By ROBERT KNOX. London: BRITISH MUSEUM PRESS, 1992. Pp. 247, 21 
            figures, 133 catalogue illustrations. [pounds]40. 
            The two Buddhist sites from Andhra Pradesh, Amaravati and 
            Nagarjunskonda, have justifiably held the attention of scholars-from 
            the earliest attempts to quantify, judge, and understand Indian art 
            in the eighteenth century up until today. The two books are welcome 
            additions to this research. While neither changes radically our 
            basic views of the art and architecture of the two early Buddhist 
            sites, they contribute many nuanced shifts in, and suggest different 
            approaches to, the extensive stone sculpture from these sites. As 
            always with Indian sculpture, there is a need to set up a 
            chronological schema, and both books spend considerable time doing 
            it. In addition, looking at the two books together helps to clarify 
            the interrelationship between the art of the two sites. 
            Amaravati is the earlier of the two. Knox's book is a catalogue of 
            sculpture from Amaravati in the British Museum, intended, according 
            to Jessica Rawson's brief preface, as a "general introduction" and 
            "descriptive catalogue" to the Museum's magnificent collection, 
            which had been newly installed in the Museum galleries along with 
            publication of the book in 1992. The catalogue illustrates each of 
            the 133 sculptures in excellent photographs, many in color and 
            covering entire pages. 
            There are four chapters of introduction before the catalogue itself. 
            These chapters give a history of the site; tell of how the Museum 
            acquired the collection; describe the great Stupa at the site from 
            which much of the sculpture comes; and, finally, give a system for 
            dating and typology of the sculpture. The typology, which consists 
            of categories of sculpted architectural elements (pillars, 
            crossbars, copings, and so forth), is used to organize the sculpture 
            in the catalogue, which is arranged more or less chronologically 
            within each type. Each object in the catalogue is carefully 
            described, the most helpful feature, I think, of the book. These 
            descriptions force the reader to look carefully at the often very 
            complex sculptures, pointing out details and characteristics that 
            would otherwise most likely be missed. 
            Looking at the introductory chapters, Knox ties the form the Great 
            Stupa finally took to the coming of the Satavahana rulers and the 
            period of prosperity that they brought. The dating of the Satavahana 
            kings, however, has long been debated - and for Amaravati they could 
            have begun ruling either in the first or the second century A.D., 
            probably, Knox feels, later - around A.D. 130. The building more or 
            less ceased with their departure at the end of the third century, 
            with activities shifting to nearby Nagarjunakonda under the Iksvaku 
            rulers (the topic of Stone's book). But there was a Buddhist 
            monument at Amaravati from before the Satavahanas' coming, going 
            back to Mauryan times in the third century B.C. Knox thus rejects 
            Douglas Barrett's short chronology for the Amaravati sculpture, 
            which argued for it beginning only in the first century A.D.(1) 
            I think most scholars today would agree with Knox, but it is not 
            clear what was at the site so early. There are massive polished 
            granite pillars (one is 2.63 m tall) cut to accommodate crossbars 
            (Knox illustrates one in a photograph of ca. 1880), which appear to 
            be part of a large vedika that can apparently be dated to the third 
            century B.C. Some recently excavated fence pieces have, according to 
            I. K. Sarma, Asokan period inscriptions.(2) Knox does not say it, 
            but this would make it the earliest stone stupa fence of any size in 
            all of India, earlier by some 150 years than those of north India, 
            such as at Bharhut and Sanci, which appear to have replaced wooden 
            fences. In addition, that the pieces were polished is of great 
            importance if the polish indicates not only a Mauryan date, but also 
            Mauryan royal patronage. Add to this the find of Brahmi Prakrit 
            inscriptions on potsherds from Amaravati which relate to similar 
            finds from Sri Lanka that date by radiocarbon to 450-350 B.C.,(3) 
            and thus one or two centuries earlier than any previously known 
            South Asian inscriptions (and thus writing), and we are forced to 
            reassess the importance of South India and Sri Lanka to early South 
            Asian civilization and religion. 
            The Great Stupa at Amaravati itself was already largely destroyed 
            when Colonel Mackenzie, an Englishman, visited it in 1797. By then 
            it had been a source for building materials by local builders, and 
            the stone was also being burned to produce lime. The site was 
            cleared completely in 1880; an enormous number of loose sculptures 
            found their way over the years primarily either to England (and 
            eventually to the British Museum) or to the Madras Museum. There is 
            also considerable material today at the site museum. But with the 
            site's destruction we will never have a clear view of what was there 
            and how the sculpture was used, and much of what scholars must do, 
            as Knox attempts, is to reconstruct the monuments from what is left. 
            
            This is essentially what he does in the two chapters (III and IV) on 
            architecture and sculpture. His reconstruction of the form of the 
            Great Stupa basically follows Barrett's of 1954, and he illustrates 
            the reconstruction in drawings that follow those in the earlier 
            study. The stupa was axial, with a massive circular fence whose four 
            entrances brought the worshiper directly to the four projections of 
            the stupa's base on which five pillars were erected. These pillars 
            and projections, called ayaka-pillars and - platforms, are found 
            almost exclusively (something similar has been found in relief on 
            two votive stupas at Ratnagiri in Orissa) in Andhran stupa 
            architecture. The sculpture is on the fence and gates, and on stone 
            relief slabs that were stacked in rows against the body of the stupa 
            itself. 
            This is not the place to go into the details of the complicated 
            arrangement of the sculpture, but Knox's reconfiguration, as those 
            in the past, is largely speculative. Indeed, the organization he 
            suggests does not fit in many ways with the organization of 
            sculpture depicted on the extensive and highly detailed stupas 
            carved in relief on slabs meant to decorate the actual mahastupa 
            itself (many examples of which are included in the catalogue). The 
            often-repeated notion that these reliefs show the mahastupa raised 
            above the fence, pushed up like a stick of deodorant out of the tube 
            in an artistic convention, so that the relief carving against the 
            stupa that would actually be hidden by the fence can be seen, does 
            not, in fact, explain the organization that scholars, including 
            Knox, have attempted to propose for the loose architectural pieces. 
            Further, which figures are intended as people and which as images on 
            these stupa reliefs is also frequently not clear. None of these 
            issues is mentioned by Knox. 
            But what I feel is the most serious concern with the book is that 
            Buddhism is not introduced in the discussion. Not only is there no 
            attempt to inform the reader about even the most basic facts about 
            Buddhism, how it works, and why people might be Buddhist, but the 
            descriptions of the reliefs name the stories and their characters, 
            all of them Buddhist, but fail to tell the stories and what they 
            mean and why they are important. Knox freely uses Sanskrit terms in 
            his descriptions, but there is no glossary. I cannot imagine how a 
            reader, unless he is already very familiar with Buddhism and 
            Amaravati, will be able to make heads or tails out of these 
            otherwise helpful descriptions. 
            The second book reviewed here deals with a nearby Andhran site (some 
            100 km away as a bird flies and connected by rivers) that is closely 
            related to Amaravati in time and in style. Elizabeth Stone's book, 
            however, is a very different undertaking from that of Knox. Stone 
            has been working on Nagajunakonda since her graduate student days, 
            and it was the topic of both her Master's thesis and her Ph.D. 
            dissertation (written under her maiden name, Rosen). She has 
            published a number of interesting and important articles on the site 
            over the years. We in Indian art history have long been anticipating 
            her book-length study. 
            It turns out that its focus is on the stylistic development of the 
            site's sculpture, and it has some surprises. But before discussing 
            this "evolution," the topic of her second and much the longest 
            chapter, I will say something of the other four. The first chapter 
            introduces the site, its ruling dynasty, and its architecture. 
            Unlike Amaravati, Nagarjunakonda was recognized by scholars only 
            very late, in the 1920s; by the 1960s it was under water, at the 
            bottom of a dammed-up lake. In preparation for the inundation, the 
            site was excavated by the Archaeological Survey of India from 1954 
            to 1960. Unfortunately, only the briefest excavation reports have 
            thus far been published. Stone does not tell us where the official 
            excavation data are and why no report has been published after some 
            forty years. In an odd twist, and apparently according to a wish of 
            Jawaharlal Nehru, some of the monuments, and much of the sculpture, 
            were transported to a hilltop which, with the flooding of the 
            reservoir, became an island, and can today be reached only by boat. 
            Stone feels, as did Knox anent Amaravati, that the building activity 
            at Nagarjundakonda is directly related to dynastic patronage, in 
            this case of the Iksvakus, of which four kings are mentioned in 
            inscriptions, reigning some seventy-five to one hundred years, from 
            about A.D. 225 to the middle of the fourth century. The Iksvakus 
            were perhaps Satavahana feudatories who rose to power with the 
            Satavahanas' downfall. But unlike the Satavahana kings, the Iksvaku 
            kings were Hindus, worshippers of Siva; it was their queens who were 
            the Buddhist patrons. Thus, there are Hindu shrines and images from 
            Nagarjunakonda. Interestingly, these temples are built in an unusual 
            technique, of a stone veneer over a core of brick or rubble, that 
            Stone suggests is Roman in source. While she says that such a 
            building technique is restricted in India to Nadarjunakonda, this is 
            not exactly true, as temple 17 at Sanci is built with this 
            technique, and it brings up the intriguing possibility that this 
            enigmatic little temple, dating to about 400, or only perhaps some 
            one hundred years later than those at Nagarjunakonda, might also 
            have been a Hindu temple. 
            Besides Hindu architecture, there are also secular structures at 
            Nagarjunakonda, and these are unique in India, including a theater 
            or, according to Stone, a boxing stadium. But it is the Buddhist 
            monuments that dominate. And Stone introduces in this first chapter 
            several of her ideas, mostly already presented in her articles, 
            regarding the development of architecture, and more generally of 
            Buddhism, at the site. These include that the site shows, in its 
            monastic architecture, the development from Hinayana to Mahayana; 
            that Mahayana developed largely from lay patronage, which focused on 
            the worship of the stupa and Buddha images (following Akira 
            Hirakawa); and that the different sects at the site can be placed on 
            a scale of innovation (or movement toward Mahayana doctrine) 
            depending on the monks' acceptance or rejection of stupa and image 
            worship. 
            Jumping to chapter III, Stone deals briefly with sculpture from two 
            other Andhran sites, Gummididurru and Goli. She uses sculpture from 
            Western museums, basically unprovenanced material, to show that it 
            can be placed at one or other of the two sites because it is 
            stylistically distinct, different from sculpture from 
            Nagarjunakonda. The sculpture from Gummididurru relates, however, to 
            the early sculpture at Nagarjunakonda, while that from Goli is 
            later, perhaps post-Iksvaku (first half of the fourth century). The 
            Goli sculpture, Stone feels, particularly relates to, in fact 
            influenced (is "evidence of the transmission of the Andhra Style"), 
            the later art of north India of the GuptaVakataka era. 
            Chapter IV is very short, but takes the transmission of Andhra style 
            even further geographically, to Afghanistan, in a discussion of the 
            Begram ivories. Stone dates the ivories, also a topic of a 
            previously published article, to the third or fourth centuries. She 
            suggests that, while made in the northwest, they have close 
            stylistic and iconographic parallels in the stone sculpture of 
            Nagarjunakonda. Similar ivories must, she feels, have once actually 
            existed in Andhra, and they reveal, as do the stones, an intense yet 
            subtle awareness of classical Western art. The point here is that 
            these styles and influences are pan-Indian, reflecting a surprising 
            interchange of peoples and goods, all feeding into, Stone thinks, 
            the style that develops in the Gupta period. The final chapter (V) 
            is a brief, two-page conclusion. 
            Now I want to return to the heart of the study, chapter II, "The 
            Evolution of the Nagarjunakonda Style." Stone "will demonstrate in 
            this chapter (that) the stylistic changes in the art of 
            Nagarjunakonda are allied to the development of the architectural 
            ground-plans of the site." While this may not sound unusual, it has 
            never, as far as I am aware, been argued for any other site in 
            India. What Stone means, and goes on to demonstrate, is that each 
            monastic or stupa site at Nagarjunakonda had its own style. This 
            means, for example, that the two contemporaneous and contiguous 
            sites 2 and 3 had two different styles. That is, the artists of one 
            monument, working at the same time but a few meters from artists at 
            another monument, were working in two different styles. This differs 
            from the rest of India, where artistic styles are broadly 
            geographical and chronological, but monuments and art from the same 
            site share the same style, and do not even show sectarian stylistic 
            differences. 
            Another surprise is the relationship of the Nagarjunakonda art to 
            that at Amaravati. Stone's dating for Amaravati is roughly the same 
            as that of Knox; it was around A.D. 225 that, with the downfall of 
            the Satavahanas, the Iksvakus' first king Camtamula I came into 
            power at Nagarjunakonda. But the art she assigns to his reign is 
            not, as might be expected, a continuation of the late styles at 
            Amaravati. While some of the art does reveal an Amaravati 
            connection, much of it shows influence from a north Indian site, 
            Mathura. It is not, according to Stone, until the reign of the third 
            Iksvaku king, Ehuvala Camtamula (ca. 265-75 to ca. 290-300 A.D.), 
            that Amaravati influence becomes more powerful. The Nagarjunakonda 
            artists then changed the Amaravati art, moving it toward lower 
            relief, more decorative style, and more formalized compositions. 
            Stone makes it clear that she feels this is a decline in artistic 
            quality from that of Amaravati - a somewhat distressing tendency 
            toward personal judgment shared, to an even greater extent, by Knox. 
            
            Both books deserve to be widely read, and contain many important 
            points that I have not been able to bring up here. Nevertheless, the 
            sculpture and architecture from Amaravati and Nagarjunakonda is so 
            rich, varied, and significant, and continues to present so many 
            important questions, that scholars should be encouraged to continue 
            research on this wonderful material. 
            1 Douglas Barrett, Sculptures from Amaravati in the British Museum 
            (London: British Museum, 1954). 
            2 I. K. Sarma, "Early Sculptures and Epigraphs from South-East 
            India: New Evidence from Amaravati," in Indian Epigraphy: Its 
            Bearing on the History of Art, ed. Frederick M. Asher and G. S. Gai 
            (New Delhi: Oxford, 1985): 15-23. 
            3 See F. R. Allchin, The Archaeology of Early Historic South Asia: 
            The Emergence of Cities and States (Cambridge: Cambridge Univ. 
            Press, 1995), 176-79. 
            ROBERT L. BROWN UNIVERSITY OF CALIFORNIA, LOS ANGELES