महावीरचरितम् (महाकवि भवभूति - वीररस-प्रधान रामकथा नाटक सम्पूर्ण ७ अङ्क सटीक)

TO CHARLES ROCKWELL LANMAN MY HONOURED TEACHER निपीतं यद्विद्यामृतमनुदिवं वर्षयुगलं न तत् कामं प्रत्यर्पयितुमधिकारो न विभवः । तथापि त्वद्भक्त्या त्रिदलयुतबिल्वावलिमिमां निधत्तेऽन्तेवासी पुरत इह बिल्वावलिकरः ॥

The three parts of this work upon Rāma's Later History form volumes 21 and 22 and 23 of the Harvard Oriental Series. The contents of the three parts are distributed as follows: Part 1: Introduction and English Translation. Part 2: Text in Sanskrit and Prākrit; Indexes of stanzas, metres, &c.; Parallel passages; Extracts from sources; Prākrit glossary. Part 3: Critical account of the critical apparatus; Notes, exegetic, critical, aesthetic, &c.; Variants; Appendixes. Part 1 was printed at the University Press, Oxford, England. Part 2 was printed at the Nirṇaya Sāgara Press of Bombay, India. A number of copies of the Text alone, for use in the colleges and universities of India, will be pub- lished at a nominal price. Applications may be made to Professor Belvalkar, Deccan College, Poona, India. The work upon Part 3 was far advanced at the time of Dr. Belvalkar's departure from Harvard in June, 1914, on his way home to India. His manuscript material for Part 3 and his books were dispatched by the steamship Fangturm, which is now (June, 1915) interned at a port of the Balearic Islands. That material included his collation- sheets, which give the readings of the various codices of this drama. On account of the great difficulty of getting those codices together again, it seems best to await the release of the steamship.

CONTENTS PAGE NOTE BY THE GENERAL EDITOR OF THIS SERIES . . . xiii A METHOD FOR CITING SANSKRIT DRAMAS. PROPOSED BY THE GENERAL EDITOR Method of citing the metrical parts of a play . . . . xvii Method of citing the prose parts of a play . . . . xvii Examples of the use of this method . . . . xvii Appeal to future editors to use this method . . . . xvii Importance of a good method of citation . . . . xviii Two aspects of the method: logical character, typographical expression . xviii Logically essential features of a good method . . . . xviii Canon 1: It should utilize the logical divisions of the text . . xix Canon 2: The numbering should start anew with each division . . xix Discord results from numbering stanzas in one sequence for entire play . xix Canon 3: The citations should be precise . . . . xix Canon 4: They should apply at once to text and translation and notes . xx Canon 5: And be such as to need little change for new editions . . xx The elements of the citations: There should be: xxi For verse and prose alike, a first element, an act-number . . . xxi For verse, a second element, a stanza-number . . . xxi And, if need be, a third element, namely, a letter (a, b, c, d) . . xxi For prose, a second element, the number of the last preceding stanza . xxii And always a third, an Arabic number, to show the subdivision . . xxii The subdivision of the prose passage should be the printed line . . xxiii Typographical devices for making this system usable . . . xxiv Additional device for making this method easy: corner-citations . . xxiv Minor typographic details . . . . . xxv Inconveniences resulting from lack of system or bad system . . xxv Lack of system, good or bad . . . . . xxv Citation by pages and multiple sequences of pagination . . . xxv Combinations of bad system and systemlessness . . . xxvi Examples from editions of the Mṛichchhakaṭika . . . xxvi b [H.O.S. 21]

Contents [x PAGE PREFACE . . . . . . . . . . . xxix INTRODUCTION . . . . . . . . . . xxxv Bhavabhūti, his life, and his date . . . . . . . . xxxv

1. Bhavabhūti's account of himself . . . . . . . xxxv
2. His family-home, Padmapura in the Berars . . . . . . xxxvi
3. Its identity with Padmāvatī (of the Mālatī-Mādhava) unproved . . xxxvi
4. Further biographical details . . . . . . . . xxxvii
5. Bhavabhūti's personality as inferred from his writings . . . xxxviii
6. His date: traditional story about Kālidāsa and Bhavabhūti . . xxxix
7. His works certainly influenced by Kālidāsa . . . . . xl
8. Bhavabhūti probably later than Bāṇa (about A. D. 610) . . . xli
9. Bh.'s identity with 'a pupil of Kumārila' (say 700–750) unproved . xli
10. Bhavabhūti certainly earlier than Rājaśekhara (about A. D. 900) . xlii
11. Date of Yaśovarman of Kanauj as determining that of Vākpatirāja . xliii
12. Date of Vākpatirāja's Gaüḍavaho not far from A. D. 736 . . . xliii
13. Bhavabhūti probably a little earlier than Vākpatirāja . . . xliv
14. Bhavabhūti probably flourished near the close of the seventh century . xlvi Bhavabhūti's works . . . . . . . . . . xlvi
15. Lost works of Bhavabhūti . . . . . . . . xlvi
16. Bhavabhūti's extant works . . . . . . . . xlvii Excursus: History of the Rāma-story . . . . . . . xlvii
17. Vālmīki's Rāmāyaṇa: its present form . . . . . . xlvii 18, A to 18, Q. Rāma-story as given in Vālmīki's Rāmāyaṇa . . xlviii
18. Vālmīki's Rāmāyaṇa not the sole source of our play . . . lv
19. Three forms of the Rāma-story in the Padma-Purāṇa . . . lv
20. The later history of Rāma as given by the Padma-Purāṇa . . lvii
21. Sources of the Rāmāyaṇa: historical element . . . . lviii
22. The Rāmāyaṇa a blending of history and allegory . . . . lix
23. Later modifications of the Rāma-story: 1. Buddhistic form . . lx
24. Later modifications of the Rāma-story: 2. Jain forms . . . lx 26, A to 26, F. Later modifications of the Rāma-story: 3. Hindu forms . lxi Bhavabhūti's early plays: Mahāvīra-charita, Mālatī-Mādhava . . lxiii
25. The Mahāvīra-charita . . . . . . . . lxiii
26. Bhavabhūti's departure from the Rāmāyaṇa . . . . lxiv
27. Technical significance of these changes . . . . . lxv
28. Merits of the Mahāvīra-charita . . . . . . . lxvi
29. Its defects . . . . . . . . . . lxvi

xi] Contents PAGE 32. The Mālatī-Mādhava . . . . . . . lxvii 33, A to 33, G. Analytic statement of its plot . . . . lxviii 34. Criticism of its plot . . . . . . lxxi 35. Technique of the Mālatī-Mādhava . . . . . lxxiii 36. Its advance over the Mahāvīra-charita . . . . lxxiii 37. Defects of the Mālatī-Mādhava . . . . . lxxv The present play : Uttara-Rāma-charita . . . . . lxxvi 38. Transition to the Uttara-Rāma-charita . . . . lxxvi 39. Special features of the Uttara-Rāma-charita . . . lxxvii 40. Technical perfection of the opening act . . . . lxxviii 41. The Uttara-Rāma-charita as a psychological study . . . lxxx 42. The Recognition-scene in the sixth act . . . . lxxxii 43. Other merits of the play . . . . . lxxxii 44. Defects of the play . . . . . . lxxxiv 45. Some common characteristics of the three plays . . . lxxxv Bhavabhūti as a poet . . . . . . lxxxvi 46. His shortcomings . . . . . . lxxxvi 47. His merits . . . . . . . lxxxvii 48. Conclusion . . . . . . . lxxxviii TRANSLATION OF THE PLAY UTTARA-RĀMA-CHARITA . . 1 Dramatis personae . . . . . . . 2 Main characters . . . . . . . 2 Interlocutors . . . . . . . 3 Synoptic analysis of the play . . . . . . 4 Prologue . . . . . . . . 4 Synoptic Analysis of Acts i to vii . . . . . 5 Translation of the play . . . . . . 15 Translation of Act i, The Picture-gallery . . . . 15 Translation of Act ii, Daṇḍakā revisited . . . . 31 Translation of Act iii, The Shadow . . . . . 41 Translation of Act iv, Kausalyā and Janaka . . . . 60 Translation of Act v, Princes in combat . . . . 73 Translation of Act vi, Princes recognized . . . . 82 Translation of Act vii, Reunion . . . . . 94

इस पृष्ठ पर कोई मुद्रित पाठ उपलब्ध नहीं है (यह पृष्ठ रिक्त है)।

NOTE BY THE GENERAL EDITOR OF THIS SERIES WITHIN the last decade, the West and the Far East have become virtually near neighbours. From the responsibilities of such neighbourhood there is no escape. We must have to do with the East, and as members of the world-family of nations we must treat the East aright. To treat the people of the East aright, we must respect them; and to respect them, we must know them. And in order that we may know them, competent scholars must study their history and make it accessible to the Occident. The business of the Orientalist is something that is in vital relation with the practical and political needs of to-day. It is a fact of happy augury that scholars of the East are joining hands with those of the West in the great work of helping each to understand the other. The work calls for just such co-operation, and above all things else for co-operation in a spirit of mutual sympathy and teachableness. There is much of great moment that America may learn, for example, from the history of the peoples of India, and much again that the Hindus may learn from us. But the lessons will indeed be of no avail, unless the spirit of arrogant self-sufficiency give way to the spirit of docility, and the spirit of unfriendly criticism to that of mutually helpful constructive effort. The relation of teacher and taught is here, in an eminent degree, a reciprocal one, for both East and West must be at once both teacher and taught. The whole spiritual and material background of the life of the East differs so completely from that of the West that neither can ever understand the other from a mere study of the other's literary monuments. Such study is indeed inexorably necessary, and it must be fortified by broad and rigorous training in the many-sided methods of to-day. But that is not enough. An Occidental who would faithfully interpret the East to the West must also know the life of the East from actual observation and experience, and must look at it from the Eastern angle of vision. Accordingly, for example, the Sanskrit professor of the next generation must have resided

Note by the General Editor [xiv] in India, have mixed (so far as possible) with its people, and have mastered at least one of the great modern vernaculars, such as Marāṭhī or Bengālī. And, on the other hand, since the Hindus themselves are already actively engaged in interpreting the East to the West, it is needful also that they visit us, not merely to learn our way of doing things, but also to look at life as we look at it, and thus to find out what things—such, let us say, as repose of spirit or the simple life—the West most needs to learn of the East. Sir William Jones and Henry Thomas Colebrooke, illustrious pioneers of Indic philology, are ideal examples of the spirit and methods that were and are and must ever remain exemplary. They went to India and learned of the Hindus, and to the work of making the East known to the West they gave, with heroic devotion, all that they had to give. And ever since their day, the business of the East India Company or of the Imperial Government has taken men to India who have proved to be not only faithful officials, but also Indianists of large achievement. With the spread of Sanskrit studies upon the Continent in the second third of the last century, there arose men who, like Eugène Burnouf or the lexicographers Böhtlingk and Roth, accomplished great things without ever visiting the Land of the Rose-apple. More recently, the ease and speed of ocean-travel have brought it about that the Indianists who have not visited India are few indeed. They feel that they must go; for thus, as so often in modern life, do greater opportunities make obligations greater. But it is not only for a better understanding of each other's hopes and aspirations that the Indianists of the two hemispheres must join forces. They must realize, far better than they now do, the importance of turning to account the modern methods of organization and business efficiency, and the modern progress of the graphic arts. The value of organization, and of combining the labours of isolated scholars for well-considered ends, is illustrated by the Kāvya-Mālā of Bombay and the Ānandāçrama Series of Poona. Moreover, as regards wide circulation and usefulness, complete works issued in such large groups or series as those, and in such form as only an adequate printing establishment can give them, have an enormous advantage over works issued singly or in incomplete parts, and at some obscure press, and in a small edition. Let me add that the work of eminent printers, such as the late Jāvaji Dādāji of Bombay, seems to me to be a very real service to science, and as such to deserve generous recognition from scholars.

xv] Note by the General Editor His Excellency the Prime Minister of Nepal sent lately to Oxford, to be copied photographically for the use of scholars, seventy manuscripts of the Nepal Durbar Library. Such enlightened co-operation between East and West deserves mention, not only by way of grateful acknowledgement, but also as an admirable example which—let us hope—may be followed by others. It suggests the establishment, perhaps at the Nirṇaya Sāgara Press in Bombay, of a central agency to which Hindu Librarians may safely entrust manuscripts for mechanical reproduction, without sending them to Europe. The requirements for such an establishment are by no means too hard to provide. They are (1) a safe place of storage, (2) the photographic apparatus, and (3) a competent operator.—See Postscript on the next page. It is with hope and with gladness that I lay before students of the history of the East the work of Doctor Belvalkar. I am glad that a Hindu, well versed in the learning of his native land, should think it worth while to learn of the West. And I hope that his residence in America may make his Eastern learning vastly more fruitful for us Occidentals than it ever could be, if he had not come hither to find out (as I said) what lessons from his country's past may best be taught to us. And I hope again that many in the coming years may follow his example, establishing thus most valuable relations of personal friendship and co-operation between Indianists of the Orient and the Occident. Aeschylus—as Athenaeus¹ reports—said that his dramas were fragments of the great banquet of Homer's table. And so of Sophocles: his Ajax, for example, is essentially the Ajax of the Iliad. In like manner, the material of the greatest Hindu dramatists, Bhāsa and Kālidāsa and Bhavabhūti, is in large measure taken from the Hindu epics.² Thus the play before us, Rāma's Later History, is, as appears from its title, a dramatization of the Rāmāyaṇa. Its author, Bhavabhūti, for knowledge ¹ viii. 39 . . . Αἰσχύλου, ὃς τὰς αὑτοῦ τραγῳδίας τεμάχη εἶναι ἔλεγε τῶν Ὁμήρου μεγάλων δείπνων. ² The Mahā-Bhārata itself, using the very same metaphor as Aeschylus, declares (i. 2. 389) that 'all the best poets live upon this story' of the Bhāratans. And it adds (385, 388) that 'the inspiration of the poets comes from this incomparable narrative', and that 'there is not a story in the world that does not go back to the Bhārata'. इदं कविवरैः सर्वैराख्यानमुपजीव्यते । इतिहासोत्तमादस्माज्जायन्ते कविबुद्धयः । अनाश्रित्यैतदाख्यानं कथा भुवि न विद्यते ॥

Note by the General Editor [xvi] of what was to him the classics, and for his use of that knowledge, may justly be compared with John Milton; and for the exquisite finish of his diction, we might call him the Thomas Gray of India. For the student of comparative literature, in particular of literary genetics and the history of the drama, this play has many aspects of interest which I hope the editor has not failed to put in a clear light. But it is for the student of the history of religions that the Rāma-legend is especially instructive: its hero has become a pattern, often quite divine, of all that is good and noble; Sītā is wifely fidelity incarnate; and the Rāmāyaṇa—translated from Sanskrit into the great vernaculars, Bengālī, Hindī, Marāṭhī, and so on, and especially the Rāmāyan of Tulsīdās—has become to untold millions literally a Bible to which they look for examples of virtue and for inspiration to righteous living. Our drama is accordingly one of the many forms into which the Indian genius has cast the story of its greatest popular hero, its deified hero; one of the many expressions of its loftiest ideals. May the knowledge of those ideals so enlarge our minds and our sympathies that, when we pass judgements on the East, we shall realize the truth of the Sanskrit proverb, ‘The whole world is one family’. C. R. Lanman. Harvard University, December, 1914. Postscript.—The prism-process was described by K. Krumbacher in an article entitled ‘Die Photographie im Dienste der Geisteswissenschaften’. This appeared in 1906 in the Neue Jahrbücher für das klassische Altertum, 17. 601–659, and was also published separately by Teubner, price marks 3.60. For further articles on the photographic reproduction of manuscripts, see The Nation, of New York, for March 5, 1908 (procédé Graffin), for March 19, 1908 (excellence of ‘good ordinary photographs’, and ease of making them), and for December 30, 1909 (apparatus of Carl von Arnhard of Munich). The Photostat, manufactured by the Eastman Kodak Co., is sold by the Commercial Camera Co. of Rochester, New York, for $500 or $750 or $1,100 according to the size. The Camera Co. issues a pamphlet entitled ‘The Photostat and its operation’. ‘The photostat is a simple and practical apparatus for the rapid photographing of printed or written docu- ments...directly upon the surface of sensitized paper with the image in correct position.... No previous knowledge of photography is necessary.’ For a great library or scientific society which is not too painfully hampered by lack of funds, and in which the apparatus will be much used, the photostat is unquestionably the best and in the long run the cheapest apparatus.

A METHOD FOR CITING SANSKRIT DRAMAS PROPOSED BY THE GENERAL EDITOR OF THIS SERIES Method of citing the metrical parts of a play.—These are cited by act, and by stanza as numbered from the beginning of that act, and by quarter or pāda, the pāda being indicated by a or b or c or d. Method of citing the prose parts of a play.—The prose lines between any two consecutive stanzas are numbered from the prior to the latter stanza, starting always anew from the prior stanza. A given line is cited by number of act and by number of prior stanza (counted as above) and by number of line as counted from said prior stanza.—If there is no prior stanza (that is, if the act begins with prose), a zero is put in place of the stanza number. Examples of the use of this method.—Thus, for this play, the Uttara- Rāma-charita, the citation iii. 2ᶜ means quarter or pāda 3 of stanza 2 of act iii, mohe mohe Rāmabhadrasya jīvam. Thus again iii. 2³ means prose-line 3 as counted from stanza 2 of act iii, Muralā—Katham iva. And iii. 0³ means prose-line 3 at the beginning of act iii, Muralā— Bhagavatī Tamase, &c. Appeal to future editors to use this method.—We might suppose a method of citation to be a matter wholly unworthy of the serious con- sideration of a great scholar, if we judge by the way in which such great scholars as Böhtlingk or Pischel have failed to consider it. And from the brevity of the foregoing three paragraphs, one might suppose the matter of a method of citation to be so simple as to call for no elaborate discussion. It is simple, but the simplicity is the outcome of a laborious consideration of complexities arising from bad methods or lack of method, which com- plexities are a constant annoyance to the scholar and a hindrance to his progress. The essentials of a good method are discussed on the following pages, and the need of a good method is made clear by showing the faults of some bad ones. A bad method puts the student to a deal of worse than useless trouble, and wastes his time, just as a good method saves it. A c [H.O.S. 21]

A Method for citing Sanskrit Dramas [xviii good method should be so good as to commend itself to general acceptance by scholars and future editors, since uniformity is in itself highly desirable. For these reasons I earnestly hope that this system will be widely accepted and followed by scholars of India and the Occident. Importance of a good method of citation.—‘A simple and sufficient means of citation is absolutely indispensable for any text of mingled verse and prose that is of consequence enough to be studied and cited at all. The editor who fails to provide such means is guilty of flagrant neglect of plain duty and of gross disregard for the time and convenience of students and of his colleagues. By way of punishment he may count upon the seriously circumscribed usefulness of his book and the silent maledictions of those who are forced to use it.’—This was my plea, in 1900, before the court of intending editors, on behalf of a good method of citation, and made in the preface of the translator of Konow’s Karpūra-mañjarī, Harvard Oriental Series, vol. 4, p. xvi. The system tried by me in 1900 proved unsatisfactory in one particular, the method of citing the prose-passages. The one now proposed is substantially the same as that of 1900, excepting the feature which concerns the prose-passages. Two aspects of the method : its logical character and its typographical **expression.—**A good method must be good in that it is logical and there- fore easy to understand and to remember, and good also in that the typographical devices for its expression are simple and are easily and quickly caught by the eye. These general requirements are so obvious as to need no argument. Nobody disputes them—chiefly (I fear) because nobody stops to think of them. Hence the need of bringing them to the attention of future editors. The excellences of a good method appeal in part to the mind and in part to the sense of sight, and the means for securing them will accordingly be treated under two heads (which, however, cannot be wholly separated in discussion), namely, The logically essential features of a good method, and (p. xxiv) Typographical devices for making this method usable. **Logically essential features of a good method.—**A good method should utilize the native divisions of the text so far as these are natural or logical or generally accepted.—To illustrate. For the English Bible we have the generally accepted divisions into books and chapters and verses. These are always accordant among the innumerable editions, while the pagination (unless by some strange chance) is always discordant. To cite the English Bible by pages, ignoring those divisions, would be futile folly. And it is no less wrong in principle to ignore the natural divisions of a Sanskrit play. Hence

xix] Proposed by the General Editor of this Series Canon 1. A good method should utilize the logical divisions of the text—that is, the divisions of the play into acts, and the divisions of the acts into stanzas and prose-passages. Canon 1 forbids us to cite the play by pages, since this ignores the division into acts, and results in endless discord as between different editions. Again, it forbids us to number either the speeches or the lines from the beginning of the act, since this ignores the division of the act into prose and verse, and results in the very undesirable possibility of citing the same verse in more than one way (by stanza-number or speech-number or line-number). Canon 2 (corollary to canon 1). The numbering should start anew with each division.—The numbering of stanzas should start afresh at the beginning of each act; and the numbering of prose-lines should start afresh from the last stanza preceding the prose-passage. Discord from numbering stanzas in one sequence for entire play.— The numbering of all stanzas in one sequence from beginning to end of the play has this result : that any differences in the number of stanzas in prior acts as between different editions become operative or even cumulative for any later act or acts. The serious disadvantages of this bad practice are strikingly illustrated by four European editions of the Śākuntala, namely Böhtlingk's of 1846, Williams's of 1876, Pischel's of 1877, and Cappeller's of 1909. In all these four editions act vii has thirty- five stanzas (no more and no less), and these stanzas (apart from mere variants) are identical. In Böhtlingk they are numbered from 160 to 194; in Williams, from 165 to 199; in Pischel, from 187 to 221; and in Cappeller, from 155 to 189. This is confusion worse confounded, and (with all deference to these honoured names) it is the more inexcusable because it is wholly needless. The discord results primarily from the fact that the numbering of stanzas goes in one sequence from the beginning of the play to the end, and secondarily from the fact that, for example, act iii contains twenty-four stanzas in one text (so Böhtlingk, Cappeller) and forty-one in another (so Pischel). This wholly wrong method of numeration continues the discord of act iii through all the remaining four acts. If only the numeration began anew with each act, the stanza-numbers of act vii would be wholly accordant in all four editions, and the same references would have been precisely correct for each and all. The discord in the other acts is on the whole so slight as to cause little trouble, and even that little is avoidable (see below, p. xx). Canon 3. The citations should be precise.—A citation which makes it necessary to search an entire page or long prose-passage is not precise. It should be so precise that you can instantly find the exact word intended.

A Method for citing Sanskrit Dramas [xx It should therefore consist of three elements, and, if possible, of not more than three. A citation consisting of two elements only (act-number and stanza-number) may be used to designate an entire stanza, but not to designate a prose-passage. Canon 4. They should apply at one and the same time to text and translation and notes.—The system should be such that the same citation will serve equally well for finding either a passage in the text or the corresponding passage of the translation or of an appurtenant native com- mentary or the corresponding exegetical or critical notes. Thus a given matter may be found by the same reference in any one of three or four different places. The method here proposed by me is usable at once in these three or four ways, and this is a very great convenience. Canon 5. The system should be such that the citations will need little change for new editions.—That is, it should be applicable to edition after edition of the same play with the very least possible change and inconvenience. This requirement also is met by the system here proposed. First as to the stanzas. If a stanza rejected from the text of an earlier editor is to be inserted in the text of a later editor, let the later editor insert it in its place and (without changing the number of all the subsequent stanzas) assign to it the number of the preceding stanza and distinguish the repeated number by a star. On the other hand, if a stanza inserted by an earlier editor is to be rejected by a later editor, then let the later editor leave out not only the stanza but also the number. For example, act ii of the Śākuntala has, in each of the four above- named editions, the same eighteen stanzas, all numbered with the utmost possible discord. Next after the stanza śamapradhāneshu (ii. 7) some MSS. give a stanza nirākṛita-, &c. Böhtlingk and Williams give this in their notes, while Pischel receives it into his text, and Cappeller rejects it. If the stanza-numbering had begun anew with each act, Pischel might have inserted nirākṛita- with the numbering ii. 7*; or, if Pischel’s edition had been the editio princeps and he had numbered nirākṛita- as ii. 8, Cappeller might have left the stanza out and have numbered śamapra- dhāneshu as ii. 7, and (skipping ii. 8) have numbered his next stanza as ii. 9. In this way, even in spite of the difference in the total number of stanzas, the numbering of them could very easily have been kept entirely harmonious. As to the prose-passages, this may be added. Although the prose- passages are occasionally long, they are for the most part short. As between different editions of the same play, the numbering of the prose- lines will not correspond exactly; but since the numbering starts anew with each new prose-passage, the discordances will never be cumulative and thus they will usually be so slight as to cause very little inconvenience.

xxi] Proposed by the General Editor of this Series At this point a word in defence of Mr. Belvalkar's procedure in giving, immediately after a Prākrit passage, its explanation in Sanskrit (or chhāyā), set off by double verticals and by a difference of type. It is true that this increases the length of the prose-passages, and consequently the possible discord in the line-numbers as between one edition and another. But this disadvantage is seldom considerable, and it is more than offset by the convenience of always having the explanation just where it is needed, instead of having to search for it at the bottom of the page or the end of the volume. Moreover the breaking up of one printed page into two or three or four typographically varying parts for text and chhāyā and variants and comments is very unsightly. The elements of the citations.—These may now be discussed in systematic order. It is only the second and third elements of the citations for the prose-passages that occasion doubts or difficulties. For metrical and prose-passages alike, the first element of the citation should of course be the act-number, as canon 1 requires. For metrical passages, the second element should of course be the stanza-number as counted from the beginning of the act. The stanzas are short, and of a length (two or four printed lines) which is uniform as compared with that of the prose-passages, so that stanzas are precisely indicated and easily found by a single number. For metrical passages, the third element should indicate the natural subdivisions of the stanza. A Sanskrit stanza consists of four quarters or pādas. A quarter of a stanza may therefore be indicated by a letter (a, b, c, d). If an entire stanza is intended, the third element, the letter, should be omitted. In this edition, anushṭubh and āryā stanzas are printed as two lines, each line consisting of two undivided pādas; and in referring to such a line, the letters ab are used together without any comma¹ between them, or else cd. Metrical passages containing interjected prose.—Prose-lines some- times intervene between parts of a stanza. This complication is met in a manner so logical and natural as hardly to need description: namely, the stanza-number and pāda-letter together are treated as one element of the citation, that is, as the second element, and the letter is not set as a 'superior'. Thus between iv. 24ᵇ and iv. 24ᶜ of Uttara-Rāma-charita a prose-passage of two lines and a complete stanza (iv. 25) intervene. The prose-lines are cited as iv. 24 b¹ and iv. 24 b². The prose-lines fol- lowing iv. 24ᵈ are cited as iv. 24 d¹, iv. 24 d², and so on (although iv. 25 precedes them !). ¹ If there is a comma (e. g. ii. 25ᵃ, ᵇ), the ferred to are not printed in the same punctuation indicates that the pādas re- line (are not anushṭubh or āryā).

A Method for citing Sanskrit Dramas [xxii] It even happens that a bit of prose intervenes¹ between two fractions of a pāda: so between the first and second words of pāda d of Mahāvīra- charita vi. 7 (durgo, &c.). Here the line of prose may be cited as vi. 7ᵈ¹. For prose-passages, the second element of the citation should be the number of the stanza immediately preceding.—This solution of the matter is not obvious, and that is the reason why different editors use so per- plexingly different methods.—The indication of the prose-passage ought to be logically parallel with the indication of the stanza. This rule requires that each prose-passage, like each stanza, be regarded as a unit and be numbered from the beginning of the act. Editors have not usually treated each prose-passage as a unit,—apparently for the entirely invalid reason that a prose-passage unit often consists of two or more speech-units, and perhaps also because they did not see how to indicate typographically the numbering of units so complex and of such irregular length. Act i of the Śākuntala begins with a stanza. This is followed by a prose-passage unit consisting of several speeches; but there is no difficulty in citing them all as ‘act i, prose-passage 1’. And since each one of stanzas 1–9 of act i is followed by a prose-passage, it is evident that each of these nine prose-passages may be cited by the number of the stanza immediately preceding it. The like is true of the remaining twenty-three prose-passages which follow stanzas 11–33 (st. 33 is the last of act. i in Pischel’s posthumous ed.). There is no prose-passage immediately following stanza 10. But it is not necessary on this account to give to the twenty-three prose-passages which follow stanzas 11–33 the numbering 10–32 just for the sake of a strict arithmetical sequence, since this produces a discord through most of the act. It is simpler to number them as 11–33, that is, to skip the number 10. In this way there is no need of any special or separate typographical indication of the number of a prose-passage. In short, it suffices to give to any prose- passage the number of the stanza immediately preceding. For prose-passages, the citation should always contain a third element, an Arabic numeral, that is, an element which will distinguish a prose-citation from a stanza-citation. A stanza-citation consists of an act-numeral (Roman) and stanza-numeral (Arabic), followed, if necessary by a letter (a, b, c, or d) to indicate the quarter or pāda. The distinguishing element of a prose-citation should therefore be an Arabic number (1, 2, 3, 4, 5, &c.). Thus i. 5 means act i, stanza 5; and i. 5ᵈ means act i, stanza 5, ¹ Since the citations of the system here pro- bare name of a dramatis persona) from a posed are intended to show on their face line which contains a verse-portion. Thus whether they refer to prose or to verse, the stage-direction after iii. 39 of this play it is necessary that an editor should should not be printed in the same line rigorously exclude all prose (except the with iii. 39ᵈ.

xxiii] Proposed by the General Editor of this Series

pāda 4. But i. 5⁴ means act i, prose-passage immediately following stanza 5, subdivision 4.—It will be observed that this method has the incidental advantage (a very considerable one) of showing on its face whether a line of poetry or a line of prose is intended.—The prose-passages vary in length from a single word to several pages. Examples of long prose-passages (forty to sixty lines) in the Śākuntala are those following i. 24, vi. 1, vi. 24, vii. 20. Precision of reference (and not alone the need of distinguishing prose-citations from stanza-citations) requires a third element indicating the subdivision.—What then shall this subdivision of the prose-passage be? The subdivision of the prose-passage should be the printed line.— The natural subdivisions of a prose-passage are the single speeches. But these are very uneven in length, and the numbers employed to indicate them would come at such irregular intervals and places on the page as not to be readily caught by the eye (a grave disadvantage). Moreover, a single speech is sometimes so long that a single number would not show precisely enough where a word or phrase is. In the case of prose-passages, therefore, it is best to use the printed line as the basis of subdivision, and to place the line-numbers at brief and regular intervals, and always in the same relative position on the page, that is, in the left-hand margin, and never in the right-hand margin. The numbering of the lines of the prose-passage must of course start anew with each new passage (canon 2, p. xix). The numbering of the lines or of the speeches from the beginning of the act is ruled out as wholly illogical and as a source of confusion (canon 1).¹ In Konow's Karpūra-mañjarī, Cambridge, 1901, the numbers of the act, and of the stanza as counted from the beginning of the act, are clearly shown. For the prose between two consecutive stanzas, I have numbered each clause (not speech and not line) of prose, counting from the prior stanza. The disadvantage of this method is that the numbers have to be put at irregular intervals and in irregular positions, and are therefore not easily caught by the eye. The disadvantage is so serious as to condemn this my former method.

¹ The Mālavikāgnimitra of Mr. S. P. Pandit, Bombay, 1889, numbers the stanzas from the beginning of the act. But it also numbers the speeches from the beginning of the act, without distinction of prose and verse or of Sanskrit and Prākrit. Mr. Pandit makes a great gain in sim- plicity by using the same numbers (many, of course, are skipped) for reference to his chhāyā and his variants on the lower part of the page and to his notes at the end of the volume. His speech-numbers are very easily caught by the eye because they are always in the same relative position, a free vertical column on the left-hand edge of the page, no matter whether it is a right-hand page or a left- hand page. So far as convenience goes, I have long deemed this to be the best edition of its time among Sanskrit plays. Only the speeches are sometimes too long for a precise citation, and the numbering of speeches from the beginning of the act is wrong in principle.

A Method for citing Sanskrit Dramas [xxiv Typographical devices for making the present system usable.—No matter how easy the system is for the mind, it will be a failure unless it is easy also for the eyes. That is, the system must not only be well thought out, but there must also be adequate typographic devices for its expres- sion.—Device 1. The number of the act should be clearly stated in the headline of the right-hand page.—Device 2. The number of the stanza as counted from the beginning of the act should be given at the end of each stanza.—Device 3. The number of every third prose-line as counted from the last preceding stanza should be shown by a figure close to the beginning (not the end) of that line, but nevertheless standing free in the left-hand margin, and thus very readily caught by the eye. The number should be placed exactly opposite the middle of the type-line. The numbering by threes is far more convenient than the numbering by fives. The reason why the marginal number should be on the left and not on the right is that many prose-lines contain only a word or two, so that the eye has to travel a long distance over blank paper to find the number. Any one may convince himself that this is a real inconvenience by examining the line- numbers of a wide-paged book (for example, p. 70 of Hertel’s Tantrākhyā- yika, Berlin, 1910). If the marginal numbers for prose-lines in all editions were uniformly put on the left-hand side, the gain of convenience would be very great. Any one may assure himself of this by looking at Mr. S. P. Pandit’s Mālavikāgnimitra, where the speech-numbers are thus given (see above, p. xxiii, note). The eye catches them with wonderful ease. Additional device for making this method easy : corner-citations.— In the outer upper corner of the left-hand page is given the citation of the line with which that page begins; and in the outer upper corner of the right-hand page is given the citation of the line with which that page ends. A dash and square bracket are placed after the left-hand citation, and a square bracket and dash are placed before the right-hand citation, and the two citations are intended to be read together. Thus the two corner-citations vii. 31⁵—] [—vii. 35ᵇ are to be read as follows: ‘From vii. 31⁵ to vii. 35ᵇ.’ The purpose of the corner-citations is to show just what parts of the text are to be found included on any two pages which face each other at an ‘opening’ of the book (beim Aufschlagen des Buches, à l'ouverture du livre). The object of putting the citations in the outer corners is to enable the reader, when ‘thumbing’ the volume, to find any passage with the least possible opening of the leaves, and therefore with the minimum of time and trouble. If they were put, as they often are, on the inner corners, the reader would have to open the leaves almost completely in order to make use of the corner-

xxv] Proposed by the General Editor of this Series citations.—The page-numbers are relegated to the inconvenient inner corners in order to discourage as far as possible the employment of them for citation. (The page-numbers have their use for the printer and book- binder.) Minor typographic details.—As to the corner-citations: The numerals and letters and dashes should be set close together, without spacing, to the left or right of the single square bracket. The act-numerals should be Roman numerals (not Arabic), and lower-case (not capitals). The stanza- numbers should be Arabic. The letters which indicate the verse of the stanza should be ‘superiors’ (that is, be set higher than the rest of the type-line) and Roman (not Italic). The same holds for the figures which indicate the prose-line. The corner-citations should be set thus: vii.31⁵—] [—vii.35ᵇ They should not be set as follows: not thus, [—vii. 35. b; nor thus, [—VII. 35. b; nor thus, [—VII. 35.ᵇ I have no quarrel with one who insists on using Arabic figures for the number of a maṇḍala of the Rigveda, of a parvan of the Mahā-Bhārata, or of a kāṇḍa of the Rāmāyaṇa. But for an act of a play, the Roman numeral has undoubted advantages. And if the Roman numeral is used, let it be in lower-case, not in capitals. As eminent printers have observed (so Bänsch-Drugulin, De Vinne), the use of many capitals on a page ‘stört die Ruhe des Satzbildes’, ‘spots the page’, gives it an unrestful look. For punctuating the elements of one citation, moreover, periods are much better than commas: thus, iii. 1¹, iv. 2ᶜ, v. 6¹, vii. 35ᵇ. In a mass of references, the overworked commas give the line a serrated or very ragged effect: thus, III, 1, 1; IV, 2, c; V, 6, 1; VII, 35, b. The contrast between the serrated style and the restful style is so great as to make comment needless. Inconveniences resulting from lack of system or from a bad system.— These might well be adduced by way of comment on the foregoing principles and canons and devices, or to serve as arguments therefor. Their relevance is clear. Lack of system, good or bad.—As an example may be cited the old edition of Uttara-Rāma-charita, of Calcutta, 1831. This gives the number of the page, but not the line of the page, nor any stanza-numbering (good or bad), nor even the number of the act except at the beginning and end of each act. Since the edition is long out of print and the page-numbers of later editions have no determinable relation with the page-numbers of the old one, all references made to the old one by writers on the drama and by lexicographers and grammarians and others have of course become worthless. Citation by pages and multiple sequences of pagination.—An extreme d [H.O.S. 21]

example of the futility of citing by pages is seen in Ghāṭe's edition of this play (Nagpur, 1895). Here the pagination begins anew in one and the same volume no less than seven times: namely, for preface (1–2), intro- duction (1–12), text (1–192), translation (1–60), notes (1–51), glossary (1–8), and corrections (1–3), to say nothing of two sequences without pagination. Thus for example there are (if we disregard the differences of Arabic, Roman, and Nāgarī numbers) two pages for every page numbered ‘60’ or less, three pages numbered ‘51’, four pages numbered ‘12’, five pages numbered ‘8’, and seven pages numbered ‘2’. In Kāḷe's Śākuntala, Bombay, 1898, there are no less than ten different sequences of pages! two sequences are without pagination, and eight sequences begin with page 1. Multiple sequences of pagination are very frequently found in books printed in India, and I call upon Hindu editors to recognize the inconvenience thereby occasioned, and to exclude such sequences from the works which they edit. Combinations of bad system and systemlessness.—These need hardly be described except to show the bewildering perplexities that result. Of these an example or two. The Mṛichchhakaṭika of Calcutta, 1829, gives in the headlines the name of the play (343 times repeated, but with no indication of the act!) and the page-numbers. There is no numbering of the stanzas and none of the prose-lines. Stenzler's edition of Bonn, 1847, gives in the headlines the number of the act and the page-numbers, and every fifth page-line is indicated by a marginal number. Lassen's Prākrit grammar refers to the Calcutta edition. In order that students may use Lassen's book with Stenzler's text, Stenzler puts the Calcutta page-numbers in his margins. The general effect is confusing, and the confusion is aggravated by the numbers at the end of every speech (not line) in Prākrit (not Sanskrit) counted from the top of the page! Böhtlingk's translation of this play gives no help for finding a desired passage except the number of the act at the beginning of each act. In order to use it with reasonable facility, I had to write the number of the act on the outer corner of the uneven pages, and the number of the stanza in the margin. Goḍabole's Mṛichchhakaṭika of Bombay, 1896, gives the act-number in the headlines. It numbers the stanzas from the beginning of each act, and gives these numbers between double verticals at the end of each stanza. But it also gives a number between European parentheses at the end of each speech in Prākrit (not Sanskrit) for reference to the explanation at the foot of the page. This double numbering is, to use an English nyāya, like having a big hole for the big cat and a little hole for the kitten. Moreover, it gives a marginal numbering of the lines counted (without distinction of prose and verse or of Prākrit and Sanskrit) from