Canaletto and the Moon: Part 3 — La vigilia di Santa Marta
This is Part 3 in a series of articles examining paintings by Canaletto that depict the moon at night. Can we use lunar position and phase to help date them?
Tomorrow, July 29, is the Festa di Santa Marta, one of the less well-known Venetian festivals and one that is no longer celebrated in a part of the city that is perhaps the most changed since Canaletto’s day.
If you’d been there in the 18th century for the celebration held the previous evening, July 28, La Vigilia di Santa Marta, it would have been fish supper of sogliola with saor for dinner. Festivalgoers would take to the water around the area of Santa Marta and head out to catch sole, returning to cook and eat it, accompanied by music and dancing well into the night.
Canaletto painted the scene as part of a commission for Sigismund Streit, a retired German merchant who had made his life in Venice. He wished to send a collection of paintings, books and manuscripts to his former school in Berlin, the Gymnasium zum Grauen Kloster, that the pupils there should know something of life and opportunity in the Venetian Republic.
As with all four paintings in the commission, the precise dating is uncertain. The period 1758–1763 is the established range in modern literature, and that is what is displayed alongside the painting where it hangs in Berlin today.
Santa Marta today
If you have visited Venice in recent years, probably the closest you may have come to Santa Marta is if you arrived by sea. The area around the church, at the far western tip of the city, has been continuously expanded and repurposed and now provides docks for arriving cruise ships and associated service organisations, including the port authority. Modern residential blocks occupy the central area, away from the waterfront.
The church of Santa Marta, never considered one of Venice’s great architectural glories, still stands in somewhat reduced form. It is almost the only surviving building of its time in the immediate area.
Look — another moon!
As with La Vigilia di S. Pietro (see Part 1) the possibility of applying astronomical dating techniques suggests itself: the painting depicts a well-known location on a fixed day of the year (July 28) showing the moon in a particular phase and position. A good match might support dating the painting to a particular year.
In theory…
Again, using Lodovico Ughi’s 1729 map of Venice we can get a sense of the Santa Marta area as it was in Canaletto’s day. We can overlay the relevant part of the map on a modern satellite map in The Photographer’s Ephemeris and align things by referencing three churches which still stand today: Santa Marta, San Nicolò and San Sebastiano (named S. Bastian on Ughi’s map), marked with yellow stars below.
Reducing the opacity of the Ughi map, we can see how much land reclamation has occurred since the 18th century in this part of Venice:
Taking the building depicted in the right foreground of the painting as our clue, we can identify the viewpoint Canaletto depicts:
We can establish an approximate azimuth for the moon using clues in the painting. An island can be seen offshore in the lagoon on the horizon towards the left. This is San Giorgio in Alga. In 1782, before the church and campanile were demolished, Doge Paulo Renier welcomed Pope Pius VI there, an event recorded in a wonderful painting by Francesco Guardi that shows San Giorgio in some detail.
In Canaletto’s painting, a full moon hangs in the sky above San Giorgio, slightly to the left. The bearing from our viewpoint to the island is ~245°. Knowing that the approximate diameter of the moon is 0.5°, and assuming it is depicted at the correct scale, with a little on-screen measuring the azimuth of the moon can be estimated to be ~244°, i.e. southwest. Using the same technique and assumptions, the altitude above the horizon is approximately +7°. Finally, the moon is shown full, as seen in this closer view:
All the usual caveats apply
There are all sorts of reasons why this sort of lunar analysis may be inappropriate (see Part 1). Let’s quickly recap the main ones.
Canaletto did not paint large canvasses in oil at night in the middle of a street party. He did make sketches, and some sketches of Santa Marta survive, but they are studies of the buildings on the right of the painting, indicating position, heights, colours, and materials. No moon appears in the sketches.
As with the San Pietro moon, it may depict another evening altogether or it may be entirely invented to suit the needs of the artist. If Canaletto decided he needed a moon to light the dark shoreline, or if his client had requested a moon, he would naturally have endeavoured to make it artistically successful.
What if, for example, the moon on the night had been hanging in the sky above Santa Marta itself? That would be compositionally weaker and would have provided no opportunity to paint the reflection in the water — reasons for the artist to ‘improve’ reality.
As it is shown, the moon over San Giorgio in Alga forms strong compositional triangles to lead the eye through the painting, along the row of buildings from the right, towards San Giorgio, and up to the moon. The moon sat over the water allows Canaletto to paint bright reflections, forming another triangle for the eye to follow: from the foreground to the woman staring out at the viewer beneath the tenting.
OK — on to the lunar analysis.
Scouting out the view
On a visit to Venice last year, we thought it would be fun to visit the Santa Marta area and see what remains of the views shown in the painting. What did we find? Not much.
While the moon phase and direction can easily be checked on a map, we thought it would be fun to show the view towards San Giorgio using augmented reality to simulate how the moon would have appeared in the 1750s and 1760s.
Canaletto’s viewpoint is not so easily accessible today. New buildings to the south block the view out towards San Giorgio in Alga. The photo below shows the remaining deconsecrated building of Santa Marta, looking along Calle Dietro Ai Magazzini (roughly meaning “alley behind the warehouses”, which tells you pretty much everything):
We had to “walk” along the line from the painting’s viewpoint towards San Giorgio in order to obtain a clear line of sight:
From there, it’s easy to explore the moon position using augmented reality in The Photographer’s Ephemeris to see if there are any likely matches. Here’s 1758:
Not even close. The giant grey pin in the camera view is sat atop San Giorgio. The orange line that bisects it represents +6° above the horizon (remember, we’re looking for around +7°). To get the moon even close to aligning with San Giorgio, we need to advance to the middle of the morning of July 29, with the sun well up. And even then, the moon is far from full, and sits much higher in the sky. You can check it out yourself using The Photographer’s Ephemeris.
1759 fares little better: the sun has not yet set when a waxing crescent moon is over San Giorgio.
1760? The moon is just past full, but it’s setting too far south. You’d need to stay up for a very late night indeed to see it anywhere close to what’s depicted.
In a similar vein, 1761 and 1763 also miss the mark. 1762 is slightly closer, but it’s not a full moon.
What about 1757?
Donald W. Olson in “Further Adventures of the Celestial Sleuth” makes a case for 1757, based on finding a full or nearly full moon in the southwestern sky at an altitude below 15° between the hours of midnight and morning twilight on the night of July 29.
He writes: “Canaletto was inspired to create Night Festival at Santa Marta as a nearly full Moon was sinking toward the southwestern horizon in the early morning hours of July 29, 1757”.
Here’s where the Moon would have appeared at around 2am (Venetian clocks would likely have been set a few minutes different to today):
The moon is indeed mostly full (waxing, 94% illuminated), but it is around 13 degrees farther to the south than the position depicted in the painting.
While you might argue that such a difference can be explained by artistic license, exactly that argument can also be invoked to ask why the painting should be assumed to depict an actual observed moon at all, let alone the early hours of July 29 1757 specifically.
Seeing Stars
Canaletto has scattered stars throughout the night sky. They are all painted in what may be Naples Yellow — a pale yellow made from lead antimony oxide that he is known to have used.
There are no constellations that I can confidently identify, except perhaps for one. Towards the left half of the sky is what might be a distorted painted-from-memory rendition of Scorpius:
Scorpius is prominent in the southern sky during evenings in the height of summer and is one of the more visually memorable constellations. However, the tail curves in the wrong direction and the position in the sky is not realistic (it is shown too high in the sky to be to the west southwest at this latitude).
Whether this is indeed a version of Scorpius or not, the painting of the stars suggests that Canaletto was no more concerned with astronomical accuracy than we might expect him to be.
If it is intended to be Scorpius, the stars and the moon are mutually inconsistent, astonomically speaking.
Canaletto’s Lunar Precedent?
There is one other well-known painting of the Festa di Santa Marta.
When I write “well-known”, that should be read as “visitors to the Ca’ Rezzonico museum may recall seeing it and scholars of settecento vedute likely know of it”.
It is by Gaspare Diziani (1689–1767), a contemporary of Canaletto better known for painting religious and allegorical subjects than vedute or “reportorial views”[1]. You can view the painting in high resolution online via Google Arts & Culture.
The scene is almost a mirror image of Canaletto’s painting. Buildings to the left, festivalgoers on the banks and in the boats, an island offshore in the mid-ground, and a full moon. But this time, the full moon lies to the extreme right of the composition, apparently in the skies above the distant Alps.
It is dated 1750–1755 in the literature, earlier than any currently accepted date for Canaletto’s version. Another candidate for astronomical dating, perhaps?
The topography of the painting is confusing. It is identified in the literature as depicting buildings on the island of Giudecca on the left, with the boats setting sail to head into the lagoon to the northwest. But there are several aspects which make this location hard to rationalise.
Where is the painter’s viewpoint located? On the north side of the island of Giudecca itself? It’s hard to identify a plausible location that would provide this angle of view onto the buildings.
Is it somewhere on Dorsoduro or San Marco, on the other side of the Giudecca Canal? If so, then the perspective is extremely compressed or distorted.
If we’re looking at buildings on Giudecca, then the church cannot be Santa Marta but must be San Biagio. But why depict San Biagio in a painting celebrating La sagra di Santa Marta?
The island in the lagoon appears also to be San Giorgio in Alga, but in reality, there’s no clear sightline from Giudecca that would also incorporate San Biagio. Rather, you would expect to see the Punta di Santa Marta itself to the northwest.
Perhaps instead, this is not painted from the Giudecca at all, but is a view from near Santa Maria Maggiore just north of the Punta di Santa Marta. This would make more logical sense — the painting would then be depicting the Santa Marta area itself — and it also accounts for the clear line of sight to San Giorgio in Alga. However, the buildings depicted don’t match up terribly well to what is shown by Ughi. Additionally, one of the large buildings to the left is identified[2] as the Palazzo Vendramin on Giudecca near to San Biagio.
Given the contradictions of the topography painted by Diziani, it would, in the words of Sir Humphrey, be a most courageous decision to attempt to date it based on the position of the Moon.
Erich Schleier, whose research established the “not before” 1758 date that is generally applied to Canaletto’s paintings for Sigismund Streit, suggests[3] that Streit may have been inspired to commission Canaletto to paint the Festa di Santa Marta on account of having seen Diziani’s recent version.
If so, then it seems quite possible that Canaletto’s depiction of a full moon, rather than being inspired by an actual observed event, is, in addition to its use as a light source and compositional aid, simply a nod to Diziani’s prototype.
I admit I started this journey excited at the prospect of potentially being able to improve the dating of the Streit Canalettos using the moon. But the deeper I dig, the less probable the whole notion seems.
The history of the dating of these paintings is a long and tortuous one — only one hundred years ago, the night scenes were not thought to be by Canaletto at all.
While the Gemäldegalerie and most sources nowadays stick to the 1758–1763 dating, others hint at more precise dates but the basis is not always apparent. Indeed, the original source for Erich Schleier’s own 1758 terminus post quem (see Part 1) has so far proved elusive. Perhaps further research will put that in doubt and bring Olson’s 1757 date back into contention — if it does however, I don’t think it will be on the basis of a full moon over San Giorgio.
[1] Peter Björn Kerber, Eyewitness Views. Making history in Eighteenth-Century Europe, The J Paul Getty Museum, Los Angeles, 2017, p. 1.
[2] Giovanni Mariacher, Recenti restauri a Cà Rezzonico Dipinti di Gaspare ed Antonio Diziani in Arte veneta, V (1951), pp. 173–176
[3] Erich Schleier in Venedigs Ruhm in Norden, Forum des Landesmuseums Hannover 1991–1992, pp. 35–37
