333 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 



promontories to construct it with a great variety of aspects ; and in any 

 case there should be a sharp turn at the northerly exit, so that the exit 

 may not let in the north winds. 



The chief advantage of a valley is that it can be made to afford so 

 many aspects, and if it winds and has slopes of varying steepness it 

 escapes the monotony which is the chief fault of the valley form. 



A valley, of course, must be made by two ranges, and where there 

 is not unlimited space the inner slopes of these ranges should be, in the 

 main, gradual, and the outer more steep. Thus there will be more room 

 in the protected part of the rock garden than in the unprotected. But 

 here and there should be steep jutting promontories of bold rock- work 

 on the inner slopes, wdth subsidiary valleys running up between them, 

 and one of these subsidiary valleys would be a good place, I imagine, to 

 try one of Mr. Farrer's moraines, in which he grows the more difficult 

 alpines with such success. It is a common fault of rock gardens, I 

 think, that their slopes are too steep. It is true that a steep slope 

 gives sharp drainage, but that, where necessary, may be provided much 

 better by broken rocks under the soil. A valley with steep inner slopes 

 is apt to suffer from lack of air, and the steeper the slopes the more 

 rocks there must be and the less room for plants. The best plan, I 

 think, is to make the outer slopes steep and to cover them with easier 

 plants, reserving tbe inner slopes for the more difficult. 



The valley form of rock garden cannot be sharply distinguished from 

 the hollow form, for a valley is only an elongated hollow. Which you 

 ^vill have depends mainly upon the shape of your ground. If you have 

 a square, flat piece of ground, form a hollow wdth slopes partly below the 

 surface of the ground and partly above it. How far down you will 

 dig must depend upon the nature of your natural drainage, unless you 

 provide elaborate artificial drainage. I have myself no experience of 

 badly drained sites. My own is only too well drained. 



Much of what I have said about the valley applies to the hollow. Its 

 inner slopes should be gradual and its outer more steep. Or, if you 

 are cramped for space, make the southerly slopes gradual and the 

 northerly more steep. In any case vary the contour with promontories 

 of bold rock-work, and let there be little valleys leading out of the 

 hollow between lengthened mounds, so that the rock garden does not 

 begin and end too suddenly. The problem of paths is more difficult 

 in a hollow than in a valley. The main path necessarily runs along 

 the bottom of a valley. In a hollow it can be placed where you choose. 

 If it runs straight across, it looks very artificial. It will be most useful 

 and least ugly if it is made as unobtrusive as possible, winding about 

 here between bold rocks and there between flat spaces, and being paved 

 with rough flat rocks which will both mark its course and seem to give a 

 reason for its existence. But this pavement must not, of course, suggest 

 flagstones. The rocks should not be fitted too close or regularly 

 together. Low plants, such as the Stonecrops, or Cainpanula 

 caespitosa or Thymus Serpyllum, should break in between them here 

 and there, but not so frequently or thickly as to make the pavement 



