Brunelleschis Dome (13 page)

One of the keys to raising the dome was the precise calculation and measurement of each horizontal layer of brick or stone as it was added in a gradually contracting sequence. But how would these measurements be taken? How could the curvature of the eight individual walls be controlled during the process of construction? The difficulty was made even more acute by the fact that each wall had to incorporate two shells rising in tandem, as well as their supporting ribs. A deviation of only several inches in one of these ribs — each of which was over 100 feet in length — meant that the connection, like that at Saldae, would not be achieved.

The teams of masons at work on the dome had certain basic measuring devices at their disposal. Most of these had not changed significantly for a thousand years. For checking the perpendicularity of walls, for example, they used a plumb line: a string on which a weight, usually a ball of lead, was suspended. The string would be braided like a fishing line to prevent the weight from rotating in the breeze. And to ensure the stones were laid in perfectly horizontal courses or layers, a mason’s level was employed. This instrument was shaped like the letter
A
: a plumb line hung from its apex, while the horizontal crosspiece was inscribed with a graduated scale. The plumb line would come to rest in the center of the crosspiece when the stone or brick was on a level plane.

As they are neither perpendicular nor horizontal, vaults such as arches and domes obviously demanded a more complex system of measurement. The master builders of the Gothic cathedrals regulated the curvature of such structures by first plotting them full-scale, like a giant set of blueprints, on to special tracing floors. These floors were covered in plaster of Paris onto which life-size geometrical designs of, say, a vault’s ribs would be drawn. Once these drawings were complete, carpenters used them to devise the wooden templates from which the stone for the ribs was shaped by the masons working at the quarry. The gypsum floor was afterward wiped clear and the next set of drawings incised into its surface. If facilities for tracing floors did not exist, an area of ground would be cleared and the designs sketched in the soil. In 1395, for example, the plan for the timber trusses of the roof at Westminster Hall were set out on a patch of ground near Farnham in Surrey.

It was this latter method to which Filippo resorted in the summer of 1420. Downstream from Florence he had a large area of the Arno’s bank leveled, an expanse roughly half a mile in every direction, and in the sand he traced a full-scale plan of the dome.
¹
It is most likely that the templates for each of the eight vertical ribs were made from this enormous geometrical design. These models, cut from pine, were eight and a half feet in length and roughly two feet wide. Sheets of iron were used to stiffen them and prevent their warping. They were fitted onto the outside wall of the inner shell, allowing them to serve as guides for both shells, which were built with identical inclinations. Moved progressively upward as the dome rose, they ensured that the eight massive ribs would ultimately converge at the level of the fourth stone ring. In order for these ribs to serve as guides for the rest of the dome’s vertical curvature, they were built first: that is, only after several courses of the bricks for the ribs were laid did the masons begin filling in the intermediate sections.

Regulating the curvature of the ribs was not the only problem that confronted Filippo and the master masons. The cupola’s bricks were not laid in horizontal courses but rather at ever-increasing angles to the horizontal plane, with the final layers leaning inward at a steep 60-degree angle. A method therefore had to be found of guiding and controlling this gradual inclination. A related difficulty was calculating the radial dispositions not only of the bricks but also of the transverse sandstone beams of the second and third stone chains: all of this masonry had both to tilt inward and to radiate from the vertical center of the dome. Under such circumstances, traditional tools such as plumb lines and mason’s levels were quite useless.

How exactly Filippo calculated the disposition of the bricks and the massive stone beams is another of the dome’s mysteries. However, in
Historia Florentinorum
, written sometime during the 1490s, the humanist poet and historian Bartolomeo Scala offers a hint as to how he might have done it: “For, when the centre [of the dome] was pinpointed and marked,” Scala claims, “Filippo stretched a cord from the centre to the circumferences. Carrying on this process around in a circle, he determined in what order and according to what curvature the bricks and mortar were to be placed on the wall by the masons.”

That is, in order to guide the laying of the bricks, Filippo ran a cord (what the documents call
corda da murare
, “building string”) outward from the center of the dome to the inside edges of the masonry. This cord, which could be swept 360 degrees around the cupola, would have risen and then progressively shortened as more courses of bricks were added and the dome’s radius shrank from 70 feet at its foot to only 10 feet at the top. The inclination of the bricks as well as their radial positionings could therefore have been carefully monitored.

Scala’s account is supported by Manetti, who claims that Filippo used this same procedure when vaulting the Ridolfi Chapel. In this experiment the
capomaestro
used a cane that was fixed at one end and “circled upwards, gradually narrowing as it pressed constantly on the bricks on its unfixed side.” This device anticipates the modern-day trammel used by bricklayers to set out circular walls. The trammel consists of a horizontal wooden plank that pivots on an upright metal bar fixed at the center of the wall’s curvature. Describing an arc as it rotates round this axis, the plank indicates the position in which each individual brick should be laid.

Still, the curvature control device at Santa Maria del Fiore was obviously a much larger instrument: if it was to stretch from the center of the dome to its circumference, the
corda da murare
must have been 70 feet long at least. This unwieldy size would have presented certain problems. How, for example, was the cord prevented from sagging in the middle and therefore causing inaccurate measurements? Was a system of pulleys used? Or was it tautened and then smeared with wax like the ropes used by surveyors in the Middle Ages?

But most perplexing is how the cord was fixed at the dome’s center. A wooden pole would have had to have been 180 feet high to reach from the ground to just the base of the cupola, and almost 300 feet high to reach its top. The average height of a mainmast for a ship in the British Navy during the eighteenth century was 120 feet, and such masts could be built only with wood from the forests of the New World — Quebec, Maine, and New Hampshire — since no trees of sufficient height were to be found anywhere in Europe.
²
As one commentator has observed, “One would have to fantasize an enormous trunk of a California sequoia hoisted onto a central tower or suspended platforms.”
³

Whatever Filippo’s method of controlling the curvature of the dome, it had its critics. Not surprisingly, the most insistent of these came from the camp of Lorenzo Ghiberti. Late in 1425 Lorenzo’s deputy, Giovanni da Prato, appealed to the wardens that Filippo was failing to observe the terms of the 1367 model. As
capomaestro
, Filippo had of course sworn his allegiance to this sacred structure, just like all of the
capomaestri
before him. Yet Giovanni was dissatisfied. He voiced a number of complaints, the most serious of which was a claim that Filippo was not building the cupola according to the proper profile, the
quinto acuto
or “pointed fifth” curvature established by Neri di Fioravanti.

This pointed profile was important to the cupola both structurally and aesthetically. The pointed arch was, of course, the favored Gothic method of spanning space: the arches in the nave of Santa Maria del Fiore are pointed, for example, as are those in the naves of most Gothic churches. The pointed arch has two distinct advantages over the rounded or semicircular one that would come to dominate architecture throughout the Renaissance. The first has to do with proportions, for a pointed arch rises higher than a semicircular one of equivalent span. In 1367 this factor no doubt influenced the thinking of the wardens of Santa Maria del Fiore, for a dome with a pointed profile was able to stand as much as a third higher than a semicircular one built over a tambour of equal diameter. Only with a pointed curvature, that is, could the cupola reach the desired height of 144
braccia
.

The second advantage of a pointed arch is structural. The horizontal thrust of an arch or dome varies inversely with its rise, and since a pointed arch rises higher than a rounded one, it naturally generates less thrust. In fact, the architects at the Cathedral of Milan believed that pointed arches produced no horizontal thrust whatsoever. They were mistaken, of course, though a
quinto acuto
arch does generate as much as 50 percent less radial thrust than a shallower, semicircular one. It therefore requires less abutment and has a lower tendency to crack or burst at its base.

The
quinto acuto
profile demanded in 1367 is a geometric figure produced when the radius of curvature in the intersecting arches is four-fifths of the resulting span. The radius of curvature in a semicircular vault, by contrast, is only one-half the diameter, leading to a much shallower and more rounded profile. It was this figure about which Giovanni da Prato raised the alarm. In a submission to the Opera del Duomo he maintained that the dome had been “falsely built” because Filippo was constructing it “half round” and not as the specified pointed fifth. It was being built, that is, as a
mezzo acuto
, halfway between a semicircular arch and a pointed fifth. The result would be a dome incapable of reaching the required height. And Giovanni attributed this error not to some fault in the system of curvature control but, instead, to Filippo’s ignorance.

The
quinto
acuto
arch.

“I, the aforesaid Giovanni,” he wrote with some indignation, “declare that it seems to me that the angle chosen and suggested decades ago ought not to be changed or modified by lowering it for any reason whatever.” The dome would be aesthetically marred otherwise, he insisted, not to mention structurally unsound. In short, Filippo’s error in deviating from the established curvature would “brazenly spoil and endanger the church.” This submission ends, unsurprisingly, with a bitter personal attack on Filippo:

This has happened because of ignorance and presumption on the part of those to whom the execution has been entrusted, and who are being well paid and compensated for it. And I have written this so that if it befalls which all reason tells me must befall, and the building is spoiled and put in danger of ruin, I shall be excused and blameless. For God’s sake, be prudent, which I am certain you will be. Think of the danger that befell the cathedral of Siena for trusting a dreamer incapable of reasoning.

The tone is that of a biblical seer predicting future calamities should his words go unheeded. The unfinished extension to the cathedral of Siena to which he refers was a gargantuan folly that was partially torn down in 1357 after plague struck and funds ran out — a catastrophe that had for obvious reasons haunted the minds of the builders of Santa Maria del Fiore.

The motives behind Giovanni’s submission to the Opera were not perhaps the purest, especially since there was no basis for his claims. The shells, as built, conform exactly to the specified
quinto acuto
profile, and no corrections appear to have been necessary at any point during the construction.
⁴
The fact that Giovanni was so mindful of his own reputation (“I shall be excused and blameless”), as well as so obviously resentful of the fact that Filippo was “well paid and compensated” for his work, leads one to suspect a motive of jealousy.

In 1425 Giovanni had good reasons to be jealous of Filippo, even though he was himself an accomplished man, a respected humanist scholar who had composed a famous philosophical treatise, the
Paradiso degli Alberti
. However, like Lorenzo he had thus far failed to make an impression in the field of architecture. Filippo’s model for the wooden chain had been selected over one of his own, winning the
capomaestro
the very substantial prize of 100 gold florins. In the same year Filippo also won the competition for the design of the
castello
. Meanwhile, of course, his ox-hoist had been a great success, and the sandstone chains, the second of which was being laid in 1425, were working according to plan. Filippo’s reputation, in short, had never been better. To complete his triumphs, a few months earlier Lorenzo Ghiberti had been suspended from his duties as
capomaestro
, possibly because of Filippo’s intrigue involving the wooden chain. He would be reinstated shortly afterward, albeit with reduced responsibilities and powers. The Ghiberti faction had reached its lowest ebb.