THERE is general agreement among students of the art and epigraphy of the south-east Deccan regarding the chronology of the Middle and Late Phases at Amaravati. This is based on the well-known inscription of an official of Siri Sivamaka Sada carved on a Middle Phase fragment of the coping of the rail of the Great Stipa. Though first reported by Robert Sewell some little distance from the site of the Stupa, this fragment does indicate that the mature style of the Middle Phase had already been achieved by the reign of the Sdtavdhana Siva Sri (about A.D. 159 to 166) or Sivaskanda Satakarni (about A.D. 167 to 174), whichever identification is preferred. (Those who place the later Satavahanas two decades earlier than the dates accepted in this paper may make the necessary adjustments.) On the date and duration of the Early Phase, however, there seems to be no agreement whatever. The First Period of C. Sivaramamurti lasts from about 200 to 100B.C.: a view based on stylistic comparisons with other sites in the north-west Deccan and north India and on epigraphy. I have placed roughly the same sculptures in my Early Phase, and, also arguing from style and epigraphy, have dated them to the second quarter of the second century A.D. and a little earlier. Philippe Stern and Mireille Benisti, arguing from style and iconography and insisting that their chronology is relative merely, seem to occupy an intermediate position. In his first treatment of the problem Stern1 provisionally placed his Jaggayyapeta Style (my Early Phase) together with Kdrla, Junnar, Kondane, and the early paintings in Caves IX and X at Ajanta, between the toranas of Sanchi I and III and the 'Mathura Style'. Of Jaggayyapeta he said that some may think it contemporary with Bharhut but 'Son aspect fruste ne permet donc pas de conclure a une relative anciennete'. With the Pompeii Ivory in mind, he also permitted himself the cautious statement that the Sanchi toranas were not very much earlier than A.D. 79. This view, not very different from my own, was radically modified in the French scholars' book on Amaravati (Stern and Btnisti). There, certain pieces of the First Period (the Jaggayyapeta Style of the earlier book) were related to Bharhut and early Sanchi (their Plates I-III), while others 'un peu plus evoluees' were related to the toranas of Sanchi I and III and the 'Mathura Style' (their Plates IV-X). Others (their Plates XI-XV), the end of their First Period and the beginning of my Middle Phase, were later again. Thus while most of their First Period pieces were advanced to a point where my absolute dating was easily accommodated, some (their Plates I-III and related pieces) were taken back to the period of Bh.rhut and early Sanchi, that is, in the opinion of most scholars, to about 100 B.C., if not earlier.
- 1. J. Hackin et al. Nouvelles recherches archlologiques a Begram Paris, 1954. 2