1. Disguised symbolism and Panofsky's views about iconography / iconology are now part of the much larger issue of Panofsky himself; see for example the papers of the Paris symposium, Erwin Panofsky: Cahiers pour un temps, ed. Jacques Bonnet, Paris, 1983; Michael Ann Holly, Panofsky and the Foundations of Art History, Ithaca, N.Y., 1984, with reference to earlier criticism; Silvia Ferretti, Cassirer, Panofsky, and Warburg: Symbol, Art, and History, trans. Richard Pierce, New Haven, Conn., 1989, and the works cited by Bedaux, 19 n. 1. Anecdotes circulate that Panofsky was at times distressed by applications of his theory of disguised symbolism (see Holly, 164; Bedaux, 14).

2. Because the English edition renders these and other passages rather freely (e.g. the reference to the picture as "a kind of snapshot" is not justified by the German text), see Friedländer (1934-37), 1 (originally published Berlin, 1924), 56: "Stimmung sakramentaler Symbolik" and 1:57: "Auch dem geringsten Dinge wird . . . Bedeutung und Wert eines Kultgegenstandes verliehen." For a similar instance of the eclectic genesis of Panofsky's iconographic ideas, see Lubomír Konecny * , "On the Track of Panofsky," Journal of Medieval and Renaissance Studies 4 (1974): 29-34.

3. Panofsky, 1934, 126.

4. Ibid.

5. Panofsky, 1953, 1:203, 180.

6. Ibid., 142.

7. For example, among the Romanesque capitals of Saint-Pierre in Chauvigny, one that depicts Saint Michael weighing a soul in a balance has inscriptions identifying the two main figures: "Micael Arcangelus" and "Hic est diabolus." Across the top of another portraying the Adoration of the Magi on its principal surface the artist has inscribed his name, "Gofridus me fecit," in letters that are clearly legible from the nave and choir. Both by content and grammatical construction, these simple inscriptions closely parallel those found on Van Eyck's works.

8. Panofsky, 1953, 1:137, 143.

9. This interpretation of the towel has become a commonplace, but see, for instance, Dhanens, 93; and Albert Châelet, Van Eyck, New York, 1980, 22 and pl. 21.

10. Joannes Molanus, De historia ss. imaginum et picturarum, ed. Joannes Paquot, Louvain, 1771, 274-75.

11. See Chapter 1, nn. 23-24; the works cited in n. 1 above; and the review of Holly (cited in n. 1) by Paul Crowther in American Historical Review 91 (1986): 87: "The question of criteria for determining the validity of iconological interpretations is not given adequate consideration. Without such criteria, the way is clear for an empiricist scepticism toward the worth of iconology as such."

12. Dhanens, 198, 203.

13. Brooke, 282.

14. Myriam Greilsammer, L'envers du tableau: Mariage et maternité en Flandre médiévale, Paris, 1990, 46-49.

15. Robert Baldwin, "Marriage as a Sacramental Reflection of the Passion: The Mirror in Jan van Eyck's Arnolfini Wedding," Oud Holland 98 (1984): 57-75.

16. Jan Baptist Bedaux, "The Reality of Symbols: The Question of Disguised Symbolism in Jan van Eyck's Arnolfini Portrait," Simiolus 16 (1986): 5-26 (for the passage referred to, 13), now reprinted in Bedaux, 21-67. There is some confusion and ambiguity in Bedaux about the theory and practice of medieval marriage. For example, he misunderstands arrha (29), and his use of expressions like "to contract a marriage" are ambiguous in the context of the fifteenth century, for, as I have tried to explain, without further qualification it is not clear whether the reference is to a betrothal contract of future marriage or a marriage contract of present consent. Consequently, Bedaux is misunderstood by other writers, e.g. Harbison, 1990, 252 n. 7, and Roskill, 65, who both think he is writing about the picture as a betrothal. In his book, however, Bedaux is quite specific: his doubts about disguised symbolism, he says, developed "as it became increasingly clear that [the London panel] was fairly easy to interpret within the context of the contemporary marriage ceremony. In fact, it became more and more apparent that the Arnolfini portrait was an accurate depiction of a specific wedding" (11). In Bedaux's view, the ceremony depicted in the double portrait follows what he terms, on the basis of a lecture he attended, "the ancient codification of secular law," which, he says, did not preclude an ecclesiastical rite. He further suggests the ceremony was not clandestine and that Arnolfini likely "would have had his marriage consecrated afterward" (25, 29). Bedaux then proceeds to accommodate "the left-handed gesture" of the double portrait to what he presumes would have happened at such a ceremony by misreading various manuscript miniatures, including the Froissart illumination that is my Fig. 33 (see also the comments above in Chapter 2, n. 65). Although he does not say so explicitly and apparently is not fully aware of what he implies, Bedaux seems to interpret the picture as an instance of the Italian rite before a notary. If such is indeed the case, he misunderstands that by the twelfth century this was the equivalent of the northern ceremony in facie ecclesiae and constituted the sacramental rite that created the matrimonial bond. The only major difference between the two ceremonies in this regard was that in the north what Bedaux calls the "consecration" (i.e. the nuptial blessing of the old Roman liturgy) followed immediately inside the church, whereas in Italy the church rite, deferred to another day, remained part of the nozze and might even be omitted. And as I have explained, the gesture used in the wedding ceremony before the notary was not a joining of hands at all but rather the placing of a wedding ring on the bride's finger by the groom.

17. See for example Barbara G. Lane, "Sacred versus Profane in Early Netherlandish Painting," Simiolus 18 (1988): 106-15; and the reply of Craig Harbison in 19 (1989): 198-205.

18. For the passages quoted or cited, see Harbison, 1991, 16, 17, 36, 44; and Harbison, 1990, 261-64.

19. Harbison, 1990, 288; and Harbison, 1991, 46.

20. To suggest an analogy with the evolution of language, the interpretation of paintings, like the meaning of words, can change with the passage of time. For example, egregious was long used in English with both good and bad connotations ("prominent," "remarkable" / "gross," "flagrant"). Although the Latin root and the modern Italian egregio have only positive connotations, in modern English the historical or etymologically correct meaning has been completely displaced, and only the antithetical negative meaning is now current. One can say therefore that the meaning of the word has changed, but surely no one would insist on applying the inverted English meaning to a Latin or Italian context or the etymologically correct usage to a modern English context.

21. Seidel, 1989, esp. 57-58, 78; and Seidel, 1991, esp. 26.

22. Seidel then moves on to new speculations that are better pursued in her own exposition. But the gist of the argument is that Giovanna might not have been of marriageable age when the picture was painted, and thus "the imminence of consummation . . . enshrined in the panel" was something hoped for in the future rather than realized in the present. The dowry might then have been advanced to Arnolfini as a disguised, and in some way irregular, loan for his commercial activities during the interval between "Ring Day" and the much later consummation of the marriage, in which case the panel, "as the clever commission of a shrewd banker," would become "a financial transaction that is enacted as a marriage ceremony." For the passages quoted and cited, see Seidel, 1989, 66-73, 77, 79; and Seidel, 1991, 34. As disguised loans were a commonplace of contemporary Italian mercantile practice, had the two men wished to negotiate such a contract, they could have done it with a simple notarial instrument; it is thus inconceivable that they would have indulged in the complex and legally invalid ruse Seidel suggests.

23. See Julius Kirshner and Anthony Molho, "The Dowry Fund and the Marriage Market in Early Quattrocento Florence," Journal of Modern History 50 (1978): 403-38. See also Klapisch-Zuber, 191-92.

24. Retaining Panofsky's notion that the panel was intended as a legal document, Seidel (1989, 68-69)—arguing that the mirror, "a seal-like object," functions as Van Eyck's seal—misappropriates information about the Flemish institution of the erfachtige liede, which required that a document, to be authenticated, be witnessed by two urban landowners (to own a house was not sufficient); (see Murray, 162). The document cited (Weale, xxxviii) to establish Van Eyck as a landowner concerns the annual redditus, or rent, he paid the Church of Saint Donatian in Bruges for a house. Van Eyck is believed to have owned this house, but the land on which it was built evidently belonged to the church; this explains both the rent and the amount of the redditus, which is too small for a year's house rent (cf. Weale, xxxvi, for the earlier rental of Van Eyck's house in Lille). Thus what evidence there is suggests than Van Eyck did not own land in Bruges. Seidel also conflates the two medieval ways to authenticate a document, by signature and by seal; normally the use of one precluded the use of the other by the same individual. A mirror, moreover, is not a seal, Van Eyck was unqualified to authenticate a document in either way, and Italian merchants like the Arnolfini and Cenami, accustomed to notarial instruments for centuries, would hardly have considered the London panel an appropriate substitute.

For the latest version of Seidel's ideas, see her Jan van Eyck's Arnolfini Portrait: Stories of an Icon, Cambridge 1993. Since the publication of this book coincided with the typesetting of mine, I can add only one further comment on her methodology. Seidel notes at one point (255 n. 17) that she has "knowingly nudged the evidence" in her concern for "the plausible rather than the probable." One instance of such nudging exemplifies her approach. Arguing that the "intertwined design with crosses immediately above" Van Eyck's name in the signature inscription (i.e. the ornamental flourish of the inital letter) is Jan's sign as a notary, Seidel continues (132-33): "Shaped like a chandelier and furnished with a lighted candle, the sign additionally alludes to the contemporary adjudication of claims by fire, known in contemporary parlance as 'la chandelle éteinte,' the extinguished candle." Two documents are cited in support of this assertion from archival materials in Bruges published by Gilliodts-van Severen. At first it appears that Seidel has discovered more satisfactory evidence for an earlier interpretation of the lighted candle proposed by Dhanens (199) and Bedaux (60 n. 14). But what these documents refer to is "adjudication à la chandelle [éteinte," a form of auction (as the editor explains in a footnote on one of the pages Seidel cites) common in northern Europe from the fifteenth to the eighteenth century: at such an auction bidding continued until a short piece of lighted candle went out, with the successful bid being the last made before the candle extinguished itself. For further reference to this practice, see Antoine Furetière, Dictionaire universel, The Hague, 1690, s.vv. "chandelle" and "adjudication"; and Jacob Grimm and Wilhelm Grimm, Deutsches Wörterbuch, 5 (Leizpig, 1873), 617; Samuel Pepys describes two sales of ships "by an inch of candle" (see his Diary for 6 November 1660 and 3 September 1662). Seidel's reference to "adjudication" by "la chandelle éteinte" is thus irrelevant to the double portrait, unless one wishes to argue that the woman is being auctioned by the man and the mirror reflects the images of two bidders. This example epitomizes what I see as a continuing methodological problem in the application of historical evidence to iconographic scholarship: the uncritical symbolic reading of something commonplace (taking a calligraphic flourish for a chandelier with a lighted candle); the application of that reading to circumstances that are historically untenable (the argument that Van Eyck was a notary and the painting a legal document); and the attempt to justify the argument by irrelevant historical documentation (the contextually meaningless reference to auction sales "by the candle"). It should also be noted that the jeweled pendant in Boston that Seidel (104-5) relates to the gesture in the double portrait is a nineteenth-century imitation; see Hanns Swarzenski and Nancy Netzer, Catalogue of Medieval Objects in the Museum of Fine Arts, Boston: Enamels and Glass, Boston, 1986, 152.

25. Kenneth Clark, The Nude: A Study in Ideal Form, Princeton, N.J., 1972, 317-21, sees this "curve of the stomach" as a characteristic of the International Gothic style. An often reproduced miniature from Brussels, Bibliothèque Royale ms. 9296 (Prevenier and Blockmans, fig. 98), depicts Margaret of York performing the seven corporal works mercy. Although in each scene the duchess wears the same gown, her appearance resembles that of the woman in the double portrait and the Dresden Saint Barbara only in the scene where the front of the gown has been drawn up to the waist and held with one hand.

26. Christine de Pizan, Le livre des trois vertus, ed. Charity Cannon Willard, Paris, 1989, 157-59. This chapter (11) is entitled "Ci devise de celles qui sont oultrageuses en leurs abiz, atours et abillemens."

27. An amusing reflection of these conceits is found in the Vienna Girart de Roussillon, ms. 2549, fol. 164r (Thoss, pl. 46), which illustrates the simultaneous building of several churches. The stone carvers assert their rank at the top of the hierarchy of workers depicted in the miniature by their fashionable clothing, including hats with feathers and poulaines with long points whose awkwardness they overcome by sitting down while working. By contrast, the footgear of the lowly assistant who mixes mortar is so worn as to reveal his bare heel and toes. For a similar miniature by the same master, see Dogaer, pl. 5, where the length of the shoes is inversely proportional to the mobility required by the work being done.

28. Panofsky, 1934, 126; Panofsky, 1953, 1:203.

29. Joseph continues to be portrayed wearing clogs in the Marriage of the Virgin until the end of the fifteenth century, as in a painting by the Master of the Tiburtine Sibyl in Philadelphia and in a number of anonymous panel paintings; for the latter, see the exhibition catalogue Primitifs Flamands anonymes, Bruges, 1969, nos. 55-56.

30. For the Vatican prototype without the pattens, cf. Meiss, 1967, figs. 7 and 10; all seven Arsenal miniatures depicting figures with pattens are by the Guillebert de Mets Master, who is known to have been a Fleming. Both the stylish straw hats and the selective use of these very elongated clogs by the duke himself and those closely associated with him can also be seen in the Eyckian painting in Versailles of some festivity at the court of Philip the Good; on this painting, see Anne van Buren Hagopian, "Un jardin d'amour de Philippe le Bon au parc de Hesdin," Le Revue du Louvre 35 (1985): 185-92.

31. Commission royale d'histoire, Recueil des chroniques de Flandre: Corpus chronicorum flandriae, ed. J.-J. De Smet, Brussels, 1837-65, 4:57; exactly what Philip did with the pattens is somewhat obscure from the chronicler's description: "à mettre sus et clouer pattins."

32. Philip wears pattens while seated in a canopied chair of state in the presentation miniature of the Vienna Girart de Roussilon (see Thoss, pl. 1) as well as in Brussels ms. 9095 (see Winlder, pl. 30); for other examples, see Georges Dogaer and Marguerite Debae, La librarie de Philippe le Bon, Brussels, 1967, pls. 42, 43, 47, 49.

33. Meyer Schapiro, Late Antique, Early Christian, and Medieval Art, New York, 1979, 1-19.

34. See for example Chapter 3, n. 49, and R.-A. d'Hulst, Jacob Jordaens, London, 1982, 280.

35. Panofsky, 1953, 1:203; Panofsky, 1934, 126, 118 n. 10, 125 n. 27.

36. Mâle, 426-27. Panofsky's citation of Barbier de Montault's Traité d'iconographie chrétienne leads to no more than a statement that the dog, as the most faithful of animals, is fittingly "le symbol vivant de la fidélité" because fidelis is derived from fides . Panofsky also cites Ripa's Iconologia, where the dog and lighted candle are attributes of the personified theological virtue Fides, but since the lighted candle does not appear as an attribute of Fides in this sense before about 1470 (see Mâle, 311-12) and the dog apparently not before the seventeenth century, Ripa's symbols are irrelevant to a picture painted in 1434. Eventually (though Panofsky makes no mention of it) the dog does appear as an emblem of wifely fidelity in Alciati's Emblemata (Emblem 191: "In fidem uxoriam," as in the Lyons edition of 1550, 205), while in a drawing by Maerten de Vos, Fidelitas is personified with a dog as an attribute (see E. de Jongh, Portretten van echt en trouw: Huwelijk en gezin in de Nederlandse kunst van de zeventiende eeuw door, Zwolle and Haarlem, 1986, 232, pls. 28c and 52c). But it would be impossible to infer the intended meaning of the dog in either case without the accompanying captions, and the imagery, because of its late date, is as anachronistic as Ripa for interpreting a work from the fifteenth century. Even when a dog appears in seventeenth-century double portraits, a symbolic reading based on Alciati is dubious because this emblem imagery also refers textually—and usually visually as well—to apples as the fruit of Venus (the couple often sit under an apple tree). Thus if a painter had really been inspired by Alciati, it seems likely that both aspects of the emblem imagery would have been incorporated into a picture.

37. Meiss, 1969, 31-32; see ibid., figs. 34 and 114, for representations of the duke, accompanied by small dogs, kneeling in prayer before the Virgin. See also Mercer, 21.

38. See for example Winkler, pls. 8, 12, 16, 18, 25, 27, 36, 38, 40, 43-45, 50, 52; and Dogaer, figs. 32, 33, 44, 47, 61, 63, 64, 67, 81.

39. Guennol volume, fol. 142v; John Plummer, The Hours of Catherine of Cleves, New York, n.d., pl. 77.

40. Eileen Power, Medieval People, New York, 1963, 90-91; see also the same author's Medieval English Nunneries, Cambridge, 1922, 305-7, 588, 662.

41. Panofsky, 1934, 126; or in Panofsky, 1953, 1:202: "symbol of the all-seeing Christ." Panofsky gives no documentation whatsoever for the candle as a symbol of God, and none of the sources he cites provides acceptable evidence linking the candle either to marriage customs or to fifteenth-century practice in swearing oaths. The much later custom of swearing an oath before a crucifix with lighted candles on either side is hardly analogous to what is seen in the double portrait. And since the phrase "iurare super candelam" can only refer to a corporal oath, this too is inapplicable to the London panel, for the male figure does not touch the candle; moreover, since the expression comes from the satirical tale of the roguish fox, Reinke de Vos, "to swear on a candle" was apparently intended as a joke, which explains why the lexicographers of medieval Latin make no mention of such an oath (cf. Du Cange, 4:451-63). Bedaux, 60 n. 14, summarizes other suggestions, all very casually documented, none of which specifically concern marriage contracts. And although it seems plausible to relate the lighted candle to the swearing of a promissory oath, neither contemporary visual documentation (cf. Figs. 32-35 and Plate 10) nor any fifteenth-century texts known to me provide evidence for the use of a lighted candle in such a context. See also n. 24 above.

42. See for example the exhibition catalogue Masters of Seventeenth-Century Dutch Genre Painting, Philadelphia, 1984, pls. 52, 55, 71, 110, 122; a single candle in a chandelier is seen in the central panel of an eighteenth-century copy in the Rijksmuseum of a famous lost triptych of Gerard Dou (ibid., 186; for a better illustration, see Peter Thornton, Seventeenth-Century Interior Decoration in England, France, and Holland, New Haven, Conn., 1983, pl. 268).

43. Bennett, 67, 96.

44. For reference to the chronicler who mentions the sale of the melted-down wax, see Robert L. Mode, "San Bernardino in Glory," Art Bulletin 55 (1973): 62-64.

45. Mirot and Lazzareschi, 99.

46. Panofsky, 1953, 1:203, also linking the "crystal beads" with the mirror, both of which are said to be (although without any justification) "well-known symbols of Marian purity."

47. Heinrich Schwarz, "The Mirror in Art," Art Quarterly 15 (1952): 96-118; idem, "The Mirror of the Artist and the Mirror of the Devout," in Studies in the History of Art Dedicated to William E. Suida on His Eightieth Birthday, London, 1959, 90-105, especially 99-100 and 104. For early criticism of this tendency (but not with respect to Van Eyck's double portrait), see Peter Schabacker, "Petrus Christus' Saint Eloy: Problems of Provenance, Sources, and Meaning," Art Quarterly 35 (1972): 103-4.

48. Panofsky, 1953, 1:148.

49. Bernardinus de Busti, Mariale, Strasbourg, 1496 (Goff B-1334), fols. f7r-g6v. Elsewhere (fol. rr1r) in this massive compendium of Marian theology drawn from patristic and scholastic sources, Busti says that the Virgin is the "speculum dei" because God's image, that is, Christ, shines forth from her, and he goes on to explain that just as a mirror receives into itself the image of its maker, "so the blessed Virgin received into her womb him who created her" ("Sicut autem speculum recipit in se figuram eius qui fecit illum, ita beata virgo in suo utero recepit eum qui creavit ipsam").

50. The Immaculate Conception is actually the central theme of this iconography, for a figure of Christ at the top of the composition is invariably accompanied by the text, "Tota pulchra est . . . et macula non est in te," taken from the Song of Songs. For the Grimani Breviary miniature, see Carol J. Purtle, The Marian Paintings of Jan van Eyck, Princeton, N.J., 1982, pl. 81. This imagery is found in various printed Parisian Heures for about two decades, beginning in 1505, after which it disappears. Other early examples include a polychrome wood relief in the cathedral of Bayeux and a Tournai tapestry at Reims cathedral. It is found again, with the emblems still labeled, in a work of the Jesuit scholar Petrus Canisius, De Maria virgine incomparabili, et Dei genetrice sacrosancta, Ingolstadt, 1577, but later, as in paintings of this subject by Ribera, the identifying texts are omitted.

51. Ritzer, 251-52; Molin and Mutembe, 212-13, esp. 213 n. 14.

52. For this miniature, see Thoss, 177, and pl. 47.

53. For other examples of the moralizing use of mirrors, see Marrow, 162-64; and Ilja M. Veldman, "Lessons for Ladies: A Selection of Sixteenth and Seventeenth-Century Dutch Prints," Simiolus 16 (1986): 113-14.

54. For example a miniature by the Master of the Saint Bertin Altarpiece in the British Library's Trésor des histories (Cotton ms. Augustus V., fol. 334v) depicts an interior with a motto written on the rear wall and the painted inscription "Ave Maria" on the fireplace canopy. Similar wall inscriptions, sometimes with dates, are so characteristic of another Flemish miniaturist active about 1480 that he is known as the Master with White Inscriptions; see Dogaer, 124-25.

55. See Andrew Martindale, The Rise of the Artist in the Middle Ages and Early Renaissance, London, 1972; and John Larner, Culture and Society in Italy, 1290-1420, New York, 1971, 264-84. For the Paris inscription, see Eugène Viollet-le-Duc, Dictionnaire raisonné de l'architecture française du XI ^e au XVI ^e siècle, Paris, 1854-68, 1:111.

56. A nineteenth-century survival, this forced and implausible chronographic reading requires the arbitrary suppression of one of three D 's in the first part of the inscription, that W be read as 10, and that J, I, and Y all count as 1. And because both dates are already given in Arabic numerals, the supposed chronograms would in any event be pointless. Panofsky's claim that "Leeuw," if written out, would have added an L, V, and W, thus destroying the chronographic reading, is taken from Weale, 87; Panofsky, 1953, 1:198.

57. The thirteenth edition consists of three supplementary volumes to the twelfth edition of 1922; the biography of Freud is found in 2:116.

58. See for example Susan Koslow, "The Curtain-Sack: A Newly Discovered Incarnation Motif in Rogier van der Weyden's Columba Annunciation," Artibus et historiae 7 (1986): 9-33. What the author calls a "curtain sack," i.e. the drapery panel of a hung bed rolled up for convenience' sake when not in use (the fabric of a hung canopy for a chair of estate was also sometimes folded up in exactly the same way; see Fig. 39), is interpreted symbolically on the basis of visual similarities with a curd sack and the shape of the womb.

59. See in particular Panofsky, 1953, 1:179-80.

60. Weale, xlii-xliii; see also lix for the description of Van Eyck as an "excellent maistre en art de painture" by the author of the eyewitness account of the Burgundian embassy that made the arrangements for the duke's marriage to Isabel of Portugal.

61. For the Latin text of Fazio's De viris illustribus, see Baxandall, 103.

62. Fazio's characterization of Van Eyck ("litterarum nonnihil doctus. geometriae praesertim et earum artium quae ad picturae ornamentum accederent") seems no more than a reminiscence of Pliny's comment on Pamphilus ("primus in pictura omnibus litteris eruditus, praecipue arithmetica et geometria, sine quibus negabat artem perfici posse"; Historia naturalis 35-76), with the important difference, however, that Pliny makes explicit his point about the learning of Pamphilus, whereas Fazio, writing of Van Eyck, does not.

63. Panofsky, 1953, 1:2, 179-80. Panofsky's translation of Fazio is unduly free so as to support his point of view; it should be compared with that of Baxandall, 102.

64. L. Bril and E. Lejour, "Les oranges dans nos provinces au XIV ^e et au XV ^e siècles," Archives, bibliothèques et musées de Belgique 26 (1955): 56-59.

65. On such carpets as symbols of status and the evidence that some at least were of European manufacture in the fifteenth century, see Walter B. Denny, "Rugs and Carpets," Dictionary of the Middle Ages, 10 (New York, 1988): 546-52.

66. Baldass, 76.

67. See for example Meiss, 1969, 31-32.

68. Marrow, 161.

69. Despite controversy about the details, there is general agreement that Van Eyck achieved his spatial constructions empirically rather than geometrically. For the recent literature, see James Elkins, "On the Arnolfini Portrait and the Lucca Madonna: Did Jan van Eyck Have a Perspectival System?" Art Bulletin 73 (1991): 53-62.

70. As is well known, Petrus Christus was the first northern painter to employ a single vanishing point, but he was also Jan's principal follower in the use of objects rationally disposed to create the illusion of pictorial space (see Fig. 40). Elements of the same technique are readily apparent in the anonymous Turin Virgin and Child (see Plate 12).

71. See for example, Bosshard (as in the following note), 10.

72. Emil Bosshard, "Revealing van Eyck: The Examination of the Thyssen-Bornemisza Annunciation," Apollo 136 (1992): 4-11. For references to other examples of alterations by Van Eyck to enhance the three-dimensional quality of his works, see ibid., 11 n. 15.

73. Pliny, Historia naturalis 35.81-83.

74. Baxandall, 102-3.