A small composition of the Virgin and Child byJan Gossaert, known from at least six painted versions, was sufficiently celebrated by 1589 to be described as 'insignis tabella' in an engraving by Crispijn de Passe the Elder (Fig.2) which bears the legend: 'EFFIGIES HAEC SCVLPTA EST PER CHRISPIANVM VANDE PASSE AD IMITATIONEM INSIGNIS ILLIVS TABELLAE DEPIC- TAE PER IOANNEM A MABEVGE'. The painted version at Munich (Fig.3), though it is inscribed 'IOANNES MALBODIVS PINGEBAT 1527', is not invariably considered to be Gossaert's original.' Another, formerly in the Descamps collection, is inscribed IOANNES MALBODIVS INVENIT 1527',2 while a third, in Vienna (Fig.4), is believed to be a picture listed in an inventory drawn up in Vienna shortly after 28th June 1619 as '1 taffel ist Vnser Lieben Fraw mit dem kindl, von Johann de Mabusius'.) A fourth painting of the composition (Figs. 1 and 11), previously dismissed as a 'late copy' 'perhaps of the seventeenth century',' is in the National Gallery, London, and is the subject of this article. The remaining examples are not of particularly good quality and appear to be of limited interest

When the Rembrandt Research Project was initiated in 1968, the issue of Rembrandt's workshop was not nearly as acute as it is now. At that time it was generally believed that Rembrandt's oeuvre was contaminated on a considerable scale by later imitations and forgeries, and that paintings by pupils should be relatively easy to spot. There was great hope that scientific investigation would afford an objective alternative to classic connoisseurship in the sifting out of later accretions to Rembrandt's oeuvre.' Yet scientific investigation, above all dendrochronological analysis, has made it clear that there are far fewer Rembrandt forgeries or pastiches circulating than had been thought.2 Evidently, there were always enough paintings from Rembrandt's own workshop to hand, whether or not bearing false signatures, that in later times could be brought into circulation as 'genuine' Rembrandts. As it can safely be assumed that Rembrandt and his shop assistants used the same materials and procedures, it was inevitable that the members of the Rembrandt Research Project (RRP), like their predecessors, had to rely on connoisseurship in their attempts to distinguish autograph from workshop produc- tions. Given the project's working method, connoisseurship could be based on far more visual information than was available to earlier scholars. Often the paintings could be studied under ideal conditions, and most could be examined in situ in the relatively short (five-year) period of extensive travel at the outset of the project. Assessment of the condition of the paintings refined perceptions of their individual stylistic characteristics, while X-radiographs afforded additional information about working method and brushwork. Vast quantities of detail photographs and slides formed a rich visual databank.

An important painting by the early sixteenth-century Antwerp master Joachim Patinir has recently surfaced, or rather resurfaced, and is now in a private collection in Germany; it depicts two hermits in a vast landscape (Fig.34). M.J. Friedlander first published the picture as by Patinir without an illustration in 1931 in his Alt- niederlandische Malerei,' and it was finally reproduced from a poor, almost unreadable black and white photograph in 1973 in the English translation of that work.2 In 1947 Friedlander reduced his previously generous attributions to Patinir to just a dozen or so 'unzweifelhaft eigenhandiger gemalde'. In my own catalogue raisonne, published in 1968, I listed nineteen originals,' to which I would now add another, this newly rediscovered picture, which has recently been cleaned and has undergone a technical examination." The painting is in excellent condition, with only a few small touches of repaint. A dendrochronological analysis of the oak panel permits a dating after c.1509. I would place it among the mature works,c. 1515-20, perhaps around the time of the Landscape with St Jerome, now in the Louvre (Fig.35).

Since 1945 the Pavlovsk Palace-Museum has housed a Mars and Venus trapped by Vulcan by Maarten van Heemskerck, signed and dated 1561 (Fig.36),' which arrived there with the collections from the Gatchina Palace, Paul I's summer residence near St Petersburg, when they returned from evacuation during the Second World War. The painting may have been in the Gatchina Palace since the 1770s, the time of Count G. Orlov, Catherine the Great's favourite and the building's first owner. Under Emperor Paul the work was housed in the Picture Gallery of the palace, concealed because of its indecency beneath a Salvator Rosa, presumably with some mechanism by which it could be revealed.2 In 1853, after the new inventory was made of all the imperial collections, Nicholas I presented the Mars and Venus to the Tsarevich, the future Alexander II.: Nevertheless, the picture remained in Gatchina up until 1941 and, despite its importance, it has never been published or exhibited since the time of Paul I.

A small painting on copper of A landscape with figures (Fig.43) in a private collection can be firmly associated with the work of the Dutch Italianate landscape painter Bartholomeus Breenbergh (1598-1657).' In the background of the landscape is a castle perched on a cliff, its neat buildings silhouetted against the sky. In the foreground, somewhat awkwardly arranged, are pastoral fig- ures and a herd of goats, while a clump of trees frames the view on the left. On the back of the panel are inscribed the words 'Original Bartolommeo', clearly referring to Bartholomeus Breenbergh, the Dutch artist who went to Rome in 1619 and there came under the influence of Paul Bril (1554-1626). Breenbergh stayed in Italy for ten years, from 1619 to 1629, before returning to his native Holland.

Johann Liss was probably less than 35 years old when he died at Verona in 1631 and his oeuvre is small. Thus any addition to the number of his autograph paintings adds to our knowledge of his artistic development, even when the composition is well known. This is the case with the Temptation of St Mary Magdalen formerly at Edgcote, Northamptonshire, and sold at auction in 1994 (Fig.45).' A canvas of the same composition in the Dresden Gallery (Fig.47), first recorded there in 1765, was universally accepted as autograph until the monographic exhibition of Liss's work at Augsburg and Cleveland in 1975,2 when some scholars began to doubt its authenticity. Richard Spear, in his excellent review of the exhibition, was the first to publish this opinion, calling the Dresden painting a 'copy of the lost original'.3 It is now clear that the original is the Edgcote picture, which probably reached England in the eighteenth century, and remained in the collection of the Cartwright family at Edgcote, where it was little known,4 and was assumed to be a copy of the Dresden picture until its appearance in the sale-room. Its earlier provenance is unknown, and it cannot be established whether it is the picture referred to by Arnold Houbraken as a life-sized Magdalen by Liss in the collection of Siewert van der Schelling in Amsterdam, for this description could equally apply to Liss's Repentant Magdalen now at Slavkov in the Czech Republic (Fig.46)

In 1937 the Wallraf-Richartz Museum, Cologne, purchased a small painting on copper of the Temptation of St Anthony (Fig.52). Robert Eigenberger had attributed it to Liss in 1930,' and it was first published in 1951 by Otto Benesch, who connected it with the Temptation of St Anthony described by Sandrart as being in the artist's studio in Venice in 1629.2 Most scholars have followed Benesch's opinion that it is a late work, completed during Liss's second sojourn in Venice, c. 1626-29,' one of the few dissenting voices being that of Richard Spear who dated it to Liss's Roman period, prior to the second Venetian sojourn, in his review of the 1975 Liss exhibition.' Spear pointed to its stylistic affinities with the Merry company outdoors in a private collection, and to the similarity of the young temptress in the Cologne painting with the principal female figure in the Game of Morra at Kassel, which, he believed, might represent the same model." He also noted Sandrart's rather enigmatic chronology, in which the discussion of Liss's early rural scenes is followed by the sentence: 'Nach diesem hat er verfartiget eine Tentation S. Antonii ...'.