Sant’ Agnese fuori le mura is my favorite church in Rome. I treasure memories of cool early morning visits on otherwise hot summer days (figure 1, 2011), and I likewise treasure the memory of introducing this complex of ecclesiastical buildings to classes of students I’ve brought there.
This is emphatically not the place to try to teach you about this (aptly named) complex, which illuminates a great deal of social, political, archaeological, ecclesiastical, and art history of the late antique and early medieval periods. But I will expend a few paragraphs showing you one of the gems of the complex, the imperial mausoleum of Constantine the Great’s daughter, Constantina, now called the church of Santa Costanza. We see today the relatively intact remains of a second phase of the building that dates to around 350 CE.
The mausoleum can be seen in Piranesi’s engraving from Le antichità Romane (figure 2). It’s the cylindrical structure with labels G, H, P, and R. Two cylinders form the nucleus of the building, a taller central one surrounded by a lower, wider one. An annular concrete vault (with wonderful original mosaics) runs between the two cylinders. The inner cylinder is not solid but stands on a series of columns like stilts. Just where the inner cylinder rises above the roof of the annular vault running between it and the outer cylinder there are clerestory lights that give the interior a luminous quality. The inner cylinder is capped with a cupola. Opposite the entrance is a sort of apse of rectangular plan that rises above the level of the clerestory. But this is all very dry prose. Your heart should beat a little faster when you see the interior (figure 3).
The point of view is from the entrance while standing under a portion of the annular vault I mentioned. The furthest wall you see at ground level with various niches in it is the outer cylinder. Twelve pairs of columns, arranged radially, support the inner cylinder, which spans the columns with arches. The clerestory lights can be seen above. The gray-streaked ring of pavers marks the footprint of the inner cylinder were it to come all the way down to the floor.
The radial central plan is typical of the direction late-antique architecture was taking; there is a surviving major church with a central plan, Santo Stefano Rotondo, not far away in Rome. It dates to the middle of the fifth century CE, a hundred years after Santa Costanza (figure 4).
Characteristic of late-antique, and especially paleochristian architecture, is the reuse of parts from earlier buildings. Close examination will reveal irregularities betraying such reuse even in the relatively low resolution images in figures 3 and 4. There are various practical and symbolic reasons for such reuse which needn’t occupy us here. What is important here is that on the splendid paired columns, which are unique in their radial arrangement in classical architecture, we find what are called composite column capitals (figure 5).
Composite means here a fusion of features of Ionic and Corinthian capitals. You see the acanthus leaves that so strongly mark the Corinthian, and above them the scroll-like volutes, with egg-and-dart molding between them, of the Ionic. Such capitals carry a notion of triumph about them, as they are typical features of Roman triumphal arches. An example from Santa Costanza can be seen in figure 6.
The close observer will note that this pair actually has only one composite capital; the outer (left) column is the sole example of a regular Corinthian capital in the colonnade. Even imperial builders couldn’t always command a perfect used set of 24 capitals. If you go there, it’s fun to point this out: it’s the second pair as you go around the ambulatory to the left from the entrance.
There is another church that bears a more than passing resemblance to Santa Costanza. That is San Michele Arcangelo in Perugia, also called Sant’ Angelo, which dates to the 5th or early 6th centuries, CE. Here is a cutaway drawing (figure 7):
Sant’ Angelo is a bit larger than Santa Costanza, quite a bit smaller than Santo Stefano Rotondo. Rather than having Santa Costanza’s concrete annular vault, the Sant’ Angelo architect employed a series of buttress-like arches to span the space like tendons between the two cylinders that make up the bones of the structure. Like Santa Costanza, the central cylinder is supported on columns, but this time single ones, not pairs. Over every other column similar ribs spring to form arches that support a timber roof above the drum. In these last features the church resembles Santo Stefano Rotondo more than Santa Costanza.
Sixteen reused columns, an ill-sorted lot, support the drum in Sant’ Angelo. They are nevertheless attractive colored marbles and granites (figure 8).
The Sant’ Angelo architect had to content himself with Corinthian capitals, since he was in the boondocks and composite capitals are comparatively rare. Yet in looking over the interior when I first saw it in 2015, I noticed that there was one example of a composite capital in Sant’ Angelo. More interestingly, in a church of the same basic plan and dimensions as Santa Costanza (they’re both about 80 feet in overall diameter, not counting lost external colonnades, etc.), we find that one composite capital is atop the third column along on the left when you start from the axis of the original entrance (figure 9 shows column 3 with composite capital and column 4 with a Corinthian capital).
Santa Costanza’s twelve pairs of columns are spaced at 30 degrees, whereas Sant’ Angelo’s sixteen columns sit at 22.5-degree intervals. That’s not quite precise, but it will serve for this argument. Do the calculation and you’ll find that if we take the radial axis bisecting the original entrance as zero degrees and count rotation (and columns) clockwise, the sole Corinthian capital in Santa Costanza, number 2, lies at 45 degrees, whereas the sole composite capital in Sant’ Angelo, number 3, lies at 56 degrees. The position of number 2 in Santa Costanza, at 45 degrees, bisects the difference between numbers 2 and 3 in Sant’ Angelo, 11.25 degrees from each. What this means is that an architect wanting to “cite,” as it were, his design source at Santa Costanza, could not have put his one outlier capital closer to where Santa Costanza has it than where he did, as it is a 50-50 split between columns 2 and 3 in Perugia.
I might add, though it is common enough that it would not have convinced me of a deliberate citation of Santa Costanza by itself, that the intercolumniations on the principal axis and the cross axis are slightly wider, and the arches on those axes slightly taller than the rest. This privileging of the principal axes, imposing a notional cross on the radial plan, is something Santa Costanza famously does as well. Indeed, even the use of arches over the colonnade was hardly inevitable. The Santo Stefano Rotondo architect uses a massive architrave, for example. In all of these details the Perusine architect further cited his Roman source, and the sole off-kilter capital makes sure we get the correct reference. The use of a striking onion-skin marble column under the one special capital directs our attention to that one in the first place.
Of course, if you look closer, you can find all sorts of differences, too. I’d expect that, as the amount of money the Perusine architect had to work with was almost certainly dwarfed by that available to an imperial commissioner; I suppose peculiarities of the site may also have played a role, as well as the provincial location of the building and origin of its workers. Still, when I mentally did the math for the radial position of those one-off capitals while I was in Sant’ Angelo, I was convinced that our architect had studied Santa Costanza as a model, used its basic plan, and left a winking citation in the composite capital. Some day I’ll go back with a tape measure, check historical records, and work the argument out precisely, but it’s too fun not to publish an anteprima here.
The slender resources of the interwebs on Sant’ Angelo suggest that no one else has seen this before—but you never know, almost every idea has occurred to someone else before . . . .