The Arch of Constantine
The Arch of Constantine

The Arch of Constantine

Assendelft – 1597 – 1665 – Haarlem

“The Arch of Constantine”
Oil on panel
41 x 51 cm

Attribution to Saenredam
The quality of the present painting is considerable, warranting the quest for its author.

Stylistically it is no doubt seventeenth-century Dutch.

The paramount focus on architecture and the degree of realism borne out in the array of finely nuanced colour and light gradations on the arch’s worn stone surface with complex sculpted reliefs, clarify that this artist must be a specialist in architecture.

The only Dutch seventeenth-century architecture painters that painted existing Roman subjects, are Pieter Saenredam and Jacob van der Ulft. Since Van der Ulft painted in a very different, later style with more focus on figures Saenredam remains.

These are observations/ considerations that seem to render Saenredam less likely:
- The painting is not signed whereas Saenredam usually signed.
- The walnut support is unusual for Saenredam.
- The condition of the arch is troublesome and the artist emphasized the many signs of wear and decay. Saenredam often eased them out in his known Roman views and other exterior views of buildings.
- There could be other artists that have been overlooked as candidates.

All these points can be countered.
- Signatures:
1 - About nine paintings by Saenredam that are undisputed are unsigned.
2 - We cannot know whether our painting was originally unsigned. Our painting could have been signed but the signature could have been lost in a cleaning.
3 - The lack of a signature in itself is no argument against attribution.
- Walnut support:
There simply are no Dutch seventeenth-century artists that preferably painted on walnut. If we accept that the painting is Dutch and seventeenth-century, we should also accept the possibility that the painting was executed by Saenredam.
- Saenredam’s presumed avoidance to depict conspicuous signs of wear in architecture: The early mention of a “ruin” by Saenredam and the choice to depict Roman subjects argue against the assumption of a natural inclination on the artist’s part to avoid depicting ruined structures.
- There could be other artists that have been overlooked as candidates:
Hardly possible. Architectural painters not specialized in Roman subjects who theoretically could have done a view such as ours as a one off are Bartholomeus van Bassen, Hans van Baden, Jan van der Vucht, Nicolaes de Giselaer, Dirck van Delen, Salomon de Braij, Gerrit Houckgeest, Willem van Vliet, Daniel de Blieck, Anthonie de Lorme and Emanuel de Witte.

Van Bassen was excluded (email) by Axel Rüger who works on the catalogue raisonné of Van Bassen. Van Baden, Van der Vucht, Giselaer and Van Delen never displayed the same concern for realism let alone the attention for surface qualities of the materials of which the structures are made. Furthermore, the compositions by these artists are more conventional and the style is more primitive. De Bray’s few architectural pieces form a stylistically consistent group of imposing interiors that are in mint state and this

painting does not fit in it. Houckgeest’s early works is similar to Van Bassen and are architectural fantasies while his works from c. 1650 are existing Dutch churches: this work does not fit into his oeuvre. The same goes for Delorme. The younger artists: Van Vliet en Witte closely followed the later work of Houckgeest: this work does not fit into their preserved output.

Painters specialized in Southern including Roman subjects but not specialized in architecture, who theoretically could have painted a view such as ours as a one off are Willem van Nieulandt, Barthomeus Breenbergh, Cornelis van Poelenburgh, Daniel Vertangen, Pieter van Laer, Jan Both, Jan Asselijn, Willem Schellincks, Thomas Wijck, Johannes Lingelbach, Herman van Swanefelt, Hendrick Mommers and Adriaen van Eemont.

Nieulandt painted in a much more old-fashioned style and the architecture is no more than an important backdrop, never the real subject and never represented so fastidiously detailed and realistically. The latter aspect also applies to Breenbergh en Poelenburgh. Vertangen displayed even less interest in Roman architecture than his model Poelenburgh. Van Laer, Van Wijck, Lingelbach and Mommers all were experienced in painting figures and the figures in our painting don’t look like any of their figures. Furthermore, they were only interested in architecture as a setting, focussing on a lively atmosphere. Both, Asselijn, Schellincks, Swanefelt and Eemont are landscape painters that never showed any interest in portraying a structure for its own sake, let alone without a strong sense of realistically rendered atmospheric effects. This work is alien to their stylistic idiom.


In terms of quality our painting matches Saenredam’s other works. Saenredam’s known Roman views each show a different approach.

One, formerly in Orléans, shows a panoramic view with an obelisk in the middle zone and the Colosseum and other buildings in the distance.

A second in Washington shows the church of Santa Maria della Febbre as a mishmash of old partly ruined buildings in the middle distance, fashioned as receding sequence of walls and a panoramic view opening up on the left.

A third in Berlin shows the Colosseum in the middle distance but with a large repoussoir at the extreme left in the immediate foreground.

A fourth in a private collection shows the entrance of the Pantheon from within with imposing columns in the foreground and no landscape view whatsoever.

All these four Roman views are signed, dated and based on drawings from a sketchbook by Maerten van Heemskerck.

Our (now) unsigned and undated painting also is based on drawings by Van Heemskerck. The idea to depict the arch from that close by is unusual but the drawings by Heemskerck show this approach and Saenredam had access to these drawings and

later owned them, making it even more likely that he is the maker of our painting and effectively ruling out other artists.

Since Saenredam took his lead from Van Heemskerck’s drawings and not from own sketches made on the spot his four Roman subjects do somewhat stand out as a group within the rest of his oeuvre consisting of Dutch subjects, and, among themselves. Still, there are parallels between our painting and Saenredam’s known Roman subjects and his Dutch subjects.

The idea to make a diagonal composition featuring a receding façade as the actual subject focussing on surface qualities and the suggestion of depth is seen in our painting and in the painting in Washington, but also for instance in his painting of the Sint Laurenskerk in Alkmaar (Utrecht, Catharijneconvent).

The subtle discolorations on the stone wall in our painting, especially the pinkish patches, is Saenredam’s trademark and they are not encountered in this way and so abundantly with other architectural specialists. Also the brushwork, including broad vertical strokes, that is used to suggest texture on the stone surface of the inside of the arch is highly comparable to the stone walls of the Sante Maria della Febbre in Washington.

The figures in our painting immediately call to mind Cornelis van Poelenburch. Comparison of the plump figures with numerous in securely attributed paintings and drawings by Poelenburch confirm that he must be the painter of the figures in our painting. Similar are the sturdy proportions of his figures, especially the thick arms that don’t seem to have wrists and the claw-like hands. Also the body positions, especially the figures seen on their back with their big arms slightly bowed at the elbow and the simple pleating in the garments as well as the cursory treatment of the faces are all typical of works by Poelenburch. The figures of early followers of Poelenburch such as Daniel Vertangen and Dirck van der Lisse are more refined, worked-out and have more slender limbs. Jan van Haensbergen can be ruled out as well because was only born in 1642 and stylistically our painting can be dated to 1630-60, and he probably only started to paint in Poelenburgh’s style late in his career. The figures of Gerard Hoet look very differently; more refined and idealized.

Since Saenredam had probably also collaborated with Poelenburch on the Pantheon painting, the presence of Poelenburch’s figures here can be taken as a potential extra argument that our painting is another collaboration between Saenredam and Poelenburch.

There are also technical similarities to our painting and the Pantheon work. Saenredaem used in both cases U-shaped protective wooden strips. Both panels correspond to the Utrecht duim. When Saenredam painted the Pantheon painting in 1643 he was busy with painting Utrecht churches, conceivably for Utrecht patrons. Given the Utrecht sizes of our painting, the collaboration with Poelenburch, who resided in Utrecht and the fact that Saenredam did not date other paintings in 1643, it is quite possible that he painted our work in that year or at least around that time, when he was absorbed by Utrecht subjects and probably in touch with Utrecht collectors. The use of pure oils and precious ultramarine finally argue in favour of a commission.

Finally the IR photo reveals a working practice consistent with Saenredam, namely the underdrawing which appears to have been prepared as a compositional drawing on a separate sheet first. After that the drawing was traced on the preparation layer. In raking light the remnants of tracing can be made out.