230 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 



The first portion to be built was the main, or top waterfall. I may 

 say that I had hoped the planting generally would have been more 

 forward ere this, as, even now, the rocks are very bare. There is a 

 fine lot of Semperviviim in the crevice between two stones here (fig. 74), 

 which stones, by the way, have been so placed as to appear to have 

 slipped from the mother rock, and form what is known as a geological 

 "fault." In a bed of soil here is- Primula deflexa, very much at 

 home, the bank behind being damp and fairly shady. Primula 

 frondosa (Hort.) also grows well in a similar position; this was only 

 planted early in the spring, and flowered well a few weeks later. 



The bank adjacent to the waterfall was steep, consequently the 

 rocks had to be constructed to obtain a cliff or wall-like effect. At 

 the foot of this cliff it was more or less damp, so opportunity has 

 been taken to introduce various Primulas, including Primula jap onica, 

 P. rosea, P. denticulata, and Haberlea rJiodopensis. Fig. 75 shows 

 a distant view of this fall, the cliff referred to, and one or two 

 nearer falls, while fig. 76 is a view taken from a little lower down, 

 nine months later, where the cliff shows plainly. The distant effect of 

 this portion should be particularly bold and telling later on when the 

 rocks are furnished, and the massive cliff is surmounted by Ehododen- 

 drons, with a background of Pines, whose dark colour throws into 

 relief the brighter foliage and the grey of the rocks in the foreground. 



Other portions were, of course, also being constructed (figs. 77 and 

 78) at this time. These two figures are interesting from the fact that 

 they show the actual formation of the rock. They also give a very good 

 idea of the slope of the ground. 



The stream was followed down from the top, and is concreted so as 

 to appear quite natural from end to end. Every section, or length 

 between each fall, is provided with a proper valve, in order that it can 

 be emptied separately and cleaned out, whenever occasion may arise. 



Did someone start at the mention of the word " concrete " just 

 now? I am well aware that there is a more or less acute difference of 

 opinion on this subject, many advocating puddle in preference to con- 

 crete, because, in their opinion, the latter looks hard, and the edges 

 have a way of intruding themselves, making the stream or pool look 

 artificial — a most undesirable result. But to those who hold this view, 

 I may say that, for efficiency, there is no comparison between the two, 

 and, if concreting is properly undertaken by those experienced in its 

 uses and abuses, it is infinitely more effective, and better in every way, 

 especially for cleaning-out purposes, as small streams soon get filled 

 with leaves, weeds, and rubbish, particularly in the autumn, and it is 

 difficult to clean them out, when puddled, w^ithout the risk of damaging 

 the bottom or sides and losing the water. It is also urged by the 

 advocates of puddle that a pond or stream holds water better if so 

 treated, for does not concrete crack and cause leakage? I must admJt 

 that such is occasionally the case, but it is generally only from a lack 

 of knowledge of the best materials and the way to use them, which 

 knowledge only years of careful study and constant experience can 



