However, the figures in the central lights in Saint-Remi are slightly taller than those made for the regular single lancets of Braine, so the cartoon would have had to be adapted for the lower panels. Unfortunately we cannot know how closely the rest of Abiud followed the design for David because it is missing. The nineteenth-century restorer who reassembled the Braine glass in Soissons, either before installing it there of before putting it on the market, supplied Abiud with a mix-and-match body and lower extremities from another figure in the series! But, the lower parts of the king David in Reims and of the ancestor of Christ, Amminadab, that comes from Braine have much in common: For each one, the green lines follow the side borders (albeit with an extra fillet in Reims), one leg, and the contour of the throne with adjacent skirt.  A simple horizontal tuck has shortened Amminadab’s shins:

Fig. 9a.  Saint-Remi, N.IVb, lower part of David.


Fig. 9b. Braine (actually in the Cathedral of Soissons), lower part of Amminadab.

This is also a case of the second type, since the legs and skirts of David were adjusted to fill the width of the window without the red fillet.  The width of the throne is increased for Amminadab. Furthermore, the distance between the knees of Amminadab is greater than with David, and the lower legs are angled slightly differently.  The shaded areas on Amminadab show the alterations:  it was simple to let in a vertical strip, but a triangular dart was required to change the angle of the legs. Marked differences in the rendering of draperies and throne almost mask the very close relationship between these designs.  The figure from Braine is more monumental, and the perspectival treatment of the footstool has been eliminated. Such traits confirm the later date of Amminadab. 


Further adaptations were made to produce the series of figures from Braine. For example, the throne back and the chin of Abiud were repeated for Amminadab (in orange here), even though the latter’s head is in three-quarter view:


Fig. 10a.  Braine, upper part of Abiud.


Fig. 10b.  Braine, upper part of Amminadab, with countors from Abiud superimposed.


The center and lower panels of Amminadab were adapted from the panels now attached to Abiud. The orange lines show the extent of the duplication, and the purple indicates that the broad outlines of the drapery on the upper part of the right coincides, even though the position of the hands is changed. This selective reuse is of the third type. The adaptation is clumsy, in that Amminadab’s right forearm appears disjointed (we will see why below):


Fig. 11a.  Braine, Amminadab, lower and center panels.

Fig. 11b.  Braine, lower and center panels mounted with the head of Abiud.


These changes, and a completely different color scheme, mask the partial repetition of the cartoon.  Another modification is that the angle of Amminadab’s legs – already modified from the cartoon for David in Reims – is altered by taking out two triangular darts to produce the more elegantly draped lower part attached to Abiud (the yellow lines are otherwise identical).


The pair of lateral lancets that flank the central ones of the type with King David in Saint-Remi are shorter and narrower, as in Fig. 4.  Each figure that fills the lateral lights is composed of two instead of three panels of glass, mounted on the horizontal bars of the armature.  Yet the overall dimensions of these lateral figures are close enough to those at Braine for the cartoons to be adapted for reuse. The design for one of the ancestors of Christ from Braine, Amminadab, corresponds to that of David’s companion Micheas, but it has been reversed except for the head: These figures are identical in the basic outlines of the haloes, heads, faces and necks (in yellow), and the shoulders and upper arms, left shins, and the rise of the footstools are the same reversed (in purple):


Fig. 12a.  Saint-Remi, choir N.IVa (left lancet), Micheas. 


Fig. 12b.  Braine, Amminadab,  reversed to show that the design for Micheas was flipped over, except for the head.


But the cartoon had to be lengthened a few inches for Amminadab, and this was done by letting in a horizontal strip at the figure’s waist level, essentially adding to the top of the lower glazing panel of Micheas. The forearm positions were redrawn to disguise this, but Amminadab’s left hand and arm are oddly disproportionate because the hand has to reach down to the drapery fold that Micheas held effortlessly by resting his right hand on his knee. Amminadab’s right leg was inexplicably shortened, and it was thinned down by using a drapery fold in the Rémois design as a contour; this indicates that the cartoon was marked for painted detail as well as for cutting the glass. A throne back for Amminadab helps to fill the wider window opening; its definition coincides with the limit of the blue ground behind Micheas.  Amminadab’s throne and footstool are extended to the left in the lower panel for the same reason.


The cartoon for Micheas also served, but in the same sense this time, for Naashon in Canterbury. The horizontal armature falls in the same place, and the template for the window head in the Trinity Chapel at Canterbury was adjusted to approximate that of the lateral lights in Saint-Remi (this represents a dramatic change at Canterbury, in that the masons had to step up the roof level to accommodate the more pointed window openings for the rest of the series to the east!):


Fig. 13a.  Saint-Remi, Micheas.


Fig 13b.  Canterbury, Trinity Chapel clerestory (N.X), lower figure, Naashon.


Fig. 13c.  Photograph of Nashon.

The overall dimensions of the figure had to be reduced for Naashon, however, because of a shorter window and a wider ornamental border at Canterbury.  A horizontal tuck taken out at ankle height led to shortening Micheas’s robe by removing a false hem! The horizontal bar was shifted down slightly.  In the upper part, a dart taken in the center altered the angle of the upper arms, causing the hands to cross in Canterbury. The result is that Naashon’s right hand no longer rests on his knee but holds his mantle in a rather mannered gesture.


Naashon’s companion, Amminadab, in Canterbury shares several contours with one of the Frankish kings made for the clerestory lancets in the nave of Saint-Remi. The basic outlines coincide, except that the shins of Amminadab have been shortened by a horizontal tuck, the angle of his left upper arm adjusted by a dart, and the edge of the blue ground is tilted to outline a sceptre:


Fig. 14a. Saint-Remi, nave clerestory (N.XXIV), lower figure, Frankish king.


Fig. 14b. Canterbury, Trinity Chapel N.X, upper figure, Amminadab.

The odd-looking gesture of the king’s left hand results from a dart taken in the chest of Naashon (drawn on the king) that eliminated his foreshortened forearm. And this time it is the Rémois figure that grasps a fold in his mantle instead of making the speaking gesture of the Canterbury figure.  In general the execution of this series in Reims is much coarser than the figure paintings done for Canterbury, or even Braine. It looks very much as if the design for Canterbury’s Amminadab (itself probably derived, as we have seen, from a figure in the retrochoir of Saint-Remi) returned to Reims with the glaziers who adapted it for the nave window.


The ancestors painted for the windows further east in the Trinity Chapel at Canterbury are much smaller than Naashan and Amminadab.  Inexplicably, despite the lancet shape recently imported form France, the designer decided to fill each window with varied geometric frames such as lozenges or quatrefoils, thus drastically reducing the size of the seated figures placed in them.  When the design used for Jesse at Canterbury was adapted for King Chilperic in the nave of Saint-Remi, the Canterbury figure had to be enlarged to give the needed dimensions:


Fig. 15a. Canterbury, Jesse, from the Trinity Chapel cleresztory (N.VIII).


Fig. 15b. Pattern enlarged for King Chilperic, Saint-Remi, nave S. XXIV.

Jesse’s right hand rests on his knee in a dramatic pose, arm akimbo, while his left hand holds a large scroll. Chilperic, on the other hand, does not hold his scepter securely in his similarly positioned right hand, and the restorer did not know how to give him a coherent left hand without a scroll.  Triangular pieces were let in to adjust the angle of the right arm in relation to the vertical edge of the mantle, and to widen the legs.  


Despite the great difference in size within the Trinity Chapel series at Canterbury, the pattern for Amminadab was adapted for his small neighbor, Obed.  This time only isolated elements could be reused, so the close relationship is almost imperceptible:


Fig. 16b.  Canterbury, Amminadab, photograph.


Fig. 16b. Canterbury, Obed, photograph.

Fig. 17a. Canterbury, Amminadab.



Fig. 17b. Canterbury, Obed, fromTrinity Chapel N.VIII, design pieced together from elements of Amminadab.


A triangular insert has separated the hands and changed the angles of the shoulders, but Obed is otherwise much thinner.  His face and hands appear disproportionately large, especially since the abdominal area is reduced in height (his left knee raised up to his hand).  A changed color scheme completes the camouflage, The only immediate sign that we are dealing with the same painter is the very distinctive rendering of the frowning face and the sinews of the hand:


The pattern for Amminadab in Canterbury is also closely related to that of ‘MELEA’ from Braine (which has an entirely new lower panel). Like the last example, the adaptation is very complex, though without the necessity for a major change in size.  Three sections of the mid-part are adjusted with tucks and darts – or more probably, pieces were let in to achieve the bulkier figure of Melea:


Fig. 18a.  Canterbury, Amminadab.


Fig. 18b.  Braine, an ancestor of Christ (?Melea) ; the lower panel is modern, but the mid-part may be adapted from the pattern for Amminadad.


The pattern appears dismembered and reassembled.  But there is another possibility. We have seen that Abiud from Braine is a replica of King David in Reims, at least in major contours (Figs. 8a&b), and that Naashon in Cantorbéry (companion of Amminadab) depends on Micheas in Reims (Figs. 12a&b).  So I hypothesize that the same Rémois pattern was used for « Melea » in Braine (without changing the size), and for Amminadab in Canterbury, and that this figure is among those missing from the hemicycle of the choir of Saint-Remi.  



Following these observations, I am in a better position to decide what carrier the glass designers had used. What is required is a material that is reversible, pliable so that it could be taken in or let out to change the size of the design, and that eventually lent itself to stitching together an assemblage of parts. My one-time professional engagement with dress-making led me quickly to the supposition that this must be a fabric; after all it is perfectly normal to adjust the fit of a garment by using tucks and darts, or by  letting in similar pieces. The contours (cut lines) could have been drawn on fabric with charcoal, and reinforced in order to transfer them to the glass painter’s tabletop. Turning the pattern face-down would reverse the design, as with Amminadab from Braine in relation to Micheas in Reims (Figs. 12a&b).

Or, to keep the design in the same sense – which is important if an archbishop is to keep his crosier in his right hand – it is easy to lay the cartoon on a surface that has been dusted with soot or charcoal powder, and trace over the lines firmly. They are then visible on the reverse side of the fabric, and can be transferred to the tabletop by tracing them one more time.  Or, a general advantage of fabric is that it becomes transparent when waxed, and oiled linen has been mentioned in eighth-century accounts in relation to the glazing of windows in York Minster. And by the late fifteenth century painters were used to working on linen:  Records indicate that good-quality linen was purchased for the painter Henritz Heyl in 1476 in order to make full-size cartoons for the glass painter Konrad Rule.

The examples demonstrated in this paper show an extraordinarily fluid adaptation of patterns by an atelier working in three different sites over a period of twenty years.  The preferred method of figural composition within this atelier involved repetition of elements in various combinations, without any concern for the proportion or volume of the figures.  In other words, the essentials of mimetic figure-drawing as practiced in other periods are completely lacking. The aesthetic of these “Gothic” assemblages seems post-modern to us now, but it was understood by the glass-painter-restorers of the nineteenth century, who extended the practice of recombining parts of different figures by mis-matching the actual panels when they put them on the market. In effect, the color scheme and the painted detail that gave character to the medieval figures – something we might recognize as artistic interpretation – masked the predilection for re-using cartoons.


Click here to return to previous page