his drawing is sometimes meagre and wiry. The single quality in his work that never varies is the workmanship. This is always wonderful, even when carrying out the least pleasing aberrations of his genius.

It seems that no other painting on parchment by him is known to exist, and the use of parchment as a ground for painting appears to be extremely rare. This use of an uncommon material in itself is surely more like the experiment of a great master than the mistake of a copyist, who would naturally employ a ground like that of the original, especially if it was in everyday use, as prepared panels then were. If an original picture were an elaborate piece of painting, as all Dürer portraits are, it is hard to imagine that any copyist would run the risk of wasting a large amount of time and labour by experimenting with unusual materials, when those used by the original painter were ready to his

hand.

If, however, the treatment of the head be compared with that of a Dürer drawing" in the British Museum (supposed to have been done some four years later), there can be little doubt as to the identity of their authorship. It is only necessary to point out the tremulous suggestion of the wrinkles and veins round the eyes, the drawing of the eye sockets, the treatment of the nose, and the emphatic statement of the furrowed flesh about the jaw in both works.

The details of the cloak can be studied only at Trafalgar Square, for the photogravure gives no idea of the transparency and lightness of the picture. Attention has already been called to the force and impetuosity of the treatment of the left sleeve. The right sleeve is equally inter

6 Mr. Herbert Horne informs me that there is no known instance of an Italian panel picture on a parchment ground earlier than 1550. Mrs. Herringham (who considers that there may be a thin gesso ground under the Dürer picture) is equally definite on this point.

7 Reproduced on page 437.

esting, for if examined closely it will be seen to have an underpainting in monochrome, done with rapid and accurate brush strokes in the exact manner of Dürer. This should be compared with the sleeves in the Uffizi portrait (1490) and in that of Oswolt Krel at Munich (1499). The treatment of the hand should be compared with the right hand of Imhof in the Prado. The prominence at the root of the fingernails (as Mr. Charles Ricketts pointed out to me) is repeated in Dürer's portrait of himself once in the Felix collection at Leipzig.

The remaining objections to the work are based on the inscription. It must be at once admitted that this may not be from Dürer's hand, as the style of lettering differs slightly from that on the Oswolt Krel (where Dürer for once uses Roman capitals), and there is no warrant for the spelling 'Thurer.' For this reason Mr. Dodgson, following Dr. Friedländer, points out that the inscription on the Munich picture has a more genuine ring about its wording, and corresponds absolutely in style with that of the Prado portrait. He recognizes that the Munich picture is a bad copy, but suggests that its lettering represents that on Dürer's original.

Yet if this original were so lettered, why did Greenbury, the reputed painter of the Syon House copy, give an entirely different and far less convincing wording, and why did the German painter who painted the older copy at Frankfort do just the same? If the original picture had borne Dürer's monogram and an interesting gothic inscription, it is incredible that these two painters, working at different times and in different countries, should have omitted it, and agreed in substituting another and less obvious one.

The original work must thus have borne an inscription practically identical with that

8 Even this objection, however, now seems to be groundless. Mr. Dodgson has recently sent me an extract from a letter of Pirkheimer to Conrad Celtis dated March 14, 1504.- Turer te salutat.'

now upon the National Gallery picture. This inscription was evidently added some time after the portrait was executed, when the background had thoroughly dried, and runs thus:

1497 ALBRECHT. THURER. DER. ELTER VND. ALT. 70 JOR.

When I first copied the inscription I wrote the date as 1494, since the 7 was made exactly the same size and shape as the 4 preceding it, except for the horizontal crossbar. The difference even in a strong light was so small that I overlooked it until it was pointed out by a friend who was with me. The mistake seemed to explain the reason for the date 1494 in the Frankfort copy.

It immediately occurred to me to test the Syon House inscription in the same way, where VI ID. AET is substituted for VND. ALT., and at once the cause of the mistake became apparent. The crossbar of the N in the National Gallery picture has sunk into the ground till it is almost invisible, so that VND has become VI ID. The Syon House copyist has read it thus, and presuming perhaps that 'ID' stood for idem,' read the next word also as Latin—‘aet' for 'aetatis.' If the original frame came close to the picture, as I have previously suggested, the top of the L would be invisible, and the mistake under the circumstances a very natural one. These two slips certainly seem to indicate that the National Gallery picture was the original of the Frankfort and Syon House copies, and, coupled with its general correspondence to the inventory of 1637, show that it can hardly fail to be the picture which was once in Charles the First's collection.

The history of the picture may thus be somewhat as follows:-It was painted by Dürer in 1497, but for some reason or

other it was not carried to a high degree of finish. After his death it remained in Nuremberg, and was copied carefully by the painter of the Frankfort version, and less skilfully by the painter of the Munich version, who forged an inscription and signature to match that on the portrait of Dürer himself now in the Prado. With the Prado picture it was presented (after some retouching) by Nuremberg to Charles I. Then the portrait was copied by Greenbury (if he was indeed the author of the Syon House version) and catalogued by Van der Doort. When the royal collection was dispersed the damaged condition of the picture was repaired by its new owner, who had it mounted on a sound panel, and retouched again. A label on the back indicates that it was in some English collection towards the end of the eighteenth century (where perhaps it was numbered 208), before it passed into that of Lady Ashburton.

The available external evidence thus all seems to indicate that this picture is the original of the three other versions of the subject, and is identical with the picture in Charles the First's collection, which was presented to him by Dürer's native city, together with another magnificent work by that city's greatest master. It is hard to believe that a copy or a forgery would have been sent under such circumstances, even if the picture were not in itself one of the most lively and emphatic specimens of German portrait painting which the nation has hitherto acquired.

This retouching has made the background redder than that of the other versions. A comparison of the Frankfort and Syon House versions seems to indicate that the high lights and darker lines on the hair were added before the portrait left Nuremberg. The under painting of the hair is typical of Dürer's method. The surface, too, has everywhere been rubbed in cleaning so that the original work is often blurred, as the characteristic high lights on the bridge of the nose show. That the repainting was necessary may be judged from the fact that the forehead had actually cracked right away from the cap.