Systems of Sanskrit Grammar § 44-1 speech.' The Chandra terminology with slight changes is that of Panini. The mode of presenting the subject is also artificial, after the fashion of Panini. The gra- mar goes by the nickname of ar, perhaps because the s are not here treated of separately, but probably be- cause wherever in his sutra Pāṇini has used the word #ar Chandragomin uses the word ' 60 45. Accessory treatises of the Chandra grammar.-In addi-- tion to the stras in six adhyayas Chandragomin has put together an Unādi list in three parts, a Dhătupātha in ten sections (both published by Dr. Liebich), as also Lingakā- rikäs or Liñganuśāsana, Ganapatha, Upasargavritti, and Varna sūtras. The Upädis differ from those belonging to the Paniniya school principally in their mode of present- ation, the suffixes being here arranged according to their final letter. In a few cases Chandra also derives the words differently. The Dhātupatha, as we saw (p. 52, above), is referred to by Kshirasvamin and was subsequently in- corporated in the Kätantra system. The Liñganusāsana is referred to by Vāmanācharya, Ujjaladatta, and Raya- mukuta (see above, p. 53). As to the Ganapatha no separate work of the kind has yet been discovered, but we must assume the existence of such a work as we find it. embodied in the sutra-vritti, just as the Kāśikā has done with regard to the Pauiniya Ganapatha. The Upasarga- vritti is found in Tibetan version only, and explains the meaning and use of about twenty upasargas. Finally, the Varpasūtra (Ms. no. 289 of 1875-76 in the Deccan College collection) is a very short treatise correspond- ing to the Paniniya Siksha and gives in about 40 sutras 1 Compare Chandra sutras i, 2,30, i. 3. 77, ii. 2. 14, &o, with Panini's iii. 2. 46, iii. 3. 174, ii. 1. 21 &c. A few cases 2 I tako this occasion to publish do occur, however, where the work entire on the basie Chandra permits the use of the word e. g. Chandra i. 1. 128 Panini iii. 1. 112.
