Vitruvius, in all of which the side walls presented no difficulty to the use of vertical windows, if the architects chose to introduce them. That they did so in some cases, especially in the Ionic order, we know from extant examples, for the details of which we must be content to refer to Mr. Fergusson's pages, as well as for the exceptional cases in which the smallness of the temple made the light from the entrance sufficient, sometimes aided by windows on each side of the door.

Few persons, who have given the least thought to the teaching of their own experience, will question the vast superiority of upright windows to skylights (even with all the help of glass), both for the ordinary uses of light, and much more for its effect in illuminating the statues and other works of art, which glorified the Greek temples.* But the latter consideration at once suggests the superiority of light admitted from above, to side windows in the main walls of the cell; and thus the question of lighting becomes involved in that of the construction of the roof. Now it is an extraordinary fact, that on neither of these essential points of construction does Vitruvius give us any information (putting aside for the moment, the solitary passage on which the hypethral theory is based); and we have only a few incidental allusions to the roofs of temples in other writers. But these are enough to prove that the usual form was the ordinary ridge-roof, which is plainly denoted by the triangular pediment surmounting the portico-an arrangement which at once suggests an absurdity, if the roof thus indicated covered the portico only; and, further, that the roof was constructed of timber, covered with tilest (many of which have been found among the ruins), or plates of stone or marble, and in some cases of gilt bronze. Among the many testimonies to the general use of wooden roofs, two are of special interest for our present enquiry. Pliny cites as an example of the durability of certain woods the timber of the temple of Diana at Ephesus, which had lasted 400 years (B.C. 330 to A.D. 70), and especially the cedar planking of its roof; and Strabo, expressly referring to the fact that the roof was the part destroyed in its successive conflagrations, asks the question which bears closely on the hypethral theory, after the conflagration, and when the roof was destroyed, who would have wished to have a deposit lying there, with the sacred enclosure open to the sky' (èv vпailpw τự mx)?—its normal state on the common view!

The obvious difficulties arising from the want of window-glass could be met by an arrangement of grilles, supporting blinds or curtains.

Many interesting particulars about these roofing-tiles will be found in the Dictionary of Greek and Roman Antiquities,' art. TEGULA,

The details of such a roof, as worked out by Mr. Fergusson's artistic and constructive skill, are too technical to follow here; but one point is essential for our discussion. Most expositors have been content to regard the triglyphs in the Doric frieze as the ends of solid joists, carved with a decorative pattern. But Mr. Fergusson refuses to charge the Greek architect with such a needless waste and weight of solid timber, and with convincing ingenuity he shows how much better the whole roof is constructed, if we regard its sloping supports as made of three planks bolted together by wooden trenails; the ends of these planks forming the triglyph, and the mutules on the under surface of the frieze (which alwavs preserve the same slope as the roof) representing the trenails that fastened them to the wall-plate. Be this as it may, it is evident that either kind of timber construction would leave those open spaces between the triglyphs, which retained in the stone-work the significant name of opa (holes) ;* and these would serve as windows for the admission of light in the three simpler forms of temple thus far described. For the exact mode in which this may have been managed, we must be content to refer to Mr. Fergusson's work, while we pass on to his new solution of the more interesting and difficult problem raised by the peristylar class of temples.†

When a more magnificent form was given to the temple by widening the portico from four columns to six or eight, and carrying a colonnade of one or two rows of pillars round the sides of the cell, with the roof of course projecting over the whole, it is evident that the spaces between the triglyphs could no longer serve for the admission of light into the cell, and in fact they were always filled in with those sculptured slabs, called metopes, the examples of which from the Parthenon adorn our Elgin Gallery. We have therefore an entirely new form of the problem, to discover by what construction of the roof light was admitted into the sanctuary; for it was certainly through the roof, unless we are content with the utterly improbable hypothesis, that the large temples were lighted only through the door. Here Mr. Fergusson first takes his stand on the principle of common sense, that, whereas it appears difficult for British architects to consider their Grecian brethren as other than incompetent bunglers,' we ought to give these

* There is a very common and not unnatural confusion of language between these openings (ope) and the metopes (metopa i. e. 'coverings of the holes') with which they were filled in the perfected form of the temple.

+ We follow Mr. Fergusson in using this general term to include the peripteral, the dipteral, and the pseudo-dipteral of Vitruvius, for all which the problem of lighting is identical.

But to

masters of art full credit for mere mechanical skill. discover how they used their skill, their extant works must be thoughtfully investigated, and not merely copied with a complete disregard of their original use and meaning.

'It is, indeed, this false system of art that lies at the root of all our perplexities on this question. If, instead of puzzling themselves with obscure or corrupt texts and false analogies, architects had set to work to discover. from existing remains, how Greek temples could best be lighted, the question would long ago have been solved. The Greeks were neither foo's nor savages, but on the contrary the cleverest architects we know of and we have every reason to believe that the interiors of the temples were as perfect as we know the exteriors to have been. To contend, therefore, that they alone, of all people in the worl, could not put a weather-tight roof on the temples, while admitting the requisite quantity of light for their illumination, seems one of the most monstrous propositions that ever was put forward. There are many ways in which the end might be accomplished, without much taxing their ingenuity. One of the most obvious was to introduce a range of openings bigh up the cella walls under the peristyles. Windows So situated would have been perfectly protected from the weather in all circumstances, and the light introduced so situated as, according to our ideas, to meet all the artistic exigencies of the case. If it was not adopted-as we know it never was-it must have been that the Greek architects knew of some better expedient, which was mechanically as perfect, and artistically was better, and this they adopted in preference to what appears to us the most obviously practical mode of introducing light-what that was, it is the object of this treatise to explain,'

To give the result of Mr. Fergusson's most elaborate and ingenious discussion in one word, the method of lighting for which he contends was by a clerestory to borrow the name from the Gothic church, to the internal arrangement of which that of the Greek temple presents, on this theory, a very remarkable analogy. The attempt to do justice to Mr. Fergusson's argument in detail would require us to reproduce most of the substance of his moderate-sized quarto, without the aid of his numerous and elaborate illustration. It must therefore suffice, in making the theory clear to our readers, to indicate the main supports on which it rests. In all the great Grecian Doric temples, with very few exceptions, we find internal rows of columns, the use of which is generally unintelligible and in some cases quite inconsistent with sound principles of construction, if regarded only as supports to the roof, which did not need their aid, while they only encumber the internal area. But their introduction is at once explained, if they upheld an inner wall beneath the roof, pierced with openings for the

admission of light. The interior colonnade might be of two stories, or of one, especially by the use of the taller proportions of the Ionic pillar; but that which may be regarded as the normal arrangement, as exemplified in the Parthenon, appears to be such as the following. Entering at the eastern door-for, strange to say, all the extant temples in Greece Proper contradict the assertion of Vitruvius, that the principal front was to the west-there was on each side a range of Doric columns,-the places of some of them being still visible on the floor,-surmounted by a second range of the same order on a smaller scale, which supported the internal wall pierced with windows, probably with pilasters over the columns. The middle story formed an internal gallery (úñeρm̃mv) -like the Gothic triforium-the roof of which served at the same time as the floor of an external gallery under the main roof of the edifice and between the clerestory wall, on the one side, and the back surface of the frieze of the cella, on the other. This external gallery is called by Mr. Fergusson the opaion (onatov), as being the essential arrangement for the admission of light through the windows (onat), by means of corresponding holes in the roof above. Such holes might be partly defended from the rain flowing down the roof by means of tiles provided with a fender-edge (imbrex); and Mr. Fergusson gives a drawing of such a tile, found by Mr. Cockerell at Bassæ, in its place in a restored view of the roof. Any rain or snow passing direct through these perforated tiles would fall on the floor of the opaion, that is, the roof of the internal gallery, and none would enter the temple. The drainage of the opaion might be effected in various ways, which Mr. Fergusson discusses. He finds the origin of this system of lighting, as of the Doric order itself, in Egypt, where, as he holds, the central avenue of the great hypostyle hall of Karnak was provided with a clerestory, from which the light was thrown into the colonnades on each side, these latter having flat roofs, suited to the rainless climate of Egypt. He very ingeniously adduces as a connecting link the temple of Demeter at Eleusis, with its triple aisles on each side of the central nave, its ground-plan forming a square equal to just half, of the great double square of the hall at Karnak. A form so remarkably different from the prevalent plan of the Greek temples is the more likely to have been derived from Egypt, considering the connection which Herodotus (ii. 59) recognizes between Isis and Demeter, and the belief of Lactantius and others, that the Eleusinian mysteries were closely allied to those of Isis.

For further particulars of the theory, and a full discussion of

the detailed arrangements and difficulties, of which he evades none, we must be content to refer to Mr. Fergusson's work. The system has, at all events, the one merit of bringing back the method of lighting Greek temples within the domain of common sense, and of presenting an harmonious and admirable result, which has stood the test of experiment in a model of the Parthenon, which Mr. Fergusson has constructed on a large scale.* Taking as a test the example of the Parthenon-not only the crowning glory of Greek art, but artistically the most perfect edifice raised by the hands of man, he asks us to go with him in quest of interior arrangements worthy of the sanctuary itself and of the chryselephantine statue there enshrined as the climax of Greek sculpture.

These principles furnish a test of that other theory-the hypäthral, or open skylight-which has been generally accepted during the last half-century, but which we have purposely reserved till last, from the conviction that it has no fair 'prerogative' claim to notice. When, in 1815, Quatremère de Quincy undertook his celebrated restoration of Phidias's chryselephantine statue of Jove at Olympia, he cited the general agreement of preceding writers, that the Greek temples were lighted only by the door; but he himself adopted a suggestion already thrown out by Stuart in a vague form, that light was admitted by an opening in the roof, which he figured as semicircular in order to give headroom for the statue. On this theory Mr. Fergusson observes that to cut a square opening in a waggon-vault, to remove the keystone in fact-is an architectural solecism which, we may feel sure, the sense of architectural propriety in a Greek would never have tolerated in any vault, either in stone or wood. Besides this, the wandering light from a naked skylight would have been most inartistic and disagreeable; not only because of the glaring sun which must have shone on the worshippers at mid-day, but of the rain and snow against which, on this system, it was impossible to provide any protection. Everything, in fact, combined to render this mode of lighting the temple most objectionable, while the gain in height was insignificant. Throughout the middle part of this century the theory was warınly debated by

* Apart from this model, any reader may see the system of lighting practically exhibited in the building constructed by Mr. Fergusson, in Kew Gardens, for Miss North's most wonderful drawings of tropical plants-a monument of the liberality of both. Of course the lights are here in direct contact with the open air; but in other respects they give a very pleasing reproduction of the opaion, with its pilasters and other decorations; and the ground glass may fairly represent the curtains which tempered the light.