338 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. j 



phylliim ; stronger stoiiecrops such as Seduvi albuiii are apt to smother ^ 

 them. They look their best iiesthng in patches close to small shrubs. | 

 There is nothing so important for the general effect of the rock | 

 garden as the right choice and placing of shnibs. In my opinion too 

 large shrubs are often used and misused. It seems to me that the larger j 

 Oistuses, such as Cistus cyprius or G. laurifolius and the larger brooms, i 

 are out of scale, except where used as a background in very large rock li 

 gardens. 



"What is needed is shrubs which, whether prostrate or not, are com- 

 pact in habit and, at least, look as if they were mountain plants. Only ' 

 the more prostrate shrubs, I think, should be planted on the top of a ; 

 mound ; upright shrubs look like .monuments in such places ; and we 

 must remember that in the mountains the heights are wind-swept, 

 and so covered only by the most prostrate plants. The beautiful 

 prostrate Savin never looks better than when it grows high up in a flat j 

 place surrounded with some grey stonecrop or with patches of Sern- i 

 pervivum ; and it is well to repeat the effect in another flat place a little 

 lower down and a few yards to one side of it. On the other hand, 

 rounded or upright shrubs look their best half-way down a slope with a , 

 background of bold rock-work; and, as I have said, they should have 

 no large plants near them to put them out of scale. In fact, the 

 more I see of rock gardens the more I am convinced of the importance of 

 filling them only with plants that are not only rock plants but look like 

 rock plants. We hear a great deal about the need for variety in rock 

 gardens, but there is, I think, more often too much variety than too 

 much monotony in them, and it is easy to get enough variety in colour, 

 leafage, and habit of growth with a selection of plants that are all [ 

 thoroughly in scale. The height can be varied, not only with the [ 

 smaller, upright, and rounded shrubs, but also with rock plants that 

 throw up erect flowering stalks, such as many Saxifrages, the Semper- ; 

 vivums, many rock pinks, the smaller Achilleas, AetJiionema grandi- j 

 florum, Lithospermum graniinifolium, Papaver alpmum, Aster alpiniis, ' 

 or Campanula harbata. And the smaller the rock garden the more 

 carefully scale should be observed. In very large rock gardens, with 

 very bold rock-work, the larger mountain plants may be grown, but 

 even then they should be kept well apart from the smaller alpines, 

 and particularly from Alpine shrubs. I have seen OstroivsMa luatfuifica \ 

 recommended as a rock plant, chiefly, I suppose, because it is con- 

 sidered difficult to grow, and Incarvillea Delavayi used to be often planted 

 in rock gardens, while it was still a novelty. Both these plants seem i 

 to me quite unsuitable, .unless we are to regard a rock garden only 

 as a nursery for rare or new or difficult plants. If we use it as a 

 means of achieving a kind of beauty not otherwise obtainable, w^e should 

 be careful to eliminate any plants w^hich, however beautiful in them- 

 selves, are not in character with that peculiar kind of beauty. 



