[ix]
according to the rules, in order to make more tangible the difference between homonymous determinatives and possessives. 1
What form should be given to the headings of articles is often a question. For denominative verbs, I have chosen the stem-form (e.g., mantraya rather than mantray) ; to this is prefixed the root-sign (V), merely in order to catch the eye ; the sign must not be understood as meaning that such stems are in any wise co-ordinate with roots. The stems of the perfect active participle and of the primarj- comparatives are given as ending in vans and yans, but without any implication that these are theoretically better than the weaker forms. Roots with medial or final ar or r are given in the latter form, and so are the stems in ar or r. 2
The synopses of conjugational forms which follow each verbal root are based on the collections, still in manuscript, made by Professor Whitney. 8 These were placed at my disposal by him with the greatest kindness. They include all the verb-forms cited by the St. Petersburg Lexicon and Boeht- lingk's Abridgment, as occurring in the actual literature, besides very extensive gleanings made independently by Professor Whitney from texts represented in the Lexicon, 4 and from others published since its completion, 5 or even not yet published. 6 In the description and classification of the forms, I have followed Whitney. It often happens that there are several forms in actual use for the same tense ; in such cases, the commonest one has been given, or else the one prescribed by the Hindu Root-book (dhatu-patjia) , or sometimes more than one form. Although in the case of many roots the aorist is confined almost exclusively to the Vedic language, I have neverthe- less given the aorist in such cases in order to fill out the conjugational scheme, since this seemed desirable from a pedagogical point of view. For pedagogical reasons, also, the secondary conjugations have been for the most part omitted. Many roots which are conjugated regularly in only one voice show forms of the other voice in the Epos, especially where the metre demands them. It is very difficult to say just how far such forms should be included, and my course in accepting or rejecting them has been, I fear, not wholly consistent.
The Notes, which form the third part of this work, will be issued as soon as is practicable. It is designed that they shall be as brief as possible, but shall render ample assistance in the interpretation of difficult passages and
1 Compare bhimaparakrama and bhima- 4 For example, the (Jatapatha and Aita- parakrama, p. 206. reya Brahmanas.
2 See Whitney, §§ 107 and 108, and com- 6 Especially Garbe's edition of Apastam- pare § 370. ba's Qrauta Sutra, and von Schroeder's
8 See Proceedings of the American Oriental Maitrayani. Society for May, 1882, p. xiii. 6 The Jaiminiya Brahmana.
