DESCRIPTION OF PLATES AND WOODCUTS. Xlll 



Plate IV. — (continued.') 



Fig. 3 shows how little green is required to brighten up a design, 

 as I have already shown in the description of Plate i. (See p. 63, 

 and Blue, B 2, p. 132.) 



Fig. 4 is an instance of white on a red ground, generally so 

 heavy (see p. 149, and description of Plate in. Jigs. 13, 14), which 

 is made perfectly agreeable by the addition of blue and yellow in 

 this design. (See Blue, B 6, p. 133.) 



Fig. 5 is another instance of the manner in which red and blue 

 may be prevented from appearing purple by the intervention of 

 yellow and white, and how blue should be separated from yellow in 

 a pattern (as in a carpet) by a black line. (See the combination 

 in Blue, C 7, p. 134.) 



Fig. 6 is given as an instance of the propriety of making the 

 patterns rather irregular than exactly symmetrical and of equal 

 size, particularly in a carpet. The advantage of this is more 

 obvious in a large expanse, and the inaccuracies being there 

 smaller in proportion than in the figure here given, they do not 

 appear so evidently to the eye, though they have the desirable effect 

 of preventing that monotony which fatigues it in an exactly symme- 

 trical design. A better instance of this is given in Plate v. Jig. 2. 

 The colours are blue, scarlet, orange, black, white, yellow, and 

 purple.* (See Blue, E 8, p. 136.) 



Plate V. Jig. 1 is a carpet border of blue, scarlet, green, yellow, black 

 and white. (^ee Blue, D 1, p. 135.) Though the combination is 

 harmonious, the arrangement of the red next to the green is not 

 such as would be generally recommended ; nor should yellow border 

 the red ; but those defects are here remedied by the distribution 

 and proportion of the other colours ; and the whole is well balanced 

 and agreeable. 



Fig. 2 is intended to show how much more important is the 

 effect of the colour in a carpet than that of the pattern, as I have 

 observed in p. 20 ; and how much more agreeable is that irregularity 

 in certain parts of the pattern met with in Eastern carpets, than 

 the formal and symmetrical exactness thought so necessary in our 

 own. As I have said in the description of Plate iv. Jig. 6, the 

 irregularities are made more apparent in this small design than they 

 would be in the large expanse of a carpet, where they would give the 

 required variety without being actually apparent to the eye ; but 



* By mistake the yellows as well as the orange, have all been printed o r the 

 latter colour in the plate, instead of being alternately so. 



