The Indus Valley Civilisation: what we know and what we still do not

Harappa, in what is now Pakistani Punjab, had long been known as a mound from which bricks were taken. Systematic excavation there began in 1921 under Daya Ram Sahni of the Archaeological Survey of India, and in 1922 R. D. Banerji began digging at Mohenjo-daro in Sindh. In 1924 the Survey's director, John Marshall, announced to the world that these sites belonged to a previously unknown civilisation, and the news was a sensation.

The mature phase of the civilisation is usually dated from about 2600 to 1900 BCE. Well over a thousand sites have been found, spread from Baluchistan and Sindh in the west to Haryana, Punjab, Rajasthan and Gujarat in India, and north to Afghanistan. Among the largest are Mohenjo-daro and Harappa in Pakistan, and Rakhigarhi in Haryana and Dholavira in Gujarat in India.

The cities were laid out on grids, with main streets crossing at right angles. Houses were built of baked brick, often with a bathing room, and wastewater ran into covered drains along the streets, a standard of sanitation that was remarkable for its time. Bricks were made to a common proportion across hundreds of kilometres, and carefully made stone weights follow a standard system, which suggests a shared set of rules for trade.

At Mohenjo-daro, a large brick-lined tank known as the Great Bath is one of the best-known structures, though archaeologists disagree about what it was used for. At Dholavira, which UNESCO made a World Heritage Site in 2021, a chain of reservoirs and channels shows how carefully water was managed in a dry land. A structure at Lothal in Gujarat has often been described as a dock, a reading that some scholars question.

The people grew wheat, barley, peas and sesame, and cotton for cloth, and kept cattle. Craftworkers made beads of carnelian and other stones, copper and bronze tools, and fine pottery. Small square seals, usually of steatite, carry an animal such as a unicorn-like bull and a line of signs, and were probably used to mark goods. Seals and beads from the Indus have been found in Mesopotamia, where texts speak of trade with a land called Meluhha that is usually identified with the Indus region.

The signs on seals and other objects form the Indus script, which has about 400 distinct signs. The inscriptions are very short, most of them only a handful of signs, and no bilingual text, like the Rosetta Stone for Egyptian, has been found. Scholars do not even agree on whether the signs write a language, or serve as a system of symbols. Claims to have deciphered the script appear every few years and none has been accepted by the field, so it is wise to treat such headlines with caution.

Unlike Egypt or Mesopotamia, the Indus cities have no clear palaces, royal tombs or monumental temples, and there are few signs of warfare. Some researchers conclude that power was shared among merchants, kin groups or councils, others that evidence of the rulers has yet to be recognised. We do not know what the people called themselves, what language they spoke or what they believed.

From about 1900 BCE the great cities were gradually abandoned, and people moved towards the east and south. There is no evidence of a single catastrophe. Hypotheses include a weakening of the monsoon and a drier climate, changes in the courses of rivers, and a decline in trade with Mesopotamia, and many scholars think that several causes worked together. New evidence from sediments, ancient climate records and DNA keeps adding to the picture, which is why news about the civilisation is frequent and why many of its stories change.

• How the Indian monsoon works, and why forecasts matter

  The south-west monsoon brings most of India's yearly rain. What drives it, how it moves across the country, and how to read the headlines about it.

• How science works: hypotheses, evidence and peer review

  A short guide to how scientific claims are tested and checked, and why a headline that says a study shows something deserves a closer look.