Drawings and Watercolours

Dürer’s drawings invite us to change the way we look. A face can emerge from fine grey lines; a wing can fill a sheet with colour; a few marks can gather people into a story. Sometimes the smallest passage demands the closest attention. Sometimes the interest lies in an unfinished arrangement, where one form has been explored carefully and another remains barely indicated. The variety makes drawing an especially revealing way into his art.

A mirror, training and tools

Charles Ephrussi’s 1882 study already described the ALBERTINA’s silverpoint Self-Portrait at Thirteen of 1484. He transcribed the later mirror inscription and noticed the contrast between the carefully handled face and the less natural pointing hand. His account is useful partly because it shows how quickly description becomes judgement. He praised the young artist’s individual grasp of form, yet compared the hand to a borrowed saint’s gesture. We can take those as a nineteenth-century reader’s responses, without making the drawing prove a future career. The added inscription, too, looks back: it records an older artist identifying work made in childhood, rather than a child announcing what he would become. Winkler already observed revised contours around the cap, sleeve and fingertips. ^{[1] [4]}

Dürer’s own account of his education gives the sheet a practical setting. In the family recollection printed in William Martin Conway’s 1889 Literary Remains , he says that his father taught him goldsmithing after he had learnt to read and write. When he could work neatly, he asked to become a painter. His father regretted the earlier training time but agreed, binding him to Michael Wolgemut for three years in 1486. Drawing belonged within this movement between crafts, before the young man had established a painter’s workshop of his own. ^[5]

The older Italian manual of Cennino Cennini offers a useful account of the discipline such tools required. In Mary P. Merrifield’s 1844 translation, Cennini begins with a prepared tablet and a finely polished silver point. The pupil should draw almost invisibly at first, then repeat strokes to deepen the shadow, leaving the lights more lightly touched. He recommends a little work every day. This is historical teaching context, not evidence that Dürer read Cennini or followed his recipe. It explains how a fine, apparently even line could support gradual decisions about darkness and shape. ^[6]

The Metropolitan Museum’s double-sided sheet of 1493 combines a self-portrait, a left hand and a pillow; six pillow studies occupy the reverse. It is drawn in pen and iron gall ink. Closely repeated strokes build shadows while leaving paper visible between them. This hatching makes soft cloth legible through an accumulation of separate lines. ^[7]

Cennini’s instructions also make pen drawing an acquired skill. He advises beginning lightly and returning repeatedly over the shadows, with a little diluted colour added by brush if desired. His next chapter explains cutting a goose quill and making its point broad or fine according to need. The pen’s mark therefore depended partly on a tool the artist shaped. The manual links this practice to the ability to invent designs: repetition trained the hand for more than copying an outline. ^[6]

Pictures, clothes and acts of copying

Copying could be an active encounter with another maker’s work. Ephrussi discussed Dürer’s pen copies of Mantegna’s Bacchanal and Battle of the Sea Gods together. He thought the figures had become more prominent against their surroundings and that the engraved, sculptural effect had changed in the drawings. This is his comparison, not a description of an automatic transfer. It asks us to look at what happens between media even when the arrangement remains close. A design can circulate without its maker being present; copying an Italian print therefore cannot establish that Dürer visited Mantegna. The Bacchanal with Silenus in the catalogue below is the same pen copy. Winkler described the rounded, crossing strokes, but treated tracing as established. Nuremberg remains possible; tracing is unproved. ^{[1] [4] [8] [9]}

The Venetian Woman of 1495 rewards that same attention to what a drawing selects. Ephrussi described the richly patterned underskirt exposed by lifted fabric, the jewellery and hair, and the much smaller rear view. His account demonstrates an enduring attraction of the sheet: it lets a reader compare different parts of a costume, and different degrees of finish, within one rectangle. He also linked the costume with the woman riding the beast in Dürer’s Apocalypse . That is a proposed connection between images, not a social identity for the person who wore the clothes. The descriptive evidence and the identity question should remain distinct. Vecellio’s 1590 costume book cannot establish whether she was a noblewoman, courtesan or bride. ^{[1] [8] [10]}

In the front view, the patterned underskirt and the heavier hanging fabric create two rhythms. The exposed pattern opens out over broad areas; the raised cloth gathers into narrow dark channels. Jewellery, hair and the joined sleeve sections introduce smaller units above. Beside this densely worked figure, the much lighter rear view lets long folds run almost uninterrupted towards the hem. These unequal degrees of finish encourage us to compare the construction of the clothes, rather than treat the sheet as one uniformly completed portrait.

A place observed and arranged

Innsbruck Seen from the North is a watercolour with opaque colour and white heightening, only 12.7 by 18.7 centimetres. The ALBERTINA describes an expanded river, a compressed skyline and probable separate painting phases, with the upper sky added last. These are interpretations of the object’s making, not conclusions obtainable from a small reproduction alone. Its current label gives about 1495; the extended architectural essay instead argues for a return journey around 1496–97, using the Wappenturm and related courtyard views. The dating should retain that qualification. The panorama’s specificity cannot make it an automatically dated, single-sitting record of Dürer’s first journey towards Italy. ^[11]

The river gives the towers an unusually long approach. Small touches of warm colour separate roofs within the pale architecture, while the water breaks their reflections into narrow horizontal bands. Above, the clouds occupy a large open field. At this scale, the eye continually moves between the town as one silhouette and the tiny differences that make its buildings legible. Looking at the watercolour slowly changes its apparent size: what first seems a compact view contains many separate opportunities for attention.

Animals, plants and convincing appearances

The Young Hare of 1502 was already central to Ephrussi’s account of the nature studies. He praised the alert eye, erect ears and soft fur, but then imagined an effortless, single burst of execution. His own vivid response cannot document how the sheet was made. That gap between a convincing animal and a recoverable working process matters: an image can give a powerful impression of life while leaving the model’s actual circumstances unknown. The window reflection cannot prove continuous live modelling. The ALBERTINA’s expressly credited 2014 collecting history traces the hare through the Imhoff and imperial collections, the Court Library in 1783 and Duke Albert’s collection in the 1796 exchange of prints for drawings. ^{[1] [8] [12] [13]}

The Great Piece of Turf , dated 1503, also appears in Ephrussi’s account of the ALBERTINA’s nature studies. He grouped the herbs and small plants with highly finished animal pictures, rather than making drawing mean a rapid outline alone. His account helps recover a long history of close attention to these sheets. It does not establish that everything pictured grew together in front of Dürer. Small studies can be persuasive pictures, and the appearance of growth can coexist with deliberate arrangement. Workshop assembly from separate plant studies is proposed but undocumented. ^{[1] [8] [4]}

The Wing of a European Roller also shows why an old description needs checking. Ephrussi listed the dead blue bird and its separate wing under 1512. His catalogue is evidence that the inscription was being read as a date in the nineteenth century; it cannot settle the date of execution. The NGA’s 2013 brochure describes watercolour and gouache on vellum, heightened with white. Dated about 1500 on fine, smoothed, polished parchment, the wing bears 1512, probably added by another hand; direct reuse in Dürer’s figures is unproved. ^{[1] [2] [8] [4]}

Berlin’s Lobster , dated 1495, broadens the material range with pen, wash and white heightening. The 1999 catalogue From Schongauer to Holbein notes its abraded, discoloured surface and treats it as study material with a pictorial character. ^[14]

Christof Metzger’s 2019 ALBERTINA exhibition interpreted major nature studies as virtuoso demonstrations for workshop visitors. The proposed audience adds a possible purpose to their close description. The argument concerns what these sheets could do; finish alone documents neither a particular display nor the presence of a prospective patron. ^[3]

Margaret Doyle’s 2013 National Gallery of Art brochure likewise stresses that nature studies were not necessarily preparations for other compositions. The category of drawing includes highly finished pictures as well as working sketches. Watercolour, covering paint and a fine brush could serve both kinds of purpose. ^[2]

Coloured grounds, fragments and copies

The museum’s 2025 exhibition distinguishes Venetian carta azzurra , coloured during manufacture, from grounds applied to sheets in Nuremberg. Similar colours can conceal different material structures. Curators Achim Gnann and Christof Metzger also interpret finely finished detail studies as an archive of analytical observation and workshop showpieces. This attributed view questions their immediate preparatory usefulness while leaving their identifiable relationships to paintings in place. ^[3]

Cennini describes the principle of working from a middle tone with unusual clarity. The ground supplies one part of the tonal range; ink darkens the shadows and white adds the lights. In chapter 31 he recommends gradually deepening diluted ink, then using white both broadly and with a pointed brush for the brightest passages. He ends by reinforcing small dark details. His account helps explain the terms “wash” and “heightening”: a wash spreads diluted colour over an area, while heightening adds light. These are general operations described in an earlier manual; the passage cannot identify the actual pigments or working sequence of an individual Dürer sheet. ^[6]

Ephrussi recorded a hand holding a book among the studies for Christ among the Doctors . He placed the brush drawings with the work of 1506, identifying a relationship in paper and handling with the Feast of the Rosary studies: blue-grey paper and added white belonged to that group’s way of modelling. His catalogue also distinguished individual sheets from drawings that still shared a support. That distinction remains useful when we approach a surviving fragment. A rectangle containing one detail can look complete to a museum visitor while preserving only part of an earlier arrangement. Albert’s inventory pairs Hand with Book and Two Hand Studies ; cutting date and commercial motives remain speculative. ^{[1] [8]}

The Praying Hands belong to the Heller Altarpiece’s kneeling apostle. Ephrussi’s list assigns the head, joined hands and sleeve to the same figure. He describes the studies as brushwork on prepared blue-grey or greenish paper, with lights added in white. His list makes the body present through separate pieces: head, hands and clothing were studied at different scales and now invite different kinds of attention. In a footnote he records a smaller Dresden copy of the hands, on green prepared paper, bearing S. B. and 1576; he also reports disagreement over its attribution. This is historical evidence of copying, not proof of the original sheet’s cutting date. Winkler’s 1937 catalogue describes the 1508 hands as brushwork with white on blue-prepared paper. He identifies them as the right half of the head study, using a continuing shoulder-line and a shared fold to reconstruct their relationship. Missing strips matter: today’s rectangles cannot simply be pushed together to recover the original sheet. Opaque white heightens grey and black wash; its cutting date is unknown. Ephrussi identified Jobst Harrich as the painter of the surviving copy. That substitute preserves a broad composition, rather than Dürer’s actual painted surface. Wölfflin’s 1908 book corrects an earlier 1674 date to 1729. The centre was destroyed in 1729 by fire. ^{[1] [4] [8] [15]}

Three Drapery Studies keeps different portions of an apostle’s clothing together. Ephrussi’s Heller list describes the sleeve and part of the same apostle’s mantle immediately after the joined hands. Read alongside his general account of brushwork and white on coloured grounds, that entry helps us treat the clothing as an investigation in its own right. The pieces need not form a miniature complete body to be useful to a painter. Ephrussi nevertheless made the sweeping claim that the surviving studies had been used unchanged in the painting. The lost original makes such certainty especially difficult to sustain. Green lies over pink; Harrich’s different arm and knee lighting cannot securely establish original changes. ^{[1] [8]}

Dürer’s letters to Heller bring another kind of evidence to this exceptionally careful work. On 24 August 1508 he described the central panel as outlined with great care and at considerable cost of time. In November he resisted a demand for the greatest conceivable degree of finish: a single face, he protested, might occupy him for half a year, and such minuteness across an altarpiece would scarcely be visible. His complaint distinguishes effort from what a viewer could actually see. Read as a painter negotiating with a patron, it gives substance to the practical problem of how much attention one part of a larger image required. ^[5]

Ephrussi’s Heller inventory offers a further lesson in what a catalogue can preserve. A head, a pair of hands and a sleeve were already separate entries in 1882, while a copy could place the hands beside another detail on a single support. That arrangement belongs to the copyist’s sheet. It cannot be carried back automatically into Dürer’s original. Ephrussi’s descriptions help distinguish a surviving drawing, a recorded copy and a reconstructed painted figure: three kinds of evidence whose relationships need to be established rather than assumed. ^[1]

Charcoal and an individual presence

Berlin’s portrait of Barbara Dürer is dated 1514: charcoal on paper, 42.2 by 30.6 centimetres. Michael Roth’s catalogue account describes the concentrated head, neck and chest giving way to looser clothing near the edges. Barbara had lived in her son’s household since 1504 and had suffered a prolonged illness; the portrait preceded her death in May 1514 by about two months. Those documentary circumstances deserve attention without turning the drawing into a clinical record. The emphasis on age and bodily particularity supplies neither a diagnosis nor evidence that her son’s sympathy depended upon softening what he saw. The paper preserves a specific encounter with a specific woman. ^[16]

Across the forehead, thin wrinkles run between much broader curves of cloth. The line of the mouth is small and compressed; the head covering opens out behind the neck into a large pale fold. The shoulders have only a slight presence below. These changes of scale hold the face near us while allowing the surrounding sheet to feel comparatively open. The drawing’s force can be felt before its inscription or family history has been read.

Drawing within a painting and towards a print

In the Metropolitan Museum’s unfinished Salvator Mundi , about 1504–05, exposed and infrared-revealed underdrawing shows modelling with pen and brush. Ainsworth’s catalogue records later completion paint and its removal. The visible state has undergone interventions; the reason Dürer stopped remains unknown. ^[7]

Paper could also establish a design that changed as it became a print. Doyle compares the Visitation drawing of 1503–04 with the resulting woodcut: the central encounter remains, but architectural surfaces gain detail and the landscape changes. The printed composition reverses the drawing’s orientation. Preparation established relationships while leaving room for further decisions; a finished print need not reproduce every feature of its preliminary sheet. ^[2]

A travel book and drawings exchanged

A dispersed silverpoint travel book accompanied the Netherlandish journey of 1520–21. One leaf, Berlin’s KdZ 33, measures 12.1 by 17.1 centimetres: an Antwerp portrait and later Andernach landscape occupy the front, lion studies the back. Hans Mielke proposed Aert Praun as the sitter and associated the lions with Ghent, interpreting them as two views of one animal. These remain proposals, not identities proved by the sheet alone. ^[14]

The diary provides a separate record of drawing among the journey’s encounters. At Ghent, Dürer says he saw lions and drew one with a metal point. He also records making a metalpoint portrait of Jan Provoost before leaving Bruges. The entries place looking at animals and portraying fellow artists within the same travelling practice. ^[5]

Back at Antwerp, drawing was also a way of doing something for another person. Dürer recorded making Hans Lieber’s portrait in charcoal and refusing the offered payment. Elsewhere he accepted money for a portrait, received an alabaster bowl after designing three sword hilts, and recorded a goldsmith’s payment covering a seal design and two likenesses. He listed portraits and Netherlandish costumes in black and white on grey paper, alongside narrative scenes on half-sheets. The variety belongs to his own account of his activity; it need not be inferred from the handful of surviving objects most familiar today. ^[5]

The diary’s compact entries resist a single division between private sketch and commercial work. A portrait could acknowledge hospitality, a design could answer a practical need, and a gift could enter an exchange without a fixed cash price. Dürer also complained that much of his work for others brought no return. These are the artist’s descriptions of particular transactions, not a complete balance sheet. They nevertheless show drawing circulating between people while its maker was alive, giving the medium a social usefulness beyond the eventual museum collection. ^[5]

Late studies and changing arrangements

For the Lisbon Saint Jerome studies of 1521, Ephrussi’s catalogue gives another kind of foothold. He describes a lectern on a table, an open book, two closed books beneath and a small box alongside. The following page dates these brush-and-white studies on violet paper to 1521. Even small accessories were available for separate scrutiny. His description lets us reconstruct what a nineteenth-century reader encountered on the sheet, without making an inventory of objects equivalent to a record of every decision in the painting. A preparatory drawing can preserve more than the finished composition eventually required. The painting strengthens forearm foreshortening, shortens books and omits the box; transfer and reasons remain unproved. ^{[1] [8]}

The Met’s Music-Making Angels of 1521 and a head study of 1522 have been associated with an unexecuted altarpiece project. No corresponding commission is documented. ^[7]

The horizontal drawings often called the Oblong Passion raise a larger question about late invention. Dana E. Cowen’s 2014 dissertation argues for their independence from the traditionally proposed unfinished print cycle, and for their usefulness as models. Her interpretation connects the widened spaces, restrained emotion and Netherlandish compositional examples with Dürer’s engagement with Reformation ideas. She also notes the substantial difference between the drawn Last Supper and the woodcut of 1523. This is a scholarly reading, not a documented programme or a consensus that every related sheet had the same purpose. It keeps their narrative choices visible instead of reducing the late drawings to abandoned steps towards prints. ^[17]

The Adoration of the Magi drawing of 1524 gives late invention a spacious setting. Ephrussi described the kneeling king before the seated Virgin and child, Joseph standing nearby, and the standing king guiding the third visitor forward. He identified the ALBERTINA sheet as pen work dated 1524. Those relationships give us a route through its gestures without confusing it with Dürer’s painting of 1504 or a woodcut of the subject. His nearby discussion of the drawn Last Supper also notices that Christ occupies a different position in the woodcut. Related compositions can change significantly between one treatment and another. Winkler noted pale ink at left, black at right and yellow on the gifts. Cycle or workshop models remain possible; yellow is probably later. ^{[1] [4] [8]}

Observation, measurement and invention

Dürer’s Four Books on Human Proportion appeared posthumously in 1528. His writings put observation beside the construction of forms. Conway’s edition illustrates a freely studied hand next to a measured one, then prints Dürer’s remark that no two fingers have the same shape. He describes fixing lengths, breadths and thicknesses before completing the outline. Measurement required the discrimination of differences: the fingers were not interchangeable units. A hand could be examined for its particular appearance and drawn again to articulate a system of relationships. These are connected uses of drawing, even though one produces an immediate likeness and the other a diagram. ^[5]

The theoretical passage on page 247 carries that connection into invention. Dürer recommends close attention to nature, allows geometry a place in demonstrating an artist’s work and insists upon experience. The imagination, he argues, becomes capable through what study has stored within it. An experienced artist can then draw without copying each individual figure from life. In this historical account, learning from the visible world does not oblige an artist to remain forever dependent on a model standing in front of him. The products of observation become resources for making further images. ^[5]

The distinction helps explain why a measured diagram and an invented figure need not oppose a portrait. They ask different things of the knowledge the artist has acquired. Proportion can make relationships explicit; recollection can release a pose or detail for another setting. Dürer’s argument is a theory of trained invention, not a claim that imagination works without preparation. The same passage belongs to an older discussion containing racial hierarchies that cannot serve as neutral standards of beauty. Its practical insight here is narrower: repeated study increases what an artist has available to draw. ^[5]

Copies, collections and using the catalogue

The drawings’ later lives complicate the name “Dürer” on a museum label or an old catalogue page. The Germanisches Nationalmuseum’s 2022 Hans Hoffmann exhibition distinguished precise copies, variations and independent compositions. Hoffmann had access to the Imhoff collection; his works might carry his own initials or Dürer’s monogram. A monogram consequently cannot establish authorship by itself, and a copy need not be dismissed as a fraudulent object. The exhibition also described his reuse of the hare in more elaborate natural settings. Repetition could preserve a celebrated model while creating a different composition, for another audience and at a later moment. ^[18]

For the selection gathered here, start with what interests you and then compare how it is made. Follow a contour across a face; notice where lines gather or separate; look for the intervals that remain light. Compare the same kind of form in another sheet. A sleeve can be a few long folds or a dense field of tiny marks. A hand can point, grip, pray or spread out its fingers. Those differences offer more useful routes through the drawings than expecting every example to perform the same task.

The catalogue is a selection, not a complete authenticated list of everything Dürer drew. Individual records are especially valuable where an old title masks a changed attribution or several objects share a subject. Use the institution and inventory number to keep the sheet distinct. The drawings discussed above show why this matters: front and back, joined and cut, original and copy can change the questions an image invites. Their variety rewards returning from the general name to the particular object.