92 ON COLOUR. Paet I. 



of which any one who has a correct eye for colour may judge. 

 Indeed there are so many exceptions to the rules already sug- 

 gested that nothing can yet be accepted as a reliable guide ; 

 and I have already shown how erroneous a conclusion has been 

 arrived at on the subject of accidental colours (pp. 60, 74), 

 and how theory has propagated the error. But if I differ 

 from the opinions of some who have attempted to lay down 

 rules before they understood the subject, it is my desire to do 

 this with perfect respect for them, without any intention to 

 condemn their praiseworthy endeavours to impart instruction 

 to others; but having the same object, I cannot, on public 

 grounds, see opinions gain favour which are totally at variance 

 with sound experience, without entering my protest against 

 their hasty adoption. However plausible a theory may be, if 

 founded on erroneous notions it can only mislead; and I 

 appeal to all the practice of the Asiatics, the Egyptians, the 

 Greeks, the Moors of Spain, the Italians, the French, and 

 others who have been noted for their success in ornamenting 

 with colours, and to the convictions of those who have a 

 natural perception of their concords, to second me in my 

 assertion, that such combinations as these accidental colours, 

 viz. green and russet, orange and olive-green, purple and 

 citrine, are not concords, but offend against true harmony. A 

 yellow-orange with olive-green would be less objectionable 

 than a red-orange, and would at least accord with it by 

 analogy ; and an orange with some other hues of green would 

 be far preferable to orange with olive-green. 



And in order to show the effect of the secondary purple and 

 its accidental tertiary colour citrine ; and of the secondary 

 green and the tertiary russet; I introduce these colours as 

 specimens of discords, though they have been put forth as 

 harmonious combinations. (See Plate in. figs. 9, 10.) Again, 

 it has been said that each primary, with its accidental 

 secondary, and the tertiary which is accidental to this last, 



