554. It is so difficult to make associations of hues which will have a satisfactory effect, that I prescribe _in general_ the mutual association of flowers whose colours belong to neighbouring scales. We must then separate Pink flowers from those that are either Scarlet or Crimson. Orange ” ” Orange-yellow, Yellow ” ” Greenish-yellow, I shall even go further in advising the separation of— Red flowers from Orange, Blue or Pink from Violet. _Where the Plants are indiscriminately mixed._ 555. Flowers which only present contrasts of hues and which spring from seeds sown thickly in borders, or in beds, will not have the same objection as when the roots are planted at a distance from each other. 556. Flowers, presenting a disagreeable contrast of hues, may still produce a good effect, when their assortment makes part of an arrangement of contrasts of colours, strongly opposed; in this case being no longer seen isolated, they become in a manner the element of a picture. SECTION II. ASSORTMENTS AS TO HARMONY OF ANALOGY. 557. I only reckon shrubs susceptible of this harmony; because only perennial plants afford to the horticulturist a guarantee that the flowers of one year will be identical with those of the preceding; consequently, if we plant woody shrubs in such manner as to secure a regular gradation of tones, the successive annual flowerings will be constantly according to that order. We may apply this kind of arrangement to standard roses. 558. But I do not advise any one to attempt submitting annuals to this arrangement, because of the uncertainty that exists in the tones of their colour. _Harmonies of Analogy of Hues._ 559. If I have spoken against the associations of _contrast_ of hues (554), I am more inclined to speak against the associations of _analogy_ of hues, always remembering the restrictions I have stated above (Section 1). We must not forget that my intention is to describe assortments, whose good effects are certain. Now the more the colours contrast, conformably to our law, the more latitude there will be, although the colours of individual flowers associated vary in tone and in hue. 560. An objection might be addressed to me _that the green of the leaves, which serves as a ground to the flowers, destroys the effect of their contrast_. But it is not so; and to be convinced of it, it is sufficient to fix upon a screen of green silk, two kinds of flowers in conformity with the arrangement of the coloured bands (_pl. 1. fig. 7_), and to look at them from a distance of about ten paces; for when the eye is fixed on two well-defined objects simultaneously, surrounding objects, especially distant ones, produce but feeble impressions. ON THE ART OF ASSORTING LIGNEOUS PLANTS IN GARDENS, SO AS TO DERIVE THE BEST POSSIBLE ADVANTAGE FROM THE COLOUR OF THEIR FOLIAGE.