another precaution to be always carefully taken in the management of circular buildings, and this refers to

the unpleasant effect of overhanging segmental architraves or soffits. A is the elevation of a window B in a bow projection. B is its perspective appearance from one side. This is not less objectionable in respect to its constructive weakness than in regard to its ugliness; for it is only by concealed management that a flat soffit arch on a curved plan can be made to stand at all. The case is still worse when the window-head is a curve, and, in short, this practice is only allowable when the curve of the plan is so large, and the openings so narrow, as not to leave perceptible the defect of the overhanging segment. Thus, in the vast curved outline of the Coliseum the arched colonnade is unobjectionable. In the closely set peristyles of St. Paul's dome, and of the Temples of Vesta at Rome and Tivoli, it is equally so; but where the curve of the plan is small, and the openings, or spaces between the columns, proportionally large, it is a grievous fault. Where small bay projections are desired, they should always be semi-hexagons or semi-octagons, with the windows in the flat faces, unless indeed the required bow window may be so subdivided by mullions or pilasters as to remedy the objections stated. The semicircular portico, fig. 1, may be sufficiently pleasing in its front view; but a glance at fig. 2 will

Fig. 1.

Fig. 2.

[* For similar reasons, circular corners to buildings of small radius are very objectionable, and arches constructed in them defective in stability. Such corners should be canted.]

show the necessity of studying, not fronts only, but profiles also.

A

Triangular Plans.-While on the subject of the different appearance of objects in different points of view, it may be as well to refer to the triangle as a form of plan frequently, and most injudiciously, adopted in pyramids, obelisks, and pedestals. Viewed directly in front, on the lines ab or c-d, it is well enough, as shown by figs. 1 and 2; but who that sees its appearance on the line e-a, as shown by fig. 3, does not at once observe that no pyramid or obelisk should ever have an odd number of sides?

Plan.

Fig. 1.

For the same

reason,

All this goes to prove the necessity-if not absolutely of models of the perspective effects which buildings

will have from all points of view. The architect, pleased with his front elevation, may find reason to alter it the moment he turns the corner.

SOLIDS AND VOIDS.-A different proportion in the solids and voids of a main and return elevation will be fatal to good effect. A material difference in the distances between the common angle and the windows of the front and return façades, or a much more crowded position of windows in the one than the other, will be offensive; and it may be here remarked, that the proportions of solid and void which hold good in the case of a simple façade with no dressings to its doors and windows, will not equally serve when those dressings are to be supplied, since it is only the plain part of the pier, or of the space between the lower and upper apertures, which will "tell" in the matter of breadth.

WINDOW DRESSINGS.-The architraves and flanking columns of a window must be regarded as the window itself; and as a general rule, it may be said that there cannot be an adequate expression of breadth, unless the plain part of the pier be equal to the entire width of the window and its dressings united. The same law holds good in the horizontal spaces, which should exhibit, in a large and ornate building, the same amount of plain masonry above and below the architraves, strings, cornices, &c., which, in a small and plain building, would intervene between the sills and soffits of the windows and the strings or cornice below or above them. This, it may be remarked, amounts to little less than saying that architectural decoration is more applicable to large than small buildings, and it is true; for doors and windows do not increase in the same ratio that the size of the building increases. On the contrary, they generally bear a much greater relative proportion in small houses

than in princes' palaces; and assuredly, where they do not leave, at the very least, such a breadth of pier as will allow the width of the opening to intervene between the dressings of two adjacent windows, the latter had better be left without dressings. On no occasion whatever ought the breadth of a pier to be less than the width of a window opening; i.e. of course supposing the window to be a single one, and not triple with intervening mullions.

JUNCTION OF MAIN HOUSE AND OFFICES.--If the error of making a discordant difference between the fronts of the main house be so serious, not less so is the total discordancy often seen between the main house and the offices. Now it is, in fact, very rarely that the offices are not, from several important points of view, seen in conjunction with the principal mass of the structure; and the difference therefore between the two should strictly be one of degree only. A handsome cornice along the eaves of the one will be illaccompanied by a common eaves gutter along those of the other. Correctly speaking, it should be of the same form, reduced proportionally in scale, and-if required-without the enrichments of the main cornice. Above all things, the young architect should avoid the common mistake of reducing the beauty of the chimneys; for those of an office range, springing usually from a lower roof and having a greater relative altitude, will very likely be more conspicuous than the others. In short, an aptitude for chimney design is most important to an architect engaged in villa building. Let not the anticipation of chimney-pots escape his consideration. On the contrary, let him design them, and show them in his elevations, as likelihoods, which, if ultimately necessary, may not be absolutely disfiguring. He will further remember, that, where

the flues in any one stack are numerous, it may be better to place them in united parallels than in one continuous range; and he will be also cautious in so arranging his fire-places as that the various flue-stacks may be as nearly as possible of one size. This uniformity, at all events, should be observed in corresponding pairs of stacks.

CHIMNEYS. He need not be reminded, that, in Gothic structures, chimneys are not only admissible, but are often advantageous in their attachment to outer walls-especially when they rise with the gables. In the free and irregular style of the Italian villa they may also occasionally be connected with the outer walls. In the severer Roman style they may rise from the angles (as shown in Barry's Reform Club House); but in no style (saving only the Gothic) should they rise from the eaves if it can possibly be avoided.* The inordinate height required to raise them above the ridge of the roof, their insecurity (involving often the application of iron struts to sustain them), the difficulty of a satisfactory management of the main cornice beneath them, and the plumbing required to make weather-tight their union with the slates ;-all these circumstances make it most desirable the plans should be so arranged that the chimneys may ride, as it were, upon the ridges of the roof.

Flues.-The occasional practice of making flues run a long raking course in the thickness of walls, and of

These observations and restrictions as regards chimney-stacks are correct as far as they agree with the conventional rules of taste laid down by architects, though the young practitioner is advised to exercise his own common sense and discretion in designing such important functions of a building untrammelled by the mere pedantry of academic rule. Stacks should be collected as much as possible, and placed in positions whence they can rise uninterruptedly, without sudden bends, through the roof. Chimneys attached to walls, and projecting therefrom, often relieve a blank front.]