Gothic chevets
it was then that I saw the Pendulum. The sphere, moving at the end of a long wire attached to the choir vault, described its ample oscillations with isochronous majesty.
Umberto Eco(Foucault's Pendulum)
It's interesting to study the construction of the polygons that form the apses. On many monuments, you'll notice that the diagonal of the last bay of the choir directly prepares the angle of the polygon. By design, by the choice of an initial proportion, the final layout is anticipated. For those in doubt, note that Villard de Honnecourt drew these lines with the same pencil stroke (Fig. 1),

Fig. 1 - Chevet plans (Villard de Honnecourt - folio 29) D.P.
I've developed a number of procedures for drawing this type of cut. Here's a particularly interesting one (fig. 2). On the last bay of the choir, a square is defined, the diagonal of which is equal to half the width of the main nave (as shown by the dotted arc).
The height of this square is used to mark the apse's boundary. At the upper ends of the rectangle thus formed, I mark the pillars, and trace the semicircle of the apse with center O. From the lower corners, I draw two lines that intersect at O, defining two further pillars and cutting the circle.
The distance between the pillars we've just defined determines two of the sides we're looking for. All we need to do is transfer them symmetrically to complete the circle's partition.

Fig. 2 - Dividing an apse into five parts
The procedure we've just described is used to design the last bay of the choir and the chevet of many Gothic buildings. It can be found in Reims, the Abbey of Saint-Martin-aux-Bois and the Royal Collegiate Church of Saint-Quentin.
The chevets, crowned by seven chapels (fig. 3), are built in the same style. The width of the central nave is divided into three parts, creating a three-by-one rectangle inside the apse. The intersection of the diagonals (point O) will provide the origin of the circles, while their extensions will create the starting angle and fix the position of the columns.

Fig. 3 - Dividing an apse into seven parts
Knowing that half the length of the straight line is equal to the value to be transferred to the circle to form the angles, it's easy to draw the pseudo-heptagon. These are, in fact, one-by-two triangles, arranged in a fan-shape. The process is used in Beauvais and Amiens, among others.
Both are generic systems. Some cathedrals, such as Chartres, have their own specific layout.
I'll use the abbey church of Saint-Martin-aux-Bois in the Oise region to illustrate this point (fig. 4). The building features a transeptless aisle, enclosed by an apse similar to that of Reims.

Fig. 4 - Choir of Saint-Martin-aux-Bois
We've just seen that the side of the square determines the width of the last bay, while its diagonal is equivalent to the half-length of that same bay. To plot the height of the nave bays, the architect simply used this half-width to form equilateral triangles. There are two such triangles per bay. This segment (the red lines) is common to the entire building. It alone explains the eurhythmics.
In this example, it's hard to tell where the layout of the choir and nave begins. Here, they are part of the same nature. You can see a laser scan of this building at saint-martin-aux-bois.org
Traces of the apses of Mont Saint-Michel abbey and Notre-Dame d'Évreux
In the report "Mont-Saint-Michel - Le labyrinthe de l'archange" at 1h 15' 52'' , medieval art history professor Yves Gallet highlights a correspondence of proportions between the apse of this abbey and that of Évreux cathedral. He also notes that some of the choir's piers are lined with Romanesque supports, as are the crypt's columns. Understandably, the architect was forced to follow a pre-existing Romanesque rhythm.
The question then arises: since the chevet of Troyes cathedral is similar to that of Le Mont, isn't it itself a reflection of a Romanesque process? The history of Notre-Dame d'Évreux seems to invalidate this hypothesis. According to Philippe Gavet, the ancient Romanesque cathedral of Évreux is similar to the Abbatiale de la Trinité in Caen, whose chevet is segmented into five parts rather than seven. However, this element alone cannot dismiss the hypothesis.
In terms of geometric layout, the segmentation of the chevets at Évreux and Le Mont is the simplest imaginable. The width of the nave is divided into seven parts, seven units. The semicircle of the choir is set one unit apart. From one edge of the chancel, an arc with a radius of two units, an arc equal to half the width of the nave, and an arc equal to five units are drawn. By intersecting the semicircle, these arcs determine the exact centers of the piers.
as can be imagined, the segmentation obtained by this process is not regular, but the exact same deviations can be seen on the plans.

Mr. Jean Pierre Bourcier, a specialist in technical lines and stereotomy, checked my geometric demonstrations from a digital point of view. I'd like to share with you his analysis and opinion on the tracings we've just seen.
Analysis of the plot in figure 2 (created with AutoCAD)
"Let (A,B) be the width of the nave and (U,V) its axis. From point B, trace (B,M) such that the angle (A,B,M) is equal to 45°. Draw the arc of a circle with radius (B,U) and center B. It intersects (B,M) at I. From I, draw a line (D,C) parallel to (A,B). This line marks the boundary of the choir.
Let O be the midpoint of (D,C) and draw the line (O,B). It intersects the semicircle of radius (O ,D) at P, which is the center of the second stack, D being the position of the first stack. To trace the centers of the other stacks, simply transfer the chord (D,P) to (P ,Q) and then (Q,R). As you can see, the QT chord is larger than the DP chord.
If we keep the point P and its symmetrical W as a reference, and the point Q and its symmetrical T defined by transferring the chord (D,P), then the chord (Q,T ) is greater than the other four.
Let DP=CW=WT=PQ with QT > DP, which for a nave 12 m wide gives DP =3634.87 QT=4000, while the side of the decagon inscribed in the circle of radius OD is 3708.3
If we consider point T, the deviation between this plot and the plot of the theoretical decagon is 145.33 mm along an axis perpendicular to the nave axis and 49.48 mm along an axis parallel to the nave axis, with the center of the pile remaining on the circle of radius OD. In view of the function of the pile and the measuring instruments used at the time, the deviation is quite acceptable."

Figure 3 plot analysis:
"Let AB be the width of the nave and UV its axis AB is divided into three equal parts. The rectangle ABCD with short sides AB/3 is drawn. The two diagonals are drawn, and at their intersections the point O is the center of the circular axes of the piers. This point O is also the point of intersection of the radiating axes of the piles.
Let the circle of radius OE be the center of the first stack, its point of intersection I with the diagonal OD.
Let the center of the stacks located on the circle of radius OF be drawn. The diagonal OC is extended are intersection with the circle of radius OF defines the center of the first stack at J.
If K is the midpoint of OJ, the segment JK is the chord to be plotted on the circle of radius OF to obtain the points L and M, the centers of the next two piles JL= JK LM = OJ/2.
By symmetry with respect to the nave axis, we define the points N P and Q, centers of the following piers. The intersection of the straight lines radiating at O and passing through the centres of the piers of the circle of radius OF with the circle of radius OE defines the centers of the piers belonging to this circle.
We then see that JL=LM =NP =PQ with MN < JL, which gives for a nave 12 m wide and OF = 10616 (any value) JL=5308 MN=5012.34 the side of the 14-sided regular polygon inscribed in the circle of radius 0F being 4724.56
If we consider point M, the deviation between this layout and the layout of the 14-sided regular polygon inscribed on the circle of radius OF is 143.9 mm along an axis perpendicular to the nave axis and 33.9 mm along an axis parallel to the nave axis, with the center of the pile remaining on the circle of radius OF. In view of the function of the pile and the measuring instruments used at the time, this is perfectly acceptable."
Commentaires
David Orbach (Architecte - Ingénieur structure - Enseignant à l’Université Populaire de Caen de Michel Onfray)
Jean-Michel Mathonière - Directeur éditorial chez Éditions Dervy - Historien des compagnonnages
Cathédraloscope
Site : lescathedrales.wordpress.com
Jean-Pierre Bourcier - Spécialiste du trait
Olivier Petit - Médiéviste
Jean-François Lecompte - écrivain
Luciano Xavier - Maquettiste en cathédrales gothiques
Arcana Les Mystères du Monde - Youtubeur (Chaine Arcana)
troph38
Jean-François Lecompte - écrivain
John Brown
Armand Priest (ESTP) - Commentaire Facebook
Anthony CRESTIN - La géométrie et le mythe
Joël Supéry - Site tuskaland.com
Asso Fermat-Science
M. Moldovan
Catherine Leschenne





Dominique Gury