ORNAMENTATION OF CARPETS COLE. 133 



outer border show near affinity to ornament on old Chinese bronzes ; 

 the smaller details in the field are derived for the most part from 

 plant form. This cut-pile rug was probably made in the neighbor- 

 hood of Yarkand. 



Two rugs are shown on the next slide. That on the left (pi. 3, 

 fig. 3) is of cut pile and has three disks, each of which is surrounded 

 by curA^ed and spiral versions of the swastika. The border has a 

 variety of circular blossoms or emblems. This rug comes from 

 Yarkand. The second one (pi. 3, lig. 4) has three octagonal panels 

 instead of disks occupying the larger part of the field, which is else- 

 where filled in as the border is with various more or less abstract 

 details like those we have already discussed. Amongst them are a 

 few Chinese symbols of simple type. The rug is of closely-stitched 

 needlework, and is considered to be a Soumak rug, which is, I believe, 

 a corruption of Semaka, a town in Caucasia. The work corresponds 

 with that of some of the saddlebacks from this district. 



Of less interest in the history of carpet design is that of the familiar 

 red and green modern Turkey carpets. In these comfortable cut-pile 

 floor coverings, the unintelligible forms are, I think, remotely related 

 to those of the Asiatic rugs mingled with others distantly derived 

 from patterns that were being designed before or about the time when 

 Marco Polo traveled in Asia Minor and noted the fine carpets made 

 there. These were doubtless of a type of Mohammedan style, the 

 gradual development of which in Egypt on the one hand, and the 

 Mesopotamian districts of Persia on the other hand, commenced soon 

 after the eighth century. About then and for some time later on, 

 Asiatic rugs such as we have seen were used at the courts of the 

 Khalifs and Mohammedan governors in Egypt, Syria, Sicily, and 

 Spain, whence germs only, of the later taste for rugs and carpets, 

 were sparsely diffused in Eurojje. 



I have already said that at a period shortly preceding the Moham- 

 medan conquest, the ornamentation in Syria and western parts of 

 Persia, and to some extent in Egypt, was largely of a degenerate 

 Roman character with occasional traces of ancient Assyrian feeling. 

 It had but little Chinese flavor, and to give you a bare impression of 

 its character I have a few slides made from Coptic and Perso-Roman 

 specimens. 



The first is from a Coptic tapestry weaving, with an Egypto-Roman 

 style of ornament of the fifth or sixth century A. D. picked out in 

 needlework. It may have served as a couch or stool cover. The 

 greater part of its ornament consists of ingenious variations of the 

 Roman Guilloche. The intertwistings fall into repeated circles, 

 within which are blossoms, and from such may have descended the 

 fully developed plan of pattern seen in silk weavings of the period, 

 in which the repeated circlas were much larger and more widely sep- 



