176 enjoyment." Each work of literature will theoretically have one flavor dominant throughout; and there are rules stating which ancillary flavors may be used in a work dominated by each particular flavor without disrupting the total effect. The word rasa is used in a conceit at the end of stanza 24, and I owe the ingenious translation to Michael Coulson. Poetic theory also recognizes the use of "ornaments," corresponding to our "figures." Ornaments can be of sense or of sound, and include simile, metaphor, zeugma, etc. In early theory rasa was merely one such ornament. According to another theory, which came to be held in conjunction with the rasa theory, true kavya is literature in which that meaning which is suggested not directly expressed, is the more important. I made this translation while studying Sanskrit literature under Professor Daniel H. H. Ingalls at Harvard in 1962-63, when there were hardly any readable translations of kavya in print. At that time Professor Ingalls was putting the finishing touches to his monumental work of translation and exposition, An Anthology of Sanskrit Court Poetry,^- which came out in 1965. Since then Dr. Barbara Stoler Miller has published her translation of Bhartrhar^s Centuries,^ and Professor John Brough his Poems from the Sanskrit.^ 1 can, therefore, refer the reader who is further interested in the theory and practice of kavya to these excellent books; in particular, Professor IngallsT introductory essay, "Sanskrit Poetry and Sanskrit Poetics" (pp. 2-29), makes it unnecessary for me to say more on this topic. Professor Ingalls and Dr. Miller present only the single-stanza genre of kavya; Professor Brough, too, concentrates on isolated verses, though he also gives longer extracts from a play and from a major poem in many cantos (mahakavya). The poem here translated is a specimen of the medium-length poem, or khanda-kWya^ of which Kalidasa^ Meghaduta (cloud messenger) is the most famous example. Unfortunately no authoritative text of this poem exists^ and I have had to edit my own. The poem has been published in three main recensions, but between these the choice was simple, for only one has just the fifty stanzas of the title, and these all beginning with the same word, plainly an original feature. However, even of this recension not all editions have the same fifty stanzas. Forty-eight of my stanzas are from one text which seemed to me the best; I have emended it where necessary. My stanzas 12 and 40 are from another text. Stanza 40 in the best text is corrupt beyond intelligibility or redemption; stanza 44 is hardly better; but my translation, numbered 44, combines all intelligible content from both main extant versions. My stanza 12 is introduced for another reason: stanzas 12 and 13 of the best text are so empty and repetitious, partly of notions difficult to render tasteful to a European, that I have combined them in my stanza 13. This is the only place where I have allowed myself such a liberty, ^Editor's note: Barbara Stoler Millers critical edition of the text, in two recensions, was published since the writing of this article, in April, 197! (New York: Columbia University Press).