Tombs, Landscape and Society in Southern Madagascar - Part VI

Mike Parker Pearson

Excavations in Androy, Southern Madagascar 2003

Over the last twelve years the Androy project has been investigating the long-term archaeology and history of southern Madagascar, from its earliest beginnings 2000 years ago until the present day. Five main themes have emerged during this research:

  • the origins of Madagascar's occupation and inhabitation,
  • the extinction of the Aepyornis (the Elephant Bird, largest of the flightless ratites),
  • the rise and fall of the 10th-13th century Manda civilisation,
  • the European impact, specifically the story of Robert Drury who was enslaved here in the early 18th century,
  • the origin and development of megalithic and monumental tombs from the 18th century to the present day.

The story of our research between 1991 and 2000 is told in In Search of the Red Slave: shipwreck and captivity in Madagascar written by Mike Parker Pearson and Karen Godden, and published in 2002 by Sutton.


In 2003 we concentrated on excavating four settlement sites. One of these is Andaro, a proto-urban settlement enclosed within two stone walls on either side of a river valley. This is one of the largest sites left by the mysterious Manda civilisation (manda means enclosure) which collapsed around AD 1300. In the next two centuries population numbers fell dramatically and most of the settlements were located for defence. We excavated one of the largest of these, a clifftop village at Ampahijolike. In 2000 we excavated a royal village from the period after this, the 16th and 17th centuries. This was the village of Montefeno, where French troops visited king Andriamififarivo in 1649. Returning in 2003, we excavated another two royal villages nearby at Anjankitike and Mionjona. These are probably the two 'towns' where Robert Drury was kept as a slave between 1703 and 1709.


The earliest site that we excavated in 2003 is one of the largest in southern Madagascar - a pair of stone-walled enclosures covering 21 hectares. From surface surveys within both enclosures we were able to plot the distribution of Islamic and Chinese imported ceramics and glass (115 findspots) together with a few beads and other trade goods amongst the very high density of locally made pottery. Both enclosures had substantial residential areas whilst the larger enclosure had an open space in its interior, possible for penning cattle. Excavation of the stone wall of the large enclosure revealed that it was built as a single phase but well into the period of occupation, replacing a palisade of closely set timbers. Two areas within the interior were excavated, covering an area of 140sq m and producing evidence of rectangular wooden houses slightly larger than those inhabited between the 14th century and today. These house forms indicate the appearance of a specifically Malagasy architecture by the 10th-11th centuries even though it lay within the ambit of the Arab-Swahili world.



This is one of the largest settlements of the 14th-15th centuries in southern Madagascar and is defended on three sides by river cliffs. We were unable to detect any defences on the fourth side because this field was under cultivation and could be fieldwalked but not excavated. Nonetheless, a field survey identified this northern edge and revealed dense sherd scatters which probably indicate three east-west rows of houses. There were also dense concentrations of iron slag and animal bones. On the edge of the field we were able to excavate one of these houses, whose layout and dimensions are identical to those of the 20th century. Further indications of the emergence of today's cattle pastoralist lifestyle around this time come from the animal bones which included substantial parts of a young cow deposited as an offering within the abandoned house. In contrast to the hundreds of imports from Andaro, this site yielded just five sherds of Chinese Longquan celadon and two shards of probably Islamic glass, indicating a steep decline in imported goods.



This site was partially excavated in 2000 and we returned to complete this work. Most of this 18th-19th century settlement site was already destroyed by ploughing but we were able to reconstruct its layout from the scatter of sherds, animal bones and small finds on the surface of the ploughed fields. With a cattle penning area in the centre, rows of houses were arranged east-west in rows on either side. From the restricted distribution of metal artefacts and graphite-burnished sherds we could identify the area of the royal residence in the southeast corner whilst the one zone with no special finds at all, in the west end, probably marks the slaves' quarters. European imports included musket balls, blue-on-white factory-made wares and metal ornaments and tools.


Located within a kilometre of Mionjona, this is probably the first of the two royal villages that Robert Drury lived in as a slave between 1703 and 1709 (the other being Mionjona). We found burnt posts across the site which tally with his account of its being sacked in an intra-dynastic feud. The village was smaller than Mionjona but its layout was almost identical, with the royal residence marked by a concentration of gunflints, graphite-burnished ware and stone tobacco pipes, and the slaves' quarters in the west by a low ratio of bowls to cooking pots. Along the south side we identified the presence of a diviner/healer's house from incense burners and also excavated a house which preserved the pattern of its interior activities from micro-artefacts within the house floor - a pattern identical to contemporary uses.

Mike Parker Pearson, Karen Godden, Georges Heurtebize, Ramilisonina, Retsihisatse, Jean-Luc Schwenninger, Helen Smith, Zoe Crossland and Brian Boyd