1. J. Williams, ''Sarnath Gupta Steles." My "From the Fifth to the Twentieth Century and Back," moreover, suggests a rationale for what might be called ethno-art history.

2. Another dimension of the problem is the diverse placement of the picture: displayed on the wall (rare, but not unknown, in Rajasthan or in Orissa); collected with unrelated pictures in an album (a Mughal tradition that flourished in the later Rajput context); and as part of a book or numbered sequence of pictures (the case for the Rasikapriya and for Orissan illustrated manuscripts).

3. For a lucid analysis from the viewpoint of the maker, see W. Eisner, Comics and Sequential Art . The Amar Chitra Katha comics of India, which deal with a wide range of traditional myth, literature, and history, deserve a comparable account of their own.

4. A. K. Ramanujan, "Three Hundred Ramayanas^[*] ," 46.

5. For example, N. Goodman, "Twisted Tales; or, Story, Study, and Symphony." There is much scholarly writing about narration in ancient and early Christian art. Because such works raise doubly complex questions about applying both verbal and pictorial models from the West to Indian forms, I shall not allude to them, preferring to make my own mistakes.

6. This is in general the stance of Vidya Dehejia, who concludes her systematic account of this topic in early Buddhist sculpture as follows: "Ajanta's narrative networks must have required a competent 'reader' in order to function in the manner in which they were intended" ("On Modes of Visual Narration in Early Buddhist Art," 392).

7. R. Barthes, "Introduction to the Structural Analysis of Narratives," 258-59.

8. Barthes, "Introduction," 251. Yet in distinguishing articulation (of separate units) from integration (gathering these "into units of a higher rank") he reverts to an Aristotelian hierarchy Art historians may recall a comparable distinction between additive and organic treatments of the human body, a distinction that likewise accepts the formula of Athens or of Renaissance Florence as superior to all alternatives. In general Barthes struggles against the weight of Western narratological (and aesthetic and historical) assumptions in an instructive manner.

9. Barthes, "Introduction," 252-53, where he refers to "the storyteller's (the author's) art, talent, or genius—all mythical forms of chance"; and 293-95.

10. G. Genette, Narrative Discourse: An Essay in Method ; see also his Narrative Discourse Revisited .

11. M. Bal, Narratology , 102-14; cf. Genette, Narrative Discourse Revisited , 72-78.

1. P. Richman, Many Ramayanas^[*] ; R. Thapar, "The Ramayana Syndrome."

2. I find Sahoo's work more helpful than other surveys of this kind in Hindi and Oriya.

3. Bulke acknowledges his debt to K. C. Sahoo, his student, for Oriya references (238-43 in the 3rd edition).

4. W. L. Smith focuses upon Balarama^[*] Dasa and Visvanatha Khuntia^[*] for Orissa, dismissing Upendra Bhañja as very difficult (34). His entire framework seems to privilege Valmiki as canonical. Nonetheless this work includes invaluable references to "apocryphal" vernacular variants in eastern India.

5. A. K. Ramanujan describes this role as faithful or iconic ("Three Hundred Ramayanas^[*] ," 44). I shall not use the semiotic terminology of Charles Sanders Pierce because "iconic" has such divergent connotations in common usage.

6. These dates, and some alternatives, are cogently presented by Robert Goldman in his introduction to vol. 1 of the: Princeton University Press translation of Valmiki.

7. Valmiki, Ramayana^[*] , ed. G. H. Bhatt.

8. The eastern recension is most easily accessible in the text and Italian translation of Gaspare Gorresio.

9. For a fuller discussion of the theory of eight (or nine) rasas and the distinction between this kind of abstracted sentiment and bhava , or actual human emotion, see D. H. H. Ingalls, An Anthology of Sanskrit Court Poetry , 13-17.

10. W. L. Smith, Ramayana^[*] Traditions in Eastern India , especially chap. 3. Smith's argument for the Oriya reflects his focus upon Balarama^[*] Dasa, and Smith's summary of Valmiki's structure includes precisely those events upon which the eastern vernacular versions do agree, omitting others.

11. At least twenty-five appear under the title Arsa^[*] (= Risi^[*] ) Ramayana^[*] in the Orissa State Museum: K. Mahapatra, Descriptive Catalogue of Sanskrit Manuscripts of Orissa , 3:6-15. One, catalogue no. 13, was sent to Baroda for the compilation of the critical edition of Valmiki's Ramayana^[*] . The term Arsa^[*] Ramayana^[*] may also designate translations of Valmiki into Oriya.

12. B. C. Mazumdar, Typical Selections from Oriya Literature , 1:xxiii. It has been argued, however, on the basis of passages that are directly translated from the Sanskrit, that Balarama^[*] Dasa must actually have known Valmiki in written form: K. C. Sahoo, "Indian Rama Literature and Jagamohan Ramayana," 176.

13. P. Lutgendorf, "Ramayan: The Video." Lutgendorf stresses the relationship to the Ramcharitmanas of Tulsi Das but discusses ways in which the serial departed from it as it both drew upon diverse traditions and remained embedded in concerns of the 1980s.

14. P. C. Bagchi, Introduction to Adhyatmaramayanam^[*] , 76. Cf. K. Bulke, Ramkatha , 166-67; F. Whaling, The Rise of the Religious Significance of Rama , 113 (denying the tradition that Ramananda came from the south); J. L. Brockington, Righteous Rama: The Evolution of an Epic , 252-57; B. L. Baij Nath, The Adhyatma Ramayana (although this translation is not always accurate, because of its accessibility my text citations follow its numbering system).

15. K. Bulke, Ramkatha , ch. 12.

16. G. S. Ghurye, Indian Sadhus , 172-74. Whaling, Rise of the Religious Significance of Rama , 177. While the Adhyatma Ramayana^[*] has been ascribed to Ramananda, the founder of the sect, this attribution is open to question. Evidence about the Ramanandins is conflicting; for instance, they are said to follow a qualified dualism, although the Adhyatma is unabashedly monastic (Ghurye, 165).

17. F. Whaling, Rise of the Religious Significance of Rama , 217, and chaps. 14 and 16 in general.

18. P. C. Bagchi, Introduction, 60.

19. F. Whaling, Rise of the Religious Significance of Rama , 198.

20. K. C. Sahoo, "Oriya Rama Literature," 175-76. I have worked from the Sharada Press translation of Gopala's translation of the Adhyatma Ramayana^[*] , published in Berhampur, n.d.

21. The illustrated one, from the Dasavatara Matha^[*] near Jajpur, Cuttack District, is discussed below in Chapter 2. A second, from Madhupur, also Cuttack District, is in the Orissa State Museum, Bhubaneswar, no. P/182A: K. Mahapatra, Descriptive Catalogue , 3:38.

22. The only copy I have seen is in the Library of the City Palace, Jaipur (Rajasthan). G. N. Bahura, Literary Heritage of the Rulers of Amber and Jaipur , 65 (no. 2767). This Brahma Ramayana^[*] , like the illustrated Oriya copy, comprises only five books (numbered 13 to 17), although Bulke's account suggests that more complete manuscripts must exist.

23. K. Bulke, Ramkatha , 178-79.

24. Lutgendorf discusses this religious movement, known as rasik sadhana , making clear that the form devoted to Rama is not derivative from the more familiar Krisna^[*] form ("The Secret Life of Ramcandra," 219-28).

25. M. Mansinha, History of Oriya Literature , 50-69. I concur with K. C. Sahoo ("Oriya Rama Literature," 31-36) in rejecting the Vilanka Ramayana^[*] as a work of the same author. In the case of both Sarala and Balarama^[*] Dasa, I make no pretense of having read the entire work in Oriya.

26. For example, Dasaratha is cursed by the divine cow in the same way that Dilipa was Kalidasa's Raghuvamsa^[*] (W. L. Smith, Ramayana^[*] Traditions in Eastern India , 20).

27. C. Das, A Glimpse into Oriya Literature , 65.

28. M. Mansinha, History of Oriya Literature , 62-63.

29. Smith, Ramayana^[*] Traditions in Eastern India , 56 (although Sarala describes the demon as meditating in an anthill, not a mound of earth as Smith says). I cannot share Smith's assumption of direct influence from the Jain Paumacariyam , as opposed to shared oral sources.

30. K. Bulke, Ramkatha 238-39.

31. W. L. Smith, Ramayana^[*] Traditions in Eastern India , 38-40.

32. B. C. Mazumdar, Typical Selections from Oriya Literature , 1:xxiii.

33. K. C. Sahoo, "Indian Rama Literature," 176. The tellers (Brahma, Siva, Jagannatha, and Valmiki) are not identical with Tulsi Das's. Cf. Philip Lutgendorf, "The View from the Ghats."

34. Sahoo, "Oriya Rama Literature," 93, and "Indian Rama Literature," 176-78. In general Sahoo stresses the plurality of sources synthesized by Balarama^[*] Dasa.

35. K. Bulke, Ramkatha , 241. W. L. Smith, Ramayana^[*] Traditions in Eastern India , 121, 125. For example, the curse of the Phalgu River for the rapacious behavior of the Gaya brahmans (an antiestablishmentarian theme) occurs in Bengali written texts. The Sabari's gift of fruit she has tasted and the assistance of a squirrel or mouse in building the bridge to Lanka^[*] are found in Telugu folklore.

36. M. Mansinha, History of Oriya Literature , 96.

37. W. L. Smith, Ramayana^[*] Traditions in Eastern India , 167.

38. K. C. Sahoo, "Oriya Rama Literature," 67-70; K. Bulke, Ramkatha , 239-90.

39. L.D. Institute (Ahmedabad) Ms. 20, discussed below in Chapter 2.

40. Nilambara Dasa, sixteenth century; Mahesvara Dasa, mid-seventeenth century W. L. Smith, Ramayana^[*] Traditions in Eastern India , 33.

41. K. C. Sahoo, "Oriya Rama Literature," 36. W. L. Smith, Ramayana^[*] Traditions in Eastern India , 142-44. This episode had occurred in the Sanskrit Adbhuta Ramayana^[*] .

42. K. C. Sahoo, "Oriya Rama Literature," 167.

43. The headquarters of Ghumsar was called Russelkonda from 1837 until after India's independence, when it was renamed Bhanjanagar. Royal families titled Bhañja go back to the fourth century A.D.

44. K. C. Sahoo, "Oriya Rama Literature," 138-39.

45. Smith gives Upendra's dates as 1670 to 1720, a year that seems too early for his death ( Ramayana^[*] Traditions in Eastern India , 33). Kedarnath Mahapatra suggests that he lived until 1753 ( Khurudha Itihasa ), as does Dandapani Behera ( Freedom Movement in the State of Ghumsar in Orissa , 6). Upendra's claims to have undertaken tantric practices to obtain poetic power need not be taken at face value (M. Mansinha, History of Oriya Literature , 116). Mansinha presents a generally positive view of the poet's work, whereas B. C. Mazumdar is broadly damning ( Typical Selections from Oriya Literature , 2:xvi-xxv).

46. M. Mansinha, History of Oriya Literature , 118.

47. I have found no correspondence between this story and various Sanskrit tales in which the names Lavanyavati^[*] and Chandrabhanu appear.

48. The performer is called an indrajalaka , usually translated "magician," although the illustration in Raghunath Prusti's manuscript of the Lavanyavati^[*] , which I discuss later in Chapters 3-5, shows a troupe of people (Figure 193), perhaps something like the Gujarati bhavai , which consists of conjuring and other entertainments, including folk drama.

49. B. C. Mazumdar, Typical Selections from Oriya Literature , 2:xii. Sahoo, "Oriya Rama Literature," 158-61. Smith, Ramayana^[*] Traditions in Eastern India , 34.

50. W. L. Smith, Ramayana^[*] Traditions in Eastern India , 148-51.

51. Daniel H. H. Ingalls originated this translation (a propos of Krisnalila^[*] ), the implicit pun ("jest") referring to the literal meaning of lila , "sport."

52. Dashahra designates the tenth day of the bright half of the lunar month of Asvina. Ramanavami is the ninth clay of the bright half of the lunar month of Chaitra. A life of Chaitanya suggests that in the sixteenth century Ramalila^[*] took place in Purl at Dashahra time (N. Hein, Miracle Plays of Mathura , 109-10).

53. K. C. Sahoo mentions him in "Oriya Rama Literature." In 1983 I was told that Vikrama Narendra was from Ghumsar. The text on paper now in use was copied from a palm-leaf manuscript about twenty years ago.

54. N. Bisoi, Dasapalla Itihasa (in Oriya), 10. The temple bears a plaque that dates its consecration to December 23, 1903. From 1884 to 1886 the Oriya writer Fakirmohan Senapati was diwan of Dasapalla, and his autobiography says that most of the rural population of the state was Khond or Khaira, i.e. aboriginals ( My Times and I , 81).

55. For the origin of this Laksmana-rekha^[*] in a minor work by Tulsi Das and its propagation as part of restrictions put on women, see U. Chakravarti, "The Development of the Sita Myth: A Case Study of Women in Myth and Literature," Samya Shakti 1, no. 1 (1983), 73.

56. Local legend has it that the headman of Bisipada 160 years ago had seen the Dasapalla Ramalila^[*] and wished to emulate that. Rama came to the brahman Janardana Dasa, who agreed to write a text and required that he be allowed to meditate for twenty days uninterrupted in a temple near the town. At the end of that time, the villagers found that he had vanished, leaving a completed manuscript.

57. This canopy was made by a visiting artisan from Pipli, the Muslim appliqué center between Bhubaneswar and Puri. The canopy must already have been old in 1923, the date of repair embroidered on it. Concerning the Pipli tradition in general, see B. C. Mohanty, Appliqué Craft of Orissa .

58. See Chapter 2 for the story of Jagannath Mahapatra of the pata^[*] tradition. Vaisya Sadasiva may have lived in the later eighteenth century (K. C. Sahoo, "Oriya Rama Literature," 192-93); his text is widely available today.

59. In no Orissan version have I seen the systematic conversion of an entire town into the landscape of the play that characterizes the peripatetic Ramalila^[*] of Ramnagar in Uttar Pradesh. Cf. S. Bonnemaison and C. Macy, "The RamLila in Ramnagar." Thus in Dasapalla, the stage serves as Chitrakuta^[*] until Rama leaves to shoot the illusionary deer, when suddenly audience and actors move to a spot 300 meters away and the hut is relocated. After this, Lanka^[*] is generally located 500 meters further down the road, although some action in Lanka^[*] also takes place on the original stage.

60. This Ramalila^[*] begins on Aksaya^[*] Tirtha, the third day of the lunar month of Vaisakh (following Chaitra), a significant moment in the Jagannatha ritual calendar. My source for this information is Purna Chandra Mishra of Purl.

61. My source for this information is the late lamented Dhiren Dash of Bhubaneswar.

62. The manuscript in the National Museum, New Delhi, is discussed in Chapter 2. Cf. K. C. Sahoo, "Oriya Rama Literature," 167.

63. My source for this information is Raghunath Panigrahi of Chikiti, who had himself acted under the guidance of his father, a major singer ( gayaka ) and exponent of the tradition, in which the last king's son also took part. The paper manuscript in use in the 1950s, preserved by Raghunath Panigrahi, appears to be a version of Krishna Chandra Rajendra's text, simplified by a later ruler. After the Ramalila^[*] ceased, Jatra troupes continued to perform occasionally in Chikiti.

64. I visited this village on April 14, 1990, and my information comes from several local people, who said that the performance used to be based on Vaisya Sadasiva, Vikrama Narendra, and Ananga Narendra (another nineteenth-century author). The patrons were local Khandayats, but there was no royal family. In Dasapalla, the Hanumana mask is likewise kept in the Mahavir shrine and worshiped.

65. The Temple Endowment Board kindly allowed me to check their schedule of performances, which goes back to 1977. Some variation might be explained by weather, although in 1990 in the event of rain they read the text while canceling the enactment, sticking to schedule. Moreover, Thursday is an inauspicious day and performance was often, but not consistently, suspended then, spreading the entire Ramalila^[*] over a longer period.

66. Such variations occurred both in Dasapalla, as documented by records, and in Bisipada, where in 1983 I saw the breaking of Siva's bow combined with Rama's marriage, whereas in 1990 the first event had taken place on the previous night and the marriage led directly to the encounter with Parasurama.

67. In the printed program, Bharata's visit was listed for the seventh night, but this was said to be a mistake. Some such variations may result from the direct reading of the text by the gayakas , who literally have the last word, whereas the program is the work of the temple administrator. This division need not imply disagreement; it may rather indicate an accepted diversity of function.

68. The kidnapping must occur on the full moon of Chaitra, a requirement in Danda^[*] Jatra as well.

69. Killing the rhino is in fact a major event in Bisipada, where it occurs on tire main stage, rather than in procession. Gania, near Dasapalla, borrowed Vikrama Narendra's text but does not perform this event. The rhino itself is probably not the key to the incident. Although rhinos have not been found in Orissa for at least a century, people still speak of rhino meat as a precious substance suitable for sraddha . Those early dharmasastra texts that permit the consumption of meat mention that the ancestors are particularly gratified by the offering of rhino flesh: P V. Kane, History of Dharmasatra , 4:422. Cf. J. Bautze, "The Problem of the Khadga^[*] ."

70. L. Hess, "Ram Lila: The Audience Experience." R. Schechner, Between Theatre and Anthropology , chap. 4, "Ramalila of Ramnagar." Lutgendorf makes clear the variety of performances in the Banaras area, demonstrating that the smaller ones are not merely scaled-down versions of Ramnagar ( The Life of a Text , chap. 5).

71. Maya Sita must be played by a brahman in Dasapalla, thus intensifying Ravana's^[*] sin, that he kidnaps a brahman woman. The narrow stage in this portion of the performance reminds me of Prahlada Nataka^[*] , the drama of Narasimha's victory over Hiranyakasipu, which flourished in Ganjam District, where the innocent victim Prahlad paces up and down in a similar passageway in the midst of the crowd.

72. The effort of hoisting the image involves a large number of youths and carpenters and is part of the whole drama of Ramalila^[*] . The effigy is not supposed to be raised before dusk; one year when it was raised earlier in the day to accommodate a visit by the chief minister, it fell over.

73. My information is dependent upon the late Dhiren Dash of Bhubaneswar (who kindly told me about the form initially) and upon Omshankar Sarangi of Asureshvar, which I visited during the daytime. I have not actually seen the evening performance, nor has anything been written about it.

74. N. Hein, The Miracle Plays of Mathura , 17-30; P. Lutgendorf, The Life of a Text , 256-57.

75. There are thirty masks for rent today, according to the temple authorities. The vow begins with the kidnap of Sita on the full-moon day, when Hanumana's services become necessary.

76. D. Mukhopadhyay, "Sahi Yatra." Mukhopadhyay suggests that Sahi Jatra is linked to many of Jagannatha's vesas , or festive ornaments, throughout the year, although his suggestion is not borne out by my own observation in Puri. It is difficult to get a clear account of the performance from the temple priests, for they are not centrally involved. My impression is that Sahi Jatra varies considerably from year to year, its form determined in part by the initiative of the sahis and akhadas^[*] themselves. The performances are related to the military skills of the town, in the past fostered by the Purl Temple for its own defense; hence the prominence of Naga warriors.

77. The repertoire here has been shaped by cosmopolitan royal dancers and patrons, influenced by Uday Shankar and Javanese troupes that visited Calcutta. I am unable to see any particular connection between Seraikela or Baripada Chhau as they survive today and one illustrated manuscript produced in this area, the Baripada Vaidehisa Vilasa^[*] of Satrughna, discussed in Chapter 2. For example, in Seraikela Chhau, Risyasringa's^[*] mask consists of a human head with the ascetic's hair twisted upward to form a conical knot that vaguely resembles a horn; Satrughna depicts the sage with the full head of a two-horned blue antelope.

78. D. Dash, Jatra , 13-15, and Pala Itihasa Pala , an ingenious history of Pala in Pala-style Oriya verse.

79. J. Pani, Ravana Chhaya . The late puppet master Kathinanda Das preserved a set of puppets from the Mahabharata, although the equation of the entire genre with the name "Ravana's^[*] Shadow" suggests the centrality of the Ramayana^[*] . I saw performances in 1982 and 1983.

80. J. Pani, Ravana Chhaya , 21-31, provides a detailed summary of the episodes correlated with the text. Pani notes that the Uttara Kanda^[*] used to be performed, although it is not today

81. N. K. Sahu, state editor, Balangir District Gazeteer , 493-96. There are other shrines to Lankesvari^[*] in the region, for example a larger temple at Sambalpur. While one may not accept the historical identification of Sonepur with an actual Lanka^[*] in the epic, this theory points to an interesting cluster of local beliefs.

82. P. Richman, "E. V. Ramasami's Reading of the Ramayana^[*] ."

1. For a full account of attitudes associated with this seemingly archaic form and of the process of bookmaking, see J. P. Das, Chitra-pothi , chaps. 2 and 5. Incised manuscripts, generally without pictures, were made throughout south India, although at least one illustrated fifteenth-century Adhyatma Ramayana^[*] is known (personal communication, Sobha Menon, Dept. of Art History, M.G. University, Baroda). The illustrated manuscripts of Bali (Indonesia) are engraved with a knife rather than the pointed stylus used in Orissa.

2. It is possible that another person added pigment to the palm-leaf pictures, although sometimes the scribe/illustrator himself did so, as in the case of Balabhadra Pathy, discussed on page 55 (J. Williams, "Jewels from Jalantara").

3. John Beams's statement that not one man in a hundred was literate in Orissa ( Comparative Grammar of the Modem Aryan Languages of India , 1:89) is an assessment of ability to write orthographically and grammatically correct (i.e. standard) Oriya; Beams did not consider the nature of a vernacular language in an age before printed tools provided standards for uniformity Oriya, moreover, interfaced with other languages on all three sides of Orissa. Surely the audience of readers was fairly wide, however varied their ability to write.

4. The 1936 account of how Raghunath Prusti's illustrated Ragamala (N.B.: not a sacred text) was kept by his heirs vividly describes this smearing with sandal paste: Kalicharan Patnaik, Kumbhara Chaka , 258-59; J. P. Das and J. Williams, Palm-Leaf Miniatures , 6.

5. Cf. J. Williams and J. P. Das, "Raghunatha Prusti^[*] : An Oriya Artist," plates 29-31 (a work by an artist most of whose oeuvre fared better).

6. J. P. Das, Chitra-pothi , chap. 4. Unillustrated manuscripts were made for Rs 1.25, in exchange for a window frame, or for rates such as an anna (1/16 of a rupee) per chapter. Brajanatha Badajena^[*] asked the Raja of Dhenkanal for Rs 100 for an illustrated work. J. P. Das and J. Williams, Palm-Leaf Miniatures , 8.

7. J. P. Das, Chitra-pothi , 36, 48-49; J. P. Das and J. Williams, Palm-Leaf Miniatures , 21.

8. " Tikam^[*] puttalikanvitam^[*] vililikhitam^[*] Srigitagovinda ." This manuscript is the work of Dhananjaya^[*] , discussed later in this chapter in connection with its date. J. P. Das, Chitrapothi , 38.

9. J. Williams, "Jewels from Jalantara" (the work of Balabhadra Pathy discussed below).

10. P. K. Mishra, The Bhagavata Purana^[*] ; J. P. Das, Chitra-pothi , 41. The colophon concludes, "E pothi astama^[*] navama charita pittula mo kama. Ghanasyama je putra mora ta, lekha pus-taka aksara^[*] ."

11. Other karana^[*] illustrators included G. B. Pattanayaka and Chakradhara Mohanty. See S. Pani, Illustrated Palmleaf Manuscripts of Orissa , 81-83.

12. Orissa State Museum Ext. 166; J. P. Losty, The Art of the Book in India , 137; S. C. Welch, India , 62; S. Pani, Illustrated Palmleaf Manuscripts , 19-22;J. P. Das, Chitra-pothi , 38; J. P. Das and J. Williams, Palm-Leaf Miniatures , 40. That the aksara^[*] a occurs only in the archaic form in this work does not rule out a late date but makes it less probable. The aksara^[*] a occurs in its modern form in the Radhakrisna^[*] Keli manuscript, British Library Or. 11612, which appears to be another work of Dhananjaya^[*] , as Losty has pointed out (loc. cit.). The use of such an a again is improbable but not impossible in the late seventeenth century.

13. Orissa State Museum Ext. 46, now renumbered Ext. 334. K. Vatsyayan, "The Illustrated Manuscripts of the Gita-Govinda from Orissa," 279, plate CVIIIa. The older form of a occurs in the text here, but the modern one occurs in the colophon.

14. His anka^[*] 4 (= year 3) is also given as the date when the manuscript was completed. The sole Harikrisna^[*] Deva of Khurda ruled from 1715 or 1716 until 1720. This entire colophon remains problematic and may possibly be a later addition to the text.

15. D. P. Ghosh, "Eastern School of Mediaeval Indian Painting (Thirteenth-Eighteenth Century A.D. )," figures 215-17, 330-36. These paintings on wood are also undated.

16. See note 10 above. The colophon has been read as the twenty-third regnal year (nineteenth actual year) and equated with November 20, 1799: S. Pattanayak, Brajanatha Granthavali^[*] , viii. This information is repeated by J. P. Das ( Chitra-pothi , 41), by P. K. Mishra ( The Bhagavata Purana^[*] , 12), and by J. P. Das and J. Williams ( Palm-Leaf Miniatures , 40). In 1795 the seventh day of the dark fortnight of Karttika began on Wednesday, as the colophon says; but in 1799, the date Pattanayak gives, it did not (cf. L. D. Swamikannu Pillai, An Indian Ephemeris , 6:393, 401). In any case 1799 is not the twenty-third regnal year of any Khurda ruler, and Brajanatha is usually said to have died in 1798.

17. The former is in fact known as a karani^[*] form; it is used exclusively in the archaic scribal script employed on paper.

18. Mishra, The Bhagavata Purana^[*] . I am indebted to Dr. J. P. Das for patiently checking my observations about handwriting and for discussing these issues with me.

19. The "modern" a also occurs in a Radhakrisna^[*] Keli (British Library Or. 11612), which has been ascribed to the same Dhananjaya^[*] on the basis of style (J. P. Losty, The Art of the Book in India , 137).

20. S. Kramrisch, "The Hundred Verses of Amaru Illustrated," 226. Cf. S. Mahapatra and D. Pattanayak, Amarusatakam , 2. Kramrisch is to be applauded for recognizing the interest of this entire subject. It is small wonder that for lack of a local framework in which to place the manuscript, she resorted to a pan-Indian one.

21. J. P. Das, Chitra-pothi , 37. In addition to the four manuscripts discussed here, at least one more Adhyatma Ramayana^[*] by Sarathi Madala has been partially published in Bansidar Mohanty's Odiya^[*] Sahityara Itihasa , 1966.

22. Orissa State Museum Ext. 81, 136 folios, 37.5 × 4.2 cm. The nineteenth anka^[*] is equivalent to the sixteenth year. Thus because the reign of Divyasimha III of Khurda began in A.D. 1859, this is the same as samvatsara (here indicating Christian year) 1875, when the solar date, lunar date, and day of the week also concur. Sana (Muslim or Hegira era, adding 593) 1284 (= A.D. 1877) appears to be miscalculated.

23. Spencer Collection, Indian Ms. 14, 146 folios, 29 × 4 cm. The colophon gives the dates samvat 1891, fasli 1298 (= A.D. 1891), Saka year 1812 (= A.D. 1889/90), anka^[*] 12 of Mukunda Deva of Khurda (the beginning of whose reign is often taken as 1880), and anka^[*] 51 of Laksminarayana^[*] Deva of Badakhemundi.

24. The C. L. Bharany copy has 182 folios, 26.5 × 4.5 cm. The colophon gives the dates 1301 fasli (= 1894, presumably wrong), samvatsara 1891, December 2, Wednesday (the last two concurring in 1891), eighth day of the bright half of Margasira (not concurring with December 2 or Wednesday in either 1891 or 1894), fifty-second anka^[*] of Laksminarayana^[*] Deva of Badakhemundi.

25. OL 121, 88 folios (the last numbered 155), 37 × 4.5 cm. This colophon reads " samvat 1992 [which appears to be a slip of the pen for 1892], thirteenth anka^[*] [= eleventh year] of Kripamayadeva^[*] [of Badakhemundi], the first of Kumbha, the fourth day of the bright half of Magha, Tuesday." Kripamayadeva^[*] apparently came to the throne in 1891 (after December 2). In 1902 the first day of the solar month Kumbha was indeed the fourth day of the lunar month Magha, but this was a Wednesday Here, as in most of Sarathi Madala's colophons, there seems to be confusion about some part of the complex dating system; such confusion occurs more consistently in his work than in that of any other scribe I have encountered.

26. National Museum 75.536, 53 folios, 33 × 4.5 cm.

27. C. L. Bharany, 237 folios, 36.5 × 4 cm. This version is more sparsely illustrated than any other Adhyatma Ramayana^[*] I have seen.

28. Thirteen folios preserved, original pagination 189-201 (which was not the end of the text), demonstrating that the earlier books were also included; 26.3 × 4 cm.

29. L.D. Institute Ms. 20 (currently catalogued as Adhyatma Ramayana^[*] , A.D. 1703), 34 folios, 24 × 4 cm.

30. National Museum 75.556, 58 folios (the last of which is paginated 60), 39.5 × 4.5 cm.

31. This manuscript has 281 folios, 28.5 × 3.8 cm. The Jubel Library in Baripada recorded that this manuscript was collected in Kuamara (Mayurbhanj District) by Padmasri Padmananda Acharya in the 1930s. It was mentioned by Klaus Fischer in his article "Orissan Art in the Evolution of Postmediaeval Indian Culture," 27. It is now in the Orissa State Museum in Bhubaneswar.

32. The closest parallel I know is the Zurich manuscript of the Rasika Haravali , where in two cases a band of foliage frames a rounded section of text (E. Fischer and D. Pathy, Die Perlenkette dem Geliebten , 34).

33. J. P. Das and J. Williams, Palm-Leaf Miniatures , plate 76.

34. Samvat here must indicate the Muslim or fasli era, beginning in 593. The nineteenth anka^[*] is equivalent to year 16, and Ramachandra Deva III of Khurda came to the throne in 1817. Here the colophon dates referring to two different eras can be reconciled, unlike Sarathi Madala Patnaik's.

35. The Rietberg folios are 39 × 4.4 cm. (as are the Victoria and Albert and the New Delhi leaves). R. Skelton, Indian Miniatures from the Fifteenth to Nineteenth Centuries , 23-25, plate 5. E. Fischer, S. Mahapatra, D. Pathy, Orissa , 245.

36. The Victoria and Albert leaf is unpublished. One side bears text alone; the other depicts Rama and Sita in Panchavati and the encounter with the crow.

37. The New Delhi leaves are unpublished; folios are numbered 22 (Ravana^[*] and gods), 66 (my Figure 100), and 104 (battle between monkey and raksasa^[*] in chariot).

38. The poem has 110 folios, 17.5 × 4.2 cm. The name Karana^[*] Satrughna is legible on the final leaf.

39. K. Das, Kavisurja Granthavali^[*] , 9. I am indebted to Dr. J. P. Das for this reference and for invaluable assistance in reading the colophons.

40. The first Vaidehisa Vilasa^[*] by Michha Patajoshi has 215 folios, 40.5 × 5.5 cm. The artist is identified as Michha, son of Krishna Mishra Sharma. In one place samvatsara 192 appears, and in another Christian Era 1102 is written. I would deduce that Michha was young at this point (not yet given the title Patajoshi), hence mistakes in the figures for which he was responsible. The date also includes anka^[*] 24 (= year 20) of Mukunda Deva of Khurda, which might be anytime between A.D. 1896 (counting from 1876, his actual accession) and 1908 (counting from 1888, his father's death).

41. Michha Patajoshi's second work has 151 folios preserved (originally 245 according to the pagination of colophon folio), 41.6 × 4.6 cm. Thins manuscript was acquired from the historian Kedarnath Mahapatra in Bhubaneswar. The 1911 date, definitely in Michha Patajoshi's own handwriting on the final folio, says that the patha^[*] (literally "reading") of the book was finished then. Afterward, a slightly different hand has written "the inscribing [ lekha ] was completed on the twelfth day of the dark half of Margasira, 1914." Possibly this indicates that the illustrations of the manuscript took over three years to complete.

42. Asutosh Museum Ms. 5, 302 folios preserved (the major manuscript had 283, according to the pagination of the colophon), 39 × 4.5 cm.

43. E. Fischer, S. Mahapatra, and D. Pathy, Orissa , 250. One wonders whether these and the interpolated leaves in the more: complete copies might not come from the same copies. My suggestion that these represent eight different versions comes from some duplication of the same events in various sequences, as well as a comparison of the rate of narration as indicated by the pagination. For example folio 117 in the Asutosh 1918-28 manuscript shows Rama and Laksmana^[*] with Kabandha, whereas an interpolated page 117 shows Rama, Laksmana^[*] , and Sita with Viradha, an event covered about 17 pages previously in the main sequence; thus this interpolated set would presumably have been substantially longer than 283 pages.

44. One is Bhubaneswar, Orissa State Museum, Ext. 13, 12 joined folios, measuring 51 × 31.5 cm. as a whole. The second, slightly smaller, is in the Asutosh Museum and is composed of ten joined leaves. Notes on both describe them as patas^[*] made by Michha Patajoshi of Balukeshvarpur; they are: dated only by month.

45. In all these works, there is consistency throughout the manuscript in the use of pigment as well as in figure style and composition. Thus I would infer that they were painted as part of the original production of the manuscript, perhaps by a single maker, rather than considerably later, and hence that the resemblance is not merely adventitious. Two of these works are erotic texts, and in the remainder many erotic scenes appear, perhaps as the forte of this artist or group of artists. One Chausathi^[*] Rati Bandha is in the Orissa State Museum (J. P. Das, Chitra-pothi , 93; S. Pani, Illustrated Palmleaf Manuscripts of Orissa , 63), and a second erotic text is in the collection of Sachi Rautroy in Cuttack. A Chitrakavya Bandhodaya is in the Orissa State Museum (OL 530; J. P. Das, Chitra-pothi , 83-84; S. Pani, Illustrated Palmleaf Manuscripts , 61-64).

46. That illustrated here in Figure 140 National Museum 72.109, measures 25.5 by 4.25 cm. British Library Or. 13720 measures 35.5 by 3.5 cm. In both cases the manuscripts are complete, and it is clear that the scribe chose to leave no space for pictures.

47. All measure ca. 41 × 4.5 cm. No pagination is preserved; I assume that only one folio from the Ramayana^[*] sequence (Surpanakha^[*] , the magic deer) is missing, although it is possible that more have been lost. In addition to the Ramayana^[*] sequence discussed here, folios from other parts of this manuscript are found in the following collections:

National Museum, New Delhi

Poonam Bakliwal, New Delhi

Jagdish and Kamala Mittal Museum, Hyderabad

Museum für Indische Kunst, Berlin I/5675, no. 130a, b (E. Waldschmidt and R. L. Waldschmidt, Indische Malerei , 196)

British Museum, London, 1963-11-11

Metropolitan Museum of Art, New York, 1974.220

Edwin Binney 3rd, San Diego, 53 N, O

George Bickford Collection, Cleveland (S. Czuma, Indian Art from the George E Bickford Collection , no. 132)

Anonymous private collection (S. C. Welch, Indian Drawings and Painted Sketches , 30-31)

Elvehjem Museum, Madison, Wisconsin (P. Chandra, Indian Miniature Painting in the Collection of Earnest C. and Jane Werner Watson , 10-11; J. P. Das and J. Williams, Palm-Leaf Miniatures , color plate L)

48. Some leaves lack any pigment. The great diversity of transparent washes and intense, opaque colors suggests that at least some pigment may have been added long after the manuscript was made.

49. For a discussion of previous daring (and confirmation of an early nineteenth-century position), see E. Fischer and D. Pathy, Die Perlenkette dem Geliebten , 53-54.

50. National Museum 80.1276, 197 folios, 33 × 4.5 cm. This Lavanyavati^[*] was acquired from the collection of Sachi Rautroy of Cuttack.

51. L.D. Institute Ms. 15, 28 folios, 28.5 × 4.5 cm. J. Williams, ''Jewels from Jalantara." Since writing this article, I have found another Vaishnava work by the artist in the Elvehjem Museum, Madison, Wisconsin (66.13.3).

52. National Museum 72.108, 32 folios, 26 × 4 cm. The British Museum has a folio from this same manuscript (1968-12-04), as does the Chandigarh Museum. The Usabhilasa^[*] is firmer in drawing and writing than the Lavanyavati^[*] or the Gita Govinda , and the facial type is slightly less elongated, suggesting that this may be an early work by the arrest, whose subsequent drawing is yet more fanciful and daring. Folios 12v and 15r show the distinctive treatment of a storm, although without the diagonal lines for some gusts of rain in Plate 10 (in this book), which seems to me an advance in dynamism.

53. I am grateful to Dr. John Emigh of the Department of Theatre Arts at Brown University for this information. See D. Dash, Jatra: The People's Theatre of Orissa , 26-27.

54. Dr. Dinanath Pathy informs me that Pathy is a very rare brahman name in Orissa, and that only one family of that name survives in Digpahandi and two in Bhadrak. One might infer that the artist was brought from the outside to the court of Jalantara.

55. For illustrations of comparative material, see E. Ray, "Documentation for Paithan^[*] Paintings," and V. Stache-Rosen, "On the Shadow Theatre in India."

56. In later Javanese temple reliefs, often described as in a "wayang" style, the interstices between figures are similarly filled in by natural and even abstract forms that have no direct counterpart in the shadow theater of Indonesia.

57. J. Mittal, Andhra Paintings of the Ramayana , e.g. figure 23 (the seven sal trees, with no serpent). Here royal figures consistently wear tall southern-style crowns, absent in Orissa generally and in Balabhadra Pathy's work in particular.

58. J. Williams and J. P. Das, "Raghunatha Prusti^[*] : An Oriya Artist"; J. P. Das and J. Williams, Palm-Leaf Miniatures . An additional work, a copy of the Artatrana^[*] Chautisa recently discovered in Digpahandi, has been published by E. Fischer and D. Pathy, Die Perlenkette dem Geliebten , 55-56.

59. Folios of this Lavanyavati^[*] measure 32.5 × 4.5 cm. All remain in the Subuddhi house in Mundamarai, except for two in the Bharat Kala Bhavan, Varanasi (10927 and 10928, Figures 189-92 here), and ten in the Rietberg Museum, Zurich (R. Skelton, Indian Miniatures from the Fifteenth to Nineteenth Centuries , 25-28). Since the original pagination is clear, we can be certain that folio 79 is missing.

60. J. Williams and J. P. Das, "Raghunatha Prusti^[*] " 138-39; J. P. Das and J. Williams, Palm-Leaf Miniatures , 32-38.

61. National Museum 72.165, I-II, 80 folios, 33.5 × 4 cm. Western pagination has been added along with the museum's acquisition number, visible at the left in my Figures 1-40. The new numbers are out of sequence and in many cases obliterate the original pagination, which I have reconstructed from those cases where it is visible. Fortunately there is no question about the actual sequence of folios, for the verse numbers are clearly visible in the text.

62. That colophon is part of National Museum 72.165, III-IV, 48 folios, 35.3 × 4 cm. The date is 1305 Sala^[*] , or A.D. 1898. This second manuscript lacks the Ramayana^[*] portion. The eyes of its figures open wide, with a dot for the pupil isolated in the center.

63. This work, 66 folios, 4 × 27 cm., belongs to the Dasavatara Matha^[*] of Jajpur subdivision. I am indebted to R. P. Das of Bhubaneswar for showing it to me and enabling me to photograph it. A notation in the text suggests that it was made by sakhi , which in the context of Visnu-bhakti^[*] implies a generic devotee rather than a female companion. (Williams and Das, "Raghunatha Prusti^[*] ," 137, plate 48.)

64. S. Pani, Illustrated Palmleaf Manuscripts , 31-38; J. P. Das, Chitra-pothi , 71, 101.

65. The most accurate account of the Raghurajpur procedures is J. P. Das, Purl Paintings , chap. 8. See also E. Fischer, S. Mahapatra, and D. Pathy, Orissa , 231-35 (using playing cards as an example); and B. C. Mohanty, Patachitras of Orissa , 11-15. My description here is in the past tense because the procedures today are rarely followed exactly as I describe them. For example, bazaar pigments are used except to demonstrate the older procedures to visitors; lac is rarely applied, and when it is, it takes the form of commercial shellac.

66. J. P. Das, Puri Paintings , 36-39, plates 3-6, color plates 1-3; B. C. Mohanty, Patachitras of Orissa , 6, plates 1-4; E. Fischer, S. Mahapatra, and D. Pathy, Orissa , 220-23.

67. These images follow more general rules for image making not peculiar to Jagannatha, who appears here as a standard divinity The frontal view of the heads follows patterns for depicting iconic deities that do not apply to the lesser figures usually shown in profile in patachitras^[*] .

J. P. Das, Puri Paintings , 81-83.

68. J. P. Das, Purl Paintings , 36.

69. J. Williams, "Marriage Paintings in South and Central Orissa."

70. J. P. Das, Purl Paintings , 55.

71. D. Pathy, Mural Paintings in Orissa , 7-8; Pathy draws on present oral tradition of the painters. The technique appears to resemble the fine plastering with ground shells known as gotai^[*] in Rajasthan.

72. J. Williams, "A Painted Ragamala from Orissa."

73. The chitrakaras' contributions to the Jagannatha Temple included painting parts of the carts built each year for Ratha Yatra.

74. In the 1950s, most painters were working as agricultural laborers and masons (J. P Das, Puri Paintings , 88).

75. J. P. Das, Puri Paintings , plates 23, 24, 26.

76. J. Williams, "Criticizing and Evaluating the Visual Arts in India: A Preliminary Example," 10.

77. J. Williams, "A Painted Ragamala from Orissa," figure 10.

78. Jagannath Mahapatra of Raghurajpur has a large notebook that includes 120 images; he made it roughly thirty years ago. Another made in 1928 by Kenduri Mahapatra of Nayagarh is in the hands of his nephew, Udayanath Mahapatra.

79. J. P. Das, Purl Paintings , 51.

80. The account I present here is documented in the records of the Virañchi Narayana^[*] Matha^[*] itself and is also found in T. C. Rath, Ghumsar Itihasa , 36-38. Virañchi (or Viriñchi, as it occurs in other parts of India) is an epithet of Brahma. For a systematic discussion of the Buguda murals, see D. Pathy, Mural Paintings in Orissa .

81. J. P. Das, Puri Paintings , chap. 9. Cf. figure 31 there and sages at Buguda such as Visvamitra or Parasurama (Figure 207 here).

82. D. Pathy, Mural Paintings in Orissa , 82.

83. That convention appears in the elephant-hunting friezes on the base of the Sun Temple at Konarak and in one illustration of the Amaru Sataka (S. Mahapatra and D. Pattanayak, Amarusatakam , verse 92). This resembles the Jain Mahapurana^[*] painted at Palam in 1540 (K. Khandalavala and M. Chandra, New Documents of Indian Painting — a Reappraisal , 74).

84. J. P. Losty, Krishna , plate 2; S. Czuma, Indian Art from the George P. Bickford Collection , plate 125.

85. E. Fischer, S. Mahapatra, and D. Pathy, Orissa , figures 456, 464, 578-84; D. Pathy, Mural Paintings in Orissa , plates 23-28.

86. E. Fischer, S. Mahapatra, and D. Pathy, Orissa , figures 457-58, 585-90; D. Pathy, Mural Paintings in Orissa , plates 27-37.

87. J. P Das, Puri Paintings , 54, plates 9-10. D. Pathy, Mural Paintings in Orissa , 86, plates 39-41. An inscription reads Sam[vat] 1323 , which Pathy ascribes to the Saka era (= A.D. 1400), although that seems unlikely; possibly the Sana or Muslim era is intended (= A.D. 1916). Pathy also reports recent repainting. In his plate 40, the mountains to the left are much more regular than those of Buguda.

88. G. N. Bahura and C. Singh, Maps and Plans from Kapardwara (Jaipur), no. 260. It remains questionable whether the royal worshiper depicted is Man Singh of Jaipur, who had ruled a century before the painting was collected by the Rajasthani brahman Manikram Paliwal. A pata^[*] of Jagannath dated 1670 in the Victoria and Albert appears to me to be entirely the work of Nepali artists.

89. J. Williams, "From the Fifth to the Twentieth Century and Back," figure 7. British Museum 1940-7-13, 0152, 32.4 × 58 cm. It was not included in Moor's book of 1810, The Hindu Pantheon . No catalogue of his own collection is preserved, and it remains possible that some other member of the family added it later.

90. J. P. Das, Puri Paintings , plate 27.

91. Boston, Museum of Free Arts 25.525, given to the museum in 1925.

92. M. Archer, Indian Popular Painting in the India Office Library , 123;J. P. Das, Puri Paintings , 171. Over the past fifteen years, I have found that painters rarely keep examples of old, completed work in their possession.

93. Bibliothèque Nationale 1041, 270 × 150 cm. L. Feer, "Introduction du Nouveau Catalogue," 371 (information that the collection was presented in 1894), 379 (description of this image, whose original provenance Feer did not know). The image is reproduced in color in S. Gole, Indian Maps and Plans , plate 23.

94. British Museum 1880-303, c. 65 × 35 cm. The reverse bears the initials of A. Franks, keeper of Oriental Antiquities, who retired in 1896.

95. See J. P. Das, Puri Paintings , 120-21. Other patas^[*] include Rama's pursuit of the magic deer or the cutting of Ravana's^[*] umbrellas in the upper right comer.

96. J. P. Das, Puri Paintings , chap. 7.

97. J. P. Das, Puri Paintings , plate 22 (the Ganesa^[*] ). Halina Zealey herself must have played a major role in the award, and her taste may be discerned in the pale, grayish palette of this picture, which is exceptional even for Jagannath Mahapatra's work. The refinement, however (for example in the middle border composed of rats), is not new or exceptional—cf. the Bibliothèque Nationale nineteenth-century painting (Figure 216).

98. Jagannath Mahapatra's notebooks contain an unillustrated list of thirty-four Ramayana^[*] subjects, all of which are included in the final set of seventy-five. The Uttara Kanda^[*] is entirely absent in the list. It is not clear whether this list goes back to before Zealey's days or is a subsequent condensed version of the large set.

99. I have been unable to locate this set in Delhi. Figure 220 is a photo of a photo that Jagannath Mahapatra had kept.

100. K. C. Kar, S. K. Das, and K. S. Sahu, Ramayana in Ordisi Pata Painting , Jagannath Mahapatra's role is not mentioned in this book. In 1989 he claimed that his paintings had been directly published, although I am inclined to believe his account from 1982 that they were copied.

101. I paid Rs 2,000 for the entire set of seventy-five. Jagannath Mahapatra said that Zealey was to have paid Rs 2,000 for her set (the deposit was 200) and that Subas De paid the same amount on behalf of the Handicrafts Board. I am aware that these figures may have been cited to justify the price my research assistant negotiated for me. The 1982-83 set was surely in part the work of Jagannath Mahapatra's assistants, whom I occasionally saw at work on it. Dinabandhu Maharana of Raghurajpur claimed that he had executed some of the initial scenes in 1981. Less than half of the set was completed in early 1983, when the approaching marriage of Jagannath Mahapatra's daughter may have provided an incentive for its prompt completion (which meant final payment); the master himself was thus involved more toward the end. The style is remarkably uniform throughout, demonstrating the "quality control" of workshop production.

102. This 1983 set was based more directly upon the 1954 Zealey sketches. Only one photo of the now missing 1965 scene of this particular event (Figure 220) was in Jagannath Mahapatra's hands.

103. J. P. Das, Puri Paintings , 73; B. C. Mohanty, Patachitras of Orissa , plate 13.

104. My documentation of the last, which was being painted during my visit, is largely verbal (backed up by bad photographs).

105. Here I have in mind particularly Raghunath Prusti, three of whose works have similar yellow covers with delicate floral designs. (J. P. Das and J. Williams, Palm-Leaf Miniatures , plate B). Even here it is possible that the similarity results from continued collaboration between the illustrator and a single professional painter.

106. The cover (Figure 257) and a second cover that belongs to a paper manuscript are in the Orissa State Museum, Bhubaneswar. The palm-leaf Lavanyavati^[*] in the British Library, Or. 13720, in which, paradoxically, the text sequence of the Rama story is not illustrated, bears on its covers simple scenes of the bending of Siva's bow, the release of Ahalya, Sita garlanding Rama, and the coronation.

107. Bhagavata Maharana was the elder brother of Banamali Maharana of Raghurajpur. This box was under consideration by the purchase committee of the National Museum in New Delhi in 1982 as an older object. When I showed the photos in Raghurajpur, there was a consensus that it was Bhagavata Maharana's work, and he concurred himself that he had made it about ten years before.

108. J. P. Das, Puri Paintings , 50-51, 163; J. Williams, "The Embassy." Both these discussions treat the paintings of Nayagarh, Ranpur, and Itamati as within the mainstream of Orissan style, in part because of their refinement, a viewpoint I would now question. It would seem that more objects were made for courtly patrons in Nayagarh than elsewhere. This is the home area of Jadumani, the chitrakara poet and wit who served the rajas of Nayagarh and Ranpur.

109. B. C. Mohanty, Patachitras of Orissa , 7. I know of only one example of a pata^[*] proper from Sonepur; the painter who owned it said it was roughly forty years old, an unpublished wedding scene in a private collection, found tied around a post in the artist's house. Unlike Puri patas^[*] , it has no painted background. The figure style is a large version of the playing cards discussed below.

110. B. C. Mohanty, Patachitras of Orissa , 11; R. von Leyden, Ganjifa: The Playing Cards of India , 108.

111. S. Mahapatra and D. Pattanayak, Amarusatakam , verse 32. This detail might suggest that this elusive manuscript was made in western Orissa.

112. E. Fischer, S. Mahapatra, and D. Pathy, Orissa , figure 594. The cycle was in the Radhakantha Matha^[*] , which was whitewashed in 1981.

113. Victoria and Albert IS 46-1963, 9.5 cm. diameter. R. von Leyden, Ganjifa , 102-3, color plate 9. The Victoria and Albert Museum records that this set of cards was donated in 1918 by Rev. H. W. Pike, a retired Baptist missionary. The cards show no signs of wear, and I see no reason to assign a nineteenth-century date to the set.

114. The difference between Pun and Parlakhemundi painting is visible if one compares photographs of the process of work on playing cards (often relatively coarsely painted) in Raghurajpur with similar photographs of the process of making a Parlakhemundi wall painting (E. Fischer, S. Mahapatra, and D. Pathy, Orissa , figures 516-27 and 620-53). In the playing cards the background is painted first, opaque areas of color are filled in, and delicate outline is added only as a final stage. In Parlakhemundi wall paintings, some initial work is linear and sketchy, no background is provided, and more time is spent on lively outlines.

115. Here as in Parlakhemundi I interviewed several chitrakaras about their genealogies, the nature of their work, and iconographic details. Gopinatha Mahapatra was one of the most informative in Jeypore.

116. J. P. Das, Puri Paintings , plate 48.

117. E. Fisher, S. Mahapatra, and D. Pathy, Orissa , figures 585-93. My information about dates comes from Apanna Mahapatra, mentioned below, as well as a second local chitrakara .

118. M. Mansinha, History of Oriya Literature , 130-32.

1. K. C. Panigrahi, Archaeological Remains at Bhubaneswar , 28; possibly latter half of the sixth century; V. (Dehejia, Early Stone Temples of Orissa , 183:650-80; T. E. Donaldson, Hindu Temple Art of Orissa , 1:31: late sixth century The scenes in question are reproduced in Donaldson (figure 29) and in D. Desai, "Narration of the Ramayana Episode—Vali-Vadha—in Indian Sculpture," plate 40.

2. Inconsistency of direction is visible on the Great Stupa at Sanchi, where most lintels read from left to right, but a few in reverse (e.g. the Vessantara Jataka on the North Gate). In some architectural friezes, the direction of circumambulation may explain the reading of reliefs from right to left, but this does not explain the inconsistency at Sanchi. In tiered panels, boustrophedon organization (e.g. some Sarnath Gupta steles, or the epic friezes of the Kailasanatha at Ellora) may respond to the efficiency of eye movement back and forth.

3. K. C. Panigrahi, Archaeological Remains at Bhubaneswar , 69: a cognate member of the Parasuramesvara group; S. C. De, "Svarna Jaleswar, One of the Early Temples of Bhubaneswar": same period as Satrughnesvara, late sixth century; V. Dehejia, Early Stone Temples of Orissa , 88: formative group, seventh century; T. E. Donaldson, Hindu Temple Art of Orissa , 1:43-50, figures 12-56: first decade of the seventh century.

4. T. N. Ramachandran, "The Kiratarjuniyam^[*] or Arjuna's Penance in Indian Art," 22-26; M.-A. Lutzker, "The Celebration of Arjuna—the Kiratarjuniya and the Arjunawiwaha in South and Southeast Asian Art," 38-41. I concur with Ramachandran in reading this section from left to right.

5. The case for apasavya , or counterclockwise, circumambulation as part of the Pasupata ritual has been made by C. D. Collins, The Iconography and Ritual of Siva at Elephanta , 136. Devangana Desai has suggested an occasional form of worship involving counterclockwise movement up to the water chute on the north side of the shrine, followed by return to the east, clockwise movement, and final counterclockwise movement: "Placement and Significance of Erotic Sculptures at Khajuraho," 148. The early Chola temple at Puñjai combines clockwise sequence (north and south walls) and counterclockwise sequence (west and part of south walls) in its reliefs: D. T. Sanford, Early Temples Bearing Ramayana^[*] Relief Cycles in the Chola Area: A Comparative Study , 141.

6. Gérard Genette connects prolepsis, relatively rare in Western narrative, with predestination ( Narrative Discourse , 67). In the present case, one might expect Rama's victory over Ravana^[*] , rather than one of his dubious deeds, to figure first by philosophical or religious logic.

7. Flashbacks occur, for instance in the Uttara Kanda^[*] of Valmiki or in the Uttararamacarita of Bhavabhuti (the retelling of the entire story seen painted on the wall), but not, as far as I know, in reverse sequence or inverting the order of books 3 and 4. Reversed-sequence narration does occur in Indonesian shadow-puppet theater and might be found in some performance or oral tradition in India also.

8. Kramrisch, Unknown India , plate 33.

9. This section is clearly visible in T. E. Donaldson, Hindu Temple Art of Orissa , vol. 3, figure 4156. His excellent plate shows the Svarnajalesvara^[*] as it existed before 1979, and my older photos confirm that placement. The present restoration has moved this section from the corner to the central facet, or raha portion of the wall (where no frieze existed originally), and one stone with three monkeys has been shifted to the left end of the flanking anuraha , visible at left in Figure 274 (bottom).

10. D. Desai, "Narration of the Ramayana Episode—Vali-Vadha—in Indian Sculpture," 82.

11. T. E. Donaldson, Hindu Temple Art of Orissa , vol. 3, figures 4160-62.

12. V. Dehejia, Early Stone Temples of Orissa , 116-20, 185: formative phase B, 680 to 750 (which seems very early); T. E. Donaldson, Hindu Temple Art of Orissa , 1:166, figures 369-400: late ninth century

13. Donaldson identifies one scene from the east end of the south side of the porch as Sita, Rama, and Laksmana^[*] , presumably because two male figures hold bows ( Hindu Temple Art of Orissa , 3:1416, figure 4172). I find it difficult to identify the remainder of this continuous frieze (e.g. three Ganesas^[*] side by side) with the Ramayana^[*] . Some figures with tails may indeed represent monkeys. Perhaps one large figure attacked by monkeys represents Kumbhakarna^[*] , in which case another large reclining figure to the left may represent the giant's awakening (Donaldson, figure 4171). Ramayana^[*] or not, this seems to illustrate an unfamiliar story.

14. D. Mitra, "Four Little-Known Khakara Temples of Orissa": between the Parasuramesvara and the Muktesvara; V. Dehejia, Early Stone Temples of Orissa , 125-28: advanced transitional phase, up to 850; T. E. Donaldson, Hindu Temple Art of Orissa , 1:274-81, figures 693-713: first quarter of the tenth century

15. A. Boner and S. Rath Sarma, Silpa Prakasa , xxiv. Donaldson notes the role of erotic imagery elsewhere on the temple ( Hindu Temple Art of Orissa , 1:280-81).

16. Donaldson identifies Rama and Laksmana^[*] seated on the ground, although these slightly corpulent figures without bows at the left of my Figure 281 (bottom) may represent Ravana^[*] and Maricha ( Hindu Temple Art of Orissa , 3:1416, figure 4177). To the right Laksmana^[*] may appear twice, seated, conversing with Sita.

17. In a careful examination with binoculars on the spot, I saw no human aspect in these trees, mentioned by D. Desai ("Narration of the Ramayana Episode—Vali-Vadha," 83); cf. Donaldson ( Hindu Temple Art of Orissa , 3:1417, figure 4180.

18. Donaldson suggests that repeated scenes of a monkey conversing with a human in front of a small building with a flag ( Hindu Temple Art of Orissa , 3:1417, figure 4181) represent the monkeys reporting back to Rama the results of their fruitless search. It seems equally possible that these are Hanumana's exploits in Lanka^[*] .

19. Cf. the discussion of the Sonepur Lanka^[*] Podi^[*] and the Asureshvar Danda^[*] Jatra in Chapter 1. Among later illustrations, Balabhadra Pathy's Lavanyavati^[*] , in particular, picks up this thread. Not all images of monkeys invoke the Ramayana^[*] —witness Pañchatantra illustrations on the Muktesvara Temple (K. C. Panigrahi, Archaeological Remains at Bhubaneswar , figure 73).

20. There are detached panels representing Rama, Laksmana^[*] , and Hanumana from Suklesvara. Friezes on the Sisiresvara at Bhubaneswar and the Pañcha Pandava^[*] at Ganesvarpur (two archers with monkeys) may represent the Ramayana^[*] , although no full identification of particular scenes has been proposed (T. E. Donaldson, Hindu Temple Art of Orissa , 3:1169, figures 4175-76).

21. D. Mitra, Konarak , 67. It is difficult to identify this as Rama's marriage per se, for of course monkeys were not yet associated with him at that point in his life.

22. Panigrahi identifies one figure as Vibhisana^[*] , but to me it appears to be the crowned Sugriva ( Archaeological Remains at Bhubaneswar , 86, figure 137).

23. These elements occur in a tenth-century image from Benusagar whose: details are clearer (T. E. Donaldson, Hindu Temple Art of Orissa , vol. 3, figure 3784). It is tempting to suggest an identity for these figures appropriate to the Gandhamadana incident, perhaps the demon Kalanemi and the woman-crocodile who attempt to destroy Hanumana in the Adhyatma Ramayana^[*] and many Oriya versions. Yet the generalized use of this iconic type makes it equally possible that these are simply generic opponents. Outside Orissa separate male and female figures may appear under each foot: K. C. Aryan and S. Aryan, Hanuman in Art and Mythology , plates 2, 52, 56, 57, 59, 62, 63.

24. T. E. Donaldson, Hindu Temple Art of Orissa , 2:726, and 3:1371, figure 378l. The absence of Sita rules out identification as the coronation.

25. This relief is roughly a meter square. Similarly refined wood carvings from. Nayagarh have been published in E. Fischer, S. Mahapatra, and D. Pathy, Orissa , 205-6; see also J. P. Das, Puri Paintings , plate 12. Those carvings were preserved by a descendant of chitrakaras , who said they were made by his ancestors about 150 years ago.

26. A scene of Ravana^[*] in his chariot from Jajpur appears to me to have been made at least as late as the seventeenth century (T. E. Donaldson, Hindu Temple Art of Orissa , vol. 3, figure 3783). Rama's coronation occupies the lintel of a wooden door, probably nineteenth century, that recently appeared on the New York art market ( Sotheby's Indian and Southeast Asian Art , New York, June 2., 1992, 177).

1. R. P. Goldman, ed., The Ramayana^[*] of Valmiki , 1.75.139-57.

2. B. L. Baij Nath, The Adhyatma Ramayana , 9.

3. Sarala Dasa's Mahabharata makes Risyasringa^[*] the son of a demoness and makes the courtesan Jarata the mother of his child. In Balarama^[*] Dasa's Jagamohana Ramayana^[*] , he is the son of a celestial apsaras and Iris wife, Santa, the daughter of Dasaratha, as in many Oriya versions (apparently based on some recensions of Valmiki). Here the seduction and the boat journey are described at length. (K. C. Sahoo, "Oriya Rama Literature," 213-15).

In the Ramalila^[*] performances I know, Risyasringa^[*] appears as a sage, his jata^[*] mukuta^[*] suggesting a single horn; but his dalliance with the courtesans is not prolonged, perhaps because this is a prelude to the actual birth, which must take place on Ramanavami itself, the first night of the entire cycle. For this event I would not suggest that performances outweighed the ornate literary tradition in guiding many artists.

4. In this relationship, most Oriya texts follow the Bengali recension of Valmiki as opposed to other recensions, where Santa is the daughter of Rompada (or Lompada).

5. F. A. Marglin, Wives of the God-King . Marglin presents the Risyasringa^[*] story as explaining the nexus between auspiciousness and political status (pp. 100-101).

6. In Kanungo's version, published in 1977, the boat does not have a swan prow and the figures are asymmetrically arranged (K. C. Kar, S. K. Das, and K. S. Sahu, Ramayana in Ordisi Pata Painting , figure 3). This event follows Sita's birth in the published form, as it does in the Vaidehisa Vilasa^[*] , whereas in Jagannath Mahapatra's sets it occurs first.

7. The boat of the vesyas is more fully described in the Sanskrit Mahabharata and in most Oriya texts than in Valmiki's Ramayana^[*] .

8. The 1902 manuscript devotes 5 1/2 folios (13a through 18a) to this sequence. From the 1914 version, folios 15 (showing the vesyas' arrival) and 17 are preserved in the Rietberg Museum, Zurich. Folio 20v (in Delhi) depicts Dasaratha's sacrifice, which suggests a total of six pages for the whole episode. The 1926 version devotes 10 1/2 pages to the same events.

9. R. P. Goldman, The Ramayana^[*] of Valmiki , 1:75.

10. Balarama^[*] Dasa, Jagamohana Ramayana^[*] , 1:1.

11. K. C. Sahoo, "Oriya Rama Literature," 211.

12. B. J. Baij Nath, Adhyatma Ramayana , 14; Upendra Bhañja, Vaidehisa Vilasa^[*] , 6, 43. In Oriya the name generally appears as Tadaki^[*] , although Tataka^[*] or Tadaka^[*] (often transcribed Taraka) appear in Sanskrit; her genealogy and previous history vary in each account. One interesting addition by Balarama^[*] Dasa to Valmiki (also found in the Vaidehisa Vilasa^[*] , in Krittibasa's Bengali, and in folk tales of the Birhors) is that Visvamitra urged Rama to take a road around the grove of the demoness, a suggestion the hero rejected (W. L. Smith, Ramayana^[*] Traditions in Eastern India , 174). Michha Patajoshi faithfully illustrates this interesting variation, which serves to emphasize Rama's divine mission, in his versions of the Vaidehisa Vilasa^[*] .

13. In Valmiki twenty-one chapters intervene between these two events. In the Adhyatma Ramayana^[*] , the Vaidehisa Vilasa^[*] , and the Lavanyavati^[*] the defeat of Subahu and Maricha is briefly mentioned as a lead-in to the Ahalya episode ( Vaidehisa Vilasa^[*] , canto [ chhanda ] 7). In Valmiki's version, Ahalya is rendered invisible rather than turned into stone (1.47.29).

14. The episode of the boatman, absent in Valmiki, is widely known from the Ramcharitmanas , where, however, it occurs after Rama's exile in the next volume of the story The emphasis upon bhakti appears in many late versions.

15. E.g. Deogarh, sixth century (M. S. Vats, The Gupta Temple at Deogarh , plate XVa— Ahalya); Virupaksa Temple, Pattadakal (H. Cousins, Chalukyan Architecture of the Kanarese Districts , plate XLIV, upper left—Tadaki^[*] and Ahalya); Nageshvara Temple, Kumbhakonam and other early Cola reliefs (D. T. Sanford, Early Temples , 79).

16. The branching lines on the limbs may derive from striped clothing commonly worn by demons (cf. Figure 184, top). The spotted face and sari vaguely resemble Persian conventions for divs , by this time widely adopted for raksasas^[*] in Rajput painting.

17. The Brahma Ramayana^[*] itself simply says, "He released the fallen Ahalya from the curse of Gautama" (verse 64). Arjuna Dasa's Rama Vibha the first Oriya kavya version of the Ramayana^[*] , probably sixteenth-century and not widely popular later, describes Ahalya as turned into a square stone (K. C. Sahoo, "Oriya Rama Literature," 252).

18. The text on this folio (the reverse of the leaf) comprises verses 10 to 16 of the seventh chhanda , which describes the charms of the forest. Presumably the text of the Ahalya episode followed on the next leaf, not preserved in Zurich. R. Skelton, Indian Miniatures , 24, identifies our Figure 9 as folio 3r.

19. I am deeply indebted to Mrs. Suresvari Mishra (Reader in Sanskrit, Puri Women's College) for deciphering this verse after I and other Oriya friends had given up on it.

20. The 1902 manuscript in Koba devotes about four folios to this sequence. The 1914 version in New Delhi covers it in seven pages (including scenes of other demons killed after Tadaki^[*] ), whereas the 1926 version in Calcutta devotes nine pages to it. The general selection of scenes is similar; for instance a full page of unillustrated text precedes the episode of the boatman in all. The disguise of Indra as a cat, found in the Padma Purana^[*] and in Punjabi folktales as well as in Upendra Bhañja's work, may derive from the Sanskritic term for cat, marjara , which may also imply "paramour" (K. C. Sahoo, "Oriya Rama Literature," 255).

21. Some versions of Krittibasa^[*] include Bharata and Satrughna in Visvamitra's expedition, but they are eliminated in the incident of alternative routes mentioned above in note 12 (W. L. Smith, Ramayana^[*] Traditions in Eastern India , 174). The label in Figure 177 spells his name Bharatha, as does that in Figure 183, right, the standard visit of Bharata to Rama.

22. According to A. K. Ramanujan, there is an implicit relationship between Ahalya and Tadaki^[*] in Kampan's^[*] Tamil version of the epic ("Three Hundred Ramayanas^[*] ," 31).

23. S. I. Pollock, The Ramayana^[*] of Valmiki , 2:64-73.

24. S. I. Pollock, The Ramayana^[*] of Valmiki , 2:487-89; Valmiki, Ramayana^[*] , ed. G. H. Bhatt, 2, Appendix 1, 26. In Balarama^[*] Dasa, Sita, marked with ocher, clasps Rama, leaving a red mark, which makes them both laugh. All the crows are blinded, but Sita persuades Rama to restore their sight on the condition that they remain squint-eyed (K. C. Sahoo, "Oriya Rama Literature," 322-23).

25. Cantos 18 and 19. The poet introduces the latter as "a garland of jamakas ," a generally untranslatable trope in which one word is repeated with different meaning. In Valmiki and the Adhyatma Ramayana^[*] , Bharata's visit is followed more or less directly by the exiles' visit to yet another illustrious sage, Atri, and his pure wife, Anasuya.

26. The wall forms a continuous composition behind the niche, broadly similar to Prusti's design in Figure 183. Conceivably the central niche was a later addition, although this speculation is not borne out by details of painting or construction.

27. There are four in Orissa State Museum Ext. 81. In the remainder only three are preserved, but pages are missing in each, and in general Sarathi Madala Patnaik includes a picture on each page. The scenes are not always the same.

28. The Vaidehisa Vilasa^[*] reads simply, "He worshipped the kings shoes." In the Baripada manuscript I am unable to locate the sequence of Bharata's visit.

29. Victoria and Albert IS 24-1967 seems also to illustrate the shooting of the crow.

30. Michha Patajoshi's 1902 Vaidehisa Vilasa^[*] also depicts Rama and Laksmana's^[*] transformation, although there the brothers stand while tying their hair (Figure 106). This scene, as well as the relative elaboration of architecture in this early work, leads me to wonder whether as a youth Michha was not aware of Prusti's work.

31. He concludes that he will gain both moksa^[*] and the ease of dying earlier than Ravana^[*] : "As when heat dries up a pond, first the fish and then the crocodile die."( Vaidehisa Vilasa^[*] , 24.15).

32. Kathleen Erndl describes Surpanakha's^[*] role in general: in the Adhyatma Ramayana^[*] as the object of a prank by the brothers, elsewhere as Sita's dark alter ego ("The Mutilation of Surpanakha^[*] ").

33. Jatayu^[*] swallowing Ravana's^[*] chariot is a regular scene in Bengali paintings (J. Williams, "Jatayu the Valiant Vulture").

34. K. C. Sahoo, "Oriya Rama Literature," 383-87, suggesting an origin in folklore.

35. K. C. Sahoo, "Oriya Rama Literature," 377-79, suggesting parallels in the folklore of the Kols. Cf. L. Hess, "The Poet, the People," 247.

36. In this section, I must confess particular uncertainty about my identification of the scenes labeled (in Figure 200) Surpanakha^[*] denosed by Laksmana^[*] , Surpanakha^[*] before Ravana^[*] , Laksmana^[*] and Sita, and the blue figure and woman at the bottom of wall B.

37. E.g. J. P. Das, Puri Paintings , color plate 23 (on the mid-right edge). For recent independent versions of the scene, see J, Williams, "Criticizing and Evaluating," figures 14-22 (plates 15 and 16 there are reversed, although the captions are in the correct order).

38. See Appendix 1. The creation. of Maya Sita might well occur on missing leaves in Sarathi Madala's Bharany and Utkal University manuscripts. It is definitely absent, however, in the two Adhyatma Ramayana^[*] manuscripts not by him—National Museum 75.536 and the second copy in the Bharany collection.

39. The figures in Figure 59 are labeled from the left, Sita, Ramachandra, Ramachandra (with no explanation for his change of costume), and Maya Sita.

40. See Appendix 1. Fruit appears in Orissa State Museum Ext. 81 and National Museum 75.536, although it is impossible to tell if it has tooth marks (not mentioned in the Adhyatma Ramayana^[*] itself). The less frequent images of Kabandha always show him with extremely long arms that stretch horizontally across the scene, suggesting some shared pictorial tradition not found at Buguda or known to the illustrators of other texts.

41. J. P. Das, Puri Paintings , 135. This is the only written explanation for Navagunjara^[*] I know, although similar images occur outside Orissa.

42. Satrughna depicted the Raslila^[*] similarly in his illustrated Braja Bihara .

43. In even the version of Prusti, often indebted to Buguda, we see a four-wheeled ratha (Figure 187). The form of Puspaka^[*] Vimana in the reliefs of Loro Djonggrang in Java a millennium earlier shows a similar head below, not necessarily indicating any visual continuity (W. Stutterheim, Rama-Legenden , vol. 2, plate 35).

44. The form standard on pats , with a large-headed Jatayu^[*] holding the chariot in his open beak, also occurs in the 1775 Ramcharitmanas made in Midnapore District (N. Goswami, Catalogue of Paintings of the Asutosh Museum Ms. of the Ramacaritamanasa , plate XIVb; J. Williams, ''Jatayu the Valiant Vulture").

45. The Sabari's costume, blouseless, with the sari over her right shoulder, is reserved in this manuscript for servants and may indicate humble, if not necessarily tribal, status.

46. Cf. his empty rooms in illustrating the Sangita Damodara (K. Patnaik, Raga-citra , 38, 42).

47. Ravana^[*] observes both these scenes in the form of an ascetic, as he does in Prusti's version (Figure 186, far left). While this is a conceivable twist, I know of no text that mentions his presence.

48. J. Williams, "From the Fifth to the Twentieth Century and Back," figure 7. It is also interesting that the "centaur" pattern, with the torso of Maricha emerging from the shoulders of the deer, does not, as far as I know, occur in later Orissan images, although it is common in Bengal in terra-cottas on temples, in the 1775 Ramcharitmanas made in Midnapore (N. Goswami, Catalogue , plate XIVa), and in the pats of itinerant entertainers (S. Kramrisch, Unknown India , plate XXXIII).

49. N. C. Mehta and Moti Chandra, The Golden Flute , plate 1 (a page from a Balagopalastuti , fifteenth century); J. M. Nanavati, M. P. Vora, and M. A. Dhaky, The Embroideries and Beadwork of Kutch and Saurashtra , 93, plates 36, 104, middle (traditional embroidery and beadwork); E. Fischer, J. Jain, H. Shah, Tempeltücher für die Muttergottinnen in Indien , figure 250, no. 29. In Rajasthan the motif occurs in late, rough wall paintings (Orchha, Shekhavati) and in the padhs^[*] used in village performances, but I have not discovered it in courtly miniatures. The two-headed deer is also (more rarely) associated with a goddess, in Gujarat (E. Fischer, J. Jain, and H. Shah, Tempeltücher , 233) and in the terra-cottas made at Molela in western Rajasthan.

50. I would interpret the motif here as a form of doodling, for this artist gratuitously inserted other small animals in the same way. Several three-headed deer appear as decoration (along with one- and two-headed forms) in the wall paintings of Shekhavati. The discovery of this unexpected detail in the C. L. Bharany manuscript and my reaction to it call to mind Ogden Nash's immortal poem "The Lama" ( I Wouldn't Have Missed It: Selected Poems of Ogden Nash [Boston: Little, Brown, 1975], 16).

51. Vana Parva, verses 336-42 in the edition of the Orissa Government Department of Culture (Bhubaneswar, 1974).

52. Sri Ramalila^[*] , 63, in which the deer is simply described as covered with gold and two-headed.

53. In the example from Dasapalla, one head is removed when Prima returns with the corpse, "making it easier to carry." This unique case in Orissa could be reconciled with the Gujarati explanation discussed below, that the second head suggests the action of the live deer only Bisipada and Amarapura in Ganjam also retain a two-headed wooden deer with both heads fixed in place (J. Williams, "From the Fifth to the Twentieth Century and Back," figure 9). In Gania and Belpada today the deer is played by a human actor without mask. In the Jeypore area actors wear a mask in the form of a hat with two tiny deer heads on top, like antennae. One of these on display in the Indian Museum, Calcutta, is illustrated in Indian Archaeology 1978-79 — a Review , plate LIV, c (mistakenly labeled "Chhau mask'').

54. J. M. Nanavati, M. P Vora, and M. A. Dhaky, Embroideries and Beadwork , 93. Gujarati block-print textile makers of the kind discussed in E. Fischer, J. Jain, and H. Shah ( Tempeltücher ) have given me a similar explanation, conflating the scene of Rama with another story that involves seven sisters and a deer that looks back. The padh^[*] painters of Bhilwara and Shahapura in Rajasthan, however, explain the two heads as showing the deer's extraordinary maya nature.

55. Tala^[*] (palm) trees are mentioned in the Bengali and northwest recensions of the Valmiki and are depicted in sculpture in preference to the deciduous sala^[*] (Valmiki, Ramayana^[*] , ed. G. H. Bhatt, vol. 4, chap. 11, verse 47). Only one Oriya version, the eighteenth-century Rama Krisna^[*] Kelikallol of Tripurari Das, refers to talas^[*] (K. C. Sahoo, "Oriya Rama Literature," 406). Because salas^[*] appear in all the texts and in most images considered here, 1 use sal.

56. Balarama^[*] Dasa, Jagamohana Ramayana^[*] , Kiskindha^[*] Kanda^[*] , 70.

57. K. Bulke, Ramkatha , 3rd ed., 471. In the Ananda Ramayana^[*] the snake has an adversarial relationship to Valin and is uncoiled by Laksmana^[*] .

58. D. Desai, "Narration of the Ramayana Episode," 84, figures 47-49. The image of a serpent appears as early as Loro Djonggrang (tenth century) in Java, and the serpent is mentioned in some Southeast Asian texts such as the Seri Ramayana .

59. Upendra composed a collection of such verses, Chitra-kavya Bandhodaya . (J. P. Das, Chitra-pothi , 81-84). These remain popular today, and I have heard stories that the last generation of Oriyas used bandhas in writing affectionate private letters. It is to be underscored that the bandha appears even in unillustrated manuscripts of the Vaidehisa Vilasa^[*] as well as in modern printed texts. Chitra-kavya did, admittedly, exist in other vernaculars, from Tamil to Gujarati, as well as in Sanskrit: K. Jha, Figurative Poetry in Sanskrit Literature .

60. The Bengali and northwestern recensions of Valmiki provide sources for Tara's curse (K. C. Sahoo, "Oriya Rama Literature," 414). Oriya Ramalilas^[*] maximize Tara's role as advisor and underscore her curse with emphatic music.

61. That the composition resembles Hoysala reliefs rather than earlier Orissan ones need not be used to infer specific connection with thirteenth-century Karnataka.

62. Sarathi Madala Patnaik's 1875 manuscript (Orissa State Museum Ext. 81), does not illustrate this event at all. The 1892 version (Utkal University Library) is missing a folio here.

63. Rama specifically says, "Monkeys are usually considered ugly, but this one surpasses the nymph Rambha in beauty" ( Vaidehisa Vilasa^[*] 28.149-50). The jumbled condition of the Baripada manuscript makes it impossible to be sure whether Satrughna illustrated her curse or not.

64. The cock on the left is mentioned in the Lavanyavati^[*] after the rainy season and before Rama gets news of his wife (see Appendix 3, no. 27). Two monkeys in Figure 189 are identified by labels as minor followers of Sugriva.

65. The initial meeting with Sugriva, which occupies a full folio (Figure 150) includes the unexplained figure of Jambavan (labeled Rsimukha^[*] ) to the right of Laksmana^[*] . Could this also be a reinterpretation of the scene at Buguda of Laksmana^[*] straightening the arrow (Plate 11), as I have suggested in the text for Prusti's version of Mount Malyavan in Figure 189?

66. For another scene of rain in Pathy's work, see J. Williams, "Jewels from Jalantara," plate 7.

67. In this leap, Hanumana encounters the mountain Mainaka and the demonesses Surasa and Singhika (Valmiki 5.1).

68. Vaidehisa Vilasa^[*] canto 36.

69. One might expect to find this portion among the badly damaged paintings, but I see no likely candidates. On the front wall (L in Figure 200), local tradition identifies one scene of figures beside a hut as the visit to Bharadvaja's hermitage, which occurs soon after the exile in book 2 of the epic. The occurrence of this portion in the final position is puzzling, but virtually all possible identifications pose that difficulty I find no images of Ravana^[*] , a sine qua non if this wall depicted book 5.

70. The Orissa State Museum copy, Ext. 81, included fifteen pictures in the Sundara Kanda^[*] . Both the New York Public Library and the Utkal University copies included six pictures, and missing folios may account for more in each case. An untraceable copy, some illustrations from which were reproduced in Bansidhar Mohanty's History of Oriya Literature (1977), included at least five pictures in this section. The Adhyatma Ramayana^[*] illustrated by another artist (National Museum 75.536) as well as the illustrated Ramalila^[*] of Krishna Chandra Rajendra (National Museum 75.556) both lack their later portions and hence cannot be discussed from this point on.

71. I have been unable to locate this sequence in the Vaidehisa Vilasa^[*] by Satrughna from Baripada.

72. One would, however, expect that Pathy copied the entire text leaving blank spaces for pictures before he went back to execute the illustrations, a procedure documented by the colophon of his Gita Govinda now in the L.D. Institute in Ahmedabad (J. Williams, "Jewels from Jalantara").

73. K.K. Handiqui, Pravarasena's Setubandha; R. Basak, Pravarasena's Ravanavahamahakavyam^[*] . This event does not figure vividly in Oriya Ramalilas^[*] , where today it is more likely to be chanted alone than to be enacted.

74. B. L. Baij Nath, Adhyatma Ramayana , Yuddha Kanda^[*] chap. 4, line 1.

75. Adhyatma Ramayana^[*] , trans. Gopala, 173.

76. K. Bulke, Ramkatha , 546-47. The existence of this story in Tamil and Bengali suggests its ubiquity Islam preserves a similar tale that cats with stripes on their heads have been blessed by Muhammad.

77. Vaidehisa Vilasa^[*] 40.34-38. Nala's curse is also present in Balarama^[*] Dasa.

78. This couplet is a pun and can also be translated, "Lovely herds of horses came and went; shady women lost their fear [i.e. were protected by the noise?]."

79. This is not true of the Dispersed Lavanyavati^[*] (Figure 153), where, however, one other folio might read from right to left. (Figure 149). In both of Sarathi Madala Patnaik's 1891 manuscripts, two monkeys are moving to the left, although the placement of Rama and Laksmana^[*] as well as the Siva shrine would suggest movement to the right (Figures 64, 71).

80. He creates an illusionary head of the hero in at least some versions of the text (K. Bulke, Ramkatha , 552), a case in which the theme of maya (perhaps moha ) goes back to the Sanskrit.

81. Adhyatma Ramayana^[*] 6.5.44. In the southern recension of Valmiki, Rama assists Sugriva by shooting Ravana's^[*] single umbrella (K. Bulke, Ramkatha , 554).

82. K. C. Sahoo, "Oriya Rama Literature," 487-88 ( Jagamohana Ramayana^[*] 6.57-60).

83. K. C. Sahoo, "Oriya Rama Literature," 490. This variant appears to be peculiar to Orissa, although the same term is used for mushroom and umbrella in other languages also ( chhattra or a close variant).

84. Vaidehisa Vilasa^[*] 41.23.

85. J. E Losty, Krishna , plate 83. For the Kanchi-Kaveri story, see J. P Das, Puri Paintings , 120.

86. The label says "Indrajita bound Hanu with the naga-pasa [hence the snake above] and brought him to Ravana^[*] ." One might read the coils beneath as part of the snake. I am not aware of such a textual variant at this point in the story In the non-Orissan examples mentioned here, snakes are not involved. Possibly the Parlakhemundi card shows a conflation of the two incidents.

87. J. Mittal, Andhra Paintings of the Ramayana , plate 33; H. Cousins, Chalukyan Architecture of the Kanarese Districts , plate XLVI, upper left.

88. The use of the monkey's tail as a seat figures in the Ramalila^[*] of Dasapalla but is apparently not enacted with much dramatic effect.

89. This may be an interpolation, according to K. Bulke ( Ramkatha , 558).

90. Adhyatma Ramayana^[*] 6.5.65-74 and 6.6.1-6.7.37. It is clear that the initial directive to pick up the entire mountain is borrowed from the subsequent episode, in which this action is explained by Hanumana's inability to locate the herbs. Kumbhakarna's^[*] death follows both episodes in this version.

91. W. L. Smith, Ramayana^[*] Traditions in Eastern India , 182-83. The Kalanemi episode also occurs in the Bengali recension of Valmiki (G. Gorresio, Ramayana , 6.82, 91-98).

92. W. L. Smith, Ramayana^[*] Traditions in Eastern India , 145-51.

93. This form in which the Garuda^[*] is largely human, with wings, leathered arms, and a snake raising its hood above a crown, occurs often in patas^[*] .

94. Valmiki 6.74.

95. Conceivably the hero is Bharata shooting at Hanumana, but the demons do shoot at him, in which case this scene would seem to represent the death of Kumbhakarna^[*] .

96. One exception is the early Chola temple at Pullamangai; see D. Sanford, "Miniature Relief Sculptures at the Pullamangai Siva Temple."

97. In Vaisya Sadasiva's Ramalila^[*] , the text does include an Uttara Kanda^[*] .

98. See Chapter 2, note 87, for the evidence about their date. This is the only Ramayana^[*] subject there. The small figures in the upper left are presumably the seven sages, whose presence would be generally appropriate at such an auspicious event.

99. J. Williams, "Marriage Paintings in South and Central Orissa." The coronation is more likely to be replaced by the wedding of Rama in southern Ganjam District. In the Purl area, subjects such as the Navaguñjara and guardians visible in Figure 262 indicate that not all themes are precisely linked to marriage.

100. Jagannath Mahapatra's version of Sita's ordeal does not actually depict flames but rather piled wood and a rain of flowers (Figure 243). Possibly even in the 1950s in Orissa, where sati was not in practice, there was reluctance to encourage widow-burning, evoked by an image of a woman within a fire. In the recent television version of the Ramayana^[*] , this was an explicit concern.

101. In the published set, executed by Bibhuti Kanungo, Hanumana is the only monkey, and he has five hillocks wrapped in his long tail (K. C. Kar, S. K. Das, and K. S. Sahu, Ramayana in Ordisi Pata Painting , no. 71). In the verse accompanying the publication, Hanumana and Jambavan are mentioned (in Oriya and Hindi as well as English), which is not the case for the version Jagannath Mahapatra has supplied as his "text." In all versions, this verse centers on Lava's hitting Rama with an arrow, causing him to swoon.

102. In the Bibhuti Kanungo publication, this figure is more clearly in a dancing pose (K. C. Kar, S. K. Das, and K, S. Sahu, Ramayana in Ordisi Pata Painting , no. 75). In his version of the verses apsaras are not mentioned, although they are in Jagannath Mahapatra's.

103. In Figures 54 (1875), 68 (December 1891), and 86 (1899) the extra sets radiate from the main arms, whereas in Figures 58 and 65 (January 1891) and 74 (1902) they are attached to the torso. The reader is reminded of the evidence in Chapter 2 that these are indeed one artist's work.

104. Since what survives is folios 189 through 200, we may presume that the earlier sections were as amply illustrated as in Sarathi Madala's copies. In both Figures 80 and 81 details of the text, such as Laksmana's^[*] transformation into the serpent Sesa^[*] are omitted. It would be interesting to know how this book has been separated from the rest of the manuscript and whether the selection reflects some questioning of its contents.

105. The 1902 manuscript in Koba devotes one leaf to the coronation per se and a separate one to festivities that are shown below the throne in Figure 123, including music, swordplay, and the purna^[*] ghata^[*] with fish.

106. It is possible that a previous page is missing, which might have included the death of Ravana^[*] . It is equally possible that the artist of the Dispersed Lavanyavati^[*] , often unconventional in his narrative choices, omitted that seemingly central event.

1. Vivarana^[*] and varnnana^[*] are the entries under "narration" in standard English-to-Oriya dictionaries. Both come from varna^[*] meaning fundamentally "covering" or "color." The same set of meanings exists in Sanskrit and other Indo-European north Indian languages. The term kathana , "telling,'' also exists (see the subsequent discussion of katha , "tale," in the text) but is less commonly used to name an abstract process.

2. G. Lukacs, "Narrate or Describe." I must confess my reluctance to see "description" and "narration" as mutually exclusive binary opposites, even in English. As Tzvetan Todorov put it, "Description alone is not enough to constitute a narrative; narrative for its part does not exclude description, however" ( Genres in Discourse , 28).

3. Katha also means "speech," which includes other forms in addition to narration of a story Nor can story and narrative be exactly equated in English. As Gerald Prince puts it, "Although any story is a narrative, not any narrative is a story" ( Narratology , 170 n. 8).

4. "Able to be carried by articulated language, spoken or written, fixed or moving images, gestures, and the ordered. mixture of all these substances; narrative is present in myth, legend, fable, tale, novella, epic, history, tragedy, drama, comedy, mime, painting (think of Carpaccio's Saint Ursula ), stained-glass windows, cinema, comics, news item, conversation" (R. Barthes, "Introduction to the Structural Analysis of Narratives," 251). There is some slippage between "narrative" and the French term " recit^[*] ,'' more easily equated with a genre.

5. P. Lutgendorf, "The View from the Ghats," and Life of a Text , 18-29.

6. Aristotle in the Poetics (chap. 6) defines six parts of tragedy (plot, characters, language, thought, spectacle, and melody) and devotes five chapters to plot, with only one chapter to the characters and even less than a chapter to the remaining parts. Thought would usually be considered a non-narrative element, and spectacle and melody may characterize the theater more than other narrative forms of literature.

7. A. B. Keith, The Sanskrit Drama , 277.

8. See Chapter 1, note 9.

9. Thus Upendra Bhanja is depicted at the beginning of Michha Patajoshi's Vaidehisa Vilasa^[*] , along with other poets who had created their versions of the same story. In the north Indian illustrations of the Aranyaka^[*] Parvan of the Mahabharata, King Janamejaya's sage, Vaisampayana appears frequently at the top of a larger picture (K. Khandalavala and Moti Chandra, Illustrated "Aranyaka^[*] Parvan "). Cf. G. Prince, Narratology , 15-16 (on multiple narrators); and M. Bal, Narratology , 143 (on the related form of embedded texts, common in the Near East as well as South Asia).

10. The Vishnudharmottara^[*] (Part III ), trans. S. Kramrisch, 59-60; A. K. Coomaraswamy, "Reactions to Art in India."

11. M. Bal's terminology for verbal narrative may be one way of formulating this distinction: "fabula," the actual series of events; "story," the fabula presented in a certain manner; and "text," a story told by a narrator ( Narratology , 5). Even for an entirely verbal situation, I find this model unnecessarily complex. We know the fabula and story only from the text, and they remain our own artifact. Why not stop there?

12. One notable exception is the image of Bharata worshiping Rama's footprints within a circle, and even here Bharata appears alone in the second work of 1891.

13. It is likely that he worked from an unillustrated manuscript, but I have frankly not checked the consistency of his texts to be certain of this.

14. Simple statistics of the number of pictures do not make this point, for pages are missing in some cases (almost the whole second half of the last two works). Moreover, the physical dimensions of the leaves may affect the ratio of text to image. Hence the need to refer to Appendix 1 to verify my assertions.

15. Cf. J. P. Das, Chitra-pothi , 37; J. E Das and J. Williams, Palm-Leaf Miniatures , 79.

16. The cartouches above Rama and Laksmana ^[*] do not contain their words or thoughts, nor does there seem to be a specific rationale in the placement of other verses in the scene.

17. J. P. Das, Chitra-pothi , 41, 42, 56.

18. The argument that the character of this manuscript corresponds to that of the Vaidehisa Vilasa^[*] itself is borne out by my general impression that Satrughna's illustrations of a rather different text, the Braja Bihara belonging to the History Department of Utkal University, create a greater sense of order. That work awaits careful study, however.

19. Because there is remarkable consistency in Michha Patajoshi's choices, although subordinate parts of the principal scenes are added, I have omitted a chart comparing his major manuscripts like the one in Appendix 1 for Sarathi Madala Patnaik.

20. I believe that this proportion of pictures to text holds even in the case of his 1902 copy, which had 218 folios originally, versus Satrughna's Baripada work of 281 folios, for Michha Patajoshi's was 41.6 cm long whereas Satrughna's was only 28.5. Some of Satrughna's multifolio pictures obviously devoted more space to a single image than did any of Michha's. The two artists' handwriting was similar in scale, a factor in the physical bulk of text. Of the 151 surviving folios in the 1911 pothi , only five sides comprise text alone, which is an unusually low proportion.

21. Some manuscripts, such as Satrughna's, lack such captions, and in others the captions are fewer and shorter. Most scribes, such as Sarathi Madala Patnaik and Raghunath Prusti, began each caption "This is . . . [e.g. Ramachandra]"; Michha Patajoshi, however, launched directly into a descriptive sentence.

22. Perhaps a battle may have occurred on a missing subsequent leaf, which might have included the death of Ravana^[*] and Vibhisana's^[*] coronation. It remains equally possible that these too were omitted, for without pagination this manuscript is hard to reconstruct. Likewise my assumption that the encounters with Surpanakha^[*] and the golden deer were present on a missing folio is based on my own causal preconceptions.

23. Another possible explanation is that the more auspicious episode occurs to the right. This is contradicted in Figure 153 by the placement of Ravana^[*] to the right of the building of the bridge.

24. I have toyed with the possibility that we should regard these figures as what Gerard Genette has termed "focalizers," the source of the vision presented, distinguished from the narrator ( Narrative Discourse , 189). In general Indian literature and art seem to me to show what Genette terms "zero localization," largely as a result of the conscious impersonality of the arts. In the present situation, I see no reason to regard Lavanyavati^[*] as more focal or critical than the various male narrators to the interpretive, subjective nature of the entire account. The artist of this manuscript was daring and sophisticated in his narrative choices, but he worked within a classical Indian tradition.

25. Hence a relatively small number of episodes is depicted in this work. This is not the case with the depiction of the Lavanyavati^[*] proper in this same manuscript.

26. The one exception is Risyasringa's^[*] marriage to Santa.

27. The first page is numbered 72 and the last preserved is 81. Surely one and probably two more pages followed the Gandhamadana incident. Here as in the Dispersed Lavanyavati^[*] , no text is interwoven with the pictures. Possibly it was concentrated at the beginning or the end, a suggestion I cannot verify Whether the text was present or omitted, in both manuscripts the pictures seem to have a kind of independence, with no immediately accompanying verses.

28. In the New Delhi copy of Michha Patajoshi this portion of folios is missing; Satrughna's Baripada manuscript seems to lack this scene, but its jumbled condition makes it difficult to be certain. Other scenes paired in Prusti's Lavanyavati^[*] that immediately resemble Buguda are the battles with Kumbhakarna^[*] and Indrajita (Figures 191, bottom; 192, top), which resemble the badly preserved murals at the top of wall H (Figure 212). Such dense barrages of arrows do, however, occur in other manuscripts and cannot be used as prima facie evidence for the direct impact of Buguda (cf. S. Pani, Illustrated Palmleaf Manuscripts of Orissa , 47).

29. Conceivably the Buguda artist, who was not entirely consistent in chronological sequence, intended this scene to follow the cutting of the umbrellas, in which case Buguda is equally hard to explain textually It is worth noting that Prusti adds more monkeys to the extreme right, but with no suggestion of Sugriva's palace.

One additional scene probably borrowed from Buguda but problematic in meaning is Rama's meeting with Visvamitra and a pupil in Figure 184. I know of no direct textual reference for this, although there are many general allusions to ascetics on Mount Chitrakuta^[*] that would make it possible. The comparable scene at Buguda includes only a tall blue ascetic and a youth, their hands raised in respect (Figure 209). Here the identification is made possible only by Prusti's neatly labeled version of the same pair.

30. Vaidehisa Vilasa^[*] 17.35-36. Guha is of course absent from the brief version of the Lavanyavati^[*] . Balabhadra Pathy shows Guha as a boatman by a river. Sarala Dasa generically identifies many forest dwellers as Sabaras, as does the Ramalila^[*] of Dasapalla.

31. New York Public Library, Spencer Collection ms. 11, folio 15. Michha Patajoshi also worked the motif into the final coronation of Rama (Figure 123).

32. The text reads: "The person who hears the Rasotsava^[*] of Raghav-Sita becomes a devotee of Hari and gets a place near him. The seventeenth chapter of the Brahma Ramayana^[*] is complete here. This chapter concerns the desires of mature ( purna^[*] ) girls in the Ramakrida^[*] . . . ." Conceivably the orb is in fact a full ( purna^[*] ) moon. The Oriyas I have consulted are equally puzzled by this image.

33. Anantasayana^[*] appears in a related form in the nineteenth-century wall paintings of Dharakot. E. Fischer, S. Mahapatra, and D. Pathy, Orissa: Kunst und Kultur , figure 556. For the coronation, see Chapter 4, "The Conclusion."

34. Admittedly, I cannot be certain that either Mrs. Zealey or the professor did not dictate the inclusion or omission of some particular event just because the artist does not recall that. At least we can be sure much was left up to him and that he has internalized the choices that were made.

35. J. P. Das, Puri Paintings , figure 3.

36. All were in the process of being painted, and my understanding of subjects and order is indebted to my conversations with the painters as they worked. The very large pata^[*] at Danda Sahi represented in Figure 289D was sold unexpectedly; hence the absence of a photo here.

37. The omission of Hanumana with Mount Gandhamadana (except possibly in the large Danda Sahi work, only partially painted) is puzzling. Perhaps this image is hardly thought of as narrative at all, having become the most common form of image of the monkey alone in small patas^[*] .

38. The four corner events in Figure 289B seem to read counterclockwise, although it is hard to be sure about them. In any case, they are not clearly in any sequence, serving to lead from the tight order of the central circle to the final battles at top and bottom.

39. As I suggest in Chapter 3, one small hill occurs at the top of wall A, in a position where the most likely subject is Rama and Sita on Mount Chitrakuta^[*] . Conceivably the artists at this particular point saw the suitability of larger scenes set on mountains, in which the setting could be developed.

1. U. K. Le Guin, "It was a Dark and Stormy Night," p. 188.

2. After writing this chapter, I discover that these three questions correspond to C. S. Pierce's division of the field of semiotic inquiry into three parts: semantic (symbolic), syntactic (concerned with the production of meaning), and pragmatic (concerned with efficacy). M. Bal and N. Bryson, "Semiotics and Art History," 189.

3. A. Eschmann, K. Hermann, and G. C. Tripathi, The Cult of Jagannath; H. Kulke, "'Ksatriyaization^[*] ' and Social Change in Post-Medieval Orissa."

4. K. C. Sahoo, "Oriya Rama Literature," 20.

5. Philip Lutgendorf makes clear that the rasik sadhana centered on Rama should not be viewed as derivative from that of Krisna^[*] ("The Secret Life of Ramcandra," 228-30).

6. Studies of this potent theme in Indian thought in general include P. D. Devanandan, The Concept of Maya ; R. E Goldman, "The Serpent and the Rope on Stage"; and W. O'Flaherty, Dreams, Illusion, and Other Realities .

7. E. Dowling, " Apate, Agon , and Literary Self-Reflexivity in Euripides' Helen ."

8. A. K. Coomaraswamy, "Reactions to Art in India," 108; J. Williams, "Siva and the Cult of Jagannatha: Iconography and Ambiguity," deals with ambiguity largely in terms of multiple interpretations by the audience. For Indian images it is more difficult than for texts (with commentaries) to fix the response of the "actual" audience. I must reaffirm my conviction that some interpretations may be wrong, although many may be right.

9. Robert Goldman provocatively explores the extension of the philosophical cliché to the theater and to ordinary social discourse in "The Serpent and the Rope on Stage."

10. E. C. Dimock, "A Theology of the Repulsive," 195.

11. Philip Lutgendorf likewise questions claims that the television series has "put paid to Ramlilas" or homogenized the diverse Ramayana^[*] traditions of India as a whole ("Ramayan: The Video," 166-70).

12. R. Barthes, "Introduction to the Structural Analysis of Narratives," 253, 293-95.

13. The writings of Ernst Gombrich, such as Art and Illusion (New York: Bollingen Foundation, 1960), deal with the psychology of representation with a sophisticated awareness of the complexity of the constructed systems of illusionism and ways in which the beholder resolves ambiguities.

14. The passage from Edward Dimock quoted in the text (see note 10) moves from the image of a two-headed deer to a Picasso painting of a woman with two eyes on one side of her head as something not in our experience. I am not convinced that Picasso was asserting multiple realities.

15. M. C. Beach, Early Mughal Painting , 134.

16. Wendy O'Flaherty does not really consider the pictures seriously, but she suggests this interpretation in general ( Dreams, Illusion, and Other Realities , 280-82).

17. A better candidate than Jahangir's Yogavasistha^[*] might be a well-known painting in the Freer Gallery, which includes a note that it represents a dream experienced by the emperor in a well of light. (R. Ettinghausen, Paintings of the Sultans and Emperors of India , plate 12.) Jahangir embraces a diminutive Shah Abbas of Persia, with whom he was at war when this work was painted and whom he never met. Combined with the meticulous representation of surface detail that characterizes painting of this period, there are various features that resist direct reading as visual phenomena. The rulers stand upon a lion and a lamb nestled against a globe that reveals the Safavid and Mughal empires. A large emblematic gold disc shines in the background. Here, as with Salvador Dalí's limp watches or allegorical Elizabethan paintings, one is encouraged to read the image as potent or super-real, which is not to say that other Mughal painting consistently follows European illusionistic conventions. In this case, "dream" must be equated with vision, not with illusion, however.

18. E. Isacco, A. L. Dallapiccola, et al., Krishna the Divine Lover , plate 175; E. Moor, The Hindu Pantheon , plate 93; J. P. Das, Puri Paintings , 135-36. The motif resembles what has been called a chimera in western India, both figures challenging the beholder's ability to identify them.

19. J. P. Das and J. Williams, Palm-Leaf Miniatures , 4.

20. K. K. Patnaik, Kumbhara Chaka , 258-59. The passage is quoted in full in J. P. Das and J. Williams, Palm-Leaf Miniatures , 6.

21. See p. 52. One might wish that Das and Patnaik had recorded actual prices of manuscripts, but the interpersonal and negotiated nature of such transactions in a village would make the use of such information problematic.