In

from it, the level is an abrupt termination, and a little rin narks it distinctly. To cover that rim, a short sweep at the bot of a swoll, a small rotundity at the entrance or a hollor, must be interposed. every instance, when ground changes its direction, there is a point where the change is effected, and that point should never appear; sone other shapes, uniting easily with both extrenes, must be thrown in to conceal it. Eut there must be no unizomity oven in those connections; if the same sweep be carried all round the botton of a swell, the sanc rotundity all round the top of a hoolow, though the function te poffect yet the art by which it is made is apparent, and art rust never appour. The manner of concealing the separation should itself be disguised; and different decreas of cavity or rotundity; different shapes and dimensions to the little parts thus distinguished by decrces; and those farts, breaking in one place ore, in another loss, into the principal fornis which are to be united; produce that variety with which all neture abounds and without which ground cannot be natural.

V. The relation of all the parts to the whole, when clearly marked, facilitates their junction with each other; for the common bond of union is then perceived, before there has been time to examine the subordinate connections; and if these shouli to doficient in rose niceties, the dofect is lost in the general pression. But any part which is at variance with the ratt, is not barely a blemish in itnolf; it spreads dicorder as far as its influence extends; and the confusion is in proportion es the other parts are more or los adaptod, to point out cry particular direction, or to mark any peculiar character in to ground.

If in round all descending one way, a picco is twisted another, the general fall is obstructed by it; but if all the parts incling in the sane direction, it is hardly credible how small a declivity will seem to be considerable. An appearance even of steepness ay be given to a very centle descent, by raising hillocks upon it, which lean to the point, whither all the rest are tending; for the eye measures from the top of the highest, to the bottom of the lowest ground; and when the relation of the parts is well preserved, such an effect from one is transfused over the whole.

But they should not, therefore, all lie exactly in the same direction: sone may seen to point to it directly, others to deling vory Lach, others but littlo, sono partially, some entirely. If the direction be strongly marked on a few principal parts, great liberties may to taten with the others, provided none of them are turned this contrary woj. general idea Lust, however, be procervod, clear even of a doubt. hillock which only intercepts the sight, if it docs not contribute to the principal effect, is, at the best, an unnecessary excrosus (cu; and even an interruption in the coneral tendency, though it hide nothing, is a blemish. On a descent, any hollow, any fall, which has not an o.tict to lower ground, is a hole; the eye skips over it. instead of being continued along it; it is a cap in the composition.

There may indeed be occasions, when we should rather with to chock than to promote, the cen rel tendency. Cround any proceed too hastily towards its point; and we have equal power to retard, or to accelerate, o fell. We can slacken the precipitancy of a steep, by breaking it into parts, some of which shall incline less, than the whole before inclined, to the principal direction; and, by turning then quite away, we may even change