weaving(although sometimes ascribed to India)Michael Franses has argued that it was a gift of  Shah Tahmasp to the Ottoman Sultan Suleyman.The largest piece is said to have come from the Suleymaniye Mosque,although there is no proof of this.Just why the Sunni Sultan would want to place a carpet from a Shia monarch in his new Mosque is not clear.The Saphs previously examined in the Turkish section are stylistically world`s apart from the Classic Safavid patterning,but do echo the spirit of the Ottoman Society.The Persian carpet would surely have seemed out of place.Be that as it may,the four fragments were all at one time in Istanbul.One large piece is still in the TIEM,the others are in Berlin,Kuwait,and the former Wher Collection.The rhythmic patterning varies from row to row,alternating between medallion field and arabesque.The arabesque is not unlike that found on the Stora carpet,the smaller version of the Ardabil,with a counterpoint of large lotus leaves and small palmettes

92-The Stora Medallion Carpet,Keir Collection


An attribution to North-west Persia,on technical grounds,seems clear.The carpet was woven in a non-perpendicular style,thus necessitating a very large loom.

94-Istanbul,TIEM 105,after Gans-Ruedin

95-Istanbul TIEM 105-Hali ICOC Istanbul guide




There is a considerable difference in the duktus of the Berlin carpet`s mihrab,indicating that it may have been part of a second example


A last fragment,once with Hugh Moss,passed to the Wher Collection.It resembles the Berlin fragment

100-Grosvenor House catalogue 1973(65)

Two Saphs have been kept at the shrine of Imam Ali at Najaf,Irak,for hundreds of years,and are presumably still there if sectarian violence has not displaced them.Local tradition has it that they were donated by Shah Abbas,and one is signed with the typical phrase,"Abbas,dog of this shrine".The carpets,a complete and fragmentary piece,are closely related,as are two further border fragments in the collection.They are woven in wool with metal thread brocading(in the ground of the mihrab and spandrels)thus locating them in the Salting group zone.Their ultimate publication in colour with full analysis will one day be a major event.


There is more than a hint in the direction of the vase carpets with their split palmettes and Lotus whirligigs

102-McMullan Collection


Leading on from the Najaf pieces are two fragments,both of which were once with Bernheimer

104-MET,ex Ballard


The whereabouts of the second fragment are unclear.They are said to have been made in either Central or North East Iran


A well published pair of Saphs are now attributed to Khorasan.Several fragments have appeared on the market,starting with the largest,the Godard-Keshishian


This mighty fragment first appeared at  Christies London on 28 April 1995 where it sold for $47,840.It was accompanied by a corner piece which fits quite nicely in the missing corner,although it does not belong there:the indentation is thus woven,presumably to circumnavigate an architectural obstruction.The corner piece was sold as the next lot for $3680.Just why it was woven is a mystery.


109-The Saph with its ghost-corner

The major whiteground border seems to have been inspired by a type seen on NW Persian medallion carpets,such as the MET`s Blumenthal

110-MET wb-14.100-2

The carpet was re-auctioned at Sothebys New York,sans corner,in 2010,and this time its importance was rewarded with a sale price of $218,500.

Another large fragment,presumably a pair to the above,made a brief appearance at Boisgirard`s sale in Paris on 19 March 2004,where it was bought in against a bid of €105,000.(see hali 134-13)This was unfortunate,as it later made the same price but in dollars,at Christies London on 4 October 2011(200)It is said to have been in a Basle Collection

111-Christies 2011

Interestingly,the remains of an indentation can be seen in the upper right hand corner,showing that the original carpet had been wedge-shaped(or actually,like a rectangular mihrab)Here a simple reconstruction of the two pieces together,although the originals would have faced off-not to forget the perpendicular weave with the mihrabs woven sideways


Perhaps there exists somewhere a mosque with such a lacunae,either in East Persia or India.

Some splendid fragments exist from these carpets,including an item offered by the Textile Gallery in Hali 1-4

113-Textile Gallery

and a piece once with Herrmann in Munich

114-Herrmann,ATT 1(7)

A small fragment is in the Keir Collection

115-Keir Collection 2(T30)