the same general basis, the two ends of the large beds are cir cular. Either of these four figures, but especially the last two, would be of service where a sort of running border of flowers is desired. In respect of the beds with pointed ends, however, it may be suggested that the specimens placed between them should be of the dwarfest and most trailing habits. An additional step in the way of variety is made at fig. 68, PPPPI Fig. 68. where the ends of the beds are turned to the walk, and a more flowing outline occasioned. Very small shrubs are likewise inserted in the circular ends of each alternate bed. And if beds of this pattern cannot be much commended, on account of the trouble demanded to fill them nicely, and to keep them correctly cut out, they may yet be useful in making a species of scrolllike fringe to a walk, where only one description of plant (such as Verbenas of different colours) is intended to be grown in Fig. 69. them The next plan (fig. 69) is both simpler and more artistic; and the ends of the beds might be made square, if preferred, te adapt them more thoroughly to the line of the walk. Figs. 70, 71, and 72 belong to a more elaborate class, and may appropriately finish the series. The first of them could easily be varied by squaring or rounding the ends of those beds which Fig. 70. stand at a right angle with the walk. The shrubs shown in the circles must, of course, be of the smallest description, and should be upright-growing, if possible, to keep them from being injured by the summer flowers. Probably the dwarf Box, trained and kept regularly clipped into the shape of a thin cone, would be most serviceable. In fig. 71, a strict adherence to the Fig. 71. line of the walk is maintained in the flower-beds, and a similar conformity is observed in fig. 72, some little play of margin being accomplished by the interposition of dwarf shrubs at regular intervals. It will be obvious that specimens of this character might be multiplied to an almost infinite extent, if any sufficient object could be served by their introduction. In those already given, however, and which are intended rather as hints than as models, the reader will probably find enough of suggestive ness to render a further instalment unnecessary. B Fig. 72. Towards the outside of a formal garden, or in parts sufficiently separated from the house, or from the view obtained at its principal front, there will be little objection to the use of masses of plants, with a more irregular outline, or specimens scattered about in the mixed style; provided a kind of connexion be kept up by the help of circular or other regular beds in the centre or at the corners of such compartments, or, when irregular lines are adopted as a fringe round the boundary, they are made as inconspicuous as possible from the house, and do not thrust themselves into notice anywhere, or weaken the effect of the more artistic parts. They will not disturb the harmony of the place unless they are obtruded. One most important requirement in a formal garden is, that the ground should be quite smooth and level. No undulations or unevenness of surface can be for a moment allowed. Regular and easy slopes, or dead levels, are as essential as straight lines in a house or in the walks. A perfectly flat surface is unquestionably the best for the purpose, as the lines will appear longer. When a line slopes away from the point of view, it is, to some extent, foreshortened. If the ground should, by any unfortunate chance, rise as it recedes from the house, it may be kept flat to as great a width us is possible, and then be formed into one or more terrace banks, (fig. 73,) as it may require; the walks to be carried up these banks by flights of steps, and the change of level effected Fig. 73. by grassy slopes, or by low architectural walls. The latter of these is represented by fig. 74, and the former will be seen in fig. 73. Fig. 74. When, in addition to a slope from the house downwards, or apart from it, the ground also slants naturally in a cross direction, this will demand some modification. As far at least as either of the main fronts of the house is concerned, the ground, to the full breadth of those fronts, and of any additional terrace-bank by which they may be supported, must be brought into a perfectly level platform. There should be no cross slopes,-no oblique inclination of the ground in a direction parallel with the front of the house. The level basement-line of the house would in no way accord with a diagonal or sloping ine in the ground; the latter being sadly out of harmony with the squareness of the style. Indeed, the side of a house out of natural level of the ground, and the shaded line the level to which it should be reduced.) The change of level from this point, in a line taken precisely at a right angle from the house across the garden, should be effected, whether ground rises or descends, by a terrace-bank of grass, the upper edge of which is kept quite square; or by a low wall; carrying the walks up or down either by flights of steps. The steps resulting from any such alteration of levels, will, if rightly treated, and adorned with small vases, materially contribute to sustain the general character of th place; though they should never be without more or less massive edgings or kerbs of stone, or some living substitute for these in the way of low dense evergreen hedges. Any extreme slope of the ground away from the house can be converted into terraces, as suggested for rising ground. But many terraces on a descending slope ought not to be used unless really necessary; for they serve to lessen the apparent size of the place. Water, if admitted at all into the geometrical style, takes the shape of basins with an architectural rim, or fountains, or larger |