5
Narrative Strategies

To apply the terms "narrative" and "strategy" to the making of pictures requires some defense. "Strategy" may seem to imply deliberation on the part of the artist. Yet in the case of the Indian artisan, other forces are usually assumed to override the will or problem-solving role of the individual maker. On the one hand, one might expect the high-caste patron or a learned intermediary such as a raj-pandit to play a guiding role. Chapter 2 should have made it clear, however, that such guidance is demonstrable neither for the artisanal chitrakara nor for the Orissan book illustrator. On the other hand, one might expect revered artistic tradition, known in India as parampara , to guide particularly the professional painter. In fact, even for him there were choices among models, although he may not always have articulated the way he reached his own decision. The palm-leaf illustrator was more likely to work without specific visual models. The very diversity of illustrated books produced in one small area and at one time is a central theme of this book. At the very least, "strategy" may be understood as a shorthand for choices implicit in the work itself and the way it addresses its problems, whatever the human locus and process by which such decisions were made. Precisely because we might wonder whether choices are not made in response to earlier images, it will be necessary to keep visual models in mind.

The problems "narration" presents are great enough when applied to words, let alone to images. From the Latin root narro , "to tell" the term has diverse implications, much discussed in the subfield of narratology. In Indian languages there seems to be no exact equivalent. In Oriya perhaps the closest word is vivarana^[*] , which, however, is more literally translated "coloring," hence "description."^[1] Yet in English, "description" has been thought to concern a static condition, diametrically opposed to "narration," which concerns action.^[2] ç This polarity does not really exist in India, hence a less consistent concern there with action, sometimes equated with "narrativity" in Western critical writing. At the same time Oriya, Sanskrit, and most north Indian languages freely use the word katha , from a verbal root "to tell," which can be equated with "story," a basic form of narrative.^[3] ç Thus it is possible to consider Indian literature as narration without expecting the categories of Western analysis to fit precisely Let me say at the outset that I do not

conceive of narrative as a distinct genre but rather as a procedure (telling a story) found in various types of poetry and prose, both read and heard, including theater.^[4] In an effort not to impose inappropriate standards from a foreign evaluative tradition, or even from other Indian traditions, I shall avoid using "narrativity" as a criterion for artistic success.

Surely some fundamental elements in the structure of telling are universal. In the first place there is a narrator , the teller of the tale. The narrator has long been recognized as a feature of Indian literature, implying an awareness of the subjectivity of the account that has only recently been made explicit in the West. Thus in the Sanskrit epics authors are encoded in the action (Valmiki in the case of the Ramayana^[*] ), in drama the director (sutradhara ) appears onstage, and in Indian stories a commentator often links the sequence of episodes. In the case of Tulsi Das's Hindi version of the epic, there are four symmetrical narrators, who in turn correspond to four interpretations of the same events and accommodate the pluralism of the Indian tradition.^[5] The Adhyatma Ramayana^[*] is comparable, if less symmetrically constructed. Oriya literature seems even less systematic in general, yet the Ramalilas^[*] may present as their teller Valmiki, Siva, and also the particular historical author, such as Vikrama Narendra, whose name recurs constantly in the chanting at Dasapalla.

In the second place, there is the tale itself, with subcomponents such as plot, characters, and language. Here we may turn to the theoretical writing about drama, admitting that not all narrative is drama and that at least some recent drama is minimally narrative. The component of plot has preoccupied Western theoreticians from Aristotle onward, with attention to time span and sequence (both actual and narrated).^[6] Indian theoretical literature dwells more on language and an elaborate system of tropes, whereas plot is a secondary element.^[7] ç Yet in both traditions there are some normative codes governing action, which vary with genre and with the taste of particular periods.

The third major component of narration is the hearer of the tale. Indian literary theory accords a more important role to the audience than has usually been granted in Western writing, albeit a role responsive to the work of literature itself. The point of classical Indian literature, whether narrative or not, is to evoke sentiments in the hearer. These sentiments, usually numbering eight or nine, are known as rasa , literally "flavor" or "essence," for they are distillations of ordinary human emotions. Experiencing them is in itself pleasurable and uplifting for the ideal connoisseur (rasika ); their effect is thus unlike catharsis, which purges the audience's undesirable and real feelings of pity or fear. Suspense plays a small role in general in a tradition such as India's, where most plots are already familiar. Yet the mood of awe and wonder (adbhuta ) is comparable to suspense when an outcome is remarkable.

Rasa is one key to the Sanskrit Ramayana^[*] , often described as built around the distilled sentiment of pity or pathos.^[8] What is important is not whether cut-and-dried systems in fact govern all works of literature, but rather the way in which Indian tradition has long incorporated the listener into the tale, as in recent Western reception criticism. Moreover, recognizing the centrality of mood helps us to understand the prolongation of sequences that present relatively little infor-

mation or action but convey emotion in a way that the West has sometimes dismissed as sentimental.

Can we transfer such a framework to visual images when we talk about narrative pictures? Obviously it would be a mistake to project the elements of one artistic medium upon another, even though they may at times intersect in function, such as in telling stories. India generally recognized an immense gulf in social status between the makers of images and authors, as well as a gulf in the way images and texts are viewed. Still I shall argue that pictures may imply the same three fundamental components as verbal narration.

There is a teller, ultimately the artist, who may call attention to his presence by a signature or, in the recent West, by his very brushwork, as in action painting. In India the intrusion of the artist is often minimal, yet one of the two Orissan traditions considered here, the palm-leaf manuscript, frequently records his name in the colophon. The fact that he is identical with the scribe makes clear his status as narrator, repeating an oft-told tale. Moreover, the author of the text may appear in Indian painting when he plays a prominent role in the story.^[9]

Likewise pictorial narration assumes an audience. This component of narrative may seem more largely implicit for images than for words, although a device such as linear perspective: literally predicates where the ideal viewer stands, which in turn affects the emotional impact of the image. If the beholder's viewpoint is at ground level, the scene is almost inevitably awesome and overwhelming. At least one revered Indian text on painting, the Chitrasutra of the Visnudharmottara^[*] Purana^[*] , describes rasas in particular images and prescribes some moods as suited to pictures on the walls of ordinary houses, some only to those of palaces.^[10] Again the palm-leaf book that includes poetic text and images seems devised for an informed audience, the rasika , or connoisseur. The pata^[*] chitra , on the other hand, addresses any visitor to Puri.

The central element, the tale, may be equated not only with sequences of images, which are the principal concern of this book, but even with individual pictures. In that case we may be tempted to distinguish the narrative picture from the iconic (in India) or from the descriptive (in the West), although there are many mixed examples of both. If only one moment is depicted, multiple events may nonetheless be evoked. Because some elements of the story seem almost inevitably more ambiguous in a single image than in a linear verbal account, one may doubt whether such images are ever truly narrative. Plot sequence cannot necessarily be deduced when we see two figures side by side. Yet it should be remembered that even in a verbal story the order of telling need not correspond to the order in which events occur and that deliberate ambiguity is a legitimate literary device. Likewise we have some rules by which an image is to be read, although many pictures flout these rules. Sequential images are usually seen in a particular order, although we are not prevented from altering that order, whereas reversal is difficult in hearing a story, viewing a play, or reading a novel. Sequence in turn may have implications for causation and for interpreting what happens; hence words often produce a stronger causal certainty than do images.

Other constituents of the tale, such as Aristotle's categories of character and spectacle, are on the other hand often less ambiguous in a picture than in verbal narrative. It is almost impossible to avoid saying something about where an event

takes place or what a character looks like in a picture. The element of "language" is obviously replaced by the physical materials of the image or, more generally, by color and line, which follow their own systems of metaphor and associated meaning.

Finally, in the case of images explicitly accompanying a text, as in an illustrated manuscript, one might expect the text itself to determine narrative choices. "Illustration" implies some dependency upon a pre-existent written story, which leads me to avoid that loaded term as much as possible.^[11] This study is fortunate in presenting several situations that test this assumed dependency. We have four examples of a single text (the Adhyatma Ramayana^[*] ) made by the same artist, as well as a second and quite different text (the Durga Stuti ) illustrated by the same person. There are also multiple examples of another text (the Vaidehisa Vilasa^[*] ) by two different artists. Finally we have four examples of the same concise text (the Lavanyavati^[*] ) illustrated by four purposeful and very distinctive artists. For the theoretically inclined, all these examples may seem unnecessary to my basic contention, that the artist must be regarded as a narrator himself. For the empiricist, however, they may at least demonstrate the variety of relationships possible between the artist and his text. The issue here is not whether he understood the text but rather whether he told his story in the same manner that it did. With this framework in mind, let us examine each of our works or cycles of images as a tale in which events and details are selected, presented in a sequence, and in the process emphasized and interpreted for the viewer. Here, unlike the previous chapter, we will look at each cycle as a whole or in long sequences, rather than dwelling on individual pictures.

Sarathi Madala Patnaik's Adhyatma Ramayanas^[*]

I shall begin with cases where we can discern the artist's role as narrator most clearly Sarathi Madala Patnaik's copious oeuvre from the late nineteenth century includes four illustrated manuscripts of the Adhyatma Ramayana^[*] . As the table in Appendix 1 shows, there is surprisingly little consistency among the four in the choice of scenes to be depicted. When a scene does appear in several versions, it is rarely very similar in composition or detail, as I argued in Chapter 4.^[12] Clearly Sarathi Madala did not retain one of his own works, nor did he work from another illustrated manuscript that served as a model.^[13] ç He may be compared to an oral performer who varies each telling of a tale, improvising rather than attempting to replicate a previous version.

Given this procedure, what particular slant can we discern in his various performances? First, the pace of illustration in his manuscripts follows a certain pattern. There are many illustrations in his works, certainly more than in the work of a second artist who illustrated this same text (C. L. Bharany Collection, ms. no. 2; see Appendix 1, column 6, [Bharany Collection anon.], Figures 78-79), although roughly the same number as in the work of a third (anonymous, National Museum 75.536; Appendix 1, column 5, Figures 75-77).^[14] Sarathi Madala's first manuscript has the most pictures, often one on each side of every folio. In all his manuscripts the ratio of images to text decreases toward the end, when more and more pages are devoted to text alone. Like other scribes, he seems to have copied

the text first, leaving blanks for the images; in his third manuscript, several blanks in the last quarter of the text have not been filled in. Thus the slackening pace of pictures is in part deliberate and not directly the result of exhaustion on the artist's part. One gets an impression that Sarathi Madala produced illustrated books primarily as commodities, although he gradually lost interest in the pictures during the course of work on each manuscript and during his own lifetime.

Some of the distinctive characteristics of the Adhyatma Ramayana^[*] outlined in Chapter 1 are not particularly visible in any artist's illustrations for this text. Perhaps the abstract doctrines of monistic Vedanta would have been difficult to depict, and it is no wonder that long devotional sermons such as Ahalya's paean appear as uninterrupted passages of text. In one manuscript Sarathi Madala does show the creation of Maya Sita, albeit not in a visually self-explanatory form (Figure 59), and the illusory deer is repeated emphatically with the two heads that might serve as an emblem for illusion. Nonetheless, flipping through the illustrations of this text, one would not guess it is a particularly philosophical or devotional version of the story of Rama. Nor do all of Sarathi Madala's pictures tally with what the text says. For example in his image of the seven sal trees done in December 1891, the knotted serpent must be drawn from some separate illustration of a sarpa-bandha and has no particular place in the Adhyatma Ramayana^[*] account. In short, Sarathi Madala seems to have paid scant attention to the words he repeatedly copied.

The impact of his pictures is fairly uniform, more so than is that of two anonymous illustrators of the same text. He is more likely than the other two to omit actions that might seem central to the story; such as Laksmana^[*] cutting off Surpanakha's^[*] nose (absent in both 1891 versions) or Jatayu's^[*] battle with Ravana^[*] (absent in both 1891 versions and in 1892). When action is depicted, the effect is rather static. For example, Rama merely confronts the magic deer, and in the one case when he shoots it, the two stand squarely and a horizontal arrow connects them (Figure 51). Many scenes consist of several figures confronting each other in a row. Nor does a sense of action emerge from successive pictures that include the same figure in a different pose. Settings are rarely elaborated unless they are an essential part of the event, as in the building of the bridge to Lanka^[*] . Many symmetrical compositions center on Rama, corresponding to particular situations but emphasizing their iconic character. And there are many "frame" scenes in which we see the narrator of the story, generally Siva and Parvati. Such compositions produce an effect like that of storytelling that abounds in repetitions, so that the content is easily assimilated. While plot development and action are not stressed, the result is a legitimate form of oral narrative.

I must also grant that all I have said applies less fully to Sarathi Madala's manuscript of January 1891 than to his other works. Here settings are more elaborated and compositions more varied, with few simple series of standing figures. Some scenes are memorable, such as the mutilated and disheveled Surpanakha^[*] kneeling before her distraught brother Ravana^[*] , framed by twisted trees (Figure 58). We do not know the precise circumstances under which Sarathi Madala produced this manuscript, the patron or the price that might have encouraged a greater concern with individual images. There seems no reason to look to some

biological framework of the artist's development for an explanation, for both earlier and later works fall into the more repetitive type. Perhaps the folklorist concerned with constant variations in oral traditions that have multiple existence or the musicologist concerned with performance traditions such as India's, where improvisation is mandatory, could provide a better framework than the art historian for such diversity At any rate, Sarathi Madala does emerge as a village storyteller, much like the itinerant performer of rural Orissa today.

Sarathi Madala Patnaik's Durga And Hanumana Stutis

This prolific artist was by no means limited to a single text. The stutis to Durga and Hanumana, while based on Ramayana^[*] themes, do not present a coherent version of the plotline, nor do they include the metaphysical underpinnings of the Adhyatma Ramayana^[*] . A generic devotee shown worshiping Hanumana in the final illustration is conceivably to be read as Sarathi Madala himself, for the last verse of the text leads directly into the colophon, which begs indulgence for any mistakes by the scribe-illustrator.

The pace of illustration in this manuscript is entirely regular, with pictures on both sides of every page. This may be partly explained by the brevity of the text, with two stutis combined as they are in one pothi . Many illustrations are necessary to bring it up to the standard thickness of a book, in this case thirty-three folios. Similarly, the reduced number of illustrations in some Adhyatma Ramayanas^[*] , especially toward the end, may have to do with the cumbersomeness of the pothi bundle itself, particularly when the dimensions of each leaf are small, so that many leaves are required to cover the text.

Some details in the illustrations of the stutis bear out Sarathi Madala's awareness of the verses he had copied, for instance the depiction of Hanumana's tail ending in a serpent head (Figure 86). At the same time, stock motifs are added that are not called for by the text, such as the two-headed deer (Figure 82). The serpent-weapon with which Indrajita enthralls Rama and Laksmana^[*] appears in the air as the familiar sarpa-bandha of the episode of the seven sal trees, which makes less sense here than would a composition in which the knot enveloped the brothers (Figure 83). This image is not easily read in connection with the next three pictures, in which Rama and Laksmana^[*] pray, standing before Durga, and are subsequently felled by a number of actual serpents (Figure 84). In short, the artist borrows imagery from other kinds of narratives in an allusive way, without attempting to adapt it to the particular action of this story

The compositions of these pictures are as repetitive and static as in any of Sarathi Madala's work. In the many scenes of the goddess or monkey with worshipers, such effects are entirely appropriate to the repetitive verses of the text itself. Scenes of confrontation, such as the serpents engulfing the brothers, are presented conclusively (Figure 84); this is not the indeterminate, tragic struggle of the Laocoön. Figures tend to simple vertical or horizontal positions. Such principles also govern Sarathi Madala's other works, such as the lyric poetry of the Dasapoi or the devotional images of the Artatrana^[*] Chautisa .^[15] Thus despite the

improvisational character of his particular images, he was remarkably consistent in overall effect. As a storyteller he showed an enthusiasm and simplicity that would not tax the audience.

The Ramalila^[*]

The unassuming manuscript of the Ramalila^[*] composed by Krishna Chandra Rajendra must be considered briefly here as comparable to the work of Sarathi Madala Patnaik (Figures 87, 88). Like Sarathi Madala's shorter books, this one is illustrated profusely and with regularity, for a single picture occurs on each side of every leaf that survives. The unknown artist appears to follow the idiosyncrasies of this particular text, for example in substituting Navaguñjara for Jatayu^[*] as the creature that attempts to stop Ravana^[*] (Figure 87).

The compositions, simple and with minimal settings, are similar to those of Sarathi Madala. None are as ambitious as some parts of his Adhyatma Ramayana^[*] of January 1891. Nor are they quite as repetitive as most of his work. The figure drawing is slightly more adept and the poses slightly more varied. Nonetheless such a work shows what may have been the most common pattern of manuscript illustration—motivated by a desire to turn out illustrated works rapidly. These must have been acceptable if they reproduced the desired text and enlivened that with pictures. But no single picture seems to have caught the attention of the artist for long or to have been designed for the viewer/reader to linger over. Although Krishna Chandra Rajendra's text was utilized in performances, the illustrations of this manuscript have no more (and no less) to do with actual Ramalila^[*] enactments than do other images under discussion.

Satrughna's Vaidehisa Vilasas^[*]

Upendra Bhañja's long, demanding poem, filled with verbal conceits, is not the sort of text one might offhand expect to be depicted. In fact such ornate kavyas in general and his work in particular were more commonly illustrated in Orissa than were many straightforward pieces of writing, perhaps reflecting both the general popularity of Upendra's work and the predilections of particular artists. Two artists preferred to work on the Vaidehisa Vilasa^[*] over other texts. The two differ radically in what they make of this complicated literary work.

Satrughna, a karana^[*] of northern Orissa, working early in the nineteenth century, was apparently unique in the way he organized his pictures. They frequently overlap several folios, surround irregular cartouches in which the text floats, and are oriented so that the entire manuscript must be turned perpendicular to the normal direction of reading. Each novel characteristic has its own effect in narrative terms.

The use of joined leaves to extend the composition beyond a single narrow palm frond enabled later artists to make complex iconic scenes in which a large number of figures gather around a deity and little or no text is included. The extended height of such pictures provides a break from the restricted format that had encouraged repetitive rows of figures or sequences of action. Indeed Satrughna also created many large iconic scenes that stand out as "showstoppers"—

emphatic, courtly moments in the story such as Rama and Sita on Mount Chitrakuta^[*] , Rama's grief in the rainy season, Sita's captivity in Lanka^[*] , or the final coronation (Plates 6, 8; Figure 101). His use of color adds to the broad resemblance between such compositions and the hieratic paintings of the chitrakara , although there is no particular connection in figure style.

Yet Satrughna's work is not consistently static to the extent that Sarathi Madala's was. The Baripada version of Tadaki^[*] lunges across five folios, dwarfing the brothers and creating wonder about the outcome (Plate 3). In many scenes a river winds across several leaves to create an unusually developed landscape, in the same way that Upendra Bhañja uses the setting as a mirror and foil to human action. There are several strikingly diagonal compositions (Figure 90).

Moreover, the device of orienting some scenes perpendicular to the normal way of holding the leaf produces variety, complexity, and at times confusion—the last surely deliberate in some of the final battles, where it is hard to pick out individual participants in the melee (Plate 7). In general the shifting orientation of various parts of the manuscript creates an effect analogous to the illusion that is one of the poem's concerns. In the same way, the vivid image of Maya Sita in the fire underscores this theme (Plate 5). The dispersed manuscript of the same text discussed above as an earlier work by Satrughna lacks such effects, at least in the pages so far located. If my historical hypothesis is correct, the artist would seem to have achieved this bold and deliberate chaos in his maturity

The prominence of color as well as the sheer wealth of images makes the Baripada Vaidehisa Vilasa^[*] exceptional among Orissa's manuscripts in the way the pictures tell the story. This effect is clearly not achieved at the expense of the text, as I have already suggested. The floating cartouches of writing play a more integral physical role in the impact of the book than does the usual format of separate panels that can be read independent of the pictures (or vice versa). In Plate 6 they function uncannily as clouds or mist. Their content, which of course must be grasped by holding the leaves perpendicular to the picture's orientation, is more effective in linking sky and earth than it would be if the writing issued from the mouths of the actors like the balloons in a comic strip.^[16] In Figure 99, a line of commentary rises out of Ahalya's rock. I find it hard to imagine the production of such a manuscript by the normal procedure of completing the text first with blanks for the pictures that are executed in a second "campaign" for the blank spaces here are clearly designed with specific compositions in mind, which might well be forgotten after a lapse of several months.

It is difficult to assess today the full impact of this manuscript, many of whose leaves are out of sequence. Even originally, it must have been difficult to read the text on joined leaves, which seems to require that one scrutinize the front and back in sequence, although the pictures compel one to look at one group together without turning them over. On the one hand, this may be viewed as an experiment, perhaps as an unsuccessful one, for no other artist took up the same format for a full manuscript. Later examples employ joined leaves as a coda to a conventional manuscript (Figure 55) or as a single one-sided picture with a brief text, composed of only a few leaves.^[17] Satrughna's own use of the adjective vichaksana^[*] , "ingenious" to describe himself, may be a more positive way of presenting the same phenomenon of novelty. At the very least, his work demonstrates that even

the seemingly conservative genre of palm-leaf manuscripts could tolerate experiment.

On the other hand, one might view the demanding nature of this work as a deliberate puzzle, analogous to Upendra Bhañja's poem. The audience was challenged to read the bandha (knot) in the same way that Rama was challenged to straighten out the seven sal trees. The audience was yet further challenged to follow this manuscript, to figure out which way to hold it and which side of the leaves to look at next.^[18] It worked like a complicated anagrammatic crossword puzzle, in which a literary passage is rearranged and can be reconstructed by solving partial clues and by a broader familiarity with the background text. If narrative is conceived exclusively in terms of action, this work shows inconsistent narrativity and would frustrate hasty attention to plot. Yet if narrative is a tale told to an audience, this work must have been effective in involving the viewer in its rich and subtle imagery

Michha Patajoshi's Vaidehisa Vilasas^[*]

This same text seems to have exercised an even greater attraction for the brahman Michha Patajoshi, who worked in southern Orissa in the first third of the twentieth century. The description by Kulamani Das quoted in Chapter 2 mentioned the illustrator's pleasure in showing an illustrated copy of the Vaidehisa Vilasa^[*] that he had made and in chanting its verses, suggesting that he both depicted and performed his story. Moreover, parts of eight versions of this text by Michha Patajoshi survive, and it appears that he kept previous copies with him, so that his pictures were consistent over time. The other two known works by this artist, patas^[*] on joined leaves depicting the Purl temple, were far less ambitious than the illustrated Vaidehisa Vilasas^[*] that he manufactured prolifically

Michha Patajoshi's pictures for each copy of this text were copious. His books were large in scale, being among the thickest and longest Orissan manuscripts, as if he felt none of the restrictions that may have led Sarathi Madala Patnaik to reduce the number of illustrations toward the end of most of his works to keep them manageable. In fact Michha's pace of illustration is consistent in each work, and his later copies grew progressively longer as he expanded the number of pictures for some episodes.^[19] There are more pictures in proportion to the amount of text than even in Satrughna's highly pictorial work, to judge from the number of scenes devoted to each event.^[20] ç One has the impression that this zesty illustrator saw no physical or artistic reasons to limit the number of pictures he could include.

Michha certainly did not struggle against the constraints of the palm-leaf format but rather welcomed them at a time when printed books must have formed an alternative, even in a traditional village like Balukeshvarpur. There is no difficulty in following his story, which proceeds regularly from left to right, front to back of every folio. Flashbacks are not distinguished—for example Ahalya's tale of sin, which follows her liberation from the curse that followed it (Figures 104, 105)—for these may be assumed to be familiar to the audience. The pictures present events straightforwardly in terms that would make sense to people in rural Orissa during the early decades of this century A detail such as the stretcher on

which Ravana^[*] is carried away is not so much an anachronism as a logical solution to the problem of transporting the body (Figure 138). Top hats and violins appear already in the 1902 manuscript. In case the images themselves are not clear, they are supplemented by the scribe's own chatty captions, which differ from Upendra Bhañja's text in their colloquial language and even in the use of more modern script forms. These captions float unframed within the pictures, whereas the text forms a counterpoint of more regular panels demarcated by small decorative borders.^[21]

As the examination of individual scenes demonstrates, Michha Patajoshi seemed fully to follow the meaning of his favorite poem. There are virtually no discrepancies between pictures and text, although details may be embroidered, such as the squirrel at the building of the bridge to Lanka^[*] (Figure 120). In canto 19, a "garland of wordplay" (jamakas ) devoted to the charms of Mount Chitrakuta^[*] , the artist does not attempt to render the poet's wordplay but does present two tender actions of Rama—marking Sita's brow with ocher (Figure 107) and shooting the crow that attacked her. The image of Maya Sita leaving her true counterpart in the fire is emphasized by the bold, dark form of the flames, if not by the scale of the main actors as in Satrughna's picture (compare Figures 92 and 109). There is generally a sense of zany good humor rather than solemnity in Michha Patajoshi's figures. One imagines him repeating his favorite tale with gusto, aware of the literary complexity of the written text but opting for a good story in his own visual telling.

I cannot help seeing the difference between these two versions of the Vaidehisa Vilasa^[*] as a matter of personal temperament, for there is no uniform regional or period style in the small part of Ganjam District that produced Raghunath Prusti, Sarathi Madala Patnaik, and Michha Patajoshi in quick succession. At the most, Satrughna's remarkable originality might be ascribed to his having worked in isolation at a time and in an area where there were few other palm-leaf illustrators.

Four Versions of the Lavanyavati^[*]

Upendra Bhañja's second and yet more frequently illustrated work provides a particularly neat situation in which to consider various pictorial strategies for presenting the same literary story. In the first place, the Ramayana^[*] is told here with virtuosity but succinctness, and the reader can easily consult my relatively pedestrian English translation in Appendix 3 to get a sense of the words. In the second place, we have four ample, distinct, interesting, and fairly well-preserved sets of pictures of the Lavanyavati^[*] , whose narrative choices can be seen at a glance in the table in Appendix 2.

At the outset, we must recall that this Ramayana^[*] is embedded in a longer and less familiar tale. The magic performance described has a function in the longer plot, serving to arouse the love of the heroine, who identifies the hero with Rama and herself with Sita. Moreover, the presentation of the well-known epic as an ephemeral and illusory performance serves as a commentary upon the very nature of that story. We must also remember, in our enthusiasm for this subject, that it was not mandatory that this brief episode be lavishly illustrated. It was not

depicted at all in some manuscripts of the Lavanyavati^[*] , and in others it was reduced to a single formulaic scene of Rama's coronation (Figure 140).

One remarkable artist whose attention was caught by this subplot was Balabhadra Pathy, working in Jalantara at the southern extreme of Orissan culture. He devoted more folios to the Ramayana^[*] than did any other illustrator, elaborating upon the terse text of the Lavanyavati^[*] yet not pursuing most of its verbal conceits. For example, in the line "He made short work of the monkey (sarabha ) with an arrow (sara )," the central pun and the speed of Rama's action are not depicted. Yet the death and the monkey's reaction are prolonged over several folios (of which Figure 169 is only the last) with events not spelled out by the poet. Pathy indeed seems to be completely carried away by the episode of Hanumana's visit to Lanka^[*] , which figures in the text only with the single line "To get news of his wife he (Rama) sent a messenger." It is tantalizing that the manuscript ceases at this point, so one will never know how or even if Pathy returned to the substance of the poem. At least we can say with assurance that this episode was emphasized in the context of his generally expansive treatment of the Ramayana^[*] .

The puppet-like character of figures in this work has already been mentioned, although no one would argue that the prolific scenes are literal transcriptions of shadow theater as we know it today in Asia. The exaggerated poses have a generally theatrical effect, similar to that of many Indian dramatic forms in which gesture and movement surpass facial expression in conveying emotion. If Sarathi Madala Patnaik and Michha Patajoshi bear comparison with oral storytellers, Pathy orchestrates his pages like the director of a drama, and it is tempting to search for connections with the genre of Prahlada Nataka^[*] , which originated in his home town, Jalantara. The massed figures that enlarge and comment upon some scenes (Figure 173, right) form a chorus. Although the chorus as an institution is not part of the Indian theatrical tradition, this visual device is comparable to the group of gayakas , or singers, who repeat the lines read by the chief chanter in the Ramalilas^[*] of Orissa.

One distinctive aspect of Pathy's narrative forms is certainly pictorial or perhaps broadly literary and not borrowed from the relatively simple staging of Indian theater. More than any of our artists, he uses setting inventively to reflect or enlarge upon the human situation. Thus he gives us three vivid scenes of Mount Malyavan in the rainy season, vignettes devoid of human figures, in which the natural elements are not only appropriate (frogs and a rainbow in Plate 10), but also independently expressive of Rama's sorrow. Pulses of lightning dart and wriggle like living organisms against the black sky. Dotted lines of rain suggest tears and rage, particularly when they cross (Figures 170, 171). These scenes come as close as those of the most subtle Pahari painting to evoking the karuna^[*] rasa by means of landscape, a phenomenon akin to the pathetic fallacy of Western Romanticism. Thus while the artist does not choose to illustrate the poet's verbal tropes literally, he creates; his own visual metaphors that have a comparable effect. There is also artful variety in the scenes, which include starkly ordered buildings (Figure 161), the tangled thicket in which the magic deer moves (Figure 164), and the immense conflagration in Lanka^[*] (Figure 174). In general, Pathy's version of this story is energetic, emotional, at times overwrought, and wearing for its audience.

The anonymous artist of the Dispersed Lavanyavati^[*] probably devoted at most ten folios to the Ramayana^[*] sequence (as opposed to Pathy's more than sixty) and illustrated only events mentioned in the poetic text. Thus the pace of narration here is rapid-fire. This version is nonetheless idiosyncratic for reasons that go beyond the accident of pages missing today. On the eight folios I have located, there are several unexpected omissions. The jump from the birth of the princes to the killing of Tadaki^[*] , occurring on a single side of a page, omits only one incident mentioned in the text but leads the viewer suddenly from Rama as a baby to the full-grown hero (Figure 143). On the next-to-last surviving folio, depictions of Garuda^[*] , Kumbhakarna^[*] , and all fighting are omitted (Figures 153, 154).^[22] Most amazing is the jump from the encounter with Parasurama to that with Bharata, side by side on a single leaf (Figure 147), omitting any explicit image of the banishment, which one might take to be a critical turning point in the plot.

There are also several alterations of the sequence of events presented in the Lavanyavati^[*] , largely prolepses, or flash-forwards. The two most striking cases are the inversion of Sita's kidnap with the Sabari's gift (Figure 149), and that of Lavanyavati's^[*] final observation with Rama's coronation (Figure 156). In this assertion, I assume that the pictures are intended to be read from left to right like the writing, as is true generally in manuscript illustrations. In earlier Indian sculpture, however, we have seen inconsistent directionality (see Chapter 3). It is probable that this artist was his own scribe and hence literate, therefore predisposed to compose his work from left to right. But perhaps our expectation of consistency is at fault. What alternatives are there to saying he acted entirely on the basis of whim?

In all these narrative surprises, I discern several principles at work, of which the last example may exemplify the simplest—a choice to "foreground" the episode shown first. Obviously Lavanyavati^[*] is watching the coronation while giving her final interpretation; placing her first may simply bring us back emphatically to the primary plotline of the poem. This interpretation does not contradict our reading of the opposite side of the same leaf, where Sita's test is shown first, to the left, as the remarkable event that Rama watches.^[23]

A second principle that may enter into the surprising omissions is that of deliberate pairing on a single page, to make a point of similarity, contrast, or another connection that images convey more subtly than words. Thus the jump from the infant Rama to his killing Tadaki^[*] makes the point that he was precocious. The encounters with Parasurama and Bharata are comparable cases of submission to Rama's virtue by a potential opponent. The inversion and pairing of Sita with the Sabari underscore the piety of both women, who gave generously to strangers, with quite different results. Similarly on two sides of a single folio we move from Tadaki^[*] to Ahalya (both cursed, one evil, one virtuous) and from Ravana^[*] being well advised by Mandodari to the heroes ensnared by serpents. What is important is not the precise point of each comparison, which must remain conjectural, but rather that some such comparison would explain the selection and placement of events.

In one case, we have noted an insertion not found in any verbal account of the story I know: Valmiki's presence as an actor in Dasaratha's sacrifice (Figure 142). Whether or not this scene is the invention of this particular illustrator, his treatment of the sage seems purposeful—paired with the king at the center of the

folio, presenting the sacred porridge dramatically just above the string-hole, and emphasized by the way the second sage, Risyasringa^[*] , turns to look at him. On the final leaf of the Ramayana^[*] sequence a sage-like figure appears to the left of center, but this is surely the magician talking to the king of Simhala at the end of the performance. In both cases a narrator of the tale is gratuitously brought into the telling, calling attention to the fictive status of this telling of the sacred tale. In each scene the folio begins on the left with a small, demure female figure who observes the scene and mediates between the male narrator and his story.^[24]

In general the goal of this artist seems to be interpretative: he does not wish to tell the story, assuming its familiarity to his audience, but rather aspires to use it inventively. And his use reflects his visual medium, as the comparisons above suggest. In the same way, he embroiders many scenes with natural details that are appropriate but do not necessarily allude to other events. Squirrels climb the seven sal trees as well as the bridge to Lanka^[*] . There are never more than two events on a page in this section of the manuscript, and space surrounds the large figures, giving the scenes a certain mythic grandeur.^[25] There is a general alternation between scenes of action, with strong diagonals, and more sedentary episodes, which, however, are not stock tableaux but rather embroidered with delicate and varied detail. The elegant drawing and composition of this manuscript may explain why it has often been selected as the acme of Orissan manuscript illustration. This judgment, while perhaps unfair to other illustrators who had other goals, reflects the independence of this artist from his text and his great accomplishment as a draftsman.

The illustrator of the Round Lavanyavati^[*] , also anonymous, provides an extreme contrast with the artist of the dispersed set. Here twelve pages are used to tell the story, and these bear all ten stanzas of the text, unlike the dispersed set, where this portion of text presumably occurred on a previous folio. Thus in an equally restricted space, this third artist worked in almost every event mentioned in the text.^[26] No selection occurred here, and only one minor inversion of events—Sita's test occurring to the left of Vibhisana's^[*] coronation, for which I see no particular explanation (Figures 35, 36). The seven sal trees precede the lifting of Dundubhi's bones (Figure 26) as in the Lavanyavati^[*] itself, a reversal of the standard sequence of these exploits even in Upendra Bhañja's other poetry and unlike other illustrated versions of this text. The only episode included here that does not appear in the text is the minor and ubiquitous addition of the squirrel at the building of the bridge to Lanka^[*] (Figure 29). In short, this was an unusually literal illustrator.

The pace of his pictures is consistent, with two to four scenes on a page. Thus the exceptional single image of Ravana's^[*] death in isolation stands out climactically and effectively, by Aristotelian standards. The Ramayana^[*] sequence as a whole is emphasized in the manuscript, not as in the dispersed set by the mythic quality of its images, but rather by their snappier pace, for elsewhere there are usually one or at most two scenes per page, with longer passages of text. The small scenes of the Ramayana^[*] sequence are separated by frames at the beginning, but those disappear after the exile, accelerating the speed of the action.

While in general one can find a visual continuity between the writing of most texts and accompanying images executed by the same artist, there is particular

harmony between the rounded hand of this scribe and his drawing of the figures. Individual pages are neatly composed, for instance the framing of Visvamitra's calm sacrifice by the slaying of frenzied demons on either side (Figures 6, 7). Staccato beings surround the kidnap of Sita (Figure 19). Each part is pleasing, but no detail is embroidered at the expense of clarity.

The maker of this manuscript certainly worked with great artistry; I would characterize him as the optimum "illustrator." Thus he seems devoted to his text, as was Michha Patajoshi to the Vaidehisa Vilasa^[*] . His interest, even more than Michha's lay in the plot rather than in Upendra Bhañja's wordplay. He was also perhaps more adept as a craftsman; but in preferring not to let any distinctive slant in interpretation intrude, he also dispensed with the whimsy that gives Michha's work its charm.

Finally Raghunath Prusti, working near Sarathi Madala Patnaik in central Ganjam District in the last two decades of the nineteenth century, made the Lavanyavati^[*] his masterpiece. This volume was longer and more profusely illustrated than any he produced. The Ramayana^[*] portion, comprising eleven or twelve folios, is as dense with pictures as any we have seen.^[27] Among the events mentioned in Upendra Bhañja's poem, only the rain in Lomapada's kingdom and Risyasringa's^[*] marriage are not depicted. These omissions are worth mentioning because so many episodes not in the Lavanyavati^[*] text are included. In fact Prusti shows extraordinary independence of the text, although his motivations are different from those of Balabhadra Pathy or from those of the artist of the dispersed manuscript.

A significant factor in Prusti's images was the model of Buguda, roughly 30 kilometers away. Our security in dating both the illustrator and the wall paintings, as well as the esteem in which the Viranchi^[*] Narayana^[*] Temple continues to be held, makes unusually convincing the suggestion that this particular monument directly influenced him in some instances. One striking case is the image of Rama and Laksmana^[*] binding up their locks of hair as they begin their exile (Figure 183). This incident, mentioned in the Vaidehisa Vilasa^[*] and not in the Lavanyavati^[*] , is depicted at Buguda (Figure 201, tier 4, left) and in Michha Patajoshi's first copy of the Vaidehisa Vilasa^[*] (Figure 106).^[28] Prusti showed the two brothers in poses very similar to those at Buguda. He did not, however, simply reproduce the entire scene but spread it out to accommodate the palm-leaf format, shifting Sita to the left so that she introduces the group, as in the mural, which was read from right to left in this tier.

Prusti's Lavanyavati^[*] seems also to borrow from several large and vivid compositions set on hilltops at Buguda. Prusti is able to present the scene of the lovers on Mount Chitrakuta^[*] and Bharata's visit in the slightly unusual sequence in which they occur in this poem while retaining the composition of the mural, which was read from right to left beginning with Bharata, the sequence of the Vaidehisa Vilasa^[*] (compare Figures 183, 205). Prusti condenses this hill into a single dome and only once depicts Rama, shown marking Sita's brow with ocher in precisely the same pose as at Buguda. Thus Laksmana^[*] alone receives Bharata in the manuscript, holding an arrow that suggests his initial alarm at the approaching party. Here again we see a skillful adaptation of the almost square mural composition to the narrow, horizontal palm frond. Prusti also includes the incident of the crow (mentioned in the Lavanyavati^[*] ), omits the three mothers, and converts the small ele-

phant from the bottom of the wall into a more integral part of the courtly procession and hence of the story

Prusti's second scene set on a hill depicts the brothers' meeting with vanaras on Mount Malyavan, together with the end of the bridge to Lanka^[*] , presumably Mahendragiri (Figure 189). The former scene is somewhat indeterminate as a moment in the story; its presence is explained by its mirroring a composition at Buguda that is also paired with the building of the bridge. Here the later artist has reversed the sequence of Buguda, which was viewed in circumambulation from right to left. At the same time he retains the direction of movement on the bridge toward the left, a point of irrationality that perhaps would trouble only the most pedantic viewer.

Prusti's third hill scene follows the Buguda paintings most precisely, presenting problems in interpretation and suggesting that the later artist was as puzzled as we are about the precise meaning of events to the right (compare Figure 190, Plate 11). Questions arise because the artist here reproduces the mural composition, which places this episode after the scene of cutting Ravana's^[*] umbrellas and hence presumably too late in the plot for Laksmana's^[*] threat to the monkeys to make sense. I know of no version in which Laksmana^[*] would straighten an arrow proffered by Jambavan during the Yuddha Kanda^[*] .^[29] In the scene to the left, Prusti extends the mural's dramatic central grouping across the narrow format of the page, thereby rendering the falling umbrellas more vivid than at Buguda.

Some scenes in the Mundamarai manuscript are clearly not borrowed from Buguda. Nor are they mentioned in the Lavanyavati^[*] itself, although they occur in other texts, which may account for Prusti's inclusion of them. For example, the three Sabaras in Figure 182 are explained if we remember that in the Vaidehisa Vilasa^[*] Guha, whom the exiles meet at this point, is called Guhasabara.^[30] The pair of vases labeled purna^[*] kumbha in Figure 181 may allude to the long metaphor in that same text that compares Kaikeyi to a milkmaid, Ayodhya to a pot, and the peace of the city to milk that is churned (17.32). This conceit is implicit in the wordplay of the Lavanyavati^[*] , Manthara-manthana (literally, her churning), although that poem does not mention a pot in particular. A similar vase, flanked by auspicious fish (a common enough motif in wedding paintings in rural Orissa) had appeared in Prusti's illustrated divination cards, the Prasna Chudamani^[*] .^[31] ç Possibly the appeal of the crisp emblem supplemented the artist's admiration for Upendra Bhañja's metaphor.

Two more "added events" present slightly greater problems. No version of the death of Tadaki^[*] includes Bharata along with Rama and Laksmana^[*] (Figure 177). Perhaps Prusti drew on some lost oral tradition, or perhaps he invented this variant himself. His scene of monkeys running over the roof of the palace in Lanka^[*] is another interpolation, whose rationale in terms of design has been suggested in Chapter 3 (Figure 191). Here it is also possible that he incorporated some image of Hanumana's escapades from the Sundara Kanda^[*] , otherwise absent here.

In two places we can be sure this inventive illustrator departed from the simple narrative sequence of any version of the story to compose his pages artfully, without direct dependence upon the Buguda model. One is the kidnap of Sita, where I have argued that logic of place takes precedence over that of time, so that scenes at the two ends of the leaf occur on either side of Sita's hut (Figure 187).

The second example is Valin's death, where Rama and the fighting monkeys frame the dying Valin, the result inserted between two components in the action (Figure 189, left). One can imagine Prusti reasoning, "It is so clear what happened in each case that I can throw chronological caution to the winds."

And yet I must confess that of all versions of the Ramayana^[*] , I have greatest difficulty following and remembering Prusti's. This may be partly a result of his very inventiveness, as he gave the story new twists. It is also the result of his complexity as an artist and the several distinct principles that underlie his work. He must have admired the Buguda paintings and copied many of their dramatic compositions and rich details. But he also adapted the model to the narrow format of the palm leaf, normally read from left to right. He further enriched the visual detail and devised his own vivid forms. Moreover, he presented the narrative inventively, so that neither the story nor the images were repetitive. The story is told in fits and starts, but at the same time no static iconic intrusion interrupts the plot. This entire portion is denser, less embroidered with clever contemporary touches, and less realistically descriptive than his illustrations for the Lavanyavati^[*] story proper. Once again this performance is set off from the poem in which it is embedded. To sum it up simply, Prusti seems to have been impressed by the elegant paintings of Buguda but to have adapted their model quite freely to his own visual and narrative requirements.

On the whole, the four versions of this text considered here all set the Ramayana^[*] sequence apart from the rest of the story and emphasize its role by various formal means. In all, the pace of narration picks up as the rate of illustration increases vis-à-vis the condensed text. The artist of the Round Lavanyavati^[*] followed the plot of the text most faithfully and conveys with clarity every part of the action actually mentioned. Balabhadra Pathy was moved by narrative gusto to go beyond Upendra Bhañja's words and is perhaps most captivated by the moods of the story, which his pictures convey. The artist of the Dispersed Lavanyavati^[*] felt yet more independence from the text, selecting and rearranging the events to create new points by visual means. Raghunath Prusti was perhaps most independent of his text in this section, captivated by an effective visual model as well as by his own perception of the problems of the palm-leaf form.

The Brahma Ramayana^[*]

Our last manuscript is the least narrative literary work considered here in the sense of presenting a sequence of actions. This by no means leads to a lack of images, for it is profusely illustrated. Of 132 pages, 63 bear pictures with no text. While 23 bear text alone, on many of these the amount of writing is small, occasionally a single line surrounded by empty space and ornate end-borders. The designer was clearly concerned with the visual elegance of every page.

Because the standard story of the Ramayana^[*] is confined to folios 12 to 26, one might regard this portion, like the Lavanyavati^[*] account, as a capsule version of the entire epic, albeit truncated. Unlike the versions of Upendra Bhañja's poem that we have just considered, however, this portion of the story does not particularly stand out in the manuscript. The artist of the Brahma Ramayana^[*] resembles that of the Round Lavanyavati^[*] in his fidelity to his text, which is illustrated verse

by verse, a simple phrase such as "He liberated the distressed Ahalya from the curse of Gautama," being fleshed out with an effective image of Rama extending his foot toward a square rock followed by a worshiping woman (Figure 196). The enumerative series of events in the verses is comparable to the discrete, abruptly separated pictures.

The general nature of the forms and their import, however, are perhaps as close to those of Balabhadra Pathy as to any work we have seen. Thus the figures are stylized, if less radically so than Pathy's, drawn with assurance and not repetitive. Elements of setting are detailed and related in design to the human action, although not quite in Pathy's expressive manner. For example the tangled arrows of fighting in Figure 195 and the flowers showering upon the wedding in Figure 197 convey distinct dramatic situations.

Other devotional parts of this manuscript may use setting with even greater inventiveness. To the right in Figure 194, the panel sprigged and surrounded by small plants forms an elusively haunting image of Brahma's garden, appropriate enough but not described in the actual verses. The next-to-last folio bears one of the most charming and decorative landscapes in Orissan manuscript illustration (Figure 199). Possibly this is a rising sun, appropriate to the sacrifice performed by the brahmans to the right, or possibly the sun is setting to indicate that the Brahma Ramayana^[*] is complete (samapta ).^[32] In any case the reference is obscure and the image itself haunting.

The bulk of this work is broadly similar to the Gita Govinda in expressing devotion to God in the guise of a romantic relation between sakhis and lover. Thus as with many illustrated versions of Jayadeva's poem, we see a series of scenes of the hero with one or more devotees in amorous situations. Variations are small and subtle but deliberate, and most scenes are carefully composed. Clearly plot is irrelevant to the artist, although these images do revolve around the literary sentiment of Love (sringara^[*] rasa ). While not strictly narrative in Western usage, these pictures develop the emotional element that is present in many Indian stories.

Jagannath Mahapatra's Painted Set

Turning to the pata^[*] tradition, we begin with the set of seventy-five paintings made by a living artist as most straightforward in storytelling (Figures 223-52). This set was to some extent designed as a book, to be seen sequentially with text probably appearing below and certainly clearly formulated in the mind of the painter. Yet it differs from a palm-leaf manuscript in the way the picture has primacy, dominating every page, so that only one image is visible at a time. Moreover, the "text" was the creation of the artist, on the basis of oral and theatrical versions of the Ramayana^[*] . Jagannath Mahapatra was indeed the narrator of this tale, like any village storyteller.

This painter was not especially indebted to Buguda, although that was in some sense part of his professional heritage. Nor have I found any other particular extended group of images from which he borrowed. Visnu^[*] Anantasayana^[*] and the coronation of Prima were common enough subjects for the chitrakara , and his compositions resemble previous versions of these subjects painted on walls.^[33]

Certainly he worked within the visual language of his tradition. On the whole, however, he invented this set for a specific patron, Halina Zealey, who was generally interested in the story and brought in a professor from Bhubaneswar for guidance, but who ultimately seems to have given Jagannath Mahapatra carte blanche.^[34] I would urge that this situation not be understood as unique to the twentieth century but rather as a circumstance that might have existed at times in the past. Even the "lowly" artisan was able to devise his own narrative and visual strategy.

In addition to the selection of some unusual parts of the broad cycle to illustrate (Mahiravana^[*] , the Uttara Kanda^[*] ), as already discussed, the pace of this set of paintings is its most striking narrative feature. Often two or more scenes illustrate what might be thought of as one episode, sometimes an episode that even the palm-leaf illustrators, less constricted in space, showed within a single frame. Thus the birth of the four sons occupies three quite similar pictures. Kumbhakarna's^[*] waking and his death are separated. Rama's decision to hold an Asvamedha takes place in two scenes of no great distinction in terms of image or plot. On the whole one senses more "stretching" of the story at the beginning (when the planner perhaps worried how he would fill seventy-five frames) and at the end (when he found himself left with thirteen frames for the Uttara Kanda^[*] ), whereas in the middle he chose to omit the shooting of the seven sal trees. In short, his spatial constraints were a bit like the time constraints of an oral performer who must fill an agreed-upon period but not go beyond.

In the treatment of individual scenes, there often seems to be a deliberate alternation between the static and the active. This resembles Indian drama, which traditionally includes nonactive sequences that develop the mood but not the plot. Some scenes are symmetrically composed like the tableaux of the theatrical tradition of jhanki^[*] , to be savored and even worshiped in iconic form, rather than played for action (Figures 223, 240, 244, 249, 250-52). At the same time the static does not dominate, and the artist is clearly motivated by a desire to achieve visual variety as well as to get significant parts of the story across.

Finally, it is interesting that some part of the audience did not regard the sequence of events as an unalterable causal chain. The Cuttack publication of the same seventy-five episodes begins with Anantasayana^[*] , as a standard frame, whereas Jagannath Mahapatra had logically placed this episode after Risyasringa's^[*] sacrifice, when Visnu's^[*] decision to be born as Rama was brought about. Likewise, his placement of Sita's birth before Rama's is true to the Oriya tradition, whereas the Cuttack authors followed the more generally acceptable Indian pattern, in which the bride is younger. We see how this freedom of attitude toward plot, characteristic of India, arises from multiple versions of one story and from the absence of the compulsive reasoning post hoc, ergo propter hoc .

The Parlakhemundi Playing Cards

Sequence is yet more problematic in the set of playing cards made in Parlakhemundi before 1918, which belongs to the professional chitrakara tradition in the technique by which it was painted and in its general appearance. Yet the version of the story clearly has little to do with that of Jagannath Mahapatra, lacking

for example the Mahiravana^[*] episode and the entire Uttara Kanda^[*] . Its eighty-eight numbered cards cover the story up to Sita's banishment, to which the Raghurajpur set devoted sixty-three scenes. Since there is no "stretching" of a single episode over several pictures, substantially more events are included in the Parlakhemundi set. Nor does the specific narrative selection of this set particularly resemble that of the manuscript of Balabhadra Pathy, which was probably produced 50 kilometers away There was no south Orissan consensus. Qua story, this version follows a widely shared outline that cannot be specifically traced to any written text, with some idiosyncratic touches—the names of demons Hanumana encounters on his visit to Lanka^[*] (Chhaya and Grama Devi), or the way he is seated before Ravana^[*] on something that may combine coiled tail and serpent (Figure 266), as discussed in the preceding chapter. Thus it seems that the painter of these cards drew and perhaps embroidered upon a local oral version of the Ramayana^[*] in presenting his version of the story.

The images themselves are remarkably varied and dramatically effective, considering their small circular format, to which the compositions are neatly adjusted. Ravana^[*] , with one eye in each of his ten faces, resembles a sinister beetle (Figure 266). Hanumana's exuberant tail winds like a vine around demons in an unusually heraldic version of the burning; of Lanka^[*] , suitable to the small card, where rectilinear architecture would work against the round format (Figure 267). Parts of some scenes are tilted to suggest onward progress against obstacles, as well as to accommodate the round format (Figures 264, 265). While some motifs are widely shared in the professional painting tradition, I am struck by many unique solutions in these cards.

The sequence of episodes seems to proceed in chronological order, from card one through card ten of the successive suits—except for the end of the second suit. There, after Dasaratha's death we find on card ten the label, "Sri Rama, having been given the turban [i.e. crowned] sits on the throne with Sita in his lap" (Figure 268). The picture is a standard version of the coronation, including monkeys, which one would expect to conclude the Ramayana^[*] . The third suit begins with four scenes that would in standard sequence precede the coronation—the revival of Laksmana^[*] , the fight between Rama and Ravana^[*] , the coronation of Vibhisana^[*] , and Sita's test by fire. The story then resumes with a second scene of Bharadvaja's ashram and proceeds m more accepted order. The final card of the eighth suit shows Hanumana carrying the mountain, and subsequent events already shown in the second and third suits are not repeated.

Two explanations present themselves. On the one hand, this sequence might represent some accidental circumstance or miscalculation on the part of the artist that we cannot precisely reconstruct. On the other hand, it may constitute a deliberate prolepsis, perhaps a prediction made in Bharadvaja's ashram (hence the exceptional repetition of this scene) of the hero's ultimate future. The coronation also forms a visual parallel to the scene of Rama and Sita in the same pose on Mount Chitrakuta^[*] , alone except for Laksmana^[*] and one unidentified figure in a chariot.

In either case, this unusual sequence must also be understood in view of the actual use of these pictures as playing cards. In fact, as the cards were shuffled and used, the story would not normally be seen in sequence. One need not discuss the

intricacies of the Orissan game to imagine wild conjunctions of parts of the plot in the course of ordinary play. Inevitably the narrative structure here would be more fluid than in other texts or sets of images, and it may be a simple matter of convenience that most of the deck is in chronological order: the artist would have worked this way to be sure of getting all the major events in. The arrangement of these images was in part determined by the user, making them like computer hypertext or children's books that offer alternative plots, described as "make your own story." This is literally a playful version of the sacred tale used to embellish the afternoon hand of cards, no doubt glossed with wisecracks on the part of the players.

Large Patas^[*]

The traditional thia badhia^[*] , or detailed plan, of the Jagannatha Temple included one scene from the Ramayana^[*] in the upper right corner, often the final fight with Ravana^[*] (Figures 216, 217). In such patas^[*] , scenes from the life of Krisna^[*] sometimes flanked the temple in vertical rows (Figure 216), and these also occurred in a circle in the type known as Krisna's^[*] Birthday pictures.^[35] From such sources, paintings devoted to telling an extended story in multiple panels have become particularly popular in recent years. Perhaps the circular format of the small panels can be attributed to the model of playing cards as well. I have selected four examples, admittedly at random among the wide number produced today, to demonstrate diversity of patterns of organization (Plate 12, Figures 253, 254, and 289 in Appendix 5).^[36] ç

One feature shared by all the modern story patas^[*] of the Ramayana^[*] I have seen is the placement of the coronation in the center. This may seem natural enough as the auspicious conclusion, especially if the Uttara Kanda^[*] is omitted (as it is not in the last two examples). The regular execution of wedding paintings on the walls of houses by the chitrakaras and the centrality of this subject there (Figure 262) may have made it particularly appropriate as the focus of the pata^[*] . It was the most prominent iconic form for Rama. A pata^[*] , by virtue of its origin as a picture sold to the pilgrim and as a replacement image made for the Purl Temple, is conceived of as a stand-in for the murti , literally the "body," of a major divinity.

As Appendix 5 indicates, the selection of events is not consistent among the large examples, nor does it correspond to Jagannath Mahapatra's set. Among the episodes on which there is complete consensus, some may seem natural choices: Rama's birth, his encounters with Tadaki^[*] and Ahalya, the golden deer, Jatayu's^[*] attack, Hanumana and Sita in the asoka grove, and Rama's final fight with Ravana^[*] . Mahiravana^[*] and Durga appear in the larger Danda Sahi painting, in the work of Bhikari Maharana, and in Jagannath Mahapatra's set. The inclusion of these events, which were not in the mainstream of Orissan texts or manuscript illustrations, may reflect the Sahi Jatra tradition of nearby Puri. The Raghurajpur pata^[*] with twenty-one small scenes (Figure 289A) and the Danda Sahi one with twenty-two (Figure 289B) treat events through Hanumana's visit to Lanka^[*] at a fairly consistent pace, but both compress the conclusion into two or three battles.^[37]

The sequence of scenes follows some general principles. Events usually proceed clockwise, whether organized in a circle or in a square.^[38] The beginning of

the sequence is on the lower left in Figure 289C and 289D, but on the upper left in Figure 289A. Vertical panels are apparently read downward. The greatest inconsistency in sequence occurs in the multiple vertical rows, which progress inward toward the coronation in different ways in Figure 289C and 289D. A painting such as that in Figure 289B is complex and sophisticated in the way it reconciles chronology with placement on the canvas. Despite some shared principles, each pata^[*] differs somewhat. Surely we may entertain the possibility of such plural patterns in the past.

Buguda Wall Paintings

Finally we return to the oldest cycle of paintings of the Ramayana^[*] , possibly contemporary with the earliest palm-leaf illustrations as well, for it is clearly dated in the 1820s. No text accompanies the Buguda murals, nor does my analysis suggest fidelity to one written version. It is tempting to connect this cycle with Upendra Bhañja, a member of the family of the builder of the Virañchi Narayana^[*] Temple. While many scenes resemble the account in the Vaidehisa Vilasa^[*] , some do not. Thus the distinctive image of the seven sal trees knotted by a serpent is absent (Figure 203, center), and the Buguda version of Laksmana^[*] straightening his arrow outside Kiskindha^[*] is explained neither by that poem nor by the Lavanyavati^[*] (Plate 11). In general we may surmise both that the artists knew Upendra Bhañja's works along with additional texts and oral tales and that they may have introduced their own interpretations in ways we can never precisely reconstruct. Moreover, damage to some of the paintings creates difficulties in interpretation. Here, in fact, we must proceed backward to understand these images in light of the greater certainty, or at least lesser uncertainty, of more recent manuscripts and paintings.

Another complication most striking at Buguda among all our examples is the likelihood that the artist or artists worked on one wall at a time without an over-arching plan in mind from the beginning. Possibly a plan was altered on an ad hoc basis as the painting proceeded, which is not to imply that the result was therefore less successful. In fact one might argue that it was more creative to make adjustments while the work was in progress. We cannot be certain that the painting was executed in the sequence of circumambulation, and it remains possible that several parts were under way simultaneously, worked on by different painters.

Thus walls A, B, C, and presumably L follow a pattern of depicting many incidents sparely to create a running narrative. The effect of each register is similar to that of a palm leaf, and it is conceivable that the painters worked from an illustrated manuscript. If so, however, they did not follow it slavishly, The vertical sequence on the first three walls moves logically enough from bottom to top, top to bottom, then again bottom to top. Yet the horizontal movement is irregular on wall A, the most complex of the first three. Wall B does follow a boustrophedon course, and C has only one incident per tier. From this sequence one might deduce that the complexity of A was regarded as excessive. If wall L were preserved, it would confirm or deny this hypothesis. At least we can say that while working on wall A, the designer did not take it as his primary task to present the story clearly, as if from scratch. Indeed, because the viewer would surely have been presumed to

have some familiarity with the Ramayana^[*] , the artist was cast in the role of someone retelling a well-known story creatively, to entertain and please his audience.

Wails D through K make simple sequential reading yet more difficult. Thus the scenes at the top of D and E take place in the section of the story covered on wall A and might conceivably have been omitted there with such emphatic treatment in mind. Wall F follows after the events of C, and the scenes at the bottom of D and E follow that, preceding wall G. This is the point at which improvisation may have taken place, and any precise explanation for changes in plan is speculative.

Nonetheless, some general motivations seem to be at work. Perhaps the most natural explanation for the large scenes is that they are iconic, as opposed to the preceding narrative portions. In fact, only one of the actual images here corresponds to a type standard in images of worship, and that is relatively minor in its placement here—Hanumana with Mount Gandhamadana (Figure 211). In the other scenes Rama is no larger than other actors and is sometimes entirely absent (Plate 11). Thus the term "iconic" is not entirely appropriate.

An alternative literary concern is apparent, however: that of mood, or rasa , as opposed to plot. Thus the first large scene, Bharata's visit (Figure 204), is played for sentiment, mainly pathos, rather than as a simple statement of religious devotion, which would be implicit in the worship of Rama's shoes. Here one feels the sorrow of the court, including the widowed queens, bereft of ornament. Bharata confronts his brother as an equal human being, his humbly hunched pose contrasting with Rama's energetic ebullience despite his modest circumstances. This is high drama in the Indian tradition, and even though the scene lacks physical action, it is packed with emotion that tells a story

The neighboring scene on Mount Chitrakuta^[*] combines pathos with romance, as Rama decorously yet tenderly touches Sita's brow with red earth because he lacks the red lead that a princess could expect to use as ornament (Figure 206, top). The isolation of the figures underscores not only the intimacy of the moment but also their impoverished situation. The demotion of the building of the bridge to Lanka^[*] , often regarded as a pivotal event, to the position of a footnote on this wall suggests the greater interest in more emotional episodes on the part of this designer.

The next two major scenes on walls F and G lead us toward the heroic sentiment. Laksmana^[*] straightening his arrow evokes compassion again, that the brother of the heroic Rama should be ignored by a mere monkey (Plate 11). The cutting of the umbrellas is Rama's first major humiliation of Ravana^[*] , as opposed to the building of the bridge, which forms a major turning point in military strategy (Figures 208, 209). Vibhisana's^[*] dramatic gesticulation and the forlornly fluttering umbrellas underscore the triumph inherent in this episode. An element of humor is woven in as monkeys accompany the heroic event on a rural type of drum below.

The pace of battle quickens in smaller scenes of the fight with Kumbhakarna^[*] and Indrajita, laced with arrows (Figure 212). Then the epic confrontation between Rama and Ravana^[*] occupies the maximum area of any of the Ramayana^[*] scenes, its original impact only dimly glimpsed today because of the much dam-

aged condition of the south wall (Figure 213). And yet this vivid triumph of the heroic forces did not conclude the sequence, for it was followed by the smaller scenes of wall L, which, whatever their content, might seem anticlimactic. This sequence serves to remind us that the structure of the paintings, however dramatic, did not follow an Aristotelian pattern, structured by norms of unity. This is the world of rasa , or sentiment, in which the goal is emotional richness, a linked sequence of moods.

Yet surely. visual goals were also felt. The selection of four events set on hills for the major scenes of walls D through G is partly a matter of composition and balanced appearance. The form of a mountain would have interrupted the rapid sequence of scenes on the earlier walls, hence the omission of these events there.^[39] Here the hill connects the entire wall despite the presence of an architectural niche in the center. The small natural touches worked into the base of the hills, visible because they are placed close to eye level, provide a chance for the artist to show off his refinement and invention in a way less appropriate in the registers of the initial walls, which imply uniform scale all the way to the top.

On wall A, Visvamitra's sacrifice, Sita's wedding, and an unexplained scene set in a shrine are aligned, suggesting a relationship of three sacred events (Figure 201). This connection shows an appreciation of the visual properties of images, making comparison between the three ceremonial fires possible in the same way the artist of the Dispersed Lavanyavati^[*] manuscript compared Sita and the Sabari. Moreover, a vertical series of architectural forms connects the three horizontal tiers in the same way that the flames of Risyasringa's^[*] and Visvamitra's sacrifices connect the lower two tiers. The result is not a tightly constructed unified composition but rather various chains that link the wall in pieces, lest it dissolve into a mere series of unrelated registers.

In short, we may view the designer at Buguda as an inventive storyteller, whose medium, of course, was pictures rather than words. The effectiveness of individual images as elegant pieces of design is borne out by the impact they had on subsequent artists. Thus the wall painting of Manikarnika Sahi in Puri (Figure 215) and the wedding box of Bhagavata Maharana (Figures 258-61) demonstrate the exemplary. status of the Buguda murals. And at least one palm-leaf illustrator, Raghunath Prusti, seems to have used these compositions as a point of departure for illustrations in a quite different genre.

Many strategies are at work in Orissan narrative pictures, which begin with different versions of the story and follow these versions with different degrees of fidelity. Some present action as clearly as possible, creating a sequence of events that may be read as causally linked. Some take that familiar action for granted and weave around it a commentary in philosophical or emotional terms. Both story and commentary may be long-winded, embroidered, and allusive or terse and direct.