150
Montanha's transcripts, is the fact that the main altar of Madre de Deus, long lost after the 1835 fire (if not before), had statues of the four Jesuit saints in an order that echoed that of the façade. This third point implies that the design of the façade of Madre de Deus may have been conceived in relationship to that of the main retable inside. The religious and artistic interaction between a retable-façade and a main retable is a development known in some Latin American retable-façades at a later date, of which that of the Jesuit College in Tepozotlán, outside Mexico City, may be the most remarkable example.
The Architectural Structure of the Façade
One of the most remarkable features of this work's design is the use of projecting classical orders as main structural elements. However, it is not the use of the orders per se that relates this design to contemporary retables.
From the point of view of architecture, what the use of the Ionic order for the first storey, the Corinthian order for the second and the Composite for the third and fourth reveals is a rather sophisticated knowledge of the canons of ancient Roman architecture (Fig. 16). This is perfectly compatible with contemporary architectural practice. At the time Renaissance and Mannerist theorists had made these canons accessible to architects throughout Europe in a number of illustrated publications.
Whoever designed the façade of St. Paul's had evidently had a thorough grounding on these canons. How did he get such knowledge? Most likely from the sixteenth-century illustrated books by Serlio, Palladio or Vignola. Editions of their books circulated not only in Italy but also in the Iberian Peninsula. As the researches of Sylvie Deswarte show, knowledge of the Vitruvian orders had reached Lisbon, Portugal, quite likely around 1549 via Da Pintura Antigua by the Portuguese Francisco de Holanda. Holanda had underhandedly appropriated his main ideas from Serlio and published them in a luxurious edition as his own.22 In Spain, Francisco de Villalpando had already translated Books III and IV a few years before Holanda's book appeared, and his translation was known in Portugal.
These classical columns are important for my arguments because