Jean Hey
NG 4092 
The Meeting of Saints Joachim and Anne at the Golden Gate; Charlemagne

Lorne Campbell, 2014

Extracted from:
Lorne Campbell, The Sixteenth Century Netherlandish Paintings with French Paintings before 1600 (London: National Gallery Company and Yale University Press, 2014).

Fragments of an altarpiece
Oil on oak panel, 72.6 × 60.2 cm,
painted surface 71.8 × 58.8 cm

Provenance

The first known owner was Victor Decock, a Belgian painter, restorer, dealer and collector who lived in Paris. He had settled there by 1919¹ and died in or shortly before 1948. He was described as ‘one of those Flemish collectors who are inspired by a genuine passion for the art of their country as well as by real connoisseurship’.² His collection was auctioned on 12 May 1948 at the Galerie Charpentier in Paris.³ He had sold the Meeting of Saints Joachim and Anne at the Golden Gate apparently in 1923 to the dealer François, Francis or Franz Kleinberger ( c. 1858–1937), Paris,⁴ from whom it was bought on 11 July 1923 by ‘Paul Amsterdam’.⁵ Paul Amsterdam seems to be a coded name for the Amsterdam Gallery, founded in 1923, of the Berlin dealer Paul Cassirer (1871–1926). It was sold by Cassirer to the collector Hans W.C. Tietje, Amsterdam;⁶ by Tietje to Agnew’s in London; and by Agnew’s to the National Gallery on 7 July 1925 for £10,000, paid in three instalments between 9 July 1925 and 7 April 1926 out of the Temple West Fund.⁷

Version

A figure of Charlemagne on the reverse of the left wing from a small enamel triptych seems to be a free version of the Charlemagne in NG 4092. The wings are in the Wallace Collection (XII A 68; figs 1, 2); the central section is lost. On the reverse of the right wing is a Saint Louis, perhaps related to the figure that once appeared on the left of the Annunciation in the Art Institute, Chicago (fig. 3). On the obverses of the wings are Pierre II of Bourbon presented by Saint Peter and Anne of France, Duchess of Bourbon, presented by Saint Anne. These are versions of the wing panels of the Moulins Triptych. The Wallace Collection wings are enamel on gold, measure 4.5 × 3.7 cm, and were in the collection of Louis‐Fidel Debruge‐Duménil (1778–1838) in Paris. It has been suggested that these wings are nineteenth‐century fakes. They were examined at the British Museum on 22 September 2011 by Andrew Meek, who found no reason to doubt their authenticity. They are entirely genuine and distinguished enamels of the period around 1500.

Technical Notes

According to an inscription on the cradle removed in 1989–90, the panel was cradled in 1924 by B… Bommel in Amsterdam.⁸ Immediately after its acquisition by the Gallery, a small loss in Saint Anne’s robe was repaired, the unpainted area at the top of the right edge was overpainted, and strips of wood were added at the top and on the right to provide a rebate for framing.⁹ The paint has a history of flaking and has been consolidated several times. In 1989–90 the cradle was removed, the [page 755][page 756]panel was repaired, the added strip of wood was removed from the right edge, and the added strip at the top edge was reduced in width. The painting was cleaned and retouched. Since only a few fragments remain of the deepest shadows on Saint Anne’s red robe, it was decided not to match them and to give only indications of the missing modelling.

Apart from losses of paint and glazes in some of the shadows on Saint Anne’s robe, the picture is remarkably well preserved. There are losses on Saint Anne’s right cheek and lower lip, in Joachim’s hair and beard and in Joachim’s clothes. A small part of the lower right corner of the panel has broken off and has been replaced: a tiny area of original paint has been lost.

The structure of the panel is complicated and is more easily explained in an illustration (fig. 4) than in a detailed verbal description. At the top, the original unpainted edge was removed up to the barbe before the added strip was attached. An unpainted edge survives on the left, but it may have been slightly trimmed. At the right edge the panel has been cut. At the top is an area of unpainted wood 33.2 cm high and between 0.8 and 0.9 cm wide. Here an element of the original frame covered the panel. Below that, the panel has been cut through Charlemagne’s arm. The damage to the lower right corner may have occurred at this time.

The panel is made from four boards of oak, laid horizontally. Dendrochronological investigation has established that the oak was grown in France and that the third and fourth boards from the top are from the same tree. It is clear that NG 4092 and the Annunciation in Chicago, which has been cut on the left, are two sections of the same panel (figs 3, 4). The second, third and fourth boards of both panels have the same dendrochronological properties; the rings of the first boards of both panels could not be measured and are to that extent similar. The 169 rings of the second board of NG 4092 were formed between 1258 and 1426, the 144 rings of the third board between 1298 and 1441, and the 109 rings of the fourth board between 1331 and 1439. The third board continues like the rest into the Annunciation, where its 116 rings were formed between 1340 and 1455.¹⁰

The panel has been thinned. Its present thickness varies between 0.6 and 0.8 cm. The thinning process has exposed three dowels, themselves reduced to half their original thickness, one across each of the three joins. Also exposed is a shallow dowel channel, 6.3 cm high, cut into the top of the third board: as there is no matching channel in the second board, the panel‐maker must have created this channel by mistake. The Chicago Annunciation has also been thinned and cradled and its cradle is still in place. Three vertical dowel channels about 12 cm high have been found there. In NG 4092, the dowel slots are 11.7 cm, 12.4 cm and 12.0 cm high, the dowels being 10.3 cm, 9.6 cm and 9.7 cm high.

There are indications that whoever removed the original frame had some difficulty in performing the task. An incision along part of the unpainted edge on the left may have been made during preparations to remove the frame. Indentations which seem to be comparable are on the three uncut sides of the Annunciation, about 1 cm from the edges. Tiny fragments from the original frame, vertical in grain, still adhere to the unpainted edges on the left and top right of the London panel.

The ground is chalk bound with animal glue. It appears to have been sealed with glue size, over which is a very thin priming of lead white extended with some bone white (calcium phosphate).¹¹ The use of the extender accounts for the fact that this priming is less densely white than most of the other primings examined for this catalogue. Infrared reflectograms show a great deal of underdrawing and many changes (figs 5, 6). The underdrawing appears to have been made in liquid media and with brushes. There is a certain amount of hatching, especially in the draperies. Several ruled lines are visible in IRR . Some, but not all, of them are used in the painting. Diagonal lines descending from left to right across the landscape in the top right sector may indicate that the artist was uncertain about perspective or that he was trying out different vanishing points.

The rest of the underdrawing is very free and sketchy and marks approximate positions, often varied several times, for the main elements of the composition: the figures; the arches of the Golden Gate; the towers of the castle; the wall behind Charlemagne. The precise positions in which they were painted, however, are not underdrawn. Also sketched are a number of features that were not included in the final painted composition: for example, a cloud above Charlemagne’s head and additional arches near the left edge of the painting.

The poses and draperies of the figures have all been changed, often several times. Some have been redrawn; in many parts of the painting, none of the underdrawings has been followed. There are, for example, at least three underdrawn positions for the left leg and foot of Joachim, none of which is exactly followed in paint. Saint Anne, as she was first drawn, appeared to be more clearly in front of Joachim. During painting, Joachim’s right shoulder was brought forward towards Anne, his purple mantle was extended across the area where her yellow veil was drawn, and the architecture behind was taken across the veil drawn under her chin. The veil on her head was drawn higher on her forehead and may have been doubled in the underdrawing. The cowl of Joachim’s mantle was underdrawn higher behind his head; the position of his left arm seems to have been indicated undraped and then several different sleeves were drawn. Joachim was first painted with more hair across his forehead and ear and his face has been moved slightly back. During painting, his left wrist and hand have also been altered. The skirt of his robe was drawn shorter; his right foot was further advanced; his left knee was bent and his underdrawn left foot rested on its toes on our right of the painted foot. Originally there would have been a greater sense of his movement towards Anne. Charlemagne’s mantle has been altered several times in the underdrawing, especially on our left but also on our right above his elbow. His legs, especially the left foot, have been redrawn several times. His orb was larger in the underdrawing. During painting, the angle of his crown has been altered. It was once further towards our left; parts of the crown and forehead are painted on top of the background.

In a cross‐section of a paint sample taken from the figure of Charlemagne, traces of a very fine black were found between the ground and the priming. They may be drawing; a thicker layer which is more certainly drawing is present on top of the priming; and there is a further layer of drawing between the light grey base colour of the ermine and the thin white lines of the hairs of the fur.¹² This could indicate successive stages of reinforcing and refining the primary drawing but, as there are many alterations in this area, it may instead relate to changes made at different times during the execution.

Natural ultramarine has been used lavishly, sometimes in mixtures with white or red lake.

An unusual feature of the layer structure is the extent to which the artist employed a preliminary painted design in tones of pinkish brown, red‐brown and redder tones mixed from red earth colours, white and, in certain underpaints, some vermilion. These underlayers vary in colour and constitution from place to place. It has been noted that cross‐sections from the Chicago Annunciation show that a similar technique was used on that section of the altarpiece, although it is not known how extensively. This must have been some type of compositional device, rather than a plan to recast the colours of the draperies as the painting evolved. The underlayer colours involved bring to mind the quality of a sinopia design beneath a fresco; but the technical motivation here cannot be a direct parallel. Saint Anne’s dress, for example, is underpainted in red‐brown or pink (red earth and white), over which is a layer of vermilion. Her sleeves are underpainted in bright red but finished as bright blue with mixtures of ultramarine and white. Joachim’s mantle is purple, mixed from natural azurite and red lake and glazed with a thin layer of red lake to which some colourless powdered glass has been added;¹³ but the underlayer is a brighter red earth mixed with lead white. The lighter coloured pinkish‐mauve lining consists of a thin red lake layer over a pale blue of ultramarine and white. A layer made up of red‐brown earth lies beneath this. Joachim’s grey boots are over a brownish‐red underlayer. The blue of Charlemagne’s tunic is underpainted in red earth and white, which may be a base colour for the cloth‐of‐gold sections of the same tunic. The blue is ultramarine and white, with a final scumble of pure ultramarine. The saturated greens of Joachim’s drapery, the fragmentary altarcloth design and the foreground foliage (fig. 17) comprise multiple layers of paint composed either of lead‐tin yellow mixed with verdigris, or lead white mixed with verdigris, finished with translucent green oil glazes based on unmixed verdigris pigment.

The medium was identified, in all the areas that were analysed, as linseed oil. It is partially heat‐bodied in a sample from the path at the right edge at the level of Charlemagne’s knees. In the sky at the right edge, it is heat‐bodied. In the red of the cloth of honour at the right edge, it is heat‐bodied and there is an addition of some pine resin. In deep translucent green foliage paint at the right edge, the upper layer is in linseed oil, heat‐bodied and mixed with some pine resin, whereas the lower layer is in linseed oil that has not been heat‐treated but to which a small amount of pine resin has been added. The differences between the media in the two layers may explain why there has been some flaking of the upper from the lower layer, which might have dried more slowly and contracted after the upper layer had dried.

The painter has displayed considerable skill in using his brush to push away the upper layer of paint and so create the wrinkles in Joachim’s brow. He has worked blue and yellow detail wet‐in‐wet into the green paint of the bush at the right edge. The red lake glazes have been blotted (fig. 7), for example in Charlemagne’s mantle, but here more red lake, sometimes mixed with a little black, has been applied over the blotted glaze to make the deepest shadows. These final glazes have not been blotted. The contours in the flesh, for example in Charlemagne’s face and hands, are lines of red lake (fig. 8). Very similar red lake outlines are found in the head and hands of the Virgin in the Chicago Annunciation.

Description

The Golden Gate, built from pinkish and greyish stone, is decorated with images of three winged putti, who hold swags. The putti and the swags give the appearance of objects that have been gilded. Above the putto on the right is an image of a crowned head framed in an oval: it is a relief carving in grey stone of a king, whose crown appears to be gilded. Facing him is part of another such relief but only the neck is visible. The kings represented may be David and Solomon, especially if the Golden Gate is understood to be a gate of the Temple rather than a city gate. Weeds are growing out of the mortar, [page 761]on the rocks and across the path and turf in the foreground, where the plants are a plantain, a dandelion, another plantain, a daisy and a violet. Between Joachim and Charlemagne is a thistle. Saint Anne wears a ring on the wedding finger of her right hand. Her yellow snood is striped in red. Joachim’s pouch is reddish, presumably leather, and the two subsidiary pouches are closed with drawstrings. The handle of his sword is transparent, possibly crystal, and his left boot is closed with a toggle of yellow metal.

On top of Charlemagne’s crown is a black double‐headed eagle (fig. 10); its beaks and legs are red. Charlemagne’s orange armour is perhaps leather, set with yellow metal studs and other decorations. His orb is crystal and reflects the hand that holds it as well as a window, presumably in a building outside the picture. He wears a coat of mail, its lowest rings gilded and falling in points, and over that a surcoat with the imperial arms impaling France ancient and referring to Charlemagne’s titles of Holy Roman Emperor and King of the Franks. This surcoat is richly woven and embroidered. The imperial eagle is black with red legs and feet; a black arc connects the feathers of its displayed wings; its tail feathers are green. Hanging over the parapet behind him is a rich textile of red velvet cloth of gold with areas of black and green and blue flowers. The largest motif is a yellow eagle and one of its three‐toed feet grasps a green tendril. The brownish shape cut by the right edge at the level of Charlemagne’s ankles (fig. 9) is presumably a continuation of drapery worn by a figure from the lost section, or, less probably, part of a shoe or boot. It is purplish brown; on top of that are lead‐tin yellow and red (earth?); and, to make the shape near the edge, there is a similar mixture with less yellow and perhaps some fine black, so that this area is brown rather than orange.

In the background are towers and other buildings, all with very small windows. The building on the left is constructed on top of a bridge that spans the river. On the river are swans, on our left, and brown ducks, on the right of Charlemagne. More birds perch on the roofs of the fortresses in the background.

Iconography

The story of Joachim and Anne is narrated in the Golden Legend,¹⁴ itself based on ‘The Gospel of the Birth of Mary’, attributed, wrongly, to Saint Jerome.¹⁵ Joachim and Anne had been married for twenty years but had no children. The priest in the Temple refused Joachim’s offering because he was sterile and he went to live with his shepherds. An angel appeared to tell him that his wife would bear a daughter, who was to be called Mary and consecrated from infancy to the Lord, and that he would meet Anne at the Golden Gate of Jerusalem. The same angel then went to Anne and told her the same things. Joachim and Anne met at the Golden Gate and Anne conceived the Virgin Mary. Some of those who believed that the Virgin was immaculately conceived thought that the conception took place at the Golden Gate. Pope Sixtus IV in 1476 and 1480 introduced Offices of the Immaculate Conception.

The third figure is Charlemagne (747–814), who was crowned Holy Roman Emperor at Rome in 800. He is identified by his closed imperial crown, his orb and the impaled arms of the Empire and France on his surcoat: these were the arms assigned to Charlemagne in the thirteenth century. He was similarly represented in other paintings, manuscripts and tapestries. The ‘Retable of the Paris Parlement’ (Louvre) includes a closely comparable representation of Charlemagne [page 763](fig. 15), who was venerated as one of the Nine Heroes and as a saint. His canonisation by the anti‐Pope Pascal III in 1165 has never been ratified by the Holy See.¹⁶

Reconstruction

NG 4092 is the left side and the Chicago Annunciation (fig. 3) the right side of one oblong panel of four horizontal boards (fig. 4). They are not perfectly rectangular. The joins between the first and second boards and between the second and third boards run at angles to the horizontal. Since it may be assumed that the edges of the second board were relatively straight, it is possible, by continuing the lines of the joins at the same angles, to determine where they would have met and so estimate, with a fair degree of accuracy, the width of the missing central section. It was about 48 to 50 cm wide and the whole painting would have been about 160 cm wide (see fig. 4).¹⁷ The painted surface of the Annunciation measures 71.7 × 49.2 cm. X‐radiographs and infrared reflectograms show that part of a standing figure has been obliterated at the left edge. It has been scraped out and skilfully overpainted so that the Annunciation could pass as a complete picture. This was done at an unknown date but before 1906, when the Annunciation was first discussed and reproduced. It then belonged to the dealer Dowdeswell in London.¹⁸ It is said to have come from a private collection in Paris.¹⁹

The head and most of the body of the scraped‐out figure were in the lost central section. The shape of the scraped‐out area, however, some underdrawn lines and fragments of ultramarine paint near the left edge suggest that the figure was a man dressed in long blue robes. As in parts of NG 4092, the [page 764]blue drapery was underpainted with a layer containing red‐brown earth pigment. It is generally assumed that the man was Saint Louis (1215–1270, King of France as Louis IX from 1226), who was sometimes paired with Charlemagne, for example in the Retable of the Paris Parlement or in the enamel wings in the Wallace Collection (fig. 2). As the enamels are very much in the style of the Master of Moulins, the enamel Saint Louis may give some idea of the missing figure from the painting.

As the painting would have been symmetrical, the unpainted area at the upper right edge of NG 4092 would have had its counterpart in the missing section. The unpainted area indicates where an applied part of the frame was once attached: it was covered with ground at the same time as the rest of the frame and the flat surface to be painted. The lost section of applied frame would have been about 10 cm to the left of the left edge of the Chicago Annunciation and may have intersected the right arm of Saint Louis just as the applied frame on the left intruded across Charlemagne’s left arm. The two elements of applied framing would have been continuous with the real frame as it surrounded a projecting upper section above the central area of the painting. It is sometimes assumed that this upper section was rectangular or square; but an arched, semicircular projection would have agreed better with the emphatic arches of the Golden Gate, the alcove behind the angel in the Annunciation and the arch‐topped picture of the Head of Christ that hangs in the curtains of the Virgin’s bed. Altarpieces of that shape are seen for instance in the underdrawing of the Mass of Saint Giles (NG 4681: pp. 777–807) and in miniatures by Simon Marmion: for example the miniature of the Life of Saint Louis in the ‘Fleur des histoires’, Brussels, KBR , MS 9232, fol. 388.²⁰ The central section surmounted by the arch would have been about 40 cm wide. If the raised central section was indeed semicircular, the radius would have been about 20 cm and the [page 765]painting would have been about 92 cm high at its central and tallest point. One more horizontal board would have been sufficient to make the projection with the edges needed to support its framing structures.

The missing section would have included the rest of the parapet and the brightly coloured cloth of honour behind Charlemagne, his left hand, which probably carried a sceptre, and the greater part of the figure of Saint Louis. The grassy area on which Charlemagne stands may well have continued across the missing section to join the original area of greensward in the lower left corner of the Annunciation.²¹ The central section would seem not to have been enclosed; the landscape in NG 4092 probably continued behind the cloth of honour and the parapet.

The figure or figures in the central space must have related in terms of subject to NG 4092, where the meeting at the Golden Gate refers to the conception of the Virgin, and the Annunciation, representing the conception of Christ. The Assumption of the Virgin is a possible subject,²² but it would not fit well chronologically or spatially. A more logical solution would be an enthroned Saint Anne with the Virgin as a young girl. A limestone sculpture of Saint Anne teaching the Virgin (Louvre: fig. 14) may give some idea of the missing section of the composition. The sculpture is one of three slightly over life‐size statues discovered in 1845 under the stone floor of a church situated within the château of the Dukes of Bourbon at Chantelle (Allier: south west of Moulins and north west of Vichy). The other statues represent Saint Peter and Saint Susanna. They may be by Jean Guilhomet or Guillaumet, called Jean de Chartres, who, according to Lorentz, worked from designs by the Master of Moulins. They are thought to have been made for the chapel dedicated to Saint Peter and built around 1500–3 against the north side of the church; it was demolished in 1633. The three statues represent the patron saints of Pierre II of Bourbon, his wife Anne of France and their only child Suzanne; they are believed to have been commissioned for the chapel by Pierre II and Anne, who liked Chantelle. Indeed Anne was to die there in 1522.²³

As Martha Wolff has observed, the missing figures may have been echoed in an enamel of Saint Anne instructing the Virgin (The Cleveland Museum of Art: fig. 16).²⁴ It is related in technique and style to the Wallace Collection enamels referred to above, and in style to the statue from Chantelle just mentioned and to the paintings of the Master of Moulins.

Saint Anne could have been seated on a throne in front of the cloth of honour; she and the young Virgin could have been reversals of the enamel or the stone figures, so that the hems of their robes might have trailed across next to Charlemagne’s feet. The browns and purple‐browns in that area could have been parts of a fur or textile lining of the Virgin’s robes (fig. 9), which would almost certainly have been bright blue, like the Virgin’s dress in the Chicago Annunciation. The composition could have been surmounted by angels or perhaps a vision of God or Christ. A portrait of Christ as a mature man appears, surprisingly, in the Annunciation. Whether or not the subject was Saint Anne, the central figure must have been dressed in brilliant red in order to stand out against the richly patterned cloth of honour, and in order to balance the largely scarlet figure of Saint Anne at the Golden Gate, the red mantle of the Virgin and the duller red bed curtains and bedspread in the Annunciation.

Attribution

Friedländer in 1925 published NG 4092 as the work of the Master of Moulins and made an association with the Chicago Annunciation,²⁵ itself published in 1906 by Roger Fry as the work of the Master.²⁶ Friedländer’s attribution has been accepted by subsequent scholars with the notable exception of Martin Davies, who, in 1946 and 1957, stated that NG 4092, ‘taken alone might be supposed a follower’s work, the Charlemagne in particular being hard and coarse’.²⁷ He agreed that NG 4092 and the Chicago Annunciation were probably the extremities of one panel and that the Annunciation ‘is apparently autograph’; he therefore catalogued NG 4092 as from the ‘Studio of the Master of Moulins’. Since the picture was cleaned in 1989–90, however, it has been regarded as the work of the Master himself, without much intervention by assistants. The equation of the Master with Jean Hey is gaining increasing support.²⁸

The Patron, the Original Setting and the Date

The presence of Charlemagne, the possible inclusion of Saint Louis and the prominence accorded to Saint Anne suggest that the patron may have been Anne of France (also known as Anne of Beaujeu), Duchess of Bourbon, wife of Pierre II, Duke of Bourbon. The Bourbons were interested in the Immaculate Conception and Anne, who was born in 1461 and married in 1473, did not give birth to her only surviving child Suzanne until 1491, when she was thirty and had been married for more than seventeen years. Anne may have regarded her patron saint with particular affection if she had prayed for her help in conceiving. Lorentz and Regond have suggested that the London and Chicago pictures formed part of an ex-voto offered in gratitude for the birth of Suzanne on 10 May 1491.²⁹ It could have been painted for the altar of the Immaculate Conception in the Collegiate Church of Our Lady at Moulins (now the Cathedral): this altar had been endowed in 1474 by Jean II, Duke of Bourbon, an older brother of Pierre II.³⁰

The style is close to that of the illumination attributed to the Master of Moulins showing Saint Michael appearing to Charles VIII from a manuscript of the Statutes of the Order of Saint Michael (Paris, BNF , MS fr. 14363, fol. 1), given by Pierre II of Bourbon to Charles VIII, perhaps during the winter of 1493–4.³¹

Albert Châtelet, Susan Foister, Pierre‐Gilles Girault, Peter Klein, Andrew Meek, Ashok Roy, David Saunders, Jeremy Warren, Faye Wrubel and Martin Wyld have most generously helped with this entry. I am particularly grateful to Martha Wolff, without whose assistance I could not have completed it.

Notes

Infrared reflectography was carried out using a Hamamatsu vidicon. For details see p. 18, note 22.