INTRODUCTION XXV is an improvement on the existing usage (pratisamskrtaih). In the fourth stage these selected words are classified according to subjects, Varga-s, such as those dealing with celestial objects, etc. And finally the words are listed along with genders as completely as possible (sampurnam ucyate). Although Amara himself is quite definite in the division of the kośa into three Kända-s and the various Varga-s, some scholars, notably in Germany, have further subdivided ¹ the Varga-s into Kiṭaprakaraṇam (p. 337), Aśvaprakaraṇam (p.515), Pāśupālyam (p. 588), etc. In the third Kanda alone there is a difference of opinion; the commentator Svāmin consi- dered the section on homonymous avyaya-s as part of the follow- ing Avyaya section, whereas other commentators have generally added this to the Nänärtha section. The commentator Bommaganti Appayārya has discussed this question at length in the beginning of the Nānārtha Avyaya words. Setting forth arguments on both sides he has come to the conclusion that the avyaya-s are separate words and both the sections of the avyaya, homonymous and synonymous, should go together and that the Nãnärtha section promised at the beginning of the third Kāṇḍa by Amara begins with känta words and ends with hänta words. The nanärtha avyaya-s are also arranged by their finals, although not shown as such in any MS. or edition. Purușottama expressly mentions this in his varnanir- deśana and it has been discussed by Ramānātha and others.2 But as the beginning of the homonymous avyaya-s is avyayāḥ pare (III. 3. 239) and as it forms the latter part of the line in the verse dealing with nānartha, it is hardly likely that it forms a separate Varga. In the first Kanda alone there is a serious difference of opinion. The commentator Rãyamukuța considers that the Svargavarga extends up to the Natyavarga; the commentaries 1 These sub-sections or Prakaraṇa-s are mentioned by Sūrin. 2 Ak., Borooah's ed., p. iv.
