The precise relationship of The Three Maries at the Tomb (Fig. 1) in the Boymansvan Beuningen Museum in Rotterdam to the work of Hubert and/or Jan van Eyck has proved difficult to establish, mainly because relatively little is known about their output apart from Jan van Eyck's signed paintings of 1432-41. The provenance of the Rotterdam picture has been traced back to the mid 18th century (Note 2), while the coat of arms, a later addition at bottom right, has been identified as that of Philippe de Commines, who has thus been posited as the earliest known owner (Note 3). Since the beginning of this century the panel has generally been ascribed to Hubert van Eyck on the basis of a comparison with his contribution to the Ghent Altarpiece, but doubts have also been expressed about the attribution to the Van Eycks (Note 5), while later dates have been suggested on the grounds of the view of Jerusalem (Note 6, 7) or the arms and armour (Notes 8, 9) . However, Panofsky remained convinced of the early date and kept to the attribution to Hubert, while suggesting that Jan had worked over certain details (Note 10). The restoration of 1947 (Note 11) revealed some gilded rays on the right side, which gave rise to suggestions that the panel had once formed part of a friezelike composition or a triptych (Notes 12-14). Recent opinion still remains divided, Sterling seeing the panel as having been painted by Jan van Eyck after 1426 (Note 15), Dhanens as the work of a follower around 1450-60 (Note 16). Scientific examination appeared to be the only way of obtaining new data, while the recently published results of a similar examination of the Ghent Altarpiece (Note 17) offered an additional incentive. An earlier scientific examination was carried out by Coremans in 1948 (Note rg), while the work had previously been examined by infrared reflectography by the authors in 1971 (JV ote zo) . Tfie 1)(inel on which the picture is painted consists rf three horizontal planks with dowelled joints (Note 21). The four corners are bevelled off at the back, which suggests that any later reduction in the panel can only have been slight. On the back is a sealed statement by D. G. van Beuningen to the effect that the painting had not suffered from being stored underground during the war (Fig. 2, Appendix 2) . The paint surface is in a reasonably good state, but exhibits heavy craquelure, which has played a part in the aesthetic assessment of the picture (Note 23) . Dendrochronological examination (Appendix I) showed that the two oaks from which the planks came were probably not felled before 1423. Since recent research has shown that the gap between felling and usage was not likely to have been much more than fifteen years in the 15th century (Note 25) and there is nothing to support the hypothesis that an old panel was reused here (Note 26), it is highly improbable that the picture was painted at the end of the 15th century. The most likely date is C. 1425-35 i.e. the period when the Ghent Altarpiece was painted or slightly later. No other results of dendrochronological examination on Van Eyck panels are available for comparison as yel. Examination by infrared reflectography (Note 28) revealed detailed underdrawing in virtually all parts of the picture and this was very carefully followed during painting with changes only in small details (cf. Figs.3, 5, 7). Stylistically the underdrawing accords with what is known about underdrawing in Van Eyck paintings today, this exhibiting a considerable difference from that of other Flemish Primitives, so that the Rotterdam panel is certainly a Van Eyck work. Among the most striking similarities to the central panel (x) and that with the Knights of Christ (IX) in the Ghent Altarpiece (Note 30) are the underdrawing of the drapery of the angels (Figs. 7-9), the city in the distance (Figs. 3,4, Note 31) and the minutely detailed armour (Figs. 14, 15, Note 33). Types of hatching that appear to be characteristic of the Van Eyck style are that of the shadows, which is sometimes overlapping and generally parallel to the main contours (Figs. 5,8) and a more rarely used type with short lines at an angle to contours (Fig. 9). The x-radiographs (Note 35) give a good idea of the damage to the paint surface (Figs. 16, 17) , which isfound mainly in the sky, along the crack in the top plank and on the bottom edge on the left. There is also a great deal of abrasion on the edges of the craquelure. The x-radiographs confirm the fact that no radical changes were made in the original, generally underdrawn, composition and reveal that the soldiers and their arms were left in reserve during the painting of the rocks and ground, a detail which likewise indicates continuity during the painting process. The underpainting of the rocks in large light blocks with simple contours shown up by x-ray photography is very close to that in panel IX in the Ghent Altarpiece (Note 38). Examination by stereomicroscope (Note 40) generally already gave an impression of the layered structure of the paint. It also showed up some minute details scarcely distinguishable by the naked eye : two horsemen and somefigures in tlae square on tlte leji qlthe city, a .slalue in a niche in the doorway in the zvall in tlae certtre (Fig. 18; possibly a reminiscence of the Golden Gate, Note 56) and a number of ship's masts with crow's nests on the horizon on the right (Fig. 19). Part of the vegetation was shown to be very finely and precisely rendered (Figs. 20, 21), while the rest was not so fine. Similar differences appear in the two bronze-coloured ointment jars in this painting and also in the bottom zone of the Ghent Altarpiece (Note 41). These may reveal two different hands or the somewhat hasty finishing of some areas. The paint samples (Note 42) revealed the presence of an oleaginous isolating layer over the chalk and glue ground comparable to, but thinner than that on the Ghent Altarpiece (Note 45). The only other Flemish Primitive in whose work such a layer is found is Dirc Bouts (Note 50). The paint layer also exhibits many similarities to that of the Ghent Altarpiece, not only in the number and thickness of the layers, but in the composition and overall structure of the paint. For example, the skies in both works are built up in three layers from light to dark on the basis of lead white with increasing amounts of azurite and sometimes a bit of lapis lazuli, the vegetation consists of two layers of green with a glaze over them and the structure of the red mantle of one of the Maries resembles similar areas in the Ghent Altarpiece. This technique again makes it very unlikely that the panel was painted at the end of the 15th century or later. A final point is that the gilded rays ( Fig. 22), like the coat of arms (Fig. 23), prove to be a later addition. Finally, renewed consideration was given to certain iconographical aspects which have been used as dating criteria. The arms and armour have been seen as grounds for a later dating by Squilbeck in particular, but it seems quite likely that many of the forms are purely imaginary, while other experts do not agree with Squilbeck in dating certain elements to the 16th century (Note 53). The arms and armour are in any case an integral part of the painting. The detailed view of Jerusalem is regarded by some as impossible before Erhard Reuwich's print of 1486, while others express surprise that it was not copied by other artists. In fact, however, it is strikingly close in many details to the view in the Ghent Altarpiece, although the latter is firmer in its spatial construction and more convincing. Whole sentences have been read into the texts on the hems of two of the Maries' garments and the soldier's cap (Note 57 ) and it has been argued that the letters are Roman, not Hebrew (Note 58), but in fact they are indispulably Hebrew and although words can sometimes be recognized, they do no form a sentence or text (Note 59). The coat of arms is certainly that of a nobleman of the Order of St. Michael, but whether he was Philippe de Commines is uncertain. The Van den Woesteyne and Van Meaux van Vorsselaer families also bore these arms, albeit in different tinctures (Note 6o). Since the arms are done, in a brownish-grey, they cannot be more precisely identified. The presence of no less than five layers of varnish between the green meadow and the coat of arms could indicate that the arms were added much later than previously thought, possibly in the 16th or even the 17th century (Note 47). While the present study has shown that the Rotterdam painting is quite an early Van Eyck, its precise position in the Van Eyck oeuvre cannot be determined until results of examinations of other works in the group are available.