previously attributed to Cosimo Tura and Michele Pannonio and, Taddeo Crivelli, St George and the Dragon

Photo courtesy of Dave Penman (All Rights Reserved)


Country House
Mells Manor
St George and the Dragon
? 1460s
Medium and support
Tempera and gold leaf, punched and indented, on wood (made up to a rectangle?)
Overall height: 34.3 cm, Overall width: 18.4 cm
previously attributed to Cosimo Tura and Michele Pannonio and, Taddeo Crivelli
Catalogue Number


This lively and impressive painting was first published by Eberhard Ruhmer who attributed it to the young Cosimo Tura. His reasoning was clear: there are considerable similarities between the present panel and the much larger canvas of St George and the Dragon in Tura’s organ shutters of 1468–9 in the cathedral in Ferrara; and it is hard to deny that the two pictures must in some way be linked, with the present picture being an earlier attempt at the same subject. Perhaps insufficiently emphasised, other than by Boskovits, is the elevated quality of the Mells painting. The drawing is of great precision and the complex foreshortening, with the saint’s horse rearing forward but turning his head away, is exceptionally accomplished. The stance of St George, rising from the saddle with his foot set tensely in the stirrup, is stabilised by the successive horizontals of the bridle, his horse’s back and his left arm in the upper half of the painting; and, at the base, by the adumbrated ground and the flattened dragon, pinned down by the horse’s right hind leg. Curves are also handled in a masterly way: the play of the saint’s cloak frames the rising angle of the horse and there is an especially witty intersection of the dragon and the horse’s left hind leg around which his tail ineffectually winds. These planning and draughtsmanly features are of the highest quality and imply that the painting is by a major master. The tooling of the gilding in lines that converge more or less on St George’s waist simultaneously evokes an aura and creates a kind of notional semi-dome behind him, and the off-centre placing of George’s halo seems like a rising sun.

A work of such quality could well be thought to be a tyro effort by Tura and the ‘cragginess’ of the forms is also appropriate for him. However, the colour and tone differ from anything found in Tura or, at least, in Tura’s work as presently known. His formation, which is far from clear, seems to have had less of the decorative in it than this panel suggests and when Tura does indulge in decoration it is very spiky. Here the saint’s helmet with its elaborated plume – which in the opinion of the present writer recalls forms affected by Uccello rather than by Pisanello, who has been cited in this regard – and the colour of St George’s armour suggest a painter with a decorative bent, rather than one of Tura’s severity: this would fit with the relatively little that we know of Michele Pannonio (documented 1438–45 and 1463), to whom the painting was attributed by Boskovits. In support of this view, there are some points of contact between the Mells panel and a privately owned Crucifixion, transferred by Boskovits from Liberale da Verona to Pannonio (a transfer accepted by Toffanello). However, these do not strike the present writer as compelling and, otherwise, he sees little similarity between any of the panels included by Toffanello as by Pannonio and the single painting apparently signed by Pannonio, datable to the early 1460s, the Talia in Budapest; but the authenticity of that signature has been questioned. In short, the present writer is not convinced, firstly, that the panels attributed to ‘Pannonio’ truly form a coherent group and, secondly, that any members of the group can readily be connected with the Mells panel.

Dr Laurence Kanter has informed the writer that some years ago Dr Angela Dillon suggested to him that the ‘Pannonio’ group is, in reality, by Taddeo Crivelli (fl. 1451–d. Bologna before 1479), who was active in Ferrara for some two decades, from c.1450 to the early 1470s.1 Taddeo is best known as a miniaturist but he is also documented as painting larger pictures. The present writer has not, as yet, been able to follow up this lead and to check Crivelli’s work but notes that even should Dr Dillon’s re-christening of the group’s author prove to be correct, it would not allay his disquiet about its coherence. However, Dr Kanter informs the writer that he is inclined, from the photograph, to accept the Mells panel as by Taddeo and his view should be taken seriously. 

Concerning the inclusion of this panel in the ‘Pannonio’ or ‘Taddeo Crivelli’ group, the present writer should perhaps explain his misgivings more fully. As noted, the foreshortening of the Mells St George and the Dragon is remarkable; it is accomplished without apparent effort. Nothing similar is to be found in the ‘Pannonio-Crivelli’ group, none of whose members displays any particular adventurousness in drawing. It might further be noted that Tura consistently favours difficult frontal views of moving animals and rearing horses, whereas most artists show them in profile. If this panel were by Pannonio, and painted at the date suggested by Tofanelli, c.1450, he would be deploying a dramatic and sculpturesque manner not to be found elsewhere in his work, and one that requires a fluency in drawing that he does not elsewhere display. Furthermore, one would have to accept that the Mells St George influenced Tura. If, however, the attribution to Taddeo Crivelli is followed, then one would have to accept either that it preceded Tura – which seems improbable, for Taddeo shows no particular aptitude for complex foreshortening –  or that it was influenced by Tura’s Ferrara Cathedral St George and, therefore, would be a late work. 

If, however, one wished to restore this panel to Tura, one would face the problem that the colour and the cursive sense of line do not seem to resemble his at any point, and that one would need to allow for an otherwise unknown youthful phase. On balance, this is the position that the present writer favours but he is uneasy at countering the views of so many able scholars. 

While there has been much discussion of authorship there has been little of the painting’s function. It is now framed with a thin dark surround which converts a round-topped field into a rectangle. This surround seems not to be original but the present writer was unable to determine whether it is a physical addition, as Ruhmer states, or whether it is simply overpaint applied to a panel that has not been enlarged. But whether this framing element is an addition or simply a passage of paint designed to cover a raw area of panel, its inner edge is lit from the right. This cannot have been a necessity – the obvious expedient if it was desired to set off the composition would have been to surround the image with a flat frame. If it is a confection, then it was presumably created in order to customise the insertion of the panel into a specific context. If it either is, or reflects, the original framing of the panel, then it may provide some insight into its function.

What can this function have been? It can hardly have been a self-standing domestic object or it would surely have retained some trace of framing. But if it formed part of a larger ensemble, what might this have been? The shape would be odd but not impossible for the side panel in a small triptych, but the writer knows of nothing similar. Neither would it be inconceivable in a panel forming part of a predella but if this was its role, no companion piece has been found and one would have to assume that it was an upright panel set among horizontal ones. If it was part of an ensemble, it is perhaps more likely that it was a pilaster panel for which its dimensions would be appropriate and Dr Kanter is inclined to accept this possibility. The fact that it is gilded would also be appropriate for such a function. But, once again, no companion pieces have been found, either by the same or another hand.

It is not known when William Graham acquired the present painting, as it does not appear in the 1882 inventory of his collection. It is labelled ‘W. Graham 33’, suggesting that it was a comparatively early acquisition.

by Paul Joannides


Eberhard Ruhmer, Cosimo Tura, London : Phaidon, 1958, p. 170: Tura

André Chastel, 'Tura and the Secrets of Ferrara', Art News, vol. 57, no. 50, February 1959, pp. 40–1, 62–3: questions attribution to Tura

Carlo Volpe, 'L’apice espressionistico Ferrarese di Liberale di Verona', Arte antica e moderna, vol. 13, no. 16, 1961, p. 155: Liberale da Verona

Rosemarie Molajoli, L’opera completa di Cosmè Tura e i grandi pittori ferraresi del suo tempo, Milan : Rizzoli, 1974, no. 63, p. 9: uncertain attribution

Miklós Boskovits, 'Ferrarese Painting around 1450: Some New Arguments', The Burlington Magazine, vol. 120, 1978, no. 993, pp. 382–3: Michele Pannonio

Thomas Tuohy, From Borso d’Este to Cesare d’Este: The School of Ferrara 1450–1628, exh. cat., London : Matthiesen Gallery, n.d. [1984], no. 1, p. 58: as Pannonio, but with doubts

Joseph Manca, Cosimo Tura: The Life and Art of a Painter in Estense Ferrara, Oxford : Oxford University Press, 2000, no. 14, p. 171: not Tura

Oliver Garnett, 'The Letters and Collection of William Graham: Pre-Raphaelite Patron and Pre-Raphael Collector', The Walpole Society, vol. 62, 2000, d241, p. 327, Pannonio

Cosmé Tura e Francesco del Cossa: l’arte a Ferrara nell’età di Borso d’Este, ed. Mauro Natale, exh. cat., Ferrara : Palazzo dei Diamanti and Palazzo Schifanoia, 23 September 2007–6 January 2008, no. 42, p. 258–9 (entry by Marcello Toffanello): Pannonio, c.1450.


  1. Dr Laurence Kanter, private communication, 2020.


Related catalogue items from Mells Manor

  • Castle Howard Mells Manor

    Cicely Margaret Horner (1883–1972)

    Edward Burne-Jones, 1893

  • Castle Howard Mells Manor

    Thomas Horner (1737–1804)

    Thomas Beach, c.1778–80

  • Castle Howard Mells Manor

    Portrait of Charles Edward Stuart (1720–1788)

    after E. Gill, ? c.1745–6