256 Mr Sang on ihc recent hnprovcments of' 



found, while scarcely any gradual shading of the tints depend- 

 ing on the nature of the figure is to be seen. A still closer 

 examination explains at once the soui'ce of these imperfections. 

 The carpet is found to consist of two contiguous webs, inter- 

 mingled with each other in such a manner as to produce the 

 pattern : each of these webs, if woven singly, would have a 

 striped appearance, being parti -coloured in the woof. One set 

 of coloured stripes is thus imposed upon another; and in de- 

 signing the colours of the pattern, no selection beyond what is 

 afforded by the judicious arrangement of these stripes can be 

 made. The number of full colours is thus very limited : these 

 can only be obtained where woof traverses warp of the same 

 colour. To bring up then a part of the figure full red, red 

 warp of the whole breadth must be traversed by red woof of 

 the whole length of the spot ; these colours can be immediately 

 concealed by sending the threads to the other web, but were 

 they to remain long there, both webs would become monoto- 

 nous. It is therefore extremely difficult to avoid a strong ten- 

 dency to striping in the colours, and, except on the principal 

 part of the figure, the colours can hardly be well managed, the 

 secondary embellishments being almost matter of chance. Yet 

 in the face of all these difficulties, patterns of great beauty are 

 daily formed on the carpet loom. 



Tlie invention of the triple carpet by Mr Morton of Kil- 

 marnock, has almost removed these difficulties. This carpet is 

 composed of three webs, which interchange their threads in 

 order to produce the pattern. The primary object in the in- 

 troduction of the third web, appears to have been the obtain- 

 ing of greater variety and brilliancy of colouring ; but another 

 cui'ious effect has followed, that the two sides of the carpet are 

 not necessarily counterparts to each other. To a certain ex- 

 tent the figure of the under must depend on that of the upper 

 side, since threads may be needed from the under web to pro- 

 duce what is wanted in the chief pattern, but there still remains 

 the choice of an interchange of threads between the two infe- 

 rior webs. It is obvious that the tendency to striping must be 

 much less on this than on the common carpet, and that the de- 

 signer having a far greater choice of colours, may produce ef- 

 fects that could not before have been attempted. It appears 



