ditriglvph, and were necessary from their situation, being opposite to the principal entrances. Indeed, front the massive and bold character of the Grecian Doric, the monotriglyphic succeeded best ; but in the Roman Doric: it would not be convenient, because the passage between the columns would be too narrow, especially in small buildings ; and therefore the ditriglyphic is to be prefer red.
When the solid parts of the masonry of a range of arcades are decorated with the orders, the intercolumns necessarily become wide, and the intercolumniation is re gulated by the breadth of the arcades and of the piers. Vignola uses the same intercolumniation in all his or ders ; and though this practice is condemned by some, it is founded upon a right principle, as it preserves a constant ratio between the columns and the intervals.
Coupled, grouped, or clustered CO111111115, seen, not to have been used by the ancients. In the temple of Bac chus, at Rome, indeed, we find columns standing in some sort in pairs ; but it must he observed, that each pair is only placed in the thickness of the wall, and not in the front ; wherefore they may be called two rows of dis tinct columns, one almost immediately behind the other, rather than a coupled row. In the baths of Dioclesian, and in the temple of Peace at Rome, we find grained ceilings sustained by single Corinthian columns, which at once present a meagre appearance, and furnish but an inadequate support.
The moderns seldom employ more than one row of columns in either external or internal colonnades ; for in double rows, the back range destroys the perspective regularity of that of the front; and the rays of light, proceeding from both ranges, produce confusion. Pi lasters placed behind a row of insulated columns are lia ble to the same objection, except that the relief is strong er, owing to the rotundity of the one being contrasted with the flat surface of the other.
In buildings upon a small scale, as temples and other ornamental erections in gardens, the intercolumniations, cr at least the central one, must be broader than the po sitive dimensions of the pillars would admit.
offlic.—Though the earlier architects of Greece were either unacquainted with the use of pilasters, or refused to introduce them in their designs, they frequently pla ced a kind of square pillars at the ends of their walls, which they called ant,;, and which sometimes projected to a considerable distance from the principal front, form ing the pronaos, or vestibulum. The breadth of these
alum w as always much less on the flanks of temples, than on the front ; and sometimes they had columns between them, in which case, the return within the pronaos was of equal breadth with the front. The capitals of the ants never correspond with those of the columns, though they always retain some characteristic marks, by which the order may be distinguished.
Pilasters are rectangular prismatic projections, advan cing from the surface of the wall, furnished with bases and capitals similar to those of the columns whose enta blatures they assist in supporting. They differ from co lumns, in having their horizontal sections of a rectangu lar figure, whilst the sections of columns are either com plete circles, or sections of circles, equal to, or greater than semicircles. They are probably of Roman origin ; there being but few Grecian buildings, and those of the latter ages, (except the monument of Thrasyllus,) in which they arc repeated at regular intervals, as in the monument of Philopapus ; but of their application in Roman works we have numberless instances. Vitruviw.: calls them narastata.
When ranged with columns under the same entabla ture, or behind a row of columns, they have, as we have already observed, their bases and capitals like those of the columns, with the corresponding parts at the same heights : when placed at the angles of buildings, the breadth of the return is equal to that of the front. The trunks, also, have frequently the same diminution as the shafts of columns, as in the arches of Septimius Severus and of Constantine, the frontispiece of Nero, and the temple of Mars Ultor, at Rome. In these cases the top of the trunk is equal to the soffit of the architrave, the upright face of which rests on the capital, in the same perpendicular with the top of the pilaster. When the pilaster is undiminished, and of the same breadth as the bottoms of the columns, the face of the architrave rest ing on the capital, retreats within the head of the trunk, as in the pantheon of Agrippa.