length of the days, the general clearness of the sky, the wants and customs of commerce and of life generally. In hot climates the windows are always few in number and small in dimension. As we approach those regions where the sun has less power and the winter is longer, we observe always an increase in their size and number, so as to enable the inhabitants to take as much advantage as possible of the sun's light and rays. It seems, therefore, almost impossible to give general rules on this subject. We shall on this account endeavour, in the rules that this section contains, to confine ourselves to the sizes which seem suitable in this climate, as respects the proportion of light necessary for the comfort of an apartment.

2747. It is a matter of experience that the greatest quantity of light is obtained for an apartment when lighted by an horizontal aperture in the ceiling. Of this a very extraordinary verification is to be found in the Pantheon at Rome. This edifice, whose clear internal diameter is 142 feet 6 inches, not including the recesses behind the columns, is nearly 74 feet high to the springing of the 'dome, which is semicircular. The total clear number of cubic feet in it may therefore be taken in round numbers at 1,934,460 cubic feet. Those who have visited it well know that it is most sufficiently and pleasingly lighted, and this is effected by an aperture (the eye, as it is technically called,) in the crown of the dome, which aperture is only 27 feet in diameter. Now the area of a circle 27 feet in diameter being rather more than 572 feet, it follows that each superficial foot of the area lights the astonishing quantity of nearly 3380 cubic feet. Independent of all considerations of climate, this shows the amazing superiority of a light falling vertically, where it can be introduced. But in a majority of cases the apertures for light are introduced in vertical walls; and the consequence is, that a far greater area of them for the admission of light becomes necessary. In considering the question it must be premised that a large open space is supposed before the windows, and not the obstructed light which it is the lot of the inhabitants of closely-built streets to enjoy. Again, it is to be recollected that in the proportioning of windows it is the apartments on the principal floor that are to be considered, because their width in all the stories must be guided by them, the only variety admissible being in the height. In this country, where the gloom and even darkness of wet, cloudy, and foggy seasons so much prevails, it is better to err on the side of too much rather than too little light, and when it is superabundant to exclude it by means of shutters and blinds. We are not very friendly to the splaying of windows, because of the irregularity of the lines which follows the practice; but, it must be admitted, it often becomes necessary when the walls are thick, and in such cases a considerable splay on the inside increases the light in effect by a great diminution of shade. It is well, if possible, to have an odd number of windows in an apartment; nothing wherein contributes more to gloom than a pier in the centre.

2748. We do not think it necessary to advert to the rule of Palladio for the dimensions of windows given in the first book of his work, chap. 25. ; because, were it true for the climate of northern Italy, it would not be so for that of Great Britain; neither are we at all satisfied with that which in his practice Sir William Chambers says he adopted, and which is as follows, in his own words :--" I have generally added the depth and height” we suppose width "of the rooms on the principal floor together, and taken one eighth part thereof for the width of the window; a rule to which there are few objections: admitting somewhat more light than Palladio's, it is, I apprehend, fitter for our climate than his rule would be." This rule is empirical, as indeed is that on which we place most dependence, and to which we shall presently introduce the reader, being ourselves inclined to the belief that in the lighting a room there is a direct relation between the area of the aperture admitting the light and the quantity of cube space in the room. Indeed the law which we are about to give is one founded on the cubic contents of the apartment; and if the results bore a regular ratio to that quantity, the discussion would be at an end, for we should then have only to ascertain the cubic contents, and, knowing how much an area of light one foot square would illuminate, the division of one by the other would supply the superficies of windows to be provided. Our own notion on this subject is, that I foot superficial of light in a vertical wall, supposing the building free from obstruction by high objects in the neighbourhood, will in a square room be sufficient for 100 cube feet if placed centrally in such room. It will, however, immediately occur to the reader, that this rule cannot in many cases satisfy the requirements of an apartment as respects the quantity of light necessary for its proper illumination. The subject is beset with numerous difficulties, which to overcome requires the greatest skill. In the case of an apartment, long as compared with its width, it is well known to every practical architect that windows of the same collective area at either of the narrow ends of such apartment will light it much more effectively than if the same area of light were admitted on either of the long sides, and most especially so, if it should happen that on such long side there were a pier instead of a window in the centre of such side. In illustration of what we mean, let us refer the reader to the ball room at Windsor Castle, an apartment 90 feet long, 34 feet wide, and 33 feet high. This room is lighted from the northern narrower side by a window nearly occupying the

width, and is supplied by an abundance of light. But had the same quantity of light been admitted from either of the long sides of the room, so many masses of shadow would have been introduced through the interposition of piers, that its effect would have differed most widely from the cheerful and airy aspect it now presents. We have taken this as an example that more presently occurs to us, but the reader from his observation will have no difficulty in supplying instances in corroboration of our impressions on this subject.

But we shall now proceed to give, in the author's own words, the rules of which we have spoken. That author is Robert Morris, and the work quoted is Lectures on Architecture, consisting of Rules founded on Harmonick and Arithmetical Proportions in Building. London, 8vo. 1734. "There are rules, likewise, for proportioning of light according to the magnitude of the room by which any room may be illuminated, more or less, according to the uses of them, and at the same time preserve an external regularity; which, as it is on an uncommon basis, I shall explain to you as well as I conveniently can. Let the magnitude of the room be given, and one of those proportions I have proposed to be made use of or any other; multiply the length and breadth of the room together, and that product multiply by the height, and the square root of that sum will be the area or superficial content in feet, &c. of the light required."

-Breadth 16 ft.

Fig. 964.

Breadth 12 f--

Fig. 965.

2749. "Example. Suppose a room (fig. 964.), whose magnitude is the arithmetical proportion of 5, 4, and 3, and is 20 feet long, 16 feet broad, and 12 feet high, the cube or product of its length, breadth, and height multiplied together is 3840, the square root of which sum is 62 feet. If the height of the story is 12 feet as before mentioned, divide that 62 feet into three windows; each window will contain 20 feet 8 inches of superficial light, and those will be found to be 3 feet 2 inches broad, and 6 feet 5 inches high, which are windows of two diameters."

2750. "Let us now suppose another room on the same range whose height is 12 feet, as the preceding example is, and its proportion (fig. 965.) shall be the cube. The product of that cube is 1728, and its root is 41 feet 4 inches, or thereabouts: divide that 41 feet 4 inches in two parts for two windows, and each will be 20 feet 8 inches of superficial light, and those will be two diameters in height, and the magnitude the same as the preceding room.'

2751. "For example sake, I will only suppose one more room (fig. 966) upon the same range, and 12 feet in height,

whose proportion shall be the arithmetical of 3, 2, and 1; that is, its height being 12 feet, the breadth will be 24 and length 36, the product of those numbers multiplied together will be 10368, and its root 101 feet 8 inches, or thereabouts: divide this room into five windows, each window will have 20 feet 4 inches superficial light, and the magnitude will be near or equal to

Fig. 966.

the others, and if the proportion be 6, 4, and 3, and coved, the light is the same." 2752. "There is," says the author, rather perhaps simply, "but one objection to this rule to make it universal for all kinds of proportioned rooms on the same floor, and that is, the square root doth not always happen to be exact enough for to make them alike; but as the variation will be so small, it may be made use of; and if the area something exceeds the standard of the principal room, that room may be converted to a use which requires more than standard light, and the necessities of families sometimes require it. But, how. ever, the rule will serve for the purpose near enough for any practice."

2753. "If you extend the rule to larger rooms, the same methods will be preserved even if the height be continued through two stories, if the upper windows be made square,

and to have two tire" (tiers) " of windows. Let us suppose the room (fig. 967.) with two tire of windows in height, to be 50 feet long, 40 feet wide, and 30 feet high, the arithmetical proportion of 5, 4, and 3, the product of those numbers multiplied together will be 60000, the square root of which sum is 245 superfical feet; divide that sum for the tire " (tiers)" of windows into three parts, or take one third of it, and that makes the attic or square windows 81 feet 8 inches superficial light; divide this into 5 windows, and they are 4 feet and half an inch square, and the five lower windows, consisting of 163 feet 4 inches superficial light, being what remains out of the 245 feet, the root, each of these windows is 4 feet and half an inch by 8 feet 1 inch, or two diameters, which 245 feet, the whole sum of the square root of the room, will sufficiently illuminate the same."

2754. The extreme piers should not, if possible, be less than half the width of the principal piers. This cannot always be obtained, but a much less width causes great irregularity, and that more especially when one of such end piers falls opposite a chimney breast, besides causing a great mass of shadow on the other side of the chimney, which has a tendency towards making the room dark and gloomy.

2755. Windows in the same story should be similar. There may be an occasional deviation for a great central window, but such deviation must be used with much caution. Another practice, most properly reprobated by Chambers, is that of intermitting the architrave and frieze of an order in the intervals between the columns to make room for windows and their enrichments, as on the flanks of the Mansion House in the city of London; a practice from which Sir Christopher Wren was, unfortunately, not exempt, as may be noticed in St. Paul's Cathedral.

2756. What are called Venetian windows are occasionally allowable, when so ranged and introduced as not to interfere with the composition, a task often difficult to effect. They should not be much repeated, as in the front at Holkham, where they become actually disgusting. Though in the examples which follow there be two which are composed with semicircular-headed centres, we do not approve of the general use of examples designed on such principles, and would advise the student rather to study the composition of the Venetian window, when required, as in fig. 968., which we do not present as one of beauty, but rather of propriety, where the want of light to the apartment renders a Venetian window expedient. The method of making sashes, shutters, and the other accessories of windows has been described in a previous section; we therefore proceed to offer a few of the most celebrated examples of windows. It is not necessary, after the investigation relative to the voids and solids of doors, to pursue the inquiry into the relative proportions of windows as respects that part of the subject. They are, in a measure, in regard to windows, subject to the same principles, and this, by trial, will be immediately apparent to the student; and we therefore shall not stop for such investigation.

Fig. 968.

2757. Fig. 969. is after the lower story of windows at St. Peter's at Rome, by Michael Angelo, and is rather less than the double square in height. The architrave is one seventh

of the aperture's width, being the same as that of the pilasters. The length of the consoles is one third of the width of the aperture, and the entablature one quarter of its height. 2758. Fig. 970. is from the Mattei palace at Rome, and is the design of Bartolomeo Ammanati. It possesses, though rather heavy, considerable beauty, and well deserves the attention of the student. Chambers, from whom we have selected many of our examples in this and others sections, says, "the parts made somewhat less would succeed better, as would also a pediment instead of the sloped covering at top: " but we entirely disagree with him, and are of opinion that what he proposes would ruin the design.

2759. Figs. 971. and 972. are the compositions of Bernardo Buontalenti. The apertures are a double square, or something less, the architraves a sixth or seventh of the apertures, and the pilasters may be about the same. The height of the entablature should not be more than a quarter that of the aperture, nor much less. The greatest length of the consoles should not exceed half the width of the aperture, nor should their least length be less than one third of it.

2760. Fig. 973. is from the old Louvre at Paris, and is by the celebrated Pierre Lescot,

abbot of Clugny in the reigns of Francis I. and Henry II. Its proportions are not much dissimilar from the two last examples.

Fig. 973.

Fig. 974.

Fig. 975.

2761. Fig. 974. is a window constantly used by Palladio. The opening is a double square, the breadth of the architrave equal to one sixth of the aperture, and the frieze and cornice together equal to double the height of the architrave. The breadth of the consoles equal to two thirds the width of the architrave. The breaks over the consoles in the bed mouldings of the cornice are perhaps not strictly correct, but are deviations from propriety which may be tolerated. The breaks in the upper vertical parts of the architrave would perhaps be better omitted. The practice generally should be avoided, except in cases where a greater length of cornice is wanted for the purpose of filling the bare walls to which the windows are applied.

2762. Fig. 975. is from the Banqueting House at Whitehall, by Inigo Jones. The aperture is a double square, the entablature one fourth of its

height, and the architrave somewhat more than one sixth of its width.

2763. Fig. 976. is by Michael Angelo, and executed at the Farnese palace at Rome. It possesses all the wildness and fancy of the master, and though abounding with faults, is redeemed by its grandeur and originality.

2764. In fig. 977. is given the design by Ludovico da Cigoli of a window from the ground floor of the Renuccini palace in Florence. It can scarcely be properly estimated without its connection with the façade, to the character whereof it is in every respect suitable.

2765. Fig. 978. is a design of Palladio, nearly resembling that executed in the Barbarano palace at Vicenza. It has been imitated by Inigo Jones, and perhaps improved on by him, in the flanks at Greenwich Hospital.

Fig. 976.

Fig. 977.

2766. Fig. 979. is also by Palladio, and executed by him in the Porto palace at Vicenza. 2767. Fig. 980. is the design of Raffaelle Sanzio, and worthy of the reputation of that