half and three) equalizes an upward slope | of the ground at the south-western corner; the ground is quite level on the western side. Deflection produced by the cutting of masonry of course cannot be referred to sinking. Fig. I. presents certain additional facts which preclude any resort to this theory, the cathedral being surrounded by a carefully graded raised pavement, sloping downward from the walls to the five steps by which it is ascended, and the plinth supporting the walls nowhere sinks below it nor breaks its line. Mr. Ruskin, as already quoted, has remarked that the cornice (marked 2 in Fig. III. See also Fig. II.) suddenly leaves the tops of the four westernmost arches but is evidently not aware (since he questions whether these four arches fall or the cornice rises) that the cornice rises throughout its whole length at nearly the same angle and that the remaining eleven arches rise with it. In the wooden model of the cathedral kept in one of the galleries, its oblique direction is accurately copied. How is it that so marked an obliquity should be unnoticed by an eye engaged in noting ir

VOL. VIII-28

regularities and with attention directed toward the cornice?

Obliquity of horizontal line is a perspective effect to which the eye is accustomed from every standpoint except the single one in which it is equally distant from two extremities of a line, and is therefore easily mistaken for it. The above cut is copied from the ordinary photographic view of the Cathedral. It is evident, for instance, that from the standpoint here taken, the heavy cornices falling toward the junction of transept and main building, appear to fall from perspective effect. The Cathedral simply appears larger than it is.

There may be an additional reason why this obliquity is not generally noticed. It has been shown (figures II. and III.) that the distortion in the stripes of the Cathedral is connected with the inclination of the façade, and this distortion is doubtless a device to deceive the eye into believing the façade to be upright by causing the stripes to enter its corner pillar at a right angle. But is it not possible that the distortion has an effect beyond this? viz. :

that of making the cornice appear horizontal by leading the eye to believe that its divergence from the line of the stripes is owing to the palpable bend downwards of the latter. Should this be so, the divergence of the line of the arches from the line of the cornice would be also explained; it would be an attempt to make the eye believe that the cornice does not rise, but that the arches fall-an attempt we must call at least a partial success in the case of Mr. Ruskin, for he says, as already quoted,."the cornice rises (or seems to rise), leaving at any rate, whether by the rise of the one or the fall of the other, an interval of more than two feet between it and the top of the western arch." Doubtless this divergence and that of the stripes assist in preventing the eye from perceiving the perspective deception.

The story above the cornice repeats nearly, though not quite, its divergence from the horizontal. With regard to the roof line of the main building it is not quite so easy to determine, but of the transepts it is certain that the roof lines follow the obliquity of the cornice, and that the obliquity of cornice in both transept and main building is owing to the arrangement of

blocks are cut and fitted at the necessary angles. (See figure IV., 2.) It is evident that this change in the direction of the southern wall, amounting to twentyeight centimeters, gives an effect of greater width to the whole church to any one standing within and near the entrance door, for the standard of the nearer part is always taken by the eye as a standard for the whole. From any point of view outside, south of this portion of the southern wall, the wall is nearer than the eye believes it, appears larger, and gives a false standard for more distant parts.

Is there not an architectural illusion produced by these irregularities aside from the apparent increase of material size, and, if so, was this fact recognized on principle by architects in Pisa? That variety in corresponding parts, and deviations from strict symmetry are the soul and the life of all decorative art and all architecture, is a fact which no æsthetically trained person pretends to ignore; as little will any such one deny that herein lies the superiority of the Greek temple and the medieval church over all modern copying and supposed imitation. I have, however, evidence to show that not only was this law of "life" a recognized principle of building in Pisa to an extent undreamed of even by its most ardent advocate,-Mr. Ruskin in the "Seven Lamps," but that a still more subtle effect was habitually studied. Say, for instance, that an observer stands opposite the oblique cornice of the Pisa Cathedral at the point where it appears to be horizontal. With an ordinary building, the point where its horizontal lines are not subject to perspective obliquity is exactly opposite the center. As there is an actual obliquity in the line of the cornice in question, it will not appear horizontal unless the observer stands at some distance to the right of the real central point (in looking at the southern wall), and here there is a marked inequality in what, from the horizontal cornice line, the eye supposes to be two equal halves of the building. I have assumed that the spectator stands at the point where the oblique cornice appears horizontal, in order to illustrate plainly the optical contradictions between. apparent fact and actual appearance-but these must exist from any point of view in which the cathedral is seen. Is there not in a building so constructed an architectural illusion springing from this mystification of the eye.

Mystification of the eye is the secret of all "life" in art, is the secret of the charm which the variations of the Greek scrollwork, of the arabesque pattern, of the medieval detail, from dead symmetry have for all of us.

Here the whole building is made to vibrate in an architectural illusion of the same character. On me, at least, the Pisa Cathedral, as seen from the Lucca railroad, made the effect of a ship under sail.

Let us take again the inclined façade. Why should so delicate and expensive a device be employed, if it is not for the life breathed into this wonderful creation by reason of the fact that the eye, assuming the façade to be upright, is mystified in making the unconscious rectification required by its inclination. I query if there is not a continual wavering of the eye between the innumerable possible lines of rectification lying between base and summit, resulting in a suggestion of mirage.

With regard to the curve described I would also put the inquiry if the continual variation in the lines of the building which the curve produces, does not in like manner mystify the eye, and thereby again contribute to the optical effect?

Thus, having first established the facts. of an intentional inclination, an intentional obliquity, and an intentional curve in the Pisa Cathedral, I have raised the question if their optical effect is not something more than the pure and simple apparent increase of material size undeniably present in the two latter irregularities. I propose to debate this point by offering some cumulative evidence as to other intentional irregularities and leaving it to the reader to determine if they can have any other purpose.

The very perceptible inclination of the

Pisa Baptistery is not generally remarked by travelers. On figure V., 1, may be found a series of measurements that go far to show that this inclination was intended. It must be borne in mind that the polygonal surrounding pavement rises from its edge. toward the circular wall of the Baptistery, being as carefully joined at the edges of its sections as a piece of veneered furniture. On the face of things, therefore, the preservation of this delicate upward sloping grade on all sides is good presumptive proof against sinking. But, still farther than this, the foundation layers above the surface of this pavement are cut in gradually lessening height in the direction of the greatest inclination. These foundation layers (fig. V., 4) of masonry are three in number, arranged in step fashion; above them is a string-course cut in so rich a profile that wedging (i. e., cutting the block in converging lines) would be impossible, and above this is the wall proper.

The measurements of the three foundation layers at each particular point are given as added together. Observation will reveal the fact that the Baptistery has a very perceptible inclination in the direction where these measurements would lead us to expect it.*

May not this leaning of a round building have an optical effect? The eye not especially directed to the fact of the lean

*Figure V4 gives a view of the foundation layers, figure V3 of the delicate provisions for drainage, the want of which is the general cause of sinking foundations, the gutter below each polygonal side of the pavement curves upward toward the center; (C, c-12 centimeters. B, b=151⁄2 centimeters). In fig. V. 1, the marks on the circle show the position of the pillars; the position of each measurement denotes approximately the spot where it was taken. The measurements are at unequal distances, and must be tested at exactly the same

positions for the following reason: The rising slope of the polygonal pavement makes the lines at the junction of the segments rise higher because they rise farther. (Compare, in drawing V., cut 3, the lines a, B and e, F). At each segment line, therefore, the first foundation layer is cut to equalize this variation. (A, a-16 centimeters, and E, e-18 centimeters.) The measurements were, therefore, taken invariably at the center of each segment, and were omitted where this falls at the base of a pilaster, and opposite the doors, because in these places the line of the foundation layers is broken into.

ing will insensibly partially or wholly rectify it, and the result of seeing the building in a position it does not occupy may produce an effect of optical vibration similar to that I have suggested as existing in the case of the façade, the cornice and the curve. It would appear that all those who have looked at the Pisa Baptistery without perceiving its pronounced inclination must have made this unconscious rectification. (See fig. VI.)

We will now return to the ground plan of the Cathedral, in order to remark the fact indicated in figure IV., that the wall of the main building, on the southern side, meets the transept wall at an acute angle.

The corresponding angle in the surrounding pavement (41⁄2 centimeters in 56)

is more than twice as acute as the angle of the walls (2 centimeters in 56). The neat cutting and fitting of the masonry at both points is proof of an intention. The question rises again here, was the small increase in apparent length of the building hereby gained the only motive, or was a still more subtle optical effect designed? The divergence of the line of the pavement from the line of the building, which the difference in the angle produces, is palpably a device to conceal the strong curve at the western end of the wall by making the outer line of the pavement straight throughout its length without sacrificing approximate equality of width at the two ends.

It not only furnishes a second incontestable proof of design in the curve, but also enables us to measure it: (5m. 84-5.76)

Another

+(6m. 4-5m. 84) = 28 cent. irregularity may be noted before passing to the interior. The line of the wall proper both on the northern and southern sides does not correspond with the line of the plinth on which it rests, but curves in and out along its whole length. In the diagram the wave line of the wall is indicated, without representing the line of the plinth. Measured on the northern side the distances (in centimeters) of the wall proper from its supporting plinth, vary in fifteen equal distances as follows. 32-4-4-4%31⁄2-31⁄2-5-4-51⁄2-6-61⁄2-5-6-3

-42.

Those who know the Pisan masonry will neither here nor elsewhere be disposed to speak of carelessness. It is an instance of the carefulness in detail of the architects of these buildings, that of all the pilasters of this wall not one has a base profiled like its neighbors, although all profiles are of Athenian grace.

I have preferred to rest my case for the numerous external curves on one which admitted of direct personal examination of the masonry. Of two others it can be said, at least, that they are not the result of sinking, the cornice and first roof line of the northern side (reproduced in fig. III) of the southern side, curve strongly in opposite directions, (compare lines 2 and 3).

Passing to the interior, I would first call attention to a perspective deception of very obvious effect. Of the two large arches (fig. VII) spanning the nave at its junction with the transept, that nearest the choir is round, and the other pointed; the apparent width of the transept space in the direction of the choir is thereby very much increased. Of a more subtle character is the arrangement of the galleries. The northern gallery rises from the transept toward the western end, as far as the seventh arch from the transept, twenty centimeters-here the line of the gallery is broken abruptly downward, and falls twenty centimeters in the remaining three arches; the western gallery rises from north to south twenty-three centimeters, and from its southern extremity the south

ern gallery rises eleven centimeters in the first three arches, where a break similar to that in the corresponding point on the northern side takes place, but not at so marked an angle; the line of the southern gallery from this point to the transept being nearly horizontal. Thus at the third arch from the western entrance the southern gallery is fourteen centimeters higher than the northern (eleven + twenty-threetwenty), and at the transept is thirty-four centimeters higher, (approximately fourteen twenty).

The hypothesis of sinking would here shipwreck not only on the perfect joints of the masonry, and well-preserved substructure and plinth of the north-western angle, but also on the fact that the unbroken oblique cornices without correspond to the broken oblique galleries within; the cornice forbids us to assume sinking for the gallery, and vice versa.

This variation in the height of the galleries is made still more perplexing for the eye, because there is a stripe of white masonry above the northern arches, but none above the southern (see VII), the latter being, therefore, not higher in proportion but disproportionately higher. As this inequality in the height of the galleries would be a perspective effect from any stand-point south of a central line drawn east and west through the nave with galleries of equal height, it follows that in any part of the nave the eye is deceived

FIG. VI.-THE BAPTISTERY AT PISA.