of the sun at Daçapura (now Mandasor). A detailed examination of this inscription not only leads to the conclusion that in the fifth century a rich Kāvya literature must have existed, but in particular shows that the poem has several affinities with Kālidāsa's writings. The latter fact renders it probable that Vatsabhaṭṭi, a man of inferior poetic talent, who professes to have produced his work with effort, knew and utilised the poems of Kālidāsa. The reign of Chandragupta Vikramāditya II., at the beginning of the fifth century A.D., therefore seems in the meantime the most probable approximate date for India's greatest poet.
Besides the epigraphic evidence of the Gupta period, we have two important literary prose inscriptions of considerable length, one from Girnār and the other from Nāsik, both belonging to the second century A.D. They show that even then there existed a prose Kāvya style which, in general character and in many details, resembled that of the classical tales and romances. For they not only employ long and frequent compounds, but also the ornaments of alliteration and various kinds of simile and metaphor. Their use of poetical figures is, however, much less frequent and elaborate, occasionally not going beyond the simplicity of the popular epic. They are altogether less artificial than the prose parts of Harisheṇa's Kāvya, and à fortiori than the works of Daṇḍin, Subandhu, and Bāṇa. From the Girnār inscription it appears that its author must have been acquainted with a theory of poetics, that metrical Kāvyas conforming to the rules of the Vidarbha style were composed in his day, and that poetry of this kind was cultivated at the courts of princes then as in later times. It cannot be supposed that Kāvya literature was a new inven-
