Download as pdf
Download as pdf
You are on page 1of 13
The Geometries of S. Andrea al Quirinale JULIA M. SMYTH-PINNEY University of Kentucky ‘The unusual complexities ofthe oval plan of Bernini's S. Andrea al (irl result from a series of variations funded in traditional Re naissonce geometric method. Through the comparative analysis of new seasured drawings ofthe church and existing archival evidence, the development of the design is traced from its initial Serlian beginnings to its final innovative resolution. Analysis of the drawings in con- Junction with historical document confirms Bernini useof conservative ‘geometric procedures and reveals the reasons for his subsequent departure _from strictest practice. The manipulation of geometry, proportion, ond dimension arose not only out of Bernini's concern for conceptual clarity and theoretical orthodoxy, but aso from a desire to use geometry in the support of spatial organization. The geometry ofthe fnal design ilu aminaesand underlines the essential simplicity ofthe interior and pues sa beter understanding of Bernini's intentions forthe viewer's visual experienc. AwaLysis oF tis design and construction sequence of Ber- nini’s famous church of S. Andrea al Quirinale in Rome through the examination of archival material has advanced significantly in the past 20 yeas.' Nevertheless, our knowledge of Bernini's design procedure sufers from the lack of accurate moder draw- ings ofthe church? By making new measured drawings (Figs. All drawings and analytical diagram are by the author. Grants fom the American Academy in Rome andthe National Insitute for Archi- tectural Education, the Southern Regional Education Board, and the College of Architecture, University of Kentucky, made it possible ro complete the research and ae gratefully acknowledged. My thanks also to Werner Seligmann and Joel Sanders for advice and help with mea- sing the building. 1. Since the publication of many primary documents by H. Brauer and R. Wittkower in Die Zichnungen des Gianlorenzo Bernini, Belin, 1931, other important stodies have included: F. Bosi, La chise dS Andres al Quirinale, Rome, 1967; G. Bauer, “Gian Lorenzo Bernini ‘The Development ofan ArchitecuralIconography,” unpublished Ph.D. diss, Princeton University, 1974; J. Connors, “Bernini's S. Andrea al, ‘Quitinale: Payments and Planning,” JSAH, XLI, 1, March 1982, 15- 37; CL, Frommel,“S. Andrea al Quirinale: genes eserutura,” Gian Lorenzo Bern archteteVarciteur europea dl seen, G. Sages and M.Fagiolo, editors, Rome, 1983, 211-256; and T. Marder, "The Evolution of Bernini's Designs forthe Fagade of Sant Andrea al Quit- inal: 1658-76," Architctrs, forehcoming. °2. To cite a conspicuous example, all of the measured drawings in Bors, S. Andres, are unforeunaely published without sale measure~ rents, or with graphic cles that are approximately 25 percent inac- crate, thus ditoring the building's tue size JSAH XLVIL53~65. MARCH 1989 1, 2, and 3) and comparing them in both scale and geometry with the known drawings in the archives, the geometric order of the final design is revealed, and Bernini's interests and in- tentions can be illuminated. ‘The essential clement of S. Andrea's form is the oval, a geo- ‘metric figure introduced into church architecture in the late Renaissance by Peruzzi, Serlio, and Vignola” It was Serli, in his treatise The Five Books of Architecture, who codified and pop- ‘ularized the geometric rules for the construction of ovals.‘ How- cever, the geometry ofthe S. Andrea oval is not strictly Serlian, and this has confounded previous authors. Some have arbitrarily imposed the standard Serlian ovals upon the plan.* Others have recognized the lack of correspondence with Serio but have been tunable to explain the geometry that Bernini devised or to dis- cover the reasons for his departure from accepted practice.* ‘An oval is constructed from the arcs of two pairs of circles (Fig. 4). For its construction, the four center points of those circles (A, A’, B, B') can be located by describing the pair of isosceles triangles (A, A’,B, and A, A’, B!) chat ae formed about 3, The oval is a goometic approximation ofthe tue ellipse. The cllpte has two unighe fixed centers and its perimeter ac changes con- stanly. The oval is constructed with the ares of two paired circles of| diferent radi. During the Renaissance and Baroque periods, the oval ‘waz habitually used in preference to he elipe becaute itis much simpler to construct. Fora survey of the history and geometry of the oval, see ‘T. K. Kitao, Circle and Oval in the Square of St. Per; Bernini's Art of Planning, New York, 1974, 31-38, and Appendix 1, 71-73. Much recent research on oval plan designs in the Renaissance depends on ‘Wolfgang Lote, “Die ovalen Kirchenriume des Cinguecento,” R- smiscesJaebuc fr Kuntgechiche, VIL, 1955, 7-99. 4. S.Serlio, Tue pera darchiettur et prospati, Venice, 1584, and numerous other editions. 5. P. Askew, "The Relation of Bernini's Architecture tothe Archi- tecture of the High Renaisance and of Michelangelo,” Marys, V, 1950, 57, fig. 21. Bos, S. Andie, 46, calls Askew's diagram “totally axbitary.” Kitao, Circe and Oral, 108, n. 127 also points out thatthe diagram is obviously inaccurate. 6. Borsi, 5. Andres, 46-50. Bors i unable eo discover the precise geometric changes inthe design drawings and the final church as bile. He concludes his book by making general comments on centrality and bifcalsm and the “substantial conservatism” of Bernini which, while accurate, ate not specific enough to be of great vale. Kits, Cirle and (Oval, 108, 127, stats that “Bors analysis of what he calls ‘traces regolator demonstrates conclusively thatthe pla [fS. Andrea] evolved incermally as well as externally either witha geometrical web so intricate that i hardly qualifies asa regulatory sytem, ofa is more likely, with no clear geometrical system 53 54 Jsan x1) 1:1, MARCH 1989 L Fig. 1. Gianlorenzo Bernini, S. Andrea al Quitinale, Rome, 1658-1676, measured plan (auchos) the major axis of the oval. By further describing the length of| the radii of the pairs of circles (AC and BC, respectively, for the smaller and larger circles), a specific oval is generated. Ie will have a fixed proportion (the ratio of minor to major axis lengths, X:¥), as well as a characteristic roundness, and will only be variable in overall size. Extending the sides of the tric angles to the oval’s perimeter will describe the four points of tangency between the arcs (C, C’, D, and D). Serlio gives four examples of oval constructions in his treatise ‘The first oval (Fig. 5) illustrates the standard construction pro- cess, and the resultant series of possible concentric oval based ‘upon equilateral 6O~degree triangles. Because he does not set the lengths of the radii for the circles, no fixed proportion for an oval is described.” 7. Serlio explains that “the neater that the figures come to their centers, 20 much the longer they are: and tothe contrary the further that they aze from theie centers the rounder they are...°(The Five Books of Architecture, reprint of 1611 English edition, New York, 1982, chap. I, book I fol 10 snd 11). ‘The other thre Serian examples are specifi ovals with unique eometric properties. The second oval is generated by two tan gent circles (Fig. 6). It is characterized by 45-degree angles and is the longest or slimmest ofthe specific ovals. The third oval is generated by two tangent squares, and is ako characterized by 45-degree angles (Fig. 7). The fourth oval, the ovato rondo, is generated by two circles whose centers cach lie on the cit- cumference ofthe other, and isa specific case of the fist oval (Fig. 8). I is characterized by 60-degeee angles. Each of the three specific ovals can be described by a mathematical ratio of the minor to major axis lengths (X:Y in Fig. 4), a8 follows: the second oval’s ratio is V2:2, or as a fraction, 0.7071; the third 8. Th onto ono isthe oval Bernini used asthe bass for the design of the Piazza Sin Piro, beginning in 1656, just prior to S, Andre, Kitao, Cire and Oval, 34, states that “the ato fondo was, in shor, the standard oval in architecarlpractice—at latin Bernini's Iely.” How- ver, Bernini did not use che oto tondo aS. Andrea. Interestingly, che cuca constraints imposed on the piazza design by existing site con- disions (ee Kitao, Circe and Oval, 37-38) are analogous to those that -Berini again faced at S, Andrea predetermined center line along what became the val’ minor ais, and imitation of length in that direction, SMYTH-PINNEY: GEOMETRIES OF 5. ANDREA AL QUIRINALE 55 Fig. 2. Measured section through the entry and apse (thor). oval’s ratio is 2~ VE: + V%, or 0.7574; the fourth oval’s ratio is 2(2-sin 60 degrees):3, or 0.7560. ‘One of the drawbacks of the Serlian ovals is their inability to be translated into simple mathematical and numerical ratios. ‘These complex fractions do not generate the simple whole num- bers that ate preferable for building construction purposes or for traditional Renaisance-based humanistic designs? For ex- ample, when expressed in terms of the real scale of the oval interior ofS. Andrea thes ratios become 63.6:90 Roman palms for the second ova, 68.2:90 palms fr the chird oval, and 68.0: 90 forthe fourch oval. “The design of S. Andrea canbe traced through two drawings that record two gcometially diferent preliminary oval schemes 9, Vignolas “commensurable oval” is an example of an sttempr to eal withthe problem of incommensurabilty in Serlo's ovals (Kita, Circle and Oral, 109, n, 128 and fg. 51) 10. One Roman palm = 2232 m. ~.7321 Although I measured and drew S. Andres using metic measurements, dimensions in this paper wil be given in Roman palms, Since the Bernini drawings were {done in alm, the cacial dimensions in mos cases are round nmbers {in palms, while in meters they are complex decimals. Thus, numerical relationships and ratios are easier to dscera by giving measurements in pals for the church. Bernini’ ist oval scheme is found on a papal chieograph dated October 1658, and preserved in the Chigi archives atthe Vatican (Fig. 9). This ist design is based upon Seclio's second oval (Fig, 10). The particular choice of the second oval can be understood by noting several ofits charac- vevitic features, The slim proportions of the second oval would best ft the site's restricted dimensions from the street wall tothe existing convent building, Furthermore, when coordinated with the ae- tual dimensions of the church, the second oval has the distinet advantage of becoming commensurable, and generating casily divisible whole-number dimensions in palms, as follows: the ratio of /2:2 results in dimensions of 64 and 90 palms in round ‘numbers fo the interior axis lengths; the concentric oval ofthe 11, Codex Chigi, P VIL 13, 40v~4 Published by Brauer and Wite- ower, Zeichuungen, figs. 167 and 168, the dating ofthe drawing i discussed in Connors, "Bernini's. Andrea," 17-19, and Frommel, “, Andrea," 216-217 12. The diagram is verified by the holes mase by the compass om the Vatican original, 2s well as on 2 copy of the papal chirograph preserved inthe Doria Pamphil archives, and published by Fromme, 5, Andrea," 219, fig. 4. Compass holes that exist on the original drawings are circled in my figs 10, 11,13, 17, and 21. 56 JsaH, xivitts1, MARCH 1989 Fig. 3. Measured section looking toward the altar (autho). Fig. 4. Diagram ofa standard oval construction (author) building's exterior, with the chapel rings thickness of 20 palms, results in axis lengths of 104 and 130 palms (again a simple ratio, inthis ease 4:5); and the center points for describing the ares ofthe oval are located at 22.5 palms from the axes’ crossing, Finally, che 45-degree angle inherent inthe second oval can be sed to establish the location of the intermediate piers on the diagonal. All of these advantages would have been quickly 2p- parent to Bernini and his staff, experienced as they were with oval geometries, and explain the rejection of the chied and fourth Serlian ovals as possible starting points for the design of S. Andrea. ‘Bernini’ initial decision to follow historical precedent as rep- resented by the Serlian oval is consistent not only with his conceptual preference for traditional procedures, bu also with the exceptionally short time involved in the design.'* However, Serlio’s second oval also has drawbacks relative to Bernini's intentions. The major innovation of the S. Andrea plan, that of locating piers not only atthe ends of che long axis but also at the 45-degece angles, forced Bernini to elaborate on Serlo's 13. The fire oval design was apparently completed in sbout one snonth's time, between August and September 1658. See Fromme, “S. Andeea,” 216. SMYTH-PINNEY: GEOMETRIES OFS. ANDREA AL QUILINALE 57 Fig 5 Setlo's fst oval (author, after Selo), Fig. 6, Serlio’s second oval (autor, after Sei), diagram right from the stare (Fig. 11). "The center points which generate the small circles for the oval had to be different from the points where the chapel centr lines intersect the major axis of the oval, and the 45-degree angle forthe center lines ofthe piers had to intersect that axis at still another point. Bernini also inserted an uncharacteristic equilateral triangle into the oval to gencrate the center lines for the four secondary chapels ‘The frst design apparently did not satisfy Bernini and sub- sequent revisions, previously unexplained, were made. The fist papal chirograph of October 1658 was used to lay out the foun dations and to begin to build the masonry shel of the church, completed a yea later in November 1659. However, itis fairly 14, Every other previous oval church used 45-degree or 60-degree angles to locate centerlines of voids, as for example in S. Anna dei Palafrenier, . Giscomo deli Incorbili, S, Carlo alle Quatro Fontane, and oval church projects by Serio, Raimaldi, and Peruzzi. See Kitso, Circle ond Ovals. 47-50, and Lotz,"OvalenKirchentiume,” gs. 28, 0, 45,51, and 52. SIN 60°, 2- SING Fig. 8 Serlo's fourth oval, che ovo tnd (author, after Seri). certain that Bernini modified the design of the church even during this inital construction year. Some of the modifications are recorded in a second version of the chirograph, now pre- served in the Rome State Archives (Fig. 12). Although this chirograph also bears the date of October 1658, internal evi- dence suggests that it was made later, between 1659 and 1660." 15, Archivio di Stato, Disegni e mappe, catlla 88, N. 476 TL. The text and che base deawing of the convent are copied exaey from the first chirograph. Frommel,"S. Andrea,” 223, argues convincingly that the major proportional and geometric changes to the main space and side chapels shown inthe second chiograph were already wnder con- sideration by August 1659. These changes were nat offically sanctioned ‘anti 30 September 1660, however, when the pope signed the second chirograph (Frommel,"S, Andres,” 233). The second chirograph also includes information on both the base sheet and an overlay sheet from the period afer 1660. Bauer, “Bernini,” 95, free noted the sense ofthe second chitogrph asa drawing i process. Ie includes, for example, che hgh ala chapel desiga that "was not even sketched out until 1668" (Connors, "Bernini's S. Andra,” 19) andthe travertine fagade that was ‘designed only in 1670 (Fromme, “S. Andre,” 241), 58 JsaH, xt M1, MARCH 1989 Fig. 9. Detail ofthe plan ofS. Andres in the fist papalchirograph (Biblioteca Apostolca Vaticana, Codex CChigi P VIE 13, 417) z0__| 0 FAME 20 Fig. 10, Serio's second oval compared tothe plan inthe frst papal chirograph (author). SMYTH-PINNEY: GEOMETRIES OF 5. ANDREA AL QUIRINALE 59. Fig. 11, Analysis of the plan inthe fis chirograph (author). Fig. 12, 8. Andrea, detail ofthe plan inthe second papal chirograph (Archivio di Stato di Roma, Dsegai e mappe, cartel 84, N. 476 I). Certain changes recorded in the second chirograph, including the enlargement and reshaping of the chapels and the reposi- tioning of their center lines, must have occurred before the masonry shell was completed, To make those particular changes akeer November 1659 would have required a major reduction in the size of the piers, which isan unlikely and undocumented operation." 16, Other changes made during tis period include modifications to ‘the windows in the main vaul (shown a8 ovals inthe section of the 60. JSAM, XLVITIT, MARCH 1989 Fig. 13. Analysis of the plan inthe second chirograph (autho). — Age 1 CENTER FONTS ARCS » CENTER ronts ARCS CENTER Fons SANTANDREA OVAL! ——- Fig. 14. Comparison of V22 ovals (author) {ist chirograph but buil asthe arched rectangles show in the second, chirograph, and the reduction in the numberof columas ordered for ‘the interior (ftom the eight required by the plan ofthe fst cirogeaph, ‘o the four shown in the second) by 2 contrat signed in March 1659 (Grommel, “S. Andre,” 220-221), ‘Analysis of the plan of the second chirograph shows changes from, but no real simplification of, the geometric complexities, of the fist chirograph (Fig. 13). The angles of the chapels’ center lines (25 degrees and 60 degrees respectively) remain constant, but the points where the center lines intersect the rminor and major axes change. Those points along the axis still do not coincide with the center points for the construction of the oval. Both drawings do rely, however, on whole-number dimensions in palms to set most of the points for the plan's construction. The oval in the second chirograph is now a vari- ation of Serli’s second oval: it retains the ratio of minor to ‘major axis lengths (64 to 90 palms), but its roundness has in- creased. Bernini's variation of Serio’ second oval is possible because, unlike a true ellipse, an oval with a set ratio between the minor and major axis lengths (inthis case \/2:2) can be constructed from an infinite number of possible locations forthe four center points of ts ares (Fig. 14). Iecan be relatively rounder or more pointed, with rolatively smoother or more abrupt points of tan- ‘gency. Serlo's second oval is relatively pointed and has quite distinct fla areas at the points of tangency. Bernini chose to vary Serio's second oval to make a rounder oval while retaining the original ratio ofthe axis lengths to each other. At S. Andrea 17. The minor and major axis dimensions ofthe S. Andee interior ‘oval (64 and 90 palms) remain constane in both papal chirographs, and. inthe church as buile, SMYTH-PINNEY: GEOMETRIES OFS, ANDREA ALQUIRINALE 61 reo PALS Fig 15, Analysis of the S. Andrea messured plan showing the fnal oval configuration (author) the redesign of the chapels most have led to the revisions in the geometry of the basic oval. As the chapel spaces were mod ified, the oval had to become slightly rounder."* These design changes would have been made futly soon after the start of construction in November 1658, because they were incorpo- rated into the masonry shell. ‘The third and final geometric revisions to the body of the church ate revealed by an analysis ofthe church plan as it exists today (Fig. 15). This final geometry is also depicted in a third contemporary document executed to display a design for the pavement and usually attributed to Bernini's assistant Mattia cde'Rossi (Figs. 16 and 17).* The geometric simplicity of the 18, The actual dimensions involved are almost insignificant. Math- matically calculated, Serio’s second oval would have a circumference ‘of 281.34 palms; the circumference ofthe rounder oval ofthe second chirograph would be approximately 245 palms, an increase of searely 4 palm (838 inches) in each quadran 19, Also in the Archivio di Stto, Disegni e mappe, catlla 84, N. 4476 1, Fromme, “S. Andres" 241, states tha this drawing also served tw obtain papal permission forthe addition ofthe travertine faade. The

You might also like