The Rothschild Prayerbook

A Flemish book of hours of about 1505 to 1510, 252 leaves carrying sixty-seven full-page miniatures by the leading illuminators of the last generation of the craft, and the most expensive illuminated manuscript ever sold at auction. It has set that record twice, in 1999 and again in 2014.

Between those two sales lies the reason it came to market. Appropriated from the Rothschild palace in the Theresianumgasse in Vienna in 1938, it was handed after the war to the Austrian National Library under the export-permit system by which the Republic extracted donations from the families it had failed to protect, and catalogued for half a century as Codex Vindobonensis Series Nova 2844. Austria's restitution law of December 1998 sent it back to the heirs, who sold it the following July.

Nobody knows who it was made for. There is no coat of arms, no emblem and no donor portrait anywhere in it. The arms on the binding belong to a later owner and to a set of silver-gilt mounts made a generation after the book.

252 leaves, sixty-seven miniatures, and four leaves gone

The book measures 228 by 160 millimetres and runs to 252 leaves. Christie's catalogued sixty-seven full-page arch-topped miniatures with surrounding borders and complementary borders on the facing pages, twelve full-page calendar borders framed in camaieu d'or with roundels of the feasts and the zodiac signs and full-colour miniatures of the occupations of the months, five small miniatures with full-page borders of their own, and two further text pages with full borders.

Four leaves are missing, three of them with miniatures and one with a full-page border. There are slight pigment losses in the backgrounds of two miniatures, on folios 120 verso and 124 verso, and small smudges at the edge of borders on four other leaves. Otherwise the catalogue describes it as in immaculate condition.

That is worth stating flatly because the sale coverage did not. The most expensive illuminated manuscript ever auctioned is an imperfect copy, and three of its full-page miniatures have been cut out.

The Master of James IV of Scotland is Gerard Horenbout

Anyone reading about this manuscript will meet two lists of artists that do not match, and the reason is that one of the painters has been catalogued for a century under two names.

The illuminator called the Master of James IV of Scotland, named from a portrait in a book of hours in Vienna, is generally recognised as Gerard Horenbout, a documented Ghent artist. Christie's states the identification as settled. A record that lists both as separate contributors is counting one man twice, and several accounts of this book do exactly that.

Horenbout became court painter to Margaret of Austria, Regent of the Netherlands, in 1515. He designed tapestries and stained glass as well as painting and illuminating. In the 1520s he moved to England with his family and appears in the accounts of Henry VIII's household between 1528 and 1531. Christie's argues that the Prayerbook predates his entry into Margaret's household, and that it was on the strength of this work that she appointed him.

A father, a son, and a guild register

Horenbout's principal collaborator is known as the Master of the Older Prayerbook of Maximilian I, after a manuscript made for the emperor, and is now widely believed to be Alexander Bening.

The identification rests on the correspondence between the career the pictures imply and a documented life: Alexander, or Sanders, Bening joined the painters' guild in Ghent in 1469 and died there in 1519. He was a friend of Hugo van der Goes and of Justus of Ghent, and is regarded as the most productive pupil of the illuminator known as the Vienna Master of Mary of Burgundy. What supports the identification in this particular book is that his work is so closely integrated with that of his son Simon Bening, born about 1483 and dead in 1561, who became the most celebrated illuminator in Europe.

Simon's most individual contribution here is the Vision of St Bernard on folio 245 verso. His Miracle of the Ass and St Anthony of Padua, on folio 240 verso, is a version of a miniature his father had painted for the Breviary of Eleanor of Portugal, a design that Gerard David then took up and reversed for a panel painting. Three generations of the same composition, in two media, inside one family workshop.

A fourth hand, the Master of the Prayerbooks of c. 1500, contributed several miniatures for the Hours of the days of the week and some of the infancy scenes in the Office of the Virgin. He is better known for secular manuscripts, above all the Roman de la Rose in the British Library. His Nativity at the opening of prime, on folios 108 verso and 109, is bordered by further episodes of the story and faces a scene of shepherds dancing at the angels' announcement - four painters' work interlocking across one opening.

The one miniature that is a painting

On folio 197 verso, introducing the Seven Joys of the Virgin, is a three-quarter-length Virgin and Child on a crescent moon, isolated against a golden radiance inside a narrow pearl-studded frame, with two small angels holding a crown above her head. It is handled in a completely different manner from everything around it and is accepted as the work of Gerard David, the Bruges panel painter.

The catalogue's own reasoning is unusually candid about how such an attribution is made. The drapery and figure style suggest the composition was drawn by the Older Prayerbook Master, and the angels may be his too. What is given to David is the mother and the child, on the ground that the handling conveys their relationship as well as their appearance, and that this is comparable to his Virgin among Virgins in the Morgan Library. On the same grounds two further miniatures, a St Catherine and a St Clara, are proposed for him.

A single miniature by a named panel painter inside a manuscript is exactly the kind of attribution that moves a price, and it is an argument from quality rather than a document. This record marks it as such.

Flowers that cast shadows on the page

Over half the book carries significant decoration outside the text. The borders are the reason it is famous, and they are systematically deceptive.

They are described as richly varied trompe l'oeil: sprays of acanthus and strewn flowers with insects and vignettes, camaieu d'or architectural surrounds with sculptural figures and reliefs, jewels and enamels against coloured grounds, and individual borders that replicate cloth of gold and peacock feathers. Naturalistic flowers lie as if scattered on the surface of the vellum and cast shadows onto it. On some pages the border carries a narrative of its own, extending or answering the subject of the miniature.

The device is not this workshop's invention. It goes back to the Vienna Master of Mary of Burgundy and to the margins of the Hours of Engelbert of Nassau in the Bodleian, usually dated to the 1480s, and Alexander Bening was among the earliest to use it. Two of the most striking borders here, the peacock feathers on folio 225 verso and the skulls on folio 164 verso, trace back to that source.

A record of the Mass as it was actually celebrated

The first sequence of devotions, the abbreviated Offices and Masses for the days of the week, presented a problem: there is no conventional set of images for it. Horenbout's answer was to paint, for each feast, a moment from the celebration of its Mass.

The result is a series of liturgical scenes devised individually rather than copied, showing vestments, altars, celebrants and congregations in real architectural space. Christie's calls it a fascinating record of contemporary liturgical practice and setting, and it is: a book made for private devotion contains one of the better visual records of how the public rite was performed in the Netherlands around 1505.

That is what the object was for. A book of hours is not a picture gallery; it is an instrument for saying the right words at the right hours of the day, and the pictures are there to fix attention on what is being said. The Offices of the Virgin and of the Dead, which carry Horenbout's most prominent miniatures, are the two sequences a lay owner would have used most.

Nobody knows who it was made for

For a production this expensive the silence is remarkable. There are no arms, no emblems and no donor portrait anywhere in the manuscript.

Christie's identifies one component that may reflect a commissioning patron: the Suffrages to Saints Vincent, Benedict and Anthony of Padua and two prayers to the Virgin, on folios 239 to 246 verso. These appear to be modifications to the book as originally written and planned. They are in a different scribe's hand from the Suffrages before them and the Athanasian Creed after them, and they are laid out with no allowance for integral miniatures - yet the miniatures supplied on single leaves are by the same artists as the finest work earlier in the book. In one of them, on folio 238 verso, the stained glass behind St Vincent carries coats of arms, including a displayed eagle and a shield of gules with a chevron and three small charges or.

That is the closest thing to a clue in the entire volume, and it has not been resolved. The manuscript belongs to a group of luxury books made around 1490 to 1520 for an international clientele and for members of the Habsburg court in the Netherlands - the London Hours in the British Library, the Spinola Hours in the Getty, the Grimani Breviary in Venice - and Christie's notes that there is no firm evidence for the intended owner of any of them.

The binding belongs to a different owner

The red velvet is renewed. What is old on the outside of this book is the metalwork: mid-sixteenth-century silver-gilt cast and chased centrepieces, cornerpieces, clasps and catches, attributed to the workshop or circle of the Nuremberg goldsmith Wenzel Jamnitzer, who died in 1585. The leaf edges are gilt and gauffered to a diaper pattern.

The centrepieces carry the lion rampant of the Palatinate and the diaper of Bavaria, which is why the manuscript is usually said to have passed through the house of Wittelsbach. That inference comes from the binding mounts and from nothing else. Christie's is explicit that nothing supports Trenkler's suggestion, made in the commentary to the 1979 facsimile, that the arms are those of Herzog Ernst von Wittelsbach and that the book had a later provenance in the Palatine library at Heidelberg. Accounts that send this manuscript through Heidelberg are repeating a rejected proposal.

Two of the miniatures inside the book show bound volumes with edges gauffered to the same diaper pattern that survives on this one, which tells us the leaves were not trimmed when it was rebound, and raises the possibility that red velvet with gilt cornerpieces and clasps was the book's original appearance, replicated for a mid-sixteenth-century owner. It is equally possible the metalwork joined the manuscript in the nineteenth century. Both readings are in the catalogue and neither is settled.

Into the Rothschild collections, and what it was worth in 1906

Anselm von Rothschild, who died in 1874, laid the foundation of the Austrian Rothschild collections and bought Netherlandish pictures in particular. The Prayerbook was number 597 in Schestag's 1872 catalogue of his collection. Number 595 in the same catalogue, the Bening prayerbook made for Cardinal Albrecht von Brandenburg and now in the Getty, had been bought in 1868, which suggests this one was acquired at about the same time.

It passed to his son Baron Nathaniel von Rothschild, and appears as number 452 in the February 1906 inventory of Nathaniel's estate, hanging in the Galerie of the palace in the Theresianumgasse. It was valued there at 150,000 Kronen. The Brandenburg prayerbook, listed immediately after it as number 453, was valued at 80,000.

That pair of numbers is the most useful thing in the whole nineteenth-century record, because it is a valuation made by a family for probate rather than a price made in a saleroom, and it shows that they already regarded this book as worth nearly twice the other. Both are now among the most celebrated Flemish manuscripts in existence.

Nathaniel's palace, and presumably the manuscript with it, went to Baron Alphonse von Rothschild, who died in 1942. Two later inventory numbers of the palace are still on the book: 927 on a label at the upper corner of the lower cover, and AR3390 in pencil on the front flyleaf. It stayed in the Theresianumgasse until 1938.

1938, and the donation that was not a donation

Germany annexed Austria in March 1938 and the Rothschild collections in Vienna were appropriated.

The part of what followed that the sale coverage does not carry is what the Republic of Austria did afterwards. Objects recovered at the end of the war were returned to Austria and in principle to their owners, but a family that wanted to take restituted property out of the country was expected to donate significant works to Austrian state collections in exchange for export permits. The Rothschilds, who were not staying, gave up a great deal on those terms. This manuscript went to the Oesterreichische Nationalbibliothek and was shelfmarked Codex Vindobonensis Series Nova 2844.

It sat in Vienna for half a century, published in a full facsimile edition in 1979 as an Austrian national treasure, having been obtained by pressure applied to people the same state had failed to protect. Calling that a donation without qualification is the error this record exists to correct.

1998 and 1999: the law, and the sale eleven days later

Austria passed its art restitution law in December 1998 after sustained international pressure. The manuscript was among the property returned to the Rothschild heirs.

They sold at once. Christie's offered the collection of the Barons Nathaniel and Albert von Rothschild in London on 8 July 1999, and the Prayerbook was lot 102. It made GBP 8,580,000, reported at the time as USD 13,378,558, then a record for an illuminated manuscript, against a high estimate reported as USD 4.9 million. The sale as a whole raised close to USD 90 million.

Christie's published prices include the buyer's premium, so GBP 8,580,000 is a total. A hammer of GBP 7,800,000 with a flat 10 per cent premium reproduces it exactly, while the London scale of the period, 15 per cent on the first GBP 30,000 and 10 per cent above, would give GBP 8,581,500 on the same hammer. The gap is 1,500 pounds and it is unresolved, so the hammer in this record is given as derived rather than asserted.

29 January 2014: the record set again, and misreported again

Fifteen years later the manuscript returned in Christie's New York sale 2819, Renaissance, as lot 157, estimated at USD 12 million to 18 million. It made USD 13,605,000 and broke its own record.

That figure has been printed as a hammer price, including in the trade press. It is not one. USD 13,605,000 is exactly what a hammer of USD 12,000,000 becomes under a premium of 25 per cent on the first USD 100,000, 20 per cent to USD 2,000,000 and 12 per cent above: 25,000 plus 380,000 plus 1,200,000, or USD 1,605,000 in fees. The derived hammer lands on a round increment and on the low estimate, which is what real hammer prices do and what USD 13,605,000 does not.

The buyer stayed anonymous for eight months. In September 2014 the Australian media proprietor Kerry Stokes was revealed as the purchaser on Australian television.

The seller in 2014 has not been identified, and neither has the buyer in 1999. Fifteen years of this book's ownership carry no name at all, which is a gap of the same kind as the four centuries before Anselm von Rothschild, only more recent and more surprising.

Where it is now

The manuscript is part of the Kerry Stokes Collection in Perth and has been lent for public display in Australia, reported as being at the National Library of Australia in Canberra. It was published in 2015 in two books produced around an exhibition of the Stokes collection, one of them Margaret Manion's study of the Prayerbook and other works.

A privately owned manuscript that is exhibited and published is a better outcome for scholarship than most sales of this kind produce. What it does not have is an open digitisation. The images that circulate are the Christie's sale photographs, which is why the plates on this page come from a saleroom rather than a library.