have to) do without it I try to place paragraph breaks where divisions of time or scene occur in the narrative This is really not too much of a problem, for the text does sort itself into reasonable-sized chunks with markers like "us waqt" to introduce new action. As for sentences, they can quite legitimately be organized in ways that sound good in English; the difference between a semicolon and a period, since it often can't be deduced from the text, must be based on style, on the balance and feel of the sentences in relation to each other And when to put an exclamation point"? A surprisingly important decision, since the presence or absence of one changes the whole feel of an utterance The dastan-gos, who loved sudden drastic shocks, would certainly have used exclamation points if they had had them to work with. This again comes down, it seems to me, to English style- there should not be too many exclamation points, but neither should there be none at all, and they should be placed where they do the best work. In making choices like these the translator must and should rely on his own sense of styte. These choices, once made, can perhaps be illustrated more easily than they can be analyzed My sample text is a portion of volume IV of Tilism-e Hoshruba, by Muhammad Husain Jan,3 from the forty-six-volume Dastan-e AmTr Hamza published by Newal Kishore during the last two decades of the nineteenth century and the first decade of the twentieth. This passage was chosen for its liveliness, rawanT, and imaginative elegance. Paragraphs in the translation have been numbered for easy reference, and their boundaries indicated by corresponding numbers in the original text. It can easily be seen that I have aimed at something approaching Verner Elwin's simplicity. In a very few cases I have "improved" and clarified the text, mostly by replacing pronouns with proper names where the reference might be unclear, and by correcting what seem to be minor errors (the text places "gao-sawar" in Kaukab's army in paragraph [7], though it later becomes clear that these are really "nil- gao-sawar", and the number of bits of paper Kaukab throws can only be made to come out right by some fancy verbal footwork). I have also eliminated a number of introductory aurs, to make for cnsper sentence divisions. But I have tried not to do anything on a larger scale, preferring to let the text choose its own devices and work in its own way to hold the reader's attention. (In particular, I was tempted to edit the repetitive magic combats of paragraphs [9] and [10], but I resisted.) There are two specific points which I'd like to discuss with readers of the Annual. The first concerns names and epithets: when does one translate them, and when does one not9 This always seems to come down to a case by case, ad hoc decision. It is not tempting to call Kaukab "Star" or "Constellation," yet it seems right to render ""rausan-zamTr"" as "Radiant." I have always felt that "tilism" is 3 Muhammad Husain Jah, Tilism-e Hoshruba, vol IV (Lucknow Newal Kishore Press, AH 1306 [1888-89]) The edition I used was printed in 1927, and the excerpt is drawn from pp 1236-1240 I thank Shamsur Rahman Faruqi for suggesting this passage, and for his invaluable advice and comments on the translation Annual of Urdu Studies, #6 56