Luca della Robbia was recognized as early as 1436 by Leon Battista Alberti (De pictura, Prologus) as one of the principal protagonists of the ‘rebirth’ of Florentine art in the early fifteenth century: at the time, he was carving the marbles of the monumental Cantoria destined for Florence’s cathedral. He was celebrated – and this is how we still mainly remember him – for his brilliant “invention” of sculpture in terracotta invetriata: “a new, useful and most beautiful art”1. This technique was one of the most archetypal, distinctive and admired expressions of the Florentine Renaissance, and continued to be practiced for more than a century by the heirs of his prolific workshop, a family business. This charming Virgin and Childtherefore deserves our close attention: it is one of the very rare autograph works to have come onto the market in recent times, produced in the decade that saw the birth of “Della Robbian'' sculpture and its recognition as a favored choice of the most cultured and sophisticated patrons of the period.
This fascinating relief has enjoyed considerable renown since 1880 when it was exhibited at the Exposition Nationale Belge in Brussels, in a section devoted to Italian art. This contained works from the vast collection belonging to Léon Mathieu Henri de Somzée (Liège 1837 - Spa 1901): a mining engineer, a wealthy financier, a parliamentarian and a collector particularly well-known for his archaeological antiquities2. Two years later, in the book devoted to this important exhibition, the Virgin and Child was illustrated in a delicate engraving after a drawing by Gustave Fraipont (fig. 1) and attributed to Luca della Robbia,
although some doubt was expressed as to whether it could be confused with the work of other members of the master’s family. The attribution was subsequently reaffirmed and debated by the most prominent experts3. Additionally, the provenance was noted as “du monastère'' of Santa Fiora in Tuscany, from where “la municipalité” apparently sold it directly to the Belgian collector, following the suppression of worship at the church of Santa Fiora4. This provenance, later described as “de l’église des Capucins de Santa Fiora” when de Somzée’s collection was sold on behalf of his heirs at auction by the Galerie J. Fievez in Brussels in 1901, 1904 and 19075, though at times considered doubtful6, in fact appears to be very plausible, as we shall see, and has a significant bearing on the work’s historical importance: it very likely suggests a prestigious commission by the Count of Santa Fiora, Bosio I Attendolo Sforza di Cotignola – a commission that would have been very influential in terms of disseminating Della Robbian art.
The work passed from de Somezée’s widow to Rudolph Bottenwieser and left Belgium before 19137 . In 1929 (3 May) it was finally sold by Paul Bottenwieser – a Berlin art dealer operating in New York out of the Anderson Galleries Building in Park Avenue – to the Albright Art Gallery in Buffalo (from 1962 the Albright-Knox Art Gallery). Records in the gallery’s archives (Objects Files: Della Robbia, inv. no. 1929.4) show that the relief was bought with support from the Seymour H. Knox Foundation after having been on loan for approval for over a year to the Directors of the Buffalo Fine Arts Academy (from 8 December 1927). The records also confirm the traditional attribution to Luca della Robbia with additional authoritative and glowing endorsement from the elderly Wilhelm von Bode (undated), who judged it to be “ein Maisterwerke” by Luca, as well as from George Gronau (20 January 1929). These expert assessments are handwritten in pen on the back of two photographs of the same date, which are also useful for documenting a careful restoration carried out after the sale of the de Somzée collection8. This followed an earlier intervention datable to before 1880, as suggested by the detailed analyses conducted in 2005 in the Art Conservation Department of the State University College at Buffalo, prior to the recent, commendable restoration in 2006. Now, repairs to multiple fractures and a few holes in the glaze once again allow the work to be enjoyed in full: it is evident from period photographs, that the surface condition is far more pleasing and legible now than as it appeared in reproductions published in the second half of the twentieth century9, which would certainly not have allowed its qualities to be properly appreciated.
The Buffalo museum’s acquisition of this important Della Robbian Virgin and Child was immediately publicized in various illustrated journals10. In particular, a contribution by Frida Schottmüller – the respected author of substantial catalogues of Italian sculpture in Berlin museums – was published in 1930 in the periodical of the local Fine Arts Academy. In 1938 the relief was exhibited in an important exhibition of Italian Renaissance sculpture curated by Wilhelm Reinhold Valentiner at the Detroit Institute of Arts. However, although it was repeatedly described in the scholarly literature on Luca della Robbia as an autograph work datable to around 1450 (see above), the work’s compromised condition, partly resulting from inappropriate restorations, as well as its position – uncongenial to a Renaissance sculpture – in a collection that tended mainly towards contemporary art, seem to have impacted upon the attention paid to this most graceful and delicate object – as Charles Avery11, among others, lamented: “despite its severe fractures, which modern conservation could make nearly invisible, this relief is thus an important document in Luca’s oeuvreand deserves to be better known”12.
The work’s critical reception seems to be largely linked, sometimes in an unjustly subordinate role, to that of a second, better-known example of the same model, once in the Kaiser-Friedrich-Museum in Berlin, a gift from Count Donhoff-Friedrichstein (fig. 2).
This had been acquired in Florence in 188813, presumably from the art dealer Stefano Bardini (there is in fact a photograph of it in the Bardini Archives in the Comune di Firenze) and on the recommendation of Bode, who was about to become director of the Berlin museums. Bode published it with an attribution to Luca della Robbia on several occasions from 1889 onwards14. Although the Buffalo relief was already known and had been attributed to Luca since 1880, it was the Berlin relief that gave its name to the typology described as the “Friedrichstein Madonna”in the fundamental monograph by Allan Marquand (1914). Most studies have focused on this example, at least until it disappeared in May 1945, after a fire broke out in the Flakbunker Friedrichshain, the anti-aircraft bunker where many works of art from the Berlin museums were being stored. The Virgin and Child, long thought “destroyed”15, is now in Moscow in the stores of the Pushkin Museum, together with many other Renaissance sculptures that escaped that tragic event, as recently revealed in an astonishing conference16 by Vasily Rasorguev and Neville Rowley, curators of the two museums involved in the re-emergence of this hidden heritage. We are grateful for their generous assistance. Although it has been broken into pieces and its bottom section is burnt, its condition is good enough for a more detailed comparison to be made – as has often been wished17– with the Buffalo example, which has sometimes wrongly been thought a workshop “replica”18.
The two reliefs appear substantially similar, including in their dimensions, which it has now been possible to check accurately (the Friedrichstein Madonna measures 47.6 x 38.9 x 10.1 cm. (maximum depth); the depth of the frame is 6.3 cm.; the image with its related strip measures 37.7 x 28.7 x 6.7cm. (depth of relief)). It is therefore plausible that both were made in the same mold, as has been speculated elsewhere19. This would probably have been a plaster mold, perhaps taken from an “ephemeral” clay or wax model. However, there are some notable differences in the detail of the modelling, the glazing and the technique used for the back of the terracotta.
The Berlin Virgin’s veil is more fluid and linear, rippling over the forehead in a way that still has a Late Gothic feel, while the Buffalo version has folds gathered over her head that are more complex and sharply defined, with a fringed border that is also clearly visible across Mary’s breast. The twisted sash tied around her waist is also more precisely described in the Buffalo version, while the mantle looping around her left hand has a double fold overlapping the ledge of the frame which is completely absent in the Berlin version. Both reliefs have splayed box frames, embellished with similar polychrome decoration to imitate inlays of precious materials (glass paste, fine marble, pietra dura) – like several other marble or terracotta Virgins made in the same period by Donatello and Michelozzo. However, the “copper green” glaze used in the Buffalo relief has a more delicate tone, tending towards turquoise, while in the Friedrichstein Virgin the color is deeper, similar to the green serpentine marble of Prato that appears frequently in Tuscan stone inlay and in the architectural framework of later Della Robbian production. The decorative strips in the Buffalo version also connect more satisfactorily with the blue disks in the corners, wrapping around the circular forms.
Differences are even more apparent when the backs of the two reliefs are compared. The rear of the Buffalo Virgin was hollowed out before drying in the areas of highest relief (fig. 3), in accordance with the technical practices of the best Della Robbian production; this can be seen for example in the coeval Virgin and Child of the type known as the “Genoese Madonna” in the Detroit Institute of Arts, datable to around 1445-50, as well as in the slightly later Triptych in the Bishop’s Palace, Pescia. However, the back of the Friedrichstein Virgin is solid and flat (fig. 4), as indeed is also the case with two replicas, cast from the same mold, of the Virgin and Child in a Niche, of the type known as the “Bliss Madonna”: these are in the Metropolitan Museum of Art, New York and in the Museum of Fine Arts, Boston, and are also datable to the same period20.
For the moment, since there has been very little examination of the technique used in Luca della Robbia’s studio for the earliest production of glazed terracotta, it is difficult to come to any conclusions, and various plausible hypotheses present themselves, as already indicated in relation to the two replicas of the Bliss Madonna21. The Buffalo Virgin, which is distinguished by more detailed modeling, more accurate technical execution (see, for example, the way that the back of the panel is finished with chamfered edges to the borders) and a more delicate palette in the glazing, could be an entirely autograph work (its high quality and its claim to this status has been endorsed by Avery 1976 and Pope-Hennessy 1980, after seeing it with their own eyes in April 1976), while the Berlin version could be a replica delegated to an assistant – such as Luca’s young nephew Andrea della Robbia (Florence 1435 - 1525), who was documented as working in Luca’s studio from 1451 onwards. Alternatively, acknowledging the initial experimental phase of Della Robbian art, it might be that the Buffalo Virgin was executed by Luca some time later, with a perfected technique and revisions to some of the details of the modelling. However, it is not impossible that different destinations may have accounted for such variations.
Nevertheless, there can be no doubt about Luca’s responsibility for this touching model, which clearly demonstrates the “familiar, pleasing, but suspended and silent humanity” that distinguishes the master’s best work towards the middle of the century22, or indeed about his responsibility for the execution of both reliefs, substantiated by the grey-blue of the irises, which are golden yellow in work completed by Andrea on his own. This composition can be closely compared to other glazed terracotta reliefs of the Virgin dating to the same period, which were also created from a mold so that they could be replicated in several versions, for example the Virgin of Impruneta (Santa Maria, frieze in the Cappella della Madonna), one of the few images with an integral, though simpler, frame; the Bliss Madonna mentioned above, which also uses a similar range of colors in the niche (see in particular the Boston example); the Corsini Virgin (Florence, Corsini collection; Tulsa, Philbrook Art Center); and the Copenhagen Virgin(Copenhagen, Statens Museum for Kunst; formerly Berlin, Kaiser-Friedrich Museum). These last are thought to be among the first attempts at glazed sculpture, dating to around 144023.
The Friedrichstein Virgin and its counterpart in Buffalo have frequently been dated to “around 1450 or slightly earlier”24, but Pope-Hennessy thought them earlier25, even as early as “about 1438,” seeing in both examples defects in the glaze and cracks from “faulty firing”, which he attributed to a technique that was still “inexperienced”. Without getting into the detail of this subtle argument, it is possible to add that a very early date, at least in the early 1440s, when Luca’s new art was first documented (the Tabernacle of the Sacrament made for Santa Maria Nuova, now in Santa Maria a Peretola, 1441-42, and the Resurrection in Santa Maria del Fiore, 1442-44), may be supported by its very close compositional affinity with an older model from the early 1430s, known through modest replicas in painted stucco (formerly Berlin, Kaiser-Friedrich-Museum; formerly Florence, John Pope-Hennessy collection; Passirano, Brescia, private collection; etc.). This model is distinguished by the charming invention of the Child kicking his legs as he lies in Mary’s arms as though in a cradle, an image that was taken up in a painting by Filippo Lippi (Traversetolo, Parma, Fondazione Magnani Rocca) and probably suggested by a well-known classical statue, now in the Louvre, depicting Silenus with the Infant Bacchus26.
Finally, it should be remembered that specific information exists about the provenance of the present Virgin: from Santa Fiora (Grosseto), an attractive village on the slopes of Monte Amiata in southern Tuscany. This provenance, provided – as we have seen – in 1880 and 1904, when the relief was in Brussels in the Léon de Somzée collection, is of particular interest and may be considered reliable.
On the back of the relief there is a round stamp that has at its center a coat of arms with a cross. This can be identified as the official emblem of the Kingdom of Italy (the Savoy cross surrounded by ornamentation of various kinds), apparently in the format in use between 1861 and 1870. It is surrounded by the text “[…] Oggetti d’Arte - Firenze […].” It is likely that this is the stamp of the commission for the export of art objects (“Esportazione Oggetti d’Arte”) of the Regie Gallerie in Florence, applied in confirmation of the license granted to the owner to take the Virgin abroad. This substantiates the account that de Somzée acquired the relief directly from the Comune di Santa Fiora, after the suppression of religious orders in July 1866 resulted in the transfer of their assets to the State and to municipal authorities, who then had the power to sell works considered to be of minor importance or little local interest. It is therefore probable that the purchase was made while de Somzée was based in Italy as director of the gas works of the ‘Compagnie Générale pour l’Éclairage et le Chauffage par le Gaz’ (1864-1867) – gas being a resource that was particularly plentiful in the volcanic terrain around Monte Amiata. It was during these years that de Somzée accumulated most of his art collection27). “L’église des Capuchins'' where the relief was formerly located, according to the 1904 catalogue, is probably the church of Santa Chiara, linked to the Capuchin convent (of Franciscan Clarists) in Via delle Monache. The church was consecrated in 1705 but contained some much earlier works, including a fifteenth century wooden crucifix that was greatly revered28. It should also be added that a provenance from Santa Fiora is also indicated for an imposing Della Robbia arch, decorated with coffering and vegetable festoons, adapted as a niche in the "Salle Renaissance" of the Hotel Somzée in Brussels (Monuments d'Art Antique (…) composant les Collections de Somzée. Premiére partie (…), Brussels 1904, pp. XII-XIII, reproduced in the title page of the Troisième partie of the catalog of the sale de Somzèe).
This provenance is very significant, since Santa Fiora, together with the neighboring village of Radicofani, was one of the richest centers for fine Della Robbian works29, including the early, exquisite triptych (Coronation of the Virgin, St Francis Receiving the Stigmata and St Jerome Penitent) in the parish church of Sante Fiora e Lucilla, made around 1464 by the young Andrea della Robbia while he was working in his uncle Luca’s studio30. These important works were commissioned by the Attendolo Sforza family, Counts of Santa Fiora and Cotignola, encouraged by Bosio I (Montegiovi 1411 - Parma 1475), a cultured and brave soldier of fortune who was a supporter of Franciscan Observance, and who had become Count of Santa Fiora through his marriage in 1439 to Cecilia Aldobrandeschi31. Bosio had a particular fondness for Della Robbian images, as also indicated by a similar triptych in Santa Maria degli Angeli in Assisi – presumably requested by his daughter Anastasia after her marriage to the Signore di Perugia Braccio II Baglioni (1462) – and even earlier by the beautiful Virgin and Child Enthroned in the Museo Capitolare, Atri, which is attributed to Luca at the time of Bosio’s second marriage to Criseide di Capua, daughter of the Duke of Atri (1464). This suggests that he may have commissioned this very graceful Virgin, intended for private devotion, as a gift to his noble wife Cecilia, perhaps not long after 1439 and before his premature death in 1451.
1. Giorgio Vasari, Le Vite de’ più eccellenti, pittori, scultori e architettori, Florence 1568
2. cfr. Cécile Evers, Léon Somzée, “l’ingénieur qui collectionnait,” in Appropriating antiquity. Collections et collectionneurs d’antiquités en Belgique et en Grande-Bretagne au XIX siècle, ed. Athéna Tsingarida, Brussels 2002, pp. 275-298
3. Schottmüller 1913 and 1930; Marquand 1914; Valentiner 1938; Avery 1976; Pope-Hennessy 1980; Gentilini 1989 and 1992; etc.
4. de Roddaz 1882
5. Catalogue 1904
6. Nash 1979
7. Schottmüller 1913
8. cfr. Catalogue 1904, fig. p. 49
9. cfr. Pope-Hennessy 1980, pl. XXV
10. International Studio, AD, etc.: cfr. Nash 1979
11. 1976, p. 12
12. cfr. also Nash 1979
13. Schottmüller 1913, pp. 31-32 no. 68, inv. I.143
14. Wilhelm Bode, “Luca della Robbia e i suoi precursori in Firenze,” 1, in Archivio Storico dell’Arte, II, 1889, pp. 1-9: p. 6, fig. 1 p. 8; a good reproduction later appeared in Denkmaler der Renaissance-Sculptur Toscanas, V, Munich 1905, pl. 223
15. Pope-Hennessy 1980, p. 250
16. Da Berlino a Mosca, sculture italiane del 300 e del 4000 perdute e ritrovate (1945-2016), Florence, Kunsthistorisches Institut, 3 May 2016
17. Avery 1976
18. Ragghianti 1938; Planiscig 1940; cfr. Nash 1979
19. Valentiner 1938; Pope-Hennessy 1980, pp. 62, 251; Gentilini 1992, pp. 102-10
20. cfr. Abigail Hykin, “Materials and Techniques,” in Della Robbia. Sculpting with Color in Renaissance Florence, ed. M. Cambareri, exhibition catalogue, Boston, Museum of Fine Arts, 9 August - 4 December 2016, Washington, National Gallery of Art, 5 February - 4 June 2017, Boston 2016, pp. 128-143, 154-156
21. Hykin 2017, pp. 131, 155 note 8
22. Gentilini 1992, pp. 102-103
23. Pope-Hennessy 1980, pp. 62, 66; Gentilini 1992, pp. 98-99
24. Schottmüller 1913; most recently Gentilini 1992, p. 60
25. 1980, pp. 35, 62
26. Gentilini 1992, pp. 47, 159 note 56
27. Evers 2002
28. Guida storico-artistica alla Maremma, ed. Bruno Santi, Siena 1995, p. 222
29. Bruno Santi and Carlo Prezzolini, Le Robbiane di Radicofani e Santa Fiora, Siena 1993
30. Gentilini 1992, p. 172
31. Riccardo Capasso, “Attendolo, Bosio,” entry in the Dizionario Biografico degli Italiani, IV, Rome 1962, pp. 532-533
The present lot is offered with a copy of a thermoluminescence analysis report from Oxford Authentication Ltd dated 24 November 2020 stating that the sample (N120h29) was last fired between 450 and 650 years ago, i.e. between 1370 and 1570AD.